|
Activity Number:
|
333
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #310029 |
|
Title:
|
A Rank Regression Analysis of Hormonal and Cytokine Effects on Birth Outcome
|
|
Author(s):
|
Alai Tan*+ and Roberta J. Ruiz and Daniel H. Freeman
|
|
Companies:
|
The University of Texas Medical Branch and The University of Texas Medical Branch and The University of Texas Medical Branch
|
|
Address:
|
301 University Boulevard, Galveston, TX, 77555-1148,
|
|
Keywords:
|
Rank regression ; Linear regression ; Log-transformation
|
|
Abstract:
|
Normality is one of the assumptions for traditional linear regression. However, this assumption is often violated in physiological data, particularly the cytokines and hormones. Despite this violation, traditional regression is regarded as a robust method when the sample size is relatively large. Another widely used method is regression on log-transformed variables. The present study compared the rank regression analysis with traditional regression and regression based on log-transformed variables in analyzing hormonal and cytokine effects on birth outcome. We found that rank regression is more powerful in detecting the effects of hormones and cytokines on birth outcome. We propose the use of rank regression analysis in analyzing hormone and cytokine data, especially when the traditional regression and log-transformed regression fail to detect the hypothesized associations.
|
