This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
To the Editor:—
The letter of Dr. Saville has raised some pertinent points. It is difficult to define "bias" in any population sample. The studies of Vincent and Urist and Gershon-Cohen et al may be expected to represent the picture in the institutions they studied, but these people undoubtedly vary in ethnic origin, previous dietary history, occupation, and many other characteristics. There is no reason to believe that they represent a random sample of the population at large. Indeed it may be assumed that they do represent a "biased" sample.The groups we studied were individuals who consulted their physician for some purpose and who consented to participate in the study. There is indeed reason to believe that such samples are "biased" since they are people who, for some reason, consult their physician. One might expect such a sample to have a higher proportion of osteoporosis than a random sample
Bernstein DS, Sadowsky N, Hegsted DM, Guri CD, Stare FJ. Osteoporosis And Fluorides. JAMA. 1967;199(1):47-48. doi:10.1001/jama.1967.03120010091033