Deck 8: Experimental Designs

Question 1

A study lacks __________ if it fails to eliminate the possibility that an extraneous variable rather than the independent variable produced the effect on the dependent variable.

A) reliability
B) internal validity
C) external validity
D) generalizability

Answer 1

B

Question 2

If "history" or some other factor threatens the internal validity of a research design, this means that the possible effects of an extraneous variable are confounded with the

A) effects of the independent variable.
B) effects of random assignment.
C) measurement of the dependent variable.
D) constancy of experimental conditions.

Answer 2

A

Question 3

An experiment is conducted to assess the effects of two types of reading improvement lessons: speed reading versus comprehension training. The subjects, high school seniors, are divided into two groups. Members of one group receive comprehension training from their high school English teacher, while at the same time the other group receives speed reading training from a local college instructor. What variables are confounded in this study?

A) speed reading and comprehension training
B) type of improvement lesson and the particular instructor
C) type of improvement lesson and year in school
D) year in school and the particular instructor

Answer 3

B

Question 4

Suppose you have developed a treatment to help people stop smoking. To test the treatment, you identify smokers in a campus survey, and then you contact them to ask if they wish to volunteer for your smoking cessation treatment. Those who volunteer are given the treatment, and those who do not are assigned to the control group. After a month, a higher percentage of the treatment subjects than the controls has stopped smoking. Aside from the treatment, what may account for this difference?

A) testing
B) testing-X interaction
C) instrumentation
D) selection
E) maturation

Question 5

Suppose you want to study the effects of study abroad on students' attitudes toward immigration. You contact the Study Abroad Office and identify all students selected for study broad. You measure the attitudes of members of this group before they go abroad and then measure their attitudes again after they return home. If you found more tolerance after they return home, this could be due to the study abroad experience or to other factors. Which of the following threats to validity is not an alternative interpretation?

A) selection
B) maturation
C) history
D) testing

Question 6

Suppose you want to test the effects of handing out free samples of pizza on the sales of frozen pizza in supermarkets. For one week, you provide free samples in a "test" supermarket located in one area and do not provide free samples in a "control" supermarket located in another area. During this period a pizzeria located in the test area goes out of business. If you found that pizza sales increased in the test store but not in the control store, what is a likely alternative explanation to the effect of providing free hot pizza?

A) history effect
B) maturation effect
C) statistical regression
D) pretesting effect
E) testing-treatment interaction effect

Question 7

A study of the effects of the children's television program Sesame Street showed that preschoolers with the lowest initial academic skills (e.g., ability to recognize letters of the alphabet) gained more over a season of viewing than other children. The effects of the program in this study are most likely to be confounded with the effects of

A) differential attrition.
B) selection.
C) testing.
D) statistical regression.

Question 8

__________ is likely to be a threat to internal validity whenever subjects are assigned to experimental conditions on the basis of their low or high scores on the pretest.

A) Testing
B) Differential attrition
C) Instrumentation
D) Statistical regression

Question 9

Which threat to internal validity is controlled by randomly assigning research participants to experimental conditions?

A) selection
B) history
C) maturation
D) testing
E) attrition

Question 10

Random assignment of subjects to experimental conditions eliminates the possibility of _________ and __________ as threats to internal validity.

A) maturation; history
B) testing; selection
C) history; statistical regression
D) selection; statistical regression
E) testing; instrumentation

Question 11

__________ is due to random measurement error among extreme scorers, whereas __________ may be due to systematic measurement error among human observers.

A) Testing; statistical regression
B) Experimenter bias; testing
C) Statistical regression; instrumentation
D) Differential attrition; experimenter bias
E) Instrumentation; differential attrition

Question 12

__________ occurs in an experiment when one group of subjects is more sensitive to the independent variable (or treatment) than another group.

A) History-treatment interaction
B) Selection-treatment interaction
C) History-maturation interaction
D) Selection-history interaction

Question 13

__________ may occur when a pretest sensitizes subjects to the independent variable so that they react differently to the treatment than they would have had they not been pretested.

A) A testing effect
B) Testing-X interaction
C) An instrumentation effect
D) Selection bias
E) Selection-X interaction

Question 14

Which threat to internal validity is controlled in the one-group pretest-posttest design (O₁ X O₂)?

A) attrition
B) testing
C) history
D) maturation

Question 15

Which of the following would not be a potential threat to internal validity in the pretest-posttest design (O₁ X O₂)?

A) testing
B) maturation
C) history
D) selection

Question 16

Which of the following would not be a potential threat to the internal validity of a static-group comparison design?

A) testing
B) history
C) selection
D) differential attrition

Question 17

Which of the following is an example of a true experimental design?

A) one group pretest-posttest design
B) static group comparison
C) separate-sample pretest-posttest design
D) posttest-only control group design

Question 18

Suppose the effects of a treatment are tested in two different experiments. Experiment A consists of a pretest-posttest, control group design; experiment B consists of a posttest-only, control group design. The treatment is found to have an effect only in experiment
A)
A) One possible explanation of the discrepancy in findings is the effects of __________ in experiment

A) selection
B) testing
C) testing-X interaction d. statistical regression
E) attrition

Question 19

Testing is likely to threaten internal validity in a

A) posttest-only control group design.
B) static group comparison.
C) posttest only design (or one-shot case study).
D) one group pretest-posttest design.

Question 20

The posttest-only control group design generally is preferred over the pretest-posttest control group design because the latter is susceptible to the effects of

A) differential attrition.
B) history.
C) statistical regression.
D) selection-X interaction.
E) testing-X interaction.

Question 21

In a within-subjects design,

A) subjects are randomly assigned to two or more conditions.
B) the same subjects participate in both treatment and control conditions.
C) testing is controlled by eliminating the pretest.
D) only random processes could account for the difference between O₁ and O₂.

Question 22

Which of the following is not a potential threat to the validity of a within-subjects design?

A) testing
B) testing-X interaction
C) order of the treatment and control conditions
D) selection

Question 23

Counterbalancing in within-subjects designs controls for the effects of

A) the order of the treatment and control conditions.
B) lack of random assignment.
C) using different groups of subjects in each condition.
D) history and maturation.

Question 24

If a study has all the features of a true experiment, any difference between conditions (e.g., a treatment group and a control group) may be due to the manipulation of the independent variable and/or to

A) sample selection bias.
B) history.
C) random or chance variation.
D) statistical regression.

Question 25

A study is designed to test the effectiveness of a program to teach adolescents about nutrition. All research participants are randomly assigned to two groups, one which goes through the program and the other which does not. Both before and after the program, all participants take a test measuring knowledge of nutrition. What type of experimental design is this?

A) posttest-only control group design
B) pretest-posttest control group design
C) 2 × 2 factorial design
D) 2 × 4 factorial design

Question 26

Which of the following design features is present in a factorial design?

A) the dependent variable is measured both before and after the manipulation
B) a pretest measure is used in each experimental condition
C) there is more than one measure of the dependent variable
D) more than one independent variable is manipulated

Question 27

In a laboratory experiment on helping behavior, two variables are manipulated: (1) others' presence (whether participants worked alone on a task or with others) and (2) type of task (whether participants were timed or untimed on the task). The dependent measure is the number of seconds between someone's request for help and the participant's response. What type of experimental design is this?

A) posttest-only control group design
B) pretest-posttest control group design
C) 2 × 2 factorial design
D) 2 × 4 factorial design

Question 28

The two factors that are varied in a Solomon four-group design are

A) pretest and treatment.
B) pretest and posttest.
C) selection and treatment.
D) selection and posttest.

Question 29

Which of the following sets of findings is an example of an interaction effect?

A) Older men drink more than older women; younger men drink more than younger women.
B) Men drink more than women, and younger people drink more than older people.
C) Empathy enhances helping, and women are more empathic than men.
D) A negative mood such as guilt or sadness increases the likelihood of helping among adults but not among children.
E) The more bystanders present at an emergency, the less likely than anyone will intervene.

Question 30

In a study of interpersonal attraction, subjects watched a videotape featuring a person who was (1) either competent or incompetent, and (2) either did or did not spill coffee on himself (i.e., "blundered"). After watching the tape, all subjects indicated how much they liked the person in the film on a seven-point scale. Consider the following outcome:
This outcome shows

A) a main effect for competence only.
B) a main effect for blunder only.
C) main effects for both competence and blunder.
D) an interaction between competence and blunder.

Question 31

In a study of interpersonal attraction, subjects watched a videotape featuring a person who was (1) either competent or incompetent, and (2) either did or did not spill coffee on himself (i.e., "blundered"). After watching the tape, all subjects indicated how much they liked the person in the film on a seven-point scale. Consider the following outcome:
This outcome shows

A) a main effect for competence only.
B) a main effect for blunder only.
C) main effects for both competence and blunder.
D) an interaction between competence and blunder.

Question 32

Which of the following experimental designs can provide information on interaction effects?

A) pretest-posttest control group design
B) posttest-only control group design
C) static group comparison
D) factorial design

Question 33

A field experiment on labor market discrimination (Box 8.1) showed that white applicants with a high-quality resume were more likely to be called back for an interview than whites with a low-quality resume but that quality of resume made no difference in whether African-American applicants were called back for interviews. This outcome indicates

A) a main effect for race only.
B) a main effect for quality of resume only.
C) main effects for both race and quality of resume.
D) an interaction between race and quality of resume.

Question 34

Which of the following statements is false about quasi-experimental designs?

A) They use a variety of approaches to establish internal validity.
B) They are more effective than pre-experimental but less effective than true experimental designs in controlling for extraneous variables.
C) Their major deficiency is the absence of a pretest.
D) They lack one or more features of a true experiment.

Question 35

The separate-sample pretest-posttest design

A) eliminates the possibility of a history effect.
B) eliminates the threat of testing and testing-X interaction.
C) randomly assigns subjects to experimental and control conditions.
D) contains a control group but lacks random assignment.

Question 36

Which feature of a true experiment is missing from the nonequivalent control group design?

A) random assignment
B) control group
C) pretest
D) posttest

Question 37

Which threat to the validity of an interrupted time-series design is controlled in the multiple time-series design?

A) selection
B) history
C) instrumentation
D) testing

Question 38

Experimenters using quasi-experimental designs rule out threats to internal validity by doing all but which one of the following?

A) using probability sampling
B) including special design features
C) examining additional data that bear on specific validity threats
D) reasoning against the plausibility of particular validity threats

Question 39

The Clore et al. study of the effect of interracial contact at a summer camp on interracial attitudes

A) randomly assigned black and white children to segregated and integrated settings.
B) was limited by its reliance on a single, subjective measure of the dependent variable.
C) showed that the camp experience changed the attitudes of girls but not boys.
D) was unable to rule out maturation or statistical regression as validity threats.

Question 40

In Campbell and Ross's analysis of the effect of the Connecticut crackdown on speeding, they were able to eliminate the threat of history by using a __________ design, which showed that the same decline in traffic fatalities did not occur in four neighboring states.

A) separate-sample pretest-posttest
B) nonequivalent control group
C) interrupted time-series
D) multiple time-series

Question 41

An experiment's internal validity depends upon how well it controls for extraneous variables.

Question 42

The basic principle of good experimental design is allowing only one factor to vary at a time.

Question 43

"Pre-experimental designs" refer to studies involving direct observation rather than manipulation of the independent variable.

Question 44

Pre-experimental designs are subject to various threats to internal validity.

Question 45

"Selection" refers to systematic differences among subjects who are assigned to different experimental conditions.

Question 46

Selection effects may occur whenever college students are used as subjects.

Question 47

Instrumentation is a threat to internal validity whenever subjects are pretested.

Question 48

Attrition effects pertain only to studies using the aged or disabled as experimental subjects.

Question 49

Random assignment controls for the effects of selection and statistical regression.

Question 50

History is basically a problem of external validity rather than internal validity.

Question 51

Maturation effects may occur when experimental conditions are run at different times of the day.

Question 52

Statistical regression may be a threat to validity when subjects are selected on the basis of their high or low scores on the dependent variable.

Question 53

True experimental designs control for maturation and history.

Question 54

A pretest is an essential feature of true experimental designs.

Question 55

A posttest-only control group design eliminates the possible effects of testing and testing-X interaction.

Question 56

A weakness of the pretest-posttest control group design is that it lacks random assignment.

Question 57

In a within-subjects design, each subject participates in all experimental conditions.

Question 58

Counterbalancing controls for selection effects in between-subject designs.

Question 59

The Solomon four-group design may be construed as a 2 × 2 factorial design.

Question 60

A 2 × 3 factorial design has five experimental conditions.

Question 61

"Interaction effect" refers to the social relations among subjects that take place in most experiments.

Question 62

Quasi-experimental designs generally are inferior to pre-experimental designs.

Question 63

The major deficiency of most quasi-experimental designs is the absence of a pretest.

Question 64

The absence of random assignment is common in quasi-experimental designs.

Question 65

A major threat to the internal validity of the interrupted time-series design is history.

Question 66

Use of multiple time-series can rule out the threat of history.

Question 67

Consider the following experiment on helping disabled persons. One Monday, in a campus area with park benches where students often meet, an investigator placed one of two confederates: either a handicapped woman in a wheelchair or a nondisabled woman. The disabled woman was there between the hours of 9:00 a.m. and 12:00 noon, and the nondisabled woman was there between the hours of 1:00 and 4:00 p.m. While in the setting, every fifteen minutes each woman asked a randomly chosen passerby to help her find a lost earring. An observer, appearing to be reading a book while seated on a nearby bench, recorded the amount of time each subject spent helping to find the earring. If the subject had not left in fifteen minutes, the confederate pretended to find the earring.
a. What is the independent variable in this study?
b. What variable is confounded with the manipulation of the independent variable? (Identify a specific variable, not a class of extraneous variables such as history, maturation, etc.)

Question 68

2. A large cooperation develops a long-term project to test a series of leadership training courses it is designing. The training courses are to be offered after work for low-level employees. They will be run on a volunteer basis, and the workers will not be paid for staying late to attend. To evaluate the effectiveness of the course after ten years, the corporation decides that the measure of effectiveness will be how many of the individuals who take the course assume positions of leadership in comparison with those who did not take the course.
a. What is the measure of the dependent variable in this study?
b. In words or appropriate notation, identify the design of the study.
c. Assume that after ten years it is found that those who took the leadership course are more likely to be in positions of leadership within the corporation than those who did not. Identify two threats to internal validity and carefully explain how each could account for this effect.

Question 69

3. A pharmaceutical company wants to market what seems to be a revolutionary new drug: a near-cure for the common cold. Researchers in the R & D division randomly select 1,000 persons from the city where the company is headquartered to participate in a test study. Each participant is observed for six months. For the first three months, baseline data are collected. For the last three months, the participants take a weekly dose of the common-cold cure. Sure enough, 15 percent of the participants contracted a cold during the first three months, whereas only 2 percent did so in the second three months.
a. Using the notation introduced in Chapter 8, carefully diagram the design of this study. Be sure to label or specify the meaning of each "letter" in the diagram.
b. Is this a true experiment? If so, describe the design features that make it a true experiment. If not, describe what feature(s) of a true experiment are missing.
c. What is the purpose of randomly selecting participants? Does this enhance the internal validity or external validity of the study?
d. The investigators are confident that they have accurately recorded the incidence of colds and that the observed difference (15 percent versus 2 percent) is statistically significant. Therefore, they rush their findings to you, the company president, who must decide whether the data are convincing enough for the company to put the product on the market. Knowing about threats to internal validity, you consider the following: selection, statistical regression, history, instrumentation, and attrition. Which, if any, of these threats might undermine the validity of this study? Briefly explain.

Question 70

The interaction of the independent variable (or treatment) with some other variable (e.g., a variable represented by history or maturation) poses a threat to the external validity of true experiments. What are some solutions to problems of (a) sample selection-X interaction, (b) maturation-X interaction, and (c) history-X interaction? (Recall that X is an independent variable.)

Question 71

What are the three ways that rival explanations are ruled out in quasi-experimental designs? Describe one application of each of these techniques in (a) the Clore and co-workers camp study and (b) the Campbell and Ross study of the Connecticut speeding crackdown.