A Latent Allocation Model for Mindset Variables
(R2020-83486)

AUTHORS

pablo marshall, puc

KEYWORDS

LDA; Mindset; Modelling

ABSTRACT

Mindset metrics, the measure of consumer’s perceptions, attitudes, and intentions, have a long tradition in marketing. Two important mindset metrics are brand awareness and attribute importance. This paper aims to use an extension/modification of the Latent Dirichlet Allocation (LDA) model used for text analytics that incorporates Bernoulli observations instead of the multinomial specification in the LDA model. Dichotomous variables are usual in mindset metrics, and for these variables, the model extension proposed in this study is more appropriate than the usual LDA model. Two applications, the first in brand awareness and the second in attribute importance, illustrate the model.