jmp.

com/learn

JMP One-Page Guides

Updated Nov 2022

These One-Page Guides are developed by the

JMP Academic Team. If you have any suggestions or
feedback on how these can be improved, please
contact us at: [email protected].

Please visit our academic teaching and learning

resources including video tutorials, webinars, case
studies, course materials, and more at jmp.com/teach.

Copyright © 2022 JMP Statistical Discovery LLC.

All rights reserved. These guides may be distributed
freely with attribution.
 Table of Contents

1 Analysis Guides 4 Probabilities and Distributions

1.1 Analyzing a Single Variable ....................... 1 4.1 Assessing Normality ................................ 35
1.2 Relationship between a Single Factor 4.2 Fitting Distributions .................................. 36
and a Response Variable ............................ 2 4.3 Finding Standardized Values (z- Scores).... 37
1.3 Relationship between Many Factors 4.4 Finding the Area Under a Normal Curve ... 38
and a Response Variable ............................ 3 4.5 Random Sampling and Random Data ...... 39
1.4 Exploring Patterns and Features ................. 4
1.5 Collecting Data .......................................... 5 5 Basic Inference – Proportions and Means
5.1 Hypothesis Test and Confidence
2 JMP Basics
Intervals for Proportions ......................... 40
2.1 Opening JMP and Getting Started ............. 7 5.2 Two Proportions Test and Confidence
2.2 JMP Data Tables ........................................ 8 Interval ..................................................... 41
2.3 Navigating JMP in Windows ....................... 9 5.3 Chi-Square Tests for a Two-Way Table ... 42
2.4 Importing Data from Excel ...................... 10 5.4 Sample Size and Power for
2.5 Importing Text Files ................................. 11 One Sample Proportion............................. 43
2.6 JMP Tools ................................................ 12 5.5 Sample Size and Power for
2.7 JMP Tables Menu .................................... 13 Two Proportions ....................................... 44
2.8 Saving JMP Results .................................. 14 5.6 One Sample t-test and Confidence
2.9 Publication-Quality Graphics – Mac ......... 15 Interval .................................................... 45
2.10 Publication-Quality Graphics – Windows 16 5.7 Two Sample t-Test and Confidence
2.11 Sharing Results on JMP Public ................. 17 Interval ...................................................... 46
2.12 Creating Formulas in JMP ....................... 18 5.8 Paired t-Test and Confidence Interval....... 47
2.13 Excel Add-in I (Passing Data to JMP) ........ 19 5.9 One Way ANOVA....................................... 48
2.14 Excel Add-in II (Profiling Excel Models 5.10 Two-Way (Factorial) ANOVA ................... 49
in JMP) ..................................................... 20 5.11 Nonparametric Tests ................................. 50
2.15 Add-in Builder .......................................... 21 5.12 Sample Size and Power for
2.16 JMP Teaching Tools .................................. 22 One Sample Mean ................................... 51
5.13 Sample Size and Power for
3 Graphical Displays and Summaries Two Sample Means ................................. 52
3.1 Bar Charts and Frequency Tables ............ 23 5.14 Bootstrapping ........................................... 53
3.2 Pareto Plots and Pie Charts ..................... 24 5.15 Randomization Testing ............................ 54
3.3 Mosaic Plot and Contingency Table 5.16 Prediction Interval ................................... 55
(Cross-tabs) ............................................. 25 5.17 One Sample Equivalence Test for Mean ... 56
3.4 Histograms, Descriptive Statistics,
and Stem and Leaf Plot ............................ 26 6 Correlation and Regression
3.5 Dot Plots .................................................. 27 6.1 Correlation ............................................... 57
3.6 Box Plots ................................................. 28 6.2 Nonparametric Correlations .................... 58
3.7 Scatterplots .............................................. 29 6.3 Simple Linear Regression ........................ 59
3.8 Run Charts (Line Graphs) ......................... 30 6.4 Multiple Linear Regression........................ 60
3.9 Interactive Graphing with Graph Builder 31 6.5 Simple Logistic Regression ....................... 61
3.10 Bubble Plots ............................................ 32 6.6 Multiple Logistic Regression .................... 62
3.11 Summarizing Data Using Tabulate .......... 33 6.7 Stepwise Regression ................................. 63
3.12 Mapping in Graph Builder ....................... 34 6.8 Fitting Nonlinear Curves........................... 64
7 Time Series
7.1 ARIMA Modeling ..................................... 65 11.4 Attribute Control Charts -
7.2 Time Series Smoothing Methods ............. 66 C and U Charts.......................................... 90
7.3 Time Series Forecasting .......................... 67 11.5 Tolerance Intervals................................... 91
11.6 Process Capability Analysis ..................... 92
8 Multivariate Methods 11.6 Capability Analysis for
Multiple Responses .................................. 93
8.1 Clustering ................................................ 68
11.7 MSA Continuous Data - Gauge R&R ......... 94
8.2 Principal Components Analysis ................ 69
11.8 MSA Continuous Data - EMP Method ...... 95
8.3 Factor Analysis ........................................ 70
11.9 MSA Attribute Data .................................. 96
9 Mixed Models and Repeated Measures
9.1 Repeated Measures Analysis (MANOVA) 71 12 Reliability and Survival
9.2 Repeated Measures Analysis 12.1 Distribution Fitting (Life Distribution) ...... 97
(Mixed Model) .......................................... 72 12.2 Accelerated Life Testing (Fit Life by X) ..... 98
9.3 Mixed Model Analysis ............................. 73
13 Design and Analysis of Experiments
10 Data Mining and Predictive Modeling 13.1 DOE Full Factorial Design ......................... 99
10.1 Classification Trees (Partition) ................ 74 13.2 DOE Full Factorial Analysis ..................... 100
10.2 Regression Trees (Partition) ..................... 75 13.3 DOE Fractional Factorial Design ............. 101
10.3 Discriminant Analysis .............................. 76 13.4 DOE Fractional Factorial Analysis ........... 102
10.4 Support Vector Machines (Classification) 77 13.5 DOE Screening Experiment Analysis ...... 103
10.5 Support Vector Regression ...................... 78 13.6 DOE - Custom Designs ............................ 104
10.6 K Nearest Neighbors ............................... 79 13.7 DOE - Evaluate and Compare Designs .... 105
10.7 Neural Networks ...................................... 80 13.8 Monte Carlo Simulation ......................... 106
10.8 Naive Bayes .............................................. 81
10.9 Creating a Validation Column 14 Using SAS from JMP
(Holdout Sample) ..................................... 82 14.1 Connect to SAS OnDemand for
10.10 Model Comparison and Selection ............ 83 Academics from JMP ............................. 107
10.11 Text Mining - Describing Unstructured 14.2 Connect to Your School’s SAS Server
Text Data .................................................. 84 from JMP ................................................ 108
10.12 Text Mining - Analyzing Unstructured 14.3 Connect to SAS on Your PC (Windows)
Text Data .................................................. 85 from JMP ............................................... 109
10.13 Association Analysis 14.4 Setting SAS Integration Preferences ...... 110
(Market Basket Analysis)........................... 86 14.5 Using JMP to Generate SAS Programs .. 111
14.6 Entering and Running SAS Programs...... 112
11 Quality and Process
11.1 Variables Control Charts – I and
MR Charts ................................................ 87
11.2 Variables Control Charts - XBar and
R or S Charts ............................................ 88
11.3 Attribute Control Charts -
P and NP Charts ....................................... 89
 Analysis Guides
jmp.com/learn Nov 2022

Analyzing a Single Variable

This Analysis Guide provides a list of potential questions that one might consider when analyzing a single
variable or column of data. There are many statistical techniques useful in addressing these questions which
are listed in the second column. The JMP® platforms that provide access to those statistical techniques along
with One Page Step-by-Step Guides included in this manual are highlighted in the third column.

STATISTICAL ONE PAGE

QUESTIONS JMP PLATFORMS
TECHNIQUES GUIDES

Tables > Summary p. 12

Where is the data located? Mean ; Median ;
Percentiles
How much variation is there in Analyze > Tabulate p. 33
the data? Standard Deviation ;
Range ; Interquartile p. 23, 26, 28, 35,
What is the overall shape of Analyze > Distribution
Range 36, 40, 43
the data?
Proportions ; Counts Graph > Graph Builder p. 27, 30, 31
Are there any extreme values
or outliers? Histogram
Analyze > Quality and Process >
p. 24
How does the data vary over Boxplot Pareto Chart
time?
Dotplot Analyze > Quality and Process >
How much of the data is p. 92
Process Capability
beyond (or within) specific Line Chart
values? Bar Chart Analyze > Quality and Process >
p. 87, 88
Control Chart Builder
What percentage of the data Pareto Plot
is there in each value of a Analyze > Specialized Modeling >
categorical variable? Pie Chart p. 64
Fit Curve
Does the data support a Confidence Interval
Analyze > Specialized Modeling >
specific hypothesis about the (e.g., for Mean,
Time Series p. 65, 66, 67
underlying population the Standard Deviation)
data is sampled from? Analyze > Time Series Forecast
Hypothesis Test
(e.g., t-Test for Mean) Analyze > Reliability and Survival >
p. 97
Life Distribution
Process Capability

Probability
Distribution Modeling

Control Charts

Time Series

1
 Analysis Guides
jmp.com/learn Nov 2022

Relationship between a Single Factor and a Response Variable

This Analysis Guide provides a list of potential questions that one might consider when studying the
relationship between a single explanatory factor and a response variable. There are many statistical
techniques useful in addressing these questions which are listed in the second column. The JMP® platforms
that provide access to those statistical techniques along with One Page Step-by-Step Guides included in this
manual are highlighted in the third column.

STATISTICAL ONE PAGE

QUESTIONS JMP PLATFORMS
TECHNIQUES GUIDES

p. 25, 29, 41, 46,

Is there a relationship between Correlation Analyze > Fit Y by X 48, 49, 57, 59,
the explanatory factor and a 61
response variable? Scatterplot

Linear Regression Graph > Graph Builder p. 27, 30, 31

How best to graphically and
numerically describe/summarize Nonlinear Regression Analyze > Specialized Modeling >
the relationship? p. 64
Fit Curve
Logistic Regression
How does the response
variable change as the ANOVA Analyze > Specialized Modeling >
p. 65, 66, 67
explanatory factor changes? Time Series
Contingency Table
How strong is the relationship? Analyze > Reliability and Survival >
p. 98
Fit Life by X
How much variation in the
response variable is explained
by the relationship with the
explanatory factor?

Are there differences in the

response variable for different
levels of a categorical
explanatory factor?

How well does a statistical

model fit to the data predict an
outcome?

Are there any outliers or

unusual observations?

2
 Analysis Guides
jmp.com/learn Nov 2022

Relationship between Many Factors and a Response Variable

This Analysis Guide provides a list of potential questions that one might consider when studying the
relationship between many potential explanatory factors and a response variable. There are many statistical
techniques useful in addressing these questions which are listed in the second column. The JMP® platforms
that provide access to those statistical techniques along with One Page Step-by-Step Guides included in this
manual are highlighted in the third column.

STATISTICAL ONE PAGE

QUESTIONS JMP PLATFORMS
TECHNIQUES GUIDES

Graph > Graph Builder p. 27, 30, 31

Are there any relationships Correlation
between the explanatory
factors and a response Scatterplot Graph > Scatterplot Matrix p. 29
variable? Multiple Linear
Analyze > Multivariate Methods >
Which factors have the Regression p. 57, 58
Multivariate
strongest/weakest relationships Multiple Logistic
with a response variable? Regression p. 49, 60, 62, 71,
Analyze > Fit Model
72, 73, 100, 102
How best to graphically and ANOVA
numerically describe and
DOE > Definitive Screening >
summarize those relationships? Contingency Table p. 105
Fit Definitive Screening
How strong are those Mixed Models
relationships? Analyze > Multivariate Methods >
Repeated Measures p. 76
Discriminant
Does the effect that factors
Discriminant Analysis
have on a response change Analyze > Predictive Modeling >
for different levels of other p. 74, 75
Classifications Trees Partition
factors?
Regression Trees Analyze > Predictive Modeling >
How much variation in the p. 81
Naïve Bayes
response variable is explained K Nearest Neighbors
by the relationships with the Analyze > Predictive Modeling >
Naïve Bayes
explanatory factors? K Nearest Neighbors p. 79
Neural Network
What are ways to analyze data
from an experiment? Analyze > Predictive Modeling >
p. 80
Neural Network
How well does a statistical
model fit to the data predict Analyze > Predictive Modeling >
p. 77, 78
an outcome? Support Vector Machines

How do different models fit to

Analyze > Predictive Modeling >
the data compare? p. 82
Make Validation Column
Are there any outliers or
unusual observations? Analyze > Predictive Modeling >
p. 83
Model Comparison

3
 Analysis Guides
jmp.com/learn Nov 2022

Exploring Patterns and Features

This Analysis Guide provides a list of potential questions that one might consider when exploring data for the
purpose of finding features, patterns, and overall structure in the data without the need to model the
relationship between specific explanatory factors and a specific response variable. There are many statistical
techniques useful in addressing these questions which are listed in the second column. The JMP® platforms
that provide access to those statistical techniques along with One Page Step-by-Step Guides included in this
manual are highlighted in the third column.

STATISTICAL
QUESTIONS JMP PLATFORMS ONE PAGE GUIDES
TECHNIQUES

Graph > Graph Builder p. 27, 30, 31, 34

Are any of the variables Exploratory Data
correlated with other variables? Analysis
Graph > Bubble Plot p. 32
What are ways to visualize and Correlation
describe the overall features
Scatterplot Graph > Scatterplot Matrix p. 29
and patterns in the data?

Are there any outliers or Line Graph Analyze > Multivariate Methods >
p. 57, 58
unusual observations? Multivariate
Heat Map
Can the observations be Bubble Plot Analyze > Clustering p. 68
grouped into clusters?
Cluster Observations Analyze > Multivariate Methods >
Can the data be simplified to a p. 69
fewer number of variables? Principal Component Principal Components
Analysis
Analyze > Consumer Research >
p. 70
Factor Analysis Factor Analysis

4
 Analysis Guides
jmp.com/learn Nov 2022

Collecting Data
This Analysis Guide provides a list of potential questions that one might consider when trying to determine
how much data to collect. For example – in designing an experiment. There are many statistical techniques
useful in addressing these questions which are listed in the second column. The JMP® platforms that provide
access to those statistical techniques along with One Page Step-by-Step Guides included in this manual are
highlighted in the third column.

STATISTICAL
QUESTIONS JMP PLATFORMS ONE PAGE GUIDES
TECHNIQUES

How much data should be Power and Sample Size DOE > Sample Size Explorer p. 43, 44, 51, 52
collected in order to test a Determination
statistical hypothesis of
Design of Experiments DOE > Classical > Full Factorial p. 99
interest?

How many factors and how DOE > Classical > Two Level
p. 101
many different levels of those Screening > Screening Design
factors should be studied?
DOE > Custom Design p. 104
What combination of the
experimental factors and at
DOE > Design Diagnostics >
what levels should be studied
Evaluate Design
in an experiment? p. 105
DOE > Design Diagnostics >
How many replicates of Compare Design
experimental runs should be
used?

What experimental design is

best to answer specific
questions of interest?

How best to design a

sequence of experiments?

What experiment is best to

use in order to screen a large
number of factors to find the
most important ones?

What type of experiment is

best at determining factor
settings to optimize an
outcome?

5
 JMP Basics
jmp.com/learn Nov 2022

Opening JMP® and Getting Started

This guide provides instructions on creating a new data table, opening data tables, and finding help within JMP.

Opening JMP
1. When you first open JMP, you’ll see the Tip of the Day and the JMP
Home windows. On Mac you’ll also see the JMP Starter window.
o Tip of the Day gives helpful hints on using JMP.
o The JMP Home Window displays recently used files and open data
tables and windows. See the page “Navigating JMP in Windows”
for Windows-specific information.
o The JMP Starter window (click on View > JMP Starter on Windows,
including opening files and accessing JMP analyses.
2. The JMP menus, across the top, can be used to perform JMP functions.
3. The JMP toolbar, located beneath the menu bar, provides many
shortcuts and helpful tools.

Creating a New JMP Data Table

Click on File > New > Data Table or select New Data Table from the JMP Starter.

Opening a JMP Data Table

1. Click on File > Open, or click on Open from the JMP Starter
window. Navigate to the directory where your JMP, Excel or
other files are stored.
2. Click on the file name and click Open.

Getting Help
The Help menu provides many resources to help you get started:

• Searchable documentation (JMP Help and Books),

• A summary of new features,
• Tutorials,
• Sample Data,
• Indexes of statistical terms and JMP scripting functions, and more.

7
 JMP Basics
jmp.com/learn Nov 2022

JMP® Data Tables

This page gives information on the structure, components and information contained in JMP data tables.

Click on File > Open to open a JMP data table. To open a file from the JMP Sample Data directory, go to Help >
Sample Data Folder. A portion of the Sample Data file Companies.jmp is shown.

File Name
Column Name
Table
Tables Panel
Panel
Table
Saved Panel
Script Header Graph
* Click small histogram
Table Panel
Red Triangle icon in upper left of data
table to display/hide
header graphs
Columns Panel

Rows Panel

Click in this region in the upper left of the Click in this region in the upper left of the
data table to “unselect” rows. data table to “unselect” columns.

Tables Panels
• The data table name - Select ‘New Table Variable’ under the red triangle to create a note
• A list of table properties, table variables, - Click on green “play” buttons in a saved script in Table Panel to
and scripts rerun an analysis.

Columns Panel - This data table has eight columns, and one column has been
selected.
• The number of columns - Header Graphs are displayed
• The number of selected columns - Two columns are nominal (red bars).
• Column names - Six columns are continuous (blue triangles).
• The modeling type for each column - There are no ordinal columns (green bars) or unstructured text
columns (paragraph icon).
• Column properties
- Three columns have stored formulas (the plus sign after the
column name – click on the plus sign to display the formula).
Rows Panel
• Rows - This data table has 32 rows, or observations.
• Selected rows - One row has been selected (row 4).
- Three rows have been both excluded and hidden (rows 1-3).
• Hidden rows (with a mask)
- Hidden rows will not display on graphs.
• Excluded rows (with a “don’t” sign) - Excluded rows will not be included in most future analyses.
• Labeled rows (with a tag) - One row has been labeled (row 5).

Notes: Red triangles are used throughout JMP to access other commands, and gray triangles are used to
minimize display areas. Right-click in different regions of the data table (or graphs) for additional options.
Visit “Using JMP > Data Table Features” in JMP Help to learn more.

8
 JMP Basics
jmp.com/learn Nov 2022

Navigating JMP® in Windows

This guide provides information on the Windows JMP interface. For information on creating a new data table,
opening data tables, and finding help within JMP see the page Opening JMP and Getting Started.

The Home Window

• When you first open JMP, you’ll see the Tip of the Day
window.
• You’ll also see the JMP Home Window (pictured to the right),
which includes:
o Menus and toolbars (top).
o Recently used files (on the left).
o All open data tables and windows (on the right). Hover
over an item in the list for a preview.
• The JMP Starter, Home Window and other features can be
accessed via the View menu.

Navigation
• Data tables display thumbnails of open reports (bottom left).
Hover over a thumbnail for a preview.
• Each data table and report provides icons (bottom right corner) to
facilitate navigation between windows:
o Return to the Home Window (or, click CNTL – 1).
o Go to the data table.
o Select to arrange with other windows.

Tips:

• Each window has a JMP menu and toolbar.

• To view a hidden menu click Alt, or hover on the report
window where the menu would normally appear.
• To have menus and toolbars always display, go to File >
Preferences > Windows Specific and change Auto-hide
menus and toolbars from Based on window size to
Never.
• The JMP Home Window and Tip of the Day appear by
default when JMP is opened. To change the default
windows, use File > Preferences > General.

• The Preference Window enables you to specify

general and specific settings on JMP will perform
operations, styles and formats for numerical and
graphical output, among many other properties.

9
 JMP Basics
jmp.com/learn Nov 2022

Importing Data from Excel

This guide provides instructions on how to import data into JMP® from Excel.

Opening Excel Files Directly in JMP

1. Select File > Open.
2. Navigate to the directory where your files are stored.
3. JMP will display all files that JMP can import – unrecognized files
will be grayed out. In Windows, to see a complete list of file types
JMP can read, click on the arrow next to All JMP Files.
4. Select the Excel file you’d like to import. Default
settings for Windows (top) and Mac (bottom) are
shown. On Mac, you will need to click the Options
button at the bottom left to display these.
5. Click Open, or use the Excel Wizard to easily
customize the import. To use the Excel Wizard in
Windows, check the arrow next to Open and select
Use Excel Wizard. On Mac, keep the default Use Excel
Wizard as the Excel Open Method.
6. Importing data using the Excel Wizard:
• JMP will detect the file’s structure and display the appropriate
Text Import window. Edit this information using the
interactive dialog.
• Select Import at the bottom of the window when you’re
ready to import the data into JMP.
Note: If you have more than one worksheet in your Excel file, JMP
will import each sheet as a separate data table.

Copying and Pasting from Excel into JMP

1. In Excel, copy the cells you’d like to import, including column names (one row).
2. In JMP, click on File > New > Data Table.
3. Select Edit > Paste or Edit > Paste with Column Names to paste the data into the
new data table. Selecting the latter will allow you to paste column names from
your Excel file.

Tips:

• Always check to make sure that the Excel file was imported correctly.
• In particular, check that JMP has assigned the correct modeling types. By default,
numeric columns will be Continuous (blue triangles) and alphanumeric (text)
columns will be Nominal (red bars).
• To change the modeling type, click on the icon in front of the column name in the
Columns panel, or change the modeling type in the Column Info window.

Visit “Using JMP > Import Your Data” in JMP Help to learn more.

10
 JMP Basics
jmp.com/learn Nov 2022

Importing Text Files

This guide provides instructions on how to import text files (.txt, .csv, .dat, or .tsv) into JMP. Windows and Mac
differences are noted.

Importing Data from Text Files

1. Select File > Open.
2. Navigate to the directory where your files are stored.
3. JMP will display all files recognized – unrecognized files will
be grayed out.
4. Select the text file you’d like to import. In Windows, a
preview of the file can be displayed during importing.
5. At the bottom of the window choose one of the four import
methods next to Open as, and click Open.

Import Methods:

• Data, using Text Import preferences, or in Mac, Data (Using

Preferences): JMP uses the import rules set in Preferences
to open the file. To see or change your preference settings,
go to File > Preferences > Text Data Files, or in Mac, JMP >
Preferences > Text Data Files.
• Data, using best guess, or in Mac, Data (Best Guess): JMP
uses its best guess to import the text file.
• Data with Preview, or in Mac, Data (Using Preview): JMP
displays a window with the data and options for reading in
the data. Additional details are provided below.
• Plain text into Script window, or in Mac, Text: JMP opens
the file in a text-editing window.

Importing data using Preview:

• JMP will detect the file’s structure and display the appropriate
Text Import window. Click the Delimited Fields or Fixed
Width Fields radio button to change the import type.
• Select Next to see how your data will look in JMP. From here,
you can:
o Click on a column name to change it.
o Click on the icon next to each column name to toggle
between Numeric, Character and Exclude.
o Select a format for numeric columns by clicking on the red
triangle next to the column name.
• Select Import at the bottom of the window when you’re ready
to import the data into JMP.

Visit “Using JMP > Import Your Data” in JMP Help to learn more.

11
 JMP Basics
jmp.com/learn Nov 2022

JMP Tools
The Tools menu provides tools to help you interact with JMP output and data. Note: In Mac, tool bar appears at
the top of all reports by deafult. In Windows, the menus and the toolbar are hidden by default. Click the Alt key
to view. The tools can also be accessed from the tools menu.

Windows

Mac
Using the Tools
The default tool is the Arrow, or Cursor. To change tools, click on the icon from the toolbar or select from the
Tools Menu. Your curser will change from the arrow to the shape of the tool selected, and the active tool will be
highlighted in the toolbar.

Description of Tools
Arrow – Allows you to select points in plots, choose analysis options, and more. Remember to select the
arrow after you’ve finished using other tools.
Help – The help tool or question mark accesses the JMP help system. Select the help tool, then click on an
area of a data table or report on which you need assistance. Context-sensitive help tells about the items
located near the location of your click.

Selection – The selection tool, or fat plus sign. Click on the selection tool, then select rows or regions of
rows by clicking and dragging, or select portions of graphical output or analysis results to copy and paste.

Scroller – The scroller tool is for scrolling reports up or down to show only the results you want to see.

Grabber – The grabber or hand tool is for manipulating plots, axes and formula components. Click and drag
to rescale or change the range of an axis, or click and drag in a histogram to change the bin size.

Brush – The brush tool is for highlighting a rectangular area of points in a plot. Alt-click (Option-click on
Mac) to change the size of the rectangle or to extend the selection.

Lasso – The lasso tool lets you highlight an irregular area of points in a plot. Drag the lasso around any set
of points to select them.

Magnifier – The magnifier tool is for zooming in on an area in a plot. The area you click on becomes the
center of a new view. Alt-click (Option-click on Mac) to restore the original plot.

Crosshairs – The crosshair tool creates a movable set of axes on a graph useful to display the x,y coordinate
values of a point.

Ruler – Click and drag the ruler tool in any direction to measure the distance between two points on a
graph.

Annotate – The Annotate tool is for adding text to a plot, output or journal.

Line, Polygon, and Simple Shape – The drawing tools draw lines, polygons, and shapes on JMP
reports, journals, and layouts. Right click a line or shape for a menu of options to customize it.

Visit “Using JMP > Customize Menus and Toolbars” in JMP Help to learn more.

12
 JMP Basics
jmp.com/learn Nov 2022

JMP® Tables Menu

This guide provides information on the structure, components and information contained in JMP data tables
along with some tools useful for cleaning and preparing data for analysis.

The Tables Menu Functions

• Summary – Calculates summary statistics for columns in the active data table.
• Subset – Creates a new table that is a subset of the active table.
• Sort – Sorts a table by one or more columns.
• Stack – Stacks separate columns into one new column.
• Split – Splits a column into multiple columns.
• Transpose – Creates a table whose columns were the rows in the original table.
• Concatenate – Combines tables end to end (bottom to top).
• Join – Combines tables side by side.
• Update – Updates a table with values from a second table.
• Missing Data Pattern – Allows you to explore patterns in missing data.
• Compare Data Tables – Identifies differences between two open data tables.

Summarizing Data with the Summary Function

1. From an open JMP data table, select Tables > Summary.
2. Select one or more variables from Select Columns. Then, click on Statistics and select a statistic. Repeat for
each statistic desired. A Preview of the resulting data table will be displayed on the right.
3. To create a row for combinations of grouping variables, select the variable(s) from Select Columns, and click
Group.

Example:
Companies.jmp
(Help > Sample
Data Folder)

4. Click OK to create the summary table. This table is linked to the original table.

• The summary table has three rows, one for each category
of the Group variable (Size Co, in this example).
• N Rows shows the number of rows of each category that
were in the original table.
• Mean (Variable), the statistic we selected, is shown in the
third column - Mean(Sales ($M)).

Notes: Data can also be summarized using Analyze > Distribution and Analyze > Tabulate.

13
 JMP Basics
jmp.com/learn Nov 2022

Saving JMP® Results

This guide provides instructions on copying JMP output to place in other applications and for saving JMP scripts
to the data table. For information on creating publication-quality graphics, see the Publication-Quality Graphics
– Windows or Publication-Quality Graphics – Mac guides.

The Selection Tool – Copy and Paste to Another Program

Example: Car Physical Data.jmp (Help > Sample Data Folder)
1. From any JMP output window, click the selection tool ( )
in the toolbar.
2. Click on the content you’d like to copy – selected content is
highlighted. Click near the edge of the report to select all content.
To extend a selection, hold the shift key.
3. Click Edit > Copy (or Control-C).
4. Open the program where you’d like to paste the content, and
select Paste. To paste as an object, select Paste > Paste Special
and from the list select Picture (Enhanced Metafile).
If using a Mac, select Paste, or use Paste Special > PDF for
high-quality graphics.

NOTE: Use File > Export to save all tables and graphs within a report in a variety of graphical formats, including
PNG, JPEG, SVG, EPS, or sent directly into PowerPoint. Numeric table output sent to PowerPoint will be
imported in as table that is fully editable.

Save Your Work Using Scripts

You can save the steps taken to produce a report as a JSL script (JSL is “JMP Scripting Language”). This enables
you to recreate the report at any time.
1. From any JMP output window, click the top red triangle and select Save Script > To Data Table…
2. Change the script name if desired, and click OK.
3. The saved script will appear in the table panel in the top left corner of the data table.
4. To run the script, click the green triangle to the left of the Script Name. To edit the script, right click
on the script name and select Edit.
5. To save the data table with all the saved scripts, select File > Save.

Visit “Discovering JMP > Save and Share Your Work” and “Using JMP > Save and Share Your Data”
in JMP Help to learn more.

14
 JMP Basics
jmp.com/learn Nov 2022

Publication-Quality Graphics – Mac

This guide provides instructions on saving publication-quality graphics on the Mac version of JMP. For
information on creating publication-quality graphics on the Windows version of JMP, see the Publication Quality
Graphics – Windows guide. For a basic approach on saving output, see the Saving JMP Results guide.

Saving Images
Example: Big Class.jmp (Help > Sample Data Folder)

1. From any JMP output window, click the selection tool ( ) in

the toolbar or use the keyboard shortcut (S).
2. Click on the content you’d like to copy – selected content is
highlighted. Click near the edge of the report to select all
content. To extend a selection, hold the shift key.
3. Press Ctrl and Select File > Save Selection As.
(The macOS Save window will appear).
4. The default file format is .png. Manually change the file
extension to save in another format (e.g., vector format
.svg, or .eps). Save to desired location.

For printed media (presentations, journals, etc.) a vector image, such as .svg (scalable
vector graphic) is recommended. This format preserves transparency, can be edited in image-editing
software, and can be scaled or resized without losing clarity.
Tip: Applications such as Mac’s Preview and Adobe Photoshop provide tools to change properties of an
image such as the resolution and image quality.

sex
Frequencies
Level Count Prob
F 18 0.45000
M 22 0.55000
Total 40 1.00000 Size increased by 300%.
N Missing 0
F M
2 Levels

Editing Image Elements within Power Point Vector graphic being edited in Power Point

1. Save the image in step 4 above as a scalable

vector graphic (.svg).
2. After placing image in document (via. Drag-and-drop
or Insert > Picture > Picture from File), right-click on
the image and select Covert to Shape.
All the elements of the graphic can now be
individually selected, edited, and formatted
as desired.

Visit “Discovering JMP > Save and Share Your Work” and “Using JMP > Save and Share Your Data”
in JMP Help to learn more.

15
 JMP Basics
jmp.com/learn Nov 2022

Publication-Quality Graphics – Windows

This guide provides instructions on saving publication-quality graphics on the Windows version of JMP. For
information on creating publication-quality graphics on the Mac version of JMP, see the Publication-Quality
Graphics – Mac guide. For a basic approach on saving output, see the page Saving JMP Results.

Saving Graphic Images

1. From any JMP output window, click the selection tool ( ) in the toolbar or use the keyboard shortcut (S).
2. Click on the content you’d like to copy – selected content is highlighted. Click near the edge of the report to
select all content. To extend a selection, hold the shift key.
3. Select Edit > Save Selection As.
4. Under Save as Type (at the bottom of the window), select the desired
graphic format.
For printed media (presentations, journals, etc.) a vector image, such as EMF
(enhanced metafile) is recommended. This format preserves transparency,
can be edited in image-editing software, and can be scaled or resized without
losing clarity.
5. Enter a file name, specify the folder to save the file to, and click Save.

Note: An alternative to saving as a file is to paste the image directly into a document. Use Edit > Copy from
within JMP and choose Paste Special > Picture Enhanced Metafile to place image in a document.

Editing Image Elements within Power Point

Image being edited in Power Point
1. Save the image in steps 4-5 above as a an
enhanced meta file (.emf).
2. After placing image in document, right-click on
the image and select Edit Picture. Confirm choice.
All the elements of the graphic can now be
edited, and formatted as desired.
Note: You may be required to “ungroup” elements
in order to edit individual elements.

Note: It’s often best to edit the graph in JMP to the

desired appearance as much as possible
(e.g., marker sizes and/or colors, borders, axis
settings, fonts, etc.) before saving/copying and
inserting into a document.

Visit “Discovering JMP > Save and Share Your Work” and “Using JMP > Save and Share Your Data”
in JMP Help to learn more.

16
 JMP Basics
jmp.com/learn Nov 2022

Sharing Results on JMP Public®

This guide provides instructions on publishing reports to JMP Public® – a public website https://public.jmp.com
where users can share graphs and reports that others can view and interact with outside of JMP.
A SAS profile is required.

Publish Reports to JMP Public®

1. Go to https://public.jmp.com.
2. Click to Sign in. If you have a SAS profile, log in using those credentials. If you do not have a SAS profile, click
Create profile and follow the steps.
3. From within a JMP file, open the graphs or output you wish to publish. The graph “Bubble Plot by Region”
from the file: SATbyYear.jmp under Help>Sample Data Folder will be used for this exampe.
4. Select File > Publish > Publish Reports to JMP Public.
5. A window will appear listing the reports available to publish.
Select the ones you wish to publish. Here we have only one
graph (Bubble Plot).
6. Choose to Create, Add, or Replace a post and click Next.
7. Provide a Title and Description for your report, choose who to
share with and click Publish.

NOTE: JMP Public® is a public online communication channel.

Do not publish data and reports not suitable for public distribution.

8. Your report(s) will be pulibshed to https://public.jmp.com.

• Some interactive features are available to those viewing the

report (e.g., selecting data points, displaying labels for data
values, hiding and excluding data, and the column switcher
and local data filters if they were setup when the graph
and/or analysis output was created).
• You can download the associated data table(s)
and JSL scripts for a report.
• You can update or delete reports you’ve published.
• Reports can be shared via social media (Facebook, Twitter,
LinkedIn).
• Search features are available on https://public.jmp.com to
search for people or posts.
• You can also publish reports to a location on your own
computer or a network using File > Publish > Publish to File.

Visit “Discovering JMP > Save and Share Your Work” and “Using JMP > Save and Share Your Data”
in JMP Help to learn more.

17
 JMP Basics
jmp.com/learn Nov 2022

Creating Formulas in JMP®

This guide describes how to use formulas to create new variables (e.g, transformations, functions of existing
variables, perform logic operations, create patterned and random data, among others).

Method 1 (Create Formula With Formula Editor)

Keypad
Functions
1. From an open JMP data table, select Cols > New Column.
2. Under Column Name, assign a name for the column. In
this example, we name the column %profit/sales.
3. Click Column Properties, and select Formula. This takes
you to the JMP Formula Editor (shown at right).
4. To create this formula, we take the ratio and multiply by
100.
• Select Profits ($M) from the Columns list.
• Click on the divide by symbol on the keypad.
• Select on Sales ($) from the Columns list.
Example: Companies.jmp (Help > Sample Data Folder)
• Click on the outer gray box (surrounding the ratio).
• Click on the multiply symbol on the keypad.
• Type “100” on your keyboard, and click OK. This populates the data table column with the calculated %
Profit ($M)/Sales ($M) for each observation.

Method 2 (Create New Formula Columns from an Existing Column(s))

1. From an open JMP data table, right-click on a column and choose
New Formula Column. Choose from a list of functions. The new
column along with the formula will be saved to the data table.

Note: If multiple columns are highlighted, a set of Aggregate

formulas will be made available (e.g, mean, sum).

Method 3 (Temporary Columns)

1. From a launch window (e.g., Fit Y by X), right-click on a column in
the columns list. Choose desired formula. This temporary column
(available just within that current platform) can be used.

Note: To save that temporary column to the data table, right-click

on the temporary column that was created and select Add to Data Table.
This adds a new column and creates the formula, which is accessible
through the Formula Editor.

Note: To access the Formula Editor for any formula column, right-click on the
column name and select Formula.

Visit Using JMP > Create Formulas in JMP in JMP Help to learn more.

18
 JMP Basics
jmp.com/learn Nov 2022

Excel Add-In I (Passing Data to JMP®)

Use the Excel JMP add-in to transfer data from an Excel worksheet into a JMP data table or to launch basic JMP
analysis platforms from within Excel.
Note: Add-in is only available for Windows Operating System. A locally installed version of Microsoft Excel must
be on the computer at the time of installing JMP for the Excel Add-in to be added. The add-in will not be
installed if Excel is only run from a virtual environment. See Microsoft Help to learn how to install Excel locally.
If there is no JMP ribbon present after you’ve installed JMP and a local version of Microsoft Excel is also
installed, then you will need to reinstall JMP. During the JMP installation process, make sure Excel Add-In is
selected.

The Excel Add-In: Passing Data to JMP

This JMP ribbon will be accessible in Excel.
You may need to click on the JMP tab for the
JMP ribbon to appear.

1. Preferences: Use to specify

• The name to use for the JMP data table;
• Whether to use the first row as column names (checked by default);
• The number of rows to use for column name (up to four can be used);
• Whether to transfer hidden rows and/or columns.

2. Data Table: Send selected data to JMP. A JMP data table will be created containing the selected data.

Tips:
• If the data table does not automatically display, bring
JMP to the foreground.
• Review the modeling types in JMP – variables will be
assigned either Continuous or Nominal modeling
types.
• JMP will recognize dates and other Excel formats.

3. Graph Builder: Send selected data to JMP and launch the Graph Builder. Interactively create graphs,
including line plots, splines, box plots, bar charts, histograms, mosaic plots and geographic maps.

4. Analysis Platforms (Distribution, Fit Y by X, Fit Model, Time Series, Control Charts): Sends selected data to
JMP and launches the chosen platform.

Visit Using JMP > Import Your Data > Import Data Using the Excel Add-in in JMP Help to learn more.

19
 JMP Basics
jmp.com/learn Nov 2022

Excel Add-In II (Profiling Excel Models in JMP®)

Use the Excel JMP add-in to visualize and explore Excel models in JMP using the Prediction Profiler.
Note: Add-in is only available for Windows Operating System. A locally installed version of Microsoft Excel must
be on the computer at the time of installing JMP for the Excel Add-in to be added. The add-in will not be
installed if Excel is only run from a virtual environment. See Microsoft Help to learn how to install Excel locally.
If there is no JMP ribbon present after you’ve installed JMP and a local version of Microsoft Excel is also installed,
then you will need to reinstall JMP. During the JMP installation process, make sure Excel Add-In is selected.

The Excel Add-In: Profiling Excel Models in JMP

This JMP ribbon will be accessible in Excel. You may need to
click on the JMP tab for the JMP ribbon to appear.

1. Build an Excel model.

In this example, an Excel worksheet is used to model loan amortization
results. Loan (loan amount), Interest Rate and PMT (monthly payment) are
inputs, and Years to Payoff, Total Interest and Total Paid are the outputs.
The output values are derived from the input values using Excel formulas.

2. Select Create/Edit Model to define the model for JMP.

• Define the Model: Click on “+” next to Model to add a new model.
• Define Inputs:
a. Click on the “+” to add an input.
b. Type the input variable name in the Input Name field, or click
Choose to select it in the worksheet.
c. Next to Cell, click Choose to select the cell containing the
formula or value that corresponds to the input variable name.
d. Enter Minimum, Initial and Maximum values, or accept the
defaults, and hit Apply.
e. Repeat until all inputs have been entered.
• Define Outputs: Repeat steps a-c above for each of the outputs, then click OK.

3. Select Run Model to profile the model in JMP.

• Choose the model you want to profile, then click Profile in JMP.
• The Prediction Profiler window will open in JMP.
• In the Prediction Profiler, change values of the inputs by dragging the vertical
red lines to observe changes in the predicted outputs.

Notes: To access the simulator, interaction profiler, or to set desirability functions, click the red triangle next to
Prediction Profiler.

Visit Profiler > Excel Profiler in the JMP Help to learn more.

20
 JMP Basics
jmp.com/learn Nov 2022

Add-In Builder
Use a JMP® Add-In to create custom menus and easily distribute JMP scripts, applications, data tables and more.

Create an Add-In with the Add-In Builder

1. Choose File > New > Add-In (or File > New > New
Add-In on the Mac).
2. In the General Information tab, provide an Add-in
Name and an Add-in ID.

3. In the Menu Items tab, for each menu item:

• Click Add Command to add the menu item  „  ‚

„
and provide a menu item name ‚.

• Browse for the JSL script the menu item will ƒ

run. Or, copy and paste the script next to Run
this JSL ƒ.
• Click Add Submenu (if needed) and provide a
name to group items below a single heading,
or Add Separator to add a separating line „.

4. If desired, use the Start-Up Script, Exit Script, or

Additional Files tabs to add other scripts, graphics or
data tables to the add-in.

5. Click Save . Edit the file name if desired, and click Save.

By default, the add-in will automatically install under the Add-in menu in JMP when saved. Test the menu items
to ensure that everything works as intended, and fine-tune as needed. The Add-in Builder will stay open until
you close it, and saved changes will override the previously installed version of the add-in.
Share the add-in with other users. When the file is opened in JMP, all files are extracted into the appropriate
folder, and the add-in is installed in the JMP Add-in menu.

Tips and Limitations

• To disable (remove from menu) or unregister (uninstall) an add-in, use View > Add-Ins.
• To edit an add-in use File > Open, then click the arrow next to Open and select Open
Using Add-In Builder (on Mac, first click “Options” then check “Edit after opening”).
• To save an application as an add-in, select Script > Save Script to Add-In from the
Application Builder red triangle.

For more information on developing and deploying add-ins, visit “Scripting Guide > Creating Applications” in
JMP Help. To see available add-ins, visit File Exchange > Add-Ins in community.jmp.com.

21
 JMP Basics
jmp.com/learn Nov 2022

Interactive Tools for Teaching and Learning

A variety of interactive tools for teaching and learning statistical concepts are available within JMP. These tools
allow you to explore fundamental concepts covered in introductory statistics courses, including sampling
distributions, confidence intervals, hypothesis testing, ANOVA, regression, and more.

These tools can be accessed by:

1. Available by installing JMP Student Subscription.
Visit www.jmp.com/studentsubscription for instructions.
2. Downloaded and installed as a JMP menu item.
Visit jmp.com/tools for installation instructions.
3. Going to: Help > Sample Index > Teaching Resources >
Teaching scripts > Interactive Teaching Modules and
Help > Sample Index > Teaching Resources > Calculators.

Interactive Teaching Modules

A variety of simulation-based tools to help with the understanding
of important statistical concepts.
Confidence Interval for the Mean shown as an example.

Calculators
Various calculator tools to compute statistical results such as
one-sample and two-sample confidence intervals and hypothesis tests.
Hypothesis Test for One Proportion shown as an example.
These are useful when supporting textbook examples as they allow
one to calculate results using summary statistics.

Other teaching and learning tools are available such as Quick Guides, How-To Videos, Case Studies, Lab
Activities, and Datasets. Visit jmp.com/teach to access. Note: The Quick Guides and How-To Videos can be
accessed from the Student Menu under Quick Help if Student Subscription or the Add-in was installed.

22
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Bar Charts and Frequency Tables

Use to summarize and display the distribution of categorical variables.

Bar Charts and Frequency Distributions Companies.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Distribution.

2. Click on one or more nominal or ordinal variables from Select
Columns, and click Y, Columns (nominal variables have red bars, and
ordinal variables have green bars).
3. If you have summarized data (a column with counts), enter the count
column into Freq.
4. Click OK to generate bar charts and frequency distributions for each
variable.

Tips:
• To change the display from vertical to horizontal, click on the top red
triangle and select Stack.
• To change future output to horizontal, go to File > Preferences.
Select Platforms > Distribution, click Stack and Horizontal Layout,
then click OK.
• To change the graphical display for a variable, or to select additional
options, click on the red triangle for that variable.
• Click on bars in one graph to see the distribution of the variable across
other variables (dynamic linking).
• Categorical variables display in alphanumeric order. To change the display order, use the Value Ordering or
Row Order Levels column property (right-click on the column and select Column Properties).

Bar Charts – Graph Builder

1. Select Graph > Graph Builder.
2. Click, then drag and drop a nominal variable from Select
Columns to the X zone on the bottom of the graph.
3. Click on the bar chart icon above the graph.
4. Drag and drop a continuous weight variable from Select
Columns to the Y zone on the left of the graph, or a drag
and drop a count or frequency variable to the Freq field.
5. Select a statistic to be plotted from list of Summary
Statistics (bottom left).
6. Right-click on graph and select Bar > Label > Label by
Values to display the Mean Profit for each Type.

Notes: Bar charts can also be created in the Chart platform (Graph > Legacy > Chart).

Visit Discovering JMP > Visualize Your Data, Discovering JMP > Analyze Your Data > Analyze Distributions and
Essential Graphing in JMP Help to learn more.

23
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Pareto Plots and Pie Charts

Use to display the distribution of categorical variables. Pareto plots sort in descending order of frequency of
occurrence or by a weighthed value.

Pareto Plots Failuressize.jmp (Help > Sample Data Folder > Quality Control)

1. From an open JMP® data table, select Analyze >

Quality and Process > Pareto Plot.
2. Click on a nominal variable from Select Columns,
and click Y, Cause (nominal variables have red bars,
ordinal variables have green bars).
3. If you have summarized data, enter the Count column
into Freq.
4. Click OK to generate the Pareto plot .

Tips:

• To change the display or select additional options, click

on the red triangle.
• To change the display from a Pareto plot to a pie chart,
click on the red triangle and select Pareto > Pie Chart.
• To label a bar or slice of the pie, right-click on the
category and select Selected Causes > Label.

Pie Charts – Another Way

1. Select Graph > Legacy > Chart.
2. Click on a nominal variable from Select Columns,
and click Categories, X, Levels.
3. Select the variable Count, choose the Statistics
drop-down menu and select Sum.
4. Choose Pie Chart and Click OK.
5. To change the resulting display from a pie chart to a
bar chart, click on the red triangle and select
Vertical Chart or Horizontal Chart.

Notes: Bar charts can also be produced from Analyze > Distribution or Graph > Graph Builder.

For more details on creating pie charts and Pareto plots, see Discovering JMP > Visualize Your Data and
Essential Graphing in JMP Help.

24
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Mosaic Plot and Contingency Table (Cross-tabs)

Use to examine the relationship between two categorical variables. A contingency table displays the frequency
and percentages of the distribution of the data in a matrix format, while a mosaic plot graphically displays the
information.

The Contingency Table Analysis Car Poll.jmp (Help > Sample Folder)

1. From an open JMP® data table, select Analyze > Fit Y by X.

2. Click on a categorical variable from Select Columns, and click Y,
Response (categorical variables have red or green bars).
3. Click on another categorical variable and click X, Factor.
4. Click OK. The Contingency Analysis output will display.

Mosaic Plot
The mosaic plot is a side-by-side divided bar chart that allows
you to visually compare proportions of levels of one variable
across the levels of a second variable.

Interpretation:

1. The widths of horizontal bars represent the proportions of the levels

of the X variable (in this example, 65% Married, 35% Single).
2. The heights of vertical bars on the far right represent the
proportions of the levels of the Y variable (in this example, 45%
small, 41% Medium, 14% Large).
3. The cells in the plot represent the proportions for every
combination of category levels. In this example, Married/Medium
is the largest overall proportion (27.7%) and Single/Large has the
smallest (4.0%)

Contingency Table
Legend
The body of the contingency table displays:
• Count – the cell frequencies (counts).
• Total % - the cell’s percentage of the total count.
• Col % - the cell’s percentage of the count for the column.
The column variable is the Y variable, type.
• Row % - the cell’s percentage of the count for the row.
The row variable is the X variable, marital status.
• Borders display column and row totals.

Tips:
• Click on the red triangle next to Contingency Table to select or deselect display options.
• Right-click on the mosaic plot to change colors (Set Colors) or label cells (Cell Labeling).

Visit Discovering JMP > Visualize Your Data and Essential Graphing in JMP Help to learn more.

25
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Histograms, Descriptive Statistics, and Stem and Leaf

Use to summarize and display the distribution of continuous variables. Histograms and stem and leaf plots allow
you to assess the shape, centering and spread of the data.

Histograms and Descriptive Statistics

1. From an open JMP® data table, select Analyze > Distribution.
2. Click on one or more continuous variables from Select Columns, and click Y, Columns (continuous variables
have blue triangles).
3. Click OK to generate a histogram, outlier box plot and descriptive statistics.
• The percentiles, including quartiles and the median, are listed under Quantiles.
• The sample mean, standard deviation and other statistics are listed under Summary Statistics.
Car Physical Data.jmp (Help > Sample Data Folder)

Tips:

• To change the display from vertical to horizontal (as shown), click on the red triangle next to the variable
name and select Display Options > Horizontal Layout.
• Many other options on summarizing/analyzing a continuous variable is found under the red triangle
for that variable.
• To display different summary statistics, use the red triangle next to Summary Statistics.
• To change the default display, go to File > Preferences > Platforms > Distribution and select options.

Stem and Leaf Plot

To generate a stem and leaf plot, click on the red triangle for the
variable and select Stem and Leaf.

Tips:
• A key to interpret the values is at the bottom of the plot. The bottom three values are 1700, 1800, 1900. The
top 4 values in this example is 4000, 4000, 4000, 4300 (Note: values have been rounded to the nearest 100).
• Click on values in the stem and leaf plot to select observations in both the histogram and the data table. Or,
select bars in the histogram to select values in the stem and leaf plot and data table.

Visit Discovering JMP > Visualize Your Data, Discovering JMP > Analyze Your Data > Analyze Distributions
and Basic Analysis > Distributions in JMP Help to learn more.

26
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Dot plots
Use to display the distribution of continuous variables. Dot plots allow you to assess the shape, center, and
spread of the data. Dot plots are a very useful way for comparing data between groups.

Dot plots – One Variable

1. From an open JMP® data table, select Graph > Graph Builder. Car Physical Data.jmp (Help > Sample Data Folder)
2. Click, then drag and drop a continuous variable from
Select Columns to the X drop zone on the bottom of
the graph (or Y drop zone to display the dot plot on
the vertical axis).
3. To change the method of jittering the points so the
dot plot has a base on the bottom horizonal axis,
select positive grid under the jitter dropdown.

Dotplots – Adding a Categorical Variable

1. Additional categorial variables can be added to allow for comparison of the continuous variable between
groups. Below are two graphs – one with the categorial variable Type placed in the Group Y drop zone and
one with the variable Type placed in the Color drop zone.

Visit Essential Graphing in JMP Help to learn more.

27
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Box Plots
Use to display the distribution of continuous variables. Boxplots are based upon a set of summary statistics that
describe the center and spread of data. Boxplots are a very useful way for comparing data bewteen groups.

Box Plots – One Variable Companies.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Distribution.

2. Click on one or more continuous variables from Select Columns, and Click Y,
Columns (continuous variables have blue triangles). Click OK.
• An outlier box plot is displayed by default next to the histogram
(or above if horizontal layout is used).
• To display a quantile box plot, select the option from the red triangle
for the variable.
Quantile Box Plot Outlier Box Plot The Outlier Box Plot shows the box, plus:

• IQR = the 3rd quartile minus the 1st

quartile.
• Whiskers drawn to the furthest point
within 1.5 x IQR from the box.
• Potential outliers (disconnected
points).
• A red bracket defining the shortest
half of the data (the densest region).
• A confidence diamond showing the
mean (middle of the diamond) and
the upper and lower bound of a 95%
Box Plots – Two Variables confidence interval for the population
mean.
1. Select Analyze > Fit Y by X.
2. Click on a continuous variable from Select Columns, and Click Y, Response.
3. Click on a categorical variable and click X, Factor (categorical variables have red or green bars). Click OK.
• The Oneway Analysis output window will display.
4. Click on the red triangle, and select Display Options > Box Plots to display quantile box plots, or select
Quantiles to display both box plots and a numeric table with the quantile values.

Notes: Box plots for one or more variables can also be generated from Graph > Graph Builder.

Visit Discovery JMP > Visualize Your Data, Basic Analysis > Distributions > Options for Continuous Variables,
and Basic Analysis > Oneway Analysis in JMP Help to learn more.

28
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Scatterplots
Use to display the relationship between two continuous variables. Multiple scatterplots can be displayed
together in a matrix plot.

Scatterplots – Two Variables

1. From an open JMP® data table, select Analyze > Fit Y by X.
2. Click on a continuous response (or dependent) variable in Select Columns, and Click Y, Columns.
3. Click on a continuous predictor (or independent) variable, and click X, Factor. Click OK
• Initital output is a scatterplot. Further options are found under the red traingle.
Car Physical Data.jmp (under Help > Sample Data Folder)

Scatterplots – More than Two Variables

1. Select Graph > Scatterplot Matrix.
2. Select all continuous responses of interest, and click Y,
Columns. Choose Matrix Format. Click OK.
• Initital output is a display of all possible scatterplots.
Further options are found under the red traingle.

Note: Scatterplots and scatterplot matrices can also be generated from Analyze > Multivariate Methods >
Multivariate and from Graph > Graph Builder.

Visit Discovery JMP > Visualize Your Data > Compare Multiple Variables, Basic Analysis > Bivariate Analyses,
Essential Graphing > Scatterplot Matrix, and Multivariate Methods > Correlations and Multivariate Techniques
in JMP Help to learn more.

29
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Run Charts (Line Graphs)

Use to display continuous data in time sequence. Run Charts are very useful to discover trends and patterns
over time and to make comparisons to other variables or groups.

Run Charts (Graph Builder)

GNP.jmp (Help > Sample Data Folder > Time Series)
1. From an open JMP® data table, select Graph > Graph
Builder.
2. Drag a variable (or multiple variables at once) from
the Variables list and drop in the Y zone. (e.g., ‘gross
private domestic investment’ and ‘net exports of
goods and services’).
3. Drag and drop a variable indicating the time ordering
in the X zone (‘Date’).
4. Click on the Line icon in the graph pallet (top middle).

Tips:
• Right-click on the legend to change the line thickness or other line properties.
• Double-click on an axis to change axis settings.
• Float the cursor near the middle of the x-axis so that the hand icon appears. Click and drag up-and-down to
zoom in/out of time regions. Drag the hand icon left or right to move forward/backward in the time scale.

Run Charts (Overlay Plot)

1. From an open JMP data table, select Graph > Legacy > Overlay Plot.
2. Select one or more continuous variables from Select Columns and click Y.
3. If you have a column that indicates time ordering, enter the column into X, and click OK.

4. Click on the red triangle and select Show Control Panel to launch the Graph Builder platform and its
collection of tools to modify the graph.

Note: Run charts can also be produced from the Control Chart platform (Analyze > Quality and Process >
Control Chart > Run Chart).

Visit Essential Graphing > Graph Builder Examples > Example of an Area and Line Graph in JMP Help to learn
more.

30
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Interactive Graphing with Graph Builder

Use Graph Builder to interactively create graphs for one or more variables, including line plots, splines, box
plots, bar charts, histograms, mosaic plots, maps and more.
Big Class.jmp (under Help > Sample Data Folder)
Drag and Drop to Visualize Data
1. From an open JMP® data table select Graph >
Graph Builder.
2. Drag a variable from the Variables list and drop it in the
desired drop zone. In the examples (right), Weight is in the
X zone and Height is in the Y zone.
3. To add a grouping variable, drag and drop a variable in the
Group X or Group Y zone. In the example, Sex is in the
Group X zone.
4. To change the graphical display, click on a graph element
icon. Here, Line of Fit has been selected.
5. Change Summary Statistics and other display options for
the selected graph elements.
6. Click the Done button (top left) when finished.

Tips:
• Right-click in the graph to change graph properties.
• To replace a variable with a new variable, drag the new variable and drop it in the center of the drop zone.
• By default, Graph Builder displays data points. If continuous variables are in both the X and Y zones a smooth
spline will also be display by default (lambda = 0.05).
• More than one variable can be assigned to an X or Y zone, or to a group zone. Drag a variable
to either side of the existing variable in the zone – a blue ribbon will indicate where the new
variable will be placed when dropped.
• To change the modeling type to utilize different graph types,
right-click on the variable and select the new data type (if available).

Other Drop Zones:

• Drop a variable in Wrap to trellis the graph horizontally and vertically.

• Drop a variable in Color to create a legend and color by values of the variable.
• Drop a variable in Overlay to color and overlay graphs for each value of the variable on one graph.
• If data has been summarized (a frequency variable exists), drag the variable to the Freq zone.
• If a column defines a physical shape, drag the variable to Shape to create a map (shape files must exist).
• Drop a variable in Size to scale markers or map shapes according to the value of the size variable.

Visit Discovering JMP > Visualize Your Data, Essential Graphing > Graph Builder in JMP Help to learn more.

31
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Bubble Plot
Use a Bubble Plot to explore the relationship between two continuous variables and how the values change over
time. Bubble color and size can be used to incorporate additional variables.

Bubble Plot
1. From an open JMP® data table select Graph > Bubble Plot. SATbyYear.jmp (under Help > Sample Data Folder)

2. Drag a continuous variable into the Y and X roles.

3. Place one or more variables into the ID role so that labels can be
added to the bubbles (optional).
4. Add a time variable into the Time role if desired to have the
bubbles change location across time (optional).
5. Add a variable(s) into the Sizes and Coloring role to incorporate
information about those variables into the plot (optional).
6. Click OK.
7. Play, Rewind, and Forward button are available to have the
bubble positions, size, and/or color change as a function of the
Time variable.

Note: Options such as adding Trail Lines and Labels are available
under the Red Triangle.

8. If there are multiple ID Variables (as in Region and State in this

example), selecting a bubble and clicking the Split Button will
separate the selected bubble into the values for the second ID
Variable. In this example, the South Region was split into the
individual states that make up that region. Select the Combine
Button to return the individual states back into Region.
Note: The animation can be saved:
In Windows OS
1. Save as a .gif using the Record and Save Buttons.
2. Create as an Interactive HTML through File > Publish >
Publish to File or File > Export > Interactive HTML with Data.
In Apple OS
1. Create as an Interactive HTML through File > Publish to File
or File > Export > Interactive HTML with Data.

Visit Essential Graphing > Bubble Plot in JMP Help to learn more.

32
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Summarizing Data Using Tabulate

The Tabulate platform provides a tool to interactively construct tables of descriptive statistics.

Drag and Drop to Summarize Data Car Physical Data.jmp (Help > Sample Data Folder)

• From an open JMP® data table select Analyze > Tabulate.

• Drag and drop variables from the column list to the drop zone
for rows and columns.
o Country (below, left) is in the rows drop zone – the number
of observations per country is displayed.
o Horsepower (middle) is in the columns drop zone as an
analysis column – the sum for horsepower is displayed for
each country.
• Drag and drop one or more summary statistics from the middle
panel into the results area. Mean and Std Dev are displayed for
each country (below, right).

Tips:
• Click Undo to reverse the last change, or use Start Over to clear the display.
• Click and drag variables in the table to rearrange, or right-click on a variable to
delete or change the format.
• To change the numeric formats (i.e., decimal places), use Change Format at the
bottom of the window and select the desired format.
• To add new summary panels to the table, drag and drop the new variable to the
bottom or left of the table. Here, Type has been added to the bottom of the
original table.
• To add additional row or column variables, drag and drop a
new variable on either side of the current variable in the
table. Here, Type has been added next to Country and
Horsepower has been added next to Weight.
• To create a data table, click Done, then select Make Into
Data Table from the top red triangle.

Visit Basic Analysis > Tabulate in JMP Help to learn more.

33
 Graphical Displays and Summaries
jmp.com/learn Nov 2022

Mapping in Graph Builder

Use the Graph Builder to create interactive maps of U.S. states, U.S. counties, and worldwide countries and
provinces. JMP® ships with these shape files. Other shape files (e.g., ESRI) can be used or you create your own
custom maps.
CrimeData.jmp (Help > Sample Data Folder)
Basic Mapping
1. From an open JMP data table, select Graph >
Graph Builder.
2. Drag a shape variable from the Variables list (for
example, State) to main drop zone.
3. Drag variables into other drop zones until the
desired map is produced.
4. Use the Undo and Start Over buttons to try
several options. Click Done when finished.

The resulting display depends on the modeling type of the variables and the drop zone(s) used. Examples:
• Left: Region (Nominal) was dropped in the Color zone.
• Middle: Total Rate (Continuous) was dropped in the Color zone.
• Right: Total Rate was dropped in the Color zone, and Year (Continuous) was dropped in the Wrap zone.

Tips:
• Right-click on the legend to change the color gradient or transparency.
• Use the Local Data Filter accessed under the Red Triangle in a report to dynamically select, show and include
values of selected variables. The Local Data Filter applies to only the active report. See Discovery JMP > Work
with Data Tables > Example of Filtering Data in a Report and in Using JMP > JMP Reports > The Data Filter in
JMP Help to learn more.
• If your data set contains latitudinal and longitudinal data, San Francisco Crime.jmp (Help > Sample Data Folder)
you can add a background map or image. Drag these latitude
and longitude variables to the Y and X zones, right-click on the
graph, select Graph > Background Map and choose the desired
image (e.g., street map of San Francisco in the example to the
right). In this example, Day of Week was placed in the Color
Role, and “Assault” chosen from the variable Category that
was selected for the Local Data Filter).

Visit Essential Graphing > Maps in JMP Help to learn more.

34
 Probabilities and Distributions
jmp.com/learn Nov 2022

Assessing Normality
This guide provides some ways to assess the fit of a normal distribution to a continuous variable. See options for
fitting and assessing the fit of other non-normal distributions in the Fitting Distributions guide.

1. From an open JMP® data table, select Analyze > Distribution.

2. Select one or more continuous variables from Select Columns and click Y, Columns.
3. Click OK to generate a histogram (Histogram Only was selected in this example).

Car Physical Data.jmp

(Help > Sample Data Folder)

Normal Quantile Plot

Click on the red triangle for the variable, and select Normal Quantile Plot.

If the data more or less follow a straight line (fat

pen test), we can conclude that the data are
reasonably approximated by a normal
distribution.

For this example, we would conclude the

distribution is approximately normal.

Fitting a Normal Distribution

1. Select Continuous Fit > Fit Normal from the lower red triangle for the
variable.
2. In the resulting output, click on the red triangle for Fitted Normal
Distribution and select Goodness of Fit.
Two Goodness of Fit tests are performed to
evaluate if the normal distribution is a good fit
to the data (Shapiro-Wilk and Anderson
Darling).

Both have large p-values indicating that there

is no statistical evidence suggesting the normal
distribution is not a good fit to these data.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Normal Quantile Plot and
Basic Analysis > Distributions > Options for Continuous Variables > Fit Distributions in JMP Help to learn more.

35
 Probabilities and Distributions
jmp.com/learn Nov 2022

Fitting Distributions
This guide provides information on fitting various continuous or discrete distributions to data.

Fitting One Continuous Distribution

1. From an open JMP data table, select Analyze > Distribution.
2. Select one or more continuous variables from Select Columns, click Y, Columns, then click OK.
3. Select Continuous Fit from the red triangle for the variable and select a distribution (LogNormal was
selected in the example below).
4. In the resulting fitted distribution output, click on the red triangle and select Goodness of Fit (shown) or
Diagnostic Plot to assess the fit of the distribution.
Hollywood Movies.jmp (Help > Sample Data Folder)

The small p-value suggest that the

LogNormal Distribution does not
provide good fit to the data.

Fitting All Continuous Distributions

Select Continuous Fit, then Fit All from the red triangle for the variable. JMP will compare available continuous
distributions. Note: The distribution with the lowest AIC value provides the best fit to the data. Goodness of Fit
tests can be performed by selecting Goodness of Fit under the Red Triangle for the Fitted Distribution output.

Fitting Discrete Distributions

If the continuous variable contains only integer values, four
discrete distributions are available under Discrete Fit.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Fit Distributions in JMP Help to learn
more

36
 Probabilities and Distributions
jmp.com/learn Nov 2022

Finding Standardized Values (z-Scores)

This guide demonstrates three methods for calculating standardized values (z-scores) for a continuous variable.

Method 1 (Save Standardized)

1. From an open JMP data table, select Analyze > Distribution.
2. Select one or more continuous variables from Select Columns
and click Y, Columns.
3. Click OK to generate a histogram and descriptive statistics.
4. Click on the red triangle for the variable, and select Save >
Standardized. A new column of the standardized values will be
saved to the data table.
Big Class.jmp (Help > Sample Data Folder)

The column Std height

contains z-scores for
height.

Method 2 (Instant Formula)

1. From an open JMP data table, right click on the column that you want to standardize and select New
Formula Column > Distributional > Standardize. A new column of the standardized values will be created
in the data table.

To view the formula, right-click on the column header for this new column and select Formula.

Method 3 (Create Formula with Formula Editor)

1. From an open JMP data table, select Cols > New Column.
2. Under Column Name, assign a name for the column.
3. Click Column Properties, and select Formula. This takes you to the JMP Formula Editor.

A standardized value (z-score) is of the form:

4. Create the formula. This will populate the column in the data table with standardized values for the variable.
Below are two ways to create the standardized values in the Formula Editor. All functions in this example
can be found under the “Statistical” functions.

Visit Using JMP > Create Formulas in JMP in JMP Help to learn more.

37
 Probabilities and Distributions
jmp.com/learn Nov 2022

Finding the Area Under a Normal Curve

This guide demonstrates how to find the area under the normal curve using formulas and the Distribution
Calculator.

Column Formula for Area Under a Normal Curve (One Value)

1. Select File > New > Data Table.
2. Add one row - select Rows > Add Rows, and type “1”. Click OK.
3. Right-click on Column 1, and select Formula to access the Formula Editor.
4. From the function list on the left, select Probability > Normal Distribution.
This will give the following formula:
Returns the probability that a
normally distributed random variable
is less than q.

5. Click the caret on the keypad twice to add fields for

the mean and standard deviation.

6. In the fields provided, enter the value of q, the mean, and the standard deviation. Click OK.
• JMP® will populate the row in the data table with the cumulative probability value
(the area under the lower tail of the normal curve).
Note: To find the area in the upper tail, or the area between two values, use the following formulas:
• Upper tail:
• Between two values:

Note: The Distribution Calculator from Help > Sample Data >
Teaching Scripts > Interactive Teaching Modules is another method
to find areas of a normal distribution (shown on the right).

Column Formula for Area under a Normal Curve

(for a Column of Values)
1. Open an existing data table (File > Open), or open a new data table
(File > New > Data Table) and create a column of values. Big Class.jmp (Help > Sample Data Folder)
2. Select Cols > New Column to create an additional column, and rename
the column. We’ll name our column Prob Height.
3. Click Column Properties, and select Formula to access the Formula Editor.
4. Follow steps 4 and 5 above.
5. Click on the box containing “q,” and from Table Columns,
select the variable (for example, Height).
6. Type the mean and standard deviation in the fields
provided or use the Col Mean and Col Std Dev functions.
• JMP will populate the column with cumulative probability values for each value of the variable.

Visit Using JMP > Create Formulas in JMP in JMP Help to learn more.

38
 Probabilities and Distributions
jmp.com/learn Nov 2022

Random Sampling and Random Data

This guide demonstrates methods for selecting a random sample and generating random data.

Random Sampling Car Physical Data.jmp (Help > Sample Data Folder)

1. From an open JMP data table, select Tables > Subset.

2. Specify how you’d like the sample to be selected:
- Random – sampling rate (specify the proportion).
- Random – sample size (specify the desired sample size).
- To select a stratified sample across another variable,
check Stratify and select the variable.
3. Under Columns, specify All columns or Selected columns.
4. Click OK to generate the random sample.

To connect the subset to the original table, select

Link to original data table before clicking OK.

Generating Random Data

Manually Create Formula:
1. Create a new Data Table (File > New > Data Table).
2. Create the number of desired row (Rows > Add Rows).
3. Assign a name for the column (e.g., ‘Random Normal’)
4. Right-click on the column and select Formula.
This takes you to the JMP Formula Editor.
5. Click on the gray icon next to Random in the function
list on the left, and select the distribution of interest.
Here, we will select Random Normal.
6. Click on the caret once to add the standard deviation field.
Type in the desired values for the mean and standard deviation. Click OK.

• JMP will populate the new column with simulated random

normal data.

Instant Formula:
Right click any continuous column in the data table and select
New Formula Column > Random > Random Normal.
Simulated random standard normal data.
A new column containing a formula will be added to the data table. The fat plus sign next to the variable name under
To reveal or modify the formula, right click on the column header the Columns panel tells us that a formula is stored
for the new column and select Formula. in that column.

Visit Using JMP > Reshape Your Data > Create a Subset Data Table, and Using JMP > Create Formulas in JMP
in JMP Help to learn more.

39
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Hypothesis Tests and Confidence Intervals for Proportions

Use to estimate via a confidence interval and perform hypothesis tests for a population proportion.

Confidence Intervals for Population Proportions Car Poll.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Distribution.

2. Select one or more categorical variables from Select Columns,
click Y, Columns (categorical variables have red or green bars).
Note: If you have summarized data (a column with counts),
enter the column into Freq.
3. Click OK.
4. In the resulting window, click on the red triangle for the
variable and select Confidence Interval > 0.95.
JMP will produce 95% confidence intervals for the true
population proportion for each level.

Hypothesis Tests for Population Proportions

1. From the Distribution output window, click on the red triangle
for the variable and select Test Probabilities.
2. Enter the hypothesized proportions under Hypoth Prob, and
click Done.
Here we are testing the following set of hypotheses:
H0: pFamily = 0.5 vs. HA: pFamily ≠ 0.5
H0: pSporty = 0.3 vs. HA: pSporty ≠ 0.3
H0: pWork = 0.2 vs. HA: pWork ≠ 0.2

Notes: The hypothesized probabilities must sum to one. You may choose to specify some values and have
JMP rescale according to your choice of rescaling method.

• JMP will provide the results of two chi-square tests:

Likelihood Ratio and Pearson.
• The null hypothesis is that the true proportions are equal to the
hypothesized values.
• Small p-values (<0.05) indicate that at least one sample
proportion is significantly different from the hypothesized value.
• Since the p-values in this example are large (> 0.05), we cannot
reject any of the null hypotheses.

This analysis can also be performed using the Hypothesis Test for One Proportion and Confidence Intervals for
One Proportion Calculators under Help > Sample Index > Calculators or Student > Calculators in JMP Student
Subscription.

Visit Basic Analysis > Distributions > Additional Examples of the Distribution Platform > Example of Testing
Probabilities for Two Levels and Example of Testing Probabilities for More Than Two Levels in JMP Help to
learn more.
40
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Two Proportions Test and Confidence Interval

Use to Estimate via a confidence interval and perform a hypothesis test for the difference between two
population proportions. If comparing more than two proportions, refer to the Chi Square Tests for a
Two-Way Table guide.
Car Poll.jmp (Help > Sample Data Folder)
Two Proportions Test
1. From an open JMP® data table, select Analyze > Fit Y by X.
2. Choose the binary response variable for the Y, Response.
3. Choose the 2 levels variable that defines the groups for the
X, Factor. Click OK.
4. Select Two Sample Test for Proportions under the
top Red Triangle.

Two Sample Test for Proportions

JMP displays a Mosic Plot, Contigency Table (not shown) as well
as a ChiSquare and Fisher’s Exact Tests (not shown). These tests
can be used when comparing two groups and also when there
are more than two groups to compare. Choosing the Two
Sample Test for Proportions adds a Confidence Interval and
Adjusted Wald’s Test for comparing two proportions.

Here we would estimate the difference between the two

population proportions (pmarried|female - p married|male) to be 0.076
with a 95% Confidence Interval of (-0.032 , 0.182).

Here we are testing the two-sided hypothesis:

H0: pmarried|female = p married|male

HA: pmarried|female ≠ p married|male

Interpretation (using a significance level of 0.05):

The p-value for this test is 0.1686 indicating there is not statistically
significant evidence to conclude a difference in proportion of females vs.
males in the population that are married.

Note the p-value for the Wald’s Test is very similar to the Likelihood Ratio
and Pearson’s ChiSquare tests.

Notes: Relative Risk and Odds Ratio, two analyses useful when comparing two proportions, are available under
the Red Triangle.
The Z-Test approach for comparing two proportions can be peformed using the Hypothesis Test for Two
Proportions and Confidence Intervals for Two Proportions Calculators under Help > Sample Index > Calculators
or Student > Calculators in JMP Student Subscription.
Visit Basic Analysis > Contingency Analysis in JMP Help to learn more.

41
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Chi Square Tests for a Two-Way Table

Use to test for independence or homogeneity of two categorical variables. If comparing only two groups with a
binary outcome, refer to the Two Proportions Test and Confidence Interval guide.

The Contingency Table Analysis

Car Poll.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Fit Y by X.

2. Click on a categorical variable from Select Columns, and click
Y, Response (categorical variables have red or green bars).
3. Click on another categorical variable and click X, Factor.
4. Click OK. The Contingency Analysis output will display.

By defaut, a Mosaic Plot, Contingeny Table, and Chi-Square

Tests are shown.
See the Mosaic Plot and Contingency Table guide for more
information.

Chi-Square Tests
By default, JMP provides results for two chi-square tests under
“Tests” – the Likelihood Ratio and Pearson.

• If both variables can be considered as response (Y) variables,

these chi-square statistics test that the variables are
independent.
• If one variable is considered a response (Y) and the other as a
fixed factor (X), the tests are for homogeneity of Y across X.
The Expected Count
and the contribution
Interpretation (using a significance level of 0.05): of each cell results
to the Chi-Square
• P-values for the two tests are given under Prob>ChiSq.
Test Statistic was
• Since the p-values are less than 0.05, we conclude that there is a added to the
significant difference in the probability of purchasing a particular Contingency Table
by selecting that
type of car for married and single adults (i.e., car choice is not
option under the
homogenous across the two marital status groups). Red Triangle.

Note: Additional analysis options are available under the Red Triangle including
Analysis of Means for Proportions, Measures of Association, and Exact Tests.

Visit Basic Analysis > Contingency Analysis in JMP Help to learn more.

42
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Sample Size and Power for One Sample Proportion

Use to interactively explore the relationships between Power, Sample Size, and Alternative Proportions in
testing a hypothesis for a single population proportion. See the Hypothesis Tests and Confidence Intervals for
Proportions guide to learn how to test a hypothesis for a population proportion.

Sample Size and Power - One Sample Proportion

1. Select DOE > Sample Size Explorers and choose Power > Power for One Sample Proportion.
2. Choose the type of test: One-Sided or Two-Sided and Alpha (significance level for the test).
3. Choose Method (Exact Test or Normal Approximation).
4. Enter the Assumed Proportion. This is the value of the population proportion being tested and assumed
under H0. Here we are planning on testing the Hypothesis H0: p = 0.50 vs. HA: p > 0.50.
5. Enter the Alternative Proportion. This is the value for the population proportion under HA being considered.
Here we consider the scenario of the population proportion being 0.75.
6. Select parameter to solve for. Here we chose Sample Size.
7. Enter a value for the Power. Here we entered 0.80.
The solution (n=23) is shown in the Sample Size field.
Note: The Power field changed to 80.77% since sample size can only be an integer.
n=22 results in a Power of 79.07%.
8. Use the interactive cross-hair tool (or type in values) for Power, Sample Size, Assumed Proportion, and
Alternative Proportion to study the relationship between these parameters solving for many different
scenarios. The settings and solution for each analysis
performed can be saved. The table of saved
settings shows the results of five different
analyses performed when testing the
hypothesis H0: p = 0.50 vs. HA: p > 0.50
1. What sample size is needed to achieve a
power of 80% assuming an alternative
proportion of 0.75?
Answer: n = 23
2. What is the power with a sample size of 22
assuming an alternative proportion of 0.75?
Answer: Power = 79.1%
3. What is the power for a sample size of 30
assuming an alternative proportion of 0.75?
Answer: Power = 89.7%
4. What is the power for a sample size of 30
assuming an alternative proportion of 0.65?
Answer: Power = 49.9%
5. What sample size is needed to achieve a
power of 80% assuming an alternative
proportion of 0.65?
Answer: n = 67

Note: Determining sample size to achieve a desired margin of error in a Confidence Interval can be done
using DOE > Sample Size Explorers > Confidence Intervals > Margin of Error for One Sample Proportion.

Visit Design of Experiments Guide > Sample Size Explorers in JMP Help to learn more.

43
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Sample Size and Power for Two Sample Proportions

Use to interactively explore the relationships between Power, Sample Sizes, and Alternative Proportions in
testing a hypothesis comparing two independent population proportions. See the Two Proportions Test and
Confidence Interval guide to learn how to perform a statistical test comparing two population proportions.

Sample Size and Power - Two Sample Proportion

1. Select DOE > Sample Size Explorers and choose Power >Power for Two Independent Sample Proportions.
2. Choose the type of test: One-Sided or Two-Sided and choose Alpha (significance level for the test).
The Null Hypothesis is that the two proportions are equal. Here we chose a two-sided alternative which is
used to test that the two proportions are not equal.
This null and alternative hypothesis can be written using notation as H0: p1 = p2 vs. HA: p1 ≠ p2
3. Enter values for the Group 1 and Group 2 Population Proportions under HA to base the analysis on.
Here we consider the scenario where Proportion 1 is 0.50 and Proportion 2 is 0.75.
4. Select parameter to solve for. Here we chose Total Sample Size.
5. Enter a value for the Power. Here we entered 0.80.
The solution of Total Sample Size of 110 (n1=55 and n2=55) is displayed.
6. Use the interactive cross-hair tool (or type in values) for Power, Sample Sizes, and Assumed Proportions to
study the relationship between these parameters solving for many different scenarios.

The settings and solution for each

analysis performed can be saved. The
table of saved settings shows the
results of five different analyses
performed when testing the hypothesis
H0: p1 = p2 vs. HA: p1 ≠ p2
1. What sample size is needed to
achieve a power of 80% assuming
p1=0.50 and p2=0.75?
Answer: 110 (n1 = 55 and n2 = 55)
2. What is the power with a sample size
of 25 in each group assuming p1=0.50
and p2=0.75? Answer: Power = 47.2%
3. What sample size is needed to
achieve a power of 80% assuming
p1=0.50 and p2=0.65?
Answer: 335 (n1 = 167 and n2 = 168)
4. What difference from 0.50 for the
Note: Determining sample size to achieve a desired margin of error in a proportion of group 2 can be
Confidence Interval can be done using DOE > Sample Size Explorers > detected with 80% power using a
sample size of 100 in each group?
Confidence Intervals > Margin of Error for Two Independent Sample
Answer: 0.69
Proportions.

Visit Design of Experiments Guide > Sample Size Explorers in JMP Help to learn more.

44
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

One Sample t-Test and Confidence Interval

Use to estimate via a confidence interval or perform a hypothesis test for a population mean.

Confidence Interval for the Mean Car Physical Data.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Distribution.

2. Select one or more continuous variables from Select
Columns, click Y, Columns (continuous variables have blue
triangles), and click OK.
The Upper 95% Mean and Lower 95% Mean shows the 95%
confidence interval for the population mean.
Note: To change the confidence level, request a one-sided confidence limit or specify sigma, click on the red
triangle for the variable, select Confidence Interval, and select the confidence level or click Other.

One Sample t-Test for the Mean

1. From the Distributions report window (shown above), click on the red triangle
for the variable and select Test Mean.
2. Enter the hypothesized value under Specify Hypothesized Mean, and click OK.
We will test the following hypothesis:
H0: µ = 40 vs. HA: µ ≠ 40. JMP will display:
• The t-statistic.
• P-values for the two-tailed and one-tailed tests.
• A graph to aid in interpreting the p-values, showing the hypothesized mean
(center of the curve) and the sample mean (red line).

Interpretation of p-values for this example (using a significance level of 0.05):

1. Prob > |t| is less than 0.05 - reject the null hypothesis that the true mean is 40. This is the two-tailed test. Conclude that
the true mean is not 40.

2. Prob > t is greater than 0.05 - fail to reject the null hypothesis that the true mean is <= 40. This is a one-tailed test.
There no statistical evidence to reject the null hypothesis.

3. Prob < t is less than 0.05 - reject the null hypothesis that the true mean is >= 40. Conclude that the true mean is less
than 40. The confidence interval (38.0 , 39.2) provides an estimate of how different the population mean is from 40.

Notes: To explore how the p-value changes as a function of the difference between the hypothesized mean and
the sample mean, click on the red triangle next to Test Mean and select PValue animation. Use the Power
Animation tool to examine how power changes under different scenarios.
This analysis can also be performed using the Hypothesis Test for One Mean and Confidence Intervals for One
Mean Calculators under Help > Sample Index > Calculators or Student > Calculators in JMP Student
Subscription.

Visit Discovering JMP > Analyze Your Data > Analyze Distributions and Basic Analysis > Distributions in JMP
Help to learn more.

45
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Two Sample t-Test and Confidence Intervals

Use to Estimate via a confidence interval and perform a hypothesis test for the difference between two
population means. If more than two means (more than two levels of the categorical X variable), refer to the
One-Way ANOVA guide.

Comparison of Two Population Means Big Class.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Fit Y by X.

2. Click on a continuous variable from Select Columns, and click
Y, Response.
3. Click on a two-level categorical variable and click X, Factor.
4. Click OK.
The Oneway Analysis output window will display.
5. Click on the red triangle and select Means/Anova/Pooled t.

Some of the additions to the report include:

• Mean diamonds (95% Confidence Intervals) added to the
graph.
• T-test Statistic and p-values for testing the hypotheses:
H0: µ1 = µ 2 vs. HA: µ 1 ≠ µ 2 ; or HA: µ 1 > µ 2 ; or HA: µ 1 < µ 2
• Confidence Interval for (µ1 - µ 2).
• Individual Confidence Intervals for µ1 and µ 2.

Note: Means/Anova/Pooled t is the test under the assumption of equal variances.

For a test without the assumption of equal variances, select t Test under the red triangle instead.

Results

• Upper CL Dif and Lower CL Dif give the 95% CI for (µ1 - µ 2), the difference between the two population means.
Here we estimate that difference to be (-6.89 , 21.64). Since the 95% CI contains zero, we conclude that there is not the
statistical evidence needed to conclude a significant difference between the means.
• Prob > |t| is the p-value for the two-tailed test. The null hypothesis is that means are equal (the mean difference is zero).
Since Prob > |t| is greater than 0.05, cannot reject the null hypothesis (i.e., we cannot conclude that there is a significant
difference between the two population means).
• 95% Confidence Intervals for the Individual Means are shown in the Means for OneWay Anova table. We estimate, with 95%
confidence, that the population mean Weight for Females to be between 90.4 and 111.5 and to be between 98.7 and 117.9
for Males.
Note: The default confidence level is 95% (i.e., significant level of 0.05.) Select Set α Level under the red triangle to change.

This analysis can also be performed using the Hypothesis Test for Two Means and Confidence Intervals for Two
Means Calculators under Help > Sample Index > Calculators or Student > Calculators in JMP Student
Subscription.

Visit Discovering JMP > Analyze Your Data > Analyze Relationships and Basic Analysis > Oneway Analysis in
JMP Help to learn more.

46
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Paired t-Test and CI

Use to test if the populations means of two paired (dependent or correlated) samples are statistically different.
Note: The paired measurements must be stored in separate columns.

Paired t-Test Using Matched Pairs Blood Pressure by Time.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Specialized

Modeling > Matched Pairs.
2. Select two continuous variables from Select Columns, click
Y, Paired Responses (continuous variables have blue
triangles), and click OK.

By default, JMP will generate:

A graph, containing:

- The differences between the paired readings.

- The mean difference (solid red line).
- The 95% confidence interval for the mean difference (dashed
lines).
Note: A reference diamond is included on the graph when
the range of differences is greater than half the range of the data.

Statistics, including:

- The sample means for each variable.

- The difference between the sample means.
- A 95% confidence interval for the difference in the population
means.
- The t-test statistic and p-values.

Results

• Upper 95% and Lower 95% give the 95% CI for the difference between the two population means.
Here we estimate that difference to be (14.8 , 20.4). Since the 95% CI does not contains zero, conclude that there is enough
statistical evidence to conclude a significant difference between the means.
• Prob > |t| is the p-value for the two-tailed test. The null hypothesis is that the difference in the means is zero.
Since the Prob > |t| is less than 0.05, reject the null hypothesis, conclude that there is a significant difference between the
means. This is the same conclusion reached from examining the Confidence Interval. This CI, however, provides the
additional important information of an estimate of how different the means are.
Note: The default confidence level is 95% (i.e., significant level of 0.05.) Select Set α Level under the red triangle to change.

Note: A paired t-test is equivalent to performing a one sample t-test on a column of differences using the
Distribution platform, where the null hypothesis is that the mean difference is zero.
See the page One Sample t-Test and CI for instructions to perform that analysis.

Visit Predictive and Specialized Modeling > Matched Pairs Analysis in JMP Help to learn more.

47
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

One-Way ANOVA
Use to test for a statistical differences in comparing three or more population means.

One-Way Analysis of Variance Companies.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Fit Y by X.

2. Click on a continuous variable from Select Columns, and Click Y,
Response (continuous variables have blue triangles).
3. Click on a categorical variable and click X, Factor (categorical
variables have red or green bars). Click OK.
The Oneway Analysis output window will display.
4. Click on the red triangle, and select Means/Anova.

Some of the additions to the report include:

• Mean diamonds (95% Confidence Intervals) added to the
graph.
• The Summary of Fit.
• The Analysis of Variance (Anova) table.
• Means for Oneway Anova, containing summary statistics
and confidence intervals for each mean (based on the
pooled estimate of the standard error).

• The null hypothesis is that there are no differences between the population means (i.e., all means are equal).
• Prob > F is the p-value for the whole model test. Since the Prob > F is less than 0.05, reject the null hypothesis of equal means.
Conclude that there are differences between at least two of the means.
• To determine which means are different, a post hoc multiple comparison technique can be used.
Notes: The default confidence level is 95% (i.e., significant level of 0.05.) Select Set α Level under the red triangle to change.
Analysis can also be made assuming unequal variances. Select Unequal Variances under the red triangle to perform analysis.

Multiple Comparison Procedures

From the Oneway Analysis output window (shown above), click on the red
triangle, select Compare Means, and select one of the five methods.

Each Pair, Student’s t has been selected. This produces comparison

circles (shown), along with statistical output (not shown).

Click on a circle for a mean to test for paired differences.

• The selected mean will have a bold, red circle and variable label.
• Means that are not significantly different from the selected mean will
have unbolded, red circles and variable labels.
• Means that are significantly different from the selected mean will
have gray circles and gray italicized variable labels.

In this example, the mean for big is significantly different from the mean
for small, but is not significantly different from the mean for medium.

Visit Basic Analysis > Oneway Analysis in JMP Help to learn more.

48
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Two-Way (Factorial) ANOVA

Use to test and estimate the effect that two categorical factors and their interaction have on the population
mean.

1. From an open JMP® data table, select Analyze > Fit Model. Analgesics.jmp (Help > Sample Data Folder)
2. Click on a continuous variable from Select Columns, and
click Y, Response (continuous variables have blue triangles).
3. Click on two categorical variables from Select Columns, and
click Macros, Full Factorial (categorical variables have red or
green bars). This adds each main effect and the interaction
between the factors as model effects. Click Run.
The Fit Model output window will display.
The Effect Summary table shows p-values for statistical
tests for the significance of each model term. More details in
these tests are provided in the Effect Tests table.

• Terms can be removed from the model by selecting them in the Effect Summary table and clicking
Remove.
• The Null Hypothesis in these tests is that the term has no effect on the outcome and the alternative
hypothesis is that the term does have an effect.
• The accepted approach is to examine the test for the interaction(s) first. If not significant, remove them
from the model and proceed to testing the main effects.
Note: if an interaction is included in the model, the main effects must also be included regardless if the
tests for the main effects are significant or not.
• Here we’ll choose to keep all terms in the model (interaction term is significant at the 0.10 level).

Many options are available under the Red Triangle such as Show Predicted Expression, Model Diagnostics,
Multiple Comparisons, Saving results to the data table, and Factor Profilers (shown below)

Visit Discoverying JMP > Analyze Your Data > Compare Averages for Multiple Factors and Fitting Linear Models
in JMP Help to learn more.

49
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Nonparametric Tests
This guide illustrates how to perform a variety of nonparametric tests. For information on nonparametric
correlations and measures of association, see the page Nonparametric Correlations.

One-Sample Nonparametric Tests

Car Physical Data.jmp (Help > Sample Data Folder)

1. From an open JMP data table, select Analyze > Distribution.

2. Select one or more continuous variables from Select
Columns, click Y, Columns, and click OK.
3. From the Distributions report window, click on the red
triangle for the variable and select Test Mean.
4. Enter the hypothesized value under Specify Hypothesized
Mean, check the Wilcoxon Signed Rank box, and click OK.
The hypothesis being tested under the Wilcoxon Signed Rank
test is: H0: Median = 140 vs. HA: Median ≠ 140
The following results for both the one-sample t-Test and the Signed-Rank
are provided.
• The test statistics (the t-Test and Signed-Rank).
• P-values for both one- and two-tailed tests. The p-value for the two-tailed
test is next to Prob > |t|.

Two-Sample and Oneway Nonparametric Tests

1. Select Analyze > Fit Y by X.
2. Select a continuous variable and click Y, Response, and select a
categorical variable and click X, Factor, then click OK.
The Oneway Analysis output window will display.
3. Under the red triangle, select Nonparametric > A test on Horsepower amount across two country levels
Wilcoxon Test / Kruskal-Wallis Test to perform the analysis adding
results to the report.
• If the categorical variable has only two levels, both the Normal
Approximation and ChiSquare Approximation test statistics and
corresponding p-values with be shown.
• If the variable has three or more levels, only the ChiSquare
Approximation will be performed. A test on Horsepower amount across three country levels

Note: the Wilcoxon / Kruskal-Wallis Rank Sum Test is sometimes

called the Mann-Whitney Test.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Test Mean, Basic Analysis > Oneway
Analysis > The Oneway Platform Options, Basic Analysis > Oneway Analysis > Oneway Analysis Reports to
learn more.
50
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Sample Size and Power for One Sample Mean

Use to interactively explore the relationships between Power, Sample Size, and Difference to Detect in testing a
hypothesis for a single population mean. See the One Sample t-test and Confidence Interval guide to learn how
to perform a statistical test for a population mean.

Sample Size and Power - One Sample Mean

1. Select DOE > Sample Size Explorers and choose Power > Power for One Sample Mean.
2. Choose the type of test: One-Sided or Two-Sided and Alpha (significance level for the test).
Here we consider the two-sided alternative hypothesis H0: µ = µ0 vs. HA: µ ≠ µ0
3. Choose if standard deviation is known.
If “Yes” analysis will be based on the Z-Test. If “No” analysis will be based on the t-Test.
4. Enter a value for the Standard Deviation. Here we assume 2.0.
5. Enter the Difference to Detect. This is the difference between a population mean under HA considered to be
true from the value of the population mean under H0. Here we consider a difference of 3.0, which is 3.0/2.0
= 1.5 standard deviations from the value of the population mean under H0.
6. Select parameter to solve for. Here we chose Sample Size.
7. Enter a value for the Power. Here we entered 0.80.
The solution (n=6) is shown in the Sample Size field.
Note: The Power field changed to 83.25% since sample size can only be an integer.
n=5 results in a Power of 71.1%.
8. Use the interactive cross-hair tool (or type in values) for Power, Sample Size, Difference to Detect, and Std
Dev to study the relationship between these parameters solving for many different scenarios.

The settings and solution for each analysis

performed can be saved. The table of saved
settings shows the results of five different
analyses performed when testing the
hypothesis H0: µ = µ0 vs. HA: µ ≠ µ0
1. What sample size is needed to achieve a
power of 80% at detecting a difference of 3/2
= 1.5 std dev? Answer: n = 6
2. What is the power with a sample size of 5 for
the above scenario? Answer: Power = 71.1%
3. What is the power for a sample size of 10 in
the above scenario? Answer: Power = 98.7%
4. What is the power for a sample size of 10 at
detecting a difference of 1/2 = 0.5 std dev?
Answer: Power = 29.3%
5. What sample size is needed to achieve a
power of 80% at detecting a .5 std dev
difference? Answer: n = 34

Note: Determining sample size to achieve a desired margin of error in a Confidence Interval can be done
using DOE > Sample Size Explorers > Confidence Intervals > Margin of Error for One Sample Mean.

Visit Design of Experiments Guide > Sample Size Explorers in JMP Help to learn more.

51
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Sample Size and Power for Two Sample Means

Use to interactively explore the relationships between Power, Sample Sizes, and Differences to Detect for
testing a hypothesis comparing two population means. See the Two Sample t-test and Confidence Interval
guide to learn how to perform a statistical test comparing two sample means.

Sample Size and Power - Two Sample Means

1. Select DOE > Sample Size Explorers and choose Power > Power for Two Independent Sample Means.
2. Choose the type of test: One-Sided or Two-Sided and choose Alpha (significance level for the test).
The null hypothesis is that the two means are equal. Here we chose a two-sided alternative which is used to
test that the two means are not equal.
This null and alternative hypothesis can be written using notation as H0: µ1 = µ 2 vs. HA: µ 1 ≠ µ 2
3. Choose if standard deviation is known.
If “Yes” analysis will be based on the Z-Test. If “No” analysis will be based on the t-Test.
4. Enter Assumed Standard Deviations. Here we assume 2 for each.
5. Enter the Difference to Detect. This is the difference between the two population means under HA
considered to be true in order to perform the analysis. Here we consider a difference of 2.5, which is
2.5/2.0 = 1.25 standard deviations apart.
6. Select parameter to solve for. Here we chose Total Sample Size.
7. Enter a value for the Power. Here we entered 0.80.
The solution of Total Sample Size of 23 (n1=11 and n2=12) is displayed.
8. Use the interactive cross-hair tool (or type in values) for Power, Sample Sizes,
Difference to Detect, and Std Devs to study the relationship between these
parameters solving for many different scenarios.

The settings and solution for each

analysis performed can be saved.
The table of saved settings shows
the results of three different
analyses performed when testing
the hypothesis H0: µ1 = µ 2 vs.
HA: µ 1 ≠ µ2
1. What sample size is needed to
achieve a power of 80% at
detecting a difference of 2.5/2
= 1.25 std dev?
Answer: 23 (n1 = 11 and n2 = 12)
2. What is the power with a
sample size of 20 in each
group? Answer: Power = 97.1%
3. What sample size is needed to
Note: Determining sample size to achieve a desired margin of error in a Confidence achieve a power of 80% at
detecting a difference of 4/2 = 2
Interval can be done using DOE > Sample Size Explorers > Confidence Intervals >
std dev?
Margin of Error for Two Independent Sample Means. Answer: 11 (n1 = 5 and n2 = 6)

Visit Design of Experiments Guide > Sample Size Explorers in JMP Help to learn more.

52
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Bootstrapping
This guide provides instructions on the bootstrapping technique – a resampling method for estimating the
sampling distribution of a statistic as a means to generate a confidence interval. Bootstrapping is available from
many JMP reports.
Big Class.jmp (Help > Sample Data Folder)
Bootstrapping in JMP Report Windows
1. From an analysis platform report window, right-click on the
report of interest and select Bootstrap.
In this example we use the Distribution platform and
bootstrap the statistics in the Summary Statistics report for
the continuous variable ‘Height’.
2. In the Bootstrapping window (below, left), enter the desired
number of bootstrapped samples and click OK.
JMP creates a data table (below, right) with statistics for the
original sample (excluded) and each of the bootstrap samples.
The BootID• column identifies the bootstrap sample number.

3. Use the Distribution platform to explore the statistics of interest for the bootstrap samples. Bootstrap
percentile confidence intervals for different confidence levels are provided.
Hint: Click on the green triangle next to Distribution in the top left panel of the Bootstrap Results data table
to launch the Distribution platform for all bootstrapped estimates.

Note: The Summary Statistics

table was customized (under
the Red Triangle) to only
display the Mean, Std Dev,
and N.

Visit Basic Analysis > Bootstrapping in JMP Help to learn more.

53
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Randomization Testing
This guide provides instructions on performing a randomization test (also known as a permutation test) – a
resampling method for estimating the sampling distribution of a statistic to generate a confidence interval and a
p-value for a hypothesis test. Randomization Testing is available from many JMP reports.
Big Class.jmp (Help > Sample Data Library)
Randomization Testing
Here, we describe how to conduct a randomization test for
two means using Fit Y by X.
1. From an open JMP data table, right-click on the column
header for the Nominal X variable (in this example, sex)
and select New Formula Column > Random > Sample
With Replacement.
This creates a new formula column, Resample[sex].
2. Conduct a 2-Sample t-Test using the Fit Y by X platform.
For this example, the Y, Response is Weight and the X, Factor
is sex. See the page: Two Sample t-Tests and CIs for
information on how to conduct this test and interpret results.
3. In the analysis report window, right-click over the statistic of
interest and select Simulate. Here we right-click on the column
of output containing the Difference (between means).
4. In the Simulation window, select the column to switch out (sex) and the
column to switch in (Resample[sex]), enter the desired number of samples
(1000, in this example), and the random seed (if desired), and click OK.
• JMP re-runs the analysis for each sample. For each iteration, the values
of the X, Factor (sex) are resampled with replacement.
• The results are stored in a data table with statistics for the original
sample and each of the resamples. The SimID• column identifies the
resample number.
5. Use the Distribution platform to explore the results for the statistics of interest. Confidence intervals for the
original estimate (the Difference, in this example) are provided, along with empirical (observed) p-values.

Interpretation: The empirical p-value for the two-tailed test is 0.3530. That is, 35.3% of the observed
resampled differences were as extreme or more extreme than the difference we actually observed (7.37).
The Randomization Testing Add-in available in the JMP User Community (community.jmp.com) provides a
tool to perform randomization tests for common hypothesis tests in JMP.

Visit Basic Analysis > Simulate in JMP Help to learn more.

54
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

Prediction Interval
Use to produce an interval estimate of a single observation, a sample of n observations, or the sample mean and
standard deviation of a sample of n observations.

Prediction Interval for an Individual Observation

Coating.jmp (Help > Sample Data Folder > Quality Control)
1. From an open JMP data table, select Analyze > Distribution.
2. Select one or more continuous variables from Select
Columns, click Y, Columns (continuous variables have blue
triangles), and click OK.
3. From the Distributions report window, select Prediction
Interval under the red triangle next to the variable name.
4. Select the confidence level (1-alpha) and enter 1 for number
of future samples. Choose Two-sided or One-sided interval.
Here we chose 95% confidence and two-sided.

JMP will add a table displaying the lower and upper bound
of the Prediction Interval for an individual observation.
Note: Calculations are based upon the assumption of a normal
distribution as an appropriate model for the population.

Interpretation:
It is estimated, with 95% confidence, that a randomly selected observation
from this population will be between 17.34 and 23.45.
Note: The prediction interval for the mean of n=1 is the same as that of an
individual observation and there is no prediction interval for the standard
deviation of n=1.

Prediction Intervals for a Sample of n Observations

1. Follow steps 1-4 above choosing a n>1 for number of future samples.
Here we chose 95% confidence, two-sided, and n=10.

JMP will add a table displaying the lower and upper bound
of the Prediction Interval for a sample of n=10 observations
as well as the sample mean and standard deviation.
Note: Calculations are based upon the assumption of a normal
distribution as an appropriate model for the population.

Interpretation:
It is estimated, with 95% confidence, that:
• 10 randomly selected observation from this population will all be between
15.87 and 24.92.
• The mean of the 10 observations will be between 19.31 and 21.49.
• The standard deviation of the 10 observations will be between 0.78 and 2.36.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Prediction Intervals in JMP Help to
learn more.
55
 Basic Inference – Proportions and Means
jmp.com/learn Nov 2022

One Sample Equivalence Test for Mean

Use to determine if there is statistical evidence exists to demonstrate that a population mean is within a
specified range (i.e., “equivalent”) to a hypothesized value.
Coating.jmp (Help > Sample Data Folder > Quality Control)
Equivalence Test for the Mean
1. From an open JMP data table, select Analyze >
Distribution.
2. Select one or more continuous variables from Select
Columns, click Y, Columns (continuous variables have
blue triangles), and click OK.
3. From the Distributions report window, select Test
Equivalence under the red triangle next to the variable name.
4. Specify the Hypothesized Mean, the Confidence Level,
and the difference from the hypothesized value considered
practically zero (i.e., Margin of Equivalence). Click OK.
Here we chose a hypothesized mean of 20, 95% Confidence
Level, and a margin of equivalence of +/- 1. Results with a Margin of Equivalence of +/- 1

The Null and Alternative Hypothesis in an Equivalence Test is stated as:

H0: µ - µ0 ≤ -1 or µ - µ0 ≥ 1 vs. HA: -1 < µ - µ0 < 1
* This approach is known as “Two One-sided t-tests” (TOST).
Notes:
• Rejecting the Null Hypothesis in favor of the Alternative
is essentially concluding there is enough statistical evidence to
believe the population mean is within +/- 1 of the hypothesized
value. In other words, “Is Equivalent”.
• The framework of an Equivalence Test reverses the roles of the
Null and Alternative Hypothesis in a traditional hypothesis test Results with a Margin of Equivalence of +/- 0.5
for the mean where the hypotheses are: H0: µ = µ0 vs. HA: µ ≠ µ0.
In the traditional test, if the data does not produce enough
statistical evidence to believe the Alterantive Hypothesis, it does
not mean that statistical evidence was produced to believe the
Null Hypothesis. Instead, the data simply did not produce
enough statistical evidence to reject it.

JMP will display a graph showing a (1 - 2a) Confidence Interval

comparing it to the Margin of Equivalence, and two separate
one-sided t-tests. The Null hypothesis is rejcted in favor of the
Alternarive when the largest p-value is less than a. When this is the case, the (1 - 2a) Confidence Interval
will be completely within the Margin of Equivalence.

• In the first analysis (Margin of Equivalence of +/- 1), the statistical evidence was produced to conclude equivalence.
• In the second analysis (Margin of Equivalence of +/- 0.5), the statistical evidence was NOT produced to conclude equivalence.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Test Equivalence in JMP Help to learn
more.
56
 Correlation and Regression
jmp.com/learn Nov 2022

Correlation
This guide illustrates ways to visualize the relationship between two continuous variables and quantify the linear
association via. pearson's correlation coefficient. For information on nonparametric correlations, see the
Nonparametric Correlations guide.
Body Measurements.jmp (Help > Sample Data Folder)
Correlation Between Two Variables
1. From an open JMP® data table, select Analyze > Fit Y by X.
2. Click on a continuous variable from Select Columns, and click Y,
Response (continuous variables have blue triangles).
3. Click on a second continuous variable, and click X, Factor.
4. Click OK to generate a scatterplot.
5. To display summary statistics including means, standard deviations,
correlation and a confidence interval for the correlation, select
Summary Statistics under the red triangle.
6. To display the density ellipse on the graph, click on the red triangle and
select the Density Ellipse > 0.95.

Correlations Between Multiple Pairs of Variables

1. From an open JMP data table, select Analyze > Multivariate Methods >
Multivariate.
2. Click on two or more continuous variables from Select Columns,
and click Y, Columns.
3. Click OK to produce a scatterplot matrix with density ellipses and
a table of correlations.
• The Default estimation method allows JMP to determine the
method for estimating correlations that is most appropriate for
your data set.
Tips:

• Many additional correlation options are available under the red

triangle next to Multivariate, including:
- CI of Correlations.
- Inverse Correlations.
- Partial Correlations.
- Pairwise Correlations (Pearson product-moment).
- Nonparametric Correlations (including Spearman’s rho).
• Scatterplot options are available under the red triangle next to
Scatterplot Matrix.

Note: Density ellipses can also be generated from Graph > Scatterplot Matrix and Graph > Graph Builder.

Visit Basic Analysis > Bivariate Analysis and Multivariate Methods > Correlations and Multivariate Techniques
in JMP Help to learn more.
57
 Correlation and Regression
jmp.com/learn Nov 2022

Nonparametric Correlations
This guide illustrates how to compute nonparametric measures of association (Spearman’s Rho, Kendall’s Tau,
and Hoeffding’s D).

Nonparametric Correlations
Socioeconomic.jmp (Help > Sample Data Folder)
1. From an open JMP data table, select Analyze > Multivariate
Methods > Multivariate.
2. Select two or more discrete numeric (nominal or ordinal) or
continuous variables from Select Columns, click Y, Columns, then
click OK.
JMP produces a table of pairwise correlations and a scatterplot
matrix.
3. From the top red triangle, select Nonparametric Correlation, then
the measure of interest (shown below, left).
The following results are provided (below, right):
• The calculated correlation coefficient for the pair of variables.
• The p-value, showing the significance of the correlation.
• A bar chart showing the correlation coefficients.

Spearman’s Rho (r) is similar to Pearson’s correlation, but is based on ranks rather than the original values.
Like the Pearson correlation, values range from -1 to +1, with larger absolute values indicating a stronger
relationship.
Kendall’s Tau (T) is based on the number of concordant and discordant pairs of rank-ordered data. It also
ranges from -1 to +1.
Hoeffding’s D ranges from -.5 to 1. It measures the difference between the joint ranks of paired data and
the product of their marginal ranks, and can capture nonlinear relationships.

Notes: Additional options are available under the top red triangle (shown above, left). Nonparametric Density
and other options are also available under the red triangle for the scatterplot matrix.

Visit Multivariate Methods > Correlations and Multivariate Techniques in JMP Help to learn more.

58
 Correlation and Regression
jmp.com/learn Nov 2022

Simple Linear Regression

Use to model the bivariate relationship between a continuous explanatory variable with a continuous outcome
variable. Useful to describe the relationship between the variables and to predict an outcome for different
values of the explanatory variable.
Big Class.jmp (Help > Sample Data Folder)

Simple Linear Regression Using Fit Y by X

1. From an open JMP® data table, select Analyze > Fit Y by X.
2. Click on a continuous variable from Select Columns, and
click Y, Response (continuous variables have blue triangles).
3. Select a second continuous variable, and click X, Factor.
4. Click OK to generate a scatterplot.
5. To fit a regression line, click on the red triangle and select
Fit Line.

By default, JMP will provide the following results:

• The regression equation (under Linear Fit).

• The Summary of Fit.
• Lack of Fit (if the data table includes replicates of X
values).
• The ANOVA table.
• The parameter estimates.

• For other fit options, such as polynomials, transformations

(fit special) and spline (under flexible), use the top red
triangle.
• To add a legend, change markers, or make other changes to
the graphical display, right-click on the graph.
• Double-click on the axes to change axes settings.
• Other analysis steps common in linear regression (e.g.,
plotting residuals, adding confidence bands, saving
predicted values and residuals to the data table, among
others) can be found under the red triangle next to the title
of the equation fit (e.g., “Linear Fit”).
• To fit separate lines for categories of a grouping variable,
click on the top red triangle, select Group By, and choose a
grouping variable. Then, click on the top red triangle again
and select Fit Line. JMP will fit separate lines and provide
results for each level of the grouping variable.

Visit Discovering JMP > Analyze Your Data > Analyze Relationships and Basic Analysis > Bivariate Analysis in
JMP Help to learn more.

59
 Correlation and Regression
jmp.com/learn Nov 2022

Multiple Linear Regression

Use to model the relationship two or more continuous or categorical explanatory explanatory variables has with
a continuous outcome variable. Useful to describe the relationships between the variables and to predict an
outcome for different values of the explanatory variables.
Big Class.jmp (Help > Sample Data Folder)

Multiple Linear Regression Using Fit Model

1. From an open JMP® data table, select Analyze > Fit Model.
2. Click on a continuous variable from Select Columns, and click
Y (continuous variables have blue triangles).
3. Choose explanatory variables from Select Columns, and click
Add. To add an interaction term, select the variables and
click Cross. Here we included the sex*height interaction.
4. Click Run.
By default, JMP will provide the following results:

• Effect Summary
• Actual by Predicted Plot.
• Summary of Fit table.
• Analysis of Variance table.
• Parameter Estimates table, and more (not shown).
JMP also provides Leverage Plots for each explanatory variable in the
model, and for nominal and ordinal variables, the least squares
means tables.

• Select Estimates > Show Predicted Expression under the top red
triangle to display an equation for the fitted model.
• To save the prediction formula, predicted values, residuals and other
statistics to the data table, click on the top red triangle, select Save
Columns. JMP will create new columns in the data table.
• Model Diagnostic plots can be found under Row Diagnostics.
• To view indicator parameterization (using 0, 1 coding), select
Estimates > Indicator Parameterization Estimates from the
top red triangle.
• To view the effects of the explanatory variables on the
predicted response, click on the top red triangle,
select Factor Profiling and choose Profiler. In the
Prediction Profiler, click and drag the vertical red line
for a variable to change the level or value. The
predicted mean response and 95% Confidence
Interval are displayed.

Visit Discovering JMP > Analyze Your Data > Analyze Relationships and Fitting Linear Models in JMP Help to
learn more.

60
 Correlation and Regression
jmp.com/learn Nov 2022

Simple Logistic Regression

Use to model the relationship a continuous explanatory variable has with a categorical outcome variable. Useful
for estimating the probability of the occurrence of an event for different values of the explanatory variable.
Car Poll.jmp (Help > Sample Data Folder)
Logistic Regression Using Fit Y by X
1. From an open JMP® data table, select Analyze > Fit Y by X.
2. Click on a categorical variable from Select Columns, and click
Y, Response (nominal variables have red bars, ordinal
variables have green bars).
3. Click on a continuous variable, and click X, Factor
(continuous variables have blue triangles). Click OK.

By default, JMP will provide the following results:

• The logistic plot, with curves of cumulative predicted
(fitted) probabilities.
• The whole model test for model significance.
• Parameter estimates for the fitted model.

Tips:

• When the response is nominal, a nominal logistic model will

be fit. When the response is ordinal, as in this example, an
ordinal logistic model will be fit.
• To color points and add a legend, right-click in the graph and
select Row Legend. Select a variable under Mark by Column,
and select Markers to change the marker, and click OK.
• To save the probability formula or request other options, click on the top red triangle and select the option.
• To find the fitted probability for a given value of X, select the cross-hair tool ( ) from the toolbar. Click on
the graph dragging the cross-hair to the desired point on the curve.

Interpretation (for this example, X = buying age and Y = car size):

• The bottom curve represents the predicted probability
that for a given age, someone will buy a large car.
• The second curve represents the probability that
someone will buy a large or medium car.
• The distance between the two curves represents the
probability that someone will buy a medium car.
• The distance between 1.00 and the top curve represents
the probability that someone will buy a small car.
• The cross-hairs show that the predicted probability that
someone aged 39 years will purchase a large car is 0.191.

Notes: Simple nominal and ordinal logistic regression can also be performed from Analyze > Fit Model.
Visit Basic Analysis > Logistic Analysis in JMP Help to learn more.

61
 Correlation and Regression
jmp.com/learn Nov 2022

Multiple Logistic Regression

Use to model the relationship two or more continuous or categorical explanatory variables has with a
categorical outcome variable. Useful for estimating the probability of the occurrence of an event for different
values of the explanatory variables.
Car Poll.jmp (Help > Sample Data Folder)

Multiple Logistic Regression Using Fit Model

1. From an open JMP® data table, select Analyze > Fit Model.
2. Click on a categorical variable from Select Columns, and
click Y (nominal variables have red bars, ordinal variables
have green bars).
3. Choose explanatory variables from Select Columns,
and click Add.
4. Click Run Model.
By default, JMP will provide the following results:
• The Iterations history (not shown).
• The Whole Model Test.
• Lack of Fit (not shown).
• Parameter Estimates for the model.
• Effect Likelihood Ratio Tests (not shown).

Tips:

• When the response is ordinal, an ordinal logistic model will be

fit. When the response is nominal, as in this example, a
nominal logistic model will be fit.
• To save the predicted probabilities to the data table, click on
the top red triangle, select Save Probability Formula.
• To fit a model for grouped or summarized data, use Freq in
the Fit Model Specification window - specify the variable that
contains the frequency (count) for each level of the response.

• To view the effect of an explanatory variable on the

predicted probabilities, click on the top red triangle and
select Profiler.
In the Prediction Profiler, click and drag the vertical red line
for a variable to change the level or value. The predicted
probabilities are displayed.

Visit Fitting Linear Models > Logistic Regression Models in JMP Help to learn more.

62
 Correlation and Regression
jmp.com/learn Nov 2022

Stepwise Regression
Use to perform automated variable selection in multiple linear or logistic regression models. The method is
particular useful when there is a large number of candidate explanatory variables.

Stepwise Regression
Car Physical Data.jmp (Help > Sample Data Library)
1. From an open table, select Analyze > Fit Model.
2. Select a response variable from Select Columns and click Y.
3. Select predictor variables and click Add.
4. If desired, select a validation column (JMP® Pro only).
5. Select Stepwise from the Personality drop-down menu.
6. In the resulting Stepwise Fit window (shown below):
• Select a Stopping Rule.
• Select the step Direction (forward, backward or mixed).
• To run the regression automatically, click Go. To proceed
manually, click Step.

Click the red triangle for cross-

validation, all possible models
(all subsets regression), model
Use arrows to stew
averaging and other options.
forward/backward one
Choose Stopping Rule and term at a time.
Direction of adding/removing
terms. Select Go to run the
full stepwise routine
Results for current model. according to the
chosen Direction and
Check/uncheck Entered terms Stopping Rule.
to change the model. Locked
terms are used (or not used) in P-values are displayed
later steps. under Prob>F.

Each time the model is

changed, a new line is added
to the Step History panel.

• For Forward regression, remove all terms, then click Step or Go.
• For Backward regression, enter all terms, then click Step or Go.
• The Mixed direction is only available with the p-value stopping rule.
• To run the model shown in the Current Estimates table, click Run Model.
JMP generates the Fit Model report.
• See the Multiple Linear Regression or Multiple Logistic Regression one-page guides for more details.

Visit Fitting Linear Models > Stepwise Regression Models in JMP Help to learn more.

63
 Correlation and Regression
jmp.com/learn Nov 2022

Fit Non-Linear Curve

Use to build non-linear models describing the relationship between an explanatory variable and an outcome
variable.

Fit Curve
Algae Mitscherlich.jmp (Help > Sample Data Folder > Nonlinear Examples)
1. Select Analyze > Specialized Modeling > Fit Curve.
2. Select a continuous variable from Select Columns, and
add to Y, Response.
3. Select a continuous explanatory variable for X, Regressor
Add a categorical variable to Group to have a separate
model fit for each value of a grouping variable (optional).
Here we have eight specimens sampled (two from each of
four treatments).
4. Click OK.
JMP displays a plot of the response vs. predictor values
for all the data, and separate graphs for each group if a
grouping variable was used.
Note: The red triangle next to Fit Curve displays a
collection of non-linear models that can be fit.
See “Statistical Details for the Fit Curve Platform” in JMP
Help to see visual representations for different families of
nonlinear models to help choose an appropriate one for
your data.
The 4-parameter logistic sigmoid model was chosen here.
Additional models can be fit and compared.
5. Select Custom Inverse Prediction under the red triangle.
Insert values for the response variable (Algae Density) to
produce predictions and corresponding confidence
intervals for the value of the predictor variable (Days)
estimated to result in a given amount of the response
variable. Here we chose an Algae Density of 2.
JMP displays predicted values and confidence intervals
for the explanatory variable. An Analysis of Means graph
comparing these predictions across groups is produced if
a grouping variable was used.
Additonal options, such as Actual by Predicted and
Residual by Predicted plots, Save Prediction and Residual
Formula, are under the red triangle for each fitted model.

Visit Predictive and Specialized Modeling > Fit Curve and Nonlinear Regression in JMP Help to learn more.

64
 Time Series
jmp.com/learn Nov 2022

ARIMA Modeling
Use ARIMA (Auto Regressive Integrated Moving Average) time series models to examine autocorrelation,
describe patterns (trends and seasonality), and forecast future time periods.

ARIMA Modeling Steel Shipments.jmp (Help > Sample Data Folder > Time Series)

1. From an open JMP® data table, select Analyze > Specialized

Modeling > Time Series.
2. Select a continuous variable from Select Columns, and click Y,
Time Series (continuous variables have blue triangles).
3. Select a time and click X, Time ID (optional). Click OK.
Note: Data must be sorted by time and equally spaced. If no time
variable is used, JMP will assume equal spacing.
The autocorrelation (ACF) and partial autocorrelation (PACF)
plots suggest an ARIMA model with a seasonal component of AR
(1) and a non-seasonal component of AR (2).
4. Click on the top red triangle and select Seasonal ARIMA. Enter
the values as shown (right), and click Estimate. JMP displays
model results.
5. For the fitted model, check the ACF, PACF and Residual plots to
determine if a different model should be fit.
Here, we repeat Step 4 adding the “3” to q, Moving Average
Order under ARIMA.
JMP provides a Model Comparison report (shown below), which indicates that the new model fits the data
better (according to criteria such as AIC and SBC). Click and drag the slider bar at the bottom of the report to
see all of the statistics.

• To simultaneously fit a range of ARIMA or Seasonal ARIMA models, select ARIMA Model Group from the top
red triangle.
• Other options, such as Variogram, Spectral Density, Difference, Smoothing Models and Number of Forecast
Periods are available under the top red triangle.
• Use the red triangle for a model to save a forecast, create SAS® job (PROC ARIMA), and submit to SAS
(requires an active SAS connection).
• ARIMA models require that the time series be stationary. If the series has a trend over time, differencing will
remove the trend. If the series has a non-stationary variance, taking the log of the series may help.
• To forecast the time series with input variables, use a Transfer Function (use the Input List field in the Time
Series dialog window). Transfer function models are also referred to as ARIMA models with Input Series.

Visit Predictive and Specialized Models > Time Series Analysis in JMP Help to learn more.

65
 Time Series
jmp.com/learn Nov 2022

Time Series Smoothing Models

Use smoothing based time series models to describe patterns and forecast future time periods.

Smoothing Models
1. From an open JMP® data table, select Analyze > Workers.jmp (Help > Sample Data Folder > Time Series)

Specialized Modeling > Time Series.

2. Select a continuous variable from Select Columns, and
click Y, Time Series (continuous variables have blue
triangles).
3. Select a time and click X, Time ID (optional). Click OK.
- A time series graph of the data along with a graph
and table of the Auto and Partial Correlation values
is displayed.
4. To fit a model to the data, click on the top red triangle,
select Smoothing Model and choose the method.
Here we choose Sesonal Exponential Smoothing.
5. Choose Observations per Period (e.g., 12) and
click Estimate.
- JMP displays the Model Summary, Parameter Estimates
and a Forecast plot that shows the fit and forecasts from
the model.
6. To estimate an alternative smoothing model, click on the
red triangle, and choose another method (Winters
Method was selected for a second model).
7. Click Estimate.
- JMP provides a Model Comparison report, (shown
below) that compares the two methods. Click and drag
the slider bar at the bottom of the report to see all of the
statistics.
In this example, the Seasonal Exponential Smoothing model fits the data better than Winters Method
(i.e., AIC, SBC, MAPE, and MAE are smaller).

• The default number of forecast periods is 25. To change, enter a different value in the Time Series launch
dialog window.
• To save the forecast, select Save Columns or Save Prediction Formula under the red triangle for that model.
A new table with the actual and predicted values will be generated.

Visit Predictive and Specialized Models > Time Series Analysis in JMP Help to learn more.

66
 Time Series
jmp.com/learn Nov 2022

Time Series Forecasting

The Time Series Forecast platform builds a variety of different exponential smoothing models and automatically
selects the with the best forecast performance. The platform is designed to forecast multiple time series.

Time Series Forecast Monthly Sales.jmp (Help > Sample Data Folder > Time Series)

1. From an open JMP® data table, select Analyze >

Specialized Modeling > Time Series Forecast.

2. Select a continuous variable from Select Columns,

and click Y (continuous variables have blue triangles).
- More than one times series can be included in the Y
role and models and forecasts will be created for each.
Here we illustrate with just one time series (Sales).

3. Select the variable that identifies the time periods and

click Time (optional).
- Data must be sorted by time and equally spaced.
4. Click OK.
In Model Specifications you can view the recommended
models or modify the modeling options.
Under Complete Specification, you can choose:
• Time periods to forecast (NAhead)
• Period for seasonality (e.g., monthly)
• Model Selection Strategy (AIC, BIC, or Forecasting
Performance)
• Holdback sample and metric to assess fit
(if chooseing Forecast Performance).

5. Click Run

- Best fit time series model is displayed

(e.g., MAA: Multiplicative Error, Additive Trend,
No Dampening, Additive Seasonality).
- Time Series graph displays data, best fit model,
forecasting intervals, and future forecast as defined
by NAhead.

6. Choose to Save Results to the original data table

or to a new data table from the red triangle at the top
of the Report

Visit Predictive and Specialized Models > Time Series Forecast in JMP Help to learn more.

67
 Multivariate Methods
jmp.com/learn Nov 2022

Clustering
Use Hierarchical or K-Means Clustering to form clusters (groups) of observations having similar characteristics.
Cereal.jmp (Help > Sample Data Folder)
Hierarchical Clustering
1. From an open JMP® data table, select Analyze > Clustering >
Hierarchical Cluster.
2. Select one or more numeric variables from Select Columns and
click Y, Columns. Here we used the 13 numeric variables.
3. If available, select a Label variable.
4. Select the desired method (bottom left corner) and click OK.

...
...

...
JMP will generate:

• A dendrogram, showing the clusters formed at each step.

• A scree plot, showing the distance bridged each step.
• The clustering history, giving cluster statistics for each step.

Tips:
• To color clusters, to mark or save clusters, or to request other options, click the top red triangle.
• To dynamically change the number of clusters, click and drag one of the black diamonds left or right.

K-Means Clustering
1. From an open JMP data table, select Analyze > Clustering > K Means
Cluster.
2. Select one or more numeric variables from Select Columns and click Y,
Columns. Here we used 13 numeric variables. Click OK.
3. In the resulting Control Panel, choose K Means Cluster Under Method.
4. Enter the number of clusters. Click Go. Here we chose 3.
JMP will generate:
• A summary of the cluster sizes.
• Tables of cluster means and standard deviations for each variable.
Tips:

• To obtain biplots, parallel plots or request other options, click the red
triangle for the K Means heading.
• To perform analyses for a range of cluster sizes: In the Control Panel, enter the lower limit in number of
clusters and the upper limit in range of clusters, then click Go.
• To step through the formation of the clusters: In the Control Panel, check Single Step then click Go.

Visit Multivariate Methods > Hierarchical Cluster and K Means Cluster in JMP Help to learn more.

68
 Multivariate Methods
jmp.com/learn Nov 2022

Principal Component Analysis

This guide provides instructions on performing a principal component analysis (PCA). This analysis method is
often used to reduce the dimensionality of a data set (i.e., fewer variables) by creating a new set of variables
that are linear combinations of the original variables statistically independent of each other and that capture the
most information (i.e., variation and correlation) contained in the original variables.

Principal Components
Socioeconomic.jmp (Help > Sample Data Folder)
1. From an open JMP® data table, select Analyze >
Multivariate Methods > Principal Components.
2. Select continuous variables from Select Columns, and
Click Y, Columns (continuous variables have blue triangles).
3. Click OK.
By default, JMP displays the eigenvalues and three
Summary Plots.

• Eigenvalue Pareto Plot: The percent and

cumulative total percent of the variation
accounted for by each principal component.
• Score Plot (middle): A scatterplot of the first
two principal components. Use the options
at the bottom to select any two components
to display.
• Loading Plot: Correlations between the
original variables and the two selected
principal components. (Note: The factor loadings are unrotated.)

Interpretation:

• The first two principal components account for 93.4% (57.5 + 35.9 = 93.4) of the total variation in the data (see the Pareto
Plot). These numbers are displayed on the graph axes of the Score Plot and Loading Plot.
• All of the original variables are positively correlated with the first principal component (see the Loading Plot). Total
Population and Total Employment are positively correlated with the second principal component, while the other variables
are negatively correlated with the second principal component.

Tips:
• By default, PCA is performed on correlations.
• Click on the top red triangle to change the method of calculation, view additional results, save the principal
components to the data table, or view detailed information associated with the eigenvalues.
• Principal component analysis can also be accessed through the Scatterplot 3D platform or the Multivariate
platform.

Visit Multivariate Methods > Principal Components in JMP Help to learn more.

69
 Multivariate Methods
jmp.com/learn Nov 2022

Factor Analysis
Factor Analysis is an analysis technique that seeks to describe the variation in a set of observed variables in
terms of a smaller number of unobserved latent variables or factors.

Factor Analysis Socioeconomic.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Multivariate

Methods > Factor Analysis.
2. Select continuous variables from Select Columns, and Click
Y, Columns (continuous variables have blue triangles).
3. Click OK.
JMP displays an Eigenvalue report and a Scree Plot, along
with a Model Launch dialog.
4. Click Go in the Model Launch dialog to run the factor analysis
at the default settings.
The following results are provided (additional options are
available under the red triangles):

• Final communality estimates

• The variance explained by each factor
• Significance tests for the factor analysis
• Rotated factor loadings
• Factor loading plot

Interpretation of results:

• Factor 1 seems to be related to socioeconomic status,

as Median House Value and Median School Years have
large loadings on the factor.
• Factor 2 seems related to the labor force condition, as
Total Population and Total Employment load highly on
this factor.
• Professional Services loads on both factors.
• Factor 1 and Factor 2 account for 47.8% and 42.4% of
the total sample variation, respectively.

Tips:

• The default factoring method is Maximum Likelihood, the

number of factors to extract is 2, and the rotation method is
Varimax. These options can be changed in the Model Launch
dialog (shown above).
• To save the factor scores as new columns to the data table,
click on the red triangle and select Save Rotated Components.

Visit Multivariate Methods > Factor Analysis in JMP Help to learn more.

70
 Mixed Models and Repeated Measures
jmp.com/learn Nov 2022

Analysis of Repeated Measures (MANOVA)

Use MANOVA (multivariate analysis of variance) for a way to analyze repeated measures data. The term
repeated measures refers to data with multiple measurements taken on the same subjects, often taken over a
period of time. The MANOVA platform provides tests of between and within subject effects across the repeated
measurements.
This example involves 16 dogs assigned to different treatment groups. Blood concentration of histamine is
measured at four points in time. The data are arranged in a wide format (log-histamine measures are in separate
columns), which is needed for the MANOVA analysis.

Analysis of Repeated Measures: MANOVA Dogs.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Fit

Model. Select Manova from the Personality drop-
down menu.
2. Add the responses: From Select Columns, select the
continuous response variables (continuous variables
have blue triangles), and click Y.
3. Add model effects: Select variables and click Add
(under Construct Model Effects). To specify an
interaction term, select multiple columns, then click
Cross. Click Run.
4. In the resulting window, select Repeated Measures from the Choose Response
drop-down menu.
5. Accept the defaults, then click OK.

By default, JMP will display the Parameter Estimates, Least Squares Means, Between
Subjects and Within Subjects results and more. Additional options are available
under the red triangles.
• Within Subjects results include
the multivariate significance tests
for the differences over time for
both the whole model and each
effect.

• Between Subjects results include

the multivariate significance tests
for the differences in the
repeated measures across
subjects for both the whole
model and each effect.

See the Repeated Measures Analysis (Mixed Model) Guide to learn how to analyze repeated measures data
using the mixed models platform.
Visit Fitting Linear Models > Multivariate Response Models in JMP Help to learn more.

71
 Mixed Models and Repeated Measures
jmp.com/learn Nov 2022

Repeated Measures Analysis (Mixed Model)

This guide provides instructions on the analysis of repeated measures data using a mixed model (random and
fixed effects) with nesting. The term repeated measures refers to data with multiple measurements taken on the
same subjects, often taken over a period of time.
This example involves six animal subjects randomly selected from two species. The miles traveled by each
animal were measured over time. Since this data is in a tall format (stacked), a mixed model analysis is used.

Analysis of Repeated Measures: Mixed Model

Example: Animals.jmp (Help > Sample Data Folder)
®
1. From an open JMP data table, select Analyze > Fit Model.
2. Add the response: From Select Columns, select a
continuous variable (continuous variables have blue
triangles), and click Y.
3. Add model effects: Select variables and click Add (under
Construct Model Effects). To specify an interaction term,
select multiple columns, then click Cross.
4. Specify the nesting structure: Here, subject is nested
within species. Select ‘subject’ from Construct Model
Effects, select ‘species’ from Select Columns, and click
Nest. (If the subject ID is uniquely valued, skip this step.)
5. Specify random effect(s): Select a model effect, then select
Random from the red triangle next to Attributes.
Here, “subject[species]” is specified as a random effect.
6. Accept the defaults (the REML Method with Unbounded
Variance Components selected), and click Run.
By default, JMP will display tables of the Effect Summary,
Summary of Fit, REML Variance Components Estimates,
Fixed Effect Tests and more. Additional options are
available under the top red triangle.

Interpretation:

1. Variance Components Estimates: Show the estimated variances

for random effects and the residual error. In this example, the
estimated variation between animals is 0.76, or 38% of the total
variation.
2. Fixed Effect Tests: Show the F-test results for the fixed effects
using the appropriate variance component estimate as an error
term in the denominator. Here, species and season are both
significant at alpha = 0.05, while the interaction is not.

See the Repeated Measures Analysis (ANOVA) Guide to learn how to analyze repeated measures data using the
MANOVA platform.
Visit Fitting Linear Models > Mixed Models in JMP Help to learn more.

72
 Mixed Models and Repeated Measures
jmp.com/learn Nov 2022

Mixed Model Analysis

Use a Mixed Model for an ANOVA or regression model with at least one factor specified as a random variable.
JMP Pro® has a Mixed Model platform offering the most flexibility in fitting mixed models.
This example uses standard JMP to fit an unbalanced design involving six people chosen at random to take
measurements on three different machines

Analysis of Linear Mixed Models

Machine.jmp (Help > Sample Data Folder)
1. From an open JMP data table, select Analyze > Fit Model.

In JMP Pro, select the Mixed platform from the

Personality list to fit mixed models.
The remainder of these instructions are for fitting mixed
models in standard JMP.
2. Add the response: From Select Columns, select a continuous
variable (continuous variables have blue triangles), and click Y.
3. Add model effects: Select variables and click Add (under
Construct Model Effects). To specify an interaction term,
select multiple columns, then click Cross.
4. Specify random effect(s): Select a model effect, then select Random from the red triangle next to Attribute.
Here, “person” is specified as a random effect, so the “person*machine” interaction is also a random effect.
5. Accept the defaults (the REML Method with Unbounded Variance Components selected), and click Run.
By default, JMP will display the Summary of Fit table,
REML Variance Components Estimates, and more.
Additional options are available under the top red triangle.

Interpretation:

1. Variance Components Estimates: The total estimated

variance is 37.56, which can be decomposed into three
sources: person (22.46, or 59.8%), person*machine
(14.23, or 37.9%) and residual (0.87, or 2.3%).
2. Fixed Effect Tests: The p-value for the fixed effect
(machine) is 0.0003. Kenward-Roger F tests are used.

Tips:
• REML (restricted maximum likelihood) is the preferred estimation method over EMS (Method of Moments).
REML estimates are properly shrunk and the standard errors are properly scaled.
• Unbounded Variance Components is the default method for estimating the variance components.
Unchecking this box will restrict the variance estimates to be non-negative.
• JMP assumes a simple correlation matrix with compound symmetry (i.e., correlation is constant).

Visit Fitting Linear Models > Mixed Models and Generalized Linear Mixed Models in JMP Help to learn more.

73
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Classification Trees (Partition)

Use to build a partition-based model (Decision Tree) that identifies the most important factors that predict a
categorical outcome (classify) and use the resulting tree to make predictions for new observations.

Classification Trees
Auto Raw Data.jmp (Help > Sample Data Folder)
1. From an open JMP® table, select Analyze > Predictive Modeling >
Partition.
2. Select a nominal or ordinal response variable from Select Columns
and click Y, Response.
3. Select explanatory variables and click X, Factor.
4. If desired, enter the Validation Portion or select a validation column
and click Validation (JMP Pro only). A randomly selected validation
set of 30% was used in this illustration. Thus 70% for training.
5. In JMP Pro only, select the tree Method: Decision Tree (the default),
Bootstrap Forest, Boosted Tree, K Nearest Neighbors or Naive Bayes.
6. Click OK.
JMP initially displays a graph showing the proportion of observations
in each response level.
7. Click the Split button. The observations will be split into two nodes, or
leaves. The graph will update to reflect the split and a tree diagram
describing the split in more detail will be created.
Note: Click on the top red triangle and select Display Options > Show
Split Counts to show Counts, Rates (proportion of observations) and
Probs (predicted probabilities) in each leaf.

8. Click Split to continue additional splits. Click Prune to remove a split.

If a validation set is used, click Go to perform automatic splitting and
pruning optimizing the fit on the validation data.
Interpretation for the first two splits (Response is Claim (Y/N)):

• There are 1,179 obs in the left leaf, corresponding to AgeClass(Young). 768 of
those (65.1%) are Claim(Y/N) = N and 411 (34.8%) are Claim(Y/N) = Y.
• There are 12,758 obs in the right leaf, corresponding to AgeClass(Elder). The
response rate is 90.2% for N and 9.8% for Y in that node.
• For the 12,758 observations in the AgeClass(Elder), the second split, based
on the Rating Class variable, is at Rating Class (D,C) and Rating Class (B, A).

Notes:
For additional options, such as Column Contributions, ROC and Lift Curves, click the top red triangle. Other
options, such as Save Prediction Formula and Make SAS® DATA Step, are available from the top red triangle >
Save Columns.
Select Decision Threshold under the top red triangle to display correct and incorrect classification rates for the
model including the ability to evaluate those rates under different cutoff values. The default is 50%.
Visit Predictive and Specialized Models > Partition Models in JMP Help to learn more.

74
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Regression Trees (Partition)

Use to build a partition-based model (Decision Tree) that identify the most important factors that predict a
continuous outcome and use the resulting tree to make prediction for new observations.

Regression Trees Diamonds Data.jmp (Help > Sample Data Folder)

1. From an open JMP® table, select Analyze > Predictive

Modeling > Partition.
2. Select a continuous response variable from Select Columns
and click Y, Response.
3. Select explanatory variables and click X, Factor.
4. If desired, enter the Validation Portion or select a validation
column and click Validation (JMP Pro only). A validation set
was not used in this illustration.
5. In JMP Pro only, select the tree Method: Decision Tree (the
default), Bootstrap Forest, Boosted Tree, K Nearest Neighbors, or Naive Bayes.
6. Click OK. JMP initially displays a graph of showing the response for all the observations and a line drawn at
the overall mean ($3,971 in this example).
7. Click the Split button. The original observations will be split
into two nodes, or leaves.
In the top graph, horizontal lines are drawn at the mean
response within each leaf and vertical lines depict the leaf’s
relative size.

Interpretation for the first two splits (Response is Diamond Price in $):

• There are 1,502 obs with Carat Weight < 0.95. The mean price of
these obs is $2,281.
• There are 1,188 obs with Carat Weight >= 0.95. The mean price of
these obs is $6,109.
• For the 1,502 obs in with Carat Weight <0.95, the second split, also
based on the Carat Weight variable, is at a Carat Weight of 0.74.

8. Click Split to make an additional split. Click Prune to remove

a split. If a validation portion or validation column is used,
click Go to perform automatic splitting and pruning
optimizing the fit on the validation data.

Notes:
For additional options, such as Leaf Report, Small Tree View, Column Contributions, click the top red triangle.
Other options, such as Save Prediction formula and Make SAS® DATA Step, are available from the top red
triangle > Save Columns. For split options for a particular node, click on the red triangle for that node.
Visit Predictive and Specialized Models > Partition Models in JMP Help to learn more.

75
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Discriminant Analysis
Build a boundary based statistical model to predict a categorical outcome (classify) as a function of multiple
continuous preditor variables.

Discriminant Analysis Iris.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Multivariate Methods > Discriminant.
2. Select one or more continuous variables from Select Columns, and click Y, Covariates (continuous variables
have blue triangles).
3. Click on a categorical variable from Select Columns, and click X, Categories (nominal variables have red bars,
ordinal variables have green bars).
4. Click OK.
By default, JMP displays the Canonical Plot and
Discriminant Scores.
• The Canonical Plot shows the points and
multivariate least-squares means on the first
two canonical variables that best separate the
groups.
• The Biplot Rays on the Canonical Plot indicate
the directions of the predictors in the canonical
space.
• The Discriminant Scores report shows information used to classify
each row in the data table.
• The Score Summaries report provides a summary of the
misclassifications and tables that tabulates the number and percent
of correctly and incorrectly classified cases.

Tips:
• JMP provides Stepwise Variable Selection and three Discriminant
Methods (Linear, Quadratic and Regularized).
• Click on the red triangle to select Stepwise Variable Selection,
change the discriminant method, show canonical details, specify prior
probabilities, save results, customize plots or select other options.
• If a validation column is specified in the model dialog, the Score
Summaries table will include counts and misclassification rates for the
training, validation (and test) partitions.

Visit Multivariate Methods > Discriminant Analysis in JMP Help to learn more.

76
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Support Vector Machines - Classification

Build a boundary based statistical model to predict a categorical outcome (classify) as a function of
multiple predictor variables. SVM is able to create much more flexible boundary shapes than the Classification
Tree (Partition) and Discriminant Analysis method.

Support Vector Machines

Equity.jmp (Help > Sample Data Library)

1. From an open JMP® table, select Analyze > Predictive Modeling >
Support Vector Machines.
2. Add a nominal or ordinal response variable from Select Columns to
the Y, Response role.
3. Add candidate predictor variables to the X, Factor role.
4. If desired, enter a validation column into the Validation role as
shown in this example. Click OK.
5. The Model Launch control panel opens allowing a choice of a Kernal
Function and associated options. Default settings were used
for this example. Click Go.
JMP displays:

• Response Profile Plot displaying the classification regions and

the data values for two of the predictor variables. These can be
changed to other variables by selecting the red triangle next to
the variable name on the axes. Levels for all the remaining
predictors can be changed with the sliders above the plot.
• Model Summary (not shown) and a Confusion Matrix detailing
the classification performance.

Intepretation:

• There are 649 observations in the Vaidation Data. Of these, 45 (6.9%) where
misclassified. 45/(45+18) = 71% of the Bad Risk customers were misclassified
as Good Risk. 0/(0+586) = 0% of the Good Risk customers were misclassified as
Bad Risk.

Note: The default rule is to classify an observation in the class with the highest
estimated probability of being in that class (i.e., Prob > 0.50). It is advantageous
to evaluate different cutoff values in order to minimize a specific type of
misclassificaiton rate over another. In this example it would be much better to
choose a cutoff level to create a lower misclassification rate for Bad Risk
Customers misclassified as Good Risk while accepting a higher misclassification
rate for the other. This analysis can be performed by saving the predicted
probabilities to the data table and using the calculator tool to create a
conditional argument to conclude an outcome based on those predicted
probability values.

Additional options, such as ROC and Lift Curves, Profilers, Save Predicteds, Save Prediction Formula, Save
Probabilities, as well as Publish Probability Formulas are accessible from the red triangle.
Visit Predictive and Specialized Models > Support Vector Machines in JMP Help to learn more.

77
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Support Vector Regression

Build a boundary based statistical model to predict a continuous outcome as a function of multiple
predictor variables. SVR is able to create much more flexible boundary shapes than the Regression Tree
(Partition) method.

Support Vector Regression

Boston Housing.jmp (Help > Sample Data Folder)
®
1. From an open JMP table, select Analyze > Predictive Modeling >
Support Vector Machines.
2. Add a continuous variable from Select Columns to the
Y, Response role.
3. Add candidate predictor variables to the X, Factor role.
4. If desired, enter a validation column into the Validation role as
shown in this example. Click OK.
5. The Model Launch control panel opens allowing a choice
of a Kernal Function and associated options. Default settings were
used for this example. Click Go.
JMP displays:
• Model Summary detailing the performance of the model via
metrics such as RASE and R-Squared. Here we see that the
R-Squared is 89.6% - a rather good model at describing the
outcome variable.

• Actual by Predicted Plot displaying the actual response values

on the Y axis vs. the predicted response values on the X axis.
This graph provides a graphical means to evaluate the
predictive performance of the model, visualize the variation in
the data around the fitted model, and identify outliers.

To visualize the model and the impact each factor

has on the predicted response, launch the Profiler
found under the red triangle. Contour and Surface
Profilers are also available to visualization the
relationship between pairs of the factors and the
predicted response.

Additional options, such as Residuals Plots, Save

Predicteds, Save Prediction Formula, and Publishing
the Prediction Formula to the Formula Depot are
accessible from the red triangle.

Visit Predictive and Specialized Models > Support Vector Machines in JMP Help to learn more.

78
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

K Nearest Neighbors

Use a proximity-based algorithm to predict a categorical outcome (classify) or prediction the value of a
continuous outcome for new observations based upon the outcomes of similar observations (i.e., their nearest
neighbors).

K Nearest Neighbors Equity.jmp (Help > Sample Data Folder)

1. From an open JMP® table, select Analyze > Predictive Modeling >
K Nearest Neighbors.
2. Select a categorical or continuous response variable from Select
Columns and click Y, Response. Here, we illustrate using a
categorial response variable.
3. Select candidate predictor variables and click X, Factor.
4. If desired, enter the Validation Portion or select a validation
column and click Validation. Click OK.

JMP displays:
• Graph and table showing the misclassification rates and counts
across a range of values for K.
• Confusion Matrix detailing the classification performance for
the value of K with the smallest misclassification rate.
• Mosaic plots (not shown here) which graphically shows the
values in the confusion matrix.

Resuts of the K Nearest Neighbors to predict the risk level (Bad/Good)

• There are 1,192 observations in the Validation Data. The

misclassifcation rate is the lowest when the prediction is based on
only 1 nearest neighbor: 85/1,192 (7.1%) were misclassified. Note
that the misclassification rate increases as the number of nearest
neighbors increase. Of these total misclassifications in the Validation
Data, 3/(3+194) = 0.3% of the Good Risk observations were
misclassified as Bad Risk. 80/(80+195) = 29% of the Bad Risk
observations were misclassified as Good Risk.
• There were 1,192 observation set aside for the Test Data (Results not
shown). The total misclassification rate for these observations using 1
nearest neighbor is 6.9%. 0.4% of the Good Risk observations were
misclassifed as Bad Risk and 3.2% of the Bad Risk observations were
misclassified as Good Risk. These results are often considered to
produce the most accurate estimate of what the misclassification rate
will be in future data as these observations were not part of the
model training nor selection process.

Notes: Additional options, such as Lift Curves, Saving Predicteds, Save Prediction Formula, and Publish
Prediction Formula are accessible from the red triangle near the top next to the response variable name.
Visit Predictive and Specialized Models > K Nearest Neighbors in JMP Help to learn more.

79
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Neural Networks
Build a network based model to describe the impact that multiple predictor variables have on an outcome and
to make predictions of a categorical outcome (classify) or a continuous outcome.

Neural Networks
Diamonds Data.jmp (Help > Sample Data Folder)
®
1. From an open JMP data table, select Analyze > Predictive Modeling >
Neural.
2. Select a response variable from Select Columns and click Y, Response.
Here we chose ‘Price’.
3. Select explanatory variable(s) from Select Columns and click X, Factor.
Here we chose 6 variables (‘Carat Weight’ – ‘Cut’)
4. The option to use a validation column is available in JMP Pro only.
5. Click OK.
6. In the resulting Model Launch window:
In JMP Pro (Dialog box shown top right):
• Specify the Holdback Proportion or the number of Folds if a
validation column was not specified.
• Specify the hidden layer structure by entering the number of
TanH, Linear and Gaussian functions to use in each layer.
• If using boosting, specify the number of models and the learning rate.
• Select the desired fitting options, and click Go.
In JMP (Second from top):

• Select the validation method (Excluded Rows Holdback, Holdback,

KFold).
• Specify the Holdback Proportion or the number of Folds.
• Specify the number of Hidden Nodes, and click Go.

JMP and JMP Pro will generate fit statistics for both the training and
validation data. For categorical responses, a Confusion matrix and
Confusion Rates matrix are also generated.

Tips:

• Use red triangle options (for the model) to view estimates, save
formulas, display model profilers, or display the neural diagram (shown
right). The profilers are particularly useful for visualizing models.
• To view a saved formula: In the column panel of the data table, click the
plus sign next to the name of the desired hidden layer.

Visit Predictive and Specialized Models > Neural Networks in JMP Help to learn more.

80
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Naive Bayes

Use this predictive modeling technique to predict a categorical outcome (classify) as a function of
multiple predictor variables. The technique classifies observations by applying Bayes’ Theorem to
conditional probabilities.

Naive Bayes
Diabetes.jmp (Help > Sample Data Folder)
®
1. From an open JMP table, select Analyze > Predictive Modeling >
Naive Bayes.
2. Select a nominal or ordinal response variable from Select
Columns and click Y, Response.
3. Select candidate predictor variables and click X, Factor.
4. If desired, enter the Validation Portion or select a validation
column and click Validation as was used in this illustration.
5. Click OK. JMP displays:

• Total misclassification counts and rates.

• Confusion Matrix detailing the classification performance.
• ROC curves and AUC values (only results for the training data
shown here).
Note: Click on the top red triangle and select Profiler to display an
interactive tool that shows the predicted probability of each class as a
function of the levels of the predictor variables. Other options, such as
Assess Variable Importance can be accessed from the red triangle
next to Prediction Profiler. The profiler for the three predictor
variables HDL, BMI, and LTG shown on bottom right.

Resuts of the Naive Bayes classifier to predict the level (Low/High)

from the 442 diabetes patients:

• There are 309 observations in the Training Data. Of these, 66 (21%)

where misclassified. 41/(185+41) = 18% of the Low observations
were misclassified as High. 25/(25+58) = 30% of the High
observations were misclassified as Low.
• There are 133 observations in the Validation Data. Of these, 32
(24%) were misclassified. 25/(70+25) = 26% of the Low
observations were misclassified as High. 7/(7+31) = 18% of the High
observations were misclassified as Low.

Notes:
Additional options, such as Lift Curves, Save Predicteds, Save
Prediction Formula, Save Probability Formula, as well as Publish
Probability Formulas are accessible from the top red triangle.

Visit Predictive and Specialized Models > Naive Bayes in JMP Help to learn more.

81
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Creating a Validation Column (Holdout Sample)

Use to subset the data into a set used to build a model (training) and a set used to evaluate a model's predictive
performance (validation). If multiple models are fit, the best performer on the validation data is often the one
chosen. At times, a third set is used (test) to evaluate the chosen model's predictive performance on new data.
This is considered to be a more accurate means to evaluate a model’s future performance as the test set was
neither used in the model building nor selection process.
Using a validation column is particularly useful in building models that have a tendency to overfit the data. Some
modeling platforms in JMP provide the option to specify the validation portion when fitting the model and thus
creating a validation column is not necessary.

Creating a Validation Column (Train, Validate, Test) in JMP Pro

1. From an open JMP data table, select Analyze > Predictive Modeling >
Make Validation Column.
2. Stratification, Grouping, and Cutpoint columns can be used to tailor
the partitioning. If a simple validation column is desired, Click OK.
3. In the resulting window, enter values (counts or proportions)
indicating how the data will be allocated to the training, validation
and test sets. Choose a Random Seed in order to reproduce the same
random assignment if desired. Click OK.
A new column is created, populated with the values 0, 1, and 2 in the
proportions (or counts) specified.

• 3,576 (60%) of the observations (Training set) will be used to build (train) the model.
• 1,758 (30%) of the observations (Validation set) will be used to validate and select the best model.
• 596 (10%) of the observations (Test set) will be used to test the chosen model’s performance on new data.

Creating a Validation Column in JMP

1. From an open JMP data table, select New Column from the Cols menu.
2. In the resulting New Column window, change the Column Name to
Validation.
3. Next to Initialize Data, click on the arrow and select Random.
4. Select Random Indicator. Type in the desired proportions. Here we
chose 50% 0s (train), 30% 1s (validate) and 20% 2s (test).
5. To display the labels Train, Validate and Test rather than 0, 1 and 2,
select Value Labels under Column Properties. Enter the value and the
desired label and click Add one value at a time.
6. Click Apply to view the new column in the data table (to verify that the
column will be created as desired). Then click OK to create the column.

Visit Predictive and Specialized Models > Make Validation Column in JMP Help to learn more.

82
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Model Comparison and Selection

Use this platform to summarize and compare the performance of multiple statistical models that have
been fit to data. For details on fitting different statistical models, see the appropriate guides.

Model Comparison – Continuous Response

Example: We use the Body Fat.jmp data to predict Percent body fat. Formulas for several models, saved to the
data table, are grouped under Prediction Formulas in the column panel. The models were chosen based on the
model’s performance on validation data.
Body Fat.jmp (Help > Sample Data Folder)
1. from the data table,select Analyze > Predictive Modeling >
Model Comparison.
2. Select the prediction formula columns from Select Columns,
and click Y,Predictors. If no columns are selected JMP will
use all saved prediction formulas in the data table.
3. Select Validation from the Select Columns list, and click By.
Then, click OK.
JMP provides the following statistics for comparing the
performance of the different models (statistics for the
Validation set are displayed):
• RSquare (higher is better)
• RASE: Root Average Squared Error (like RMSE, but can
be compared for competing models)
• AAE: Average Absolute Error (a measure of the
magnitude of the errors)

Model Comparison – Categorical Response

Example: In the Equity.jmp data, we predict Bad as a Equity.jmp (Help > Sample Data Folder)
function of several predictors. In this example, the data
was partitioned into Training, Validation and Test sets.

Create the predictive models of your choice (with

Validation), then follow steps 1-3 above to generate
model statistics for each value of the Validation column.

• The misclassification rates and other measures of

performance are provided.
• Select Decision Threshold under the red triangle to launch an interactive tool to compare correct/incorrect
classification rates using different cutoff criteria.
• ROC curves, lift curves, model averaging and other options are available from the red triangles.
• The profiler can also be shown for each model to provide a detailed view of the effect each of the factors
has on the outcome and compare the predictions for various factor settings.
Model Comparison is also available from the Formula Depot.

Visit Predictive and Specialized Models > Model Comparison in JMP Help to learn more.

83
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Text Mining – Describe Unstructured Text Data

The Text Explorer platform is used to explore frequently used words and phrases in unstructured text data such
as text found in product reviews, social media posts, comment fields in surveys, incident reports, etc. Results are
summarzied through frequency tables and word clouds. Tools such as recoding, combining terms, creating stop
words, among others are available to clean the data and help target the anaysis efforts.
Note: Additional text analysis tools are available in JMP Pro. See the Text Explorer – Analyzing Unstructured
Text Data guide for more information.

Example: Pet Survey.jmp (Help > Sample Data Folder)

Summarizing Text Data
1. From an open JMP® data table, select Analyze > Text Explorer.
2. Click on a categorical or unstructured text variable from Select
Columns, and click Text Columns (categorical variables have
red or green bars and unstructured text variables have grey
lines ( ). Click OK.
By default, JMP will provide the following results:
• A summary of the number of terms and cases.
• Term and Phrase Lists, providing a summary of the most
commonly occurring terms and phrases.

Cleaning and Targeting

Many text preparation options, including stemming and
parsing, are available under the red triangle. Additional
options are available from the Term and Phrase Lists.
• To remove terms or phrases, select the terms (or
phrases), right-click, and select Add Stop Word.
Here, we selected ‘cat’ ‘cats’ ‘dog’ and ‘dogs’ as Stop Words.
• To add phrases to the term list, select the phrases, right-click, and select Add Phrase.
• To combine terms or change the values, select the terms, right-click, and select Recode.
These steps are very helpful in targeting analysis efforts to better find and quantify the importance of key
themes.

Visualizing Text Data with Word Clouds

To generate a word cloud for your prepared text data, select
Display Options > Show Word Cloud from the red triangle
menu.
Options for customizing the word cloud are available from
the red triangle next to Word Cloud such as coloring the
words based on the value of a variable. In this example, the
options Centered Layout and Arbitrary Colors were used.

Visit Basic Analysis > Text Explorer in JMP Help to learn more.

84
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Text Mining – Analyzing Unstructured Text Data

The Text Explorer platform is used to explore frequently used words and phrases in unstructured text data
such as text found in product reviews, social media posts, comment fields in surveys, incident reports, etc.
Additional tools are available in JMP® Pro for further analysis.
The text data must first be prepared for these analyses. See the Text Explorer – Describing Unstructured Text
Data guide and the JMP Help for information on using the Text Explorer platform to do this.

Analyzing Unstructured Text Example: Pet Survey.jmp (Help > Sample Data Folder)

Text analysis involves transforming prepared text data into a Document Term Matrix (DTM). Each row in the
DTM corresponds to a document (a cell in the column of text data), and each column in the DTM corresponds to
a term. The DTM is then used as an input in all analyses.

Here we selected Save Document Term Matrix under the Red Triangle and used the default specifications.
A number of analysis options are available from the top red triangle menu in the text explorer analysis report:

• Latent Class Analysis groups documents into clusters of

similar documents.
• Latent Semantic Analysis, SVD is a dimension reduction
technique similar to principal components analysis.
• Topic Analysis groups words into themes or topics,
and is similar to factor analysis.
• Cluster Terms or Cluster Documents performs
hierarchical clustering of the terms or documents.

Here we selected Latent Semantic Analysis, SVD using the

default settings. Then selected SVD Scatter Plot Matrix using
the default settings.

Saving Results
Text Explorer analysis results, such as the DTM, can be saved to the data table. The resulting columns can then
be used as predictors in other analyses.

• To save information that corresponds to documents, save the Document Term Matrix, Document
Singular Vectors or Document Topic Vectors.
• To save information that corresponds to terms, without respect to the specific documents, save the
Term Table, Term Singular Vectors or Term Topic Vectors.

Visit Basic Analysis > Text Explorer in JMP Help to learn more.

85
 Data Mining and Predictive Modeling
jmp.com/learn Nov 2022

Association Analysis (Market Basket Analysis)

Analyze transactional data such as product purchases and occurrence of events to identify those that are
dependent upon each other or tend to occur together. Metrics such as the likelihood of items/events occuring
based on the occurrence of other items/events, among others are produced.

Note that the data must be in list format, where each row identifies the customer or transaction ID (in one
column) and the item or event (in another column).
Example: Grocery Purchases.jmp (Help > Sample Data Folder)
Association Analysis
1. From an open JMP® data table, select Analyze > Screening >
Association Analysis.
2. Select a categorical variable from Select Columns and click Item
(categorical variables have red or green bars).
3. Select the variable that represents the transaction ID and
click ID.
4. Accept the default settings (lower left), and click OK.

By default, JMP provides the following tables of results:

• Frequent Item Sets: the list of item sets whose support

(proportion of occurrences) exceeds the minimum
value specified in the launch window.
• Rules: Association rules that meet the minimum
support, minimum confidence, minimum lift,
maximum antecedents, and maximum rule size
requirements specified in the launch dialog.

Here, we have sorted the tables in ascending order on Item

Set and Condition (respectively). To sort a table, right-click on the table, select Sort by Column, select the
column of interest, check Ascending, and click OK.

Interpretation:

The minimum support is 0.1 and the minimum confidence is 0.4 (see the default settings above). Thus, an item set will not appear in either list unless
that item set is purchased by at least 10% of the customers (this is support). Furthermore, it won’t appear in the rules list unless a consequent item is
purchased at least 40% of the time (this is confidence). For example, apples and avocado were purchased together 14% of the time. And, when apples
and avocados were both purchased by a customer, that customer also purchased a baguette 81% of the time.

The maximum antecedents is 3 and the maximum rule size is 4. Therefore, the rules list only contains conditions of three or fewer items and no rule
includes more than 4 condition and consequent items (combined).

The minimum lift is 1.2. Lift, which is the ratio of the confidence to the support, measures how much more (or less) likely a customer is to purchase the
consequent item when they are already purchasing the condition items, compared to the probability of buying the consequent item in general. For
example, the lift for apples, avocado and the consequent baguette is 2.058, indicating that customers are more than twice as likely to buy a baguette
when they are also buying apples and avocados.

Visit Predictive and Specialized Models > Association Analysis in JMP Help to learn more.

86
 Quality and Process
jmp.com/learn Nov 2022

Variables Control Charts – I/MR Charts

This guide provides instructions for creating I & MR (Individuals and Moving Range) control charts using the
Control Chart Builder and the Control Chart platform. I/MR control charts are used to monitor a continuous
variable where the data is sampled without subgroups.

I/MR Charts – Control Chart Builder

1. From an open JMP® data table, select Analyze > Quality and Process > Control Chart Builder.
2. Drag a continuous variable from Select Columns (continuous variables have blue triangles), and drop it in
the Y zone.
By default, JMP produces a control limit summaries table and two charts: Individuals (Individual
Measurements) and Moving Range (the moving range between each pair of consecutive points). Right-click
on the graph and select Limits > Sigma to change the method for estimating the standard deviation.
3. If available, drag and drop a variable in the Phase zone (top), and click the Done button when finished.
Diameter.jmp (Help > Sample Data Folder > Quality Control)

I/MR Charts – Control Chart Platform

1. From an open JMP data table, select Analyze > Quality and
Process > Control Chart > IMR Control Chart.
2. Select one or more continuous variables from Select
Columns, and click Process. Add a variable defining phases
into the Phase role if desired. Click OK.
Notes:
Additional options, such as tests for special causes, are available
from the Red Triangles in the Control Chart platform (or Right-
click on the graph in the Control Chart Builder).

Visit Quality and Process Methods > Control Chart Builder and Legacy Control Charts in JMP Help to learn more.

87
 Quality and Process
jmp.com/learn Nov 2022

Variables Control Charts – XBar & R/S Charts

This guide provides instructions for creating XBar & R or XBar & S control charts using the Control Chart Builder
and the Control Chart platform. XBar & R or XBar & S control charts are used to monitor a continuous variable
where the data is sampled with subgroups.

XBar & R Charts – Control Chart Builder

1. From an open JMP® data table, select Analyze > Quality and Process > Control Chart Builder.
2. Drag a continuous variable from Select Columns (continuous variables have blue triangles), and drop it in
the Y zone. Here we used ‘Diameter’.
3. Drag and drop a subgroup variable in the Subgroup zone (at the bottom). Here we used ‘Day’.
By default, JMP produces a control limit summaries table and two charts: XBar (subgroup average) and
Range (subgroup range).
Right-click on the bottom graph and select Points > Statistic to change from a Range chart to an S Chart, and
use Limits > Sigma to change the method for estimating the standard deviation.
Diameter.jmp (Help > Sample Data Folder > Quality Control)
4. If available, drag and drop a
variable in the Phase zone (top), and
click the Done button when finished.
Here we used ‘Operator’.

XBar & R Charts – Control Chart Platform

1. From an open JMP data table, select Analyze > Quality and Process >
Control Chart > XBar Control Chart.
2. Select one or more continuous variables from Select Columns, and
click Process.
3. Enter the variable that defines the subgroup into the Subgroup role.
Click OK.

Notes:
Additional options, such as tests for special causes, are available from the red triangles in the Control Chart
platform. Right-click on the graph in the Control Chart Builder for many of these same options.

Visit Quality and Process Methods > Control Chart Builder and Legacy Control Charts in JMP Help to learn more.

88
 Quality and Process
jmp.com/learn Nov 2022

Attribute Control Charts – P and NP Charts

This guide provides instructions on creating P and NP attribute control charts. P charts are often used to plot the
proportion of nonconforming (defective) items per subgroup, while NP charts are often used to plot the number
of nonconforming items p er subgroup.

P Charts
1. From an open JMP® data table, select Analyze > Quality and Process > Control Chart > P Control Chart
2. Select one or more continuous variables from Select Columns, and click Process.
3. Select the column defining the number of trials (i.e, sample size) for the n Trials role. Click OK.
Note: If sample sizes are not constant, the control limits will vary.

Washers.jmp (Help >

Sample Data Folder >
Quality Control)

NP Charts
1. From an open JMP data table, select Analyze > Quality and Process > Control Chart > NP Control Chart
2. Select one or more continuous variables from Select Columns, and click Process.
3. Select the column defining the number of trials (i.e, sample size) for the n Trials role. Click OK.

Washers.jmp (Help >

Sample Data Library >
Quality Control)

Tips:

• The process variable must be sorted in time order.

• Many options, such as tests for special causes and capability analysis, are available from the red triangles.

Visit Quality and Process Methods > Legacy Control Charts in JMP Help to learn more.

89
 Quality and Process
jmp.com/learn Nov 2022

Attribute Control Charts – C and U Charts

This guide provides instructions on creating C and U attribute control charts. C charts are often used to plot the
number of nonconformities (or defects) in a subgroup that usually, but not necessarily, consists of one
inspection unit, while U charts are often used to plot the number of nonconformities per unit, where the
number of units per subgroup can vary.

C Charts
1. From an open JMP® data table, select Analyze > Quality and Process > Control Chart > C Control Chart.
2. Select one or more continuous variables from Select Columns, and click Process.
3. Select the column defining the unit size (if not a constant size) and place in the n Trials role. Click OK.
Note: If sample sizes are not constant, the control limits will vary.

Fabric.jmp
(Help > Sample
Data Folder >
Quality Control)

U Charts (or DPU Charts)

1. From an open JMP data table, select Analyze > Quality and Process > Control Chart > U Control Chart.
2. Select one or more continuous variables from Select Columns, and click Process.
3. Select the column defining the unit size (if not a constant size) and place in the n Trials role. Click OK.
In the example below, the number of defects (# defects) per unit inspected (Unit Size) is plotted on a U
Note: Since the Unit size is not constant, the control limits vary.

Example:
Braces.jmp (Help
> Sample Data >
Quality Control)

Tips:

• The process variable must be sorted in time order.

• Many options, such as tests for special causes and capability analysis, are available from the red triangles.

Visit Quality and Process Methods > Legacy Control Charts in JMP Help to learn more.

90
 Quality and Process
jmp.com/learn Nov 2022

Tolerance Interval
Use to produce an interval estimated to contain a specified proportion of a population.

Tolerance Interval for Proportion (Normal Distribution)

Coating.jmp (Help > Sample Data Folder > Quality Control)
1. From an open JMP data table, select Analyze > Distribution.
2. Select one or more continuous variables from Select
Columns, click Y, Columns (continuous variables have blue
triangles), and click OK.
3. From the Distributions report window, select Tolerance
Interval under the red triangle next to the variable name.
4. Select the confidence level and specify proportion of the
population you desire the interval to cover. Choose Two-sided
or one-sided interval. Here we chose 95% confidence,
90% proportion to cover, and two-sided.
5. Choose to Assume Normal Distribution if you believe
that is an appropriate model to use to describe the distribution
of the variable for the population. Click OK.
Note: A goodness-of-fit test (not shown) indicates that a normal
distribution is appropriate.

JMP will add a table displaying the lower and upper bound
of the Tolerance Interval based upon assumption of a normal distribution.

Interpretation:
It is estimated, with 95% confidence, that 90% of pins produced by this process that
these data can be considered a sample from will have a Weight between 17.3 and 23.5.

Tolerance Interval for Proportion (Non-Normal Distribution)

Process Measurements.jmp (Help > Sample Data Folder)
1. Follow steps 1-3 above.
2. In the Tolerance Interval Dialog box, choose Nonparametric
method. Click OK.
Note: A goodness-of-fit test (not shown) indicates that a
normal distribution is not an appropriate model to use.
JMP will add a table displaying the lower and upper bound
of the Tolerance Interval for any continuous distribution.
Note: The actual confidence level achieved may be larger than the
value specified. Here it’s 97.6% confidence.
Interpretation:
It is estimated, with 97.6% confidence, that 90% of this process that these data
can be considered a sample from have values between 5.23 and 17.2.

Visit Basic Analysis > Distributions > Options for Continuous Variables > Tolerance Intervals in JMP Help to
learn more.
91
 Quality and Process
jmp.com/learn Nov 2022

Process Capability Analysis

This guide provides instructions on performing a capability analysis using the Distribution and Control Chart
platforms. A process capability analysis is used to evaluate the performance of a process for a continuous
variable to specifications (Lower and/or upper specification limit, and/or target). See Capability Analysis for
Multiple Responses guide for performing process capability analyses on multiple process variables
simultaneously.

Capability Analysis – from Distribution Platform Coating.jmp (Help > Sample Data Folder > Quality Control)

1. From an open JMP data table, select Analyze > Distribution.

2. Select one or more continuous variables from Select Columns, click Y,
Columns. Here we chose ‘Weight’. Click OK to generate a histogram
and summary statistics.
3. From the red triangle next to the variable name, select Process
Capability.
4. Enter the specification limits.
5. Specify options such as defining a Subgroup variable, using a historical
sigma value, choosing a specific distribution to use. Click OK.
• Two sets of capability metrics are produced – Within Sigma
Capability based on the within subgroup measure of variation (Within
Sigma = 1.052) and Overall Sigma Capability based on the total
variation in the data (Overall Sigma = 1.475).
• Estimates of the % of a process that would exceed the specification is
provided based upon the fitted distribution (normal in this case) and
the two different estimates of variation (Within and Overall). An
additional estimate of this is provided based upon the actual number
of observations in the data set that exceeded the specification limits.
• Interactive Capability Plot is available under the red triangle next to the variable name.
The specification limits, mean, and the overall standard deviation can be changed and impact on the
capability metrics and PPM outside of specification.

Note: Process Capability Platform can also be accessed from Analyze > Quality and Process > Process Capability.

Visit Quality and Process Methods > Process Capability in JMP Help to learn more.

92
 Quality and Process
jmp.com/learn Nov 2022

Capability Analysis for Multiple Responses

This guide provides instructions on performing a process capability analysis for multiple continuous process
variables simultaneously. This platform produces a set of reports and graphs that are designed to compare the
performance of the different process variables to each other.

The Capability Analysis Dialog Semiconductor Capability.jmp (Help > Sample Data Folder)

1. From an open JMP® data table, select Analyze > Quality

and Process > Process Capability.
2. Select one or more continuous variables from Select
Columns (continuous variables have blue triangles), click Y,
Process, and click OK.
If specification limits have not been entered as a column
property for each of the measures, or if there is an error,
you will receive an error window. Click OK. The resulting
window allows you to correct specification limits or select a
data table with the limits stored.

Additional Options:

• To skip the spec limits dialog, click No (skip columns

with no spec limits) under Show Spec Limits in the
opening Dialog Box.
• To specify subgroupings use Process Subgrouping.
• To specify sigma use Historical Information.
• To perform capability studies for nonnormal data use
Distribution Options.

The Process Capability Analysis Report

JMP produces three graphical summaries of the capability of
each measure. For these graphs, measures have been
standardized to the spec limits.

• Goal Plot: Measures with a Ppk > 1.0 are plotted inside the triangle, and measures with a Ppk < 1.0 are
plotted outside the triangle. The position of each point relative to the axes provides information on spread
(y-axis) and centering (x-axis) of each measure relative to the spec limits. Choose Shade Levels under Red
Triangles to add color.
• Capability Box Plots: Box plots for each variable.
• Capability Index Plot: The Ppk of each point.

Notes: Use the sliders to change the cutoff for acceptable Ppk. Additional options, such as Individual Detail
Reports and Summary Reports, are available from the top red triangle menu.
Visit Quality and Process Methods > Process Capability in JMP Help to learn more.
93
 Quality and Process
jmp.com/learn Nov 2022

MSA Continuous Data – Gauge R&R

This guide provides instructions on creating a variability chart and performing a gauge R&R (Repeatability and
Reproducibility) measurement system analysis (MSA). This analysis is designed to evaluate and certify the
performance of a measurement system that produces continuous data. For information on using the EMP
(Evaluating the Measurement Process) method, see the MSA Continuous Data – EMP Method guide.

Measurement Systems Analysis: Variability Chart Gasket.jmp (Help > Sample Data Folder > Variability Data)

1. Select Analyze > Quality and Process > Variability/Attribute

Gauge Chart.
2. Select Variability under Chart Type.
3. Click on a continuous variable from Select Columns, and click
Y, Response (continuous variables have blue triangles).
4. Select one or more grouping variables and click X, Grouping.
5. Select a part or sample variable and click Part, Sample ID.

By default, JMP displays two plots:

• The Variability Chart for Y shows the individual measurements and
range bars for each combination of the part and the grouping
variables. To better visualize sources of variation, select Connect
Cell Means, Show Group Means and Show Grand Mean from the
red triangle.
• The standard deviations plot shows the standard deviation for each
combination of the part and the grouping variables.

Performing a Gauge R&R:

1. Select Gauge Studies > Gauge R&R from the red triangle.
2. In the Variability Model pop-up window, select the model type
and click OK (the default is Crossed).
3. If specs are available, choose the tolerance entry method, enter
either the Tolerance Range or the Spec Limits, and click OK. For
this example: Tolerance Range (USL-LSL) = 100.
The resulting Gauge R&R and Variance Components for Gauge R&R
reports quantify sources of measurement system variation.

Interpretation:

1. The % Gauge R&R is 23.86%. This is the percentage of the total

variation being attributed to the measurement system.
2. The Precision/Tolerance Ratio (or the P/T Ratio) is 0.3396.
3. Variance Components for Gauge R&R: The measurement system
(Gauge R&R) accounts for 5.69% of the total variation in the study.

Notes: To conduct Bias and Linearity studies, specify a standard (or reference column) in the dialog window.
Additional options are available from the top red triangle.

Visit Quality and Process Methods > Variability Gauge Charts in JMP Help to learn more.
94
 Quality and Process
jmp.com/learn Nov 2022

MSA Continuous Data – EMP Method

This guide provides instructions on performing a measurement system analysis (MSA) using the EMP (Evaluating
the Measurement Process) method. This analysis is designed to evaluate and certify the performance of a
measurement system that produces continuous data.

Measurement Systems Analysis: EMP Method Gasket.jmp (Help > Sample Data Folder > Variability Data)

1. Select Analyze > Quality and Process > Measurement System

Analysis.
2. Click on a continuous variable from Select Columns, and click
Y, Response (continuous variables have blue triangles).
3. Select a part or sample variable and click Part, Sample ID.
4. Select one or more grouping variables and click X, Grouping.
5. Ensure that EMP is the MSA Method, and click OK.
(Notes: If desired, change the Chart Dispersion Type to Standard
Deviation to base analyses on standard deviations instead of ranges,
and change Model Type if required.)

By default, JMP displays an Average Chart and a Range Chart:

• The Average Chart shows the average measurement for each

combination of the part and the grouping variables. Out-of-control
points are desirable, as they indicate the ability to detect part-to-part
variation.
• The Range Chart shows the range for each combination of the part
and the grouping variables. Having all points within the control limits
is desirable, as this indicates similar measurement error across
groups.
(Note: A Standard Deviation Chart displays if the Chart Dispersion
Type has been changed to Standard Deviation.)

Many additional options are available under the top red triangle:
• Parallelism Plots, for assessing interactions between the parts and
the grouping variables.
• EMP Results, for evaluating and classifying the measurement system.
• Effective Resolution, for determining the resolution of the measurement system.
• Bias Comparison and Test-Retest Error Comparison plots, for assessing
differences in grouping variable means and errors, respectively.
• Shift Detection Profiler, for dynamically exploring the probability of a warning.
• Variance Components and EMP Gauge RR Results, for quantifying sources of
measurement system variation.

Visit Quality and Process Methods > Measurement Systems Analysis in JMP Help to learn more.

95
 Quality and Process
jmp.com/learn Nov 2022

MSA Attribute Data

Use to evaluate the performance of a categorical based measurement system (e.g., pass or fail, classification of
parts into multiple categories, ordinal ratings). This analysis is often used to assess the ability of observers
(appraisers) to correctly and consistently classify inspected items.

Attribute Measurement Systems Analysis

Attribute Gauge.jmp (Help > Sample Data Folder)
1. Select Analyze > Quality and Process > Variability /
Attribute Gauge Chart.
2. Select categorical variables (usually raters) from Select
Columns, and click Y, Response (categorical variables have
red or green bars).
3. Select a grouping variable (often part), and click X, Grouping.
4. For an expert agreement study, select the standard (expert
or reference) and click Standard.
5. Ensure Attribute is displayed under Chart Type
(it will change by default) then click OK.

JMP displays a Gauge Attribute Chart and several reports to assess agreement:
• The Gauge Attribute Chart (bottom, left) displays the percent agreement within each part (top) and within
each rater. Here we see that parts 6 and 14 were problematic for the raters (low agreement), and that raters
A and B had slightly higher overall agreement than rater C. The Agreement Report (not shown) summarizes
rater and overall agreement.
• The Agreement Comparisons report (bottom, middle) provides Kappa statistics for each pair of raters and
for each rater against the standard. Kappa is a measure of non-chance agreement. Generally, Kappa values
of 0.6 or higher indicate acceptable agreement.
• The Effectiveness report (bottom, right) summarizes rater agreement with the standard. It includes a
Misclassification Matrix and a Conformance Report, which provides the probabilities of false alarms (Type I
errors) and misses (Type II errors) for each rater.

Visit Quality and Process Methods > Attribute Gauge Charts in JMP Help to learn more.

96
 Reliability and Survival
jmp.com/learn Nov 2022

Distribution Fitting (Life Distribution)

Use to model a single continuous variable with a probability distribution and estimate a variety of parameters
(e.g., mean, percentiles, probabilities). This platform is designed to fit and compare many different distributions
and can analyze data that is censored (incomplete). Note: Though frequently used to analyze time-to-event data,
the analysis methods can be used to model any continous variable (e.g., force-to-event).

Distribution Fitting (Life Distribution) Fan.jmp (Help > Sample Data Folder > Reliability)

1. From an open JMP® data table, select Analyze > Reliability and
Survival > Life Distribution.
2. Select a continuous time variable from Select Columns, then click
Y, Time to Event (continuous variables have blue triangles).
3. If the data contain censored values, select the censoring variable
and click Censor. Change the Censor Code if needed (the default
is Censored).
4. Select the Confidence Interval Method (the default is Wald), and
click OK.
JMP will display:
• An Event Plot (click on the gray icon to open), which graphically
depicts failures (denoted by X’s) and censored events (denoted
by triangles).
• The Compare Distributions panel, which allows you to compare
the fit of various distributions.
• The Statistics panel, with summary statistics,
parameter estimates and profilers for each distribution selected.

To Compare Distributions:
- Click a radio button under Scale to apply the scale for a
distribution. If the selected distribution fits the data well, the
plotted points will approximate a straight line.
- Check a box under Compare Distributions
to fit a distribution and display a fitted line, shaded confidence
bands and a Profiler for the distribution.
- Select Fit All Distribution or Fit All Non-negative under the top
red triangle to compare all distributions using model
comparison criteria such as AIC and BIC.

Tips:
• To compare the same life distribution across groups, use the Compare Groups option in the Life
Distribution dialog. Other options, such as show survival curve and view tabbed report are available
from the top red triangle.
• If studying competing failure causes, enter the variable in the Failure Cause field in the launch window.
• Continuous distributions can also be fit in Analyze > Distribution.

Visit Reliability and Survival Methods > Life Distribution in JMP Help to learn more.

97
 Reliability and Survival
jmp.com/learn Nov 2022

Accelerated Life Testing (Fit Life by X)

Use to model the relationship a factor has on the time-to-event. It is often used to model a factor that
accelerates the event as a means to predict the time-to-event for normal operating conditions where failures do
not frequently occur. This platform is able to analyze data that is censored (incomplete).

Accelerated Testing (Fit Life by X) Devalt.jmp (Help > Sample Data Folder > Reliability)

1. From an open JMP® data table, select Analyze >

Reliability and Survival > Fit Life by X.
2. Select a continuous time variable from Select Columns,
then Click Y, Time to Event (continuous variables have
blue triangles).
3. Select the factor and click X.
4. If the data contains censored values, select the
censoring variable and click Censor. Change the Censor
Code if needed (the default is 1).
5. Select the Relationship, the Distribution and the
Confidence Interval Method, and click OK.

JMP will display:

• A scatterplot of the data, with the time to event (Y) variable plotted against the factor (X). Triangles
indicate censored values. Under the red triangle next to Scatterplot select Show Density Curves and
Add Quantile Lines to compare density curves and quantiles for different values of the factor.
Other distributions can be it to the data by selecting under the top red triangle.

• The Comparisons panel, which allows you to compare various profilers and statistics for each of the
selected distributions (below, middle).
• The Results panel for each selected distribution, containing diagnostic plots, statistical output, profilers,
and Custom Estimation calculators to compute quantiles and probabilities (below, right).

Notes: Other options, such as fit individual distribution, fit all distributions, and view tabbed or individual reports
are available under the top red triangle. Click on red triangles throughout the report to access additional
features.

Visit Reliability and Survival Methods > Fit Live by X in JMP Help to learn more.

98
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE Full Factorial Design

This guide provides instructions on designing a full factorial experiment. A full factorial experiment in where
every possible treatment combination will be studied. For analysis of full factorial experiments, see the DOE Full
Factorial Analysis guide.

Create the Design (Full Factorial Design)

1. Open the platform under DOE > Classical > Full Factorial Design. To create a 23 full factorial design with one run at each
of the 8 corner points and 3 runs at the center point.
2. Specify the Response(s):

• Double-click on Y, under Response Name, to name the

response.
• If needed, change the response Goal and Upper and
Lower Limits.
• Click Add Response to add additional responses.
3. Specify the Factors:
• Click Continuous or Categorical, then the number of levels
to add a factor. Click Remove to remove a factor.
• Double-click to change the factor name.
• Tab to change the values for the factor.
• Repeat for all factors.
4. Click Continue.
5. Specify the Run Order (default is Randomize), the Number of Center Points and the Number of Replicates
(the number of additional sets of runs for each design point).
Here, we have specified an unreplicated fully randomized 23 full factorial design with 3 center points,
totaling 8 + 3 = 11 runs.
6. Select Make Table to generate the design (or Back to make
changes). In the design table:
• The Pattern column provides a key to the factor levels for
each trial.
• The factor settings are indicated in the columns for each
factor.
• The response for each trial will be recorded in the last
column (here, Percent Reacted).
• The Model script will be saved to the data table and the
design specification window stays open to change or regenerate the design.

Notes:
• Select Evaluate Design to view properties of the design.
• The Easy DOE platform (under DOE menu) provides a guided workflow to step through the process of
creating and analyzing experiments and is an alternative to the steps above for creating a design.
• Full factorial designs can also be generated from the Custom Design platform.

Visit Design of Experiments Guide in JMP Help to learn more.

99
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE Full Factorial Analysis

This guide provides information on analyzing a full factorial experiment (experiments where every possible
treatment combination is run). For instructions on designing of full factorial experiments, see the DOE Full
Factorial Design guide.

Specify the Model and Analyze Example: Reactor 32 Runs.jmp (Help > Sample Data Folder >
Design Experiment)
1. From an open JMP® table (for a completed full factorial This is a 25 full factorial design

experiment) select Analyze > Fit Model.

Note: The Fit Model platform can also be accessed from
the Model script saved to the data table made when the
Designed Experiment was created.
2. In the Model Specification window:

• Click on the response under Select Columns, and click

Y (under Pick Role Variables).
• Select the factors of interest. Under Macros, select Full
Factorial (with Degree = 2) to enter all 5 main effects
and all 10 2-way interactions into the model.
• To remove higher-order interactions, select the
interactions under Construct Model Effects and hit Remove.

3. Click Run. JMP will display the following results:

• The Actual by Predicted plot.
• The Effect Summary (shown).
• Some diagnostics plots.
• The Lack of Fit table (if replicates were used).
• Parameter estimates and effect tests.
• The Prediction Profiler and more.
• Other options are available under the top red triangle.

4. To reduce the model, remove non-significant terms bottom-up.

To remove a term:
• In the Effect Summary, select the least significant term(s).
• Click Remove.
• Effect Heredity: Keep lower-order components with containing
effects above them (indicated by ‘^’ in the right-most column).
• Repeat until the model has been reduced.
• Use the Prediction Profiler to explore the model, to optimize,
and/or to simulate response values.
• Surface and Contour Profiler, and Interaction Plots are other
graphs available to visualize the results.
• Model diagnostics and savings columns to the data table are available under the red triangle.

Note: The Easy DOE platform (under DOE menu) provides a guided workflow to step through the process of
creating and analyzing experiments and is an alternative to the steps above for analyzing the experimental data.

Visit Design of Experiments Guide in JMP Help to learn more.

100
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE Fractional Factorial Design

This guide provides instructions on designing a fractional factorial experiment using the Screening Design
platform. Fractional factorial experiments are used when it's more efficient to not perform an experimental run
for every possible treatment combination as is done in a full factorial design.

Create the Design (Screening Design)

To create an 8 run 25-2 Resolution III fractional factorial design.

1. Open the platform under DOE > Classical > Two Level
Screening > Screening Design.
2. Specify the Response(s):
• Double-click on Y, under Response Name, to name
the response.
• If needed, change the response Goal and Upper and
Lower Limits.
• Click Add Response to add additional responses.
3. Specify the Factors:
• Add the desired number of Continuous (2-level) and/or
2- or 3-Level Categorical factors.
• Double-click to change the factor names.
• Tab to change the values for each factor.
4. Click Continue > Choose from a list of fractional factorial designs.
5. From the Design List, select the desired design and click Continue.
Note that Plackett-Burman screening designs, and Incomplete
Block Designs (with a value under Block Size) are also available.
6. Specify the Run Order (default is Randomize), the Number of

Center Points and the Number of Replicates (the number

of additional sets of runs for each design point). Here we
created an unreplicated fully randomized 25-2 fractional
factorial design with 3 center points, totaling 11 runs.
7. Select Make Table to generate the design table (or Back
to make changes).
The Screening and Model scripts will be saved to the
data table and the design specification window stays
open to change or regenerate the design.
Notes:
• Select Evaluate Design to view properties of the design.
• The Easy DOE platform (under DOE menu) provides a guided workflow to step through the process of
creating and analyzing experiments and is an alternative to the steps above for creating a design.
• Screening designs can also be generated from the Custom Design platform.

Visit Design of Experiments Guide in JMP Help to learn more.

101
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE Fractional Factorial Analysis

This guide provides instructions on analyzing fractional factorial experiments (experiments where not every
possible treatment combination in a full factorial design is run) using the Fit Model platform.

Specify the Model and Analyze Reactor 8 Runs.jmp (Help > Sample Data Folder >
Design Experiment).
This is a unreplciated 8 run 25-2 Resolution III
1. Experiments designed in JMP® will have a Model script saved to
fractional. factorial design.
the data table. The model specification window will be populated
with this model. To generate the model manually: Thi

From an open JMP table (for a completed fractional factorial

experiment) select Analyze > Fit Model.

• Click on the response under Select Columns, and click Y.

• Select the factors of interest, and click Add (under Construct
Model Effects).
• To add specific interaction terms, select the variables under
Select Columns and click Cross. In this experiment, all but two
of the 2-way interactions are confounded with main effects.
These two 2-way interactions are confounded with other
2-way interactions:
Feed Rate*Stir Rate = Catalyst*Temperature, and
Feed Rate*Catalyst = Stir Rate*Concentration.
2. Check the Keep dialog open box, and click Run.
Note: If aliased effects were entered into the model, JMP will
display Singularity Details and will indicate that estimates are
zeroed or biased.
Note: If your design is saturated (there are no error degrees of
freedom) as the case here, select Estimates > Sorted Estimates
from the top red triangle to see p-values based on Lenth Pseudo
Standard Error (PSE, an estimate of residual standard error).
3. Reduce the model as desired, removing non-significant terms
starting with 2-way interactions until a final model containing
only significant terms (including all main effects that are involved
in significant 2-way interactions).

• Use the Prediction, Surface, and Response Profiler, and Interaction Plots to visualize the results.
• Model diagnostics and savings columns to the data table are available under the red triangle.

Note: The Easy DOE platform (under DOE menu) provides a guided workflow to step through the process of
creating and analyzing experiments and is an alternative to the steps above for analyzing the experimental data.

Visit Design of Experiments Guide in JMP Help to learn more.

102
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE Screening Experiment Analysis

This guide provides instructions on analyzing screening experiments (e.g., fractional factorial) using the Fit Two
Level Screening platform. Though the Fit Model platform can also be used, the Fit Two Level Screening platform
produces graphs and analysis results designed for identifying the important effects in a fully saturated
experiment. This example is a 20 run 2^5 screening experiment generated from the Custom Design platform.
This design allows estimation of all main effects and two-factor interactions.

Example: Reactor 20 Custom.jmp (Help >

Specify the Model and Analyze Sample Data Folder > Design Experiment)

Most experiments designed in JMP will have Screening and Model scripts saved to
the data table. The Screening script launches the Screening analysis platform and
automatically fits a saturated model.
1. Launch the Fit Two Level Screening platform under DOE > Classical > Two Level
Screening or Click on the green triangle next to Screening to run the script if
saved to the data table.
2. JMP fits a saturated model (here, 19 terms plus the intercept). The
Contrasts table and Half Normal Plot identify active factors using
Lenth Pseudo Standard Error (PSE).
Notes: In screening experiments, we assume that most effects are
inactive and their estimates are essentially random noise.
The line in the Half Normal Plot is drawn with a slope equal to the
Lenth PSE (an estimate of the residual standard error). Most effects
are inactive and fall close to this line. Effects that deviate substantially
from this line are labeled as active.
3. Select Run Model (at the bottom) to launch the Fit Model platform
with only the active effects. Results include: Effect Summary, Lack of
Fit (if replicated points), Summary of Fit, ANOVA table and more.
Other options – such as residuals or normal plot, profiler and
interaction plots – are available under the top red triangle.

Tips:

• In the Fit Two Level Screening platform, to highlight additional terms

hold the Control key and click on the term(s) to select.
• Individual and Simultaneous p-Values in the Fit Two Level Screening
platform are based on Monte Carlo simulation (and will vary).
• An alternative approach to running the analysis is to use Analyze > Fit
Model specifying the model or run the Model script. See the “DOE
Fractional Factorial Analysis” guide.

Visit Design of Experiments Guide in JMP Help to learn more.

103
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

DOE - Custom Designs

This guide provides instructions on designing optimal experiments using the flexible Custom Design platform.
The Custom Designer can be used for almost any experimental situation, including factor screening,
optimization, and mixture problems, and can accommodate designs with hard-to-change factors and other
constraints.

DOE: Generating A Custom Design

1. Select DOE > Custom Design.

2. Under Responses, specify the response(s):
• Double-click to rename the response (default is Y).
• Change the response goal (default is Maximize).
• Click Add Response to add additional responses.
3. Under Factors, specify the experimental factors:
• Click Add Factor, and select the factor type and number of levels.
To add several factors of the same type and number of levels use
Add N Factors (enter a number) and click Add Factor.
• Double-click to rename the factors.
• Change the factor values (the experimental settings).
• Change “Easy” under Changes to “Hard” or “Very Hard” to
generate a split-plot or split-split plot design. Click Continue.
4. Under Model, specify the statistical model to be estimated:
• To add all interaction or power terms up to a given degree, click the corresponding button.
• To add terms needed to perform a response surface analysis, click RSM (or Scheffe Cubic for mixture
designs).
• To add specific power terms, highlight one or more factors in the Factors window and click Powers. To
add an interaction term, highlight one or more factors in the Factors window and select the factors you
in the Model window that wish to cross with those chosen factors.
• To remove a term, highlight it and click Remove Term.
• To reduce the number of runs needed (at the expense of effect aliasing), click “Necessary” for a term
(under Estimability) and change to “If Possible.”
5. Under Design Generation, fine-tune the design (as needed):
• Specify the block size or number of whole plots (with “Hard” to change factors).
• Enter the Number of Center Points and/or Number of Replicate Runs.
• Select (or specify) the desired number of runs.
6. Click Make Design. The resulting design displays under Design.
7. Select the desired Run Order, then click Make Table to generate the design table (or Back) to make changes.
Model, Evaluate Design and DOE Dialog scripts are saved to the data table (top left), and the design
specification window stays open to change or regenerate the design if needed.

Note: The Easy DOE platform (under DOE menu) provides a guided workflow to step through the process of
creating and analyzing experiments and is an alternative to the steps above for creating the experiment.

Visit Design of Experiments Guide in JMP Help to learn more.

104
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

Evaluate and Compare Designs

Use the Evaluate Design platform to assess the properties of an experimental design. To compare alternative
designs, use the platform Compare Designs.

Evaluate Design
Most DOE platforms in JMP provide an Evaluate Design outline to assess a design before making the design table
to run the experiment. The platform Evaluate Design shows the same diagnostics for any existing design, and
can also be used after an experiment to assess the impact of incorrect settings or lost runs.

Example: Run the script Help > Sample Data > Open the Sample Scripts Directory > Compare Same Run Size.jsl.
Two 13-run design tables are constructed: A Definitive Screening Design (DSD) and a Plackett-Burman design.

1. Make Plackett-Burman.jmp the active window, and select DOE > Design Diagnostics > Evaluate Design.
2. Select all factors X1-X6 from Select Columns and click X,Factor. Click OK.

The Design Evaluation outline provides a number of tools to assess design properties, including:
• Power analysis for exploring effects of given
sizes
• Prediction variance profile and surface
• Alias matrix and color map of correlations
• Measures of estimation and design efficiency

Note: You can also modify the assumed model to further explore aliasing of effects.

Compare Designs
The Compare Designs platform compares two or three designs simultaneously to explore and evaluate their
performance. This allows you to select the design that best meets your requirements and budget.

1. With Plackett-Burman.jmp as the active window,

select DOE > Design Diagnostics > Compare Designs.
2. Select Definitive Screening Design to compare these
two designs, and click OK.
3. If desired, change the terms in the assumed model
and in the alias terms list.
4. Explore the many different performance metrics in the
Design Evaluation outline including Power, Estimation
Efficiency, Correlation between effects, et al.

Visit Design of Experiments Guide in JMP Help to learn more.

105
 Design and Analysis of Experiments
jmp.com/learn Nov 2022

Monte Carlo Simulation

The JMP Profiler, with the Monte Carlo Simulator, can be used to optimize process performance in the presence
of random variation. This enables you to estimate response distributions as a function of real-world random
variation. Monte Carlo simulation is available from JMP Prediction Profilers using the Simulate red triangle
option.

Example Setup
1. Use the Column Properties > Spec Limits window for the response MODULUS to add a
Lower Spec Limit = 500.
2. Run the saved script RSM for 4 Responses to simultaneously fit models for the four responses.
3. Scroll down to the Prediction Profiler, and select Optimization and Desirability > Maximize Desirability
under the red triangle to find optimal settings for the three factors.

Using the Simulator Help > Sample Data Folder > Tiretread.jmp.
This data table contains results of an analysis of a response surface design
with four responses and three factors.
1. Select Simulator from the Prediction Profiler
red triangle menu.
2. Choose to have the inputs Fixed or Random.
3. Under the Simulator outline, Add Random
Noise to all the responses. ß
4. Click the button Simulate to simulate 5000 LSL
values for each response.
Simulation results:

• Histograms and summary statistics for

the simulated values are displayed for
each response.
• The Defect rate for MODULUS is .5%.

Tips:

• Set Response Limits to optimize

responses, and set Specification Limits to
produce defect rates.
• Select distributions for input variation and noise for the response to match real conditions.
• Click Make Table to simulate values to a data table with the specified number of rows (N Runs).
• Additional options are available under the red triangle for Simulator, including Simulation Experiment,
which can optimize defect rates using Space-Filling Designs and Gaussian Process modeling.
• Right-click on the simulation results table and select Columns > PPM to report PPM values.

Visit Profilers > Simulator in JMP Help to learn more.

106
 Using SAS from JMP
jmp.com/learn Nov 2022

Connect to SAS® OnDemand for Academics from JMP®

Connect to SAS OnDemand for Academics from JMP, on a Mac or Windows machine, to browse SAS data sets,
transfer data between SAS and JMP, run stored processes, create and run a SAS program, and more.

Note: Before using the connection to SAS OnDemand for Academics, you’ll need to follow the
instructions provided at www.jmp.com/soda to:

• Register to use SAS OnDemand for Academics.

• Download the SODAJMP file, and run the installer to place the three .jar files in the
correct directory (jar files are required to access SAS OnDemand hosted servers).
• Create a SAS server profile in JMP.

Connect to SAS OnDemand for Academics from JMP

1. Open the SAS Server Connections window (File > SAS > Server Connections) and select the SAS Server
Version.
Note: If a SAS add-in has been installed, SAS Server Connections can also be accessed from the SAS menu in
the main JMP menu bar.
2. Click Connect under Metadata Server Connection. The metadata server Connection Details will update
once connected. Alternatively, click Manage Profiles and then click Connect.
3. Click Connect under Establish New Workspace Server Connection. The workspace server Current active
connection will update once connected.
4. Click Close. You are now connected to SAS!

Notes: See the SAS OnDemand for Academics website for restrictions on use of jar files and SAS OnDemand for
Academics. For more information on connecting to SAS from JMP, refer to the Using SAS from JMP one-page
guides at jmp.com/learn.

107
 Using SAS from JMP
jmp.com/learn Nov 2022

Connect to SAS® on Your PC (Windows) from JMP®

Connect to a local installation of SAS from JMP to browse SAS data sets, open SAS data sets in JMP, run stored
processes, create a SAS program, submit SAS code, and more. This SAS connection is available only on the
Windows operating system.

Connect to a Local Installation of SAS on Your PC (Available on Windows Only)

1. Open the SAS Server Connections window (File > SAS > Server Connections).
Note: If a SAS add-in has been installed, SAS Server Connections can also be accessed from the SAS menu in
the main JMP menu bar.
2. Select Connect to SAS on this machine.
3. Click Connect.
When the connection to SAS has been established, “Local” will display under Current active connection.
4. Click Close. You are now connected to SAS!

3
1

Notes: For more information on connecting to SAS from JMP, refer to the Using SAS from JMP one-page guides
at www.jmp.com/learn. Or, click on Learn more about connecting to SAS from the SAS Server Connections
window or search for “connecting to SAS” in JMP Help.

108
 Using SAS from JMP
jmp.com/learn Nov 2022

Connect to Your School’s SAS® Server from JMP®

Connect to your school’s SAS server from JMP, on a Mac or Windows machine, to browse SAS data sets, transfer
data between SAS and JMP, run stored processes, create and run a SAS program, and more.

Connect to Your School’s SAS Server

1. Open the SAS Server Connections window (File > SAS >
Server Connections) and select the SAS Server Version.
Note: If a SAS add-in has been installed, SAS Server
Connection can also be accessed from the SAS menu on
the main JMP menu. 3 2
2. Click Manage Profiles..., and select a SAS server profile
from the list.
If no profile exists, see the shaded box below.
3. Click Connect.
The metadata server Connection Details will update
once connected. 4
4. Click Connect to metadata-defined SAS server, and
select the desired workspace server.
Note: If your school does not have a metadata server,
you can connect directly to a remote SAS server. See
5
your SAS administrator for details.
5. Click Connect under Establish New Workspace Server
Connection.
6
The workspace server Current active connection will
update once connected.
6. Click Close.

You are now connected to SAS!

How to create a metadata server profile:

1. In the SAS Server Connections window, select the version for the SAS server.
2. Click Manage Profiles...
3. Click Add to add a new profile.
4. Fill in the information as required, and click Save.
Click on Learn more about metadata profiles for help. See your SAS administrator for profile information.

Notes: For more information on connecting to SAS from JMP, refer to the Using SAS from JMP one-page guides
at www.jmp.com/learn. Or, click on Learn more about connecting to SAS from the SAS Server Connections
window or search for “connecting to SAS” in JMP Help.

109
 Using SAS from JMP
jmp.com/learn Nov 2022

Setting SAS® Integration Preferences

Specify how JMP® works with SAS by modifying the default SAS Integration settings. For information on
connecting to SAS, see the Connect to SAS one-page guides under Using SAS from JMP.

Setting SAS Integration Preferences

From an open JMP window, select File >
Preferences (or JMP >Preferences on a Mac).
Under Preference Group select SAS Integration.
Selection options include the following:
1. SAS Server Version: Select the version of the
SAS server you will be connecting to. This
option does not apply when connecting to
SAS on a local machine.
I want to connect directly to a SAS
Metadata Server is checked by default.
Clear this option if you do not have a SAS
Metadata Server available.
2. SAS Submit Options
a. Select Automatically Generate ODS Results if you want to generate Output Delivery System (ODS)
results. You may specify:
• ODS Result Format: HTML, RDF, PDF or JMP report.
• ODS Style: The style name for ODS reports (Statistical is the default).
• ODS Style Sheet: The style sheet name for ODS reports.
• Prompt if results are larger than: The size in MB. If exceeded, will prompt to continue or cancel.
• Graphics Format: ActiveX (Windows only), Java, PNG, JPEG or GIF.
• Select Generate ODS statistical graphics to include statistical graphics in ODS reports.
b. Import generated SAS data sets into JMP: If selected, any SAS data sets generated by submitted SAS
code will automatically be imported into JMP. Large data sets prompt for confirmation.
3. SAS Data Import: You can select to
a. Use SAS variable labels for column names during data import (instead of SAS column names).
b. Convert SAS custom formats to JMP value labels when importing.
c. Warn before closing unsaved imported data (triggers to save before closing).
4. Stored Process Results: Specify the following formats:
a. Format (of reports): HTML, RTF or PDF.
b. Graph Format: ActiveX (Windows only), Java, PNG, JPEG or GIF.
5. Show SAS Log:
a. Always, Never or On Error.
b. Location: In the JMP Log window, or in a separate window.

Note: For more information on using SAS with JMP, search for “SAS integration” in JMP Help.

110
 Using SAS from JMP
jmp.com/learn Nov 2022

Using JMP® to Generate SAS® Programs

Automatically generate SAS programs using the JMP Fit Model and Time Series (ARIMA or Seasonal ARIMA)
platforms. SAS scoring code can also be generated from the Partition and Neural platforms (not covered here).

Using JMP to Generate SAS Programs

To generate SAS programs with the Fit Model platform:

• From an open data table select Analyze > Fit

Model.
• Specify a model: Select your response (Y) and
predictors (Model Effects).
• Specify the type of model under Personality.
• Click on the top red triangle and select Create
SAS Job (see resulting code, bottom right).
• Note: Will only produce simple PROC GLM, PROC
Logistic, PROC GENMOD or PROC Mixed code.
To create SAS code for ARIMA models from within the
Time Series platform:
• From an open a data table select Analyze >
Specialized Modeling > Time Series.
• Select the Y, Time Series column and columns for
other roles (if needed), and click OK.
• Click on the red triangle and select ARIMA or
Seasonal ARIMA. Specify the model, and click
Estimate.
• Click on the red triangle for the model and select
Create SAS Job.

JMP generates the SAS code for the specified model in a

new SAS Program Editor window. To submit code to SAS,
right-click and select Submit to SAS.
Notes:
• To submit SAS code, JMP must be connected to SAS.
For information on connecting to SAS, either locally
or on a server, see the page Using SAS from JMP.
• For more information on using SAS with JMP, search
for “SAS integration” in the JMP Help.
• For more information on using the Fit Model or Time
Series platforms, visit Fitting Linear Models and
Predictive and Specialized Models in JMP Help.

111
 Using SAS from JMP
jmp.com/learn Nov 2022

Entering and Running SAS® Programs

Write and submit your own SAS programs or submit existing SAS programs. This enables JMP® to be used as a
client to SAS, leveraging the depth and breadth of statistical and analytical techniques in SAS.

Entering SAS Programs

1. To bring an existing SAS program into JMP:

• From an open JMP window, select File > Open.

• Browse to the desired SAS program file, and
click Open. The SAS file will open in the SAS
Program Editor window in JMP.
2. To create (or cut and paste) a new SAS program
from within JMP, select File > SAS > New SAS
Program to open the SAS Program Editor.

The SAS Program Editor is color-coded for easy

reference:

• Dark Blue: PROC and DATA statements.

• Blue: Keywords.
• Magenta: Text strings.
• Green: Constants.

Running SAS Programs

To run the SAS program from the SAS Program
Editor in JMP:
• Select Edit > Submit to SAS. (If prompted to
connect to SAS, enter the appropriate details
and click OK – see the first note below).
• The code will be sent to SAS, and the SAS
output will be returned to JMP according to the
SAS Integration preference settings.
• To view the SAS log and/or output windows if
they are not already open, select File > SAS >
Open SAS Log Window or File > SAS > Open
SAS Output Window, respectively.

Notes:
To submit SAS code, JMP must be connected to SAS. For information on connecting to SAS, either locally or on a
server, see the Connecting to SAS one-page guides under Using SAS from JMP at jmp.com/learn. For
information on setting SAS preferences, see the Setting SAS Integration Preferences one-page guide.

112
For JMP sales in the US and Canada, call 877 594 6567 or go to jmp.com

JMP and all other JMP Statistical Discovery LLC product or service names are registered trademarks or trademarks of JMP Statistical Discovery LLC
in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies.
Copyright © 2022 JMP Statistical Discovery LLC. All rights reserved. 107189