Likert Scale

The Likert Scale was created by Rensis Likert in 1932 to measure consumer attitudes (Edmonson, 2005). It can be used in consumer tests by consumers respond to the degree of agreement and disagreement statements about a product. This information is mostly used for advertising claims on consumer perceptions about products as well as defending any legal challenge (Lawless & Heymann, 1998). The Likert scale was applied to many food related psychographic measurements such as Food Neophobia Scale (Ritchey, Frank, Hursti, & Tuorila, 2003), Food Involvement Scale (Bell & Marshall, 2003), Health and Taste Attitude Scale (HTAS) (Roininen, et al., 2001), List of Value Scale (Chryssohoidis & Krystallis, 2005), and Food Related Life Style (O’Sullivan, Scholderer, & Cowan, 2005). The challenge of using Likert scale is confusion on whether it is ordinal or interval. This confusion causes researchers to use inappropriate statistical methods to analyze data (Edmonson, 2005). 

Responses to a single Likert item are usually treated as ordinal data. The nonparametric tests that have been used include Wilcoxon’s signed rank test, the Whitney-Mann-Wilcoxon test, the Kruskal-Wallis test, and Fieldman’s test. In some cases responses to several Likert items are summed up and treated as interval data. Researchers need to make sure that the summed data are normally distributed. If the data are normally distributed then the parametric statistic methods can be applied to the summed data. However, the data transformation may be useful to make certain that parametric assumptions are not violated (Verbych, 2007). Parametric methods that have been proposed for analysis of ordinal data are as follows:

• The exact probability test to use with Likert-type data was proposed by Cooper (1976) and provided a table of critical values for small sample size. In the case of a large sample size normal approximation can be used but the analyst has to be careful with issues such as: the points of the Likert scale are equally spaced, consumers respond independently from each other, each category of scale has an equal opportunity to get the response from consumer (Cooper, 1976).
• Using the application of ordered probit model to treat with Likert ordinal data (Daykin & Moffatt, 2002).
• Use of cumulative logits for modeling ordinal response variables and cumulative link models for binary data (Verbych, 2007).

Reference List

• Bell, R., & Marshall, D. W. (2003) The construct of food involvement in behavioral research: scale development and validation⋆. Appetite, 40 (3), 235-244.
• Chryssohoidis, G. M., & Krystallis, A. (2005) Organic consumers’ personal values research: Testing and validating the list of values (LOV) scale and implementing a value-based segmentation task. Food Quality and Preference, 16 (7), 585-599.
• Cooper, M. (1976). An Exact Probability Test For Use With Likert-Type Scales. Educational and Psychological Measurement, 36, 647-655.
• Daykin, A. R., & Moffatt, P. G. (2002). Analyzing ordered Responses: Areview of the ordered Probit Model. Understanding Statistic, 3, 157-166.
• Edmonson D.R. (2005). Likert Scale: A History. Retrieved from CHARM database.
• Lawless H.T., & Haymann H. (1998) Sensory Evaluation of Food: Principle and Practices. MA: Kluwer Academic Publishers.
• O’Sullivan, C., Scholderer, J., and Cowan, C. (2005) Measurement equivalence of the food related lifestyle instrument (FRL) in Ireland and Great Britain. Food Quality and Preference, 16 (1), 1-12.
• Ritchey, P. N., Frank, R. A., Hursti, U., and Tuorila, H. (2003) Validation and cross-national comparison of the food neophobia scale (FNS) using confirmatory factor analysis. Appetite, 40 (2), 163-173.
• Roininen, K., Tuorila, H., Zandstra, E. H., de Graaf, C., Vehkalahti, K., Stubenitsky, K., and Mela, D. J. (2001a) Differences in health and taste attitudes and reported behaviour among Finnish, Dutch and British consumers: a cross-national validation of the Health and Taste Attitude Scales (HTAS), Appetite, 37 (1), 33-45.
• Verbych, E., (2007). A Comparison of Methods to Analyze Likert Scale Data. Unpublished master thesis, Kansas State University, Manhattan, KS.