ISOLATION OF TEST SYSTEM VARIABLES DATA VARIANCE COMPONENTS USING GAGE RR AND RANDOM EFFECTS ANOVA

摘要

The application of a random effects Analysis of Variance (ANOVA) to the assessment of test system data allows us to isolate the sources of variation in our measurements. Utilizing a traditional gage R&R approach, the paper shows how data collected in a systematic way can be used in this structured analysis. The approach allows for the determination of significant sources of variation due to the units under test, the testers themselves, and the interaction between the two factors. In addition, point estimates of the variance due to each of the three sources are computed giving the analyst insight as to where the source of the most variation in a set of measurements lies. Armed with the results of the omnibus ANOVA, the paper shows how to disentangle the specific sources of the differences using the analysis of simple effects, simple comparisons, and interaction contrasts. Making corrections to keep the family wise Type I error rate at acceptable levels is presented.