CT Study childhood trauma exposure data were examined to identify items with endorsement frequencies insufficient for inclusion in factor analyses (i.e., resulting in problematic empty cross‐tabulated cells). Items assessing more severe forms of abuse within each domain, and those with nonbinary ordinal responses were determined to have inadequate endorsement for inclusion as initially operationalized. Several steps were taken to address these issues. Low endorsement items assessing similar forms of abuse were combined. Items with nonbinary ordinal responses were recoded to be binary variables representing endorsement at any level. In the interview's CPA section, participants were asked parallel series of questions about abuse by mother, father, and other adult household members. Items assessing parallel sets of items were combined across parents; those covering nonparent household members were dropped due to extremely low endorsement. These changes reduced the number of items included in the first‐order factor analysis to 37 (CPA = 9, CSA = 13, PPA = 15).
