In the online conflict between certain specific men’s rights activists (MRAs) and certain specific fringe radical feminists, no part have their intellectual integrity fully intact. Elements on both sides misunderstands statistics and basic biology and they also appeal to pseudoscience to justify their political ideology. For example, some fringe radical feminists, particularly those with postmodern inclination, may devalue biological partial explanations or subscribe to the notion of a blank slate. Similarily, some men’s rights activists propose superficially plausible evolutionary accounts of the origin of gender roles from the 1940s, but in the end, cannot provide any scientific evidence for their highly speculative accounts (making them a just-so-story).
I think that gender equality is a moral necessity, but it is terribly tedious and tiresome to read the spiraling conflict between extremists, especially when they make embarrassing rookie mistakes.
One such mistake is made by the blogger Phil in Utah over at the A Voice for Men blog in a blog post about rape statistics that confuses life-time prevalence with incidence. Phil in Utah writes:
Statistic: “1 out of every 4 women will be raped in her lifetime.”
Truth: Ah, here’s the doozy. I’m sure we’re all familiar with the source of this statistic: a study by Mary Koss that has been discredited countless times. Around three-quarters of the women she identified as having been raped did not consider themselves victims of rape, and almost half of them had sex with their supposed attackers after the event identified as a rape had occurred.
I do not really know enough about the Mary Koss study to make an informed argument, but surely, rape has to be defined as objective as possible and not solely be based on personal opinion? So the argument that some of the women did not consider it rape, therefore it should not be counted as rape, seems wrong. Obviously you can be subjected to a crime even though you are not aware that it is considered a crime. A rose by any other name…
Let us look at some rape statistics from the CDC. In their National Intimate Partner and Sexual Violence Survey (NIPSV), they reached the following conclusion with regards to rape prevalence among women. From the executive summary:
Nearly 1 in 5 women (18.3%) and 1 in 71 men (1.4%) in the United States have been raped at some time in their lives, including completed forced penetration, attempted forced penetration, or alcohol/drug facilitated completed penetration.
Now, rape prevalence will differ depending on how inclusive the definition of rape is (varies between countries), but according to the definitions used by the CDC, it is around 18% of women in the U. S. Although not exactly 1 in 4 (25%), it is fairly close. With that in mind, let’s see how Phil in Utah tackles U. S. rape prevalence:
So, what do statistics collected from non-feminist sources say? Well, let’s try the FBI statistics. According to an FBI report, which did not account for differing definitions of rape, whether or not the rapes were convicted, or whether or not female-on-male rape was included, the United States had a rate of 29 reported rapes per 100,000 people in 2009. That’s not going to get us to 25%, but I’m feeling generous, so let’s look at the country with the highest rate of rape in the past decade–South Africa, with a rate of 116 rapes per 100,000 people in one year. Percentage wise, this is .1% of the population. Now, I’ll admit that I’m worse at math than anything else in the world, but even I know this isn’t even close to “1 in 4″.
Here is the problem with that argument: the CDC looked at rape prevalence (actually, life-time prevalence, see below), whereas the data cited from the FBI report (and the South Africa argument) looked at rape incidence. This seems like a minor quibble, but it is of utmost importance. Here is why.
Prevalence and incidence are two key concepts within epidemiology and medical statistics. Let us look at how these two concepts work when it comes to, say, HIV. The prevalence of HIV tells you the total number of HIV cases (usually in % of total population). The incidence of HIV tells you the number of new cases of HIV arising during a certain time period. As we can see, it is really important to not confuse these different metrics. For HIV prevention, we want to focus on getting the incidence down. Getting the incidence down is one way to prevent the prevalence from rising, but obviously we do not want HIV patients to die, so we cannot just focus on trying to get the prevalence of HIV down with any means necessary.
As an added complication, life-time prevalence of rape looks at what proportion of the population that has been raped sometime in their life. This means that we must separate discussions of the number of new cases of rape per time unit (rape incidence) with discussions about the proportion of the population that has been raped some time in their life (life-time prevalence of rape). Technically it is the proportion that has been raped some time during their life up to the time period the study is being carried out. Obviously, scientists cannot draw on non-existent, future statistics.
So to sum up, the reason that the argument laid out by Phil in Utah fails is that it confuses prevalence with incidence. Phil in Utah seems to understand this when making the following rebuttal to a perceived counterargument:
“But wait!” the feminists are saying, “Most rapes are never reported to the police!” Well, I’ve heard a number of different figures on just how many. Some say 45%, some say 60%, and some even say 80%. But hey, I’m feeling EXTREMELY generous, so despite the fact that feminists are basing these numbers off evidence that is dubious at best, I’ll go with the highest estimate. .1 times 5 is…half of one percent. In other words, one-fiftieth of what feminists claim it is.
Now, I hear them whining that I missed the key phrase “In their lifetime”. Okay, since empirical data shows that rates of rape drastically decrease after the victim turns 45, whether they are male or female, in prison or out, I’ll just be accounting for a 30-year window. Sorry, feminists, but even my generosity has its limits. I’m not going to pretend that the wackos who rape grannies aren’t extreme outliers. This means that 15% of South African women will be raped in their lifetimes. A grisly figure to be sure, but then again, this is South Africa we’re talking about–it has the second-highest crime rate in the world. The rate of rape in the U.S. is one-quarter of that, so in our most generous of moods, it is correct to say that 3.75% of women will be raped in their lifetimes. I’m puzzled as to how that can be mistaken for 1 in 4.
Sadly, Phil in Utah falls prey to the same problem. He is making the argument that since the rape incidence in U. S. is 1/4 of the rape incidence in South Africa, the rape prevalence in the U. S. has to be 1/4 of the rape prevalence in South Africa. This is of course wrong, as there is nothing that says that changes rape incidence in the U. S. follows change in rape incidence in South Africa. The latter has been strongly politically instable for decades, so that is a clear confounder.
The life-time rape prevalence in South Africa seems to be calculated by Phil in Utah as summing up the rape incidence in South Africa (together with a modifier for rapes that go unreported) over 30 years (0.5*30 = 15). However, this is not the same as life-time rape prevalence, which was, as we saw, what proportion of the population that has been raped sometime in their life.
Life-time rape prevalence for women in the U. S. is around 18% according to the CDC.
Phil in Utah makes the following statistical errors in his reasoning: (1) confuse (life-time) prevalence with incidence, (2) confuse (life-time) prevalence with the sum of incidence over 30 years and (3) thinks that 1/4 of the incidence implies 1/4 of the prevalence (really a version of the first statistical error).
To make convincing statistical arguments, one should preferably understand the relevant statistics. Otherwise, one risks making embarrassing statistical errors.