Wednesday, February 01, 2012

Mailvox: statistical illiteracy

dh appears to find it hard to distinguish between "X" and "virtually X":
This will be difficult for VD to walk back in the future when he wants to try to hold his various intellectual foes to account for words they use. "Virtually Impossible" = "something that happens every day". Got it.
First, as I have had to point out many times before, the purpose of an adjective or an adverb is to modify a word. Therefore, something that is "virtually impossible" is, by definition, not impossible. It is merely highly improbable. Does the state senator's comment qualify?

As I pointed out in the comments, the Center for Disease Control reported that an average of 473 white men were annually infected with HIV through heterosexual contact over the course of a four-year statistical study. Now, dh had already admitted that the chances of being infected for anyone, male or female, having heterosexual relations with an infected individual, was 30 in 10,000, or one in 333. Is that sufficiently improbable to qualify as "virtually impossible"? No, I don't think so. However, we're not done yet.

However, this does not distinguish between men and women, which is necessary because it is easier to transmit the virus from male to female than from female to male. That is why an average of 841 white women annually contracted the virus through heterosexual contact in the same study. This means that men are estimated to contract the virus at a rate that is about 60 percent of the female rate, thereby lowering the one in 333 figure to around one in 500.

Still not "virtually impossible" in my book, but we're getting closer. Now we have to take into account the fact that these one in 500 odds only apply to sex with an infected female. So, we have to return to the CDC, which tells us that there are 682,668 Americans presently living with HIV, 26,966 of whom are white females. This represents one in 3,649 of the 98,408,776 white females in America. Note that it is appropriate to include the entire female population here because children are also included in the AIDS statistics that we are citing due to mother-to-child transmission.

Since it isn't possible to contract the virus from someone who doesn't have it, this means that a white man who has sex with a white woman has approximately a 1 in 1,824,682 chance of contracting the virus. This compares rather favorably with the 1 in 58,618 chance of being legally executed, the 1 in 147,717 chance of dying in a dog attack, and the 1 in 615,488 chance of dying in a fireworks discharge.

In other words, for a normal white man like Stacey Campfield, it is, in statistical fact, highly improbable, or if you prefer, "virtually impossible", to acquire the HIV virus, which of course is necessary in order to spread it.

And before anyone starts complaining that white-on-white heterosexual relations doesn't encompass everyone, I will note that the entire context in which the statement was made concerned the significant difference in practical risk factors among different population demographics. Campfield's point, and it was entirely correct, is that it is totally absurd to discuss HIV infection as if it presents a similar risk to everyone, regardless of their race, sex, or sexual behavior, which of course is the main reason the mainstream media's dire predictions of a heterosexual AIDS epidemic, still less "a national disaster as great as a thermonuclear war", were so wildly incorrect.

And in conclusion, I will point out that in a nation of 310 million people, a "virtually impossible", one in 1.9 million event is indeed very likely to happen every single day. Or, as is the case here, about 1.3 times per day.



Post a Comment

Rules of the blog

<< Home

Newer Posts Older Posts