Related Experiment Videos
Resampling probability values for measures of ordinal variation and consensus
Kenneth J Berry1, Janis E Johnston, Paul W Mielke
1Department of Sociology, Colorado State University, Fort Collins, CO 80523-1784, USA. berry@lamar.colostate.edu
Psychological Reports
|July 30, 2005
Summary
Resampling permutation procedures offer accurate alternatives to exact tests for large datasets. These methods approximate exact permutation tests for variation and consensus measures in ordered categories.
Area of Science:
- Statistics
- Computational Statistics
Background:
- Exact permutation tests are computationally intensive for large sample sizes.
- Approximation methods are needed to make permutation inference feasible.
Purpose of the Study:
- To describe resampling permutation procedures as an alternative to exact permutation tests.
- To apply these procedures to measures of variation and consensus in ordered categories.
Main Methods:
- Resampling permutation procedures were developed and described.
- These procedures were applied to three measures of variation.
- The methods were also applied to three measures of consensus among ordered categories.
Main Results:
- Resampling permutation procedures provide accurate approximations to exact permutation procedures.
- The described methods are suitable for large sample sizes where exact tests are intractable.
Conclusions:
- Resampling permutation procedures are a viable and preferred alternative for statistical inference with large datasets.
- These methods extend the application of permutation tests to measures of variation and consensus in ordered categorical data.