"How many Muslims (or Jews or Mormons or Christians) live in the United States?" For decades the Census has not been allowed to count noses to determine who is what religiously. So observers of American religion are left to their own devices to assess the size of denominations, and of claimed affiliations and preferences.

Counting noses has come to depend on two sources. One source is poll-takers calling during the dinner hour to ask "What is your religious preference?" The other source is religious leaders, on both the local and the national scene. People who respond to telephone interviewers may have all kinds of motives for declaring themselves as part of this or that group, or no group at all. And people who report on the size of their congregations, denominations, and cohorts also have a variety of motives. These include, but are not limited to, claiming bragging rights, being able to throw weight around, or whining about decline in an unfriendly world. Congregations inflate numbers to show that they are successful. Then they trim them when their denominations start "assessing" on a per capita basis.

All this by way of background to the latest stir on the counting scene. Daniel Pipes, director of the Middle East Forum and no friend to most of organized Islam, addressed the question "How many Muslims live in the United States?" in the November second issue of The Chicago Sun-Times. He noted that in 1986 the Saudi embassy claimed 10 million. But "a large 1990 demographic survey counted 1.3 million." In 1998, a Pakistani paper put it at 12 million. Wildly: the "usually authoritative Yearbook of American and Canadian Churches" counted 527,000 American Muslims in 1996 and six times as many (3.3. million) in 1998. There can't have been that many immigrants or converts; there is something subjective here.

Muslim organizations came up with a "guesstimation" of six, now seven million. So often is this figure used that people count it as being reliable. Pipes argues, credibly, that most religions inflate numbers to gain more voice in the public sphere. So he welcomed the new American Religious Identification Survey 2001 from CUNY, which polled fifty thousand people, and figured that 1.8 million Americans are Muslim. Tom Smith at the University of Chicago reviewed all reviews and figures 1,886,000 to 2,814,000 Muslims, not six or eight million. Both polling agencies instantly became subjects of attack from angry Muslim leaders.

Pipes is correct. There are fewer Muslims than Muslim organizations claim. "How many Muslims live in the United States?" We don't know, any more than we know with precision how many Jews or Christians do. The question to ask of all statistical ventures on this front is, "In whose interest is it to inflate or deflate the figures?" There are plenty of interests these days. Handle with care.