Electing women to new Arab assemblies: The roles of gender ideology,Islam,and tribalism in Oman

Abstract

As Arab monarchies increasingly adopt and empower consultative assemblies, women’s representation varies markedly across countries. What leads citizens in these new electoral systems to vote for women? This study investigates the determinants of support for women’s representation using the first electoral survey ever conducted in Oman, prior to the October 2015 Majlis al Shura elections. It considers cross-nationally recognized factors – gender ideology and religion – and tribalism, a factor heretofore largely unexplored. Confirming prior studies, citizens with traditional gender ideology are much less supportive of women’s representation. Developing a simultaneous equations model, we show that religiosity and tribalism shape gender ideology. Unlike in Western countries, education is unassociated with attitudes, and there is no generational shift towards equality; younger men are less supportive of women’s representation than are older men. Increasing women’s representation requires not only increasing citizen demand for female leaders, but also changing informal tribal and formal electoral institutions.

Keywords

Women’s representation Arab politics tribalism

Introduction

Women’s underrepresentation in politics is a global phenomenon, one substantially limiting women’s influence in politics and society more generally. In the Arab region, where elected assemblies are relatively new and are gradually gaining in political power, the rates of women’s legislative representation are second lowest in the world – 18% (IPU, 2016). Still, there is substantial variation in women’s legislative representation even within the Arab Peninsula: from 22.5% in the United Arab Emirates (UAE) and 19.9% in Saudi Arabia, to 0.0% in Qatar (IPU, 2016).

In this article, we focus on Oman. This case presents something of a puzzle. On the one hand, Oman’s Majlis al Shura currently has the Peninsula’s second lowest rate of women’s representation, at 1.2%; only ten women have been elected over the course of the eight legislative elections in the country’s history (including incumbent reelections). On the other hand, Oman has been relatively progressive on other women’s issues and women have high levels of education. In 2003, Oman enacted universal suffrage for both men and women and granted women the right to compete for office.¹ It was ranked the second highest Arab country in women’s rights by the Thomson Reuters Foundation (Kehoe, 2013), and has a Gender Inequality Index value of 0.348, ranking 64 out of 149 countries in 2013 (United Nations Development Programme, 2014). A major reason for Oman’s dearth of female officeholders relative to its peninsular neighbors relates to electoral institutions – Saudi Arabia uses gender quotas, while the UAE’s council includes royal appointments. Still, Oman’s very low level of women’s representation indicates that citizens largely fail to support women in elected office. This low demand is striking given women’s high levels of education and prominence in other public roles.

What inhibits citizen support for female candidates? Relatively little work attempts to explain support for women’s representation specifically in the Arab region (but see Benstead et al., 2015; Bush and Jamal, 2015). Cross-national studies focus on the role of Islam (Inglehart and Norris, 2000; Norris and Inglehart, 2001). However, we argue that the informal institution of tribalism constitutes another significant barrier to women’s representation in Oman. Case studies of some Arabic countries report tribalism restricts women’s access to parliament (Al-Zyoud, 2009; Sabbagh, 2005). However, there are no individual-level studies testing the explanatory power of tribalism along with gender ideology and religion. We theorize that gender ideology mediates the effects of religion and tribalism on support for women in Majlis al Shura – see Figure 1.

Figure 1.

Theoretical model.

The Omani Consultative Council (Majlis al Shura) is a body with 85 members constituting the lower house of the Council of Oman. The Majlis was founded by Sultan Qaboos in 1990; elections were first held in 1991, and since 2003 elections are based on universal suffrage for men and women of age 21 and above. Provinces with a minimum population of 30,000 use a single non-transferable vote with a district magnitude of two, while the rest use plurality elections (Majlis al Shura, 2008). Though Oman’s Majlis does not have the full popular sovereignty or legislative authority of Western parliaments, since Arab Spring-inspired reforms in 2011 it is endowed in Basic Law with the power to propose legislation to the Sultan and with oversight over the state bureaucracy. Elections are recognized as transparent, though parties are not allowed to form. What leads citizens to vote for female candidates to this representative body?

In the next section, we develop a theory arguing that gender ideology, religion, and tribalism all affect support for women’s representation. We then describe our data, which come from the first electoral study ever conducted in Oman, during the campaign for the October 2015 elections. Our empirical analysis proceeds in three steps. First, we describe levels of support for women’s representation in Majlis al Shura, and assess the impact of support for women’s representation on actual voting for female candidates. Second, we proceed to multivariate analysis, developing a simultaneous equations model of both gender ideology and support for women’s representation. Third, we present results for male and female respondents separately, examining the extent to which men’s and women’s attitudes have different determinants.

We have several notable results. We discover that most Omanis express support for some representation of women, but very few support full parity. Support for women’s representation is very strongly related to gender ideology, which in turn is affected by both religiosity and tribalism. Support for women’s representation, in turn, is a strong determinant of both retrospective and prospective vote choice for female candidates, even though most of those who very strongly value women’s representation vote for men. In the conclusion, we argue that informal tribal institutions limit the supply of female candidates. Thus, improving women’s representation will require addressing both supply side and demand side factors.

Explaining women’s legislative representation in the Arab world

Scholars identify supply and demand explanations of women’s legislative representation cross-nationally (Randall, 1987). Supply factors prepare women to contest in elections, and influence whether they compete for office. Demand factors include the nature of political systems and electorates that increase the likelihood of female candidates winning. In this survey-based study, we of necessity focus on demand factors – in particular on what makes voters more or less likely to vote for women. In the conclusion, we return to discuss the supply-side, arguing that a complete understanding of women’s underrepresentation in the Arab region requires assessing both supply and demand.

To understand Arab voters’ support for female candidates, one might begin by examining voting behavior in the region more generally. While there is relatively little research, recent studies show that patronage, economic performance, and candidates’ personal qualities and campaigns motivate voter choices and turnout (Miguel et al., 2015; Yaghi and Antwi-Boateng, 2015). Women’s underrepresentation in Oman’s Majlis al Shura has not been examined in depth to date, though previous studies of women’s empowerment in Oman more generally touch on this issue in passing (Al-Hashmi et al., 2009; Al-Zedjali, 2009; Varghese, 2011).

Gender ideology

Cross-national studies find that gender ideology – that is, attitudes regarding women’s roles in society – strongly influences women’s representation across social and political systems (Norris and Inglehart, 2001; Paxton and Kunovich, 2003). We propose the following hypothesis:

H1: The less traditional a citizen’s gender ideology, the greater is support for women’s membership in Majlis al Shura.

However, we argue that this independent variable is so strongly related to our dependent variable as to be relatively uninformative. It is intuitive that attitudes toward women’s roles in the public sphere generally would affect attitudes toward women’s roles in the Omani Assembly, in .particular. The more interesting question, then, is the origins of gender ideology. Understanding gender ideology would enable stakeholders to develop better policies to alter traditional gender role attitudes. We argue that gender ideology mediates between a variety of cultural and structural factors and support for women’s incorporation in Majlis al Shura. Two such cultural factors are religion and tribalism.

Religion

In cross-national research, Islam is associated with low rates of women’s representation (Kenworthy and Malami, 1999). Furthermore, religious Omani men usually have reservations about women’s involvement in the public sphere, in general. Even if they allow their wives, sisters, or daughters to work, they encourage them to take female-dominated jobs such as school teachers and nurses. Thus, religiosity may influence support for women leaders indirectly, by influencing gender ideology.² We propose:

H2: The greater the religiosity, the less egalitarian will be gender ideology.

Still, a singular emphasis on Islam has been disputed. Some scholars point to women’s political presence in non-Arab Islamic countries, and in some Islamist parties (Sabbagh, 2005); others argue the culprit is local culture, combined with oil dependence (AbuKhalil, 1993; Ross, 2008). Al-Hashmi et al. (2009) found that 38% of Omanis strongly agreed that general attitudes toward women’s roles impeded women’s empowerment, while 26% strongly agreed that religion did. To date, however, no empirical research in the Middle East distinguishes among the roles of Islam and other informal institutions.

Tribalism

We argue that tribalism also affects attitudes toward women’s roles.³ Oman’s two main tribal blocs, the Qahtani and Adnani, have been competing for millennia. Though Sultan Qaboos has promoted national over tribal identity, tribalism remains a potent social force, and a source of honor, in Oman and across the Arab peninsula (Al-Farsi, 2013; Barber, 2007; Cooke, 2014). Tribalism strongly influences Majlis elections (Barber, 2007; see also Al-Zyoud, 2009 and Sabbagh, 2005 on Jordan and Yemen). Tribes and tribal blocs informally nominate candidates, sometimes forming agreements to circulate candidacies among tribes, and also mobilize voter turnout (Al-Farsi, 2013).

Tribal lineage and governance are highly patriarchal. Historically, a male tribal leader (sheikh) managed social, economic, political, and military issues in a sovereign territory (Torstrick and Faier, 2009). Omani women were unable to become leaders or judges, and still today they are not involved in formal discussions of tribal issues (Peterson, 1989). Married women keep their fathers’ tribal names, but their children belong to their husbands’ families; as a result, marriages are often arranged within tribes.

We expect that tribalism limits support for women’s participation in society and politics; as women do not have access to tribal forums for political discussion, they cannot campaign for female candidates. Rather, they are influenced indirectly by their fathers, brothers, and husbands to vote for male candidates. Thus, we hypothesize:

H3: The greater the strength of tribalism, the less egalitarian will be gender ideology.

Though we cannot test this proposition in the present study, tribalism affects women’s representation through supply as well as demand side mechanisms. Informal institutions create incentives shaping institutional outcomes (Helmke and Levitsky, 2004). In Western countries, political parties limit or facilitate women’s access to office through candidate selection and campaign support (Bjarnegård, 2013; Krook, 2009; Kunovich and Paxton, 2005). In Africa, ethnic-based patronage politics limits women’s access to cabinet positions (Arriola and Johnson, 2013). In the Arab region, tribes likewise are gatekeepers for women’s political power. Individual tribal leaders are unwilling to take the lead in nominating women for fear of stigma. The prejudicial effect is exacerbated by a formal institution: very low magnitude electoral districts. If a district can elect only one or two representatives, a tribe is unlikely to support a woman. By contrast, Bush and Gao (forthcoming) show that in the context of gender quotas, some Jordanian tribal leaders nominate women strategically.

Demographic factors

Though education is associated with gender egalitarianism and support for women politicians in many parts in the world (e.g. Hayes et al., 2000), tribalism may limit the impact of structural factors such as education, income, and employment. People at all socioeconomic levels value their tribes highly in Oman, and tribal loyalty is likely to overwhelm the effect of schooling on gender ideology. Therefore, we hypothesize:

H4: There is no relationship between level of education, income, or employment status, on the one hand, and gender ideology and support for women’s representation, on the other.

We also investigate the impact of demographic factors – sex, marital status, parenthood, age, place of residency, and ethnicity – on gender egalitarianism and attitudes towards women candidates. Studies from other regions show women have more egalitarian gender ideology and are more supportive of women candidates (e.g. Dolan, 1998; Hayes et al., 2000). Though Varghese (2011) found that Omani women in the Sohar region still prefer domesticity, we expect that group identity, interests, and a sense of shared fate lead Omani women to be more favorable than Omani men toward women’s involvement in public life. Moreover, a desire to maintain the status quo in which their identity group dominates will lead Omani men to resist women’s integration into the public sphere. We hypothesize:

H5: Men will have less egalitarian gender ideology and be less supportive of women’s membership in Majlis al Shura.

While studies of Western societies find generational shifts in gender attitudes, Norris and Inglehart (2001) find no such shift in developing societies. In Oman, in fact, older individuals may be more aware of traditions no longer practiced in which Omani women participated in village meetings with men to discuss tribal and political issues (Al-Farsi, 2013; Andriyanova, 2011). Younger individuals, by contrast, may be more affected by rising fundamentalist Islamic ideas in neighboring countries that oppose women’s roles in public life. We hypothesize:

H6. Older individuals have more egalitarian gender ideology and will be more supportive of women’s representation.

With respect to family composition, Al-Zedjali (2009) found that most Omani women in decision-making positions were married and had children, and were encouraged by their husbands throughout their careers. This could indicate a muted effect of family status on gender ideology and support for women’s representation. We hypothesize:

H7. There is no relationship between marital status and having children, on the one hand, and gender ideology or support for women’s representation, on the other.

Barber (2007) concluded urbanization has little effect on Omanis’ tribal ties. Still, all women elected to Majlis al Shura were from the capital. Tribal ties and cultural constraints are weaker in Muscat because mixed communities allow people to form new, non-tribal affiliations. Moreover, ethnic minority groups may be more supportive of women’s representation due both to distinctive historical women’s roles, and as a way to represent minority group interests.⁴ We hypothesize:

H8: Individuals who live in the capital and ethnic minorities have more egalitarian gender ideology and are more likely to support women’s representation.

Political information variables

We also expect that citizen access to political information affects gender ideology. First, we examine social influence. In many new electoral systems, citizens develop democratic norms through everyday social interactions (Finkel and Smith, 2011; Gibson, 2001). Those receiving political information from female discussants may develop more egalitarian gender ideology those receiving information only from men. This is in part because female discussants will tend, on average, themselves to have more egalitarian gender ideology. Also, female political discussants may serve as examples helping to counteract stereotypes of politics as a male domain. Thus, we propose:

H9: Individuals who discuss politics with women will have more egalitarian gender ideology and be more likely to support women’s representation. Individuals who discuss politics with men will have less egalitarian gender ideology and be less likely to support women’s representation.

Across the Middle East, women are less politically engaged than men (Coffé and Dilli, 2015). We suspect women’s lower political interest and knowledge reduce their support for women’s engagement in the public sphere. Likewise, we expect that citizens who are more politically engaged and knowledgeable about women’s underrepresentation will express greater support for women’s public roles (Sanbonmatsu, 2003). We hypothesize:

H10: Individuals with greater political interest and who are more knowledgeable about women’s current underrepresentation will have more egalitarian gender ideology and be more likely to support women’s representation.

Data and methods

This study uses primary data collected using a self-administered questionnaire the first author distributed to 500 Omani adults in the capital, Muscat, from mid-July to mid-August 2015, in advance of the October 2015 Majlis al Shura election. This is the first electoral study ever conducted in Oman. Muscat is a city for migrant labor, and about half of Muscat workers return to other areas on the weekends. This enabled us to contact both respondents from the capital city and those from other areas. Participants were recruited from 13 workplaces and institutions: nine public; three private; and one academic. The chief executive officer or the head of public relations in these organizations was contacted by providing them a letter explaining the study’s purpose and the targeted number of participants in each organization, and asking for permission to distribute the survey to employees/members. Three organizations granted permission immediately. In those organizations, the first author personally walked to the offices, made introductions, and requested employees’ participation. The remaining ten organizations distributed the survey to their employees/members, and the researcher returned later to collect these surveys. The sample utilized a gender quota of 250 men and 250 women. Generally, men were somewhat more cooperative than women. In most cases, the exact targeted number of men and fewer of women returned their surveys. Consequently, the survey was distributed to women only in some organizations to reach the quota of 250. Figures A1 and A2 and Tables A1 and A2 in the Supplementary Information present summary statistics, correlations among independent variables, and distributions of the dependent variables.

These data have some obvious limitations for studying Omani electoral behavior: the lack of a fully nationally representative sample; the underrepresentation of those outside the labor market; and that they were collected between two and three months prior to the election. The first and second limitations affect our ability to generalize to the broader Omani population. The last limitation affects our understanding of actual vote choices. The election campaign was relatively short, and in July/August many people were not yet fully aware of the candidates or certain whether and for whom they would vote. Nonetheless, despite limitations, Oman’s first electoral study provides an important opportunity to understand vote choice and the structure of public opinion in a campaign for a new assembly of the Arab region.

Dependent variables

Our theory suggests the relationship between tribalism and religiosity, on the one hand, and support for women’s representation, on the other, is mediated by gender ideology. Thus, our model has two dependent variables. Support for women’s representation is the average of standardized responses to three questions (Cronbach’s alpha = 0.83):

“Out of the 84 members of Majlis al Shura, what percentage would you prefer to be women?” Eight categorical responses were coded: None; 1–10% (1–8 women); 11–20% (9–16 women); 21–30% (17–25 women); 31–40% (26–33 women); 41–49% (34–41 women); 50% (42 women); More than 50% (more than 42 women).

“How important do you think it is to have women as members in Majlis al Shura?” Responses ranged from “Very important” (1) to “Not important at all” (5).

“How important do you think it is for women to run for office?” Responses followed the same Likert-type scale.

Since these three items have different ranges, they are each standardized on a 0–1 scale before we calculate their mean. The second and third items are reverse-coded so that higher values represent greater support for women’s representation.

Gender ideology captures attitudes towards women’s role in society. It is based on agreement or disagreement on a five-point Likert-type scale with three statements from the World Values Survey: “On the whole, men make better political leaders than women do”; “A university education is more important for a boy than for a girl”; and “When jobs are scarce, men should have more rights to a job than women.” Responses are recoded on a 0 to 1 scale, with higher values representing more egalitarian gender ideology; the index is the mean of responses and ranges from 0 to 1 (Cronbach’s alpha = 0.65).

We model support for women’s representation and gender ideology using simultaneous equations models, which assume the dependent variables are approximately normally distributed. As both variables are indices based on means of responses to several questions, they deviate from perfectly normal distributions. Nonetheless, both variables have measures of skewness close to 0.0 and kurtosis values close to 3.0, which would be found in a normal distribution. Under such circumstances, use of models such as ordinary least squares (OLS) developed for normally distributed dependent variables is justified (see Supplemental Information for the histograms of the dependent variables).

Is support for women’s representation tied to actual voting behavior? We also examine the downstream consequences of this variable for vote choice. Respondents were asked retrospectively, “Have you ever voted for a woman?” Response options were “No” (selected by 240 respondents), “Yes” (27 respondents), and “I have never voted” (233 respondents). The questionnaire also asked about prospective vote intentions: “If you have decided whom to vote for, is this person a woman?”; response options were “Yes” (n = 18), “No” (n = 155), and “there is no woman running in my district” (n = 49). Self-reported vote intention should be treated with caution because the survey was conducted relatively early in the campaign. Though 173 reported their chosen candidate’s gender, only 115 previously said that they were certain of their intended vote choice. In addition, 60.5% of respondents were uncertain how many female candidates were running in their district.

Independent variables

Independent variables relate to public opinion, political information, and demographics. Beyond gender ideology, there are two key public opinion variables. Religiosity is measured using two statements on a five-point Likert-type agree-disagree scale. Religion2 is “It is acceptable for me to listen to music,” and Religion3 is “It is acceptable for me to take personal loans from banks.” Religion3 is adapted from the Arab Barometer (Arab Barometer, 2013); Religion2 was added by these authors, as listening to most forms of music is forbidden in Islam. The questions are recoded on a 0 to 1 scale, with higher scores indicating greater religiosity. The index is their mean (Cronbach’s alpha = 0.67).⁵

Tribalism is an index calculated as the mean of responses to six statements developed by the authors (Cronbach’s alpha = 0.75). First, respondents were asked to what extent tribal nomination was important for their vote choice, on a four-point scale running from “Very Important” (1) to “Not Important” (4); this item is reverse-coded on a 0 to 1 scale so that higher scores indicate greater tribalism. The other five items use a five-point Likert-type scale, and are again reverse-coded on a 0 to 1 scale, with higher scores indicating greater tribalism.

Tribal1. “It is important for me to participate in my tribal events like attending ceremonies.”

Tribal2. “It is important to keep in contact with my tribal members.”

Tribal3. “It is important for me to be aware of my tribe’s issues.”

Tribal4. “Tribal compatibility is important for me in deciding whom to marry.”

Tribal5. “Tribal affiliation is an important part of my identity.”

Other variables include political interest and political knowledge of the actual level of women’s representation in the council. Political interest is the average of responses to two questions, regarding interest in Majlis al Shura elections, and following news about the elections; both are measured on a four-point Likert-type scale and reverse coded so higher scores indicate greater interest. Gender-related political knowledge is measured by asking participants to write the current number of women members on the council (just one) or to mark “I don’t know.” The variable is a dichotomous indicator for answering correctly.

Two dichotomous variables for interpersonal social influence distinguish between the influences of males’ and females’ opinions. Participants were asked whether they consider various people’s opinions when selecting a political candidate: father; mother; brother; sister; husband; wife; male friend; female friend; male colleague; female colleague; tribal leader; and religious leader. Male influence and female influence are dichotomous variables for those reporting any influence from the respective gender.

Demographic variables include sex, age, marital status, children, permanent place of residence, level of education, current employment status, and monthly income. Male, has children, ethnic minority, married, resident of capital, and employed are all dichotomous. Level of education and age are categorical variables on five-point scales, and monthly income is on a six-point scale; these three variables are recoded to run from 0 to 1.

Analysis and findings

Levels of support for women’s representation and voting for women candidates

To what extent do Omanis support women’s representation? In Figure 2, we present means and 95% confidence intervals for men’s and women’s responses to the first question in the index of the key dependent variable: “Out of the 84 members of Majlis al Shura, what percentage would you prefer to be women?” Almost all respondents – 85% of men and 97% of women – believe there should be at least some women representatives. The modal response among both men and women is that women should comprise between 1% and 10% of members. However, over 60% of women and over 40% of men say that they would prefer 20% or more of members to be women. At the high end, 8% of men and 12% of women would prefer women to comprise 50% or more of representatives.

Figure 2.

Preferred level of women’s representation, by the respondent’s sex.

In Figure 3, we present means and 95% confidence intervals for men’s and women’s responses to the second and third questions in the index of support for women’s representation, regarding the importance of having women as members of Majlis al Shura and of having women as candidates. As above, women are on average more supportive of women’s representation; 44% of men and 65% of women say it is “important” or “very important” to have women representatives, while 45% of men and 60% of women say it is “important” or “very important” for women to compete for office.

Figure 3.

Perceived importance of women’s representation, by respondent sex.

These three variables are used to create an index of support for women’s representation throughout the remaining analysis. But do attitudes matter at the voting booth? That is, are those who say they more strongly support women’s representation actually more likely to vote for female candidates? In Figure 4, we examine the relationship between support for women’s representation and self-reported voting for female candidates. The left pane is based on vote intention in the upcoming election; analysis is limited to the 173 people who knew of a woman running in their district and who reported their intended candidate’s gender. Of these, 10.4% had chosen a woman. Analysis in the right pane deals with voting in prior elections. It is limited to the 267 respondents who said they had ever voted, 5.4% of whom reported having voted for a woman. Despite differences in the sample compositions and sizes, results are remarkably consistent across the two analyses; support for women’s representation is strongly associated with both prospective and retrospective vote choice. Nonetheless, even among those who most strongly support women’s representation, fewer than a third are predicted to vote for a woman. We return to this point in the conclusion, where we discuss the implications for women’s representation in Oman and the Arab region.

Figure 4.

The impact of support for women’s representation on voting for female candidates

Determinants of support for women’s representation

What affects support for women’s representation? As a first approach, in Table 1 we assess the relationship between gender and the two dependent variables, as well as key independent variables. All gender gaps presented are statistically significant. Because variables are standardized on a 0–1 scale, gender gaps in various domains can be compared. Women have much more egalitarian gender ideology and higher support for women’s representation. By contrast, men have higher levels of tribalism and religiosity. Each gender receives most political influence from discussants of the same gender, though gender homophily is much stronger among men than among women. Men are also more interested in politics and more knowledgeable about women’s current representation. More precisely, 82.0% of women and 68.4% of men do not attempt the gender-related political knowledge question. Among those who respond, the median man and woman both get the right answer (that there is one woman on the council). On the high end, men overestimate to a somewhat greater extent than women; women’s mean response is 1.5 and men’s is 2.7.

Table 1.

Mean gender difference in key variables.

	Mean for men	Mean for women	Gender gap (men–women)	Significance test (p-value)
Support for women’s representation	0.467	0.604	−0.137	0.000
Gender ideology	0.348	0.544	−0.197	0.000
Religiosity	0.637	0.574	0.063	0.007
Tribalism	0.642	0.587	0.055	0.000
Male influence	0.236	0.164	0.072	0.044
Female influence	0.072	0.196	−0.124	0.000
Political interest	0.522	0.404	0.118	0.000
Political knowledge	0.196	0.112	0.084	0.009

Note: All variables are coded on a 0–1 scale.

We proceed to multivariate analysis of determinants of gender ideology and support for women’s representation. The following simultaneous equations model was developed:

\begin{array}{l} Support = f (gender ideology + sex + age + marital status + children + permanent \\ place of residence + level of education + monthly income + \\ employment status + ethnicity + tribal nomination + male influence + \\ female influence + political interest + political knowledge) \end{array}

\begin{array}{l} Gender Ideology = f (tribal participation + tribal identity + religiosity + sex + age + marital status \\ + children + permanent place of residence + level of education \\ + monthly income + employment status + ethnicity + tribal nomination + \\ male influence + female influence + political interest + political knowledge) \end{array}

The Supplementary Information presents coefficients from bivariate OLS regressions for each dependent–independent variable combination, as well as multivariate OLS models incorporating all possible independent variables. Analysis uses structural equation modeling (SEM) in Stata, which simultaneously calculates coefficients for direct and indirect relationships. SEM is a powerful tool to examine causality within a theoretical model; indices suggest the models are a good fit to the data. The Chi-squared statistic for the estimated versus saturated model is 1.240 (p = 0.538). Other fit indices also indicate good fit: the root mean square error of approximation is essentially zero (0.000), the comparative fit index is 1.000, and the coefficient of determination is 0.312. Table 2 presents standardized coefficients and standard errors.

Table 2.

Simultaneous equation models of gender ideology and support for women’s representation.

	Gender ideology		Support for women’s representation
	Beta coefficient	Standard error	Beta coefficient	Standard error
Gender ideology			0.504*	0.047
Religiosity	−0.090*	0.037
Tribalism	−0.078^	0.058
Male sex	−0.408*	0.021	−0.055	0.023
Level of education	−0.018	0.011	−0.023	0.011
Monthly income	0.116*	0.010	−0.062	0.01
Employment	−0.015	0.029	−0.012	0.03
Marital status	−0.070	0.024	−0.068	0.025
Children	0.002	0.022	0.052	0.023
Permanent residence in capital	0.072^	0.020	0.052	0.021
Ethnicity	0.093*	0.027	−0.004	0.028
Age	0.039	0.056	0.122*	0.057
Male influence	0.003	0.028	−0.027	0.029
Female influence	0.029	0.033	0.097*	0.034
Political interest	0.076^	0.039	0.133*	0.039
Political knowledge	0.103*	0.028	−0.028	0.029
Number of observations	466

p < 0.05; ^ p < 0.10.

The first key hypothesis (H1) is that those with more egalitarian gender ideology are more supportive of women’s representation, in line with many cross-national studies. Indeed, the analysis indicates that in Oman, gender ideology strongly and significantly predicts support for women’s representation in Majlis al Shura. In addition, consistent with H2 and H3, religiosity and tribalism are both significantly related to gender ideology. As seen in the Supplementary Information, neither significantly predicts support for women’s representation after controlling for gender ideology.

Many demographic and structural variables are incorporated in both equations. Consistent with H4 and many studies in other countries, men hold less egalitarian gender ideology than do women. The standardized beta coefficient for sex indicates that this variable is by far the strongest determinant of gender ideology. However, after accounting for gender ideology, sex does not significantly predict support for women’s representation. Age also matters; consistent with H6, older individuals are more supportive of women’s representation, though age is not associated with gender ideology.

Other demographic factors have weaker impact. Cross-national studies show that structural factors fail to explain country-level differences in women’s representation (Norris and Inglehart, 2001; Paxton and Kunovich, 2003). Consistent with such studies and with H5, neither education nor employment status explains gender ideology and support for women’s representation. However, monthly income is significantly associated with more egalitarian gender ideology. While the role of income should be explored in future studies, we speculate that income matters because wealthy households may tend to have working women whose income contributes to household consumption capacity. As expected, permanent place of residence is significantly related to gender ideology, but ethnic minority status, marriage, and having children do not matter for attitudes toward women’s role in society or toward women’s representation.

Finally, political information variables matter. Consistent with H9, discussing politics with women boosts support for women’s representation, but the sex of discussants is not related to gender ideology. Gender-related political knowledge is significantly associated with gender ideology, but is not significant in the model of support for women’s representation. Political interest boosts gender ideology, though not significantly; however, it is significantly directly related to support for women’s representation.

Next, we perform separate analyses for men and women. Due to gender differences in socialization and life experiences, many of the independent variables may have different effects on women’s and men’s gender ideology and support for women’s representation. For instance, marriage and parenthood may cause women’s attitudes to change in different ways from men. In addition, if young men are more exposed to fundamentalist versions of Islam than young women, and if ethnic minorities socialize men and women differently, age and minority status may matter for one gender but not the other.

Table 3 presents the first-stage equation of the SEM for both genders. Religiosity is significantly related to gender ideology among both men and women. The coefficients for tribalism and income fall below standard levels of statistical significance for both genders due to small sample sizes, yet coefficients are similar in magnitude to those from the combined model. By contrast, the effects of family composition, ethnicity, geography, and political information appear to vary by gender. Having children leads to more egalitarian gender ideology among men, yet both marriage and children lead to inegalitarian gender ideology among women. Residing in the capital is associated with egalitarian gender ideology among men, while ethnic minority women are substantially more egalitarian in gender ideology than are majority group women. Finally, political interest is associated with egalitarian gender ideology among men, and political knowledge among women.

Table 3.

Determinants of gender ideology (first stage of structural equation modeling), men and women.

	Men		Women
	Beta coefficient	Standard error	Beta coefficient	Standard error
Religiosity	−0.101^	0.052	−0.107^	0.050
Tribalism	−0.082	0.082	−0.074	0.081
Level of education	0.050	0.015	−0.068	0.016
Monthly income	0.112	0.016	0.108	0.013
Employment	0.046	0.046	−0.021	0.039
Marital status	−0.024	0.040	−0.150*	0.030
Children	0.173*	0.033	−0.166*	0.029
Permanent residence in capital	0.125^	0.029	0.012	0.027
Ethnicity	0.023	0.042	0.192*	0.034
Age	0.177*	0.076	−0.115	0.085
Male influence	−0.004	0.035	0.061	0.049
Female influence	0.037	0.057	0.027	0.045
Political interest	0.126^	0.052	0.014	0.055
Political knowledge	0.085	0.036	0.153*	0.042
Number of observations	234		232

p < 0.05; ^p < 0.10.

Turning to the second stage equations presented in Table 4, we find that gender ideology is very strongly related to support for women’s representation among both sexes. Beyond this relationship, age matters only among men. By contrast, having children appears to boost support for women’s representation only among women, after controlling for the negative relationship between having children and gender ideology. Future work should focus on teasing out this complex pattern of relationships. Last, political interest significantly boosts support for women’s representation among women.

Table 4.

Determinants of support for women’s representation (second stage of structural equation modeling), men and women.

	Men		Women
	Beta coefficient	Standard error	Beta coefficient	Standard error
Gender ideology	0.499*	0.069	0.435*	0.066
Level of education	−0.052	0.016	−0.004	0.017
Monthly income	−0.015	0.017	−0.145*	0.013
Employment	0.042	0.048	−0.085	0.039
Marital status	−0.060	0.042	−0.096	0.030
Children	0.016	0.036	0.104^	0.029
Permanent residence in capital	0.074	0.030	0.039	0.028
Ethnicity	0.007	0.044	−0.003	0.035
Age	0.121^	0.081	0.100	0.087
Male influence	−0.012	0.037	−0.026	0.050
Female influence	0.084	0.060	0.098	0.045
Political interest	0.078	0.054	0.206*	0.055
Political knowledge	−0.012	0.038	−0.048	0.043
Number of observations	234		232

p < 0.05; ^ p < 0.10.

Conclusions

This study investigated the demand-side determinants of supporting women’s representation in Majlis al Shura in Oman, using survey data from 500 Omani adults in the capital. It incorporated many cross-nationally recognized factors, including gender ideology, religion, demographics, structural factors, and social networks. We also argued the importance of considering tribalism – a factor heretofore largely unexplored – in the Arab context.

The analysis confirmed the importance in Oman of two main findings from previous cross-national studies: ideological beliefs regarding women’s roles strongly determine support for women’s representation; and religiosity leads to traditional attitudes toward women’s roles. At the same time, we showed that tribalism constitutes a second informal institution shaping attitudes toward women’s roles in Oman. Overall, these three attitudes constitute stronger explanations of support for women’s legislative representation than social/structural factors such as education or income.

The Omani economy, educational system, and bureaucratic systems have modernized dramatically in the past four decades. However, younger and older generations alike hold traditional attitudes, confirming Norris and Inglehart’s (2001) findings about developing societies. Still, the positive and significant effect of age on support for women’s representation in Majlis al Shura is striking – the older an individual, the greater the support for electing women. This effect is especially pronounced among men. Older men may be more aware of older tribal traditions in which women were involved in village discussions. By contrast, younger men may be more strongly influenced by recent global movements to promote fundamentalist interpretations of Islam.

We also find sizable gender gaps. Omani women have more egalitarian gender ideology and are more supportive of women’s representation than Omani men. However, Omani women are less interested in politics and less aware of the actual number of women in Majlis al Shura than men. This suggests the “gender affinity” effect (e.g. Dolan, 1998) may have only limited impact in Omani society. While Omani women are in principle more supportive of female candidates than are men, their lower levels of political interest may make them less likely to vote, and gender issues appear not to be particularly salient for them.

While this study showed that citizens’ attitudes affect representational outcomes, it also demonstrated the limitations of an exclusive focus on the demand side. First, attitudes may resist change (see, for instance, Alesina et al., 2013; Dilli et al., 2015). Economic development during the past 45 years has not brought massive cultural changes, despite the predictions of modernization theory. Although attitudes are not bedrock and they do change over time, the strong association between gender attitudes, religion, and tribalism in Oman may lead to slower evolution. Second, even when attitudes change, they will not be sufficient to produce equal representation. Even among citizens who very strongly supported women’s representation, fewer than a third had ever voted for a woman, or intended to vote for a woman in the upcoming election. This is likely due to the very low number of women running for office, and especially of female candidates with tribal support. As we argued, the informal institution of tribalism restricts qualified female candidates’ abilities to run for office and to receive effective support. Formal institutions also matter. Introducing a temporary gender quota might help increase the social acceptability of women in the council, on the one hand, and motivate women to contest elections, on the other hand (see Jalalzai and Krook, 2010). An increase in district magnitude could also encourage tribal leaders to nominate women. However, these steps must be accompanied by programs to shape citizen-level attitudes, including civic education on the importance of women’s legislative representation, and programs providing women with the necessary skills and experiences to run for office. Ultimately, women’s equal representation in the Arab peninsula will require attention to both supply side and demand side barriers.

Footnotes

Acknowledgements

We are indebted to the Government of Oman, and to David Anderson, Stephen Graham Sapp, and three anonymous reviewers for their helpful feedback.

Funding

This research was supported by the Government of Oman, which financed the first author’s graduate study.

Supplementary material

Supplementary material is available for this article online.

Notes

Author biographies

Ahlam Khalfan Al Subhi is an employee at the Ministry of Higher Education of Oman. She holds a Master’s degree in Political Science from Iowa State University.

Amy Erica Smith is Assistant Professor of Political Science at Iowa State University. During the 2016–2017 academic year, she is a Visiting Fellow at the Kellogg Institute for International Studies, University of Notre Dame, Indiana.

References

Abukhalil

As’Ad

(1993) Toward the Study of Women and Politics in the Arab World: The debate and the reality. Feminist Issues 13(1): 3–22.

Al-Farsi

Sulaiman H

(2013) Democracy and Youth in the Middle East: Islam, Tribalism and the Rentier State in Oman. London: I.B. Tauris.

Al-Hashmi

Sultan

Al-Musalami

Sheikha

Hassan

Sameer Ibrahim

(2009) Social Participation of Omani Women [in Arabic]. (Unpublished research.) Muscat: Ministry of Social Development.

Al-Zedjali

Khadija

(2009) Empowering Omani Woman and the Societal Challenges [in Arabic]. (Unpublished research.) Muscat: Ministry of Social Development.

Al-Zyoud

Ismail

(2009) The Tribe and Elections in Jordan [in Arabic]. Jordan: Dar Jalees Al Zaman.

Alesina

Alberto F

Giuliano

Paola

Nunn

Nathan

(2013) On the Origins of Gender Roles: Women and the plough. Quarterly Journal of Economics 128(2): 469–530.

Andriyanova

Olga

(2011) Some Observations on Women in Omani Sources. Proceedings of the Seminar for Arabian Studies 41. Archaeopress, 1–12. Available at: https://www-jstor-org.web.bisu.edu.cn/stable/41622117

Arab Barometer (2013) Arab Barometer Survey Instrument. Available at: http://www.arabbarometer.org

Arriola

Leonardo R

Johnson

Martha C

(2013) Ethnic Politics and Women’s Empowerment in Africa: Ministerial appointments to executive cabinets. American Journal of Political Science 58(2): 495–510.

10.

Barber

Kristie

(2007) Tribal Ties Among Zanzabaris in Oman. Macalester Abroad: Research and Writing from Off-campus Study (1)1: Article 4. Available at: http://digitalcommons.macalester.edu/macabroad/vol1/iss1/4

11.

Benstead

Lindsay J

Jamal

Amaney A

Lust

Ellen

(2015) Is It Gender, Religiosity or Both? A role congruity theory of candidate electability in transitional Tunisia. Perspectives on Politics 13(1): 74–94.

12.

Bjarnegård

Elin

(2013) Gender, Informal Institutions and Political Recruitment: Explaining Male Dominance in Parliamentary Representation. Springer.

13.

Bush

Sarah Sunn

Gao

Eleanor X

(Forthcoming) Small Tribes, Big Gains: The strategic uses of gender quotas in the middle east. Comparative Politics. Available at: https://sites.temple.edu/sarahbush/files/2012/08/Bush_and_Gao_CP_2016.pdf

14.

Bush

Sarah Sunn

Jamal

Amaney A

(2015) Anti-Americanism, Authoritarian Politics, and Attitudes about Women’s Representation: Evidence from a survey experiment in Jordan. International Studies Quarterly 59(1): 34–45.

15.

Coffé

Hilde

Dilli

Selin

(2015) The Gender Gap in Political Participation in Muslim-Majority Countries. International Political Science Review 36(5): 526–544.

16.

Cooke

Miriam

(2014) Tribal Modern: Branding New Nations in the Arab Gulf. Berkeley: University of California Press.

17.

Dilli

Selin

Rijpma

Auke

Carmichael

Sarah G

(2015) Achieving Gender Equality: Development versus historical legacies. CESifo Economic Studies 61(1): 301–334.

18.

Dolan

Kathleen

(1998) Voting for Women in the ‘Year of the Woman.’ American Journal of Political Science 42(1): 272–293.

19.

Finkel

Steven E

Smith

Amy Erica

(2011) Civic Education, Political Discussion, and the Social Transmission of Democratic Knowledge and Values in a New Democracy: Kenya 2002. American Journal of Political Science 55(2): 417–435.

20.

Gibson

James L

(2001) Social Networks, Civil Society, and the Prospects for Consolidating Russia’s Democratic Transition. American Journal of Political Science 45(1): 51–68.

21.

Hayes

Bernadette C

McAllister

Ian

Studlar

Donley T

(2000) Gender, Postmaterialism, and Feminism in Comparative Perspective. International Political Science Review 21(4): 425–439.

22.

Helmke

Gretchen

Levitsky

Steven

(2004) Informal Institutions and Comparative Politics: A research agenda. Perspectives on Politics 2(4): 725–740.

23.

Hudson

Valerie M

Bowen

Donna Lee

Nielsen

Perpetua Lynne

(2015) Clan Governance and State Stability: The relationship between female subordination and political order. American Political Science Review 109(3): 535–555.

24.

Inglehart

Ronald

Norris

Pippa

(2000) The Developmental Theory of the Gender Gap: Women’s and Men’s voting behavior in global perspective. International Political Science Review 21(4): 441–463.

25.

Inter-Parliamentary Union (IPU) (2016) Women in Politics: 2016. Available at: http://www.ipu.org/wmn-e/world.htm

26.

Jalalzai

Farida

Krook

Mona Lena

(2010) Beyond Hillary and Benazir: Women’s political leadership worldwide. International Political Science Review 31(1): 5–21.

27.

Kehoe

Karrie

(2013) Factbox: Women’s rights in the Arab world. Available at: http://www.reuters.com/article/us-arab-women-factbox-idUSBRE9AB00I20131112#uuVkfLKKDGevJPsv.97 (accessed 1 November 2015).

28.

Kenworthy

Lane

Malami

Melissa

(1999) Gender Inequality in Political Representation: A worldwide comparative analysis. Social Forces 78(1): 235.

29.

Krook

Mona Lena

(2009) Quotas for Women in Politics: Gender and Candidate Selection Reform Worldwide. New York: Oxford University Press.

30.

Kunovich

Sheri

Paxton

Pamela

(2005) Pathways to Power: The role of political parties in women’s national political representation. American Journal of Sociology 111(2): 505–552.

31.

Majlis al

Shura

(2008) Formation of Majlis al Shura for the Sixth Term (October 2007-September 2007) [in Arabic]. Majlis al Shura, November.

32.

Miguel

Carolina de

Jamal

Amaney A

Tessler

Mark

(2015) Elections in the Arab World: Why do citizens turn out? Comparative Political Studies 48(11): 1355–1388.

33.

Norris

Pippa

Inglehart

Ronald

(2001) Cultural Obstacles to Equal Representation. Journal of Democracy 12(3): 126–140.

34.

Paxton

Pamela

Kunovich

Sheri

(2003) Women’s Political Representation: The importance of ideology. Social Forces 82(1): 87–113.

35.

Peterson

(1989) The Political Status of Women in the Arab Gulf States. Middle East Journal 43(1): 34–50.

36.

Peterson

(2004) Oman’s Diverse Society: Northern Oman. Middle East Journal 58(1): 32–51.

37.

Randall

Vicky

(1987) Women and Politics: An International Perspective. Chicago: University of Chicago Press.

38.

Ross

Michael L

(2008) Oil, Islam, and Women. American Political Science Review 102(1): 107–123.

39.

Sabbagh

Amal

(2005) The Arab States: Enhancing women’s political participation. In: Ballington

Julie

Karam

Azza

(eds) Women in Parliament: Beyond Numbers. Stockholm: International Institute for Democracy and Electoral Assistance, 52–71.

40.

Sanbonmatsu

Kira

(2003) Gender-Related Political Knowledge and the Descriptive Representation of Women. Political Behavior 25(4): 367–388.

41.

Torstrick

Rebecca L

Faier

Elizabeth

(2009) Culture and Customs of the Arab Gulf States. Westport, CT: Greenwood Press.

42.

United Nations Development Programme (2014) Human Development Report 2014. Available at: http://hdr.undp.org/en/country-reports

43.

Varghese

Thresiamma

(2011) Women Empowerment in Oman: A study based on women empowerment index. Far East Journal of Psychology and Business 2(2): 37–53.

44.

Yaghi

Abdulfattah

Antwi-Boateng

Osman

(2015) Determinants of UAE Voters’ Preferences for Federal National Council Candidates. Digest of Middle East Studies 24(2): 213–235.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB