The C-S method in computation of the logistic normal integral

More about the book

The book explores logistic regression models affected by measurement error, focusing on the logistic normal integral's likelihood function. It evaluates the E. Crouch and D. Spiegelman (1990) method, implemented in FORTRAN, by integrating it with R and comparing its efficiency to the classic Gaussian Quadrature method. Simulation results indicate that the C-S approach significantly outperforms the Gaussian method in speed while maintaining precision, particularly with larger sample sizes, offering valuable insights for researchers utilizing logistic normal integrals in their studies.