"I did not follow the statement on p. 13 that there was a correlation between SO and sexual harassment r = .18, p = .01, but that this tendency was stronger among male, r..28, p = .002 and female r=.31, p = .002 participants. Doesn't the r value = .18 and p value = .01 derive from the combined correlation for men and women? How does this get stronger for men and women when considered separately? Please clarify."
Our first reaction in reading this was to make sure there was no error in our data. There wasn't. So, now, grab an hourglass and pause a minute to answer the following question: How is this possible?
If we try to draw a line (called a "regression" line) representing this relation, here is how it looks: