Identifying random-scoring respondents in sensory research using finite mixture regression models

Extrinsic and intrinsic quality cues in Chinese consumers’ purchase of pork ribs

“…The latent class model with a random-choice class (C5) identified that choices from 22.9% of the sample could not be explained by the underlying product quality cues. Compared to other studies (Cleaver & Wedel, 2001;Popper et al, 2004) this is a relatively large size of the random-choice class, which could be related to rapid change in the Chinese retail environment and the fact that some Chinese consumers are not used to buy by choosing between different product pork alternatives.…”

“…Latent class choice models offer particular flexibility to deal with participants' differences in choice decision strategies and choice consistency, which would otherwise lead to biased part worth utilities (Hensher, 2010;Hess, Shires, & Bonsall, 2013). The first approach of specifying a random-choice class identifies respondents for which choices cannot be related to the underlying product characteristics (Cleaver & Wedel, 2001;Popper, Kroll, & Magidson, 2004). The random-choice class is specified so that the utilities of all levels of an attribute are set equal, so that choices of respondents highly loading on this class were not affected by the product attributes.…”

A cross-national consumer segmentation based on food benefits: The link with consumption situations and food perceptions

“…That is, each respondent's general benefit importance ratings were adjusted by subtracting their mean. By centring the scores, response tendencies are cancelled out (Cleaver & Wedel, 2001;Dillon, Frederick, & Tangpanichdee, 1985). Especially in the case of a cross-national segmentation, it seems possible that consumers in different countries have different response tendencies.…”

Using non-food information to identify food-choice segment membership

“…Next, we introduce our methodology which includes state-of-the-art finite-mixture-modeling with concomitant variables (Cleaver & Wedel, 2001;Wedel & Kamakura, 2002, p.75). This modeling approach allows us to test whether the segment structure based on food-choice motives can be improved upon by incorporating information about membership of societal organizations.…”