“…CFA assists in the reduction of measurement error and allows the comparison of a priori models at the latent factor level (Jackson et al, 2009;Atkinson et al, 2011). In the present study, there were three a priori models regarding the construct validity of the RCTQ: (i) the single-factor model yielded by the STOQ-9 results (Keizer et al, 2014), (ii) the bifactorial model, referring to the racing vs. crowded thoughts theoretical distinction (Piguet et al, 2010), and (iii) the three-factor model, based on the results obtained in healthy individuals (Weiner et al, 2018). For the single-factor model, the 34 items of the RCTQ were included in the analysis (model 1); for the bifactorial model (model 2), 17 items belonging to the 'racing' a priori factor, i.e., items 2, 5, 7, 8, 11, 12, 14, 17, 20, 21, 22, 25, 28, 30, 31, 32 & 34, and 17 items belonging to the 'crowded' a priori factor, i.e., items 1, 3,4,6,9,10,13,15,16,18,19,23,24,26,27,29,33, were included; for the three-factor model (model 3), factor 1 consisted of items 1, 2, 4, 5, 7, 8 & 9, factor 2 consisted of items 10, 13, 14, 15, 18, 23, 29 & 30, and factor 3 consisted of items 12, 17, 20, 21, 22, 25, 31 & 32 (Weiner et al, 2018).…”