Sunday, November 22, 2009

Support for science and sex-related attitudes and behaviors

You guys have got to make sure you read comments that go along with posts; otherwise, you miss people like Jason Malloy. I want to summarize his analyses and comments in the post on the relationship between a Darwinian outlook and fertility.

First, belief in evolution is associated with less fertility, independent of theological positions. Second, he reminded readers of my finding that acceptance of evolution predicts a pro-abortion position after controlling for atheism and liberalism.

Third, Jason looked at the predictive power of believing that scientists always seem to be prying into things that they really ought to stay out of. It looks, however, like he is comparing the size of logistic regression coefficients. They are unstandardized estimates and so reflect the metrics of the independent variables. To double check, I'll estimate standardized OLS regression coefficients. (In the case of being in favor of abortion for any reason, I know that I'm violating the assumption of a normally distributed dependent variable, but a statistician friend assures me I can get away with it as long as the skew is no more than 75/25.)

The table shows that believing scientists do not pry is associated with a pro-choice position, but the prediction is not stronger than either political views or belief in God. Although the sign of the coefficient for being pro-scientist is in the predicted direction for family size, the relationship is not significant--an N of only 227 doesn't help (I limited the sample to those ages 45-59 for the fertility analysis). Finally, while political orientation and atheism predict number of sexual partners, a pro-science stance does not.

Like Jason's findings, these show some connection between supporting science and having liberal sexual views and behaviors, but results are not as striking or consistent as suggested in the comments of the earlier post. Jason did much more, but I'll have to look at that later.

1 comment:

  1. The similar question HARMGOOD ("does modern science do more good than harm?") has almost 6500 respondents, and, unlike the other questions, allows for an adequate sample size of those at the end of their reproductive life.

    I used those 45 and over (N = 2860), since 99% of people have finished reproducing by that age. The standardized regression coefficients indicate science support surpasses religion:

    Liberalism -.007
    Church Attendance .066**
    Science Beneficial -.097**

    However, adding DEGREE to the mix suggests education is the mediating variable.

    Liberalism -.007
    Church Attendance .082**
    Education -.160**
    Science Beneficial -.053

    HARMGOOD has independent effects on views like abortion and homosexuality, etc, but plays a smaller role than education, church, and politics.


Meta-analysis of clinical trials: Eat walnuts

I am always looking for easy eating choices that are good for you. This new meta-analysis of 26 clinical trials looked to see if walnuts ma...