Assessing callous-unemotional traits using the total score from the inventory of callous-unemotional traits: A meta-analysis

The Inventory of Callous-Unemotional Traits (ICU) is a widely used rating scale measure of callous-unemotional (CU) traits. Most studies have used a unit-weighted total score that sums all of the items into a composite, despite the consistent finding that items from this measure can be described using a bifactor model (1 general factor and 3 bifactors). We conducted a meta-analysis using the results of past published bifactor tests of the ICU, using indices to estimate the sources of variance in the total and subscale scores. We included studies in our review that tested and found adequate fit for the bifactor model of the ICU resulting in 12 published studies (M age = 9–26 from both community and justice-involved samples) using either the self-report (n = 9) or parent-report (n = 4) versions of the ICU scale (one published study conducted separate factor analyses for each version). We reported model-based reliability estimates that break down unit-weighted ICU scores into the variance due to individual differences in the general factor and the variance due to individual differences in the bifactors. The reliable variance in the unit-weighted total scale score of the ICU is largely determined by the general factor. Findings support the ICU total score as a measure of the general construct of CU traits, despite support for a bifactor model in multiple samples. Concerns about using the unit-weighted subscale scores from the ICU are raised since their reliable variance was strongly influenced by the general factor.