Estimates of Carry‐over Effects in Two‐product Home Usage Consumer Tests

Abstract

In‐house use consumer test data from four studies dealing with three pairs of household products and a pair of antiperspirant products were examined for significant carry‐over (product usage order) effects, which would confound the analysis of treatment (product) effects. In each study, two products were compared using a two‐period crossover design. One hundred twenty panelists participated in each study. A forced choice preference scale or a 9‐point hedonic scale was used to obtain responses from various sensory attributes. In all studies, the estimates of carry‐over effects were not significant at the 5% level. Transformation of hedonic scale data into preference dichotomy also gave estimates of carry‐over effects which were not significant at the 5% level, but led to a loss of test sensitivity for detecting treatment differences. The authors recommend that all comparative crossover design studies in sensory evaluation be monitored for carry‐over effects and that statistically determined sample size should be used to reduce the possibility of obtaining significant carry‐over effects.

Sequential Monadic Designs: Some Theoretical Considerations

“…Hence, depending on the magnitude and direction of the residual effects D, may either underestimate or overestimate the difference in the tested products. Estimates of D, have been reported by Gacula et al (1986) for household and personal care products.…”