| Author (Year) Country | Point of service | Indicator variables | Outcome variables | Outcome measure (s) | Covariates | Relevant findings | Effect size |
| The US | Unknown | Age Motive Behavioural health Context of event | Perpetrator is military/ non- military | Precoded variable in the National Violent Death Reporting System dataset, derived from information received from police and medical examiner reports. | Ethnicity Marital status | Retrospective study investigating factors which predicted perpetrator being military vs civilian. Chi-square tests show that significant differences were found among age, race, education, marital status and primary motive among matched military vs civilian groups. Logistic regression results show Age group (OR: 1.06, 95% CI: 1.05 - 1.08; p < 0.001), declining health motive vs other/unknown motive (OR: 2.91, 95% CI: 1.25 - 6.76; p = 0.013), physical health problems (OR: 2.65, 95% CI: 0.09 - 0.774; p = 0.015) were associated with military group membership. | Chi-quare (c2) OR 95% CI p-value |
| The US | Baseline: New recruits Follow-up: Post-2 years of service | Age Gender Family income Education Marital status | Past-year severe intimate partner violence | Responding yes to any of the severe physical violence items of the Conflict Tactics Scale (CTS) “Hit (or tried to hit) other person, but not with anything” “Hit (or tried to hit) other person with something hard” “Kicked, bit or hit with a fist” “Beat the other person up” “Threatened the other person with a knife or gun” | Gender | Premilitary (baseline): 11% of respondents self-reported severe IPV within the past year. Proportion was higher among women than men [20% vs 4%, c2 (1) = 43.99, n = 963, p < 0.001]. At baseline, age was significantly associated with premilitary SIPV for men (r = 0.12, n = 421, p < 0.05) but not for women (r = 0.03, n = 542, not significant [p value not reported]). To test for non-linear association between age and SIPV, c2 tests of association were also conducted but similarly found a significant association for men but not for women. Ethnicity was significantly associated with SIPV for men (c2 (2) = 9.51, n = 359, p < 0.01) and women (c2 (2) = 20.01, n = 476, p < 0.001). Neither of family income, education level, or marital status were associated with past-year SIPV. Second-year of service: Proportion reporting past-year severe IPV was 14%, with no significant difference by gender [c2 (1) = 2.23, n = 963, p > 0.10]. At second year of service, age was not related to SIPV perpetration for men or women. Ethnicity was not associated with SIPV for either men (c2 (2) = 0.83, n = 359, p > 0.65) or women (c2 (2) = 4.03, n = 476, p > 0.13). Neither of family income, education level, or marital status were associated with past-year SIPV. Change: Apparent increase from 11% to 14% between premilitary and second year was not statistically significant [McNemar change test c2 (1) = 3.43, n = 963, p < 0.07]. A trend appeared when stratified by gender. For men, proportion reporting past-year SIPV increased between measurement points increased [4% vs 16%, McNemar change test c2 (1) = 32.51, n = 421, p < 0.001], but for women there was a significant decrease [20% vs 12%, McNemar change test c2 (1) = 4.92, n = 542, p < 0.05]. | McNemar’s Chi-square (c2) n p-value |