quant-two-group-deplisted

# quant-two-group-dep — Two Dependent Measurements Comparison ## Objective Test whether two paired measurements differ (e.g., a pre/post study, or two conditions evaluated by the same respondents). The primary method is the permutation sign-flip test — it randomly flips the signs of individual differences to build the null distribution, making no parametric assumptions about normality of the differences. --- ## Pre-flight **Required inputs:** - `output/data/cleaned_data.csv` - `output/data/column_metadata.json` **Minimum requirements:** - Two numeric/ordinal columns containing the paired measurements - OR: one outcome column + one condition/time variable with exactly 2 levels + respondent ID - Minimum n pairs: 10 (warn), 20 (preferred) - Columns must be of matching type (both ordinal, or both continuous) --- ## Pipeline ### Step 1 — Load data and metadata ```python # See references/pipeline.py → load_prepared_data() ``` --- ### Step 2 — AskUserQuestion: data structure ``` AskUserQuestion: question: "How are your two measurements structured in the data?" header: "Data structure" options: - label: "Wide format — two separate columns (Recommended if pre/post in same row)" description: "E.g., 'satisfaction_before' and 'satisfaction_after' are two columns." - label: "Long format — one outcome column + time/condition column" description: "E.g., 'satisfaction' column with a 'timepoint' column (before/after). Requires a respon