Uninformative Parameters and Model Selection Using Akaike's Information Criterion
2010/07/26 by Todd W. Arnold · 29 citations
Computer Science · Environmental Science · #Data Analysis with R #Species Distribution and Climate Change #Bayesian Modeling and Causal Inference
paper · doi:10.2193/2009-367
Abstract
As use of Akaike's Information Criterion (AIC) for model selection has become increasingly common, so has a mistake involving interpretation of models that are within 2 AIC units (ΔAIC ≤ 2) of the top-supported model. Such models are <2 ΔAIC units because the penalty for one additional parameter is 2 AIC units, but model deviance is not reduced by an amount sufficient to overcome the 2-unit penalty and, hence, the additional parameter provides no net reduction in AIC. Simply put, the uninformative parameter does not explain enough variation to justify its inclusion in the model and it should not be interpreted as having any ecological effect. Models with uninformative parameters are frequently presented as being competitive in the Journal of Wildlife Management, including 72% of all AIC-based papers in 2008, and authors and readers need to be more aware of this problem and take appropriate steps to eliminate misinterpretation. I reviewed 5 potential solutions to this problem: 1) report all models but ignore or dismiss those with uninformative parameters, 2) use model averaging to ameliorate the effect of uninformative parameters, 3) use 95% confidence intervals to identify uninformative parameters, 4) perform all-possible subsets regression and use weight-of-evidence approaches to discriminate useful from uninformative parameters, or 5) adopt a methodological approach that allows models containing uninformative parameters to be culled from reported model sets. The first approach is preferable for small sets of a priori models, whereas the last 2 approaches should be used for large model sets or exploratory modeling.
Cited by
- Migratory Birds Advance Spring Arrival and Egg‐Laying in the Arctic, Mostly by Travelling Faster
- Divorce Rates Better Predict Population‐Level Reproductive Success in Little Penguins Than Foraging Behaviour or Environmental Factors
- Identifying the demographic pathways linking environmental covariates to population dynamics in an avian migrant
- Nest-site selection, nesting success, and the value of older pine plantations for the globally Endangered Giant Nuthatch <i>Sitta magna</i> in northern Thailand
- State-Dependent and Social Modulation of Circulating Glucocorticoids in a Nomadic Songbird, the Red Crossbill ( <i>Loxia Curvirostra</i> )
- Patterns of partial brood loss in relation to group size, brood size, and nest timing support the resource tracking hypothesis in a cooperative breeder, the Red-cockaded Woodpecker ( <i>Dryobates borealis</i> )
- Spatial variation in grizzly bear ( <i>Ursus arctos</i> ) dietary niche across British Columbia
- Interspecific carnivore competition and ungulate predation correlate with predator species richness
- Cost of migration increased during an outbreak of highly pathogenic avian influenza virus
- Effects of multiple mammalian herbivores and climate on grassland–shrubland transitions in the Chihuahuan Desert
- Habitat use and co-occurrence of two species of tody-tyrant, <i>Hemitriccus</i> (Aves: Rhynchocyclidae) in a fragment of montane forest in north-eastern Brazil
- Temporal Not Foraging Factors Affect Reproductive Success for American Oystercatchers (Haematopus palliatus) on Two Barrier Islands in Virginia, U.S.A.
- Of droughts and flooding rains: koala population occupancy in a north-east New South Wales conservation reserve, and implications for range-wide surveys
- Changes in visits to wildlife water developments by bighorn sheep and mule deer following a wildland fire
- Sex-specific colony attendance before and during the fertilization window in Nazca boobies
- Defensive coloration is not a reliable indicator of fungal infection in aposematic poison frogs
- Beyond message framing: Participant characteristics predict social acceptability of increased deer culling in Scotland
- Social and ecological influences on human coexistence with gray wolves
- Multiscale habitat selection of an imperiled subspecies of Least Chipmunk (<i>Neotamias minimus</i>): when a generalist is really a specialist
- Intraspecific leaf trait variation mediates edge effects on litter decomposition rate in fragmented forests
- Wolves on ice: habitat use and selection of water bodies by wolves during winter
- Unveiling the ghost: short-eared dog (Atelocynus microtis) distribution, activity patterns, habitat use, relative abundance, and occupancy in Bolivia
- Long-term monitoring of talus microclimates portrays crevices as critical thermal refuge for vulnerable alpine species
- Interspecific effects of invasive wild pigs (<i>Sus scrofa</i>) on native nine-banded armadillos (<i>Dasypus novemcinctus</i>)
- A year-long camera-trap dataset for assessing spatial occurrence and diel activity of sympatric ungulates in South Korea
- Direct and indirect ecological drivers of bird communities in intensively cultivated landscapes
- Long-Term Noninvasive Genetic Monitoring Guides Recovery of the Endangered Columbia Basin Pygmy Rabbits (Brachylagus idahoensis)
- Female North American passerine birds have lower apparent survival than males
- From open landscapes to forest refugia: Human encroachment and climate change as drivers of red deer (Cervus elaphus) niche shifts over 21,000 years
Related