There was a time when I might have speculated on the cause but then I learned that rather than continuously update my priors I should instead consider every piece of evidence in isolation and adhere strictly to the absolutely non-arbitrary standard of a 95% confidence interval.