I'm going to claim that these are important questions for anyone who reports on gender research to ask and to answer. And by report I'm talking to the journalists in the room and the researchers in the room because we're all reporting, just to different audiences.

• What research question was the study designed to address? That determines very clearly what conclusions are valid to draw from the research data.
• How was the sample formed and what population does it represent? In the social sciences we're always studying a sample of people. They're not representative of all boys, all girls, all men, or all women. They're representative of some particular population and unless we know what population they represent, if they even are representative of something, we don't know what conclusions we can infer about that population.
• What was actually measured and are those measures valid and reliable? To say that boys are smarter than girls because they out-perform girls on the SAT assumes that we believe that the SAT is a measure of something called intelligence, which also implies that we believe there is a construct called intelligence that is meaningful. So, what was measured, is it a valid and reliable measure of something meaningful that we're trying to talk about?
• Do the conclusions clearly follow from a chain of evidence?
• Are the differences reported meaningful and important? We can find lots of small differences about things that just don't matter. We can find lots of small differences that are very, very important and that we need to know about.

We need to really answer these questions to really think about these questions in looking at any studies of gender because, after all, studies of gender or of race or of any other differences within social engagement in the scientific enterprise is very highly politically charged and very, very important and we need to be very critical on each of these issues.