Which action best supports reliable normative interpretation when data quality is in question?

Which action best supports reliable normative interpretation when data quality is in question?

• A. Ignore the questionable data and proceed
• B. Re-test to verify results, check testing procedures, and consider excluding extreme outliers if justified
• C. Use only the best data
• D. Accept the data as is

Answer: (B)

When data quality is in question, the goal is to ensure the normative interpretation reflects true patterns rather than artifacts of measurement or collection. The best approach is to re-test to see if results replicate, review testing procedures to confirm they were applied consistently, and consider excluding extreme outliers only if there is justified, predefined criteria. Re-testing helps show whether results are stable and not just random fluctuation. Checking procedures reduces the chance that errors in how the data were collected are driving the findings. Excluding outliers, when justified, prevents a few extreme values from disproportionately skewing the norms, leading to a more accurate portrait of the underlying distribution. Other options lead to biased or unreliable conclusions: ignoring questionable data or accepting data as is can distort the norms, and relying only on the best data introduces selection bias and ignores potential issues in the full data set.

When data quality is in question, the goal is to ensure the normative interpretation reflects true patterns rather than artifacts of measurement or collection. The best approach is to re-test to see if results replicate, review testing procedures to confirm they were applied consistently, and consider excluding extreme outliers only if there is justified, predefined criteria. Re-testing helps show whether results are stable and not just random fluctuation. Checking procedures reduces the chance that errors in how the data were collected are driving the findings. Excluding outliers, when justified, prevents a few extreme values from disproportionately skewing the norms, leading to a more accurate portrait of the underlying distribution.

Other options lead to biased or unreliable conclusions: ignoring questionable data or accepting data as is can distort the norms, and relying only on the best data introduces selection bias and ignores potential issues in the full data set.