Premium
A COMPARISON OF THE DISCRIMINATING POWER OF ANOVA AND R‐INDEX ANALYSES OF HEDONIC DATA FOR VARIOUS PRODUCTS AND EXPERIMENTAL PROTOCOLS
Author(s) -
PARK HAENA,
O'MAHONY MICHAEL,
KIM KWANGOK
Publication year - 2007
Publication title -
journal of sensory studies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.61
H-Index - 53
eISSN - 1745-459X
pISSN - 0887-8250
DOI - 10.1111/j.1745-459x.2007.00106.x
Subject(s) - statistics , analysis of variance , ranking (information retrieval) , mathematics , rank (graph theory) , rating scale , paired comparison , econometrics , computer science , artificial intelligence , combinatorics
Consumers rated a set of toothpastes and a set of orange‐flavored beverages on a 9‐point hedonic scale, using two experimental protocols: Rank‐Rating where stimuli could continually be reassessed and a more traditional approach where they could not. A 21‐point hedonic scale was also used in the Rank‐Rating condition. The hedonic data were analyzed in the usual way using ANOVA with multiple comparisons and also by ranking the data and using an R‐index analysis. Regarding the numbers of significant differences recorded, the two analyses were comparable, with a very slight and nonsignificant advantage for the ANOVA analysis. Unlike with intensity scaling, the difference between Rank‐Rating and “traditional” scaling was slight but not for all products. The same was true for scale length effects. Differences among the products suggested effects due to the number of attributes that varied.PRACTICAL APPLICATIONS The key finding in this study is the use of Rank‐Rating where stimuli could continually be re‐assessed for the assessment of hedonic response of consumers in the various products. Rank‐Rating may possibly give a better discrimination than more traditional scaling, depending on the number of attributes which varied. The results of the study also recommend the use of R‐index analyses of ranked hedonic data in the analyses rather than ANOVA with multiple comparisons.