Censored probit models do not fail randomly: A Monte Carlo study

Values of the true correlation up to ±0.9 in a censored probit model do not lead to random failure owing to limit values of the estimated correlation, but a Monte Carlo study reveals other biases in estimating censored probit models.