Are Arabs sexist? The institutions, perhaps, but not the people

Arab societies are often regarded as bad places for women and girls. According to many observers, Arabic and Islamic culture can combine to foster attitudes that are inhospitable to gender equality.

The results of a survey experiment we are conducting may challenge common assumptions. Women do face special difficulties in Arab lands, which are reflected in bleak statistics about inequalities in political and economic life. But we find little evidence that popular attitudes are to blame. Our data from Lebanon, with its mix of Muslims and Christians, may be particularly illuminating.

Inequities in political power are actually glaring in Arab lands. According to the Interparliamentary Union, the proportion of seats in national legislatures occupied by women is 24 percent in the Americas, 23 percent in Europe, 21 percent in sub-Saharan Africa, and just 14 percent in the Arab states. Conditions in Lebanon are particularly unfavourable: women hold just three percent of seats in the National Assembly.

Economic power is also skewed. According to the United Nations Development Programme, the female-to-male labour participation ratio in Arab countries is about 0.4. That figure means that for every ten men employed in Arab lands, only four women are employed.

The number is even lower than the Arab average in Lebanon, where for every ten men who are in the workforce, only three women are. The Americas, Europe, and sub-Saharan Africa are in the 0.6-0.8 range, meaning that for every ten men who are employed in these regions, between six and eight women are employed.

Are these inequalities due to – or at least consistent with – popular attitudes? If so, we could say that popular preferences and situational outcomes are congruent. Perhaps Arab or Muslim culture inclines people toward especially traditional attitudes on gender roles.

There is some evidence to support this view. The World Values Survey (WVS) asks respondents whether they believe that men make better political leaders than women do. The survey also asks respondents whether they think that men should receive preferential treatment in consideration for employment when jobs are scarce.

Lebanon is not included in the WVS, but seventeen predominately Muslim countries, including Egypt, Iraq, Jordan, Morocco, and Saudi Arabia, are. We have analysed the data from the WVS carefully. We find that, even controlling for other factors that might affect attitudes, Muslims are substantially more likely than non-Muslims to say that men make better political leaders than women and that men should receive preferential treatment in employment.

On the question of political leadership we detect a separate regional effect as well, with Muslims in Arab countries even more inclined to support gender inequality than are Muslims from outside Arab lands.

The WVS provides a valuable source. But it produces only what are called “observational data”. The respondents know what the survey question is about, and their answers may be influenced by many extraneous factors.

Data drawn from experiments in which subjects are randomly assigned to a control group and a treated group are more reliable. In survey experiments, respondents do not know what the questions are about and are in no way primed to give a particular answer.

In an effort to overcome the limitations of observational data, we have recently conducted survey experiments on attitudes toward gender inequality in several countries, including Lebanon. We seek to uncover how respondents’ religious affiliation, gender, and other factors may affect their attitudes.

Our survey includes a question on candidates for public office and a question on pay-for-work. The treatment we use is the random assignment of the gender of the subject of each question.

In the item on candidates for office, we alternate the names of the candidates – one male and one female – randomly. We give half the respondents a choice between a candidate with a male name who works as a schoolteacher and aims to improve schools and sanitation services, and a candidate with a female name who is a businessperson who wants to improve the business climate and strengthen law enforcement.

The other half of respondents receives the question with the gender of the candidates reversed, so that the female candidate is the progressive and the male candidate the conservative. Respondents should base their selection on the candidates’ programs, but we can also tell if the gender of the candidate influences choices.

In the question on pay-for-work, we pose a scenario in which a person is entering the workforce and has secured a job at a convenience store. We state the average hourly wage for workers starting such a position, and then ask respondents what they think this new employee should be paid. We offer three choices, the first of which is a bit higher than the average wage, the second of which is the average, and third of which is a bit lower than average.

Half of all respondents are assigned the question with a female name used to identify the new employee and half of the respondents are given the question with a male name. We can assess whether respondents favour paying the male and female workers at different rates.

The results are remarkable. For neither candidates for office nor on pay-for-work do we find substantial gender bias in society as a whole. We find no evidence that the beliefs and values of ordinary Lebanese are responsible for the substantial inequities women encounter in political life and in the workforce.

On the question on candidates for office, male respondents tend to favour male candidates for office – but at a lower rate than female respondents favour female candidates. The findings call into question the notion that women are as likely, or almost as likely, as men are to favour male candidates in races for public office in the Arab world.

Furthermore, we find no statistically significant difference between Christians and Muslims. Muslim respondents are no more inclined to favour the male candidate than Christian respondents are. Nor is there a statistically significant gap between Shia and Sunni respondents in their attitudes toward female vs. male candidates.

If women are underrepresented in public office in Lebanon, it is not because the Lebanese people oppose having women in power. The answers to why women are underrepresented must lie elsewhere – in power structures, institutions, or elite strategies.

Our findings on pay-for-work also show no signs of popular bias against women. Respondents were not more likely to underpay the female employee or overpay the male employee. Nor do we find evidence of a statistically significant difference among Christian, Shia and Sunni respondents. Women’s participation in the workforce in Lebanon is low even by Arab standards, but we find no evidence of a popular tendency to regard a man’s labour as more valuable than a woman’s labour.

Thus, the inequities we find in actual conditions are not buttressed by popular attitudes. On the contrary, inequality in circumstances might stem less from popular culture and psychology than from institutions and power relations that survive despite – rather than because of – popular preferences.

M. Steven Fish is Professor of Political Science at the University of California, Berkeley and the author of Are Muslims Distinctive? A Look at the Evidence (Oxford, 2011).

Rose McDermott is Professor of Political Science at Brown University and the co-editor of Man Is By Nature a Political Animal: Evolution, Biology, and Politics (University of Chicago Press, 2011).

Danielle N. Lussier is Assistant Professor of Political Science at Grinnell College and the author of Activating Democracy: Political Participation and the Fate of Regime Change in Russia and Indonesia (forthcoming, 2014).