History of multimodel inference via model selection in wildlife science
Lindberg MS, Schmidt JH, Walker J. 2015. History of multimodel inference via model selection in wildlife science. The Journal of Wildlife Management. 79(5):704-707
We examined changes in the pathways used for inference in The Journal of Wildlife Management (JWM) and 2 other applied journals during recent decades. Although null hypothesis significance testing is still the main approach to inference, use of information-theoretic approaches based on Akaike's Information Criterion (AIC) has rapidly grown to be a common form of inference in JWM and related journals. We observed little growth in the use of other information criteria such as Bayesian Information Criterion (BIC). The use of information criteria for multimodel inference has addressed some of the criticisms of significance testing. However, information criteria still needs to be used appropriately with a priori hypotheses to be valid. In addition, much work remains to be done on application of information criteria to more complex models such as hierarchical and Bayesian models. Published 2015. This article is a U.S. Government work and is in the public domain in the USA.
- Type
- Journal Article
- Authors
- Lindberg, Mark; Schmidt, Joshua; Walker, Johann
- Units
- ARCN
- Keywords
- AIC, history, hypothesis testing, inference, model selection, multimodel