Well, you might want to question the amount of people tested.
For example :
Because you have to take into account the rapid local diversity that also occurs. If you would take the continent Africa for example, the people there are as different as night and day from each other. This difference was more prominent 1 or 2 centuries ago. The local tribe system( or just family features) really showed differences. I have a book where you could really see differences of 14 different kinds of humans all with their own specific local characteristics. If you would line them up you would not say "colored people all look alike". The same aspect applies for Caucasians or Asians.
Lets say you have one group of people who originally came from a local tribe where certain physical features are more prominent. Then when you look at those limited amount of people only, you would make the wrong assumption based upon skin color or facial features only.
Based upon ignorance and using the research you mentioned only, i would say some black men will be frustrated all their lives because of their small cocktail sausages, and some Asians will never wear any pants to constantly show off their horse dicks.