Deck 11: Causal Inference and Experimental Designs

Question 1

Which of the following is LEAST suited to providing clear evidence about a causal relationship between the independent and dependent variables?​

A) ​Classical experimental design.
B) Solomon four-group design.​
C) Posttest-only control group design.​
D) One-shot case study.​

Answer 1

D

Question 2

If we can establish that variable X comes before variable Q in time, then we can say:​

A) ​variable X is a cause of variable Q.
B) variable X is not a cause of variable Q.​
C) variable Q is a cause of variable X.​
D) variable Q is not a cause of variable X.​

Answer 2

D

Question 3

The effectiveness of randomization in experimentation is affected by the number of participants involved.​

Answer 3

True

Question 4

The one-group pretest-posttest design controls for passage of time.​

Question 5

An instrumentation effect occurs when:​

A) ​the instrument employed for the pretest has an effect that shows up on the posttest.
B) ​the measurement instrument is changed from the pretest to the posttest.
C) ​there is a change on the dependent variable from the pretest score to the posttest score.
D) ​there is statistical regression to the mean.

Question 6

The problem of external validity refers to the generalizability of results.​

Question 7

A school social worker administered a self-esteem test to a group of 9th graders in September. During the school year the students received intensive social work intervention designed to improve their self-esteem. In May the self-esteem test was given again and the self-esteem scores improved. Which of the following statements is true about this study?​

A) ​It s results show that the intervention was effective in improving self esteem.
B) It only controls for testing effects and statistical regression.​
C) It only controls for history and maturation.​
D) It does not permit any conclusive causal inferences to be made.​

Question 8

Which of the following statement best describes the criteria for inferring or causality?​

A) ​A relationship must exist between variables.
B) One variable must precede the other in time.​
C) A third variable did not cause the changes observed in the first two variables.​
D) A relationship must exist between two variables, with the cause preceding the effect in time, and the relationship cannot be explained away by the influence of a third variable.​

Question 9

A study evaluates the effectiveness of a child abuse prevention program by comparing the rate of child abuse among parents who chose to participate in the program to the rate of parents who declined participation. This design fails to control for:​

A) ​selection biases
B) correlation.​
C) testing effects.​
D) instrumentation changes​

Question 10

Suppose a new social casework program is offered to students with high levels of truancy. If those students placed in the program by their families have significantly less truancy after participating than the students who did not participate in the program, then we can conclude that:​

A) ​the program effectively reduces truancy.
B) a selection bias might explain away the difference.​
C) causality cannot be inferred because of the lack of covariation.​
D) some clients benefited, so the program has value.​

Question 11

Whenever two variables are correlated, we may assume that one is the cause of the other.​

Question 12

​Assume that you have developed a study technique that you believe will result in students scoring higher on research methods exams. You test your study technique with the design diagrammed below. ​
R = random assignment
0 = observation
X = stimulus
​
R 01 X 02
R 03 04
​
Using the diagrammed design, what predictions will you make?

A) ​02 should be greater than 01.
B) ​02 should be less than 04.
C) ​04 should be greater than 03.
D) ​04 should be less than 01.

Question 13

Suppose a case management program is established to reduce rehospitalization rates in a hospital and that those rates drop dramatically once the new program is established. We can conclude that:​

A) ​research has demonstrated an effective approach to practice.
B) causality cannot be inferred because time order was not established.​
C) another event occurring during the time of the study might explain away the change.​
D) the absence of covariation threatens the internal validity of any causal inferences.​

Question 14

The classical experiment with random assignment of participants controls for measurement bias.​

Question 15

A friend of yours, a senior, took the Graduate Record Exam in September and scored in the 99th percentile. In February your friend took the same exam over again. This time your friend scored in the 84th percentile. As a research methodology student, you told your friend that his/her lowered score was probably due to:​

A) testing.​
B) history.​
C) statistical regression.​
D) demoralization.​
E) ​compensation rivalry.

Question 16

Attrition effects are automatically controlled when participants are assigned randomly to experimental and control groups.​

Question 17

Studies that assign subjects to intervention groups on the basis of their extreme scores are vulnerable to regression toward the mean.​

Question 18

Random assignment to experimental and control groups controls for research reactivity.​

Question 19

The 55-mph speed limit was introduced in the U.S. in 1973. Shortly thereafter the number of auto accidents declined. We can conclude that:​

A) ​the reduced speed limit caused the decline in auto accidents.
B) the reduced speed limit may have caused the decline in the number of auto accidents.​
C) the reduced speed limit and the number of auto accidents are causally related.​
D) the reduced speed limit had nothing to do with the decline in auto accidents.​

Question 20

The problem of an interaction between the testing and the experimental stimulus is handled by​

A) ​the Solomon four-group design.
B) the one-shot case study.​
C) the classical experimental design​
D) the static-group comparison.​
E) ​a and b only are correct.

Question 21

Professor Rose was asked to design an experiment to test whether obese people would lose weight on a new diet plan. Professor Rose is not a methodologist. She asked for your help: You immediately told her that the selected design would have to be sensitive to the issues of statistical regression, diffusion or imitation of treatment, compensatory rivalry, and demoralization. Professor Rose said, "I don't understand." Explain these issues to her.​

Question 22

Pre-experimental designs​

A) ​are insufficient for drawing causal inferences.
B) control for most sources of internal invalidity.​
C) control for most sources of external invalidity.​
D) are excellent for drawing causal inferences.​

Question 23

Explain the elements of the classical experimental design. Is this design especially useful in dealing with causal relationships? If yes, why? If not, why?​

Question 24

Briefly describe how a study would be particularly vulnerable to the threat of statistical regression.​

Question 25

Explain what is meant by internal and external validity.​

Question 26

The Solomon four-group design:​

A) ​combines the classical experimental design with the posttest-only control group design.
B) combines the classical experimental design with the pretest-only control group design.​
C) ​combines the classical experimental design with the static-group comparison design.
D) ​combines the static-group comparison design with the one-group pretest-posttest design.

Question 27

Discuss how attrition can affect the validity of an experiment and describe steps to minimize attrition.​

Question 28

The classical experiment with random assignment of participants does NOT necessarily control for:​

A) ​history.
B) measurement bias.​
C) statistical regression.​
D) maturation.​

Question 29

What differentiates experimental from pre-experimental designs? Give an illustration of a situation in which an experimental design would be preferred and one in which a pre-experimental design would be preferred.​