Threats to External Validity

1. External validity addresses the question, "What are the boundary conditions of the independent variable-dependent variable relation?" Will the independent variable cause the changes observed in the dependent variable only under limited circumstances?

2. Nonrepresentative sample

A. Distinction between nonrepresentative sample and selection threat to internal validity
B. Participants who volunteer
C. Subject roles

(1) Good subject
(2) Negativistic subject
(3) Apprehensive subject
(4) Faithful subject

3. Nonrepresentative research context

A. Laboratory setting
B. Sociocultural context of the study

(Rogers & Prentice-Dunn, 2004, 1981. Source: Aronson et al., 2007)

4. Reactivity in the research setting

A. Demand characteristics

(1) When are demand characteristics problematic?
(2) Special case: Sensitization due to pretest
(3) Postexperimental assessment of suspicion (e.g., Spivey & Prentice-Dunn, 1990)

B. Experimenter expectancy effect

(1) Rehearsal and monitoring
(2) Minimizing the experimenter's role
(3) Condition blindness
(4) Avoidance of data snooping

5. Evaluating threats to external validity

A. Dimensions that are important will vary according to the content area of the research and the stage of theory development.

B. Value of the finding is not always dependent on generalization across several dimensions.

C. External validity accrues over several studies.