|
Activity Number:
|
369
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #301022 |
|
Title:
|
Performance of Some Statistical Methods for Testing Equality of Parameters in Zero-Inflated Continuous Distributions
|
|
Author(s):
|
Lei Zhang*+ and William D. Johnson and Jixiang Wu
|
|
Companies:
|
Mississippi State Department of Health and Lousiana State University System and Mississippi State University
|
|
Address:
|
Office of Health Data & Research, 570 E. Woodrow Wilson, Jackson, MS, 39215-1700,
|
|
Keywords:
|
Continuous ; Distribution ; Zero-Inflated ; Statistical ; Methods ; Equality
|
|
Abstract:
|
We evaluated the properties of four statistical methods for testing equality among populations with zero-inflated continuous distributions. These tests are based on likelihood ratio (LR), parametric jackknife (PJ), non-parametric jackknife (NPJ), and central limit theorem (CLT) statistics. We investigated their statistical properties using the simulated data with gamma, exponential, and log-normal density functions. These four statistical tests are compared regarding the empirical Type I errors and testing powers estimated through 10,000 repeated samples. The LR test and the PJ test performed better than the other two tests since their empirical Type I errors were close to the nominal 0.05 level and demonstrated good power for rejecting alternative hypotheses. The NPJ test had unacceptable empirical Type I errors. The CLT test performed reasonably well for larger sample sizes.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2008 program |
