Abuse of Statistics
To open a discussion on this article, please use the contact page to provide your comments.
I think I have made my views on same-sex marriage quite clear in earlier articles. I am fully in favour of it. I know of no logical reason why two adult men or two adult women should not be allowed to marry each other if that is their heart's desire. I have no personal interest in the issue, but I do think that it would make the world a better place.
However, there are those who adamantly oppose same-sex marriage. Currently, a group of Republican Senators, with the support of the President, is striving to bring about an amendment to the US constitution to ban same sex-marriage. While this is probably no more than a ploy to lock in the fundamentalist Christian vote for President Bush in the forthcoming election, there is a risk they may actually make advance the issue.
Senator James Inhoff of Oklahoma, speaking in support of the resolution, made the claim that statistics show that children from "intact families" who worship get 30% higher school grades then children from broken homes who do not pray.
This is rather an astonishing claim, and I really question it. The 30% difference in outcome is so large it strongly suggests a study with extreme selection bias - a study designed to produce the desired result.
But, suppose we accept that the study Inhoff referred to was a properly conducted independent study. Does it say anything to support opposition to same-sex marriage?
Of course not! The study is a comparison of successful heterosexual marriage to failed heterosexual marriage. There is no indication that the study considered any children living with same-sex parents, regardless of whether they attend church or not.
The study, such as it is, is completely irrelevant to the issue.
And if you want to stretch this flawed study to make a recommendation about same-sex marriage, it is the total opposite of what Infoff assumes. That is, if a stable churchgoing family is beneficial to children, then religious gays and lesbians should be encouraged to marry. And the non-believing ones can go along for the ride.
- I have little doubt that a properly conducted study would show a statistical difference between the children of stable marriages and of failed marriages. The difference would probably be in the order of 5% -10%. And a close analysis would show that the difference is more related to income levels (stable families tend to be financially better off) and school attended (higher income neighbourhoods tend to have better schools) than it is to marital status.