PREDICTING LOW EMPLOYABILITY GRADUATES: THE CASE OF UNIVERSITI UTARA MALAYSIA

摘要

The persistence of the graduate unemployment problem warrants the need for predicting factors that are associated with low employability graduates. It is found that ethnicity, English language profi-ciency and types of degree obtained are significant predictors of graduates' employability which is measured either by the number of days being unemployed or probability of being unemployed. Compared to the Tobit model, the Probit model which does not accommodate the right censoring bias is found to have less predictive power in terms of the number of days being unemployed.