Tag Archives: statistics

Dick Size Wars in a New Context?

Debbie says:

Laurie and I are contemplating a serious post on masculinity within the next week, but meanwhile, I found

this completely irresistible:

A researcher in Finland has uncovered a strong negative correlation between average penis size of the men in a country and that country’s economic growth between 1960 and 1985. Yes, that’s right: if the men have smaller dicks, the country is doing better.

chart of penis sizes and economic growth measures

If you don’t know how to read charts like this one (called “scatterplots”), the dots represent average penis sizes in various countries (we’ll get to what data he used) and the line represents an average for all the dots on one vertical line. So as the penis sizes go up from left to right, the economic growth measures go down, which is why the line moves downward from left to right. If the correlation was “positive,” meaning that larger penis sizes linked to greater economic growth, the line would move up from left to right.

The first thing any statistics course will teach you is that correlation is not causation. I might be able to draw a correlation between number of households in my city (Oakland) who have chickens in their back yard and number of rainstorms this year, and even if I got a clearcut measure like this one, it wouldn’t mean that chickens cause or prevent rain. So the whole thing is pretty silly from the start. This is how the Freakonomics guys do their work, and why it doesn’t mean much.

That being said, it isn’t quite junk science. The research was done by Tatu Westling at the University of Helsinki. My first concern was that the whole thing was just a joke, which it doesn’t appear to be. My second concern was that he had measurements on two or three penises per country (each dot on the chart is a country, so I could see that he was claiming to represent a lot of countries). It turns out that he used a well-known and reputable data set (though I have no idea why that data set includes penis size) that covers 121 countries. His economic data also looks respectable to me.

Westling obviously has a sense of humor. The paper is worth reading for lines like “”the statistical endurance of the male organ is also found very formidable” and “Taken at face value the findings suggest that the “male organ hypothesis” presented here is quite penetrating an argument.” Not to mention the paper’s last line: “It does seem like the ‘private sector’ deserves more credit for economic development than is typically acknowledged.” He makes some not-very-convincing attempts to come up with potential reasons for his findings–but since it’s almost certainly a coincidental correlation, they feel like window-dressing, not a serious attempt to draw conclusions.

For myself, I’m just charmed by the idea of men all over the world whipping out their dicks and worrying about whether or not they’re too big. It would certainly be more amusing (and possibly more fruitful) than the dick-size wars going on in Washington and Brussels right now.

Objectification Theory: Does Being Ogled Make Us Less Smart?

Laurie and Debbie say:

When a blatantly silly or misogynist evolutionary psychology study turns out to have a sample size of 16 or 30, it’s easy to discard not only the study but the whole idea. For once, however, we’ve run into a study with too small a sample size (25 people) that we would love to see done on a statistically useful scale:

Study participants — 25 women ages 18 to 35 — were told they were recruited to provide information on “the impressions people form about others solely based on their carriage and style of dress.”

Each was videotaped for two minutes — first from the front, then from behind — while they walked up and down a hall. To capture the experience of having their bodies evaluated while their faces (which presumably provide a better reflection of their individual personalities) were ignored, they were filmed exclusively from the neck down.

For half the participants, the person doing the filming was male; for the other half, the camera was held by a woman. After the filming, each woman watched her video, reinforcing the experience in her mind. She then filled out questionnaires measuring her levels of Trait Self-Objectification (her overall propensity to view herself through the lens of others) and State Self-Objectification (her tendency to view herself through the lens of others when triggered by a specific event, such as being stared at).

To test their cognitive skills, the women were shown a series of random letters or numbers and instructed to reorder them (putting them in alphabetical order for the letters, in ascending order for the numbers). They completed 21 such tasks, which were presented in increasing order of difficulty.

The results: When women with a tendency toward viewing themselves through the lens of others were placed in a situation where they were objectified (that is, they were videotaped by a man), they made a greater number of mistakes on the cognitive test. They did just as well as other women on the easy initial tasks, but had trouble when the difficulty level went up.

We like the idea of using the objectification scales to analyze the cognitive data, because the underlying assumption is (gasp!) that women are not all the same and don’t all have the same reactions. We like the study design, though we’d look for see more analysis (rather than assumptions) about the difference between being filmed by a man and filmed by a woman. We would, of course, also look for an acknowledgment that gender is not binary … though we’re not likely to get that from this kind of work.

We both have the gut sense that the results could easily be accurate, that a full-scale study could bear this one out.

Here’s the reason we can’t believe it yet: with 25 women being studied, only 20% (that’s five women) “have a strong propensity toward self-objectification.” Actually, if that result were to be borne out by a larger study, that would be great news! If only 20% of women are high on the self-objectification scale, something good is happening. At the same time, this means that all the results in the article are based on five subjects’ results on one test. This isn’t even remotely enough to be useful.

According to the article, the researchers are recommending “a campaign of awareness and education.” We recommend some repeatable large-scale studies, and if the results repeat, then we’ll jump on the awareness and education bandwagon with the greatest of pleasure.

Thanks to Firecat for the pointer.