The Statistical Relationship between Bivariate and Multinomial Choice Models
Melvyn Weeks, Chris Orne
Abstract
Open-access reader
Melvyn Weeks, Chris Orne
Abstract
Open-access reader
The authors demonstrate the conditions under which the bivariate probit model can be considered a special case of the more general multinomial probit model. Since the attendant parameter restrictions produce a singular covariance matrix, the subsequent problems of testing on the boundary of the parameter space are circumvented by the construction of a score test.
OpenAlex reports 13 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The authors demonstrate the conditions under which the bivariate probit model can be considered a special case of the more general multinomial probit model. Since the attendant parameter restrictions produce a singular covariance matrix, the subsequent problems of testing on the boundary of the parameter space are circumvented by the construction of a score test.
Key concepts: Multinomial probit, Bivariate analysis, Multivariate probit model, Econometrics, Multinomial distribution, Mathematics, Statistics, Multinomial logistic regression