Abstract
The authors derive the asymptotic mean and bias of Kendall's tau and Spearman's rho in the presence of left censoring in the bivariate Gaussian copula model. They show that tie corrections for left-censoring brings the value of these coefficients closer to zero. They also present a bias reduction method and illustrate it through two applications.
Original language | English (US) |
---|---|
Pages (from-to) | 427-442 |
Number of pages | 16 |
Journal | Canadian Journal of Statistics |
Volume | 36 |
Issue number | 3 |
DOIs | |
State | Published - Sep 2008 |
Keywords
- Bias reduction
- Bivariate gaussian copula mode
- Bivariate normal distribution
- Detection limit
- Kendall's tau
- Left censoring
- Pearson's correlation coefficient
- Spearman's rho