Table 2. Associations between views on gender roles/norms and IPV perpetration (N = 18 studies).
| Citation | Measurea (No. of items) | Indicator attributes | Sample description & size | Scale rangeb | Analysis method | Definition of Violence Perpetration [Male to female] | Resultsc | Indicator summary of significanced | |
|---|---|---|---|---|---|---|---|---|---|
| Gender equitable men (GEM) scale | |||||||||
| Das, 2014 [39] | Modified (m)GEM scale–(15 items) | Gender roles, acceptance of control over women, sexual entitlement, IPV inclusive | Boys ages 10–16 in urban Mumbai, India. Part of school or community-based cricket team (N = 1040) | High vs. low equity; Moderate vs. low equity (rev) | Multivariate logistic regression | Outcome 1: Perpetrated sexual or verbal violence last 3 months | Outcome 1: High v. low equity aOR: 0.29* (0.11, 0.80) | Consistently positive association | |
| Mod v. low equity aOR: 0.44 (95%CI: 0.18, 1.11)† | |||||||||
| Outcome 2: Perpetrated sexual violence last 3 months (incl. harassment) | Outcome 2: High v. low equity aOR: 0.09* (0.04, 0.23) | ||||||||
| Mod v. low equity aOR: 0.31** (0.20, 0.48) | |||||||||
| Gomez, 2011 [42] | GEM scale (24 items) | IPV inclusive, sexual entitlement | Young men ages 15–24 in urban slum of Rio de Janeiro (N = 240) | Mean = 0 (range = -3.1 to 1.5) (rev) | Multinomial logistic regression | IPV perpetration in past 6 months (physical, sexual or emotional) | aRRR: 0.69* (0.40, 0.89) | Positive association | |
| Fleming, 2015 [17] | (m)GEM scale–Brazil (11 items) | IPV inclusive | Men ages 18 to 59 surveyed in IMAGES multi-country survey (N = 7806 in pooled sample). Data from Bosnia and Rwanda are nationally representative; other countries are representative of regions/cities surveyed. | Standardized in each country, Mean = 0, SD = 1; score represents respondent’s score relative to other men surveyed in country (rev) | Multivariate logistic regression | Physical perpetration (lifetime) | Brazil: aOR: 0.99 (0.79, 1.23) | No association | |
| (m)GEM scale–Chile (15 items) | Chile: aOR: 0.87 (0.74, 1.01) | No association | |||||||
| (m)GEM scale–Mexico (11 items) | Mexico: aOR: 0.68** (0.56, 0.82) | Positive association | |||||||
| (m)GEM scale–Bosnia (15 items) | IPV inclusive | Bosnia: aOR: 0.68** (0.58, 0.80) | Positive association | ||||||
| (m)GEM scale–Croatia (13 items) | IPV inclusive | Croatia: aOR: 0.87 (0.75, 1.02) | No association | ||||||
| (m)GEM scale–DRC (13 items) | IPV inclusive | DRC: aOR: 0.92 (0.75, 1.40) | No association | ||||||
| (m)GEM scale–India (12 items) | IPV inclusive | India: aOR: 1.03 (0.85, 1.24) | No association | ||||||
| (m)GEM scale–Rwanda (13 items) | IPV inclusive | Rwanda: aOR: 0.94 (0.83, 1.06) | No association | ||||||
| Fulu, 2013 [3] | Gender attitudes scale (10 items) e | IPV inclusive, sexual entitlement | Men ages 18 to 59 surveyed in UN Multi-country study on Men and Violence sampled from a combination of urban and rural sites. Estimates are nationally representative in Cambodia only and regionally representative in Bougainville, Papua New Guinea. | Low equity vs. high or moderate equity | Multinomial logistic regression | Outcome 1: Physical IPV perpetration (ever) | Bangladesh Outcome 1: aRR 1.82* (1.35, 2.44); Outcome 3: aRR 2.22*(1.28, 3.84); Outcomes 2, 4: NR (ns) | Inconsistently positive | |
| Outcome 2: Sexual IPV perpetration (ever) | China Outcome 1–4: NR (ns) | ||||||||
| Outcome 3: Physical or sexual IPV perpetration (ever) | Cambodia Outcomes 1, 2 and 4: NR (ns); Outcome 3: aRR 2.31* (1.25, 4.28) | ||||||||
| Indonesia Outcomes 1–4: NR (ns) | |||||||||
| Outcome 4: Emotional or economic IPV perpetration (ever) | Sri Lanka Outcome 1–4: NR (ns) | ||||||||
| Papua New Guinea Outcomes 1–4: NR (ns) | |||||||||
| Nanda, 2014 [44] | (m)GEM scale (27 items) | IPV inclusive, sexual entitlement | Men ages 18–49 from 6 states in India (Uttar Pradesh, Rajasthan, Punjab & Haryana, Odisha, Madhya Pradesh, and Maharashtra), representative at each state level (total N = 9205) | Low vs. high/ moderate equity | Multivariate logistic regression | IPV perpetration (emotional, economic, physical or sexual) in past 12 months | aOR: 1.35** (1.15, 1.57) | Positive association | |
| Pulerwitz, 2015 [19] | GEM scale (24 items) | IPV inclusive | Young men ages 15–24 in Ethiopia (N = 729), part of community-engagement intervention | High equity vs. moderate or low (rev) | Multivariate logistic regression | Any IPV perpetration (physical, sexual, or emotional) | High-equity GEM scores were associated with a 34% reduction in the odds of perpetration† (95%CI: NR) | No association | |
| Verma, 2008 [23] | (m)GEM scale (15 items) | IPV inclusive | Young men ages 15–29 in Mumbai (urban site) and Gorakhpur (rural site), India (N = 660) | High, moderate, and low equity. Terciles created from continuous score (rev) | Multivariate logistic regression | Perpetration of physical or sexual IPV in past 3 months | Mumbai High v. low equity aOR: 0.69* (95%CI: NR) | Consistently positive association | |
| Mod. v. low aOR: 0.79† (95%CI: NR) | |||||||||
| Gorakhpur High v. low equity aOR:0.45** (95%CI: NR) | |||||||||
| Mod. v. low aOR: 0.73* (95%CI: NR) | |||||||||
| Verma, 2006 [30] | GEM scale (24 items) | IPV inclusive | Young men ages 16–24 in Mumbai, India (N = 107) | Continuous (range: NR) | Mean difference | Physical IPV perpetration in past 3 months | NR coefficient* | Positive association | |
| Other gender norms and belief scales | |||||||||
| Anderson, 2004 [43] | Rules about sex questionnaire (21 items) | Sexual entitlement | Male students ages 11 to 36 (middle/high school and university) in Indiana, USA (N = 137) | Continuous | Correlation | Frequency of perpetration of sexual coercion | r: 0.30** | Positive association | |
| Espinoza, 2012 [31] | Traditionalism subscale of Mirandé sex role inventory (MSRI) (17 items) | Young men age 15–18 in high school in Monterrey, Mexico (N = 75) | Continuous | Multiple linear regression | Outcome 1: physical IPV Outcome 2: emotional IPV | Outcome 1 Adj.B: -0.44** SE: NR | Inconsistently negative association | ||
| Outcome 2 Adj. B: -0.03, SE: NR | |||||||||
| Figueredo, 2001 [49] | Patriarchy scale (11 items) | IPV inclusive, male control over wealth | Men in Sonora, Mexico who were in a committed relationship during past year. Mean age = 33 (N = 106) | Continuous | Multiple linear regression | IPV perpetration (any type) | Adj Beta: -0.06, SE: NR | No association | |
| Gage, 2016 [32] | Gender stereotyping scale (7 items) | Male high school students in Port-au-Prince who had ever been on a date (N = 342) | Continuous | Multiple linear regression | Outcome 1: Psychological IPV perpetration Outcome 2: Physical/ sexual IPV perpetration (ever) | Outcome 1 Adj. B: 0.27, SE: 0.12 | No association | ||
| Outcome 2 Adj B: 0.23, SE: 0.20 | |||||||||
| Kalichman, 2007 [36] | Male role attitudes scale items [tested individually] | Men older than 18 in Cape Town, South Africa (N = 435) | NR | Multivariate logistic regression | Sexual assault perpetration (ever) | ||||
| It is essential for a man to get respect from others | aOR: 0.70 (0.40, 1.40) | No association | |||||||
| A man always deserves the respect of wife & children | aOR: 0.50* (0.20, 0.90) | Negative association | |||||||
| I admire a man who is very confident | aOR: 0.50* (0.20, 0.80) | Negative association | |||||||
| A man will lose respect if he talks about his problems | aOR: 0.80 (0.50, 1.20) | No association | |||||||
| A young man should be physically tough, even if he is not big | aOR: 0.70 (0.40, 1.30) | No association | |||||||
| I don’t think a husband should have to do housework | aOR: 1.60** (1.10, 0.60) | Positive association | |||||||
| Men are always ready for sex | aOR: 0.90 (0.50, 1.40) | No association | |||||||
| A man who does not provide for his family is less than a man | aOR: 1.10 (0.60, 1.90) | No association | |||||||
| Hostile attitudes towards women scale items [tested individually] | Men older than 18 in Cape Town, South Africa (N = 435) | NR | Multivariate logistic regression | Sexual assault perpetration (ever) | |||||
| Many women seek special favors that place them over men | aOR: 1.70* (1.10, 2.90) | Positive association | |||||||
| Most women think innocent remarks or acts are meant to hurt them | aOR: 1.10 (0.70, 1.80) | No association | |||||||
| Women are too easily offended | aOR: 1.30 (0.80, 2.30) | No association | |||||||
| Most women fail to appreciate all that men do for them | aOR: 1.10 (0.70, 1.90) | No association | |||||||
| Women who have jobs and make money should give the money to their man to pay bills | Male control over wealth | aOR: 1.20 (0.70, 1.90) | No association | ||||||
| Women only work so they can gain power and control over men | Male control over wealth | aOR: 1.70* (1.10, 2.70) | Positive association | ||||||
| Once a woman makes money she usually tries to control her man | Male control over wealth | aOR: 1.40 (0.90, 2.20) | No association | ||||||
| It is difficult for a man to work at a job where a woman is the boss | aOR: 0.80 (0.50, 1.20) | No association | |||||||
| A woman should only show her man respect in front of other people. | aOR: 0.80 (0.50, 1.30) | No association | |||||||
| Some women need a man to help them survive | aOR: 2.20** (1.20, 4.10) | Positive association | |||||||
| Reed, 2011[37] | Gender attitudes scale (13 items) | Sexual entitlement | Young men ages 14–20, seeking healthcare at clinics in Boston, USA (N = 320) | Continuous | Multiple linear regression | Teen dating violence perpetration (physical, sexual or emotional) (ever) | Total sample Adj. beta: 1.50*, SE: 0.60 | Consistently positive | |
| Sexually active subgroup Adj. beta: 2.00*, SE: 0.90 | |||||||||
| Sambisa, 2010 [41] | Attitudes about wife working outside the home (2 items) | Male control over wealth | Married men ages 15 to 49 in Bangladesh (N = 8320) | Support for wife working outside home in at least once instance (vs. none) | Multivariate logistic regression | Outcome 1: Lifetime physical IPV perpetration | Outcome 1 aOR: 0.92, (95%CI: NR) | No association | |
| Outcome 2: Past-year physical IPV perpetration | Outcome 2 aOR: 0.87, (95%CI: NR) | ||||||||
| Outcome 3: Lifetime sexual IPV perpetration | Outcome 3 aOR: 1.01, (95%CI: NR) | ||||||||
| Outcome 4: Lifetime IPV perpetration | Outcome 4 aOR: 0.90, (95%CI: NR) | ||||||||
| Santana, 2006 [50] | Male role attitudes scale (8 items) | Men ages 18–35 who are sexually active in the past 3 months, English and/or Spanish and receive services at clinics in Boston, USA (N = 283) | Continuous | Multivariate logistic regression | Physical or sexual IPV perpetration in the past year | aOR: 1.80* (1.10, 2.09) | Positive association | ||
| Shannon, 2012 [45] | Gender inequity norms scale (6 items) | IPV inclusive, sexual entitlement | Men ages 23 to 36 in Botswana and Swaziland (N = 999) | NR | Multivariate logistic regression | Rape perpetration | aOR: 2.19*, (1.22, 3.51) | Positive association | |
| Prather, 2012 [34] | Traditional-egalitarian sex roles scale (TESR) (20 item) | Male control over wealth | College students ages 18–25 in USA (N = 260; 77 men, 183 women, finding adjusts for gender) | Continuous | Multiple linear regression | Psychological IPV perpetration | Std Adj. beta: 0.25** (95%CI: NR); Respondent sex did not moderate relationship between sex role attitudes and perpetration (Std adj. beta:0.08) | Positive association | |
Notes: NR indicates not reported.
a This category is inclusive individual beliefs, attitudes and expectancies on social norms and roles considered appropriate for men and women
bScales are coded so that higher score represents less equitable beliefs, (rev): indicates reverse orientation of indicator response scale (higher score signifies more equitable views)
cWe report outcomes for the most adjusted or final statistical model using the following terminology: aOR = adjusted odds ratio; aRR = adjusted risk ratio; Adj beta = adjusted beta coefficient, Std Adj beta = standardized adjusted beta coefficient, r = correlation coefficient (unadjusted). Unless indicated the variance measure reported is the 95% confidence interval.
d For consistency across studies, indicator performance is summarized in the hypothesized direction (e.g., less equitable beliefs and greater likelihood of IPV perpetration), Inconsistent results noted when direction or level of significance varied by subgroup or outcome (if multiple reported)
eConstructed from GEM scale and Medical Research Council men’s health and relationship study.
*p<0.05
**p<0.001
† Marginal significance at p<0.10