Re: Interesting poll results

Posted by Carol on June 18, 2004 at 11:25:05

In Reply to: Interesting poll results posted by Alan on June 10, 2004 at 11:04:52:

Does this news item have an url? The reason I ask is because the results and conclusions are reported, but the methods are not. You raise a good point about how the term "evangelical Christian" was defined by the pollster. I would also want to know whether a random sample was selected. If so, how was this randomization accomplished? Truly randomized samples taken from the general population are fairly complex and expensive to do.

There's also a frequently overlooked matter called "effect size"--if the sample is large enough, a researcher can find a significant relationship between almost any two variables. Even in detailed scholarly publications, researchers frequently fail to report having done a power analysis to clarify how much the size of the sample influenced the statistical findings.

Nevertheless, the results you have reported are consistent with a number of other studies that have been done on religious affiliation & political views, so I see no reason to question the results unless the issue is how wide was the margin of error and confidence interval (which also are not reported). Still, from all the survey research I've seen reported on this topic, I'm willing to concede that although the figures cited in this particular study seem a bit high, they're still in the ballpark with other, similar studies.

The problem with a lot of social science reporting in the popular press is that the methodology, analytic techniques, and statistical assumptions are rarely reported--we get just the results and conclusions. I read a piece in USA Today just this week about a study that shows evangelical Christian men are good fathers and husbands. The article said nothing about how the study reached those conclusions. Did the researcher use a comparative sample, where evangelical men were compared to devout catholic or observant jewish or practising Buddhist men? Was it specifically the evangelical interpretation of Christian faith or a more generic religiousity that created an association with being good fathers & husbands? Just how large was the sample and the corresponding effects found between the dependent & independent variables? Also, how did the researcher define what is meant by "good" father and husband? None of these methodological issues were reported in the news story.

Nevertheless, I'm not sure I need a social science study to convince me that someone who truly loves God is also a loving father and husband. It's both logical and intuitive that love begets love.