File under: "Lies, Damned Lies, and Statistics"

A recent study in The American Journal of Clinical Nutrition had us rolling our eyes at its insistence that women who replaced sugar-sweetened beverages and fruit juices with water had a lower incidence of type 2 diabetes. While we suspected the study of being another data dredge of the over 82,000 women in the Harvard School of Public Health s Nurses' Health Study II, we called upon ACSH scientific advisor and statistician Dr. Stanley Young, who confirmed our suspicions, and then some.

Dr. Young expertly dissected the published paper with the scalpel of a seasoned statistician and, sure enough, found it chock full of error and statistical manipulation. (Or, as ACSH's Dr. Elizabeth Whelan puts it, "These researchers have publicized their hypotheses minus any supporting evidence.") While we won t quote for you his criticism of the study s range of p-values or stratification variables, we think his general summation of the study is worth citing verbatim: The relationships among the variables are sufficiently complex and the measurement of water consumption so imprecise, writes Dr. Young, that making firm claims is problematic.