The value of the article is not in its statistical likelihood, but rather, in recognizing how his assumptions blinded him. And actually listening to people.
Because he expresses uncertainty in more than one of his take-aways. He uses stronger language when he has more confidence in the idea.
"This may have..." or "I believe" or "this seems" all suggest he isn't stating fact and is instead thinking out loud about the possible implications of his experience.
In other words: it's obvious the author knows he is reflecting on anecdotes.
"The fact that one girl exhibited preferences counter to his assumptions does not mean that those assumption are incorrect." <-- is exactly how you do not do customer development. You have stopped listening to the user.
You're focusing on having things being driven by data. That is commendable. However, that only tells you a part of the story.
There's a big blind spot of people who depend on analyzing things like this. And that is, they stop listening to other people. It is usually because such people are consumed with being right, or even trying to prove themselves right. The skeptical stance gets distorted into requiring other people to challenge your assumptions. Sometimes, it is a covert (as in Jungian shadow covert) way of feeling good about being contemptuous of other people.
To really listen to other people, you give up the notion that it is all about you. In other words, it's better to assume that you don't know what you are are talking about. You don't have any preconceived notions blocking you from hearing someone out. You are not "listening" in the sense that you are waiting to for the other person to stop talking so you can tell them how they are wrong. You might even learn something.
This is exactly what customer development is about: listening to other people.
In any case, I doubt this will persuade you. That's the cool thing about this. Some people don't want to consider this, so it becomes an unfair advantage for the folks who do :-)
