Permutation approach for ordinal preference data


GÖKPINAR F., GÖKPINAR E., Bayrak H.

COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION, cilt.46, sa.3, ss.2321-2332, 2017 (SCI-Expanded) identifier identifier

Özet

Best et al. (Best, D. J., Rayner, J. C. W., O'Sullivan, M. G. (2000). Product maps for consumer categorical data. Food Quality and Preference, 11: 91-97) suggested tests based on partitioning the X-2 statistic into relevant components of location, dispersion, and skewness effects for testing equality of each effect for ordinal preference data. It is known that the chi-square approximation requires large counts for categories. For this purpose, in this study, we investigate a permutation approach for these statistics and compare the performance of these tests with simulation study. In addition, the permutation approach can be used to produce a product map that classifies the products. We illustrate the approach with a real data example.