Paired / repeated measures¶
Compares two or more conditions measured on the same respondents (within-subject), held in separate columns.
When to use it¶
When each row has several measurements to compare — for example, a pre-test and a post-test column, or three time points — rather than separate groups of people.
Inputs¶
Conditions — two or more numeric columns, one per condition/time point.
With two conditions you get a paired test; with three or more, a repeated-measures test.
Options¶
- Method
Detect automatically, Parametric (paired t / repeated-measures ANOVA), or Non-parametric (Wilcoxon signed-rank / Friedman).
- Assumption checks
Auto / Yes / No — includes a normality check on the differences. (With Detect automatically, assumption checks must be on.)
- Other
Effect size / post-hoc, Verbal indicators (in-table columns), Verbal report (dropdown for how much written interpretation: None / Key findings / Significant only / Full), Number conditions, and Plots.
Output¶
A descriptives table across conditions.
A normality table (when checks are on).
The test result with the appropriate statistic, p, and effect size (e.g. Cohen’s dz, generalized eta-squared, Kendall’s W, or rank-biserial correlation).
Post-hoc comparisons for three or more conditions.
Notes¶
Rows missing any condition are dropped (complete-case analysis), so all conditions are compared on the same respondents.