South Asian Immigrant Men and Women and Conceptions of Partner Violence

Intimate Partner Violence (IPV) is a widespread public health issue affecting all social classes and ethnicities, and impacting both women and men. The World Health Organization (WHO) defines partner violence as “any behavior within an intimate relationship that causes physical, psychological or sexual harm to those in that relationship” [1]. IPV includes acts of physical aggression, psychological abuse, forced sexual contact or other controlling behaviors. Over the last three decades significant advances have been made in assessing IPV rates, risk factors and patterns, and in evaluating treatments. However, limited understanding exists about conceptual variations in defining IPV by ethnicity for both men and women.

Ethnic diversity is on the rise due to global migration [2] and is notable in regions with a history of migration and settlement, such as Canada and the United States (US). In Canada, the ethnic mosaic has been diversifying since 1967 when immigration policy was modified, establishing a point-system based on newcomers’ skills and education and removing preferential support for Europeans. Recent cohorts of migrants to Canada include large numbers from Asia and the Middle East. In 2006, Canadians of South Asian descent became the top visible minority group [3, 4], and in 2011 they accounted for 1.5 million residents [5]. South Asia includes countries like India, Pakistan, Bangladesh, Sri Lanka and Nepal. The South Asian (SA) community is also growing quickly in the US [6].

Despite their growing numbers in North America, limited research has been conducted with the SA community with respect to IPV and its conceptualization. Some small-scale studies identify IPV as a serious issue for the community [7‐10] and others report poor health outcomes for SA women with IPV experiences [11, 12]. Further, high rates of IPV are reported for women in Pakistan, Bangladesh and India with the lifetime prevalence ranging from 40 to 66 % [13‐15]. There is a strong need to advance scholarly knowledge about the experiences and perspectives of the SA community in relation to IPV.

Socio-cultural norms and values are likely to influence perspectives about IPV. Only a handful of studies in North America provide such insights. Klein et al. [16] compared Whites, African Americans, Latinas and Asians in a US national sample using vignettes about couple interactions. The authors found that Asian women were the least likely, while White women were the most likely, to categorize certain interactions as domestic violence, such as neighbors having a fight involving loud screaming, or a cousin shoving and slapping his wife during dinner [16]. Yick and Agbayani-Siewart [17] found that Chinese women (n = 15) in Southern California minimized psychological aggression compared to physical or sexual IPV. Likewise, a review of literature by Srinivasan et al. [18] noted that some SA women may not recognize certain acts and behaviors as abusive due to their familial obligations and culturally prescribed roles. In 2004, a Canadian study by Ahmad et al. examined the influence of patriarchal beliefs on SA women’s own perceptions of abuse by using a vignette. This telephone-survey in Toronto with Urdu, Hindi or English speaking SA women (n = 47) found that SA immigrant women with stronger patriarchal beliefs were less likely to see spousal violence as abuse, while only 17 % regarded forced sex by a husband as a possibly violent act [19]. In 2008, Mason et al. conducted eight focus groups with Sri Lankan Tamil women in Toronto (n = 68) and found that participants defined IPV broadly to include physical, sexual, emotional/psychological and financial abuse, consistent with the WHO definition [20]. While some work is emerging, it remains unclear what kinds of acts and behaviors are perceived as abusive within each type of IPV. Lack of such knowledge leads to gaps in and poor socio-cultural sensitivity across IPV related services, problems which have been well-recognized for decades [21].

We sought to address this gap by using an innovative Concept Mapping methodology. The primary objective was to explore the perspectives of the SA community, by gender, regarding behaviors that constitute IPV. The secondary objective was to examine similarities and differences between gendered SA and multi-ethnic perspectives.

Concept Mapping is a participant-engaged research method involving three distinct and sequential activities: Brainstorming, Sorting and Rating, and Interpretation [22, 23]. This is a structured process with a mix of qualitative and quantitative techniques that integrate participants’ input on a single topic of interest (i.e. perceptions on what constitutes IPV), and produces an interpretable pictorial view (i.e. map) of participant generated ideas and conceptualizations. This methodology places an emphasis on purposeful sampling in order to engage expert insights about the examined phenomenon. We provide below an overview of the larger study followed by details on the linked Interpretation sub-study with the SA sample. Ethical approval was obtained from the Research Ethics Board at Saint Michael's Hopsital in Toronto, Canada.

In the larger study, adult men and women from diverse ethnic and socioeconomic backgrounds were recruited if they had English language proficiency and resided in the Greater Toronto Area. Personal experience with partner violence was not a prerequisite. Specific over-sampling strategies were used to recruit self-identified ethnic minority individuals, such as members of the SA community. The recruitment strategies included placement of flyers in organizations serving the populations of interest, and snowball sampling. Interested participants called the provided phone number and eligible individuals were invited for Brainstorming and Sorting and Rating (S&R) activities, though retention from the first to the second activity was not mandatory. Both of these activities were available via online individual-sessions or in-person group sessions segregated by gender. In Brainstorming participants were asked to answer the following focal question: “What are the behaviors or attitudes that would make up the part of the relationship characterized by severe conflict, abuse, excessive control, neglect or even violence?” A total of 67 people participated (32 women, 28 men, and seven who did not specify) in the Brainstorming and 870 statements were collected. The research team consolidated the statements by removing duplicates and merging similar items. The final list comprises 71 statements (called “items” henceforth) and was used for the S&R activity. Seventy-one people (42 women and 29 men) participated in the S&R activity, primarily via online individual-sessions. Participants sorted the 71 items into conceptually similar groups that “made sense to them” and labeled their groups in accordance with the theme of the items. Participants rated the 71 items for importance in defining IPV and in prompting a victim to seek help; details of this component are provided in a separate article [24]. The research team entered the Sorting data into the Concept Mapping software that generated visual maps (more detail below). These maps formed the bases of the Interpretation group sessions, which were segregated by gender. The group sessions were attended by 20 multi-ethnic participants (9 women and 11 men) who were purposefully selected from the S&R sessions to ensure gender and ethnic representation. The multi-ethnic Interpretation sessions led to the creation of the General Multi-Ethnic Map (GMEM); further details of this are provided in another article [25]. All individuals provided informed written consent prior to participation in their first session. Those who continued to participate from Brainstorming to S&R and/or Interpretation were asked to reconfirm their consent. Participants were offered an honorarium of $40 for each activity in which they participated.

The reported Interpretation sub-study focuses on SA participants. The larger study sample included 24 SAs in the Brainstorming (12 women and 11 men) and 20 in the S&R (8 women and 12 men) activities. Using the Sorting SA data entered into the Concept Systems software, we first created an initial Cluster Map; this was not gender specific due to the small sample. The software employs statistical techniques of multidimensional scaling (MDS) and hierarchical cluster analysis [26]. MDS arranges points, representing items, on a spatial field based on the similarity matrix of participants’ sorted items. The results create a Point Map where items sorted together by more people appear closer to one another. The hierarchical cluster analysis then uses the means of Ward’s minimum variance method to partition the Point Map into non-overlapping clusters representing conceptual domains. From this a Cluster Map is created. The research team began with a 15 cluster solution and increased or decreased the number of clusters by one successively to identify a cluster solution where separation or merger of the clusters represented the data adequately and meaningfully. Through this review, a nine-cluster map was identified having a stress value of 0.3, which falls within the acceptable range of 0.21–0.37 provided by Kane and Trochim [23]. This SA initial-map (Fig. 1) was then presented in gender specific Interpretation sessions held with SA men (n = 6) and SA women (n = 5). The Interpretation sessions were led by gender concordant facilitators. The sessions comprised participants’ viewing of the map, ensuring all items were in the appropriate clusters, labeling of each cluster, and confirming the final number of clusters.

Eleven SA participants interpreted the SA initial map. The majority of them were first generation immigrants (i.e. ten out of eleven) who had been in Canada for ten or more years. Their ages ranged from 30 to 60 years; and all were married, all were employed, and all possessed at least a university-level education.

We report here the SA interpreted final maps by gender and compare these with the final GMEM map. The final SA maps were eight-cluster for SA men and nine-cluster for SA women. The final GMEM was a seven-cluster map. The list of cluster-content items for these final maps is presented in Table 1 for the SA and multi-ethnic subsets.

The findings generated by our exploratory study advance understanding about conceptualizations of aggressive behaviors as IPV by SA men and women. We found that the SA men and women’s conceptualizations vary compared to general multi-ethnic interpretations. To begin with, there were notable similarities across gender and ethnicity for the concepts of sexual abuse, victim retaliation and controlling behaviors, with an expansion of the controlling behavior domain by the SA group compared to the multi-ethnic group. SA men and women showed high similarity in their conceptualization of verbal/emotional abuse and did not distinguish between psychological and emotional abuse in the same way that the multi-ethnic group did. SA women were unique in their attention to the public versus the private nature of abuse. Further, the conceptualization of what comprised cultural abuse was much narrower and specific for SA women compared to the multi-ethnic group and did not emerge as an important domain for SA men. These findings are discussed in light of existing literature along with implications for further research and practice.

SA men’s and women’s perceptions of sexual abuse showed several clear differences, despite overarching congruence with each other and with the multi-ethnic sample. Previous work with SA communities also reveals certain unique perspectives [18, 19] within similarly defined general IPV domains [20]. SA women in our study included physical abuse as an act of sexual abuse. Their discussion on this item during the interpretation session demonstrated their desire to highlight gender based power imbalance as a root cause of physical abuse. While our study sample was small, several other studies point to the strong patriarchal values and rigid gender roles which normalize the subordination of women within the SA community [27‐29]. Some scholars call it “three obediences” of a woman to her father, to her brother, and to her husband [30]. Because physical and sexual abuses are the types of IPV that most frequently inform research, practice and policy, the SA women’s perception of physical abuse as a form of sexual abuse may have important implications for the identification of IPV in this community. Further community specific research could deepen our understanding to enhance socio-cultural sensitivity of available programs and services. For instance, adaptations of interactive theater reported by Yoshihama and Tolman [31] could be offered to the SA community with nuanced concepts of sexual and physical abuse among men and women.

SA participants also expanded the conceptualization of controlling behaviors. They extended the concept by including items on financial control, forced work, imposed religious beliefs, and blocked access to health care providers. Further research could help clarify whether this distinct pattern indicates ethnic differences in the significance given to the controlling behaviors or the frequency of exposure. Analysis of the General Social Survey data of 1999 for a sample of 25–49 year old women in current marital or cohabiting relationship points towards the latter. Although not stratified by ethnicity, a higher proportion of emotional spousal abuse was found in the recent immigrant women compared to the Canadian-born women [32]. Our study provides additional insights specific to the SA community. The SA men and women in our study conceptualized verbal/emotional abuse similarly to each other but differently from the multi-ethnic group in that they did not separate psychological and emotional abuse. Perhaps this stems from a more inter-related conceptualization of mind, body and soul in Asian healing systems, as identified in other research [33]. These findings challenge the assumption of homogeneity across ethnic cultures in defining partner violence, informing scholarly debate on what constitutes IPV.

Traditionally, researchers and clinicians have focused on assessing only physical and sexual violence. For example, reports on family violence by Statistics Canada provide rates of spousal abuse by counting only incidents of physical or sexual abuse [34, 35]. Several screening tools in clinical settings ignore the measurement of emotional abuse and controlling behaviors [36]. Some studies with mainstream populations argue for the need to assess emotional abuse [37‐40] but a handful of studies report the experiences and perspectives of ethnic minorities. Studies by Raj et al. [41] with SA women in Boston show that partners’ controlling acts in relation to the immigration status of women increased risk of IPV. Women in the Boston study also reported reduced sexual autonomy, increased risk of unwanted pregnancy and multiple abortions [11]. Likewise, in our study the domain clusters for emotional abuse and controlling behaviors in the SA concept maps were relatively large. These findings collectively highlight the need to assess emotional abuse and controlling aspects of couple interaction within the SA community. Asking about emotional abuse and controlling behaviors in healthcare settings could promote early detection and timely management of the risks associated with IPV. In light of delayed help-seeking reported by SA immigrant women with experiences of partner abuse, this could be particularly meaningful [42].

SA women in our study gave unique attention to the public versus the private nature of abuse, unlike SA men or the multi-ethnic group. The cultural values of familism and collectivism might have played a role in this conceptual distinction. Familism places emphasis on family relationships and, thus, matters concerning a family are considered ‘private’ [43]. Collectivism prioritizes the needs and goals of a collective (e.g. community) over an individual, and this leads to an “insider” and “outsider” group separation with a desire to protect the face [44, 45]. These cultural orientations have been previously associated with minimization of experiences of partner abuse by SA women and a delay in help-seeking [42, 46]. Our study suggests a possible link between these values and the definitions of types of abusive acts perpetrated by an intimate partner. However, it is unclear why SA participant men did not distinguish between private and public acts of partner abuse. Given that all of the SA participants who interpreted the maps were immigrants, employed and educated, it is possible that the women and the men had differing levels of attachment to these values. Others report variations in the rate of acculturation by gender and an expectation that women are often cultural ambassadors for transmitting values of the culture of origin to the next generation [47, 48]. This may explain why the SA women, unlike the men in our study, included a cluster on culture-based abuse with items pertaining to dowry, gender of newborns, difficulties in seeing a health care provider of opposite gender, and using cultural, religious and moral values to justify abusive behavior. While future research is needed to examine hypotheses generated by this study, the findings clearly highlight the need to measure and assess multiple aspects of abusive behavior for gender and ethnic inclusivity.

There were several limitations in our study. Although we used a broad recruitment strategy, it was not an easy task to recruit ethnically and socio-economically diverse participants due to the sensitive topic of research. Likewise, proficiency in the English language, required to undertake the study activities, might have introduced bias in reaching the population of interest. We could not examine the difference between the perceptions of IPV by the experience of IPV or participant socio-demographic characteristics (e.g. age, education, acculturation, income levels) due to the small sample size. The selection of articulate participants for map interpretation also limits the findings to the studied group. The sorting of the brainstormed statements might have caused participant fatigue, though we limited the number of statements. Finally, volunteer bias should warrant caution in the interpretation of the findings. Nevertheless, the concept maps generated by the SA participants and its comparison with the general multi-ethnic maps provide insights for future research and services.

In conclusion, SA men and women shared similar views about sexual abuse and victim retaliation, which also generally aligned with the views of multi-ethnic participants, although several unique aspects were identifiable. SA participants expanded the concept of controlling behaviors compared to their multi-ethnic counterparts. SA women viewed some aggressive behaviors and acts as cultural and demonstrated unique sensitivity towards the private versus public nature of abuses. Further research is needed with a larger and more diverse SA sample to examine the insights gained from our exploratory study.