Laughably stupid statistical error in "The moral urgency of a good study on cold showers and depression & anxiety"
In this post: “The moral urgency of a good study on cold showers and depression & anxiety”. The two paragraphs on this study contain a terrible and frankly sophomoric error. I compared the experimental group at T1 to the experimental group at T2 without cross-comparing the control group at T1 versus the control group at T2 to make sure there wasn’t also a large difference there. I vaguely thought I had done this, but I hadn’t. Perhaps I’d stupidly implicitly assumed there wouldn’t be a large difference between the control group at the two times because it was a non-intervention control group. Overall though, I suspect the mistake was probably born of pure absentmindedness- “remembering” checking and seeing that the control group was unchanged between T1 and T2 when in reality it had changed. The real experimental effect appears to be much smaller, although the result is statistically significant. May God have mercy on my soul.