Iranians pray in Tehran during a 2013 Arbaeen observance marking the end of the 40-day mourning period commemorating the anniversary of the death of Imam Hussein, the prophet Muhammad’s grandson. (Vahid Salemi/AP)

Millions of Shiite Muslims traveled from across the Muslim world to walk in procession to the shrine of Imam Hussein in Karbala, Iraq, today for the world’s largest, yet largely unknown, annual pilgrimage. This ziyara, or visit, to the shrine in southern Iraq is known as the Arbaeen. It marks the end of 40 days of mourning for the martyrdom of Imam Hussein, a central moment in Shiite tradition. More than 10 times the size of the hajj — because Iraq does not limit the number of pilgrims — it was restricted for years during the rule of Saddam Hussein because of its potential for sectarian collective action.

Last November, we used this pilgrimage as an opportunity to survey religious Shiites from Iran and Iraq, an important but largely understudied population. Using both traditional survey instruments and experimental methods, the survey covered a wide range of issues, including religion and politics, democracy, women’s rights, regional conflict and Iran’s nuclear agreement. In a recent article, we reported key findings from this survey, including the broad support among respondents for Iranian regional policy and for the nuclear agreement with the West.

Why it’s hard to survey religious Shiites

The survey demonstrates a novel approach to research on the views of religious Shiites in Iran and Iraq. Regime restrictions in countries such as Iran limit both the ability to conduct surveys and the permissible questions within them; conflicts in Iraq, Yemen and Syria make carrying out nationally representative surveys nearly impossible.

Conducting a survey of religious Shiites is an even greater challenge than surveying Shiite populations as a whole. Western interventions often assume that these devout individuals make up the support base for their governments. Yet little data exists on their prevalence, geographic distribution, or religious practice and beliefs. Clustered sampling at places of worship is one potential corrective to this knowledge gap, but national data on mosques are not readily available. This would also under-sample important groups like women and dissenting individuals who choose to worship from home.

However, the Arbaeen pilgrimage provides a unique opportunity to sample religious Shiites from across Iran and Iraq, as the visit to Imam Hussein’s shrine is considered an ultimate expression of one’s piety and devotion. We focused on securing a regionally representative sample based on Shiite distribution in each country, rather than pursuing a representative sample of all the pilgrims.

This was also a unique opportunity to sample practicing religious women outside their homes, as religious women are active participants in the procession.

Why we survey pilgrims

The pilgrimage’s unique processional nature facilitated this regional targeting during the survey process. Able-bodied Iraqi pilgrims walk from their homes across Iraq to Karbala, with some traveling as far as about 300 miles from Basra in the south. Iranians usually travel via bus to the city of Najaf, 50 miles south of Karbala, then walk from there. During the procession, tents, or mawakib, stationed beside the path provide rest and refreshment for pilgrims. For Iraqis, these mawakib are unofficially organized by region. Iranian mawakib are less specifically targeted but still often have broader regional trends. By visiting different tents, we were able to gain a geographically representative sample.

Because our survey asked about sensitive topics, we used experimental methods to try to measure latent perspectives on sectarian tensions, Iran’s nuclear program, and attitudes toward the West, China and Russia. In addition to conjoint analysis, endorsement primes and memory primes, we also examined non-response as a way to understand topic sensitivity and knowledge.

When no response can be an important response

On the issue of sectarianism, respondents were quick to give inclusive answers to direct survey questions. About 90 percent stated they supported Sunni-Shiite dialogue to mitigate conflict and would support Sunnis and Shiites praying together in the same mosque. Yet when given choices between hypothetical neighbors and spouses in a conjoint experiment, sect was more important than religiosity, race and even prior marital status. Only alcoholism was viewed less favorably than Sunni neighbors or spouses.

Participation in the survey hovered at about 85 percent, with the primary reason for declining being that individuals were traveling with a group and did not have time. Though the median rate for question-level non-response was in the single digits, it jumped as high as 50 percent for some questions.

This variation in non-response creates an alternative method of measuring sensitivity to certain topics. The highest non-response rate was for a question asking whether Sunnis and Shiites had similar interpretations of the role of violence in Islam, to which nearly half of Iranians and Iraqis did not respond. Other topics with high non-response rates included democracy and the relationship between government and religion, as well as regional politics and Iran’s foreign policy.

While women were more likely, on average, to have higher non-response rates than men, women were relatively more likely to speak up on issues of gender. Contrary to what Western news and analysis often suggests, this trend highlights religious women’s role in supporting gender norms, rather than merely blindly following male dogma.

Religious Shiites are not monolithic. This is important.

We find that even these devout individuals express surprisingly diverse views on the proper relationship between religion and the state and are nearly divided on the perceived costs and benefits of democratic regimes.

Despite strong perceptions of U.S. bias in favor of Sunni countries, the majority of respondents favored continued engagement with the United States on topics of mutual interest but oppose U.S. military intervention. Most support the Iranian nuclear agreement and share relatively liberal views on women’s rights in government and the workplace, but less so within the family. Although overwhelmingly supportive of both Iran’s interventions and Shiite causes throughout the Middle East, we also found that many respondents were optimistic about democracy and did not necessarily see theocracy or religious political parties as the ideal political system.

The full results of the survey, as well as a detailed background on the pilgrimage and further discussion of the sampling can be found here. This approach presents a unique template for surveying hard-to-reach populations in an increasingly mobile world. The gaps between our findings from direct questions and experimental survey methods, along with the observed patterns of non-response, highlight the importance of the continued use of creative tools to gauge latent beliefs on sensitive topics.

Fotini Christia is an associate professor of political science at the Massachusetts Institute of Technology. She carried out this research while an inaugural Andrew Carnegie fellow. Elizabeth Dekeyser and Dean Knox are PhD candidates in political science at MIT.