Research Bias Evaluation - Eppp Skills

1. What is the term for outcome changes due to participants knowing they are being observed?

   Hawthorne effect. Observation awareness prompts behavioral changes as participants react to being studied, not necessarily to the intervention itself.

2. What limitation is introduced when outcomes are measured only immediately posttreatment?

   No follow-up; limited durability inference. Absence of long-term assessment prevents conclusions about the persistence or sustainability of treatment effects over time.

3. What is the name for systematic differences in how outcomes are assessed across groups?

   Detection (assessment) bias. Variations in assessment methods or rigor between groups can lead to biased outcome detection and inaccurate comparisons.

4. What is the bias called when researchers choose analyses after seeing the data?

   Researcher degrees of freedom (p-hacking). Post-hoc analytical choices exploit flexibility to achieve significance, inflating false positives and undermining reproducibility.

5. What threat occurs when changes are due to a new rater, instrument, or scoring rule?

   Instrumentation threat. Alterations in measurement tools, raters, or criteria over time can produce apparent changes unrelated to the intervention.

6. What is the term for distortion caused by nonrandom loss of participants over time?

   Attrition bias (differential dropout). Uneven participant dropout across groups can skew results by leaving non-comparable samples at study end, affecting validity.

7. What internal validity threat occurs when participants change due to time, not treatment?

   Maturation threat. Natural developmental or temporal changes in participants can be misattributed to the treatment in longitudinal studies.

8. What bias occurs when the measurement instrument does not capture the intended construct?

   Construct (measurement) validity threat. Inadequate alignment between the measure and the theoretical construct results in invalid inferences about the phenomenon under study.

9. What is the term for improvement due to natural fluctuation when selecting extreme scores?

   Regression to the mean. Extreme scores naturally tend toward the average upon retesting due to statistical variability, mimicking treatment effects.

10. What bias is present when study groups differ at baseline due to nonrandom assignment?

    Selection bias. Nonrandom assignment creates baseline imbalances in group characteristics, confounding the ability to attribute outcomes to the intervention.

11. What is the term for inaccurate recall of past events that systematically biases reports?

    Recall bias. Memory distortions, influenced by current knowledge or biases, lead to systematic errors in retrospective self-reports.

12. What is the main limitation when a study uses a convenience sample from one clinic only?

    Limited external validity (generalizability). Convenience sampling from a single site restricts the applicability of findings to broader or diverse populations.

13. Identify the bias: participants in the intervention arm receive extra attention time.

    Performance bias (attention confound). Additional attention in the intervention confounds results, as outcomes may stem from nonspecific interaction rather than treatment.

14. What is the term for a consistent instrument error that shifts scores in one direction?

    Systematic measurement error (bias). Consistent directional errors in measurement tools produce biased data, deviating systematically from true values.

15. Identify the bias: dropouts are mostly severe cases in the treatment group only.

    Differential attrition bias. Uneven dropout of severe cases biases group comparisons, as remaining samples differ systematically in composition.

16. What internal validity threat occurs when outside events coincide with the intervention?

    History threat. Concurrent external events can influence outcomes, making it difficult to isolate the intervention's true impact.

17. Identify the limitation: the sample excludes non-English speakers in a diverse community.

    Sampling bias; reduced generalizability. Exclusion of subgroups like non-English speakers creates a non-representative sample, limiting applicability to diverse populations.

18. What bias occurs when participants alter responses to appear favorable or avoid stigma?

    Social desirability bias. Participants may provide responses they believe are socially acceptable, distorting data on sensitive or normative topics.

19. Identify the limitation: a pretest–posttest study has no control group and no randomization.

    Weak internal validity; cannot infer causality. Without controls or randomization, alternative explanations for changes cannot be ruled out, preventing causal claims.

20. Which bias occurs when participants who volunteer differ systematically from nonvolunteers?

    Self-selection (volunteer) bias. Volunteers often have distinct characteristics, such as higher motivation or specific traits, leading to non-representative samples in research studies.

21. What bias is present when only statistically significant outcomes are published or reported?

    Publication bias (selective reporting). Selective publication of positive results distorts the evidence base, overestimating effects and ignoring null findings.

22. Identify the bias: a clinician rates treated clients as improved because they expect benefit.

    Observer-expectancy bias. Researchers' expectations subtly influence observations or ratings, leading to biased confirmation of anticipated outcomes.

23. What bias occurs when knowledge of group assignment changes participant behavior?

    Performance bias. Awareness of assignment influences how participants act, potentially altering outcomes independent of the actual treatment effect.

24. What threat occurs when repeated testing itself changes subsequent test performance?

    Testing (practice) effect. Familiarity gained from initial testing improves later performance, confounding interpretations of true change.

25. Identify the limitation: outcomes are self-report only for a stigmatized behavior.

    Social desirability response bias risk. Self-reports on stigmatized topics are prone to underreporting or distortion to align with social norms, compromising data accuracy.

Study these flashcards