Bar Graphs and Histograms
Bar Graphs and Histograms
Introduction
Bar graphs are created in much the same way scatter plots and line graphs are.
Histograms are a specialized type of bar graph used to summarize groups of data.
Data is entered into Excel much in the same way as it is with scatter plots and line
graphs:
Note that the independent variable is placed in the first column while the dependent
variable is placed in the second column. The headers at the top of each column are not
necessary, but they do help identify the variables.
With the data you want graphed highlighted, start the chart wizard:
If the Chart Wizard is not visible, you can also choose Insert > Chart...
Enter your titling. Also make sure to go to the Legend tab and click off the Show
Legend option. You will not need a legend with only one independent variable. Click
Next or Finish when you are done.
Return to Top
A multiple bar graph depicting data using two independent variables is created in the
same way as a simple bar graph:
• The first independent variable, mammal, is still in the first column, with the
dependent variable values in columns two and three. The second and third
columns represent dependent variable values at two different levels of the
second independent variable, week.
• Make sure to select all of the data when creating the graph. The Chart wizard
will automatically recognize you have a second independent variable.
• When you get to the last step of the Chart wizard, keep the legend turned on,
since it shows the coding for the two levels of the second independent variable.
Return to Top
The relative widths of bars to the gap between the bars can be adjusted by double-
clicking on one of the bars in the graph:
• Gap width represents the spacing between bars as a percentage of the width of
one bar
• Overlap will overlap touching bars as a percentage of bar width.
Return to Top
Introduction to histograms
In some investigations, you may find yourself collecting a large number of data points
for a single level of an independent variable. That is, you take the same measurement
over and over again. You would do this because a lack of perfect precision in your
Bar Graphs and Histograms
Page 5 of 9
measuring process would not let you get a good estimate of the true value with only a
single measurement. In this example, the fracture stress of a certain type of glass bar is
measured 24 times:
Clearly, the measured stress is not the same for each sample. In fact, the measurements
range from a low of 55 to a high of 169. How can you summarize the results of these
measurements? One way might be to simply calculate the average (mean) of all these
measurements. This would not, however, give you a good feel for how the data is
distributed. A distribution graph, or histogram, allows you to see how many
measurements fall within set ranges, or bins, of the dependent variable. A histogram is
usually depicted as a bar chart, with one bar representing the count of how many
measurements fall with a single bin.
Return to Top
To start with, it is usually a good idea to scan your data and get a feel for its overall
range. For the data above, the range is from 55 to 169. Next you will want to decide
how fine you want the increment of your bins. The finer the increment, the more bins,
and thus the more bars on your chart. For this example we will choose a bin increment
of 15 starting with 70. Depending on what you want to depict, you may want to show an
empty bin above and/or below the extreme values of your samples to show the viewer
that you are at the extremes of your data set. Type in these bin increments in a column
next to your raw data:
The frequency values should now fill the cells next to the bin increments. Note that your
first bin increment, 70, holds all the measurements at 70 and below. The next bin, 95,
holds measurements from 71-95, and so on. The result should look like this:
If only the top cell is filled with a frequency value, then you probably either didn't
highlight the range of cells next to the bin increments, or you didn't use the special key
combination to enter the function. Note that in the above figure, the last bin value, 195,
was changed to More to indicate that it represents the count for everything above 170.
Return to Top
Creating a histogram
You can now create a bar graph as you did above using the histogram summary data
rather than the raw data:
Just as you can with other data, you can create a multiple bar histogram. You can either
do this as was shown above or by superimposing two histograms (see the Advanced
module on superimposing graphs).