Cox Proportional Hazards Regression for Interval-Censored Data with an Application to College Entrance and Parental Job Loss

This study involved conducting a survival analysis by fitting Cox proportional hazards model to Korea Labor Panel data analyze the impact of parental job loss on children’s delayed admission colleges and universities in South Korea, using 376 subjects whose education levels were college-level or higher. Since are interval- right-censored, we compared three imputation methods: simple omission, as average left right values interval, multiple imputation. Their integrated areas under ROC curve (AUC) mean square errors (MSE) assess their predictive estimation performances. It was found that, within simulation, method exhibited lower MSE than other two methods. However, no difference observed iAUC values. In group where each householder had at least college degree, significantly related university first-born child regardless use interval censoring method. particular, when children experienced parents’ unemployment age 18, probability reduced nearly 53% cases they did not. implies that government should develop policy system offering psychosocial support for adolescents who cannot expect help.