Fig. 5

Histograms showing replication of the associations between participation in unique clusters of activities and depressive symptoms, based on the results of the permutation testing (The histograms show the null distribution of t-statistics generated by randomly shuffling data 1,000 times. Observed effects in the tails of or outside the null distribution demonstrate that effects are non-parametrically significant and more extreme than chance). The dashed line indicates the observed t-statistic derived from linear mixed effect models. The negative t-statistic observed for social–physical activities is expected because of the negative association between social–physical activities and depression