Rabinowitz D
Department of Statistics, Columbia University, New York 10027, USA.
Am J Hum Genet. 1996 Sep;59(3):726-30.
The Cannings and Thompson approach to correcting for ascertainment bias in family studies is extended to settings with multiple ascertainment. The extension is based on maximizing a pseudolikelihood. Two approaches to computing standard errors for the maximum pseudolikelihood estimate are described. One is especially simple to compute, while the other is more generally applicable. Simulation experiments suggest that the standard-error computations can be quite accurate.
坎宁斯和汤普森用于校正家系研究中确证偏倚的方法被扩展到多确证情况。该扩展基于最大化伪似然。描述了两种计算最大伪似然估计标准误的方法。一种计算特别简单,而另一种更具普遍适用性。模拟实验表明,标准误计算可以相当准确。