A Latent Class Analysis of Specific Family Maltreatment Behaviors Among Active-Duty Air Force Members

Background and Purpose: Insightful variable-oriented research has identified correlates of specific family-maltreatment behaviors among active-duty Air Force members (Airmen; Bowen, Jensen, & Williams, 2016). There remain valuable opportunities to employ person-oriented analytic approaches to uncover latent patterns of specific family-maltreatment behaviors among Airmen. Understanding such patterns, and their correlates, could inform prevention efforts led by stakeholders and practitioners associated with the Air Force Family Advocacy Program (AF-FAP). Consequently, the purpose of the current study was to examine holistically specific maltreatment behaviors among a representative sample of Airmen. To evaluate the practical utility of identified patterns of maltreatment behavior, we also assessed associations between patterns and theoretically plausible correlates representing various personal and family strengths that could be targeted by prevention efforts.

Methods: The 2011 Air Force Community Assessment yielded a sample of 30,541 Airmen (83% male; 45% in 26-35 age group) who had at least one child and were in a committed couple relationship. Four self-reported measures of maltreatment perpetration were used to indicate whether the respondent had perpetrated physical partner maltreatment, emotional partner maltreatment, physical child maltreatment, or emotional child maltreatment over the last 12 months. Latent Class Analysis was conducted to extract an optimal number of latent response patterns across the four maltreatment measures. The 3-step procedure, which accounts for classification uncertainty, was used to assess latent-class differences with respect to standardized latent-factor scores representing unit leader support, neighbor support, family coping, couple relationship quality, partner support, physical fitness, mental fitness, and personal resilience (α=.77-.98).

Results: Results indicated a three-class solution was optimal. Consistent with known rates of family maltreatment within the Airmen population, the vast majority of respondents clustered into a pattern marked by very low conditional probabilities of family maltreatment behaviors (n = 29,612; 97%), with child physical maltreatment being the most probable behavior (i.e., 10% conditional probability). A small subset of respondents clustered into a partner-centricpattern marked by moderate-to-high conditional probabilities of perpetrating physical and emotional partner maltreatment (n = 456; 1.5%). The remaining participants clustered into a child-centric pattern marked by moderate-to-high conditional probabilities of perpetrating physical and emotional child maltreatment (n = 472; 1.5%). Both the partner-centric and child-centric patterns were associated with relatively low levels of the personal and family strengths assessed. The partner-centric pattern especially was associated with low levels of partner support (z = -1.52), couple relationship quality (z = -1.71), family coping (z = -1.64), mental fitness (z = -1.55), and physical fitness (z = -1.00). The child-centric pattern especially was especially associated with low levels of personal resilience (z = -0.80).

Conclusions and Implications: Our findings offer a holistic perspective with respect to possible patterns of family maltreatment behaviors among Airmen. Associations between these patterns and various personal and family strengths offer practical insights related to maltreatment prevention in general, and the prevention of specific patterns of maltreatment in particular, whether those patterns are partner-centric or child-centric in nature. Study findings could inform efforts by the AF-FAP to strengthen families, prevent maltreatment, and promote mission readiness.