×

Two-way ORDANOVA: analyzing ordinal variation in a cross-balanced design. (English) Zbl 1477.62205

Summary: Variability assessment of qualitative data has an important role in diverse areas such as sociology, quality engineering, healthcare, decision making, genetics, metrology in chemistry and many others. To test the homogeneity hypothesis, we use two main categorical factors. Using a cross-balanced design, we provide a decomposition theorem of the sample total-variation into ‘intra’ (within) component and ‘inter’ (between) component for any scale. Building on this, we provide a way for decomposing the ‘inter’component itself into variation components contributed by the two factors and their interaction (when we consider more than one replicated test in the cross-design). Moreover, focusing on an ordinal scale, we offer segregation power indices for testing the significate effects of the two factors and their interaction. A two-way ORDANOVA calculator is provided to display the empirical distributions of the segregation power indices and to find critical points of the test statistics distributions at the given confidence level. Some examples are given to demonstrate the proposed method.

MSC:

62K15 Factorial statistical designs
62H15 Hypothesis testing in multivariate analysis
62J10 Analysis of variance and covariance (ANOVA)
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] CBS: Israel Central Bureau of Statistics - Tools and Databases. The site at the web address (in Hebrew). Accessed 14 2019. https://www.cbs.gov.il/he/Statistics/.
[2] Anderson, R. J.; Landis, J. R., CATANOVA for multidimensional contingency tables: Nominal-scale response, Commun. Stat. - Theory Methods, 9, 1191-1206 (1980) · Zbl 0463.62064
[3] Anderson, R. J.; Landis, J. R., CATANOVA for multidimensional contingency tables: Ordinal-scale response, Commun. Stat. - Theory Methods, 11, 257-270 (1982)
[4] Blair, J.; Lacy, M. G., Measures of variation for ordinal data as functions of the cumulative distribution, Percept. Mot. Skills, 82, 411-418 (1996)
[5] Blair, J.; Lacy, M. G., Statistics of ordinal variation, Sociol. Methods Res., 28, 251-280 (2000)
[6] D’Ambra, L.; Beh, E. J.; Amenta, P., Catanova for two-way contingency tables with ordinal variables using orthogonal polynomials, Comm. Statist. Theory Methods, 34, 1755-1769 (2005) · Zbl 1075.62057
[7] Gadrich, T.; Bashkansky, E., ORDANOVA: Analysis of ordinal variation, J. Statist. Plann. Inference, 142, 3174-3188 (2012) · Zbl 1348.62209
[8] Gadrich, T.; Bashkansky, E.; Zitikis, R., Assessing variation: A unifying approach for all scales of measurement, Qual. Quant., 49, 3, 1145-1167 (2015)
[9] Gadrich, T.; Marmor, Y. N., Supporting material for the paper two-way ORDANOVA: Analyzing ordinal variation in a cross-balanced design (2019), https://drive.google.com/file/d/1m-zzE0m17AIkntUxMo1gOXTNNaIGI_3V/view
[10] Gini, C., Variabilità E Mutabilità:à: Contributo Allo Studio Delle Distribuzioni E Delle Relazioni Statistiche (1912), Tipografia di Paolo Cuppin: Tipografia di Paolo Cuppin Bologna
[11] Giorgi, G. M., Gini’s scientific work: An evergreen, Metron, 63, 299-315 (2005) · Zbl 1416.62036
[12] Light, R. J.; Margolin, B. H., An analysis of variance for categorical data, J. Amer. Statist. Assoc., 66, 534-544 (1971) · Zbl 0222.62035
[13] Lombardo, R.; Camminatiello, I., CATANOVA for two-way cross classified categorical data, Statistics, 44, 1, 57-71 (2010) · Zbl 1291.62138
[14] Marmor, Y. N.; Gadrich, T., Two-way ORDANOVA calculator tool (2019), https://drive.google.com/drive/folders/1pV9QAkHiXVqWi-7wso2ocrPAKuN4UdHC?usp=sharing
[15] Stevens, S. S., On the theory of scales of measurement, Science, 103, 2684, 677-680 (1946) · Zbl 1226.91050
[16] Vanacore, A.; Marmor, Y. N.; Bashkansky, E., Some metrological aspects of preferences expressed by prioritization of alternatives, Measurement, 135, 52-526 (2018)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.