Friday, December 18, 2015

Same Data, Different Questions

In his recent book, Rodney Stark makes the following observation:
Contrary to stereotypes of Muslims as ardent worshippers, their numbers have been reduced almost as greatly as those for Christians when the data are limited to weekly attenders.
(Rodney Stark, The Triumph of Faith [Wilmington, Delaware: ISI Books, 2015], 15.)
This comes from the way that Stark is framing his question. What he is looking at is, if you took all the people who attend religious services during the week, what percentage of them belong to which religion. If that is the question you are asking then a typical worshiper is more likely to be Christian (39%) than Muslim (31%).

But there is another way at looking at the question. Instead of asking, "What percentage of the world's weekly worshipers belong to various religions?" we could ask, "What percentage of various religions are weekly worshiper?" That is a different question and Stark provides (on pp. 14-15 of his book) the information to answer it. Here in descending order are the percentages of adherents to different religions worldwide who worship weekly:
  • Hinduism     66%
  • Muslims     64%
  • Christians     52%
  • Others     50%
  • Buddhists     28%
  • Jews     24%
  • Secular     2%
This does not invalidate Stark's argument. It is just using the same information to ask a different question. What it shows is that there is a basis for the stereotype, since on any given week almost 2 out of 3 Muslims will attend mosque, whereas just over 1 out of 2 Christians will attend church. That is a statistically significant difference.

Incidentally, I have no idea whether Latter-day Saints would be classified as Christians or Others in this study. I would be curious to know what the specific Latter-day Saint number were, but given the geographic variation that probably exists they would be no particular help to any particular congregation. And, given the magnitude of people we are talking about in the study, whatever the Latter-day Saint numbers are, they would make a negligible difference on Stark's overall numbers.

What really impresses me are the Hindu numbers. So what are Hindus doing right? (Since we do not know what Latter-day Saint numbers are, we do not know what we may or may not be doing right compared to Hindus, but clearly Hindus are doing something right, and so are Muslims.)