To examine whether victim, offender, and relationship characteristics increase the risk and frequency of intimate partner violence (IPV). The effect of separation length on the risk of IPV is also evaluated.

Using abuse incident data (N = 4,960) from IPV victims (n = 497), who accessed a health-care facility over a one-year period, between-individual differences in the frequency and timing of abuse were assessed. Dependence of event times within respondents was accounted for with conditional risk set "gap time" models, which stratified participants across event number and failure order.

Employment of both partners and only the victim decreased the risk and frequency. Employment of only the abuser and length of separation decreased frequency but was not significantly associated with timing of abuse. Separated and non-separated respondents were not significantly different in frequency (incident risk ratio = 1.06, p > .05) and timing of abuse, (exp (b) = 0.95, p > .05).

By capturing the timing between abuse incidents, insight into the long-term risk of abuse is provided and accounts for selection effects. Findings call for an improvement in measuring relationship status indicators, including relationship length and time since separation. Future research should examine within-individual changes in separation and IPV risk.