Experts say religion study is sound, but ...

Time and again, critics of paleontologist Gregory Paul’s controversial research suggesting a correlation between religious belief and social dysfunction come back to his numbers: They don’t add up, these critics say, because he doesn’t understand statistics.

MSNBC.com asked three statistical and assessment experts to review the methodology Paul used in compiling his paper, which is published in the current issue of The Journal of Religion and Society under the title “Cross-National Correlations of Quantifiable Societal Health with Popular Religiosity and Secularism in the Prosperous Democracies.”

By and large, these experts say, the critics are wrong. Paul’s study holds up. But they also say the critics have a point.

Strictly speaking — and that is the only way academics speak — Paul’s study marshals reasonable data that credibly demonstrates, under its terms, that the United States is the most faithful of the 18 Western democracies it studies. It furthermore demonstrates that it scores at or near the bottom of a variety of indicators of social dysfunction, including homicide, juvenile and early adult mortality, teen pregnancy and abortion.

Correlation doesn’t equal causation
All the paper is saying, “in a variety of ways, is that the U.S. is different from other First World democracies in several respects,” said Dr. Joseph B. Kadane, University Professor of statistics at Carnegie Mellon University in Pittsburgh, who has been a consultant to the Justice Department.

That’s a crucial point, the experts said. All Paul is highlighting is a correlation. Nowhere does he explicitly argue that the United States’ higher rate of religiosity is a cause of the higher rate of social problems, the sin of which he is most commonly accused.

“The main caveat — and the author himself is clear on this — is that correlation is not causation,” said Dr. Larry Wasserman, a colleague of Kadane’s at Carnegie Mellon and winner of the 2005 Outstanding Statistical Application Award from the American Statistical Association. “His research suggests a correlation between religiosity and certain social ills, but he leaves open the question of whether religiosity causes these problems.”

On those terms, then, Paul’s paper “is basically sound,” Wasserman said in an e-mail analysis.

Reading between the lines
Kadane agreed, but he observed that, even though Paul draws no direct link, he plainly wouldn’t be upset if you drew it yourself. “Certainly, some of the discussion suggests that this author has one direction of causation in mind,” he said.

That appears to be the complaint of George H. Gallup Jr., the eminent pollster, who fired off a letter to the editor of The Times of London after it highlighted Paul’s paper in September.

Gallup, whose deep Christian faith is widely known, declined to talk further about his criticism, saying he was publishing a full explanation in the coming edition of the religion magazine Touchstone. But his research associate referred MSNBC.com to Tim Cupery, a sociologist of religion at the University of North Carolina, whose separate letter to The Times mirrored Gallup’s thoughts, the associate said.

“The main point of both the scholarly and journalistic articles seems to be a critique of Christianity as detrimental to the social health of the United States, a conclusion that the data do not warrant,” said the letter, which was never published.

Otherwise, Cupery said later in an interview, “why waste good research on this?”

MORE FROM U.S. NEWS

U.S. news Section Front

Deep freeze sets in; more snow on radar Accepting peace prize, Obama defends war ‘Tea party’ movement preps for 2010 vote SEC probes 9 deals for insider trading: report   Judge: Pentagon in contempt in Gitmo case   Pakistani militants rebuffed U.S. volunteers What Senate health bill means for consumers NYT: Salahis rattle D.C.'s ‘sense of decorum’   NYC cop kills man in Times Square shootout Search on for snowbound hunters in Arizona U.S. news Section Front   Add U.S. news headlines to your news reader: