There are several reasons that the statistics at the start of this thread might be wrong.
First the sample set was restricted to a group of presumably young men on college break. The might be presumed to have higher testosterone levels than older men, and better erection quality therefore leading to larger numbers. But, they were recruited for measurement at a night club at which presumably at least some of them might have been drinking heavily, leading to weaker erections and lower numbers. And out of 401 guys who agreed to participate, 101 were not included because they allegedly couldn't get an erection. WTF, over 1/4 of young guys unable to get it up? Again, maybe EtOH was a factor, but it is not unlikely that the less well-endowed used this as an excuse to avoid embarrassment in front of one or more female nurses and a doctor after peer pressure prompted them to volunteer along with their buddies when approached in the nightclub. If so, the sample set is influenced profoundly by selection bias, beyond the narrow age range.
Second, the methodology was never clearly spelled out in this study. The study was conducted by a condom manufacturer for the purposes of determining optimum condom sizing. For a condom manufacturer, one would think NBPEL is of greater significance than BPEL. But we know NBPEL can be notoriously difficult to measure consistently. We don't know if the length data refers to BPEL, NBPEL. or something in between. People have argued which it is, but we don't know. Comparing this data set with other data sets that contain both length and girth data, it seems to me that the average length in this study when compared to the average girth would be more appropriate for something in between BPEL and NBPEL than either. We also don't know where girth was measured, whether it is a mid shaft erect girth or a maximal girth. Again, for a condom manufacturer one might assume that maximal girth is of greater significance but we don't really know.
Third, all the statistical analysis assumes a normal (standard) deviation for penis length and girth and this is an assumption that may not apply. I personally believe that penis size is not normally distributed, and that there is a distinct "tail" to the distribution in the direction of the "large outliers". In other words, I do not think that the really large penises that we know do exist are completely balanced out by an equal number of really small ones. If so, larger sizes would be more common than the given statistics suggest. And there is some support for this since most penis length surveys have demonstrated a mean length that is significantly larger than the median length.
And even if something is statistically quite unlikely, it doesn't mean it doesn't happen or doesn't exist. We see things happening every day that are extraordinarily unlikely statistically. With the enormous number of men in the world, and with virtually instantaneous world-wide telecommunication, even if a given penis size is statistically extraordinarily rare, there will still be many men with that size and the odds are they will be using it to make money in porn or show it off on the internet.
