4 Jan 2018

Latent Variable Models for Market Research

Ms. Nivedita Bhaktha
Nivedita Bhaktha is a doctoral student in the Quantitative Research, Evaluation and Measurement program at the Ohio State University. She works as a graduate research associate in the Research Methodology Center and provides statistical consultation to doctoral students and faculty members. Nivedita has a Master’s degree in statistics and her research interests are in latent variable modeling, multilevel analysis and generalized linear modeling. She primarily uses R, Stata and SAS for data analysis and has conducted several workshops on using statistical software for data analysis. She has taught business statistics to undergraduate students. She has also taught univariate and multivariate courses to doctoral students. Nivedita likes learning about new quantitative techniques and applying it to social research problems.
Abstract

Latent variable models are an approach to relate manifest (observed) variables to latent (unobserved) variables. Latent variable models are widely used across different industries. Factor analysis, structural equation models and latent class analysis are some of the most commonly used latent variable models. These models are used in market research analysis and have been gaining popularity. This is evidenced by an increase in the number of articles on latent variable models in Quantitative Marketing and Economics, Journal of Marketing Research etc.

Factor analysis is used to explain the variance among correlated manifest variables in terms of fewer latent variables. This model is typically used to gain insights on buying behavior, customer demographics and in dimension reduction to name a few. Structural equation models are latent variable analogue to multiple regressions. This model is used to compare competing models. Latent class models help identify sub-groups called latent class which explain patterns of association between manifest variables. Latent class models are analogues to cluster analysis.

The presentation will provide a conceptual introduction to these models. Advantages and disadvantages of each will be provided. The presentation contains real life examples of each of the models in the field of market research.

Back