While prevalence of use of individual substances displays consistent effects of age and sex (with use increasing with age and being greater in males than in females, e.g. see Table 1 and Derringer et al., 2008; Wallace et al., 2003), these factors impact substance use at the level of an underlying dimension rather than in an entirely substance-specific way. This was indicated by model fit worsening when parameters were constrained to be equal across either age or sex (see Table 2) and by age- and sex-related trends in estimated parameters generalizing across substances (see Table 3). When examining developmental effects on individual substances, location (the level of the trait) tended to decrease as age increased, indicating that an older person expressing the same degree of poly-substance involvement as a younger person possesses less of the trait. This trend can be observed functioning at the level of the overall poly-substance dimension in the peaks of the TICs (see Figure 2), which moved toward a more central trait level as age increased. This shows that diversity of substances used provides information on
