These results are from a nine-year project funded to the second author by the National Institute on Alcoholism and Alcohol Abuse (R37AA10199). Data were collected from 1,599 cohabiting or married heterosexual couples using a multistage cluster sample from the 48 contiguous states. The men and women were interviewed separately. Eleven items were adapted from the Violence subscale of the Conflict Tactics Scale and were asked of each participant twice: once for perpetration, and for victimization. Typically, 12-month prevalence of intimate partner violence (IPV) is estimated as a deterministic proportion; either the percent of respondents where both partners have together reported at least one violent event (lower-bound estimate), or where at least one partner reported one or more violent events in the previous year (upper-bound estimate). However, another approach to the estimate of IPV prevalence is a stochastic (probabilistic) analysis. In this approach we view the partners' responses to the violence items as indicators of two or more latent classes in what is known as latent class analysis. We estimated two, three, four, and five-class models using sample weights with the twenty-two items responded to by the men and women for both male-to-female and female-to-male intimate partner violence (MFIPV; FMIPV) in the software package PANMARK. Using the Bayesian Information Criterion (BIC) for model selection, a three-class model was selected for both MFIPV and FMIPV, with estimated class proportions of 88.2%, 10.5%, and 1.3%; and, 86.3%, 11.6%, and 2.1%, respectively. These results suggest a tripartite typology of MFIPV and FMIPV. See pweb.netcom.com/~statcon/cvhtml.html
Learning Objectives: 1. Attendees will be able to differentiate between stochastic and deterministic estimation of prevalence. 2. Attendees will be able to identify and list three types of interpersonal violence.
Keywords: Domestic Violence, Epidemiology
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
Disclosure not received
Relationship: Not Received.