Fig. 4

Best-fitting model estimates for Repetitive Motor scores plotted over raw data for (a) MNLFA-derived factor score estimates and (b) raw mean scores. The model outcome for the raw mean score is converted to probability, as a logistic linking function was used. Y-axes for both plots are scaled with limits of mean outcome +/− 2.5 SDs. c Impact of sex on intercept and slope. Parents of females reported fewer repetitive motor behaviors at 18 months, and their rate of decline was more rapid relative to males