Examining Potential Bias in Large-scale Censored Data