Numbers: Why you can’t count on them

Download all chapters of "Reporting on Religion" (PDF)

Whoever said “Numbers never lie” was not a religion reporter. Beware of confidently using specific numbers about religious identification or belief. Here’s why.

  • The U.S. Census, the usual standard for counting people and their characteristics, does not ask people their religious affiliation.
  • There is no single religion survey that is considered to be the most reliable (see Page 27). The results differ depending on what options are offered, how people are contacted, how many people are surveyed and other factors. Numbers can vary widely, and many faith groups are so small that they rarely show up on surveys in proportion to their actual numbers. Some traditions – predominantly African-American denominations, for example – are typically underrepresented because of difficulty in obtaining numbers.
  • With Roman Catholics, there is one pope and a highly structured hierarchy that tracks membership. But in many religions, such as Islam, there is no official governing body and no official count.
  • Some faiths, such as the fast-growing Pentecostal movement, include people from many denominations, so there is no central record-keeping. Evangelical Christians are difficult to count because they often belong to nondenominational churches, official denominations or sometimes none at all.
  • Denominations and religions count their members differently, if at all, so it is difficult to compare their sizes. For example, the Southern Baptist Convention, which does not baptize infants, counts people who are baptized. The United Methodist Church,which baptizes infants, counts people once they are confirmed. Mosques don’t require membership, so estimates of Muslims are just that – estimates.
  • Formal affiliation with a religious group doesn’t give a complete picture. Only about half of Jews in the U.S. are affiliated with synagogues, so the National Jewish Population Survey uses four questions to determine Jewish identity.
  • Some houses of worship or faith groups can be competitive about touting numbers of adherents or members, to the point that the number of Muslims and Jews in America has become highly politicized. For example, one survey by Jewish researchers counted fewer than 3 million Muslims at a time when Muslim groups were claiming numbers as high as 6 million. In contrast, some churches, such as the Church of Christ, Scientist, do not publish statistics because its numbers have declined so much.
  • Categories are controversial.Mormons consider themselves Christian, but most Christian groups do not. Messianic Jews, who believe Jesus was the messiah Jews await, consider themselves Jewish, but most Jews consider them Christian.
  • There are only estimates of the number of Buddhists, who are difficult to count in part because immigrants and American converts practice so differently. Also, some people meditate or practice aspects of Buddhism but don’t consider themselves Buddhist, or they combine Buddhism with another faith, such as Judaism, resulting in people who call themselves “JewBus.”
  • Beware of calling any faith the “fastest-growing” in any context. Any number of faiths have been called the fastest-growing in the United States or the world, often without any documentation. Salt Lake Tribune reporter Peggy Fletcher Stack wrote a 2006 story debunking the myth that the Church of Jesus Christ of Latter-day Saints was the fastest-growing faith in the world and reporting that the Seventh-day Adventists, Assemblies of God and Pentecostal groups were growing faster.
  • Poll results differ, depending on how questions are asked. This applies not only to believers, but also to descriptions of their beliefs. For example, people’s stated beliefs about a divine role in creation vary depending on the number and types of choices they are given.
  • People frequently lie when asked about religion, perhaps out of a desire to look good and perhaps out of denial. Prominent national surveys such as those by Gallup and others generally show that about 40 percent of those in the U.S. say they attend worship services on any given Sunday, but numerous studies have found that the actual number is much closer to 25 percent.
  • Sophisticated advocacy groups promote their own polls,which support their own agendas. Beware of spin.

How to make it all add up

Don’t let the size of faith groups – either nationally or locally – overly influence your coverage. You’ll likely do more stories on Catholics and evangelicals because of their numbers, but many groups’ impact and influence outstrip their size. The Episcopal Church is small, but its battles over homosexuality are closely watched. Jews are less than 2 percent of the population but have an important voice. Buddhism has relatively few followers but permeates the culture. All the general guidelines of good journalism apply when quoting statistics in religion stories. In addition:

  • Be specific about what numbers represent. For example, specify what numbers are based on (worship attendance, membership, baptism, etc.).
  • Be careful with comparisons. If you have apples and oranges, say so and note that two groups’ definition of “member” differs.
  • Look carefully at poll questions and results yourself, rather than accepting one statistic without question. The poll, as a whole,may tell a different story.
  • Check to see if different organizations have done polls on the same subject. Polling and search engines make this easy.

Disputed statistics

1 Give a range of numbers or qualify a statistic if it is in doubt: There are 4 million to 7 million Muslims in America, according to various surveys. There are well more than 1 million Hindus, experts say.

2 Note when numbers are disputed: Say a group says it has 5 million followers, but others (specify who and why) say otherwise.

3 Quote two numbers from different sources.

4 Characterize the amount, if the specific number is not necessary. Sometimes the number is not controversial in a story and it is enough to say that the denomination has “more than doubled,” without debating whether it has tripled.