|
Activity Number:
|
298
- Model/Variable Selection and Model Evaluation
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #304930
|
Presentation
|
|
Title:
|
Maximum Likelihood Estimation of a Truncated Normal Distribution with Censored Data
|
|
Author(s):
|
Justin R Williams* and Hyung-Woo Kim and Kate Crespi
|
|
Companies:
|
UCLA and Alcon Laboratories, Inc. and
|
|
Keywords:
|
truncation;
censoring;
maximum likelihood estimation;
truncated normal distribution;
censored normal
|
|
Abstract:
|
We propose a method for obtaining maximum likelihood estimates of the parameters of a truncated normal distribution when the observations are censored. We show that our method, tcensReg, outperforms other commonly used methods such as Tobit regression or single imputation with the censoring threshold or half the censoring threshold on bias and mean squared error under a range of simulation settings. In simulations of non-inferiority testing, Type I error rates when using tcenReg were closer to the nominal levels than other methods in most scenarios. We apply the new method to analyze vision quality data collected from ophthalmology clinical trials comparing different types of intraocular lenses implanted during cataract surgery.
|
Authors who are presenting talks have a * after their name.
