Model selection and model averaging in behavioural ecology: The utility of the IT-AIC framework. Richards, S. A., Whittingham, M. J., & Stephens, P. A. Behavioral Ecology and Sociobiology, 65(1):77--89, October, 2010. ![Model selection and model averaging in behavioural ecology: The utility of the IT-AIC framework [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Behavioural ecologists often study complex systems in which multiple hypotheses could be proposed to explain observed phenomena. For some systems, simple controlled experiments can be employed to reveal part of the complexity; often, however, observational studies that incorporate a multitude of causal factors may be the only (or preferred) avenue of study. We assess the value of recently advocated approaches to inference in both contexts. Specifically, we examine the use of information theoretic (IT) model selection using Akaike’s information criterion (AIC). We find that, for simple analyses, the advantages of switching to an IT-AIC approach are likely to be slight, especially given recent emphasis on biological rather than statistical significance. By contrast, the model selection approach embodied by IT approaches offers significant advantages when applied to problems of more complex causality. Model averaging is an intuitively appealing extension to model selection. However, we were unable to demonstrate consistent improvements in prediction accuracy when using model averaging with IT-AIC; our equivocal results suggest that more research is needed on its utility. We illustrate our arguments with worked examples from behavioural experiments.
@article{ richards_model_2010,
title = {Model selection and model averaging in behavioural ecology: {The} utility of the {IT}-{AIC} framework},
volume = {65},
issn = {0340-5443, 1432-0762},
shorttitle = {Model selection and model averaging in behavioural ecology},
url = {http://link.springer.com/article/10.1007/s00265-010-1035-8},
doi = {10.1007/s00265-010-1035-8},
abstract = {Behavioural ecologists often study complex systems in which multiple hypotheses could be proposed to explain observed phenomena. For some systems, simple controlled experiments can be employed to reveal part of the complexity; often, however, observational studies that incorporate a multitude of causal factors may be the only (or preferred) avenue of study. We assess the value of recently advocated approaches to inference in both contexts. Specifically, we examine the use of information theoretic (IT) model selection using Akaike’s information criterion (AIC). We find that, for simple analyses, the advantages of switching to an IT-AIC approach are likely to be slight, especially given recent emphasis on biological rather than statistical significance. By contrast, the model selection approach embodied by IT approaches offers significant advantages when applied to problems of more complex causality. Model averaging is an intuitively appealing extension to model selection. However, we were unable to demonstrate consistent improvements in prediction accuracy when using model averaging with IT-AIC; our equivocal results suggest that more research is needed on its utility. We illustrate our arguments with worked examples from behavioural experiments.},
language = {English},
number = {1},
urldate = {2015-01-21TZ},
journal = {Behavioral Ecology and Sociobiology},
author = {Richards, Shane A. and Whittingham, Mark J. and Stephens, Philip A.},
month = {October},
year = {2010},
keywords = {Behavioural Sciences, Effect size, Inference, Model weighting, Null hypotheses, Process-based models, Statistics, Zoology, evolutionary biology, statistics},
pages = {77--89}
}
Downloads: 0
{"_id":"NNN5fM9hpsGXkQwPb","bibbaseid":"richards-whittingham-stephens-modelselectionandmodelaveraginginbehaviouralecologytheutilityoftheitaicframework-2010","downloads":0,"creationDate":"2015-09-17T14:54:17.284Z","title":"Model selection and model averaging in behavioural ecology: The utility of the IT-AIC framework","author_short":["Richards, S.<nbsp>A.","Whittingham, M.<nbsp>J.","Stephens, P.<nbsp>A."],"year":2010,"bibtype":"article","biburl":"https://api.zotero.org/users/209774/collections/BC582FG5/items?key=dqhdckbzHdCc0oD6wUOvaI99&format=bibtex&limit=100","bibdata":{"abstract":"Behavioural ecologists often study complex systems in which multiple hypotheses could be proposed to explain observed phenomena. For some systems, simple controlled experiments can be employed to reveal part of the complexity; often, however, observational studies that incorporate a multitude of causal factors may be the only (or preferred) avenue of study. We assess the value of recently advocated approaches to inference in both contexts. Specifically, we examine the use of information theoretic (IT) model selection using Akaike’s information criterion (AIC). We find that, for simple analyses, the advantages of switching to an IT-AIC approach are likely to be slight, especially given recent emphasis on biological rather than statistical significance. By contrast, the model selection approach embodied by IT approaches offers significant advantages when applied to problems of more complex causality. Model averaging is an intuitively appealing extension to model selection. However, we were unable to demonstrate consistent improvements in prediction accuracy when using model averaging with IT-AIC; our equivocal results suggest that more research is needed on its utility. We illustrate our arguments with worked examples from behavioural experiments.","author":["Richards, Shane A.","Whittingham, Mark J.","Stephens, Philip A."],"author_short":["Richards, S.<nbsp>A.","Whittingham, M.<nbsp>J.","Stephens, P.<nbsp>A."],"bibtex":"@article{ richards_model_2010,\n title = {Model selection and model averaging in behavioural ecology: {The} utility of the {IT}-{AIC} framework},\n volume = {65},\n issn = {0340-5443, 1432-0762},\n shorttitle = {Model selection and model averaging in behavioural ecology},\n url = {http://link.springer.com/article/10.1007/s00265-010-1035-8},\n doi = {10.1007/s00265-010-1035-8},\n abstract = {Behavioural ecologists often study complex systems in which multiple hypotheses could be proposed to explain observed phenomena. For some systems, simple controlled experiments can be employed to reveal part of the complexity; often, however, observational studies that incorporate a multitude of causal factors may be the only (or preferred) avenue of study. We assess the value of recently advocated approaches to inference in both contexts. Specifically, we examine the use of information theoretic (IT) model selection using Akaike’s information criterion (AIC). We find that, for simple analyses, the advantages of switching to an IT-AIC approach are likely to be slight, especially given recent emphasis on biological rather than statistical significance. By contrast, the model selection approach embodied by IT approaches offers significant advantages when applied to problems of more complex causality. Model averaging is an intuitively appealing extension to model selection. However, we were unable to demonstrate consistent improvements in prediction accuracy when using model averaging with IT-AIC; our equivocal results suggest that more research is needed on its utility. We illustrate our arguments with worked examples from behavioural experiments.},\n language = {English},\n number = {1},\n urldate = {2015-01-21TZ},\n journal = {Behavioral Ecology and Sociobiology},\n author = {Richards, Shane A. and Whittingham, Mark J. and Stephens, Philip A.},\n month = {October},\n year = {2010},\n keywords = {Behavioural Sciences, Effect size, Inference, Model weighting, Null hypotheses, Process-based models, Statistics, Zoology, evolutionary biology, statistics},\n pages = {77--89}\n}","bibtype":"article","doi":"10.1007/s00265-010-1035-8","id":"richards_model_2010","issn":"0340-5443, 1432-0762","journal":"Behavioral Ecology and Sociobiology","key":"richards_model_2010","keywords":"Behavioural Sciences, Effect size, Inference, Model weighting, Null hypotheses, Process-based models, Statistics, Zoology, evolutionary biology, statistics","language":"English","month":"October","number":"1","pages":"77--89","shorttitle":"Model selection and model averaging in behavioural ecology","title":"Model selection and model averaging in behavioural ecology: The utility of the IT-AIC framework","type":"article","url":"http://link.springer.com/article/10.1007/s00265-010-1035-8","urldate":"2015-01-21TZ","volume":"65","year":"2010","bibbaseid":"richards-whittingham-stephens-modelselectionandmodelaveraginginbehaviouralecologytheutilityoftheitaicframework-2010","role":"author","urls":{"Paper":"http://link.springer.com/article/10.1007/s00265-010-1035-8"},"keyword":["Behavioural Sciences","Effect size","Inference","Model weighting","Null hypotheses","Process-based models","Statistics","Zoology","evolutionary biology","statistics"],"downloads":0},"search_terms":["model","selection","model","averaging","behavioural","ecology","utility","aic","framework","richards","whittingham","stephens"],"keywords":["behavioural sciences","effect size","inference","model weighting","null hypotheses","process-based models","statistics","zoology","evolutionary biology","statistics"],"authorIDs":[],"dataSources":["QF2hrLefLEagsqLdY"]}