File talk:Correlation between BMI and Percent Body Fat for Men in NCHS' NHANES 1994 Data.PNG

Was this accepted in a scientific paper? It seems so strange to put a linear regression while it is obvious that it is deviating quite a lot especially in the higher and lower ranges there is a non random deviation from the linear model. If adiposity was measured as well, one could have put in a color gradient for adiposity, now the data don't demonstrate the remark about the adiposity in the different quadrants. Also the individual points are too big making it impossible to judge how many points there actually are in the blue part and how they are distributed.Viridiflavus (talk) 13:23, 18 March 2013 (UTC)