In this section I turned the question around from the previous section: Is the age difference important for a woman, and is the age difference between an older woman and her partner different than it is for a younger woman and her partner. As with the men, first I segmented the women into 5-year groups and calculated the mean age difference for each group and their respective partners. This is the resulting graph:
Notice that the mean age difference declines as the woman gets older. Performing an ANOVA on this reveals that overall the mean age difference is different for different age groups of women, F(7,181)=2.7, p<.02. But when I looked at just the five middle age groups of 21 to 45, these groups are not statistically different, F(4,164)=.56, p>.5. This means that for the majority of women (169 of 189) the average age difference of their partner was 12-14 years, which is not different statistically.
As with the men I performed a regression analysis. Although the regression showed a significant linear relationship between the woman's age and her partner, F(1,187)=14.8, p<.0005, it accounted for less than 7 percent of the variance in her partner's age (adjusted R2 =.068). The result of regressing age difference on the woman's age is shown graphically below.
In plain English what these graphs show is that the woman's age is not a very good predictor of how old her partner will be. The woman's age alone accounts for less than 7 percent of the variance (a measure of variation) in the age of her partner. This is a case of a factor being statistically significant but almost meaningless in practical terms. (Note: As with men, many factors affect a woman's selection of partner. The only raw data I had was the man's and woman's age, and based on this information alone the man's age appears to be not very important. Of course if I had more complete data this might change.)