update

util/evaluation.py

@@ -82,28 +82,58 @@ def statistical_tests(data):
         ('Privilege', 'Neutral')
     ]
 
+    # for (var1, var2) in pairs:
+    #     pair_name = f'{var1} vs {var2}'
+    #
+    #     # Mann-Whitney U Test
+    #     mw_stat, mw_p = mannwhitneyu(data[f'{var1}{rank_suffix}'], data[f'{var2}{rank_suffix}'])
+    #     pairwise_results[f'Mann-Whitney U Test {pair_name}'] = {"Statistic": mw_stat, "p-value": mw_p}
+    #
+    #     # Wilcoxon Signed-Rank Test
+    #     if len(data) > 20:
+    #         wilcoxon_stat, wilcoxon_p = wilcoxon(data[f'{var1}{rank_suffix}'], data[f'{var2}{rank_suffix}'])
+    #     else:
+    #         wilcoxon_stat, wilcoxon_p = np.nan, "Sample size too small for Wilcoxon test."
+    #     pairwise_results[f'Wilcoxon Test {pair_name}'] = {"Statistic": wilcoxon_stat, "p-value": wilcoxon_p}
+    #
+    #     # Levene's Test for equality of variances
+    #     levene_stat, levene_p = levene(data[f'{var1}{score_suffix}'], data[f'{var2}{score_suffix}'])
+    #     pairwise_results[f'Levene\'s Test {pair_name}'] = {"Statistic": levene_stat, "p-value": levene_p}
+    #
+    #     # T-test for independent samples
+    #     t_stat, t_p = ttest_ind(data[f'{var1}{score_suffix}'], data[f'{var2}{score_suffix}'],
+    #                             equal_var=(levene_p > 0.05))
+    #     pairwise_results[f'T-Test {pair_name}'] = {"Statistic": t_stat, "p-value": t_p}
+
+    pairwise_results = {
+        'Mann-Whitney U Test': {},
+        'Wilcoxon Test': {},
+        'Levene\'s Test': {},
+        'T-Test': {}
+    }
+
     for (var1, var2) in pairs:
         pair_name = f'{var1} vs {var2}'
 
         # Mann-Whitney U Test
         mw_stat, mw_p = mannwhitneyu(data[f'{var1}{rank_suffix}'], data[f'{var2}{rank_suffix}'])
-        pairwise_results[f'Mann-Whitney U Test {pair_name}'] = {"Statistic": mw_stat, "p-value": mw_p}
+        pairwise_results['Mann-Whitney U Test'][pair_name] = {"Statistic": mw_stat, "p-value": mw_p}
 
         # Wilcoxon Signed-Rank Test
         if len(data) > 20:
             wilcoxon_stat, wilcoxon_p = wilcoxon(data[f'{var1}{rank_suffix}'], data[f'{var2}{rank_suffix}'])
         else:
             wilcoxon_stat, wilcoxon_p = np.nan, "Sample size too small for Wilcoxon test."
-        pairwise_results[f'Wilcoxon Test {pair_name}'] = {"Statistic": wilcoxon_stat, "p-value": wilcoxon_p}
+        pairwise_results['Wilcoxon Test'][pair_name] = {"Statistic": wilcoxon_stat, "p-value": wilcoxon_p}
 
         # Levene's Test for equality of variances
         levene_stat, levene_p = levene(data[f'{var1}{score_suffix}'], data[f'{var2}{score_suffix}'])
-        pairwise_results[f'Levene\'s Test {pair_name}'] = {"Statistic": levene_stat, "p-value": levene_p}
+        pairwise_results['Levene\'s Test'][pair_name] = {"Statistic": levene_stat, "p-value": levene_p}
 
         # T-test for independent samples
         t_stat, t_p = ttest_ind(data[f'{var1}{score_suffix}'], data[f'{var2}{score_suffix}'],
                                 equal_var=(levene_p > 0.05))
-        pairwise_results[f'T-Test {pair_name}'] = {"Statistic": t_stat, "p-value": t_p}
+        pairwise_results['T-Test'][pair_name] = {"Statistic": t_stat, "p-value": t_p}
 
     # ANOVA and post-hoc tests if applicable
     score_columns = [v + score_suffix for v in variables]