# Size matters hence people lie – self reported vs measured sizes

A while back I had skimmed through a paper by Tatu Westling of the University of Helsinki called “Male Organ and Economic Growth: Does Size Matter”? The paper found a certain correlation between the size of male organ in a country and its GDP and was discussed on Freakonomics. I went and took a look at the data which was used in the paper, which came from here. The data was unfortunately posted only in JPEG format. So I took the .jpg files and passed them through an optical text recognition software which gave me an excel file with some faults I had to repair a bit. This all is just what I needed while waiting for several connecting flights and I was too tired to do anything useful. In the end I compiled a stata datasets which had the following variables:

• country: in non standard textual form
• remark: with two values 1. “self reported” or 2. “measured”
• length: penis length in cm
• girth: penis circumference in cm
• age: average age of male survey participant
• height: average height of male survey participant.

An amusing and, for some, expected discrepancy pops right out at you. The mean self reported penis length is 12.9% larger than that of the measured one while the mean self reported penis girth is 10.7% larger than that of the measured one. The minimum self reported penis length is 40.1 % larger than that of the measured one while for girth that same number is  25.5%.
You can control for other factors a bit but the discrepancies persist. For example there are European countries very close to each other which differ in that same way whenever they are measured on the one hand and self reported on the other. Another thing that supports the suspicion that self reported lengths are biased is the fact that the length and girth correlate much more (`$R^2=.77$`) if measured than when self reported (`$R^2=.42$`).

Lying about such an intimate measurement is in some ways different than other measurements but raises the question of what else people lie about and what does that say about self reported measurements in social sciences e.g. self reported well-being measurements. It would seem that it may be useful to have a theory of lying. I guess such a theory would have to postulate that the probability of lying about something is proportional to the advantage (you believe) the lie will bring you, and inversely proportional to the probability of getting caught.

Push up bras is the survey where women lie when size matters.

This entry was posted in data, economics, research, science, social. Bookmark the permalink.