4 Comments

Did they test more hypotheses, and did they correct for having multiple hypotheses?

Expand full comment

This is a joke?:the composite score had a significant correlation with conception risk across the menstrual cycle, r = .45, p < .0001.

I think so. Composites that are made from data that measure the same thing will always inflate correlation. "Race bias in mate attraction", "fear of out-group males" and "explicit race bias" are all virtually the same thing.

Other than that, random notes as I was glancing over the paper:

77 White college undergrads taking psychology were tested for discrimination between White and Black males. Probability of conception data obtained for a group of women with mean age 30 was applied to this group of 18-22 years old girls. There is not a single word in the paper about statistical treatment of the data, there is not a single work on how Fig.1 and Fig.2 were generated and as far as I am concerned, neither Fig.1 nor Fig.2 show actual data or give a slightest idea of how the actual data looked like. (I.e., how are these 77 points distributed across 35 days to generate the red line shown?)

There is no internal control included that would give one an idea about intrinsic stability and trustworthiness of the data.

The text says r=0.45 but Fig.1 looks like the correlation is nearly perfect - should be at least 0.9.

Also, according to Fig.1, at lest half of their lives these girls spend completely without anti-Black bias or with a strong pro-Black bias. Averaged over time, they seem to be completely devoid of racial bias. I find it almost impossible to believe and it certainly contradicts a lot of well-known data.

I think the study actually measured girls' mood swings around ovulation time, that's all.

Expand full comment

This is a joke?:

The relationship between conception risk and explicit race bias was only marginally significant, p = .08. When merged into a composite race-bias variable (standardized by variable, then averaged), the composite score had a significant correlation with conception risk across the menstrual cycle, r = .45, p < .0001.

Expand full comment

To find a fertile woman to impregnate, look for a racist

The original paper is here: http://www.cdnresearch.net/...

Statistically speaking, it looks like a joke to me.

Expand full comment