Abstract

A gauge repeatability and reproducibility study (GRR) uses analysis of variations (ANOVA) on an appropriately designed experiment to separate and quantify the components of the overall uncertainty. The authors have previously presented results of GRR studies of the measurement of airborne and impact insulation of floor-ceiling and demising wall assemblies in several apartment buildings, in which the uncertainty in the measurement method and the variability of the nominally-identical assemblies were compared. The results of two additional GRR studies on measurements of airborne noise isolation of wood stud demising walls are presented. The first study, like previous studies, evaluates the components of variance attributable to operator, repeatability, and part. The second study uses a fixed operator and part, and evaluates the variance due to loudspeaker type, position, and level on the measured noise reduction. The measurement standard (ASTM E336) gives limited guidance on the type and location of the loudspeaker used on the source side, and this study can inform whether changes in the standard with regards to the loudspeakers could reduce the uncertainty in measurement.