JUSTIFICATION OF WIFE BEATING IN RURAL BANGLADESH: A QUALITATIVE ANALYSIS OF GENDER DIFFERENCES IN RESPONSES TO SURVEY QUESTIONS

Abstract

Understanding attitudes about intimate partner violence (IPV) in cultural context is important for developing interventions to reduce IPV and its effects. This paper presents qualitative findings from research conducted in rural Bangladesh to understand men’s and women’s responses to attitudinal questions about IPV. Both men and women often responded as if the questions were about their personal behavior. A few women said that their opinion did not matter. Women’s responses were more sensitive than men's to contextual nuances in the questions, and men more often than women described their own attitudes as consistent with community norms.

Intimate partner violence (IPV) often occurs in the context of cultural ideologies that support it, and these ideologies influence what individuals believe. A contextualized understanding of attitudes about IPV is an important basis for developing strategies for prevention. National surveys such as the Demographic and Health Surveys (DHS) often collect data on attitudes about IPV. To date, the DHS have collected data on the prevalence of IPV in more than 25 countries and on attitudes about IPV against women in more than 50 countries. According to these surveys, large percentages of men and women in many countries report that wife beating is justified in various situations (Kishor & Subaiya, 2008). In 17 countries in Sub-Saharan Africa in which this attitudinal question has appeared in the DHS, 29% – 75% of women have reported that wife hitting or beating is justified, and the odds for women relative to men of reporting such views have varied widely, from less than one (0.91, 95% CI 0.83 – 0.99 in Lesotho) to more than five (5.51, 5.07 – 5.98 in Benin) (Uthman, Lawoko, & Moradi, 2010). Elsewhere, as many as 90% of women said wife-beating was justified for some reason (Yount & Li, 2009). In the Bangladesh 2004 DHS, 55% of men ages 15 – 49 said that a husband was justified in beating his wife in at least one of five scenarios (NIPORT, 2005). In the 2007 Bangladesh DHS, 36% of both men and women ages 15 – 49 said this (NIPORT, 2009).

This paper presents qualitative findings from a project that explored what men and women mean when they respond to structured attitudinal questions about IPV against women. The research aims to improve quantitative tools to understand women’s and men’s attitudes about IPV and their perceptions of norms about IPV in their communities. Specifically, this paper addresses three research questions:

1. What socio-cultural frames of reference influence the ways in which common survey questions regarding attitudes about IPV are understood and answered?

2. How consistent are responses to these questions when study participants are given the opportunity to discuss their answers and, when additional details are added to the questions, and how does the consistency of responses vary by gender of the respondent?

3. How likely are men and women to view their own attitudes as consistent with community norms?

SETTING

Reported levels of IPV in Bangladesh are high. Estimates of lifetime IPV from six rural surveys range from 32% to 72% of married women (Schuler et al. 1996; Steele et al. 1998; Khan et al. 2000; Koenig et al. 2003; Bates et al. 2004; BIDS, 2004). The research sites for this project were four of six villages where members of the research team for this study have been collecting data since 1991. Although not randomly selected, the villages and the districts in which they are located do not stand out within rural Bangladesh in terms of their socio-economic and demographic characteristics and access to services (Bates et al. 2004).

METHODS

The data for this analysis come from cognitive interviews (CIs ) and focus group discussions. Here, we discuss each source of data in turn. Cognitive interviewing is a technique designed to understand the mental processes that underlie responses to survey questions (Ericsson & Simon, 1980, 1984; Willis, 1999). We conducted CIs with women and men to expose the cognitive processes and cultural assumptions and associations influencing their understandings and responses to standard DHS questions about attitudes regarding IPV by husbands against their wives.

Three sets of CIs were conducted privately by same-sex interviewers with men 22 – 75 years of age and women ages 18 – 48 years. Our target age range was intended to be consistent with the DHS. The age range for women was intended to capture adult (>= 18 years) women of reproductive age and adult men (>=18 years of age). The lower bound of 18 years was established to abide by Emory’s IRB regarding the conduct of interviews with “adults” (>= 18 y). The first set of CIs (referred to below as "baseline CIs") was conducted with 27 women ages 19 – 48 years and 24 men ages 22 – 75 years. We began with a version of the basic DHS questions¹, as follows:

It is normal for a couple to have quarrels and disagreements. During these quarrels some husbands occasionally severely reprimand or even beat their wives. In your opinion, do you think a man would be justified (uchit²) to beat (piTh) his wife:

• If she neglects the children?

• If she argues with her husband?

• If she fails to provide food on time?

• If she visits her family or friend without her husband's permission?

Study participants were asked to relay their understandings of the overall question, words within the question, and the reasons for their response. They also were asked about their perceptions regarding prevailing norms in their community, and whether they would agree if others, such as elders or religious leaders, said wife beating was justified. Men were asked how they personally would feel if they had beaten their wife and women were asked how they personally would feel about being beaten in a particular situation. The interviewers asked pre-prepared and spontaneous questions to probe these matters. Near the end of the interview, the initial question was repeated and the participant was asked to repeat it in his or her own words.

Based on the results of the baseline CIs, a second set of CIs (referred to below as the "gender norm scenarios CIs") was conducted with 20 women ages 18 – 43 years and 23 men ages 25 – 70 years. These participants were asked new structured questions to explore how their responses might differ if the DHS-type questions added additional context qualifying each of the actions of the wife to which a husband responds with violence. Detailed scenarios were constructed based on data from the first set of CIs, which suggested that most study participants did in fact justify wife beating if they perceived the woman to be at fault. Pairs of detailed scenarios meant to tap into cultural definitions of women as blameless or at fault, depending on how well they adhered to prevailing gender norms about women’s behavior, were created for each of the DHS-type questions (the full set of questions is shown in the appendix.) For example, after study participants were asked if a man is justified to beat his wife if she visits her family or friend without her husband's permission, they were presented with the following detailed scenarios:

What if a woman is home alone? Someone comes to her house to tell her that her mother is very ill. She rushes to her parents’ house without asking her husband’s permission. Her husband beats her. Would he be justified in beating her?

And, in contrast,

What if the woman goes to her parents’ house just for fun without asking her husband’s permission, and her husband beats her? Would he be justified beating her?

For the gender norm scenarios CIs, in addition to asking about own attitudes and community norms, we asked about the attitudes of family members. At the end of the interview, we asked whether the participant thought anything should be done to stop husbands’ violence against their wives and, if so, what should be done.

The third set of CIs was conducted with 12 women ages 20 – 45 years. Here, we asked the same questions as in the gender norms scenarios CIs and also explored to whom study participants referred when they spoke about community attitudes, and who pays attention to and gossips about what others do. (These we refer to as the "supplemental gender norms scenarios CIs".)

Four focus group discussions with separate groups of men and women not included in the CIs also were conducted, using the same interview guide used for individuals in the gender norms scenarios CIs. The focus group discussions ranged in size from 5 to 8 men or women and included individuals from 20 to 65 years of age. The focus group discussions permitted exploration of men’s and women’s responses to these attitudinal questions about IPV against women in the context of a group setting.

All CIs and focus groups were conducted face to face by experienced, well-trained interviewers of the same sex as the study participant, usually in the person’s home but outside if necessary to ensure privacy. The interviewers followed WHO guidelines for the conduct of research on IPV (World Health Organization, 2001) and were trained to change the subject or terminate the interview if another person appeared during the interview. Informed consent was obtained from all participants and only one person was interviewed per household to ensure confidentiality. The CIs and focus groups were tape recorded, transcribed, and translated into English. Coding was done using the software program MAXQDA to organize the text into themes, such as misunderstandings of questions, reasons given to explain why men do or do not have the right to beat their wives, perceptions of blame, and other topics addressed in this paper. The CIs also were examined for patterns within them. Responses to questions that lent themselves to tabulation were tabulated to explore the absolute and relative frequency of specific patterns of responses across the cognitive interviews.

FINDINGS

Consistency of men’s and women’s responses

During the first set of cognitive interviews, the baseline CIs, over half (15/27) of the women changed their response to at least one of the DHS questions. Most of these 15 women changed their response to more than one question. The men, by contrast, gave more consistent responses; only 4 of the 24 men changed a response to a DHS question in the baseline CIs. This gender difference corroborates those found in other settings, in which women more often than men express ambivalent responses to attitudinal questions (Atkeson and Rapoport 2003; Rapoport 1982, 1985). In this context, this gender difference may have arisen because men in these communities are socialized to be much more sure of what they think is right. Also, women in these communities may feel ambivalent about disclosing personal beliefs that contradict what they believe to be the community norm. Alternatively, women may subscribe on the surface to perceived community norms but still feel wronged when they themselves are beaten, and so they waver in their responses.

Table 1, below, presents results based on the gender norms scenarios CIs. The individuals counted in the three center rows responded it depends. Here, we included those who said it depends when asked the unelaborated (DHS) questions in the beginning of the interview and then answered no and then yes in response to the more detailed scenario pairs tapping into prevailing gender norms. We also included those who first said yes and then changed to no and then yes, and those who said no and changed to no then yes. In other words, we counted as it depends cases where the participant said no, not justified in response to the second question in the triad, which presents a scenario in which the hypothetical woman conforms to local gender norms (where she is not “at fault” by local standards), and yes to the third question in the triad, where she violates local gender norms (and is therefore considered “at fault”). As discussed elsewhere (Schuler, Lenzi, & Yount, forthcoming), 129 of the total of 208 responses to the four questions were, either initially or ultimately, it depends (which the DHS does not code).

Table 1.

Percentages of women and men by response pattern in gender norm scenarios CIs (N=20 for both men and women^*).

Response combinations		Question triad** 1		Question triad 2		Question triad 3		Question triad 4		Average
		Women	Men	Women	Men	Women	Men	Women	Men	Women	Men
Answered Yes to baseline question, "at fault" gender norm scenario, and "not at fault" gender norm scenario	Y Y Y					5	5	10		3.8	1.3
Answered No to baseline question and both gender norm scenarios	N N N	30	70	5	45	14	60	5	20	13.8	48.8
Answered Yes, then No, then Yes	Y N Y	15	10	25	20	30	15	60	55	32.5	25
Answered No, then No, then Yes	N N Y	40	10	40	5	30	15	15	10	31.3	10
Answered It Depends then No, then Yes	D N Y	15	10	15	25	20	5	10	15	15	3.8
Answered It depends, then No, then No	D N N				5						1.3
Answered Yes, then No, then No	Y N N			15						3.8

^*Respondents who failed to respond to one or more scenarios were not included in this comparison.

^**DHS situation plus pair of gender norm scenario questions (shown in appendix).

As previously stated, in the baseline CIs, where the pairs of more detailed scenarios were not included, some study participants nonetheless changed their answers in the course of the interview, with women doing so much more often than men. Consistent with this finding, Table 1 shows that, compared to men, women in the gender norms scenarios CIs much more often changed their responses to the DHS question from no or yes to it depends when the situation referred to in the DHS question was followed by a pair of more detailed scenarios. Across the question triads, about 31% of women changed from no to it depends and 33% from yes to it depends (thus, 64% of the women changed to it depends when presented with the detailed scenario pairs). In contrast, 10% of men changed from no to it depends and 25% from yes to it depends across question triads (in total, 35% of the men changed their responses to it depends).

Across the question triads roughly half (49%) of the men, compared with only 14% of women, consistently said that beating was not justified. This suggests that men are either much less patriarchal than women in their attitudes or much more inclined to give socially desirable responses, or that what constitutes a socially desirable response varies by gender. (Based on years of research in these sites, we doubt that the men hold less patriarchal attitudes than the women.) It also suggests that women are much more sensitive to the nuances of a situation than men are in responding to such questions.

Men’s and Women’s Tendency to Personalize the Questions

Some study participants responded to the attitudinal questions with respect to a hypothetical husband and wife, as was intended. Others, especially men, responded as if the questions applied to them personally. In most of these cases the men indicated that they did not beat their wives. Eleven of 60 women (about 18%) personalized an answer to at least one of the questions, compared to 18 of 47 men (38%). Some personalized their answers to more than one question. Following are a few examples of personalized responses.

I: Do you think it is ever justified for a man to beat his wife?

R: No. There is no such problem between us. -40 year-old man, 12 years of education

Another male respondent replied to the initial, general, question as follows:

I: It is normal for a couple to have quarrels and disagreements. During these quarrels some husbands occasionally severely reprimand or even beat their wives. In your opinion, do you think a man is justified in beating his wife?

R: You have asked a very critical question. I have your answer, but [pause], you see, all men are not the same. Men’s temperament varies. If a man wants his wife to give him a glass of water after he returns from the field and she doesn’t care enough to do it, the man may lose his temper. In such a situation you too might lose your temper—am I not right? A man may beat his wife out of anger in such a situation.

I: So what is your opinion—is it justified to beat one’s wife?

R: No. It is not justified to beat one’s wife. Both of them have to understand this. This is my answer to your question. But what else can I do when I am tired and she disregards my request? Answer me. It is my question to you. Beating is not justified but you have to understand the situation. -young small businessman with a tenth grade education

Some of the men who said wife beating was not justified in a particular scenario said it was not justified because their own wives would never engage in the behaviour described. For example, one man said:

I have two babies. One is a boy and the other is a girl. As far as I’m concerned, my wife has never neglected her children, so it was and is out of the question for me to beat her for that reason. -42 year-old man with 9 years of education

As noted above, personalized answers were much less common among women. It is possible that the men, generally, were more sensitive than women to being judged by an outsider. So, some may have felt compelled to state (whether true or not) that they would not engage in wife beating. In other cases, as with the young businessman quoted above, they felt compelled to justify their violence against their wives. Women were perhaps more apt to justify IPV against women because they saw it as a norm in their communities, not because they believed they themselves deserved to be beaten. As shown in the next section, this interpretation is supported by an observed tendency among women to respond to the questions regarding their attitudes by describing what normally happens rather than whether they believe it is justified. Sometimes it was not clear whether they were referring to what happens in general as opposed to what happens in their own case.

Description versus Moral Judgment

Also common, and more common among women than among men, was a tendency to give descriptive answers about what men in their village would do in a particular situation, rather than saying that beating would be justified or not justified. Twenty-three of the 60 women (38%) gave descriptive responses to at least one of the scenario questions, compared with 7 of 47 men (15%)².

One man responded to the initial, generic question:

Obviously a husband would beat his wife if she does something wrong. Would he not beat her? -A 28 year-old man with no education

A woman was asked:

I: Is a man justified in beating his wife if she argues with him?

R: Well, wouldn’t a husband beat his wife if she argues with him? He definitely would. Some husbands have very high temper. They beat their wives if the wives even say a single word more than they are supposed to. -29 year-old woman with a fifth grade education and three children

In the following case, after the interviewer probed for a clearer answer to her question, the woman made a distinction between the rights men had in her village or society and what was, in her mind, right:

R: Every husband has the right to beat his wife.

I: Okay, but is a husband justified in beating his wife?

R: No it is not right. But what can you do when they beat you without realizing what they are doing? Nobody, even my parents, can do anything if my husband beats me. A husband finds fault with his wife so he beats her, and there is nothing you can do.

(Later) I: Is it right for a man to beat is wife if she neglects the children?

R: No, it is not right, but a husband has the right³ to beat his wife. -42 year-old woman with six children and two years of education

Thus, when asked whether wife beating was justified/right, some of the women evoked two sets of moral standards, the first relating to their own sense of right and wrong, and the second to hegemonic patriarchal local norms that they described as taking precedence in practice. We interpret the greater tendency of women, compared with men, to respond with descriptive answers rather than personal moral judgments as a function of their sense that their moral judgments are often irrelevant in the face of local norms that justify wife beating and the possible repercussions of challenging these norms.

Own versus Community Attitudes

In the baseline CIs, the generic question and the four DHS questions were repeated and the study participant was asked whether people in the community thought a man would be justified in beating his wife in these situations. Most men described the community’s attitudes as identical to their own while most women described their community as more apt to condone IPV than they themselves were. Across situations (Table 2), among the 25 women in the baseline CIs who answered all the questions about their individual attitude and their perceptions of community norms, women said men were justified in beating their wives in an average of about one situation, but said people in their community thought it was justified in an average of over two and one half situations—a difference of 1.5 situations. Of the 22 men who answered all the questions about individual attitudes and community norms, men said wife beating would be justified in just over one and a half situations in their own opinion and just over two situations according to the community, a difference of 0.7 situations (the it depends and mixed responses were counted as yes).

Table 2.

Average number of situations in which respondents said beating was justified or not justified (baseline CIs).

	Women (N=25*)			Men (N=22*)
	Individual	Community	Difference	Individual	Community	Difference
Beating is Not Justified	3.9	2.4	1.5	3.4	2.7	0.7
Beating is Justified	1.1	2.6	1.5	1.6	2.3	0.7

^*Respondents who failed to respond to one or more situations were not included in this comparison.

The situation in which wife beating was most often justified by women and men was "visits her family or friend without her husband’s permission." Twelve of the 25 women and 12 of the 22 of men initially said the man would be justified in beating his wife in this situation (including 2 women and 4 men who said it might be, depending on the specific circumstances); and 21 of the 25 women and 15 of the 22 men said others in the community thought so (some do, some don’t and it depends responses were counted as yes).

The finding that compared with the women, the men saw community attitudes as closer to their own has several possible interpretations, all of which may pertain to some extent. First, it could reflect a sense among both men and women that community norms mainly reflect men's views. It also could mean men are more conformist in their attitudes than women, and their own attitudes therefore tend to reflect their perception of community norms rather than vice versa. Finally, it may reflect a situation in which many women see IPV as wrong under most circumstances and perceive community norms to work against them. Other statements women made suggested this latter interpretation, for example:

Nobody judges a matter fairly. Rather, people in this community make judgments in favor of the husband’s family. -20 year-old woman with seven years of education

Nobody cares about whether I am happy or having a hard time. For example, here I live with my children in this house. Sometimes, I need to go to market to buy things. But people around take it as a bad thing. They look at me like I am a bad person. They complain about it to my husband. They urge my husband to do something about it. -35 year-old mother of six with no formal education

I: Do the people of your community think it is right for a husband to beat his wife if she goes to her parents’ house or to one of her friend’s houses without asking her husband’s permission?

R: People of the community don’t realize that a woman doesn’t go outside of her house for fun. She only goes out when it is very urgent and important… That’s why it wouldn’t be right for her husband to beat her, but tell me, how many people would understand this? Most people would think that she went out without asking her husband’s permission, so the husband did the right thing by beating her. -Woman age 29, 5^th grade education

CONCLUSIONS

In surveys worldwide, responses to attitudinal questions about IPV against women suggest a strong propensity to justify such violence. Notably, women often justify IPV to the same or an even greater extent than men do. The substantive and methodological reasons for this are poorly understood. This study provides rich data on the responses of rural Bangladeshi women and men to a series of structured attitudinal questions about men's perpetration of IPV against their wives. The data provide important insights about the gendered nature of response effects to these attitudinal survey questions and some of the most plausible reasons for the responses that women and men give.

Many of the responses given to the questions regarding their individual attitudes were personalized and/or descriptive. Men often responded as if the questions were about their own behavior rather than their general attitudes about wife beating; women often responded as if the questions were about community norms or their husband's behavior rather than their own attitudes. (When a woman said something like: "of course he will beat" in a particular situation, it was sometimes unclear whether she was talking about a man in general or her own husband.)

When asked first about their own attitudes and then about attitudes of people in their community, men tended to characterize the two as identical, whereas women tended to see others in their community as more likely to condone wife beating than they themselves were. This may mean that the men were more conformist than women were, describing their own attitudes as consonant with their perception of community norms. It may also reflect a tendency to identify community attitudes with men's attitudes. Either way, some women seemed to perceive that community norms worked against their own interests.

Many of the women in the study believed that IPV was a social norm in their villages, yet they themselves were ambivalent about whether it was fundamentally right or wrong. The much greater frequency of women (in contrast to men) changing their responses in the course of a CI suggests that some women may have felt uncomfortable disclosing their personal attitudes when they perceived such attitudes to be in conflict with community (patriarchal) norms about such violence. Moreover, women often seemed pessimistic regarding their power to change their circumstances, and believed their own opinions didn’t matter, which may have led them to articulate what they thought was the reality in their communities (that men can beat women), rather than what they really felt was justified. In a sense, these women may have been were giving what, to them, was the socially desirable response (the “correct” response as defined by local norms). There is some evidence that men, in contrast, tended to see “not justified” as the socially desirable response, assuming that this would be the “correct” response as defined by the outside world the interviewer was thought to represent.

When, following the basic DHS question, women were presented with the pairs of scenarios designed to reflect female behavior that conformed/did not conform to prevailing gender norms, they tended to say wife beating was not justified in the former case and justified in the latter. Whether they were merely describing local norms, as opposed to commenting on their fairness, was not always clear. In contrast, about half the men said that wife beating was not justified in response to the basic DHS questions and stuck with their original response when presented with the pair of scenarios.

The extent to which the men who responded that men are not justified in beating their wives actually meant it cannot be ascertained from this study (nor from the DHS). Nonetheless, the mere fact that they said it may provide an opening for intervention, a platform on which to build in developing interventions to undermine the norm of IPV by husbands against wives. The fact that most women did see wife beating as justified in cases where women appeared to challenge gender norms suggests a need for communication approaches that stress universal human rights principles, and interventions that raise questions about inegalitarian gender norms more broadly (not only in the context of IPV). The results of this study also suggest that surveys could elicit more accurate data regarding attitudes about IPV by developing detailed questions drawing on culture-specific gender norms.

Appendix

The pairs of questions were:

• 1a)What if a woman is over-burdened with work on a particular morning. Normally she keeps the children very clean, gives them lunch on time, and supervises their play. But today, it has been raining, and while she was working very hard to get her house work done, they went out to play in front of the house and got very dirty. She doesn’t have time to bathe them. Her husband comes home for lunch and sees how dirty they are. Would it be justified for him to beat her? Remember, I’m not asking you whether it’s justified to beat the children. I’m asking whether the husband is justified in beating his wife in this situation. Please give your own opinion.
• 1b)What if a woman regularly lets her children go around looking dirty. The husband has asked her several times before to keep them clean but she does not pay attention to what he asks. After telling her several times to keep the children clean, is he justified in beating her if she still doesn’t pay attention to what he says? What is your opinion?
• 2a)What if a man stays home out of laziness for several days, refusing to go out and work. His wife tells him they are running out of food and there is not enough money to buy food - would he please go out and work. He tells her to shut up. She argues with him. Then he beats her. Is the man justified in beating her for arguing with him?
• 2b)What if a woman often contradicts what her husband says, and argues with him. Would he be justified in beating her?
• 3a)What if a wife had a very busy morning doing the housework and tending to the children, and she is a little late in serving her husband’s meal. Normally she serves his meal on time. Would he be justified for him to beat her for being late? What is your opinion?
• 3b)What if the wife spent much of the morning gossiping with her friends and served her husband’s meal late. Would he be justified for him to beat her for being late?
• 4a)What if a wife is home alone. Someone comes to her house to tell her that her mother is very ill. She rushes to her parents’ house without asking her husband’s permission. Her husband beats her. Would he be justified in beating her?
• 4b)What if the woman went to her parents’ house just for fun without asking her husband’s permission, and her husband beats her. Would he be justified beating her?