I can't believe that I missed this study reported by BPS research last January. Way cool. It compared the performance of men working in pairs to women working pairs. The researchers placed them under performance pressure, and varied whether they drank caffeinated or decaffeinated coffee. The "caffeinated" men performed worse, while the women performed better. Here is the opening paragraph from BPS, which suggested that the stimulant has these varying effects because, when cranked-up physiologically, people tend towered their most natural and well-rehearsed behavior — which means that men get more aggressive and women become more collaborative:
If a meeting becomes stressful, does it help, or make things worse, if team members drink lots of coffee? A study by Lindsay St. Claire and colleagues that set out to answer this question has uncovered an unexpected sex difference. For two men collaborating or negotiating under stressful circumstances, caffeine consumption was bad news, undermining their performance and confidence. By contrast, for pairs of women, drinking caffeine often had a beneficial effect on these same factors. The researchers can't be sure, but they think the differential effect of caffeine on men and women may have to do with the fact that women tend to respond to stress in a collaborative, mutually protective style (known as 'tend and befriend') whereas men usually exhibit a fight or flight response.
Clearly, this is a "more research is needed" situation. But, if it generalizes to real life, the implication is that, if you are running a meeting and it is attended by all women, give them caffeinated drinks, but if it is all men, or perhaps a blend of men and women, given them the decaf if you want cooperation and better performance.
Here is the reference:
St. Claire, L., Hayward, R., and Rogers, P. (2010). Interactive Effects of Caffeine Consumption and Stressful Circumstances on Components of Stress: Caffeine Makes Men Less, But Women More Effective as Partners Under Stress. Journal of Applied Social Psychology, 40 (12), 3106-3129 DOI: