Factor-Analysis of the MCMI-III on an Offender Population

Abstract

The various versions of the Millon Clinical Multiaxial Inventory (MCMI) are frequently employed in forensic settings to inform decisions regarding the treatment and management of offenders. A paucity of research, however, has investigated the psychometric properties of this instrument with offenders, and no published studies have yet addressed the factorial structure of its most recent edition, the MCMI-III. This study investigated the factorial validity of the MCMI-III in a sample of 159 male, incarcerated offenders. Factor analytic results on the 10 clinical scales revealed two factors, accounting for 70% of the common variance in participants' responses. Factor I (Psychiatric Disturbance) consisted of 8 scales and Factor II (Substance Dependence) was comprised of 3 scales. A 3-factor solution was obtained in the analysis of the 14 personality scales. Factors I (Internalizing; 10 scales), II (Acting Out; 5 scales) and III (Self-Absorption; 3 scales) accounted for a total of 77% of the common variance. The theoretical and clinical implications of these results and directions for future research are discussed. [Article copies available for a fee from The Haworth Document Delivery Service:

An examination of the factor structure of the Millon Adolescent Clinical Inventory in a sample of detained adolescent boys.

“…8 In the present study, the doleful, conforming, borderline tendency, and self-demeaning PP scales loaded on both the internalizing and externalizing factors. This pattern of cross-loadings is consistent with prior findings with the MACI (Salekin, 2002) and the MCMI-III (Dozois & Kellin, 1999) in delinquent samples. The doleful and self-demeaning scales share similar themes such as hopelessness and pessimism (McCann, 1999), which are likely more common in troubled adolescents more generally, and, especially, in detained adolescents shortly after incarceration.…”

“…The doleful and self-demeaning scales share similar themes such as hopelessness and pessimism (McCann, 1999), which are likely more common in troubled adolescents more generally, and, especially, in detained adolescents shortly after incarceration. The observed pattern of cross-loadings identified for the borderline tendency scale are more obvious and replicate all prior factor analysis results with the MACI (Adkisson et al, 2012; Romm et al, 1999; Salekin, 2002; see also Hopwood & Grilo, 2010) and studies using the MCMI-III (Craig & Bivens, 1998; Dozois & Kelln, 1999). 9 In addition, the conforming scale loaded negatively on both factors—a finding which replicates prior EFA studies with the MACI (Adkisson et al, 2012; Hopwood & Grilo, 2010; Salekin, 2002)—and may reflect sample characteristics of a delinquent population that has already exhibited a lack of conformity.…”

“…2 However, Salekin (2002) and Adkisson et al (2012) disagreed-each citing theoretical reasons-on which of the remaining MACI scales to include in their analyses. Adkisson et al (2012) criticized Salekin's (2002) domain-specific analysis of the scales as inconsistent with traditional factor analytic methodology even though the approach used by Salekin (2002) had been conducted with the adult counterpart of the MACI (i.e., the MCMI) using a similar approach (see Dozois & Kellin, 1999;Retzlaff & Gibertini, 1987). Adkisson et al (2012) excluded the EC scales reasoning that they were more indicative of individual adolescent concerns, and, consequently, not indicative of adolescent psychopathology or personality-the key components of Millon's theory.…”

Characterising the psychiatric comorbidities among adults with ADHD who have a history of offending

“…Also the MCMI-III has been identified as having good diagnostic value and is a reliable instrument (Craig, 1999;Dozois & Kelln, 1999).…”

The Millon Adolescent Clinical Inventory in the Assessment of Juvenile Offenders

“…This approach separated the scales as Millon intended, and is consistent with much of the factor analytic literature involving the MCMI-II, in which researchers typically conduct separate analyses for clinical and personality scales (Dozois & Kelln, 1999;Retzclaff & Gibertini, 1990). One caveat concerns the predictive versus postdictive nature of this analysis.…”