A SIMULTANEOUS EQUATION MODEL OF INDIVIDUAL ATTITUDE AND BELIEF FORMATION TOWARDS SOCIAL ISSUES
Abstract
This dissertation presents a possible model for the generation of individual attitude and belief measures. The model is developed theoretically, then empirically tested with survey data collected on some alternative risky projects and activities. Analyses were done under the assumption that individuals were not homogeneous with respect to model parameters. The model is a simultaneous equation model which has several output variables being determined jointly. Because of multicollinearity in the input variables, standard statistical procedures provided unsatisfactory parameter estimates. This problem was addressed by developing and testing a new hybrid estimation technique which combined two-stage Least Squares and Ridge Regression. Subsequent to this technique's evaluation, it was used to estimate parameters and provide measures of model fit.
Degree
Ph.D.
Subject Area
Management
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.