No, Bob's hypothesis suggests increased competition, whereas the paper detects no effect of individual performance on group performance (a blow to the hypothesis) and confirms that the gender of the group is irrelevant, it's social sensitivity that matters, so a group of high-sensitivity males would outperform a group of low-SS women, which falsifies the hypothesis pretty well. Women cannot force men to 'be more serious' if there are no women present yet the SS impact is still present.
SS is just confounded with gender because it tends to be higher in women.
what would happen if a similar study were performed in an online environment? what if you had 5 groups, two consisting of all males, two consisting of all females, one consisting of mixed male/female population, one of each of the homogeneous groups half the participants would be instructed to act as members of the opposite sex (might help (in terms of portraying genuine sexual attraction) if they were gay, maybe there should be two more groups where this variable is tested).
This would have an effect if gender dynamics were important in group performance (they may be), but it's separate from this study, which simply demonstrates that social intelligence of group members is significant in determining group performance, moreso than individual intelligence. (In fact, your design is highly confounded by this factor.)
You couldn't tell people to pretend to be more socially intelligent.
Gender is only relevant in this study inasmuch as it is confounded with social intelligence. Your design seems to be focused on the impact of gender structure on the individual performance of each group member, but this study suggests that is less of a factor than a naive speculator would probably think.
