Ok, so you missed my point entirely. I have some experience in analytics and statistics, so I have a very good grasp on how this works. I'm too tired and lazy to go over it, and I found this lovely BBC link that will show you how an average of something can be misleading.
only registered users can see external links
Furthermore, that chart does not show the most common penis sizes. It solely shows the percentage of people you are longer than taking no regard for exactly how those lengths are distributed. If you are longer than 95% of people, that means that you could be 8"(or whatever) and the other 90% could be 7.9. That chart is meant to be simple, not accurate. If it was representing the actual raw data it would look more like this.
The average of this is 6.1, even though the most common of the numbers represented are 5 and 8. The fact that there are more samples above 5 than below distorts the average to 6.1. In fact if you flipped the chart, it would perfectly prove my point. 6 would be the most common number as there is 10 but no 0 so 5 would become 6. However, the average would be 5.1, even though 6 is be the most common number on the flipped chart. In this case 3 is be the second most common, and there are more samples below 5.1 than above. This distorts the data away from the most common number of 6 on the flipped chart.
Science is good, but you need to understand how to interpret data properly. Else you could misunderstand what the science is saying.
Edited to add additional data then to re add the chart.