Can women’s psychic powers affect randomisation?

February 2, 2011 § Leave a comment

Chi and Snyder recruited sixty right-handed Sydney students to perform a problem-solving task. Some of them had “transcranial direct current stimulation” during the task (either right-brain positive or left-brain positive), while others had “sham” stimulation. In the right-positive group, 12 of 20 solved the problem with the time limit, compared to 5 of 20 in the left-positive group and 4 of 20 in the sham group. A Fisher’s exact test comparing the right-positive and placebo groups gives P = 0.022. Cool result.

Look at this table though:

The randomisation is very gender-unbalanced, with a Fisher’s exact test P-value of only 0.013! This doesn’t cast doubt on the authors’ results: there was little difference between women’s and men’s performance overall: 11 out of 30 women solved the problem compared to 10 out of 30 men (though note Simpson’s paradox possibilities. Also, I would be a bit worried that the study seems to have been single-blind rather than double-blind). But the gender counts are interesting in themselves. What seems to be happening is that women have more psychic powers than men. They want to do well on the test, so they subconsciously arrange to put themselves into to the group that has the best shot at solving the problem. There’s no other rational explanation, right?

(paper via the Guardian)

Where Am I?

You are currently browsing the cheap shots category at "But it's under .05!".