Fig. 1

Results from cross-lagged model. Data from PATH youth surveys wave 2 to wave 5 (2014 to 2019). Solid lines represent paths that were statistically significant at 0.05 level. Dashed lines represent paths that were not statistically significant at 0.05 level. Observed variables are shown in rectangles and unobserved variables (error terms) are shown in circles. Similar variables are presented in rows, and variables from the same assessment wave are shown in columns. Curved lines represent covariances. Straight lines represent regression relationships. Note that we structured the data from multiple PATH waves such that the first data point corresponds to the first assessment of the individual (“Time 1”), the second data point corresponds to the second assessment of the individual (“Time 2”) and so on. As a result, a youth from the wave 2 cohort can contribute up to five data points across all waves, with up to four person-intervals between assessments. Similarly, a youth in the replenishment ('shadow sample') who became eligible in wave 3 can have up to four data points, with up to three person-intervals