[A PDF copy of this study is available here. - Adobe Acrobat or Reader required]
American Religious Identification Survey 2001
PROFILE OF THE US MUSLIM POPULATION
ARIS Report No. 2, October 2001
PRINCIPAL INVESTIGATORS
Dr Barry A Kosmin & Dr Egon Mayer
STUDY DIRECTOR
Dr Ariela Keysar
Background
The following information is based on the recently completed American Religious Identification Survey, 2001 (ARIS 2001), just released by The Graduate Center of the City University of New York.
Because of wide national interest in the size and characteristics of the Muslim population, these data are being released in advance of profiles of other religious groups in the United States.
Methodology
The American Religious Identification Survey 2001 was based on a random digit-dialled telephone survey of 50,281 American residential households in the continental USA (48 states). The methodology largely replicates the widely reported and pioneering 1990 National Survey of Religious Identification (NSRI) carried out at the Graduate Center of the City University of New York. ARIS 2001 thus provides a unique time series of information concerning the religious identification choices of American adults.
Data were collected over a 17-week period, from February to June 2001, at the rate of about 3,000 completed interviews a week by ICR/CENTRIS Survey Research Group of Media, PA, as part of their national telephone omnibus market research (EXCEL/ACCESS) surveys.
The primary question of the interview was: What is your religion, if any? The religion of the spouse/partner was also asked. If the initial answer was 'Protestant' or 'Christian' further questions were asked to probe which particular denomination.
Among the more than 50,000 respondents interviewed for ARIS 2001, 219 identified their religion as Muslim or Islam. The data reported here are based on those nationally representative responses and are statistically weighted to represent the US adult population as a whole.
1. POPULATION SIZE
The ARIS 2001 total in Table 1, which shows 1,104,000 Muslim adults, is smaller than the figures in current circulation. Yet it is twice the number reported in the NSRI survey in 1990. Allowing for a sampling error of +/-0.5 percent, the ARIS 2001 figure maybe adjusted upwards to its maximum range of 1.0 percent of all 208 million American adults (see Table 1). With such an adjustment, the total national figure for US Muslims is 2.2 million, giving a total national population (including children) of just under 3 million. By comparison, the CUNY National Survey of Religious Identification 1990 found that 0.3 percent of respondents adhered to Islam.
Click here for table
There are a number of logical, technical reasons for the ARIS 2001 findings, aside from the fact that its numbers are the only estimates in existence that are based on objective scientific measurement (Smith, October 2001). Nevertheless certain inherent difficulties have to be acknowledged. Some have to do with terminology. Others relate to varying levels of co-operation with surveys, which require sample weighting and efforts to adjust for specific problems stemming from the undercounting of recent immigrants to the US.
The Confusion of Arabs with Muslims
First, considerable confusion exists in many circles between the Arab-American and Muslim populations. Put simply: the majority of Arab-Americans are not Muslims and the majority of Muslims are not Arab-Americans.
This counter-intuitive fact may seem strange to those who are not aware of the nation's religious and immigration history. Yet recent research by the Arab-American Institute confirms the 1990 NSRI (Kosmin & Lachman 1993) as well as ARIS 2001 findings in this regard.
The Arab-American Population
Unfortunately, the term Arab-American, as an ethnic category, is not counted directly by the US Census. In addition, the country of birth data, which shows a little over one million persons born in Arab states (The Wall Street Journal, 9/28/01), obviously does not include US-born Arab-Americans.
The Arab-American Institute (AAI), claims there are 3.5 million Americans who have some Arab heritage, the majority of whom are Lebanese. More noteworthy still, AAI also reports the Christian component to be 75 percent, while the Muslim component is only 25 percent i.e. around 850,000 Arab-Americans.
Since most Muslim experts claim that around 1 in 4 American Muslims to be of Arab origin, then if we use this AAI estimate of 850,000 Arab-American Muslims, the total Muslim population logically cannot be more than 3.4 million-which, in fact, is close to the upper range of the ARIS 2001 estimate of 2.8 million.
The above AAI figures also appear to point to the fact that even recent immigrants from Egypt, Jordan and Iraq are disproportionately Christian. This is not surprising. From a sociological perspective we should expect the most westernized populations from these countries to be most attracted to settle in the US. Moreover, Arab Christians such as Egyptian Copts, Lebanese Maronites and Iraqi Chaldeans and Assyrians have long suffered from an identity crisis and insecurity in the strife-torn Middle East.
A similar pattern can be assumed to have occurred among immigrants from Iran, a non-Arab country. This explains why the religious minorities of Iran such as Christians, Jews, Baha'i, Zoroastrians and agnostics happen to be strongly represented among Americans of Iranian descent. It stands to reason that few strict believers in the Shiite Muslim faith would leave the Islamic Republic to dwell in the "Great Satan".
Non-Christian religious groups
To put misperceptions about the US Muslim population numbers in perspective, it is necessary to remember that the American public traditionally has a poor understanding of the actual demographic composition of the American population (Nadeau, Niemi and Levine 1993; Smith, 2001). People tend to greatly overestimate the size of minority groups. In fact, the smaller the minority, the greater the exaggeration-by a factor of 2 for Hispanics, a factor of 4 for Asians and nearly 10 for American Indians.
This innumeracy is understandable and particularly strong in the area of religion, where the US Census does not venture to count the population. Yet to date, no scientific national survey has found the non-Christian religious groups to account for more than 4 percent of the American population. That means that groups such as Buddhists, Hindus, Baha'is, Jews by religion, Muslims, Sikhs, Rastafarians, Scientologists and Druids, etc. altogether number less than 12 million persons out of a national population of 270 million (including children whose ultimate religious preferences are very difficult to know).
The ARIS 2001 study found that 3.7 percent of American adults adhere to these minority religions. Similarly "Other Religions" account for 4 percent of Harvard's 2000 Social Capital Survey Benchmark Survey of 3,000 respondents. This same figure also consistently appears in the religion schedules of the annual General Social Surveys of the National Opinion Research Center (NORC) at the University of Chicago.
In 1991 the Princeton Religious Research Center, utilizing accumulated Gallup Polls attempted to verify the above findings. "Where are the Muslims in the United States?", an article in the September issue of Emerging Trends reported that researchers found only 0.2 percent Muslims. They concluded that enthusiastic Muslim supporters and leaders-as well certain population groups in the US who feared the same purported trend-had put in circulation greatly inflated numbers.
Sample weighting, the 'undercounting' phenomenon and sample error
The sample weighting techniques employed by ARIS 2001 try to adjust for the known reluctance of certain sections of the population to be surveyed. This adjustment effectively recreates the sample to reflect the size and known characteristics of the American population according to the findings of the 2000 US National Census.
Obviously there is undercounting of religious minorities beyond our ability to weight the data on demographic variables. Among recent immigrants from dictatorships and fractured societies, language problems and a fear of answering personal questions about religion and other sensitive topics tend to reduce the level of cooperation with surveys in the US.
The undercounting phenomenon and the possibility of higher than average refusal rates explain the population figures of the ARIS report in Table 1. They allow for the possibility that the real proportion of Muslims in the US population may lie in the upper range of the maximum sampling error and could be twice what the survey found: 1 percent rather than 0.5 percent.
Nevertheless, there are limits to how far one can go in the upward adjustment process, given a finite US population-without reducing the size of other religious groups. In statistical terms, this means there is a 95 percent probability that the adult US Muslim population is in the range of 1.1 to 2.2 million.
2. THE RELIGIOUS CHARACTERISTICS OF US MUSLIMS
According to the National Congregational Survey carried out by Hartford
Seminary there are 1,209 mosques in the US. ARIS found that 62 percent of Muslim respondents claimed to belong to a mosque. This suggests-using the population range outlined above-that somewhere between 600,000 and 1.2 million adults are involved with a mosque giving 600 to a thousand members per mosque. The Islamic section of the Hartford project has estimated that around 2 million people are religiously involved with mosques and that 411,000 worshippers attend the weekly Friday Jum'ah prayers (Bagby, Mamiya and Nimer, 2001).
The demographic characteristics Bagby's report are in line with the ARIS findings reported in Table 1. They show a predominantly young, male, unmarried population. Mosque attendees are also overwhelmingly male (78 percent), while half are under 35 years of age. Bagby also reports that 30 percent of Mosque attendees are African-American-a figure which is almost exactly in line with the ARIS findings of 27 percent of all Muslim adherents.
Click here for table
Bagby's report also showed that two-thirds of the 19,000 converts a year to Islam are men and that 14,000 are African-Americans. The ARIS findings support this portrait of religious conversion. They show that 183,000 people have indeed converted to Islam. However, in addition, ARIS also counts people who leave the faith. They number 98,000.
In addition, ARIS asked about the religion of spouses and partners, so it is now possible to report that 27 percent of Muslims with a spouse or partner live in mixed religious families. Thus, in relation to Table 2, it is relevant to note that not every member of a "Muslim household" is an adherent of Islam, particularly given the preponderance of male adherents in this religious group.
3. SOCIO-DEMOGRAPHIC CHARACTERISTICS
The geographical distribution reported in Table 1 is in accordance with expected patterns of residence. Muslim Americans tend to be heavily concentrated in New York, Chicago, Los Angeles, Detroit and Texas.
Table 2 shows that Muslim households are larger than the national average-a figure which reflects the younger than average median age of the population (28 years of age). In terms of income Muslim households are more polarized than average: there are more poorer-but fewer middle income-families.
In the political arena Muslims have only half the national rate of voter registration. Their party preferences show that 35 percent favor the Democratic Party and 19 percent the Republican, while 39 percent declare themselves to be independent.
Finally, although ARIS 2001 shows that Muslims account for 0.5 to 1.0 percent of Americans of all racial backgrounds, they comprise approximately 2 percent of the total African-American population, and 5 percent among those who report themselves as Asians.
Concluding remarks
Given the fact that the US Constitution precludes any national census from investigating the religious preferences and characteristics of American citizens, it is impossible to obtain a precise, definitive number for the size of US religious groups. It is likewise unfeasible for any independent body to conduct an investigation with the reach and depth of the US National Census. Religious researchers are therefore limited to two possibilities. They can investigate the places of worship and religious institutions themselves. Or they can survey a representative national sample as with ARIS 2001.
To be sure, the counting of attendees of places of worship and formal members of a religious group obviously does not usually include the full range of adherents-particularly the occasional worshipper and those who only attend for life cycle events such as marriage and funerals. For this reason, religious leaders' observations and their administrative records are neither comprehensive nor scientifically reliable as to the religious self-identification of the total US population.
Consequently, there will always be a gap between the view of the religious institutions as to the size and range of their potential constituency and that provided by the members of the public themselves. Depending on their membership criteria and overall religious outlook, it is possible for religious groups to inflate-or deflate-the size of their prospective constituencies.
In the case of a minority religious group, the only way to scientifically measure the size and characteristics of the population and increase the precision over and above the data provided in Tables 1 and 2 is to sample an even larger number of representative households than the 50,281 covered by ARIS 2001.
References
"Arab Americans", The Economist, 9.22.2001
"Studies Suggest Lower Count for Number of U.S. Muslims", The New York Times, 10.25.2001
"Where are the Muslims in the United States?", Emerging Trends, September 1991.
"Why the Census of 2000 Failed to Count Arabs", The Wall Street Journal, 9.28.2001
Ihsan Bagby, Lawrence Mamiya and Mohamed Nimer, The Mosque in America: A National Portrait, Council on American-Islamic Relations, Washington DC, 2001.
Barry A. Kosmin and Seymour Lachman, One Nation Under God: Religion in Contemporary American Society, New York, Harmony Books, 1993.
Barry A. Kosmin, Egon Mayer and Ariela Keysar, American Religious Identification Survey, The Graduate Center City University of New York, 2001.
Richard Nadeau, Richard G. Niemi, and Jeffrey Levine, "Innumeracy about Minority Populations", Public Opinion Quarterly,1993, 57: 332-347.
Tom W. Smith, Intergroup Relations in a Diverse America: Data from the 2000 General Social Survey, New York, The American Jewish Committee, 2001.
Tom W. Smith, Estimating the Muslim Population in the United States, New York, The American Jewish Committee, October 2001.
|
