A series of least squares analytical models were developed in a cross-sectional epidemiologic study of observed variability in within-herd high milk somatic cell count (SCC) prevalence. The study population was restricted to Ohio (U.S.A.) dairy cow herds where post-milking antiseptic teat dipping and prophylactic dry cow antibiotic therapy of all cows were practiced. Additionally, no Streptococcus agalactiae was isolated from repeated bulk tank milk screening of these herds. The first analysis considered 12-month herd average high SCC prevalence as the dependent variable. The second analysis narrowed the herd average calculation to a 3-month or seasonal interval, while the third analysis included age adjustment of the 3-month rolling herd average high SCC prevalence. All of the models identified significant main effects and significant first order interactions of main effects. Milking order was a significant factor in all models, as were four interactions: washing teats only and rinsing teat cups between cows; washing with a common cloth or sponge and milking order; washing teats only and elapsed time between udder preparation and teat cup attachment; and elapsed time between udder preparation and teat cup attachment, and rinsing teat cups between cows.