Category: Criminal Justice / Forensics


Measurement of Criminal Thinking and Prediction of Aggression in College Students

Sunday, November 19
10:15 AM - 11:45 AM
Location: Sapphire Ballroom I & J, Level 4, Sapphire Level

Keywords: Criminal Justice | Assessment | Aggression / Disruptive Behaviors / Conduct Problems
Presentation Type: Symposium

Criminal thinking (CT) is a key predictor of criminal behavior and routinely measured in forensic populations with the Psychological Inventory of Criminal Thinking Styles (PICTS; Walters, 1995), which distinguishes two types of CT, reactive (RCT; e.g., problem avoidance, rashness) and proactive (PCT; e.g., externalizing blame, entitlement). While antisocial behaviors exist on a continuum in community settings, far less is known about the nature and role of CT in such contexts, although recent evidence (Mitchell et al., 2017) supports the validity and utility of measuring CT in young adults using a layperson version of the PICTS (PICTS-L; Walters et al., 2009). We replicated and extended this work by investigating the validity of the PICTS-L in an independent college student sample (N = 508, 73% female, age M = 19.26, SD = 1.47),  comparing  measurement models with CFA using WLSMV (MPlus) to account for non-normal categorical response data, and examining unique associations with aggression functions with the Reactive-Proactive Aggression Questionnaire (Raine et al., 2006). Fit indices supported a bifactor model (RSMEA = .03, CFI = .93) comprising a general CT factor (GCT) and separate proactive (PCT) and reactive (RCT) factors. A 2-factor PCT/RCT model offered slightly poorer fit. Omega, an index of full bifactor model reliability (.94), minus omegaH, an index of general factor (GCT) reliability (.83), indicated that 11% (.94-.83) of variance reflects multidimensionality via PCT (omega= .94) and RCT (omega= .94) factors. In regression analyses, GCT predicted unique variance in both reactive (RA; beta = .16, p < .001) and proactive (PA; beta = .30, p < .001) aggression. In separate analyses, RCT uniquely predicted RA (beta = .12, p = .009), and PCT uniquely predicted PA (beta= .30, p < .001). Findings support applicability of CT in community settings and offer potential cognitive treatment targets associated with different forms of aggressive behavior in young adults. Consistent with the conference theme, we consider how our findings can inform the extension of CBT principles to the context of assessment and treatment of criminal thinking in community settings.

Lauren Delk

Graduate Student
Virginia Tech


Send Email for Lauren Delk


Measurement of Criminal Thinking and Prediction of Aggression in College Students

Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Measurement of Criminal Thinking and Prediction of Aggression in College Students