Dr Krohn sets up a straw man by criticizing our study for not being a screening study to detect individual "responders" to aspartame. This argument rests on Bernard Weiss' (Dr Krohn's reference 5) conclusion that statistical analysis of means failed to identify "responders" in studies designed to test the Feingold hypothesis. The generally accepted sequel to that story is that Weiss' conclusion (that there is an identifiable group of responders to color additives) is a Type I error (ie, a false positive)! In any event, Dr Weiss' conclusion is not supported by the data cited. The only way to avoid such an error is a blinded clinical evaluation of each child using all available data.
- Copyright © 1994 by the American Academy of Pediatrics