An ordered generalised extreme value model with application to alcohol consumption in Australia

Area of Science:

• Econometrics
• Statistical Modeling
• Consumer Behavior Analysis

Background:

• Ordered discrete choice data presents unique analytical challenges.
• Existing models like Ordered Probit/Logit (OP/OL) and Multinomial Logit (MNL) have limitations in flexibility and accounting for choice correlations.
• The need for advanced statistical models to capture complex consumer choices is evident.

Purpose of the Study:

• To introduce and evaluate the Ordered Generalised Extreme Value (OGEV) model for ordered discrete choice data.
• To assess the OGEV model's performance against conventional OP/OL and MNL models.
• To apply the OGEV model to analyze factors influencing alcohol consumption levels in Australia.

Main Methods:

• Development and application of the Ordered Generalised Extreme Value (OGEV) model.
• Utilizing unit record data from Australia on alcohol consumption.
• Employing model selection analysis to compare OGEV with OP and MNL.

Main Results:

• The OGEV model demonstrates flexibility and is grounded in random utility maximization.
• It effectively accounts for correlations across choices stemming from unobservable individual characteristics.
• Model selection analysis indicated that the OGEV model is preferable to OP and MNL for analyzing alcohol consumption data.

Conclusions:

• The Ordered Generalised Extreme Value (OGEV) model provides a superior framework for analyzing ordered discrete choice data.
• The OGEV model's application to alcohol consumption reveals its utility in understanding the impact of economic and demographic factors.
• This study advocates for the adoption of the OGEV model in similar discrete choice modeling applications.