@ LoganRuckman - Thanks for reply, Logan. I am watching the Kentucky Derby, and typed out a monologue. Bear with me.
First, jokes are fine, but since all we have are typed words, its easy to read the comment as trying to redirect the conversation to some other controversial topic instead of resolving the conversation.
As for anecdotal evidence, yes, its one of many ways we get to know thing. If you dislike the taste of bourbon, you probably won't like mint juleps. They are very limited though. Anecdotal evidence is just your experience, and your recall of those experiences, and it can always be rebutted by someone else's contrary anecdote. Worse, it can filter incoming information, like if you come off a bad relationship, and are wary of [men/women] for a time.
Other ways are dramatic examples that make headlines, but spark reactions far in excess of what is warranted. For instance in the United States, about 35 thousand people die each year in car accidents, but it gets little attention compared to the relatively few deaths by something like terrorism. Not that terrorism is not real or horrible, but maybe we would react stronger to things like texting while driving.
Better ways still are thorough case studies, focus groups, or panels of experts. Those are good for many things, like clinical trials or marketing research, and may be useful here too. They can have serious issues though. For instance, a few years ago there was a sniper in the Washington DC area. The experts came up with lots of characteristics of the sniper, and were completely wrong on many accounts, so wrong that the actual snipers were stopped at checkpoints by police but let go because they did not meet the profile.
A fine way, perhaps the best way, to know things about a large group of things is with scientifically designed (read: statistical) design of experiments or surveys (or a complete census). Since we don't run experiments on people (usually), we are left observing them. That is still fraught with peril - how were people selected, how many responded, how was the information reviewed or adjusted, what questions were asked, in what order, how was that information collected, what about people hard to find with that method, etc. In this case, one of the big questions is how were the key terms (in particular 'gamer') defined. That's vital. Whatever the definition, if its reasonable, that's fine, but if there is another reasonable definition that would result in different results, that makes it important.
As an example, the National Crime Victimization Survey, a very professional US governmental survey, estimates 3.3 million people in the US were stalked in a recent year. You can find it here
http://www.bjs.gov/content/pub/pdf/svus_rev.pdf complete with details on the methodology. However, that number is so *huge* you should not believe it without asking questions. They define stalking very broadly, and count someone as being stalked by saying if you have experienced several of a set of experiences, you have been stalked - not that you self identify as having been stalked. (in fact, only about half of those the study says were stalked agree they were stalked) They still come up with a very small percentage of population, which is part of the problem, but multiply a small percentage by the population of the US, and you get a highly suspect number.
My long winded point (sorry for that) is that even a very thoroughly described, scientific, reputable survey can produce questionable results, so why in the world give credence to an a poorly described survey by a lobbying organization with an axe to grind. Anecdotes make for fine friendly conversation, and I'm happy to talk about female gamer we know, but its not moving the discussion forward on how numerous or important those female gamer are to the hobby and the industry. What would be nice if it companies divulged some of their proprietary information, since they probably have information about that as it affects their profits.