Open Conference Systems, 50th Scientific meeting of the Italian Statistical Society

Font Size: 
Modelling the latent class structure of multiple Likert items: a paired comparison approach
Brian J Francis

Last modified: 2018-05-17

Abstract


The modelling of the latent class structure of multiple Likert items measuredon the same response scale can be challenging. The standard latent class approachis to model the absolute Likert ratings, where the logits of the profile probabilitiesfor each item have an adjacent category formulation (DeSantis et al., 2008).We instead propose modelling the relative orderings, using a mixture model of therelative differences between pairs of Likert items. This produces a paired comparisonadjacent category log-linear model (Dittrich et al., 2007; Francis and Dittrich,2017), with item estimates placed on a (0,1) “worth†scale for each latent class. Thetwo approaches are compared using data on environmental risk from the InternationalSocial Survey Programme, and conclusions are presented.

References


Aitkin, M., D. Vu, and B. Francis (2014). Statistical modelling of the group structure of social networks. Social Networks 38, 74–87.

DeSantis, S. M., E. A. Houseman, B. A. Coull, A. Stemmer-Rachamimov, and R. A. Betensky (2008). A penalized latent class model for ordinal data. Biostatistics 9(2), 249–262.

Dittrich, R., B. Francis, R. Hatzinger, and W. Katzenbeisser (2007). A Paired Comparison Approach for the Analysis of Sets of Likert Scale Responses. Statistical Modelling 7, 3–28.

Francis, B. and R. Dittrich (2017). Modelling multiple likert items through an adjacentcategories ordinal paired comparison model. Presented at the 10th ERCIMc onference on computational and Methodological Statistics, London; 16-18 December2017.

Francis, B., R. Dittrich, and R. Hatzinger (2010). Modeling heterogeneity in ranked responses by nonparametric maximum likelihood: How do Europeans get theirscientific knowledge? The Annals of Applied Statistics 4(4), 2181–2202.

Hatzinger, R. and M. J. Maier (2017). prefmod: Utilities to Fit Paired Comparison Models for Preferences. R package version 0.8-34.

Linzer, D. and J. Lewis (2011). poLCA: An R Package for Polytomous Variable Latent Class Analysis. Journal of Statistical Software 42(10).

Magidson, J. and J. Vermunt (2004). Latent class analysis. In D. Kaplan (Ed.), The Sage Handbook of Quantitative Methodology for the Social Sciences, Chapter 10,pp. 175–198. Thousand Oaks, CA.: Sage Publications.

Vermunt, J. and J. Magidson (2013). Technical Guide for Latent GOLD 5.0: Basic, Advanced, and Syntax. Belmont, MA: Statistical Innovations Inc


Full Text: PDF