Figure 2 does not show substantial consistent sex differences in the TICs at each age, though the test information appears to span a slightly broader range of the underlying dimension of poly-substance use involvement for males than for females and provides peak information at slightly higher trait levels for females than for males. While there was limited evidence of item-level sex differences, the parameterization of the dimension as a whole differed slightly between sexes, as indicated by the fit of the model worsening when either location or discrimination parameters were constrained equal across sexes (see Table 2). As such, individual trait estimates resulting from the measure of substances used should still be estimated in a sex-specific context.
