Demographic health surveys (DHS)
This study utilises couples data from the DHS, spanning survey data collected from 2001 to 2015 in multiple countries. DHS surveys were implemented by respective national institutions and ORC Macro International Inc. with financial support from the US Agency for International Development. The DHS surveys are standardised across countries and years, at the individual and at the household (couples) level, with sample sizes between 5000 and 15,000 households. They are based on probabilistic samples originating from multi-stage cluster sampling and are stratified by rural and urban areas for different regions of the countries. The surveys are conducted on a sample of female respondents aged 15 to 49 years, and increasingly, men aged 15 to 59 years are being sampled and interviewed. Couples datasets merge data from women and their partners who live within the same household. Information on IPV has been collected since 2000 and the DHS programme has developed a standard module and methodology for the collection of data on domestic violence.
Up to 2017, 21 countries in sub-Saharan Africa have DHS surveys that contain information on experience with IPV: Burkina Faso, Cameroon, Congo Democratic Republic, Cote d’Ivore, Gabon, Gambia, Ghana, Kenya, Liberia, Malawi, Mali, Mozambique, Namibia, Nigeria, Rwanda, South Africa, Tanzania, Togo, Uganda, Zambia and Zimbabwe. For the countries in sub-Saharan Africa that implemented the IPV module, only one randomly selected woman per household was eligible to participate to maintain confidentiality and ensure the woman’s security when answering the question on the experience of domestic violence. The development of the domestic violence module was guided by available research on valid and reliable measurement of domestic violence and by World Health Organization (2001) guidelines on the safety and ethical aspects of collecting information on violence against women. The module asking about IPV is based on a modified version of the Conflict Tactics Scales [27]. It asks women if they ever experienced a series of behaviourally specific acts of physical or sexual violence by their current or most recent partner. Only women who have ever lived with a partner are selected to answer the questions about experience with IPV. Women who say yes to a particular item are then asked about the frequency of perpetration in the 12 months preceding the interview. More recent surveys also include questions on emotional abuse.
Interviewers are trained specially to handle the sensitive questions of domestic violence, and they are expected to follow a strict protocol to ensure privacy. Interviewers are instructed to check all the surroundings within hearing distance for the presence of others. Only children young enough to not understand the questions can be present. The interviews are not allowed to proceed if privacy is not ensured, and the interview is terminated if someone enters the interview area [27].
The ICF International Institutional Review Board (IRB) has reviewed and approved procedures and questionnaires for standard DHS surveys. Additionally, country-specific DHS survey protocols are reviewed by the ICF IRB and typically by an ethics committee in the host country. This secondary data analysis has also received ethical approval by the London School of Hygiene and Tropical Medicine ethical review board.
Measures
The Domestic violence module is based on a modified version of the Conflict Tactics Scale (CTS) that asks the respondents about their experience of specific acts of physical or sexual intimate partner violence. This allows for comparability across countries [27]. All ever-partnered or currently partnered women are asked if any partner ever did one of the following to them:
Physical violence
-
1.
Pushing, shaking, slapping, throwing something,
-
2.
Twisting an arm,
-
3.
Striking with a fist or something that could cause injury,
-
4.
Kicking or dragging,
-
5.
Attempting to strangle or burn,
-
6.
Threatening with a knife, gun, or other type of weapon, and attacking with a knife, gun, or other type of weapon.
Sexual violence
-
1.
Physically forcing intercourse or any other sexual acts,
-
2.
Forcing her to perform sexual acts with threats or in any other way.
In this paper, we use the indicator measuring women’s experience of physical and/or sexual violence among currently partnered women. Women were considered to have ‘experienced intimate partner violence in the past 12 months’ if they reported experiencing at least one physically or sexually violent act in the last year and ‘ever experienced intimate partner violence’ if they reported experiencing any act in their lifetime.
We measured economic empowerment using the following indicators: household wealth of the household the couple was residing in, women’s and their partners’ educational level, employment and earnings, individually as reported by the woman and her partner and in relation to each other, and income related decision-making. Household wealth is a standardised measure of the relative wealth of a household and is measured using the wealth index provided in the DHS. The index is based on household information regarding assets, type of flooring, water supply, electricity, and the ownership of durable goods. It does not include household members’ education or employment status and divides households into poor, middle and rich levels [28]. Education is captured through a categorical variable: no formal education, elementary, secondary and post-secondary education. Employment status is self-reported i.e. if the individual reports being employed, including in the 12 months prior to the interview. Relative income is measured by whether the woman earns more, the same or less than her husband. Decision-making on income is collected through the women’s survey, and based on the woman’s report on whether she decides on her own, with her partner or with someone else; if only her partner, or someone else, decides for her; or if she does not have any earnings of her own. The definition of a cluster as urban or rural is made according to the definition used in each country [29].
Analysis
Our analysis utilises merged couples data from 25 surveys of the DHS conducted in 15 sub-Saharan African countries where women and their partners participated in the domestic violence module of the DHS. The sample size is 70,993 women aged 15 to 49 years. We used weights to adjust for sampling and domestic violence survey participation.
To analyse the data, we explored descriptive statistics and possible associations between sexual and/or physical IPV in the last 12 months and the seven economic empowerment factors described above by conducting cross-tabulations and chi-square statistics to assess whether an association exists between the outcome and the explanatory variable. Next, we estimated odds ratios using multivariate logistic regression, adjusting for other economic factors in the model as well as age, urban–rural and marital status and country to assess whether the established associations were the true effect or mediated by other variables in the regression. Because these data are cross-sectional, none of the associations can be interpreted as suggesting causality. A p-value below 0.05 indicates statistical significance. All data were analysed using STATA 14. Sensitivity analyses were conducted running the regressions without controlling for country fixed effects as well.
Ethical approval for the secondary analysis was received from the London School of Hygiene and Tropical Medicine Ethical Approval Committee, approval number 14402.