Survey Methodology
Sampling
The survey analyzed data collected from a random sample of 100 mosques. This sample size provided sufficient statistical power to find a statistically significant association between most of the selected Shari‘a-adherent behaviors and violence-positive variables. Most Shari‘a adherence and violence-positive variables exhibited a strong correlation while some exhibited a weak or no correlation. A sample size of 100 mosques also allowed the survey to extrapolate to all mosques in the United States at a 95 percent confidence interval with a margin of error of +/-9.6 percent.
The survey was developed by using state-by-state estimates of the Muslim population extracted from the only extant such survey.[1] This was then used to create a listing of all states whose Muslim population represented at least 1 percent of the estimated total United States Muslim population. Fourteen states and the District of Columbia (“15 randomly selected states”) were randomly selected from the final listing to accommodate limits on physical logistics and personnel resources for the actual survey.
For each of the fourteen states and D.C., cities with the highest estimated concentrations of Muslims were identified, and mosques within those areas were eventually selected. The survey combined the data on 1,209 mosques listed in “The Mosque in America: A National Portrait”[2] with the data on the 1,659 mosques obtained online from Harvard’s Pluralism Project,[3] with duplicates eliminated. Mosques were excluded from the list if there were indications that they were no longer operating, with a final site list yielding a total of 1,401 potential mosques for the survey.
The dates and prayer times for visiting mosques were also randomly selected. If a mosque was found to be closed, abandoned, or not at the address listed, then the next mosque that appeared on the randomized list for that city was visited. When the dominant language of the subject mosque was determined to be other than English, such as Arabic, Urdu, or Farsi, the surveyor who visited the mosque was fluent in that language. Each mosque was visited twice, once between May 18, 2007, and December 4, 2008 (“Survey Period”), and then again between May 10, 2009, and May 30, 2010 (“Audit Period”). The results of the Audit Period confirmed the findings in the Survey Period in all but nine mosques.
Data Collection
A surveyor visited a subject mosque in order: (a) to observe and record 12 Shari‘a-adherent behaviors of the worshipers and the imam (or lay leader); (b) to observe whether the mosque contained the selected materials rated as moderate and severe; (c) to observe whether the mosque contained materials promoting, praising, or supporting violence or violent jihad; and (d) to observe whether the mosque contained materials indicating the mosque had invited guest speakers known to have promoted violent jihad.
Thus, the survey only examined the presence of Shari‘a-adherent behaviors, the presence of violence- positive materials in mosques, whether an imam would promote the study of violence-positive materials, and whether a mosque was used as a forum to promote violent jihad. Since there is no central body to which all mosques belong, it was difficult to ascertain that the sampling universe list was complete. This may have introduced bias into the sampling although the authors find no evidence of any systemic distortions.
References
1. Barry A. Kosmin and Seymour P. Lachman, One Nation under God: Religion in Contemporary American Society (New York: Harmony Books, 1993), pp. 96-7.
2. Ihsan Bagby, Paul M. Perl, and Bryan T. Froehle, “The Mosque in America: A National Portrait,” Council on American Islamic Relations, Washington, D.C., Apr. 26, 2001.
3. “Directory of Religious Centers,” Pluralism Project, Harvard University, Cambridge, accessed Oct. 30, 2010.