Threats to validity include mortality, maturation, experimenter bias, regression to the mean, selection, reactive measures, repeated measures, history and instrumentation. Applied Behavior Analysis (2nd Edition)
History: A major holiday season occurs during the intervention phase of a study on employee productivity. The increased distractions and potential shift in employee motivation due to the holiday season could influence the results, making it difficult to determine if the intervention was the true cause of any observed changes.
Instrumentation: A researcher observing a child’s social interactions changes their observation criteria mid-study. For example, they might become more lenient in what they consider “positive social interaction” after observing several instances of challenging behavior. This inconsistency in measurement can skew the results.
Mortality: A study on a new reading intervention program experiences high dropout rates among participants with severe reading difficulties. This attrition can bias the results, as the remaining participants may be more responsive to the intervention than those who dropped out.
Maturation: A study examining the effects of a new teaching method on young children’s language development is conducted over several months. The children’s natural language development during this period could contribute to the observed improvements, making it difficult to isolate the impact of the teaching method.
Experimenter Bias: A researcher conducting a study on a new anxiety reduction technique may unconsciously provide more encouragement and positive feedback to participants in the treatment group, potentially influencing their outcomes.
Regression to the Mean: A study focuses on a small group of students with extremely low academic performance. Due to chance, these students may have performed unusually poorly in the initial assessment. In subsequent assessments, their performance is likely to improve simply due to regression to the mean, even without any intervention.
Selection: A study compares two groups of students, one receiving a new intervention and the other a control condition. If the groups are not equivalent at the outset (e.g., one group has significantly higher initial academic skills), any observed differences between the groups may be due to pre-existing differences rather than the intervention itself.
Reactive Measure: Participants in a study on social skills training become overly aware of being observed, leading them to behave differently than they normally would. This reactivity to the assessment process can invalidate the findings.
Repeated Measures: In a study with repeated assessments, participants may become “test-wise” or more practiced at the tasks being measured. This practice effect can artificially inflate performance scores over time, making it difficult to accurately assess the impact of the intervention.
Comments