Ralph G. Edwards, Jr., Charles H. Powell and Mildred A. Kendrick, Dust Counting Variability (1966)

A nine-page reprint from the American Industrial Hygiene Association Journal for November-December 1966, by three scientists of the U.S. Public Health Service's Division of Occupational Health in Cincinnati. Seven people counted the particles in the same 30 fields of a dust sample taken "from an asbestos textile milling operation," ten times each, separating fibers from grains and using both the standard light-field microscope and phase contrast. Analysis of the 8,400 counts showed that "The biggest single source of variation in results is individual differences of the counters": counts by different people varied about twice as much as one person's repeated counts, and phase contrast gave higher counts than the light-field method "in all cases."

What it shows

Throughout the 1950s and 1960s asbestos dust was measured by counting particles under a microscope, and the resulting figures were compared with the limit of 5 million particles per cubic foot. This study showed how much those counts depended on who did the counting, and that fibers were counted less reliably than grains. The authors suggested that the best reproducibility came at 40 to 60 particles per field.