Table 5: Association of sociodemographic and military characteristics with lifetime sexual assault category among women*†
CharacteristicRelative risk ratio (95% CI)‡
Weighted n = 8900§
UnadjustedAdjusted
Lifetime non-military-related sexual assault v. no sexual assaultLifetime military-related sexual assault v. no sexual assaultLifetime non-military-related sexual assault v. no sexual assaultLifetime military-related sexual assault v. no sexual assault
Age group, yr
    < 30ReferenceReferenceReferenceReference
    30-391.66 (1.05-2.62)1.36 (0.80-2.31)1.62 (0.98-2.66)1.22 (0.67-2.23)
    ≥ 40 yr1.93 (1.22-3.05)1.64 (0.98-2.74)2.10 (1.20-3.67)1.44 (0.79-2.65)
Ethnicity
    WhiteReferenceReference- ¶
    Other1.44 (0.84-2.49)1.06 (0.54-2.11)- ¶- ¶
Marital status
    Single (never married)1.19 (0.79-1.81)1.91 (1.24-2.94)1.57 (0.98-2.52)2.58 (1.56-4.25)
    Married/common-lawReferenceReferenceReferenceReference
    Widowed/divorced/separated0.93 (0.52-1.66)0.78 (0.39-1.56)0.73 (0.38-1.41)0.62 (0.29-1.35)
Region
    Atlantic0.90 (0.55-1.46)0.77 (0.43-1.40)0.97 (0.57-1.64)1.01 (0.52-1.97)
    Quebec0.89 (0.51-1.57)0.84 (0.41-1.72)0.87 (0.48-1.57)1.02 (0.48-2.14)
    OntarioReferenceReferenceReferenceReference
    Western1.16 (0.74-1.83)1.97 (1.27-3.07)1.16 (0.71-1.88)2.34 (1.47-3.72)
Rank category
Junior noncommissioned memberReferenceReferenceReferenceReference
Senior noncommissioned member0.75 (0.49-1.14)1.76 (1.11-2.78)0.64 (0.39-1.06)1.93 (1.11-3.35)
    Officer0.64 (0.44-0.93)1.29 (0.81-2.06)0.75 (0.49-1.14)2.05 (1.21-3.45)
Element
    ArmyReferenceReference- ¶- ¶
    Navy0.81 (0.51-1.30)1.28 (0.78-2.10)- ¶- ¶
    Air Force1.00 (0.68-1.48)0.87 (0.55-1.38)- ¶- ¶
Ever deployed
    NoReferenceReferenceReferenceReference
    Yes1.25 (0.88-1.78)1.75 (1.17-2.64)0.76 (0.51-1.15)0.86 (0.54-1.37)
Count of types of lifetime nonsexual trauma1.33 (1.24-1.44)1.46 (1.34-1.60)1.34 (1.23-1.45)1.51 (1.37-1.66)

Note: CI = confidence interval.

*Multinomial logistic regression.

†Cases representing 0.2% of the population were excluded because of missing values.

‡Results are presented as relative risk ratios as opposed to odds ratios so as to provide more comparable estimates of effect size across covariates and outcomes.

§Weighted to the entire Canadian Regular Force in service in September 2012. To protect the identity of respondents, Statistics Canada permits the release only of weighted cell counts, which must be rounded to the nearest 20. We calculated prevalence rates using these rounded cell counts.

¶Not included in final model owing to the lack of an unadjusted association with the outcome.