Measuring social norms of intimate partner violence to exert control over wife agency, sexuality, and reproductive autonomy: an item response modelling of the IPV-ASRA scale

Gender based violence, inclusive of intimate partner violence (IPV) and reproductive coercion (RC), is a pervasive global problem. IPV, which can include physical, sexual, and emotional violence perpetrated by a romantic partner, is experienced by one in three women worldwide [57]. Prevalence of RC, when a male partner interferes with a woman or girl’s efforts to control her fertility via pregnancy coercion or birth control sabotage [29, 30], ranges from 8.4% in the US [6], 10.2% in rural Niger [47], and 18.5% in Cote d’Ivoire [28]. IPV and RC often co-occur, with women who experience IPV at a much greater risk of RC than those who have not experienced IPV [48, 61]. Both IPV and RC have been found to be associated with a variety of negative consequences for health and wellbeing, including child bearing at younger ages, high parity, unintended pregnancy, depression, and HIV [18, 38, 46, 57], and are driven in large part by male partner beliefs of dominance and entitlement over female partners [11, 14]—beliefs that are shaped and reinforced by social norms [31].

Reducing IPV and RC is a primary focus of the United Nations 2030 Sustainable Development Goal 5 on gender equity. To achieve that goal, shifting social norms accepting of IPV and RC is critical [53]. Social norms, understood through social norms theory and the theory of interdependent action [26, 49], are the collective, often unspoken rules about what is normal and appropriate behaviour within a group of people; they are reinforced through social rewards and sanctions. For example, in highly patriarchal contexts, gender norms, a cultural template for how men and women should each behave, often drive behavioral expectations of dominant masculinity and female passivity, creating a social context that reinforces men’s violence against women [8, 12]. There are distinct constructs inherent in social norms theory. Descriptive norms are a person’s perception of what people in their community do in a given situation. Injunctive norms are a person’s perception of whether or not people in their community approve or disapprove of a certain behavior, and are reinforced by concerns around possible social sanctions. Second-order beliefs are a person’s perception of others’ opinions or beliefs about certain behaviors [26]. While many studies aggregate individual-level beliefs as a proxy for social norms, aggregated individual beliefs fail to capture key components of social norms: social expectations of others in the community, reference groups, and social sanctions and approval [26]. Few studies have examined injunctive norms or second order beliefs specific to IPV, but those that do suggest that perceived social norms and peer behaviour are related to individual IPV behaviour and that IPV can potentially be prevented by changing the social context [35, 43, 44, 51]. Violence prevention researchers are clear that effective interventions need to target social and contextual factors, like social norms, yet these efforts are substantially limited by lack of effective social norms measurement tools [5, 16, 24].

Few validated measures of IPV- or RC-related social norms exist. While some scales assess social norms broadly related to IPV and RC (e.g., traditional gender norms), very few scales measure the perceived social acceptability of these behaviors specifically or demonstrate correlation with violence perpetration. One IPV social norms measure, the Partner Violence Norms Scale, assesses gender norms and appropriate responses of family members to a woman experiencing IPV, but this only contains one item that assesses social acceptability of IPV behaviour itself (second-order belief) [13]. Another IPV social norms scale, the Social Norms and Beliefs about Gender Based Violence Scale, was validated for humanitarian settings in South Sudan and Somalia and contains a 4-item subscale (of 30-items) on motivations of wife beating (e.g., showing love, a husband’s right, and discipline), but has not been shown to be related to IPV behaviour [39]. Both scales offer a helpful measurement tool for understanding social norms that are broadly related to IPV, yet more validated measures assessing social norms directly related to the acceptability of IPV are needed.

Most studies attempting to measure IPV social norms have used proxy measures that aggregate individual-level attitudes about IPV, primarily using the Attitudes about IPV (ATT-IPV) scale [23, 25, 55, 59]. The ATT-IPV assesses individual-level attitudes regarding the justifiability of wife beating for behaviours representing a spectrum of gender transgressions (e.g., arguing with her husband) [60], a measure that has been integrated into the core Demographic and Health Survey with some variation across 90 countries [15]. Recognizing the limitations of aggregated individual-level attitudes, one study adapted the ATT-IPV scale to measure social norms by asking participants about the number of people in their village that would agree with each of the statements in the original ATT-IPV scale [50]. This adaptation of the ATT-IPV items to social norms items was not assessed for reliability or validity as a scale but could be a useful measure of second order beliefs if shown to be valid.

The current study seeks to assess the reliability and validity of a new measure of the social acceptability of IPV when it is used to exert control over wife agency, sexuality, and reproductive autonomy (IPV-ASRA Social Norms scale). Using data from a population-based sample of husbands of adolescent wives in rural Niger, we assessed these social norms by adapting the introductory question stem of the items in the ATT-IPV scale to transform them into a scale of second order beliefs about situations in which IPV are perceived to be socially acceptable. This study provides evidence of reliability and validity of the scale and tests the ability of the scale to differentiate between husbands who have and have not perpetrated IPV and RC. This measure maximizes efficiency and is primed for use in other contexts, given that it builds on the commonly used ATT-IPV scale in the DHS questionnaire that is used in many international contexts, contains a concise number of items, and helps address the need for social norms measurement tools.

The sample contained 559 husband-wife dyads. The majority of participant households were culturally and linguistically Zarma (66%) (Table 2). At the third wave of data collection, husbands were on average 29.9 (range: 18–66) years old and wives were 20.4 (range: 16–22) years old. Most husbands had at least one child (range: 0–16 children) and one wife, however, 17% had more than one wife. Among wives, 9% reported ever having experienced IPV victimization by their current husband and 7% reported RC.

The IPV-ASRA Social Norms scale measures the latent construct of social norms regarding the perceived acceptability of IPV against wives to control her agency, sexuality, and reproductive autonomy. In all models, the scale demonstrated strong reliability, as well as internal structure and external validity. The items showed acceptable fit with the two-dimensional PCM, in which two subconstructs of IPV-ASRA social norms were represented: social acceptability of husband-perpetrated IPV if a wife is (1) not fulfilling her wifely sexual and reproductive duties, and (2) challenging her husband’s authority. Based on these findings, this brief 5-item IPV-ASRA Social Norms scale has strong potential for enhancing measurement of IPV social norms.

The IPV-ASRA Social Norms scale was associated with husbands’ perpetration of IPV against their wives, providing further evidence of validity and confirming the utility of this scale for understanding IPV behaviour. As social norms are understood to be a primary factor shaping patterns of IPV and RC behaviour within populations [24, 43], the IPV-ASRA Social Norms scale could be a critical tool for understanding contextual risk for IPV in a community and for evaluating the impact of programs intending to change IPV social norms. Specifically, Dimension 2 (Challenges husband authority) varied by IPV perpetration. This finding suggests that in this cultural and social context, the norms that sanction wives for behaviours that challenge her husband’s authority are more closely tied to the social norms that condone IPV, a finding that could be explored further as an opportunity for IPV prevention. It may be that wife behaviours that deviate from the norm of obedience to husbands are perceived as more threatening to the current gender norm structure and are therefore seen as more deserving of violent punishment from husbands to discipline this behaviour [14]. In contrast, social expectations of wives to bear children, by being sexually available to their husbands and fertile, may be perceived as less threatening to current social power structures and less linked to the norms that condone IPV. While fertility and procreation are highly valued in this cultural context and a very important part of social expectations of females [40], there may be less communal consensus on whether there are certain circumstances when it is acceptable for a wife to refuse sex or use family planning (e.g., if she already has had multiple or male children) or whether such a transgression warrants violence. Previous qualitative research in Sub-Saharan Africa has documented that wife beating is most acceptable for purposes of discipline, findings that are reflected by the stronger association between the challenging authority dimension of this scale and IPV [4, 33]. Future qualitative work could help shed additional light on the types of gender norm transgressions that are perceived to merit IPV-related punishment and the mechanisms shaping these norms within villages. Statisticians using this scale to understand how the latent construct of IPV-ASRA social norms relates to IPV behaviour will benefit from using a two-dimensional PCM.

This scale was not found to be associated with wife reports of husbands RC behavior. We suspect that this is likely due to the small number of husbands with wives reporting RC victimization and that large number of parameters in the PCM models, both of which reduce statistical power. This also could be a clear indication that the scale would benefit from more than one item specifically related to the social acceptability of reproductive autonomy that could be included in future iterations of the scale. Future research to develop an additional RC item that would complement this scale and be most appropriate in this cultural context is needed.

The reproductive autonomy item (item 1; family planning use without informing the husband) was an addition to the original ATT-IPV scale from which this IPV-ASRA Social Norms scale was developed. This item was the only item to be included in both dimensions of the scale, including the dimension that was associated with IPV behaviour (Dimension 1; Challenges husband authority). Our results demonstrate that the addition of this item is highly useful in understanding IPV-ASRA norms in this context and is likely an item worth including in subsequent use of this scale. Previous research in the region has identified covert use of family planning by women to be a strategy many young wives use to manage the conflict they may experience between their desire to delay pregnancy on the one hand and on the other, the strong social taboos against family planning use [47]. While commonly practiced, covert use of FP may be particularly risky in terms of potential husband perpetration of IPV, as evidenced by the way the perceived social norms accepting of violence in such a situation contributes to predicting IPV behaviour. Research from Niger and other settings has demonstrated a strong link between RC and IPV, stressing the importance of considering RC in understanding IPV [20, 38, 41]. Further social norm measurement development efforts would benefit from including RC-related norms and understanding how they interact, if at all, with IPV-related norms.

An additional key contribution of this scale is that it directly measures social acceptability of IPV behaviours and does so among men, those most likely to perpetrate these forms of violence. One previously identified IPV social norms scale, the Partner Violence Norms Scale, has shown an association with women who have experienced IPV victimization. That scale aims to measure the construct of traditional gender role expectations with only one item reflecting norms acceptable of IPV perpetration, a set of social norms more distally related to IPV. Additionally, the scale was assessed only among women and related to their IPV victimization, rather than among the men perpetrating violence. The IPV-ASRA Social Norms scale in the current study measures norms accepting of IPV perpetration to control wife agency, sexuality, and reproductive autonomy and was assessed among those whose behaviour is most relevant (i.e., those who perpetrated IPV), providing strong evidence of validity and utility.

While results support that the IPV-ASRA Social Norms scale is a strong measure with utility in IPV research, its primary limitation is that in this sample, it does not include enough items to capture the full continuum of the latent construct of IPV social norms; the scale contained primarily items that were severe or hard to endorse regarding perceived acceptance of IPV and lacks items that represent less severe perceptions of acceptance of IPV. The test information function graph suggests that the scale is best for populations of men with average and high perceived social acceptability of IPV (i.e., a location between about − 0.5 and 2 logits). Measure development research to expand this scale to cover more levels of the construct’s continuum would be useful to improve this scale in order to enable further differentiation of mens’ perceived social norms. Because the reproductive autonomy item was the “easiest” to endorse, the inclusion of more reproductive autonomy-related items might help expand coverage of the construct. Specifically, expanding the measure to include items representing RC social norms would be useful for understanding the norms supporting RC behavior among husbands, and their interaction with IPV social norms. This should be done by triangulating qualitative and quantitative data from this population to inform which additional items are most relevant. In the scale’s current form, the middle response option of “somewhat agree” may have been more “difficult” to endorse than a more neutral wording of this response option (e.g., “neither agree nor disagree”), which may have contributed to the skewed coverage of the construct. Further cognitive interviews with this population around interpretation of this three-option Likert scale is needed. The brevity of the current version of this scale, however, is a strength in studies where participant burden is already high or in epidemiological studies, where measures typically need to be limited in length, so the addition of a limited number of well-constructed items is recommended. Lastly, there was substantial loss-to-follow-up in this wave of data collection for husbands that may have led to selection bias in this sample. Moving forward, the generalizability of the findings for this scale will be strengthened as it is tested in more diverse, representative samples of men from this cultural context.

This IPV-ASRA Social Norms scale is a short, practical measure with strong reliability and validity evidence and is associated with men who perpetrate IPV. To date, very few measures of IPV social norms are available, and none, to our knowledge, have shown statistically significant associations with male IPV perpetration. This scale is concise and builds on a widely accepted and utilized measure of individual attitudes about justification for IPV (ATT-IPV) included in the DHS, and, with additional testing in other cultural contexts, could be a natural and useful addition to DHS-related efforts to understand the context of IPV. Enhancing current approximations of IPV social norms that simply aggregate individual attitudes, this scale directly measures social norms of IPV behaviour and could help elucidate pathways through which social norms may be impacting IPV behaviour. Moreover, as social norms are increasingly becoming the focus of IPV prevention efforts, the IPV-ASRA Social Norms scale could be used to examine areas of high need for social norms-focused prevention and to measure the impact of such efforts.