Comment by yetanotherphd | Hacker News Reader

Comment by yetanotherphd | Hacker News Reader

The problem with this article is that the author's field is full of experts who have wildly different opinions. There is no way to appeal to expertise in this situation. For example, he says

> The death of expertise is a rejection not only of knowledge, but of the ways in which we gain knowledge and learn about things. Fundamentally, it’s a rejection of science and rationality, which are the foundations of Western civilization itself. Yes, I said “Western civilization”: that paternalistic, racist, ethnocentric approach to knowledge that created the nuclear bomb, the Edsel, and New Coke, but which also keeps diabetics alive, lands mammoth airliners in the dark, and writes documents like the Charter of the United Nations.

But there are many experts in sociology who would ask us precisely to reject science, rationality and Western Civilization.

I think a far better approach is to educate all people better, so they are able to properly appreciate the knowledge of experts because they understand the field. One aspect of this would be teaching a lot more statistics, although the classical fields of debate and critical reasoning are also important.

Another problem is that many people, especially academics, think it is more important to push a particular agenda, than to make the public more informed in general. Hence people who want to promote gay rights, for example, don't want people to know that the probability getting AIDS from heterosexual intercourse with a White woman is extremely low. But in order to do this, they have to actively lower people's understanding of probability and statistics. They have to teach people to ignore their (admittedly imperfect) intuition about conditional probability.

The funny thing is, you're even more guilty of caring more about pushing a particular agenda than actually informing people. Whilst the probability of getting AIDS from heterosexual intercourse with a white woman is indeed extremely low, what matters is not that the sex is heterosexual but that it's with a white woman. Lesbian intercourse with a white woman is actually safer, whereas heterosexual intercourse with a white man is more dangerous. Of course, saying that wouldn't have helped push your anti-gay-rights agenda.

What's more, the qualifier that it has to be with a white woman is important. While heterosexual vaginal sex has a lower risk of HIV transmission than anal, it's not low enough to provide useful protection on its own. The main reason sex with white women is safe is because few of them have AIDS, which in turn is partly a result of all the high-risk heterosexual communities being very careful to protect against HIV transmission since very early on. If you look at somewhere like Africa where people weren't taught to take those precautions, the HIV infection rate from heterosexual sex is terrifying. Again, though, this contradicts your agenda of portraying HIV as a gay disease that heterosexuals are only taught to protect against in order to promote gay rights at their expense.

You claim that only the sex of the other person involved is relevant, and not whether the intercourse is with someone of the same or opposite sex.

This is wrong, and shows precisely the lack of understanding of conditional probability that I described. I am especially shocked by your use of lesbian intercourse as an example, since everyone knows the issue is men-who-have-sex-with-men. Lesbians are completely irrelevant to this debate.

Let us compare the probability of a man getting AIDS from having sex with a random man, vs a random woman. As you say, these are probably similar. However, it might be very hard for a man to convince another random man to have sex with him! So what we really want to calculate is the probability of getting AIDS from a random, actually occurring, sexual encounter (not with a prostitute) between a man and a woman, and a man and another man. This is what the statistic is intended to mean, and how people will interpret it, even if it is not always properly qualified. In this case, the probability of the man who has sex with a man getting AIDS is much higher, because gay men have a much higher incidence of AIDS than heterosexual women.

So we could add the following useless disclaimer to my earlier claim: if you are a man, sex with a randomly chosen man (including straight men) is not especially dangerous. It's only when you condition on that man being gay that it becomes statistically more dangerous.

Of course, this disclaimer would still be wrong since the sex with the man would probably be anal, which is inherently more likely to spread AIDS.

>What's more, the qualifier that it has to be with a white woman is important. That's why I included it :)

I can't comment on why AIDS is higher for heterosexuals in Africa than the US. I don't necessarily accept your claims about it. But that is not the issue here. I am talking about the probability of getting AIDS given the current situation, not any inherent link between being gay and AIDS. The actual cause is not relevant.

The best that you could say, is that the reason for these misleading statistics is not to promote gay rights, but to scare people into taking precautions that are good for society as a whole.

Reply on news.ycombinator.com