Paired Comparison Algorithm
Jump to navigation
Jump to search
A Paired Comparison Algorithm is a comparison algorithm that evaluates paired differences between matched observations or paired models.
- AKA: Matched-Pair Comparison Algorithm, Paired Statistical Test Algorithm.
- Context:
- It can (typically) analyze Paired Data Structures with dependent observations.
- It can (typically) control Individual Variation through within-subject comparisons.
- It can (typically) increase Statistical Power by reducing between-subject variability.
- It can (typically) handle Before-After Designs and matched-pair designs.
- ...
- It can (often) require Smaller Sample Sizes than independent comparison algorithms.
- It can (often) detect Subtle Differences masked by individual variations.
- ...
- It can range from being a Parametric Paired Comparison Algorithm to being a Non-Parametric Paired Comparison Algorithm, depending on its distribution assumptions.
- It can range from being a Two-Sample Paired Comparison Algorithm to being a Multi-Sample Paired Comparison Algorithm, depending on its comparison count.
- It can range from being a Simple Paired Comparison Algorithm to being a Complex Paired Comparison Algorithm, depending on its test complexity.
- It can range from being a Exact Paired Comparison Algorithm to being a Approximate Paired Comparison Algorithm, depending on its computation method.
- ...
- Example(s):
- Counter-Example(s):
- Independent Sample Test Algorithms, which assume independent observations.
- ANOVA Algorithms, which compare multiple groups simultaneously.
- Unpaired Comparison Algorithms, which lack matching structures.
- See: Statistical Test Algorithm, Matched-Pair t-Test, McNemar's Test, Sign Test, 5x2 Cross-Validation Algorithm, Comparison Task, Hypothesis Test.