1
CHAPTER 6
INDEPENDENT GROUPS DESIGNS
CHAPTER OUTLINE AND OBJECTIVES
I. Overview
II. Why Psychologists Conduct Experiments
Researchers conduct experiments to test hypotheses about the causes of behavior.
Experiments allow researchers to decide whether a treatment or program effectively changes
behavior.
III. Logic of Experimental Research
Researchers manipulate an independent variable in an experiment to observe the effect on behavior,
as assessed by the dependent variable.
Experimental control allows researchers to make the causal inference that the independent variable
caused the observed changes in the dependent variable.
Control is the essential ingredient of experiments; experimental control is gained through
manipulation, holding conditions constant, and balancing.
An experiment has internal validity when it fulfills the three conditions required for causal inference:
covariation, time-order relationship, and elimination of plausible alternative causes.
When confounding occurs, a plausible alternative explanation for the observed covariation exists, and
therefore, the experiment lacks internal validity. Plausible alternative explanations are ruled out by
holding conditions constant and balancing.
IV. Random Groups Design
In an independent groups design, each group of subjects participates in only one condition of the
independent variable.
Random assignment to conditions is used to form comparable groups by balancing or averaging
subject characteristics (individual differences) across the conditions of the independent variable
manipulation.
When random assignment is used to form independent groups for the levels of the independent
variable, the experiment is called a random groups design.
A. An Example of a Random Groups Design
B. Block Randomization
Block randomization balances subject characteristics and potential confoundings that occur
2
during the time in which the experiment is conducted, and it creates groups of equal size.
C. Threats to Internal Validity
Randomly assigning intact groups to different conditions of the independent variable creates a
potential confounding due to preexisting differences among participants in the intact groups.
Block randomization increases internal validity by balancing extraneous variables across
conditions of the independent variable.
Selective subject loss, but not mechanical subject loss, threatens the internal validity of an
experiment.
Placebo control groups are used to control for the problem of demand characteristics, and
double-blind experiments control both demand characteristics and experimenter effects.
V. Analysis and Interpretation of Experimental Findings
A. The Role of Data Analysis in Experiments
Data analysis and statistics play a critical role in researchers’ ability to make the claim that an
independent variable has had an effect on behavior.
The best way to determine whether the findings of an experiment are reliable is to do a replication
of the experiment.
B. Describing the Results
The two most common descriptive statistics that are used to summarize the results of
experiments are the mean and standard deviation.
Measures of effect size indicate the strength of the relationship between the independent and
dependent variables, and they are not affected by sample size.
One commonly used measure of effect size, d, examines the difference between two group
means relative to the average variability in the experiment.
Meta-analysis uses measures of effect size to summarize the results of many experiments
investigating the same independent variable or dependent variable.
C. Confirming What the Results Reveal
Researchers use inferential statistics to determine whether an independent variable has a reliable
effect on a dependent variable.
Two methods to make inferences based on sample data are null hypothesis testing and
confidence intervals.
Researchers use null hypothesis testing to determine whether mean differences among groups in
an experiment are greater than the differences that are expected simply because of error
3
variation.
A statistically significant outcome is one that has a small likelihood of occurring if the null
hypothesis were true.
Researchers determine whether an independent variable has had an effect on behavior by
examining whether the confidence intervals for different samples in an experiment overlap. The
degree of overlap provides information as to whether the sample means estimate the same
population mean or different population means.
D. What Data Analysis Can’t Tell Us
VI. Establishing the External Validity of Experimental Findings
The findings of an experiment have external validity when they can be applied to other individuals,
settings, and conditions beyond the scope of the specific environment.
In some investigations (e.g., theory-testing), researchers may choose to emphasize internal validity
over external validity; other researchers may choose to increase external validity using sampling or
replication.
Conducting field experiments is one way that researchers can increase the external validity of their
research in real-world settings.
Partial replication is a useful method for establishing the external validity of research findings.
Researchers often seek to generalize results about conceptual relationships among variables rather
than specific conditions, manipulations, settings, and samples.
VII. Matched Groups Design
A matched group design may be used to create comparable groups when there are too few subjects
available for random assignment to work effectively.
Matching subjects on the dependent variable (as a pretest) is the best approach for creating matched
groups, but scores on any matching variable must correlate with the dependent variable.
After subjects are matched on the matching variable, they should then be randomly assigned to the
conditions of the independent variable.
VIII. Natural Groups Design
Individual differences variables (or subject variables) are selected rather than manipulated to form
natural groups designs.
The natural groups design represents a type of correlational research in which researchers look for
covariations between natural groups variables and dependent variables.
Causal inferences cannot be made regarding the effects of natural groups variables because
4
plausible alternative explanations for group differences exist.
IX. Summary
REVIEW QUESTIONS AND ANSWERS
These questions appear in the textbook (without answers) at the end of Chapter 6, and can be used for a
homework assignment or exam preparation. Answers to these questions appear in italic.
1. Describe two reasons why psychologists conduct experiments.
2. Describe how the control techniques of manipulation, holding conditions constant, and balancing
contribute to meeting the three conditions necessary for a causal inference.
3. Explain why comparable groups are such an essential feature of the random groups design, and
describe how researchers achieve comparable groups.
4. Identify what a “block” refers to in block randomization and explain what this procedure accomplishes.
A block is one random order of all conditions in an experiment. In block randomization, subjects are
5. What preventive steps could you take if you anticipated that selective subject loss could pose a
problem in your experiment?
7
8
The mean body dissatisfaction score for the Thin Image group was -1.25 (SD = 1.07), the mean
score for the Overweight Image group was -0.75 (SD = 1.16), and the mean score for the Neutral
Image group was -0.20 (SD = .83).
A. Identify the design of this study, including the independent and dependent variables. Explain why
the researcher asked participants to judge the same neutral words in each condition, and provide
examples of factors the researcher controlled by using random assignment to conditions.
B. How would you describe the effect of the independent variable using the means for each
condition? What do the standard deviations tell you about the body dissatisfaction ratings in the
experiment?
C. The p value for the F–test for the effect of the type of image is p = .009. What claim would you
make about the effect of the independent variable based on this probability?
D. The .95 confidence interval (CI) for the thin condition is -1.71 to -0.79; the CI for the overweight
condition is -1.21 to -0.29; and the CI for the neutral condition is -0.66 to 0.26. What claim would
you make based on the estimates of the population means for the three groups in the experiment
based on a comparison of these confidence intervals?
9
E. The effect size for the difference between the means for the thin and neutral groups is d = .85.
What information does this effect size tell you about subliminal exposure to thin images compared
to neutral images, beyond what you know from the test of statistical significance and from
comparing confidence intervals?
3. Premature infants often have difficulty regulating their body temperature because of their low birth
weight and low body fat. Researchers hypothesized that immersion bathing in warm water would
improve temperature regulation relative to the common practice of sponge bathing. Eight infants were
available to test the hypothesis; four were randomly assigned to each of the two conditions
(immersion, sponge) during their stay in the neonatal intensive care unit (NICU). The dependent
variable was body temperature 15 min after the bath. Contrary to the hypothesis, no difference in
body temperature after the bath was found between the two groups.
A. A nurse noted that the two groups of infants were very different. The infants randomly assigned to
the immersion condition weighed less and were born earlier, on average, than infants in the
sponge condition. Explain how this difference between the two groups could have occurred
despite the fact that random assignment to conditions was used, and how this difference may
have impacted the results of the experiment.
B. The researchers decided to test the hypothesis again. This time, however, they considered the
weight of the infants. Twelve infants were available for the study with the following weights (in
pounds): 4.6, 5.0, 4.1, 4.0, 5.4, 5.8, 4.4, 4.9, 6.2, 5.2, 4.2, 4.6. Identify the research design you
would recommend to conduct this study and explain how the infants should be assigned to the
conditions of the experiment.
10
C. Suppose a social worker tested the bathing variable when she taught parenting skills during
home visits. She reasoned that maintaining body temperature at home is more important than in
the NICU, where incubators are used. She randomly assigned 20 parents of premature infants to
learn to immerse their babies at the proper temperature, and 20 parents were instructed about
sponge bathing. The social worker also taught parents how to take the baby=s temperature 15
min after the bath (the dependent variable). Identify a goal of the social worker=s research that
led her to focus on the setting for conducting her research. How would you describe the findings if
both experiments showed that body temperature was higher for infants in the immersion condition
relative to the sponge-bath condition?
4. A study was done to compare two diets to help diabetic male patients lose weight. Each diet reduced
patients’ intake to 1800 calories per day. In the control condition, patients chose their own meals from
a list of allowed foods, but their calorie intake was not allowed to exceed the limit. In the treatment
condition, designed to minimize fluctuations in insulin levels, patients= meals came from the same list
of foods but were planned so as to achieve specific percentages of complex carbohydrates and
proteins. Forty men were randomly assigned to each condition. All participants monitored their insulin
levels. The experiment lasted 6 months; during that time 15 men in the control condition dropped out
of the study (e.g., due to difficulties managing insulin, complications associated with diabetes). Three
men in the treatment condition dropped out of the study; two moved out–of-state, and the work
schedule for a third prevented him from attending appointments for the study. At the end of 6 months,
the average weight loss for the 25 men who completed the control condition was 11 pounds, and 10
pounds for the 37 men in the treatment condition. The researcher concluded that the treatment,
specifying the percentages of calories from different sources (e.g., complex carbohydrates, proteins),
is not helpful to weight loss among diabetic men.
A. Identify a possible threat to the internal validity of this experiment, and explain how this problem
could account for the results of the study.
11
12
15
E. A social psychology experiment was done to test the hypothesis that a person would be more
likely to comply with a request if that person had previously complied with another request than if
he or she was responding to a first request (a “foot–in-the-door” phenomenon). The experiment
included two groups of 50 undergraduates; each student had been randomly assigned to their
respective condition. In the first group, the Single Request group, students were told to complete
a problem-solving task, which they thought was the actual purpose of the experiment. After
participants had completed the task the experimenter came in and said the experiment was over.
The experimenter then asked the participant if he or she was willing to help carry some boxes
down to another room. The dependent variable was whether the participant complied with this
request or not. In the second group, the Double Request group, the participants were first asked
whether they were willing to complete the experiment by doing a problem-solving task. For ethical
reasons, those participants who did not want to do the experiment were allowed to leave and
were given credit for doing the experiment anyway. Those who agreed to do the experiment were
then given the problem-solving task, followed by the request to do the experimenter a favor by
carrying the boxes. Forty of the 50 participants in the Double Request group completed the
problem-solving task and, of course, all 50 participants in the Single Request group did the task.
Sixty percent of the participants in the Double Request group agreed to help carry the boxes, but
only 40 percent of the participants in the Single Request group agreed to help. This difference
was statistically significant. The researchers concluded that people are more likely to comply with
a second request if they complied with a previous request.
(1) Is there a problem of interpretation in this experiment because there are different numbers of
participants in each group?
(2) Would the group of 40 undergraduates in the Double Request group be comparable to the
group of 50 undergraduates in the Single Request group? Why or why not?
to the dependent variable measure (agree to experimenter’s request).
F. An experiment was done to test an explanation for the phenomenon of social loafing, specifically,
that social loafing occurs when people in a group are not accountable for their individual
response. The experiment used of a random groups design. All the students in the experiment
16
were asked to shout as loudly as they could and all the students wore headsets over which they
heard shouting. Twenty students were randomly assigned to each of three groups, differentiated
in terms of the conditions under which they did the shouting. The first group of students was
alone. This Alone group served as the control condition in which no social loafing could occur.
The second group of students performed the shouting task in a group with other students, but
students couldn’t hear the shouting of the other students in the group because they heard
shouting over the headsets. Social loafing was expected to occur in this group because the
students were not accountable for their shouting; this group was called the Unaccountable group.
The third group of students also shouted in a group of other students, but each student shouted
into a microphone that they believed could be used by the experimenter to monitor each person’s
shouting. Social loafing was not expected in this group because each student in this group was
accountable for their shouting; this group was called the Accountable group. The researchers
measured the loudness of each student’s shouting, with higher numbers indicating louder
shouting. The mean loudness scores and the standard deviations in the three conditions were:
Alone Group: M = 65, SD = 5.4; Unaccountable Group: M = 40, SD = 3.7; and Accountable Group
M = 62; SD = 4.9. Use this information to answer the questions that follow.
(1) What is the independent variable in this experiment and what are its levels?
(2) What is the dependent variable in the experiment?
(3) What information would you use to describe the results of this experiment?
(4) What characteristic of this experiment (or any experiment) is assessed most directly when
confidence intervals are determined or when a test of statistical significance is done?
(5) Compute confidence intervals for the means in each of the three conditions. A simple
computational technique to approximate the confidence interval is to use a value of 2 times
each SD to calculate the confidence interval around each sample mean. Determine whether
the three conditions differed based on a comparison of the three confidence intervals.
2. Reading Research Critically
The following two research summaries and the accompanying questions could be used in class for
small group discussion. The answers appear below. To facilitate photocopying, the research
summaries and questions appear on subsequent pages.
A. Mindfulness: Langer and Piper (1987) manipulate subtle linguistic cues to determine whether
participants in their study will respond creatively to a novel situation.
Answers to Reading Research Critically: Mindfulness
(1) Langer and Piper suggested that giving information conditionally caused participants to think
more flexibly about an object. Comment on whether this causal inference seems warranted
given their procedure and results.
(2) Langer and Piper’s findings were statistically significant. Comment critically on each of the
following statements and decide whether the conclusions are appropriate based on the fact
that the findings were statistically significant.
(a) The findings in Langer and Piper’s study have only a small probability of occurring if the
null hypothesis is true.
(b) Langer and Piper’s found that 65% of the students in the conditional group solved the
problem by identifying the object as an eraser while only 25% of students in the
unconditional group did so. Because this finding is statistically significant, do we know
that the finding is meaningful and important?
18
(c) If the probability of Langer and Piper’s obtaining their findings assuming the null
hypothesis is true was .02, then we can say that the probability that the null hypothesis is
true for their findings is .02.
B. Interpersonal Touch: Crusco and Wetzel (1984) conducted a field experiment that will be of
particular interest to students who wait tables, and to anyone who is interested in factors that
affect the size of a tip at restaurants.
Answers to Reading Research Critically: Interpersonal Touch
(1) Despite the observation that diners were apparently unaware of being touched in the Crusco
and Wetzel (1984) study, why is it possible to conclude that the touch caused the diners to
leave a larger tip?
(2) Crusco and Wetzel (1984) were interested in examining the factors that influence the amount
of money left for a tip. Explain why it was important to use the percentage of the total bill left
for a tip as the dependent variable rather than the amount of money left for the tip. What
potential variable is controlled for by using the percentage tip as the dependent variable?
20
Reading Research Critically
Read the following description of a research study to answer the questions that follow. [Based on an
article by: Langer, E. J., & Piper, A. I. (1987). The prevention of mindlessness. Journal of Personality and
Social Psychology, 53, 280-287.
Mindfulness
According to the authors, “mindlessness” is characterized by a “rigid use of information,” treating
information as if it had a single meaning and could be useful in only one way; a consequence is that the
individual fails to recognize potential novel uses of information. On the other hand, “mindfulness” is
characterized by “active distinction making and differentiation.” One who is mindful engages in the
process of creating new categories while making finer and finer distinctions. It is possible that
mindlessness arises in part because of the way that information is presented to us. For instance, people
are often educated about the world by naming objects absolutely (“this is an X”). If people were
introduced to information conditionally (“this could be an X”), perhaps they would produce more creative
use of objects and ideas.
Several experiments were conducted to test the hypothesis that subtle linguistic cues lead people to act
mindfully or mindlessly. In one experiment (Exp. 3), students were told they were participating in a study
of stress management and were taken to a room where they were asked to answer questions about
several pictures. An experimenter described various objects in the room, including an unfamiliar black
rubber object which (unknown to the participant) was the target object. The unfamiliar object was
described in one of three ways: Unconditionally (“This is a precision propel.”); Conditionally (“This could
be a precision propel.”); and Conditionally Unknown (“I do not know what this is.”). There were 20 Harvard
University undergraduates assigned randomly to each of the three experimental conditions. When the
participants had answered the questions about the pictures, they were told by the experimenter that the
answers had been recorded in the wrong section of the book and that there were neither additional forms
nor an eraser. The goal of this experimental procedure was to create a “need” for a solution to the
problem of the mis-recorded answers. The target solution was to make an eraser available (i.e.,
identifying the black object as a potential eraser). In keeping with the theory of mindlessness/mindfulness,
it was predicted that the different ways in which the unfamiliar black object had been introduced would
affect the likelihood of students meeting the need, that is, solving the problem. Specifically, the authors
predicted that more students would respond creatively in the two conditional groups than in the
unconditional group. And they did. Sixty-five percent of the students in the Conditional group and 55
percent of the students in the Conditional Unknown group solved the problem by suggesting that the
unfamiliar black object could be used as an eraser; only 25 percent of the students in the Unconditional
group came up with this creative solution to the problem. The authors suggested that the Conditional
participants had learned to think flexibly about the object.
Langer and Piper (1987) argue that too often we teach children to treat the world unconditionally, and that
cognitive flexibility, and hence, creative solutions to problems, might result were we to educate children to
think more conditionally. According to this view, giving information in absolute terms encourages the
mindless use of it.
1. Langer and Piper suggested that giving information conditionally caused participants to think more
flexibly about an object. Comment on whether this causal inference seems warranted given their
procedure and results.