Research methods common mistakes

Common mistakes

• Confusing the sign test with the Wilcoxon signed‑rank test

  Assuming that the sign test uses ranks of the differences.

  Fix itThe sign test uses only the sign of each difference; ranks are not involved. The Wilcoxon signed‑rank test, by contrast, incorporates the magnitude of differences through ranking.

• Misinterpreting p‑values

  Thinking that a p‑value below 0.05 automatically means a large or important effect.

  Fix itA p‑value indicates statistical significance, not effect size. Effect size measures (e.g., Cohen’s d) are needed to judge practical importance.

• Confusing Type I and Type II errors

  Students often think a Type I error is a false negative and a Type II error is a false positive.

  Fix itClarify that Type I is a false positive (rejecting a true null) and Type II is a false negative (failing to reject a false null). Use the symbols α for Type I and β for Type II to reinforce the distinction.

• Using parametric tests for nominal data

  Applying a t‑test or ANOVA to a nominal variable.

  Fix itNominal data should be analysed with chi‑square tests; parametric tests require interval or ratio data.

• Misidentifying the correct test for factorial designs

  Choosing a t‑test or chi‑square test when a factorial ANOVA is required.

  Fix itAlways check the number of independent variables and their levels. A factorial design with two or more factors needs an ANOVA (one‑way, two‑way, or mixed) rather than a simple t‑test.

• Confusing Pearson's r with Spearman's rho

  Using Pearson's r when data are ordinal or non‑normal.

  Fix itUse Spearman's rho for ordinal or non‑normal data; Pearson's r requires interval/ratio data and normality.