Researchers studying intimate partner violence (IPV) have numerous difficulties to face when analyzing data from one or both members of the dyad. In the case of heterosexual IPV where partners may be considered nonexchangeable on the basis of sex, two types of violence indexes can be constructed: male-to-female and female-to-male IPV. Then each partner can be classified as a victim, a perpetrator, or both. Another critical dimension that has been rarely studied to date is offensive vs. defensive IPV. These multiple classifications can leave the researcher with a bewildering array of indexes to choose as representing IPV in their models. Considerable evidence now exists that partners do not agree about the occurrence or the frequency of IPV. This suggests that the use of summary (i.e., combined-response) indexes in statistical models of IPV where data is available from both partners is problematic. Disagreement between the couples also makes the use of certain simple covariance models invalid. While more difficult to model, victim and perpetrator responses should be analyzed as separate, but related constructs. Examples will be given of how proper models may be developed, and how these models differ in substantive interpretation from those models where a combined-response index of IPV is used. For IPV researchers confronted with data from only a single member of the dyad, suggestions are also made with accompanying examples based on a modification of the weighting techniques developed by Heyman & Schlee (1997).
Learning Objectives: 1. Participants will be able to state three difficulties with combined-response indexes of intimate partner violence. 2. Participants will be able to name, and use, two methods of appropriately modeling dyadic-level intimate partner violence data. 3. Participants will be able to name, and use, at least one method of generalizing single-level partner data to the upper-bound estimate of intimate partner violence
Keywords: Domestic Violence, Methodology
Presenting author's disclosure statement:
Organization/institution whose products or services will be discussed: None
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.