Lowering sample size in comparative analyses can indicate a correlation where there is none: example from Rensch's...

Area of Science:

• Evolutionary Biology
• Comparative Phylogenetics
• Statistical Ecology

Background:

• Phylogenetic comparative methods (PCMs) were developed to distinguish functional correlations from those due to shared ancestry.
• Shared ancestry can inflate or obscure evolutionary correlations, a known challenge in comparative studies.
• Independent contrasts analysis is a common PCM used to infer evolutionary relationships between traits.

Purpose of the Study:

• To identify a specific scenario where independent contrasts analysis yields misleading results due to sample size effects.
• To demonstrate how similarity from shared ancestry can create spurious evolutionary correlations.
• To investigate the impact of clade-specific trait covariation on statistical inference in phylogenetic analyses.

Main Methods:

• Theoretical analysis of independent contrasts under specific conditions of trait covariation.
• Simulation of phylogenetic data to explore the effects of sample size and clade structure.
• Examination of a case study concerning Rensch's rule in primates.

Main Results:

• Under specific circumstances, small sample sizes in independent contrasts analysis can erroneously suggest an evolutionary correlation.
• Conversely, larger sample sizes may fail to detect a true correlation when trait covariation is concentrated within clades.
• The observed effect is opposite to the usual expectation that increased sample size enhances statistical power.

Conclusions:

• The reliability of independent contrasts analysis is contingent on the phylogenetic distribution of trait variation.
• Clade-specific factors influencing multiple traits can distort inferences about evolutionary correlations, particularly with limited data.
• Careful consideration of sample size and phylogenetic structure is crucial for accurate interpretation of comparative studies, such as Rensch's rule in primates.