We selected a six-factor structure as it minimized cross-loading between factors while including as many items as possible, accounted for the most variance (75%), and produced the same factor structure for both parent- and self-reported data (an asset for use in population-based samples). Two items, “experiences unwanted upsetting thoughts or images” (referred to as ‘upsetting’) and “spends time checking and rechecking homework” (referred to as “homework”), factored separately from the other items and were excluded from the final factor model. Upon re-examination of these two items, we considered the ‘upsetting” item too general, capturing a broad, non-specific trait, and the “homework” item, intended to capture a checking compulsion, too specific. The six factors were as follows: Cleaning/Contamination, Symmetry/Ordering, Superstition, Rumination, Counting/Checking, and Hoarding (Table 1). Distributions of these factors, as reported in Park et al.21, are shown in Supplemental Fig. 1. The results were similar when analyzed using promax and varimax rotation.Table 1Factor analysis of the TOCS (19 items)Factor loadingTOCS ItemFactor 1: Cleaning/ContaminationFactor 2: Symmetry/OrderingFactor 3: SuperstitionFactor 4: RuminationFactor 5: Counting/CheckingFactor 6: HoardingWash0.840.090.090.130.190.06Germs0.840.170.180.100.090.10Clean0.800.130.170.150.170.05Dirt0.770.310.090.110.100.03Ruined0.490.400.370.010.070.25Interfere0.210.760.170.150.180.21Not exactly0.280.660.190.420.110.16Symmetrical0.260.640.120.110.460.10Repeat0.240.600.290.240.350.13Bad luck0.170.180.790.080.270.21Special0.140.250.740.110.340.05Healthy0.350.110.640.330.070.17Guilty0.170.190.140.820.250.11Thinking0.180.230.170.800.270.12Checks0.270.120.170.250.740.07Count0.170.400.260.150.660.15Do certain0.090.200.260.220.710.15Throwing0.120.150.120.150.090.87Useless0.070.170.160.070.160.87The table shows factor loadings for each of the 19 items on the six obsessive–compulsive (OC) dimensions from the Toronto Obsessive–Compulsive Scale (TOCS)
