1

e
ut
PSYCHOLOGICAL RESEARCH

b
tri
is
The Whys and Hows of the Scientific

rd
Method and Data

CONSIDER THE FOLLOWING

,o
st
QUESTIONS AS YOU READ CHAPTER 1
po

• What is the value of research in psychology?

• Why do psychologists conduct research?
y,

• What is the difference between a population and a sample?

op

• What kinds of data are collected in psychological studies?

• What is a distribution, and how does its shape affect our analysis of the data?
tc
no

LEARNING OBJECTIVES FOR CHAPTER 1

o

• Understand that knowledge of research in psychology has value beyond

careers in research.
D

• Understand what it means to learn about behavior through observation.

• Identify different measurement scales.
• Examine data using frequency distributions.

1
Copyright ©2020 by SAGE Publications, Inc.
This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
2   The Process of Research and Statistical Analysis in Psychology

A s an instructor of an introductory psychology course for psychology majors, I have

asked my first-semester freshman students this question: What is a psychologist? At
the beginning of the semester, students typically say that a psychologist listens to other
people’s problems to help them live happier lives. By the end of the semester and their first
college course in psychology, these same students will respond that a psychologist studies
behavior through research. These students have learned that psychology is a science that

e
investigates behaviors, mental processes, and their causes. That is what this book is about:
how psychologists use the scientific method to observe and understand behaviors and

ut
mental processes and how they understand those data using statistics.
The goal of this text is to give you a step-by-step approach to designing research in

b
psychology—from the purpose of research (discussed in this chapter) and the types of
questions psychologists ask about behavior; to the methods used by psychologists to

tri
observe and understand behavior as well as the statistical tools they use to interpret the
data collected about behavior; and finally, how psychologists describe their findings to

is
others in the field.

rd
WHY SHOULD I CARE ABOUT
RESEARCH IF I DON’T WANT TO
DO RESEARCH IN MY CAREER? ,o
st
Through my years of teaching psychology methods courses, this question is often asked
by students who don’t think they want to conduct research in their careers. A few of you
po

might be bitten by the “research bug,” as I was as an undergraduate, and find research to
be an exciting way to answer questions you have about behavior. Knowing the process of
research can help you better understand the topics presented in other psychology courses
you may take because you will better understand how this information was gained.
y,

However, a majority of students majoring in psychology are interested in working as a

practitioner of psychology or may be completing a psychology minor that is related to
op

another career they want to pursue (education, social work, criminal justice, etc.) and do
not understand why research methods courses are part of their curriculum. In fact, the
majority of individuals who hold a degree in psychology do not conduct research in their
tc

jobs. Instead, the majority of individuals working in psychological areas are in helping or
other applied professions. However, what we know about behavior in everyday settings
no

comes from research findings. For example, effective treatments and counseling tech-
niques come from research in these areas. When a new treatment technique is tested, its
effectiveness is determined by the research conducted on it. Thus, just as medical doctors
do, clinicians and counselors must evaluate the latest research in psychology to deter-
o

mine whether a new treatment is one they should adopt. Knowledge of how research is
conducted can help them evaluate this research more effectively to aid their practice. In
D

addition, other popular applied areas, such as industrial–organizational psychology and

human factors, use research findings to help address issues in everyday life. Industrial–
organizational psychologists help organizations hire effective employees, prevent job
dissatisfaction, and explore the best training methods for new employees using research
findings on these topics (see Photo 1.1). Human factors professionals use research to help

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  3

understand the best way to design prod-

ucts and interfaces (such as an airplane
cockpit—see Photo 1.2) to make them
easier to use and prevent errors. Finally,
it is important that we as individuals
understand how to interpret the vast

e
amounts of information we take in each
day through media sources. Research

ut
©istockphoto.com/fizkes
findings are reported by the media every
day. Knowing the basics of how research

b
is conducted can help you decide which
of those reports you should listen to and

tri
which are best ignored. Understanding
research can also help you figure out how

is
to decide whether you believe something
you read about on social media or not.

rd
Learning how to investigate what people
know about a topic is part of the research
process.
To give you a recent example, in
debates about climate change and the
seriousness of the problem, many oppo-
,o
st
nents of climate change solutions point

©istockphoto.com/choja
out that there is disagreement among sci-
po

entists about the cause. My own father

once told me that this is the reason that
he doesn’t believe global warming is
caused by human activities—some scien-
y,

tists have stated that there isn’t enough Photos 1.1 and 1.2
evidence. As voters and consumers, it is important that we understand which evidence Knowledge of
op

research can aid

from research is the most valid (i.e., accurate) and that there will almost always be dis- in applied areas of
agreement among researchers in an area because no single study can fully answer a psychology, such
research question. In order to understand what answers research provides on a question, as industrial–
tc

organizational
we must consider the accumulation of data in many research studies (and this is what I psychology and
told my father when he stated his reasoning to me about his beliefs). We must also under- human factors.
no

stand that new knowledge is always being discovered, and we must be flexible in our
conclusions about an issue when new data suggest a different answer. Remember, there
was a time when most humans believed the sun revolved around the earth. Scientific
study revealed this idea to be false, and over time, humans adapted their beliefs to the
o

new knowledge. We must do the same when we learn new findings about the best every-
day behaviors, such as how to prevent Alzheimer’s disease or how to keep our hearts
D

healthy and live longer.

Understanding research methods can also help you better interpret research study
results that are reported in the media. In almost all cases, media sources present concise
and simplified reports of a research study and its results, leaving many questions about
the quality of the study still to be answered. When one encounters reports of research

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
4   The Process of Research and Statistical Analysis in Psychology

in the media, some important questions should come to mind. Who were the research
subjects? Was an appropriate sample tested? Was an appropriate method used to inves-
tigate the question? Were the results published in a high-quality source where other
researchers were able to critique the work? How do the results correspond to past studies
on this topic? The topics covered in this text and in your methods course will help you
ask and answer these questions as you evaluate reports in the media that you can use to

e
make decisions about your life.
Finally, the new knowledge you gain from your study of research methods can help

ut
you decide how to evaluate claims made by others in general. When you see an ad on
television for a new miracle diet pill that the ad claims has helped people lose weight

b
in studies, should you buy the pill? When your friends tell you that drinking energy
drinks helps you study better and achieve higher scores on exams, should you follow their

tri
advice? Should you believe claims that vaccines cause autism? (You shouldn’t: There’s no
valid research evidence that vaccinations cause autism.) Hopefully, one of the things you

is
will consider as you learn about research is to be skeptical about claims that seem too
good to be true. A good researcher uses the data to decide what is the best thing to do

rd
rather than use unsubstantiated advice from others who just sound knowledgeable about
a topic but who cannot provide evidence beyond an anecdote or two. Examples of how
to evaluate claims and research reported in the media are given in some of the Applying

,o
Your Knowledge sections found at the ends of the chapters in this text.
st
WHY PSYCHOLOGISTS
po

CONDUCT RESEARCH
Think about how you know the things you know. How do you know the earth is round?
How do you know it is September? How do you know that reading over your notes will
y,

help you prepare for an exam? How do you know that terrorist threats are increasing
around the world? There are probably many ways that you know these things. In some
op

intuition: Relying cases, you may know things because you used your intuition or previous knowledge that
on common sense led to deduction of these facts. For example, you may know from past experience that
as a means of
where you live, in the month of September, days tend to be warm but start to get cooler,
tc

knowing about the

world especially at night. Therefore, remembering the characteristics of the weather you are
deduction: Using experiencing and knowing you are still living in the same location as past years, you can
deduce that the month is September from your knowledge base. You can also consult
no

logical reasoning
and current a calendar online, using technology as an authoritative information source. You may
knowledge as a
means of knowing have first learned that the earth is round from an authority figure such as your parents,
about the world teachers, or text authors. You may have also observed that the earth is round by view-
o

authority: ing photographs of the earth taken from space. You may know that terrorist threats are
Relying on a increasing from authority figures as well (e.g., magazine and newspaper reporters, your
D

knowledgeable
country’s leaders’ statements). These are the primary ways that we learn new facts: intu-
person or group
as a means of ition, deduction, authority, and observation.
knowing about the Suppose something occurred that caused you to suspect that the authority figures you
world
have learned these facts from are not reliable sources of information. Perhaps they have
been caught lying about other facts. You might also consider a situation where you do not

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  5

have enough previous experience with a topic to deduce the information for yourself. In
these situations, what is the best way for you to find the facts? The answer is observation. observation:
If you had reason to believe, for example, that an increase in terrorist threats is not being Relying on what
one observes as a
represented accurately, you could examine the incidence of terrorist attacks (e.g., from means of knowing
public records) over a period of time to find out if people are representing the true con- about the world
ditions. Observing the world directly is going to give you the most accurate information

e
because you are directly gaining the knowledge yourself—you are not relying on possibly
faulty reasoning on your part or information someone may be giving you that is false or

ut
misleading. See Table 1.1 for some examples of the different ways of knowing information.
This is why psychologists conduct behavioral research; it is the best way to make

b
certain that the information they have about behavior is accurate. By conducting careful
and systematic observations, they can be certain that they are getting the most accurate

tri
knowledge they can about behavior. This does not mean that every study conducted will
yield accurate results. There are many cases where the observations collected by differ-

is
ent researchers conflict, but this is an important part of the process. Different ways of
observing a behavior may yield different observations, and these different observations

rd
help us to better understand how behaviors occur. Over time, with enough observations,
a clearer answer to the question can be found. But no single research study can “prove”
that something is true. Researchers are not able to “prove” facts with a study; the best

,o
they can do is support an idea about behavior with their data. Despite the limits of
observation as a way of knowing, it is superior to the other methods because it allows
for a more objective way of gaining knowledge. Relying on the other ways of gaining
st
knowledge can be misleading because they can be more easily influenced by biases that
people have.
po
y,

TABLE 1.1 ■ Examples of Ways of Knowing Information

op

Way of knowing Example

Intuition I’m trying to go someplace I’ve never been, but I do not know the
tc

way. I decide to turn left because it just “feels like” that’s the right
way to go.

Deduction I want to know which direction I am facing. The sun is setting to my

no

right, and I know the sun sets in the west, so I know that south is the
direction I am facing.

Authority I want to know what my pancreas does. I know that my pancreas

o

produces hormones important for digestion because that is what

my high school biology teacher told me.
D

Observation I want to know how much sleep on average Americans get per night.
I determine this by conducting a survey of Americans to learn that
most Americans get an average of 6 to 8 hr of sleep per night (e.g.,
Moore, 2004).

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 6   The Process of Research and Statistical Analysis in Psychology

Using Science to Understand

and Explain Behavior
Observation is really what sets scientific
fields apart from other fields of study.
Someone who wants to know about the
political situation during the Civil War

e
may read historical documents and use
©istockphoto.com/andresr

ut
his or her intuition to describe the sit-
uation based on these documents. He

b
or she might also read books by experts
(authority figures) on the Civil War

tri
period or books on important figures
who lived during that time. However,

is
Photo 1.3 historians typically cannot observe the
If we want to
historical event they are studying. Psychologists have an advantage in that the behavior

rd
know how much
sleep people they want to learn about is happening in humans and other animals in the world around
get, we can use them. The best way to learn about it is to just observe it (see Photo 1.3).
scientific methods
Some behaviors, such as mental processes, cannot be directly observed (e.g., attitudes,

,o
to measure this
directly or ask thoughts, or memories). Thus, psychologists have developed techniques for inferring
people to report information about mental processes through observation of specific behaviors that are
this behavior on a
affected by these mental processes. Psychologists then attempt to understand mental pro-
st
survey.
cesses through observation of these behaviors and the investigation of the factors that
influence those behaviors. That is what this book (and the course you are taking) is all
po

about—understanding the methods psychologists use to observe, measure, and under-

stand behavior and mental processes.
Research is the foundation of the field of psychology. Many people think of the helping
professions when they think about what psychologists do. This is because most peo-
y,

ple with a graduate degree in psychology work in these helping (or related) professions
(American Psychological Association [APA], 2003). However, to do their jobs well, help-
op

ing professionals, such as clinicians and counselors, need to understand the findings from
research about behavior so that they know what types of treatments and therapies can best
help their clients. The research studies conducted in psychology also help clinicians and
tc

counselors understand what constitutes “normal” behavior and what behaviors might be
considered “abnormal.” Psychological research also informs society about issues related
to these behaviors. How do we know when to trust eyewitness testimony based on one’s
no

memory and when it is likely to be inaccurate? What is the best way to set up a course to
help students remember the information over the long term? What are some daily activi-
ties that older adults can engage in that make them less likely to develop dementia? These
o

are some of the applied questions that can be answered with research in psychology.
Knowing the results of studies in different areas can help us, as a society, work out some
D

answers to these questions that will be helpful in different realistic situations.

Thinking about the field of biology may help you understand how influential research
is in the field of psychology. In the biological field, there are researchers who investigate
the way our bodies react physically to the world around us (e.g., after being exposed to
a virus). This knowledge helps other researchers determine which drugs may be effective

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  7

in helping us improve these physical reactions (e.g., reduce our symptoms as we fight
the virus). Finally, the knowledge gained in biological research helps doctors correctly
diagnose and treat their patients (e.g., what symptoms indicate the presence of a par-
ticular virus and which drugs are most effective in treating these symptoms). The field
of psychology works a lot like the field of biology (although the term psychologist applies
to both scientists and practitioners in psychology, sometimes causing confusion). Some

e
researchers investigate what causes certain types of behaviors (e.g., distraction in peo-
ple with attention deficit hyperactivity disorder [ADHD]). Other researchers investigate

ut
what treatments are effective in reducing these behaviors (e.g., rewarding someone for
staying on task). Finally, some psychologists work with clients to help them deal with

b
problem behaviors. For example, school psychologists work with teachers and parents
to develop a reward system for students with ADHD who have difficulty completing

tri
work in class because they become easily distracted. The research that investigated the
behaviors associated with ADHD and the factors that can reduce those behaviors was

is
necessary in order for the school psychologist to be able to develop an effective treatment
plan for the student.

rd
In the next section, we will begin to examine how observations of behavior occur in
research studies.

STOP AND THINK ,o

st
1.1. Think about some things you know are true to become anxious. Describe how you
po

about the world. For each of these facts, might learn about these factors by using
try to determine the way you know that observation.
information (intuition, deduction, authority,
1.3. Explain how the fields of psychology and
or observation).
y,

biology are similar.

1.2. Suppose you wanted to know about
the factors that cause college students
op
tc

POPULATIONS AND SAMPLES

Have you ever wondered about the opinion polls presented by media news sources and
no

how accurate they are? Consider a poll done on global warming by ABC News/Washington
Post taken in November 2015 (http://www.pollingreport.com/enviro.htm). People
were asked if they considered global warming a serious problem facing the country
o

(see Photo 1.4). From those polled, 63% said yes, it was a serious problem (the highest
response on this measure). We can compare this with a poll conducted in May and June
D

of 2015 by the Pew Research Center (http://www.pollingreport.com/enviro.htm). In this

poll, people were asked if global warming was a serious problem, and only 46% said it was
very serious (the highest response on this measure). Why is there such a large difference
between the reported percentages? Did many people suddenly decide that global warm-
ing was a big problem between June and November? This is one possible explanation,

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 8   The Process of Research and Statistical Analysis in Psychology

but it is not very likely. Another possi-

ble explanation could be that different
people answered the question in the two
polls. One way to determine this is to
look at the information provided about
the polls. The ABC News/Washington

e
Post poll describes that it was from
1,004 adults nationwide (in the United

ut
States) with a margin of error of plus or
minus 3.5. The Pew Research Center
©iStock/pum_eva

b
poll describes that it was from 5,122
adults nationwide with a margin of error

tri
of plus or minus 1.6. Is this information
important? What does it tell us about

is
Photo 1.4
the validity of the polls?
How many people
In fact, the information provided about the polls can be important in deciding whether

rd
believe that global
warming is a the information from the poll is accurate (i.e., valid). The Pew Research Center poll sur-
serious problem?
veyed more people, which allowed for a smaller margin of error. This means we can be
Polls can provide
more certain that the percentage of all adults in the United States who would report

,o
some information
about this, but global warming as a very serious problem is close to 46%. What we’re looking at here is
it is important
the difference between a population and a sample, and the information provided helps
to consider the
us determine how well the sample represents the whole population (see Figure 1.1). The
st
sample and
population for the population in this case is all adults in the United States. This is the group of individuals
poll to determine
we are trying to learn about with the poll. The sample is the set of people who answered
po

its validity.
the question in the poll. They were selected from the population in an attempt to repre-
sent the opinions of the whole population without having to ask the whole population
validity: The
(which would be impossible, given the population’s size). How well the sample represents
the population of interest is a function of the sample size, the way in which the sample
y,

accuracy of the
results of a study was chosen, how many people chosen actually responded or chose to participate, and a
few other factors that researchers must consider when conducting any type of research
op

population:
A group of study. Differences between the sample and the population contribute to sampling error.
individuals a
researcher seeks
Sampling error exists any time we collect data from a sample of the population because
we will never be able to get the exact population data from a sample of the population.
tc

to learn about from

a research study Each sample is a different subset of the population and will provide different scores,
sample: The group resulting in different means across samples. We’re trying to estimate the population mean
no

of individuals
chosen from the
using the sample mean, so with each sample, we will be at least a little bit wrong about
population to the population mean. This difference between the sample mean and the population mean
represent it in a is the sampling error.
research study
One way to think about this is to imagine someone trying to figure out what pic-
o

sampling error: ture is on a puzzle with only a small part of the puzzle put together. We’re not likely to
The difference
understand the whole picture from just one part of the puzzle; however, the more of the
D

between the
observations in puzzle we have together (i.e., the larger the sample size), the better we’ll be at guessing the
the population picture. But putting the whole puzzle together would be difficult and time-consuming
and in the sample
that represents
if there were thousands or millions of pieces. So instead, we make our best guess from
that population in the part of the puzzle we can put together in our study (the sample). With only a small
a study part of the puzzle put together, though, we’re not going to get the whole picture exactly

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  9

FIGURE 1.1 ■ The Sample Is Chosen to Represent the Population in a Study

e
b ut
tri
Sample

is
rd
Population

,o
right. Thus, there is some error in our estimate of the population data from our sample
data (sampling error). The margin of error reported for polls (such as those described here)
provides an estimate of the sampling error. This is an estimate of how far off the reported
st
percentage in our data is likely to be from the population percentage. Thus, knowing the
sample size and margin of error for opinion polls can help you decide if the poll is useful
po

in telling you how people really think about an issue.

It is very rare that a researcher will observe the entire population in a research study,
so samples are almost always used in order to represent the population in the study by
y,

collecting data from a realistic number of individuals. This is due to the size of most pop-
ulations; they tend to be very large because researchers want to be able to learn about the
behavior of large groups of individuals, not just a small set of people—as in our survey
op

examples where we wanted to know about how Americans felt about global warming.
Thus, a smaller sample was chosen to collect data from. If you have ever participated in
a research study, then you have been a member of a sample selected from a population.
tc

Many research studies in psychology use samples selected from the population of college
students because it is a sample that is fairly easy to obtain. However, college students is
no

too narrow a population for some studies because college students tend to be fairly edu-
cated, higher income, and young, giving us a biased sample for these characteristics from
the whole population of adults. Because of this possible bias, researchers have begun
to sample from larger populations of individuals using online technologies to deliver
o

surveys and experimental tasks to a larger population. For example, a fairly recent study
D

by Brown-Iannuzzi, Lundberg, Kay, and Payne (2015) sampled their participants from
a large population of adults using the Amazon site Mechanical Turk (MTurk). This site
rewards people with small amounts of money in their account for completing research
studies. Brown-Iannuzzi et al. sampled individuals from this site to learn how one’s sense
of one’s own wealth, relative to others’ wealth, influences one’s political ideas about the
redistribution of wealth in society (their research suggested that relative wealth does

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
10   The Process of Research and Statistical Analysis in Psychology

FIGURE 1.2 ■ R
 esults From Brown-Iannuzzi et al.’s (2015) Study Comparing
Support for Redistribution of Wealth Based on Experimental
Conditions of Relative Wealth Compared With Others’ Wealth;
Greater Support for Redistribution Was Found for Low-Status
Participants Than for High-Status Participants

e
2

ut
1
Mean Support for Redistribution

b
Low-Status Control High-Status
−1

tri
of Wealth Score

−2
−3

is
−4

rd
−5
−6
−7
−8
−9
,o
st
Status Condition
po

Source: Brown-Iannuzzi et al. (2015).

y,

affect one’s political ideas on this issue; see Figure 1.2). MTurk is becoming a popular
method for selecting a large sample from the very large population of adults in the world.
op

We will consider more issues of sampling in Chapter 4.

The population and sample of a research study are important for choosing the statis-
tical tools researchers use to better understand data. As you will see in Chapter 4, the
tc

way that you choose a sample from a population can affect the validity of your study.
Discussions in the later chapters of this text will show you how hypotheses are made
no

about populations before a sample is chosen. Finally, the type of data collected from the
descriptive sample will influence the statistics we choose to summarize the data (i.e., descriptive
statistics: statistics) and to test our hypotheses about the population (i.e., inferential statistics).
Statistics that
help researchers
o

summarize or
describe data
TYPES OF DATA
D

operational
definition: The There are many different types of data that can be collected in psychological research stud-
way a behavior ies. Researchers attempt to choose the best measure for the behavior they want to observe.
is defined in a This choice is important because it can affect the internal validity of the study—a poor
research study
to allow for its choice can mean that the data do not actually measure the behavior of interest. In other
measurement words, researchers try to determine a good operational definition of the behavior they are
Copyright ©2020 by SAGE Publications, Inc.
This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  11

interested in. An operational definition

is the way a behavior is measured in a
particular study. It also provides a way for
a researcher to measure a behavior that
is not directly observable. Operational
definitions are a necessary part of the

e
research process because many behav-
iors can be defined in multiple ways,

ut
©iStock/AntonioGuillem
and researchers need to know what to
measure from the individuals in the

b
sample when they collect their data. For
example, what behaviors should be

tri
measured to learn about one’s level of
depression (see Photo 1.5)? There are

is
Photo 1.5
many ways we could operationally
There are many
define depression: how often someone smiles in an hour (fewer smiles = more depression),

rd
ways to measure
observers’ ratings of how lethargic someone seems (lower ratings of energy level = more depression;
a researcher
depression), or a score on a questionnaire of self-reported thoughts and behaviors that we
operationally
think are present in someone who is depressed (more sleep than is typical, loss of appetite,

,o
defines the
feelings of sadness, etc.). Thus, researchers have many choices when they want to measure behavior to allow
its measurement.
depression, and they try to come up with the most valid measure (within the practical
limitations of the research study) for the behavior they want to learn about.
st
Scales of Measurement
po

Observations of behavior (i.e., data) in a study constitute what is called a dependent dependent or
or response variable. Dependent variables are measured in every research study. For response variable:
A variable that
some designs, only a single dependent variable is measured, and the behavior is examined
y,

is measured or
descriptively or causally (if the researcher is interested in a causal relationship and uses observed from an
an experiment to study this relationship). Other designs examine relationships between individual
op

multiple dependent variables. As a result, how a dependent variable is measured depends

on the data collection technique used, and what is learned about a dependent variable
depends on the type of research design used. One choice that is made by the researcher
tc

in operationally defining a behavior for a research study is the scale of measurement he

or she uses. The scale of measurement will be important in determining which statistics
are used to describe the data and test hypotheses about the data. Table 1.2 presents an
no

overview of the different scales of measurement and an example of each type.

Nominal Scales
o

Many measurements categorize observed behavior or behavior that is reported by the

D

participants. For example, if you ask someone to indicate their current mood, you might
give them response choices such as happy, sad, excited, anxious, and so on. This type of
nominal scale:
measurement is called a nominal scale because it involves nonordered categories that are A scale of data
nonnumerical in nature. You cannot order these moods from highest to lowest—they are measurement
that involves
simply different. Some types of demographic information collected in research studies
nonordered
are measured on a nominal scale. Questions about someone’s gender or major in college categorical
are good examples of these types of nominal scales. responses

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
12   The Process of Research and Statistical Analysis in Psychology

TABLE 1.2 ■ Scales of Measurement

Scale Definition Example

Nominal Unordered categories University where degree was

earned

e
Ordinal Categorical, ordered categories Letter grades earned in a course
(A, B, C, D, F)

ut
Interval Numerical categories without a true Ratings on personality surveys with

b
zero point values from 1 to 5

tri
Ratio Numerical categories with a true Age measured in days since birth
zero point

is
rd
Ordinal Scales
ordinal scale: Anytime you rank order your preferences for different things, you are using an ordinal
A scale of data scale. Rankings in a competition (1st, 2nd, 3rd, etc.) also measure individuals on an

,o
measurement that
involves ordered
ordinal scale. Ordinal scales are measures that involve categories that can be ordered from
categorical highest to lowest. However, the ordered categories are not necessarily equally spaced on
responses an ordinal scale. Imagine you are asked to report your level of anxiety today on a scale
st
that includes response choices of not at all anxious, a little anxious, fairly anxious, and very
anxious. On this scale, the difference between a little anxious and fairly anxious may be
po

smaller than the difference between fairly anxious and very anxious. Thus, this is consid-
ered an ordinal scale because the categories can be ordered from highest to lowest level of
anxiety, but the categories are not always equally spaced across the scale.
y,

Interval Scales
op

If the ordered categories on a scale are equally spaced, then the scale is known as an
interval scale: interval scale. Interval scales are used when the researcher wants to know that the differ-
A scale of data ence between any two values on the scale is the same across all values of the scale. Typically,
tc

measurement that
involves numerical
this involves numerical responses. Many rating scales on questionnaires are interval scales
responses that are because they ask participants to rate their agreement with statements or their likelihood of
equally spaced, performing specific behaviors on a numerical rating scale. An example of such a scale might
no

but the scores are

not ratios of each be to rate how much you agree with the statement “Global warming is a serious issue facing
other society today” on a scale of 1 to 10, where a higher number indicates higher agreement.
Such scales do not have a true zero point, because the values cannot be considered ratios of
o

one another. For example, because there is a minimum and maximum score on a 1 to 10
scale, there is no way to determine the ratio function between scores—a score of 4 is not
D

twice that of a score of 2 (2 is only one value higher than the minimum, whereas 4 is three
ratio scale: values higher than the minimum). The scores on the scale are not distributed in this way.
A scale of data
measurement that
involves numerical Ratio Scales
responses in which
scores are ratios A numerical scale with a true zero point allows for values that are ratios of one another.
of each other This is known as a ratio scale. On ratio scales, you can determine what score would be
Copyright ©2020 by SAGE Publications, Inc.
This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  13

twice as high as another score. Some examples of ratio scales are accuracy on a task, speed
to complete a task, and age. A score of 50% accuracy is twice as high as a score of 25%
accuracy. Ratio scales are often used in systematic and controlled measures of behavior,
a topic we will discuss later in this chapter. Note that interval and ratio scales are often
grouped together because data from these scales are typically analyzed in the same way,
and there are cases where it is difficult to determine if a scale is truly an interval scale. The

e
important difference to note for the scales of measurement and how they are analyzed is
whether they involve numbers or categories as responses on the scale.

b ut
STOP AND THINK

tri
is
1.4. For each study description that follows, the participants rate their level of
identify the population and the sample. anxiety on a 1 to 7 scale for different

rd
a. A researcher recruits students from issues that might be anxiety inducing
a fifth-grade class at an elementary (e.g., financial security, failing health).
school to examine math abilities on a d. Patients who have suffered a traumatic
standardized test in children who are
9 to 10 years old.
,o
b. A researcher recruits college students
brain injury (TBI) are included (with
their consent) in a study of how one’s
diet after the injury affects recovery
st
from the university subject pool to test time (measured in number of days they
the effect of time pressure on accuracy stay in the hospital after their injury)
po

in completing a task. from a local hospital.

c. Older adults are recruited from a 1.5. For each study described in 1.4, identify
retirement center to examine sources the most likely scale of measurement used
of anxiety in retirees. Anxiety is in the study.
y,

measured using survey items in which

op

Survey Data
tc

We have already discussed some issues with sampling from populations to collect sur-
vey data in this chapter. Using surveys to measure behavior can also limit the types of
no

measures that a researcher can use to observe those behaviors. Although any of the scales
of measurement described in this chapter can be used in surveys, the measures are lim-
ited to an individual’s report of their thoughts and behaviors. In other words, surveys
indirectly measure the behavior of interest in a study. They rely on responses to items
o

that together provide information about the behavior. The survey score is an operational
D

definition of that behavior. However, the way survey responses are presented can greatly
affect the results. Consider the surveys on global warming described at the beginning of
this chapter. Do you notice any differences in the types of responses used in the polls? In
fact, the difference in survey response options is likely the reason for the differences in
the percentage of people who agreed with the survey statement. In the first survey, people
stated that global warming was a serious problem. In the second survey, the percentage
reported was for people who very strongly agreed with this statement. It may be that a fair
Copyright ©2020 by SAGE Publications, Inc.
This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 14   The Process of Research and Statistical Analysis in Psychology

percentage of people in the second survey

chose a response of strongly agree instead
of very strongly agree. In the first survey,
there was no very strongly agree option, so
some of the people who chose the serious
problem response may in fact have felt

e
very strongly about their agreement.
Because the response choices are differ-

ut
ent in the two surveys, it can appear as if
©iStock/mediaphotos

they obtained different results.

b
In addition, there are some issues
related to the self-report nature of survey

tri
data that can influence the validity of
the data. Because surveys rely on reports

is
Photo 1.6
Survey responses
of the behavior from the participants
themselves, the validity of the measurement may not be as high as when the researcher

rd
can be prone to
social desirability directly observes those behaviors. Participants may not have an accurate perception of
bias in which
participants try to
their own behaviors, making the reports subjective. Further, participants may wish to
portray their behaviors more positively than they actually are—an issue known as social

,o
respond in a way
that makes them desirability bias. If participants respond to survey items in a way that makes them appear
look more positive.
more positive, they are reducing the validity of the behavioral measure (see Photo 1.6).
Because survey data have some issues that can affect their validity, researchers are care-
st
social desirability
bias: Bias ful in checking the construct validity of surveys and questionnaires when they are first
created in survey
used to make sure that they accurately measure the behaviors of interest. Reliability of
po

responses from
respondents’ surveys and questionnaires is also examined before they are used as measures of a behavior
desire to be viewed to ensure that they will produce consistent results when they are used in research studies.
more favorably by
others, typically
y,

resulting in
overreporting Systematic and Controlled Measures
of positive
Another type of data collected in many research studies is more systematic and controlled.
op

behaviors and
underreporting of In experimental studies, where internal validity is increased through control of the mea-
negative behaviors surement of the behaviors and the situations in which they are observed, researchers often
employ more systematic and controlled measures (see Photo 1.7). These measures are
tc

construct validity:
The degree to more direct observations of behavior than the self-reports collected on surveys (e.g., accu-
which a survey racy or speed in performing a task), which can provide more internally valid measures of
is an accurate
no

measure of behavior, but they can also have lower external validity than the behaviors measured in
interest surveys because the control imposed during the observations can influence the behaviors
reliability: The
observed. When someone knows they are participating in a research study, they may try
degree to which harder in completing a task (or may perform worse due to lack of motivation because
o

the results of they know they will get research credit regardless of their performance), changing their
a study can be
D

replicated under
accuracy and speed in performing the task compared with a more naturalistic setting.
similar conditions Consider the experiment conducted by Metcalfe, Casal-Roscum, Radin, and
Friedman (2015) to compare older and younger adults’ memories for facts. Both young
and older adults were asked to answer a series of general knowledge questions (e.g., “In
what ancient city were the Hanging Gardens located?” Correct answer: Babylon). For
each answer, they provided a rating of their confidence in their response on a 1 to 7

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  15

scale. Feedback was then given for their

answer (correct or incorrect with the
correct answer given as feedback). After
a short delay, they were then tested
on 20 of the questions for which they
made high-confidence errors (i.e., they

e
answered incorrectly but were highly
confident in their incorrect response) and

ut
©iStock/LuckyBusiness
20 questions where they made low-con-
fidence errors (i.e., they answered

b
incorrectly but were not very confident
in their incorrect response) to determine

tri
final test accuracy on these questions.
The mean accuracy for each group and

is
Photo 1.7
type of question is shown in Figure 1.3. Systematic
Older adults showed better memory on the final test—an atypical finding—especially

rd
measures
when they had low confidence in their original response. The researchers measured of behavior
sometimes
memory accuracy as a controlled measure: They carefully chose the items presented to involve collecting
the participants and tested their memory in a lab, where they could control other factors

,o
responses on
that contribute to memory other than the age group and the participants’ confidence in a computer to
allow for direct
their responses. This control increased the internal validity of the study in providing a measures of
good test of the comparison of younger and older adults. However, this control of the
st
behavior in
measure and the situation may have reduced the realism of the memory being tested. For terms of speed
or accuracy in
example, because the participants knew that the researchers were interested in their
po

completing a task.
memory performance, they may have tried harder on the task (especially the older par-
ticipants, for whom research participation may be a less-common experience) and raised
their memory levels compared with how they would perform in a less-controlled setting.
This shows how controlled observations of behavior can have high internal validity but
y,

may also have lower external validity.

op

FREQUENCY DISTRIBUTIONS
tc

Let’s now consider what the data we collect in a study might look like. When we collect
no

a set of data, we have a distribution of scores to consider. This distribution might range distribution: A set
over the entire scale of measurement (e.g., participants have used all of the values on a of scores

1 to 7 rating scale in their responses), or it might be restricted to just a small range of

scores (e.g., participants have used only the values between 3 and 6 on the 1 to 7 rating
o

scale). In addition, the scores might cluster close to one value on the scale with very few
D

values at the high and low ends of the scale. Or the scores could be equally spaced along frequency
distribution: A
the values of the scale. Thus, different distributions can have different characteristics graph or table
depending on the variability seen in the scores. A good way to examine the distribu- of a distribution
tion and see what it looks like is to create a frequency distribution table or graph. The showing the
frequency of
frequency distribution will indicate how often each value in the scale was used by the each score in the
participants in their responses or in measurements of their behavior. distribution

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
16   The Process of Research and Statistical Analysis in Psychology

FIGURE 1.3 ■ Memory Accuracy Data From Metcalfe et al.’s (2015) Study

.9

Probability Correct on the

.8

e
Final Test
.7

ut
.6

b
tri
.5

is
.4
High-Confidence Low-Confidence

rd
Errors Errors

Young Adults Older Adults

Source: Metcalfe et al. (2015).

,o
st
To create a frequency distribution graph by hand, you place the scores on the x-axis
po

of the graph and then indicate the number of times each of those scores appears in a set
of data with the bar height along the y-axis. Figure 1.4 shows a frequency distribution
graph for responses on a 1 to 7 scale that might be present from a survey question asking
how likely someone is to watch a new show on television after watching an ad for that
y,

show. The graph shows that the respondents used all the scores on the scale, but most
of the scores were clustered around the values of 5 and 6 (these were the most frequent
op

scores in the distribution). Table 1.3 shows a frequency distribution table of the same
set of scores. To create the table, the scores are listed in one column, and the frequency
count of each of the scores in the distribution is listed in the second column. For example,
tc

16 respondents rated their likelihood of watching the new show at a 6 on the 1 to 7 scale.
The other columns in the table show the percentage of all the scores in the distribution at
that value and the cumulative percentage from lowest to highest that adds in all previous
no

percentages as you move from one score to the next.

Shape of a Distribution
o

One thing we can see more easily using a frequency distribution graph is the shape of
D

the distribution. The shape of the distribution can affect the choices a researcher makes
inferential
statistics: A set in analyzing the data with inferential statistics. Look at the frequency distribution in
of statistical Figure 1.4. What do you notice about its shape? Is the distribution symmetrical, with the
procedures used half of the distribution above the most frequent score (a score of 6) the mirror image of
by researchers to
test hypotheses the other? Or are the scores clustered more toward one end of the distribution than the
about populations other? If the distribution shows a mirror image across the most frequent score, then it is

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  17

FIGURE 1.4 ■ A
 Frequency Distribution Graph for Ratings on a 1 to 7 Scale of
How Likely One Is to Watch a New Television Show

20

e
ut
15

b
Count

tri
10

is
rd
5

0
1.00 2.00 3.00 ,o4.00 5.00 6.00 7.00
st
Rating
po

TABLE 1.3 ■ A
 Frequency Distribution Table for Ratings on a 1 to 7 Scale of
How Likely One Is to Watch a New Television Show
y,

Cumulative
op

Score Frequency Percentage percentage

1 2 3.6   3.6
tc

2 5 9.1 12.7

3 4 7.3 20.0
no

4 8 14.5 34.5

5 12 21.8 56.4
o

6 16 29.1 85.5
symmetrical
D

distribution: A
7 8 14.5 100.0 distribution of
scores where
the shape of the
a symmetrical distribution. Symmetrical distributions occur naturally in some types distribution shows
a mirror image on
of data. For example, standardized test scores typically show symmetrical distributions either side of the
with an average score in the center and each half of the distribution around the average middle score

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
18   The Process of Research and Statistical Analysis in Psychology

showing fewer scores at each end of the scale in a similar pattern (i.e., a typical bell shape).
Figure 1.5 shows a fairly symmetrical distribution of data in a frequency distribution
graph. These data represent the distribution of letter grades in a college course where the
instructor has “curved” the grade distribution around an average grade of C (i.e., made it
symmetrical around the C grade).
Many distributions, however, show clustering of scores toward the top or bottom end

e
of the scale. In fact, without curving, grade distributions in many college courses show a
clustering of grades toward the high end of the grade scale—often because there are more

ut
people in a course who do well than who do very poorly. When scores are clustered at one
skewed end of the scale or the other in a distribution, it is known as a skewed distribution. Skew

b
distribution: A in a distribution can affect the comparison of different measures of what is considered
distribution of
a typical score (as you will see in Chapter 6). Speed in completing a task often shows a

tri
scores where
the shape of the skewed distribution, especially if the research participants complete multiple trials of the
distribution shows task. For example, a task speed distribution is shown in Figure 1.6. You can see in this

is
a clustering of
scores at the low
distribution that most of the scores cluster toward the low end of the scale because the
participants are trying to complete the task quickly, based on the instructions. However,

rd
or high end of the
scale there are a few scores higher up on the scale where a participant was especially slow in
completing the task. This could be due to a short lapse in attention or a particularly dif-
ficult trial that affected their speed. However, there will not be the same pattern of very

,o
st
FIGURE 1.5 ■ A
 Symmetrical Distribution of Scores—Letter Grades in a
Course That Has a Curved Scale
po

40
y,
op

30
tc Count

20
no

10
o
D

0
A B C D F
Grade

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  19

FIGURE 1.6 ■ A Skewed Distribution of Scores—Speed of Task Completion

10

e
8

b ut
6
Count

tri
4

is
rd
2

0 ,o
st
400.00 600.00 800.00 1000.00 1200.00 1400.00
Speed of Task Completion (in milliseconds)
po

fast trials at the low end of the scale because the lowest the scores can go is 0 on the scale,
y,

keeping the fast scores from spreading out at the low end. This is why most distributions
of data that measure task speed are skewed. This distribution represents a positive skew,
op

where the tail of infrequent scores is at the high end of the scale. A negative skew shows the
opposite pattern—the tail of the distribution is on the low end of the scale.
tc
no

STOP AND THINK

1.6. Explain how internal validity is increased 77, 75, 78, 56, 90, 68, 65, 63, 73, 77,
o

in a study using systematic and controlled 74, 78, 72, 79, 82, 85, 88, 52, 96, 71
measures of behavior.
D

Does this distribution appear to be symmet-

1.7. Create a frequency distribution table or
rical or skewed in shape? Explain your answer.
graph for the following set of data using
ranges of scores by 10 (i.e., frequency for
51–60, 61–70):

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
20   The Process of Research and Statistical Analysis in Psychology

Frequency Distributions in Excel

In this text, we will discuss two ways to use computer programs to produce statistics. Although
it is important for you to understand what is contributing to the statistics being produced
by the programs, most researchers use computer programs to calculate statistics for their
data. Therefore, in this text you will see both the formulas and calculations for the statis-
tics and the instructions for producing those statistics using the computer programs. The

e
first program we will consider is Microsoft Excel. Excel can be used to create frequency

ut
distribution tables and graphs. Let’s work through an example of how it works using the
distribution of data shown in Figure 1.4 and Table 1.3. These data are hypothetical ratings
for likelihood of watching a new television show after viewing an ad for the show.

b
tri
Tables
We will begin by entering the rating scores. You can follow along with the example if

is
you have Excel to work with as you read. The first step is to type the data into the data
window. This is the first column shown in Figure 1.7. The data are entered with one score

rd
per row. They are presented in order from lowest scores to highest scores, but the order of
the data entry in this column is not important; you can enter them in any order. Here are
the data to enter: 1, 1, 2, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, 5,

,o
5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7. Type these ratings into
the first column in your Excel data window. In the next column, enter in the response
st
FIGURE 1.7 ■ Excel Data Window Showing a Frequency Distribution Table
po
y,
op
tc
no
o
D

Note: Excel is a registered trademark of Microsoft Corporation.

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  21

choices—the values from 1 to 7. In some cases, you may not have exact scores to enter
here, but because these data are from a survey with a rating scale, we can enter the exact
scores that are in the distribution.
The third column will include our counts for each score. You could count these by
hand from the list of data that was just given, but for large data sets, this would be time-
consuming, and you might make a mistake. Instead, we can use the formulas in Excel

e
to calculate the counts. We will use the COUNTIF command here. To use a command
in Excel, type = and then the command; so in the COUNTS column’s first cell, you

ut
can type =COUNTIF. The COUNTIF command will include the range of scores you range: The
want to count (all the scores in the DATA column) and the specific score you want to difference between

b
the highest and
count. Your COUNTIF command will be =COUNTIF(range,score). To calculate the lowest scores in a
counts, type a (, then highlight all of the scores in the DATA column to enter the range

tri
distribution
(by dragging the cursor over the whole column), and then type ,”=“& (comma, quotation
mark, equal sign, quotation mark, ampersand). Then highlight the score in the SCORES

is
column you want to count (or you can just type in the score you want). This will then add
the score you are looking at. If you want a range of scores here, you can use COUNTIFS

rd
(instead of COUNTIF) and include sets of < or > before the & and score to indicate a
range. Close the ) and then hit Enter. If you included labels at the top of the columns as
in Figure 1.7, your first cell should look like this before you hit Enter (you can see the

,o
formula by clicking on a cell to show it in the bar at the top):

=COUNTIF(A2:A56,”=“&1) or =COUNTIF(A2:A56,”=“&B3)
st
You should see a count of 2 in the COUNTS column for a score of 1 (see Figure 1.7).
po

Repeat this formula for each score to calculate counts in your COUNTS column.
The next column in our frequency distribution table is the percentage of the total that
each count represents in the data set. To calculate this value, type =, highlight the value
y,

in the COUNTS column, then /COUNT(. Then highlight the scores in the DATA col-
umn and type ). Then type *100 and hit Enter. Your first cell should look like this (again,
if you have a column header):
op

=C2/COUNT(A2:A56)*100
tc

You should see the percentage of scores that had a value of 1 in the data set in the PER-
CENTAGE column (see Figure 1.7).
no

The last column in our table is the cumulative percentage of values for each score.
This will tell us what percentage of the scores is at a certain value or lower. To calculate
this value, again begin by typing = to indicate a formula, then highlight the next value
in the PERCENTAGE column, type +, and then highlight the previous value on the
o

CUMULATIVE PERCENTAGE column (if there is one) before hitting Enter. For the
D

first score, the cumulative percentage will simply be the percentage value. Your completed
table should look similar to the one shown in Figure 1.7.

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
22   The Process of Research and Statistical Analysis in Psychology

SUMMARY OF STEPS
• Type the data into a data window in Excel • Calculate percentages for each score using
(1st column). =score cell/COUNT(range)*100 (4th column).

e
• Type in response choices (2nd column). • Calculate cumulative percentages by
successively adding up the percentages for
• Type in =COUNTIF(range, scores) command

ut
each score (5th column).
(3rd column).

b
• Repeat COUNTIF command for each score or
range of scores to complete third column.

tri
is
Graphs

rd
Excel can also create a frequency distribution graph of these data for us. To create the
graph, highlight the SCORES and COUNTS columns of values. Then click the Charts
function window. Excel’s default is typically a bar graph for these types of data, but your

,o
settings may be different, so you may need to choose the Column Graph option. To show
the distribution shape, choose the option under Chart Layouts that shows the bars adja-
cent in the graph. You should then have a graph that looks like the one in Figure 1.8. You
st
can type in axis labels and change the fill color of the bars to format as you like.
po

FIGURE 1.8 ■ Excel Chart Showing a Frequency Distribution Graph

18
y,

16
op

14

12
tc

10
Count
no

6
o

4
D

0
1 2 3 4 5 6 7
Rating

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  23

Frequency Distributions in SPSS FIGURE 1.9 ■ Data Window for SPSS

The second computer program I will describe in this
text is IBM® SPSS® Statistics.1 SPSS is a common
program used by researchers for descriptive and infer-
ential statistics in psychology. You may be asked to
use this program in some of your psychology courses

e
to analyze data from research studies. In many cases,

ut
it will produce the statistics you want more easily than
Excel. However, this program is not as commonly
available as Excel, which is why there is instruction

b
for both programs provided in this text.

tri
Note: SPSS is a registered trademark of International

is
Business Machines Corporation.

rd
Tables
As with Excel, the first step in using SPSS is to enter the
data you wish to examine. The data window in SPSS

,o
has a similar setup to the one in Excel. However, you
will define the variable names and details in a separate
tab of the window. An example of the data window
st
in SPSS is shown in Figure 1.9. As in Excel, each row
contains data from a different participant, and each
po

column indicates a different variable. Thus, the first

column contains the ratings for desire to watch the
television show for each participant from our earlier
example as in the Excel window. However, to label this
y,

variable, you need to choose the Variable View tab at

the bottom of the window. This view will allow you to
op

name the variable (as I have done here with the label
Rating). See Figure 1.10 for the Variable View in SPSS.
Once your data are entered and labeled, you are
tc

ready to create your frequency distribution table. To

create a table, you will choose the Frequencies func-
no

tion in the Descriptive Statistics menu under the

Analyze menu at the top. Different versions of SPSS
look a bit different for these menus, so first find the
Analyze menu (or tab) at the top of the window; then
o

choose Descriptive Statistics and then Frequencies.

You should see a small window pop up that looks like
D

Figure 1.11. To create the table for the ratings, make

sure the Rating variable is highlighted (if not, click on it), and then click the arrow
to move it into the Variable(s) box. Be sure to keep the Display frequency tables box

1
SPSS is a registered trademark of International Business Machines Corporation.

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
24   The Process of Research and Statistical Analysis in Psychology

FIGURE 1.10 ■ Variable View Window for SPSS

e
ut
FIGURE 1.11 ■ Frequencies Window in SPSS

b
tri
is
rd
,o
st
po
y,
op

FIGURE 1.12 ■ F
 requency Distribution Table in SPSS, Seen in the Output
Window
tc
no
o
D

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  25

checked. The OK button should then be available. When you click OK or hit Enter, a
new Output window will appear. Your table will be displayed as shown in Figure 1.12.
The Valid column shows the scores in the distribution, the Frequency column shows
the counts for each score (including a total count), the Percent and Valid Percent col-
umns show the percentage of scores in the distribution that are at that score, and the
Cumulative Percent column shows the cumulative percentage for each score. You can

e
compare this table with those shown in Table 1.3 and Figure 1.7 to see the similarities
across the different program versions of the same table.

b ut
SUMMARY OF STEPS

tri
is
• Type the data into a data window. highlighting the variable(s) and using the
arrow in the center of the window.
• Label the variable in Variable View tab.

rd
• Make sure the Display Frequency Tables box
• Choose Descriptive Statistics in the Analyze
is checked.
menu at the top.
• Click OK; your table will be shown in the

,o
• Choose Frequencies from the Descriptive
Output window.
Statistics choices.
• In the Frequencies window, choose
st
the variable(s) you are interested in by
po

Graphs
A frequency distribution graph can also be created using the Frequencies function. If you
y,

click on the Charts option in the Frequencies window (see Figure 1.11), a new window
will open with chart options. Choose the Histograms option and click Continue. Then
op

when you click OK, both your Table and your Graph will appear in the Output window.
The graph will look like Figure 1.13. Notice that some additional descriptive statistics
are also provided alongside the graph. We will discuss these statistics in the upcoming
tc

chapters.
no

SUMMARY OF FREQUENCY
DISTRIBUTIONS
o

Frequency distribution tables and graphs are useful in helping summarize a distribution
D

of scores. They allow you to see the shape of the distribution and clustering of scores
in a particular part of the distribution. They can be created by hand, but Excel and
SPSS can create them more easily for you, reducing the chance of error if your data have
been entered into the program correctly. As we continue to discuss additional descriptive
statistics in the next few chapters, you will see that these programs are quite useful in
calculating your statistics.

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
26   The Process of Research and Statistical Analysis in Psychology

STOP AND THINK

Create a frequency distribution graph in both 1.8. Does this distribution appear to be
Excel and SPSS for the following final exam symmetrical or skewed? If the shape is
scores according to letter grade groupings with skewed, describe the skew (i.e., positive

e
90%–100% = A, 80%–89% = B, 70%–79% = C, or negative skew).
60%–69% = D, below 60% = F:

ut
1.9. About how many students received an
A on the final? How many received a D
83, 92, 100, 90, 74, 58, 84, 78, 85, 78, 72, 60,

b
or F?
67, 92, 92, 88, 88, 66, 60, 80, 88, 58, 92, 84,
84, 59, 80, 68, 78, 86, 76, 80, 64, 84, 68, 58,

tri
72, 88, 89, 72, 88, 65, 80, 84, 68, 73, 92

is
rd
FIGURE 1.13 ■ F
,o
 requency Distribution Graph in SPSS, Seen in the Output
Window
st
po
y,
op
tc
no
o
D

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  27

THINKING ABOUT RESEARCH

A summary of a research study in psychology is

given here. As you read the summary, think about

e
the following questions:

ut
1. What behaviors have the researchers

©istockphoto.com/goldy
observed?

b
2. How were the observations recorded by

tri
the researchers?
3. Were the researchers able to answer

is
their research questions with the
observations they collected? How?

rd
Photo 1.8
4. Consider the graphs in Figure 1.14. Strayer and Johnston’s (2001) study examined whether
Explain how these results address the talking on a cell phone while driving decreases driving
researchers’ question about behavior. performance.

5. What are some examples of real-world

behaviors that the results of this study
might apply to? ,o experiment. Each of the students was randomly
assigned to one of three cell phone conditions:
st
handheld phone, hands-free phone, and no phone
Research Study. Strayer, D. L., & Johnston, W. A. (radio control only). The participants performed
po

(2001). Driven to distraction: Dual-task studies a computer-simulated driving task where they
of simulated driving and conversing on a cellular moved the cursor on the screen to match a mov-
phone. Psychological Science, 12, 462–466. [Note: ing target as closely as possible, using a joystick.
Only Experiment 1 of this study is described.] Red and green lights flashed periodically during
y,

the task, and subjects were instructed to press

Purpose of the Study. The researchers were inter- the “brake” button as quickly as possible when the
red light flashed. They performed this task on its
op

ested in how use of a cell phone while driving

influences driving performance (see Photo 1.8). own in a practice segment and two test segments;
They describe previous studies that have shown a dual-task segment was placed between the two
that devices that require one’s hands while driving test segments. In the dual-task segment, they
tc

(e.g., the radio, temperature controls) can reduce were given an additional task that included one of
driving performance. In this study, they predicted the following to match the conditions listed previ-
that cell phone use would reduce driving per- ously: handheld phone conversation with another
no

formance. They tested two ideas about how cell person (who was part of the research team) about
phone use could decrease driving: (1) that the a current news story, hands-free phone conver-
handheld use of the phone would interfere with sation with another person about a current news
driving and (2) that the attention requirements of story, or controlling a radio to listen to a broad-
o

a phone conversation would interfere with driving. cast of their choice. The frequency of missing red
D

lights and the reaction time to hit the brake button

Method of the Study. Forty-eight undergradu- when a red light appeared were measured and
ates (half male, half female) participated in the compared for the three phone conditions.

(Continued)

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
28   The Process of Research and Statistical Analysis in Psychology

(Continued)

Results of the Study. The two cell phone use condi- button was longer when subjects were talking on
tions did not differ in their results, suggesting that the phone (compared with when only driving), but
driving performance in response to red lights is there was no difference in driving performance when

e
similar for handheld and hands-free phone use. subjects listened to the radio while driving and when
Figure 1.14 shows a graph for each of the measures they just performed the driving task on its own.

ut
according to the phone (combined for handheld and
hands-free conditions) and no-phone conditions. The Conclusions of the Study. The authors concluded

b
data are shown in each graph separately for driving that phone use, regardless of whether it requires
performance in the driving only segments (single one’s hands, interferes with driving performance

tri
task) and for the phone or radio task while driving more than just listening to the radio. This suggests
(dual-task) segment. The graphs show that more that the attention component of phone use is the

is
red lights were missed and time to press the brake key factor in the driving performance interference.

rd
FIGURE 1.14 ■ D
 riving Performance as Measured by Responses to Red Lights in the
Driving Task While Performing the Driving Task on Its Own (Single Task)
or While Also Performing the Phone or Radio Task (Dual Task)

0.10

0.09
A
,o
st
0.08
Probability of a Miss

0.07
po

0.06

0.05

0.04
y,

0.03

0.02
op

0.01

0.00

625 B
tc

600
Mean Reaction Time (msec)

575
no

550

525
o

500
D

475

450
Cell Phone Radio Control

Single Task Dual Task

Source: Strayer and Johnston (2001, Figure 1).

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  29

Chapter Summary
• What is the value of research in psychology? systematic and controlled responses on
Research provides psychologists with new a task are two common examples of the
knowledge about behavior, regardless of the types of data researchers collect. Four

e
type of behavior of interest. measurement scales define the type of
measurements used in the data collected:

ut
• Why do psychologists conduct research?
nominal, ordinal, interval, and ratio scales.
Psychologists conduct research because
it provides the best way to gain new • What is a distribution, and how does its shape

b
knowledge about behavior. affect our analysis of the data?

tri
A distribution is a set of scores collected
• What is the difference between a population
as data. The shape of a distribution can
and a sample?

is
take many forms, but two common shapes
A population is the group of individuals a
are symmetrical and skewed shapes.

rd
researcher wants to learn about. The sample is
These shapes will affect our choice of both
the portion of the population that participates
descriptive and inferential statistics, as we
in the research study. This is the group of
will see in the coming chapters.

,o
individuals observed by the researcher.

• What kinds of data are collected in

psychological studies?
st
Many different kinds of data are collected
in research studies. Survey responses and
po
y,

Applying Your Knowledge

op

On Facebook one day, you see a post from one of the new vitamin your friend is so excited
your friends that they have found the most amaz- about?
tc

ing vitamin supplement. They claim that they

• Suppose you came across a news item
have taken the vitamin once a day for the past
reporting that thousands of people have
no

few weeks, and they have more energy and feel

been trying the new vitamin (they include
great. They are passing on the information to their
interviews with some of these people) and
friends (including you) and urging you to try the
that overall, these people have reported
vitamin for yourself.
o

positive results. Would this convince you to

try the new vitamin? Why or why not?
D

• Why should you be skeptical of the claim

you read from your friend?

• What other information would you want

to have before deciding if you should try

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
30   The Process of Research and Statistical Analysis in Psychology

Test Yourself
1. Freud hypothesized that many of our of rats was fed the standard rat chow. The
personality traits are controlled by an rats were then tested in the maze after
unconscious conflict between aspects of being fed the assigned diet for 2 weeks.

e
ourselves—the id, ego, and superego— The amount of time it took the rats to
that we are not consciously aware of reach the food reward in the maze was

ut
(Nairne, 2009). Using what you know measured. Rats on the high-sugar diet
about conducting research, explain why took longer to run the maze on average

b
this hypothesis is difficult to support with than the normal diet rats.

tri
observations of behavior. c. A study was conducted to examine the
2. The scientific method relies on which way of effects of violence on social behaviors

is
knowing information about the world? in young children. Five-year-olds were
asked to play a superhero video game

rd
3. If I am concerned about whether the survey with mild violence (e.g., punching,
I am using in my study accurately measures throwing). Two researchers who were
the behavior I am interested in, I am not aware of the purpose of the study

,o
considering the ___________________ of my observed the children’s behavior at
study. recess. The number of social behaviors
4. For each study description that follows, seen (e.g., helping another child, playing
st
identify the most likely population of cooperatively with another child) was
interest, identify the operational definition recorded on a school day both before and
po

of the behavior of interest, and identify the after they played the video game.
scale of measurement of the dependent 5. Providing responses on a survey to make
variable. yourself look better is called __________.
y,

a. College student participants were asked a. symmetrical bias

to play a virtual ball-tossing game
op

b. skewed bias
during which some participants are
systematically excluded from the game c. social desirability bias
a short time after they began. The study d. ratio bias
tc

tested the effects of social exclusion on

6. In a research study on navigation,
the participants’ mood. The researchers
participants were asked to judge the
no

then asked the students to complete a

distance of a landmark in the environment
mood survey in which they rated their
from their current location. This
mood on a 1 to 7 scale—higher numbers
dependent variable was measured on a(n)
indicating a more positive mood.
_______________ measurement scale.
o

b. To examine the effect of diet on cognitive

a. nominal
D

abilities, researchers taught rats to

navigate a maze to reach a food reward. b. ordinal
Half of the rats in the study were fed a c. interval
special diet high in sugar; the other group d. ratio

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.
 Chapter 1 ■ Psychological Research  31

7. In a research study, you are asked to c. Nominal, ordinal

indicate your college major on a survey. This d. Ordinal, nominal
dependent variable was measured on a(n)
_______________ measurement scale. 10. Survey data are always accurate.

a. nominal a. True

e
b. ordinal b. False

ut
c. interval 11. A frequency distribution graph can show you
d. ratio the shape of a distribution.

b
a. True

tri
8. You are conducting a study that uses IQ
tests. On these tests, the participants b. False
score an average of 100. All other scores

is
12. Systematic and controlled measures are
are evenly distributed above and below this more direct observations of behavior than

rd
average. What type of distribution is this? the self-reports collected on surveys.
a. Skewed distribution a. True
b. Symmetrical distribution b. False
c. Hypothetical distribution
d. Faulty distribution ,o 13. Of the following choices, which are good
operational definitions of anxiety?
st
9. _____________ scales typically involve a. Scores on an anxiety scale
numerical scores, whereas _______________
po

b. Score on an exam
scales do not.
c. A general feeling of helplessness
a. Interval, ratio
d. Both a and c
b. Ratio, nominal
y,
op
tc
no

Visit edge.sagepub.com/mcbridermstats to find the answers to the Test Yourself questions above, as
well as quizzes, flashcards, and other resources to help you accomplish your coursework goals.
o
D

Copyright ©2020 by SAGE Publications, Inc.

This work may not be reproduced or distributed in any form or by any means without express written permission of the publisher.