To add to the significant caveats already posted by others (correlation<>causation, confounding factors, etc), diary studies of diet are known to have very poor

To add to the significant caveats already posted by others (correlation<>causation, confounding factors, etc), diary studies of diet are known to have very poor signal to noise. Also, there's a notable counter-example since diabetics are a large, distinct group which has had substantially higher average artificial sweetener consumption than the general population for many decades.