Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Model selection with multiple regression on distance matrices leads to incorrect inferences.

Ryan P Franckowiak1, Michael Panasci2, Karl J Jarvis3

  • 1Environmental & Life Sciences Graduate Program, Trent University, Peterborough, Ontario, Canada.

Plos One
|April 14, 2017
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Predictions From Evolutionary Theory for Urban Environments.

Evolutionary applications·2026
Same author

Winter inversions and summer smoke: A season-dependent approach to PM₂.₅ modeling with low-cost sensors in complex terrain.

The Science of the total environment·2026
Same author

Evolving nature-based solutions for urban resilience.

Science (New York, N.Y.)·2026
Same author

Climate Change Projected to Double the Richness and Abundance of Soilborne Phytopathogenic Fungi in Southern Maritime Antarctica.

Global change biology·2026
Same author

Roads, Soil, Snow, and Topography Influence Genetic Connectivity: A Machine Learning Approach for a Peripheral American Badger Population.

Ecology and evolution·2026
Same author

Spatially-nested topologies stabilize meta-ecosystems via cross-scale source-sink dynamics.

Ecology·2026

Model selection criteria like AIC, AICc, and BIC fail in landscape genetics when used with Multiple Regression on Distance Matrices (MRM). These methods show bias towards complex models, especially with larger sample sizes, leading to unreliable results.

Area of Science:

  • Landscape genetics
  • Population genetics
  • Ecological modeling

Background:

  • Model selection is crucial in landscape genetics for understanding genetic, geographic, and environmental relationships.
  • Information Theoretic and Bayesian criteria are commonly used with Multiple Regression on Distance Matrices (MRM).

Purpose of the Study:

  • To evaluate the performance of Akaike's Information Criterion (AIC), AICc, and Bayesian Information Criterion (BIC) for model selection within MRM.
  • To assess how sample size affects the reliability of these model selection criteria in MRM.

Main Methods:

  • Monte Carlo simulations were employed to test model selection criteria.
  • The study varied sample sizes to observe their impact on criterion performance with MRM.

Related Experiment Videos

Main Results:

  • AIC, AICc, and BIC demonstrated a systematic bias, favoring overly complex models with spurious variables.
  • These criteria erroneously assigned high support to incorrect models, with bias increasing with sample size.
  • The issues are likely due to inflated sample sizes and MRM's sum-of-squares partitioning.

Conclusions:

  • The study strongly advises against using AIC, AICc, and BIC for model selection with MRM.
  • Current model selection practices in landscape genetics using these criteria may lead to inaccurate conclusions.