| The US | Active service | Spouse is military/ civilian. | Hours worked per week Earnings per year | Hours worked: Wives of military: Self-report hours worked for pay per week. Wives of civilian: Self-report average hours worked per week over the past year Self-reported raw earnings: Wives of military: Self-report pretax personal income over the past year from all sources. Categorical measure with 20 levels. Wives of civilian: Sum of self-reported annual personal income over eight categories. Continuous measure, categorised by the authors. | Spouse’s Military branch/ service Spouse’s rank Times moved Spouse’s number of deployments | Military wives work less hours (approx. −15 hours per week; no statistical test reported) than population norms of matched census data. Military wives earn less (approx. US$17000 per week; no statistical test reported). than population norms of matched census data. None of the sociodemographic characteristics measured (age, education, minority status, children under 6 years) predicted hours gap. However, there was evidence that some characteristics of the military service-person’s service may predict hours gap, namely serving in the Navy vs Army (B = 3.00, SE B = 1.30, p < 0.05), rank of E1 - E3 vs E4 - E5 (B = −7.22, SE B = 2.06, p < 0.001), rank of E6 - E9 vs E4 - E5 (B = 3.30, SE B = 1.55, p < 0.05), and times moved (B = −2.18, SE B = 0.46, p < 0.001). Many of the sociodemographic characteristics predicted earnings gap, Age (B = −443.25, SE B = 108.18, p < 0.001), all categories of education (with bachelor’s degree or higher being associated with the largest gap (B = −14722.17, SE B = 1497.38, p < 0.001). There was also evidence that the serviceperson’s service characteristics were associated with magnitude of gap, including service in the navy (B = 2864.63, SE B = 1047.10, p < 0.01) or Air Force (B = 2781.40, SE B = 1260.10, p < 0.05) vs serving in the army; having a rank of E1 - E3 (B = −3122.17, SE B = 1455.03, p < 0.05) or E6 - E9 (B = 3373.53, SE B = 1375.80, p < 0.05) vs having a rank of E4 - E5; and times moved (B = −2261.43, SE B = 431.98, p < 0.001). Considering only the military wives who report any earnings per week, military wives have a similar number of hours worked compared to population norms of matched census data, however they earn less for these hours of work. Highly educated military wives did not appear to be disproportionally affected by these gaps. The article did not present the results of the statistical tests with these findings. | Coefficient B SE of coefficient. B p-value |
| | Current serving | Male partner is military/ civilian | Controlling attitude Physical abuse Emotional abuse Sexual abuse | Modified form of the WHO standardized questionnaire for collection of data on women’s health and domestic violence. | N/A | Whether the covariates were adjusted or not is not clear. Comparing the prevalence of the IPV categories for military-involved women vs civilian-involved women: Controlling behaviour: 37.1% vs 29.1% (p = 0.1) Physical abuse: 42.4% vs 13.4% (p = 0.001) Emotional abuse: 42.4% vs 13.4% (p = 0.0001) Sexual abuse: 9.2% vs 8.8% (p = 0.44) Majority of the women in both the military and civilian populations were married, but respondents in the civilian community were better educated 67 (62.0%) had tertiary education compared to 30 (23.1%) in the military population (p = 0.000). Mean ages for civilian and military partners were 44.9 + 9.6 and 37.75 + 5.90, respectively (p = 0.00). | % p-value |