Variation in Food Microbiological Tests Used to Evaluate Analyst Performance

Procedures for evaluating analyst performance in state and federal milk and food laboratory quality assurance programs depend on the estimates of variation for the methods used. The total components of variance for a method consist of replicate error, among-analysts variance, and analyst-sample interaction components. Eleven methods were evaluated in this study. Total variance estimates of milk analytical procedures for the Standard Plate Count, plate loop count, coliform plate count, direct microscopic somatic cell count, electronic somatic cell count and Wisconsin mastitis test were 0.01045, 0.01371, 0.01590, 0.01265, 0.00504 and 4.06, respectively. Estimates of the logarithmic total variance of food analytical procedures for the aerobic plate count, coliform most probable number (MPN), Escherichia coli MPN, fecal coliform MPN and Staphylococcus aureus MPN were 0.00853, 0.10205, 0.14705, 0.14780 and 0.24245, respectively. The study showed that current analyst performance levels used by state and federal laboratory quality assurance programs are satisfactory.