Strategies for Testing Slope Differences
- Author(s):
- Moses, Tim P.; Klockars, Alan
- Publication Year:
- 2009
- Report Number:
- RR-09-32
- Source:
- ETS Research Report
- Document Type:
- Report
- Page Count:
- 37
- Subject/Key Words:
- Slopes, Least Squares, Theil-Sen Estimates, Robust Regression, James Second-Order, Nonnormality, Residual Variance
Abstract
The robustness and power of 9 strategies for testing the differences in groups’ regression slopes were assessed under nonnormality and residual variance heterogeneity. For the conditions considered, the most robust strategies were the trimmed and Winsorized slope estimates used with the James second-order test, the Theil-Sen slope estimates used with James, and the Theil-Sen estimates used with percentile bootstrapping. The use of Theil-Sen slope estimates produced more powerful tests than the use of trimmed and Winsorized slopes.
Read More
- Request Copy (specify title and report number, if any)
- http://dx.doi.org/10.1002/j.2333-8504.2009.tb02189.x