UT study: No proof that eliminating gluten, casein from diets of autistic children is effective; "The diet doesn't have anything to do with autism," researcher says

The observations of the parents can be rather safely dismissed since absolutely no study, to date, could replicate the observations of the parents. As soon as the trials are started on groups of subjects, the effect seems to vanish.

The larger and longer and more controlled the study is, the more closely the effect-rate seems to approach that of random-chance plus or minus the margin-of-error.

The MOST biased person and the most emotionally subjective situation would be that of a parent in the role of evaluating a child that they believe they could be helping. It's the prima-facie reason for double-blinding, placeb and controlled studies.

Parents performing a diet change where they could expect a benefit will look for evidence of the benefits and will be far more likely to dismiss or ignore the evidence of either no change or any detrimental effects. On the other hand, when they stop any intervention they believe to be beneficial or begin an intervention which is believed to be detrimental, they are far more likely to dismiss any benefits, and to focus on the evidence of detriment.

That's why the researchers attempt to strongly avoid interviewing parents on subjective points, and try to stick to empirically and objectively measurable observations.