First, the difference for each gender respectively of the scores across the control and "affirmation" groups isn't what they care about. It is the difference between gender within each group.

I can't say whether the results are statistically significant, but there are error bars on the graphs and the scientists who did the experiment seem to think the results are significant.