calcSD

The Veale Study

This page details the errors that were made in the original Veale et al. 2015 meta-study published by David Veale and his colleagues, and it presents our corrected versions. Mostly because for whatever reason he hasn't published a correction himself.

Summary

The Veale study can be misleading and has numerous errors. Firstly, a lot of people believe that it measured 15-thousand men to yield its erect results, but what Veale says is that up to about 15-thousand different men where measured in total across all the studies. In actuality, upon correcting all the sample size errors: stretched flaccid length has nearly 14-thousand people total, normal flaccid length has about 10-thousand people, and for erect length and girth there were only 432 and 121 people, respectively. That's certainly not 15-thousand!

Furthermore, Veale et al. 2015 mixes bone-pressed and non bone-pressed studies despite explicitly stating:

"Flaccid or erect length was measured from the root (pubo-penile junction) of the penis to the tip of the glans (meatus) on the dorsal surface, where the pre-pubic fat pad was pushed to the bone.

Flaccid stretched length was measured as above while maximally extending the penis."

Basically, bone-pressed (BP) length measurements are pressed against the pubic bone, pushing in the fat pad. Non bone-pressed (NBP) studies don't do this, instead they measure length in front of the uncompressed fat pad. For example, person A has 6" of BP length and a 1" fat pad, while person B has 7" of BP length but 3" of fat pad. This means that the NBP measurements will be 5" for person A and 4" for person B. When comparing using NBP length, the only thing that made person B smaller than person A, is his larger fat pad. BP length controls for fat pad differences by approximating a scenario in which both were to lose their excess weight, allowing the true difference of penis size to be observed. There are reasons to prefer BP or NBP, but I think we can all agree that these two types of measurements should not be mixed together, and if mixed, should especially not be misclaimed to be all BP.

Erect Length:

One of the studies used by Veale is Wessells et al. 1996:

Wessels provides both NBP and BP erect length measurements, but Veale remarkably chooses to use the NBP results and ignore the BP results:

Of the three remaining studies providing erect length measures, only one of them (Schneider et al. 2001) actually measured BP. The other two (Promodu et al. 2007 and Sengezer et al. 2002) were both measured NBP. This issue of mixing NBP and BP studies is prevailent throughout all the length measures in Veale's study.

*Furthermore with Sengezer et al. 2002 Veale uses impossibly tiny ~0.1cm SDs due to the study misreporting SD (they are likely reporting standard error or SD as percentage of mean), this one study is given almost 30% of the total weight in the erect length category pulling the erect length SD down.

Erect Girth:

The erect girth is solely from two studies (Promodu et al. 2007 and Wessells et al. 1996). Even more remarkably if Veale had gotten the sample size of promodu correct, he would realize that the total number of people measured for erect girth is only 121 men!

Other Error Examples:

Ajmani et al. 1985: Veale uses the ±0.94mm and ±0.02mm reported in the abstract as the respective SDs of flaccid length and girth, despite: neither being stated to be SD, both being in the wrong unit (Veale did not convert them to cm), and both contradicting the actual respective SDs reported in the data tables, which yield 1.50cm and 0.97cm. Veale's SD for this study was only 48.5 times smaller than actual flaccid girth SD.

Promodu et al. 2007: Measured the flaccid length (normal and stretched) and flaccid girth of 301 people (Group 1). Then, out of these 301, it has the self-reported erect length/girth of only 93 (Group 2) of which only 41 were verified by researcher measurement (Group 3). Veale uses numbers from Group 3 for Erect Length/Girth and Group 1 for Flaccid Length/Girth, but reports the total amount of people involved as 301 for all of them. This in of itself isn't that bad, but when merging multiple averages, Veale seeks to weight them with the amount of people involved. In this case, Promodu ends up with more weight than it should, further deviating the numbers.

Savoie et al. 2003: Veale uses a sample size of 124, but there are only 63 people recorded in the study out of the initial sample (incorrect weighting). Also this study is NBP and Veale incorrectly uses stretched length 17.5 cm despite such a value never occuring in that study and the correct mean stretched length the study reports being 13.5 cm.

To be brief, there are about 20 different errors in Veale's study (not including those of combining NBP with BP). Which is about one error per study... But serendipitously, only the erect measures are appreciably deviated due to most of the errors being nullified by other studies with much higher sample sizes or due to errors cancelling out with each other.

Veale et al. 2015

The following is a comparison between the original published data in Veale's study (errors included) and a corrected version which still mixes BP and NBP, still uses sample size weighted averages and pooled SD, and does everything the same as Veale et al. 2015, but corrects the data collection errors using the original source studies:

Veale+

In the interest of having an error-free and properly BP measured version of the Veale study, I've also created Veale+, a fully corrected version that attempts to properly separate BP and NBP studies. Everything else is the same, but my version cuts the mean and pooled SD weighting halfway towards equal weighting to each study to prevent the averages from being excessively dictated by the potential biases of a few high sample size studies, as with most of my averages.

Last updated: Nov. 10, 2020