Threats to External Validity
Sample Characteristics
The effect of an intervention is demonstrated with a particular sample in a given experiment. To what extent can the results be generalized to other persons who vary in age, race, ethnic background, education, or any other characteristic? It is possible that the subjects in a given study are in some way differentially responsive (or unresponsive) to the treatment. Hence the findings may be limited to samples with specific features and cannot be extended to a larger population of interest.
Stimulus Characteristics
Equally important to sample characteristics is the extent to which the results extend across the stimulus characteristics of the experiment. The stimulus characteristics refer to features of the study with which the intervention or condition may be associated. Characteristics of the setting, experimenters, interviewers, or other features of the stimuli to which subjects are exposed may restrict generality. The findings of the study may be restricted to a particular setting, situation, or feature of the experimental arrangement.
Contextual Characteristics
The context in which the intervention is presented may place restrictions on the generality of the results. The context refers to any of the conditions in which the intervention may be embedded. There is, of course, always some context in which the intervention is presented. Yet several special arrangements may plausible contribute to the relationship between the intervention and behavior and hence restrict generality of the findings. These special arrangements include reactivity of experimental arrangements, multiple-treatment interference, and novelty effects.
Assessment Characteristics
Several different facets of assessment within a study may affect generality of the results. These facets can refer to any condition of assessment that differs from those to which the investigator may wish to generalize and that may plausibly influence the results. Within clinical research, assessment characteristics that may restrict generality readily come to mind, including reactivity of assessment, test sensitization, and time of measurement and treatment effects.
