Why do we still use stepwise modelling in ecology and behaviour?
2006/07/21 by MARK J. WHITTINGHAM, Mark J. Whittingham, Philip A. Stephens +5 · 1,518 citations
Economics, Econometrics and Finance · Environmental Science · Mathematics · #Artificial intelligence #Biology #Computer science #Ecology #Ecology and Vegetation Dynamics Studies #Econometrics #Economic and Environmental Valuation #Inference #Machine learning #Mathematics #Model selection #Regression #Regression analysis #Selection (genetic algorithm) #Set (abstract data type) #Species Distribution and Climate Change #Statistical inference #Statistics #Stepwise regression
paper · open access · doi:10.1111/j.1365-2656.2006.01141.x
published in Journal of Animal Ecology 75(5), 1182-1189 (Wiley)
openalex publication_date 2006/07/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/01
Abstract
1. The biases and shortcomings of stepwise multiple regression are well established within the statistical literature. However, an examination of papers published in 2004 by three leading ecological and behavioural journals suggested that the use of this technique remains widespread: of 65 papers in which a multiple regression approach was used, 57% of studies used a stepwise procedure. 2. The principal drawbacks of stepwise multiple regression include bias in parameter estimation, inconsistencies among model selection algorithms, an inherent (but often overlooked) problem of multiple hypothesis testing, and an inappropriate focus or reliance on a single best model. We discuss each of these issues with examples. 3. We use a worked example of data on yellowhammer distribution collected over 4 years to highlight the pitfalls of stepwise regression. We show that stepwise regression allows models containing significant predictors to be obtained from each year's data. In spite of the significance of the selected models, they vary substantially between years and suggest patterns that are at odds with those determined by analysing the full, 4-year data set. 4. An information theoretic (IT) analysis of the yellowhammer data set illustrates why the varying outcomes of stepwise analyses arise. In particular, the IT approach identifies large numbers of competing models that could describe the data equally well, showing that no one model should be relied upon for inference.
Cited by
- Risk‐sensitive allocation in seasonal dynamics of fat and protein reserves in a long‐lived mammal
- Sexual dimorphism, body size, bite force and male mating success in tuatara
- Multimodel Approaches Are Not the Best Way to Understand Multifactorial Systems
- Using Lasso for Predictor Selection and to Assuage Overfitting: A Method Long Overlooked in Behavioral Sciences
- Native trees are related to advanced bird breeding phenology and increased reproductive success along an urban gradient
- What makes a forest growth model climate-sensitive? An examination of statistical and silvicultural model needs under climate change
- The untold story of Sea Caves: Lost and found anthropogenic debris in confined marine environments
- Large field-of-view, distortion-corrected off-axis parabolic mirror relay microscope
- Anthropogenic activities influence the abandonment of Bearded Vulture (Gypaetus barbatus) territories in southern Africa
- Generalized linear mixed models: a practical guide for ecology and evolution
- AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons
- Transcriptomic evidence of cytokine storm and sepsis in little brown bats exposed to white-nose syndrome
- Life‐history characteristics of mule deer: Effects of nutrition in a variable environment
- A general theory of multimetric indices and their properties
- Errors in Statistical Inference Under Model Misspecification: Evidence, Hypothesis Testing, and AIC
- Seasonal variation in banteng density and habitat use across protected areas and buffer zones in western Thailand
- Determinants of Terrestrial and Limnic Species Richness in Germany
- Generalizing hierarchical and variation partitioning in multiple regression and canonical analyses using the rdacca.hp R package
- A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion
- Multi-generational fidelity, ecological and social determinants of roosting in a cooperatively breeding bird ( Argya squamiceps )
- Vineyard design and plant material choices effect on grapevine yield: analysis of a big dataset in the south of France
- A Closer Look at Seabird and Marine Mammal Bycatch Data in Alaska’s Longline Groundfish and Pacific Halibut Fisheries: A Reassessment with Open Access and Machine Learning Ensembles Explicit in Space and Time Shows Deficiencies
- Multimodel inference in ecology and evolution: challenges and solutions
- Environmental modeling of occurrence of dolphinfish (Coryphaena spp.) in the Pacific Ocean off Mexico reveals seasonality in abundance, hot spots and migration patterns
- Resilience in animal care professions: does the stress shield model fit?
- Fine-scale community and genetic structure are tightly linked in species-rich grasslands
- Global evidence that deforestation amplifies flood risk and severity in the developing world
- A meta‐analysis of context‐dependency in plant response to inoculation with mycorrhizal fungi
- Quantifying variation among garden plants in attractiveness to bees and other flower‐visiting insects
- Effect size, confidence interval and statistical significance: a practical guide for biologists
- Model selection and model averaging in behavioural ecology: the utility of the IT-AIC framework
- Lemurs in a dying forest: Factors influencing lemur diversity and distribution in forest remnants of north-eastern Madagascar
- Surviving in a Cosexual World: A Cost-Benefit Analysis of Dioecy in Tropical Trees
- Prevalence of avian influenza and host ecology. [europepmc]
- Can fertility signals lead to quality signals? Insights from the evolution of primate sexual swellings. [europepmc]
- Eco-evolutionary dynamics: disentangling phenotypic, environmental and population fluctuations. [europepmc]
- Pediatric triglycerides predict cardiovascular disease events in the fourth to fifth decade of life. [europepmc]
- Global drivers of human pathogen richness and prevalence. [europepmc]
- The bigger they come, the harder they fall: body size and prey abundance influence predator-prey ratios. [europepmc]
- Cryptic multiple hypotheses testing in linear models: overestimated effect sizes and the winner's curse. [europepmc]
- Baseline risk factors that predict the development of open-angle glaucoma in a population: the Los Angeles Latino Eye Study. [europepmc]
- Early warning of West Nile virus mosquito vector: climate and land use models successfully explain phenology and abundance of Culex pipiens mosquitoes in north-western Italy. [europepmc]
- Temporal variability is a personalized feature of the human microbiome. [europepmc]
- Towards a general framework for predicting threat status of data-deficient species from phylogenetic, spatial and environmental information. [europepmc]
- Reliable Change on Neuropsychological Tests in the Uniform Data Set. [europepmc]
- Comparison of subset selection methods in linear regression in the context of health-related quality of life and substance abuse in Russia. [europepmc]
- Predicting Mental Imagery-Based BCI Performance from Personality, Cognitive Profile and Neurophysiological Patterns. [europepmc]
- Egg discrimination along a gradient of natural variation in eggshell coloration. [europepmc]
- The earth is flat ( p > 0.05): significance thresholds and the crisis of unreplicable research. [europepmc]
- Identifying climate drivers of infectious disease dynamics: recent advances and challenges ahead. [europepmc]
- Glucocorticoids modulate gastrointestinal microbiome in a wild bird. [europepmc]
- A brief introduction to mixed effects modelling and multi-model inference in ecology. [europepmc]
- A simple function for full-subsets multiple regression in ecology with R. [europepmc]
- Evolutionary history of plant hosts and fungal symbionts predicts the strength of mycorrhizal mutualism. [europepmc]
- Diet modulates the relationship between immune gene expression and functional immune responses. [europepmc]
- Asynchronous carbon sink saturation in African and Amazonian tropical forests. [europepmc]
- Plant mixture balances terrestrial ecosystem C:N:P stoichiometry. [europepmc]
- Do providers use computerized clinical decision support systems? A systematic review and meta-regression of clinical decision support uptake. [europepmc]
Related