Integrated mixed logit and latent variable models

Abstract

A traditional discrete choice model assumes that an individual's decision-making process is based on utility maximization and that the systematic part of the utility function depends on some observable attributes and covariates. These attributes and covariates however can only explain part of the utility and a large part remains unexplained. In recent years, researchers have recognized that psychological factors such as attitudes, lifestyle and values can a ect an alternative's utility and hence the individual's choices. Therefore, extending the traditional discrete choice model by incorporating those latent or unobservable factors, can help to better understand the individual's decision making process and therefore to yield more reliable part-worth estimates and market share predictions.Several integrated choice and latent variable (ICLV) models which merge the conditional logit model with a structural equation model exist in the literature. They assume homogeneity in the part-worths and use latent variables to model the heterogeneity among the respondents. The current research uses the mixed logit model that describes the heterogeneity in the part-worths and uses the latent variables to decrease the unexplained part of the heterogeneity. The empirical study that we conducted to assess student preferences on mobile phone features shows these ICLV models perform very well with respect to model t and prediction.Furthermore, we compare the di erent estimation procedures that exist in the literature. Results show that, as expected, the simultaneous procedure where we estimate the structural equation model (SEM) and the choice model simultaneously, gives better model t and more accurate predictions compared to the sequential procedure that estimates the SEM rst and then the choice model taking the estimated factor scores into account. But the gain of the simultaneous procedure is relatively small. Therefore one can use the easier sequential procedure without losing much e ciency if needed.

The Impact of Perception toward the Built Environment in Airport Access Mode Choice Using Hybrid Choice Modeling

“…In addition, we compared our results to the findings of previous studies of elasticity analysis; however, such elasticity analysis does not exist, particularly in the context of ground access to the airports' literature review. As the HCM methodology is relatively new, almost all significant scholarly papers which used hybrid choice methods did not report elasticities [7,8,10,61,62,[78][79][80][81][82][83][84][85].…”

A Multiattribute Benefits-Based Choice Model with Multiple Mediators: New Insights for Positioning

“…The resulting extensions to choice models include research in marketing that explores how product attributes affect consumer perceptions about higher-order benefits (e.g., Dellaert and Chorus 2015;Kim et al 2017) and processes in which consumers use goal attainment potential to reevaluate choice options (e.g., Swait, Argo, and Li 2018). These models differ from the set of integrated latent variable choice models (ICLVs) that focus on attitudes and values; in such models the latent constructs are fixed at the individual level and do not vary across product attributes (e.g., Ashok, Dillon, and Yuan 2002;Danthurebandara, Vandebroek, and Yu 2013).…”

“…These factors can also be measured using reflective indicators, often using ratings scales, described as attitudinal inventories by McFadden. Models of this type have been used to further insights into consumer choices in marketing (e.g., Ashok, Dillon, and Yuan 2002;Danthurebandara, Vandebroek, and Yu 2013) and transportation (e.g., Bahamonde-Birke and Ortúzar 2014; Bolduc and Alvarez-Daziano 2010; Kløjgaard and Hess 2011;Paulssen et al 2014).…”

“…Reflective measures of latent variables. Previous research incorporating latent variables into choice models either explicitly observes reflective measures of the latent variables based on prior research or management input (e.g., Danthurebandara, Vandebroek, and Yu 2013;Walker and Ben-Akiva 2002) or infers the latent variables based on underlying structures of the data (e.g., Elrod and Keane 1995;Kim et al 2017;Swait, Argo, and Li 2018). Whereas the latter approach is exploratory in nature and requires additional steps after model estimation to interpret the latent variables that the models detect, our proposed model follows the former stream of research with the aim of confirming and testing existing knowledge about the latent variables, including whether they capture all variation in choices.…”