Fig 1.
Person-item threshold distribution using all items.
All 146 individuals assessed using the original ten items of the Static-99R risk assessment instrument were included. Be aware that figures generated by the software use a decimal comma instead of a decimal point.
Fig 2.
Category probability curves for items 1 and 5 of the Static-99R.
Both item 1 (“age at release from index offense”) and item 5 ([number of] “prior sexual offenses”) present four response categories each (N = 146). Be aware that figures generated by the software use a decimal comma instead of a decimal point.
Table 1.
First sequential analyses.
Fig 3.
Category probability curves for modified items 1 and 5 of the Static-99R.
Both item 1 (“age at release from index offense”) and item 5 ([number of] “prior sexual offenses”) demonstrate ordered thresholds with three response categories each (N = 146). For item 1, category 1 (“aged 35 to 39.9”) and category 2 (“aged 40 to 59.9”) were combined, and for item 5, category 1 (“charges 1, 2” or “convictions 1”) and category 2 (“charges 3–5” or “convictions 2, 3”) were combined.
Fig 4.
Person-item threshold distribution using seven items.
All 146 individuals assessed by the revised seven-item Static-99R risk assessment instrument. Be aware that figures generated by the software use a decimal comma instead of a decimal point.
Fig 5.
Threshold map during step 2 of the alternative sequential analyses.
All 146 individuals assessed using all ten items of the Static-99R risk assessment instrument, ordered by ascending item location in logits on the x-axis. Please note that at this analysis stage, the disordered thresholds for polytomous items 1 and 5 have been corrected, but the ill-fitting item 3 has not yet been removed.
Table 2.
Alternative sequential analyses.
Fig 6.
Person-item threshold distribution using nine items.
All 146 individuals assessed by the revised nine-item Static-99R risk assessment instrument. Be aware that figures generated by the software use a decimal comma instead of a decimal point.
Fig 7.
The scatterplot visualizes person location in logits before and after splitting item 1 according to affiliation with the “substance use disorder yes/no” group of the 146 individuals assessed by the revised nine-item Static-99R risk assessment instrument. Only minor differences were found, with a slope of 1.048 (CI 95% 1.03–1.07, Pearson’s r = 0.994, r2 = 0.988).