conover.test: Implementation of the Conover-Iman Multiple Comparisons Test

conover.test is a statistical software implementation for performing the Conover-Iman test for multiple pairwise comparisons following a Kruskal-Wallis omnibus test. The method, developed by Alexis Dinno, is designed to test for 0th-order stochastic dominance and accounts for tied ranks in the data. It computes k(k-1)/2 pairwise comparisons based on the Conover-Iman t-test-statistic of rank differences.

Use Cases

• Perform post-hoc analysis after a Kruskal-Wallis test based on the described method for multiple pairwise comparisons.
• Test for median or mean differences across groups when data distributions are identical except for location, as described.
• Analyze data with tied ranks using a test that accounts for this condition, as mentioned in the description.

Strengths

• Implements a specific, peer-reviewed statistical method (Conover-Iman test, 1979) for a defined analytical purpose.
• Explicitly handles tied ranks, a common data condition, as stated in the description.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Alexis Dinno

Collection Method

Software implementation of a statistical test.