Changing philosophies and tools for statistical inferences in behavioral ecology

Abstract

Recent developments in ecological statistics have reached behavioral ecology, and an increasing number of studies now apply analytical tools that incorporate alternatives to the conventional null hypothesis testing based on significance levels. However, these approaches continue to receive mixed support in our field. Because our statistical choices can influence research design and the interpretation of data, there is a compelling case for reaching consensus on statistical philosophy and practice. Here, we provide a brief overview of the recently proposed approaches and open an online forum for future discussion (https://bestat.ecoinformatics .org/). From the perspective of practicing behavioral ecologists relying on either correlative or experimental data, we review the most relevant features of information theoretic approaches, Bayesian inference, and effect size statistics. We also discuss concerns about data quality, missing data, and repeatability. We emphasize the necessity of moving away from a heavy reliance on statistical significance while focusing attention on biological relevance and effect sizes, with the recognition that uncertainty is an inherent feature of biological data. Furthermore, we point to the importance of integrating previous knowledge in the current analysis, for which novel approaches offer a variety of tools. We note, however, that the drawbacks and benefits of these approaches have yet to be carefully examined in association with behavioral data. Therefore, we encourage a philosophical change in the interpretation of statistical outcomes, whereas we still retain a pluralistic perspective for making objective statistical choices given the uncertainties around different approaches in behavioral ecology. We provide recommendations on how these concepts could be made apparent in the presentation of statistical outputs in scientific papers.

What makes a feather shine? A nanostructural basis for glossy black colours in feathers

“…The inherent loss of power associated with Bonferroni corrections and similar procedures [28,29] led us to present all our results as standardized effect sizes with their associated confidence intervals [30]. Through this approach, the magnitude of the differences and associations, as well as the associated uncertainties, can be evaluated directly, providing a more objective interpretation of both significant and non-significant results [31][32][33].…”

Untested assumptions about within-species sample size and missing data in interspecific studies

“…We preferred focusing on effect sizes rather than significance levels because models with a large number of species would likely support weak, but statistically significant effects that only explain a small proportion of the variance (Nakagawa and Cuthill 2007;Freckleton 2009;Garamszegi et al 2009a). Therefore, we focused on the strength of the relationships (effect size) by considering the precision (95 % CI) by which it can be measured given the available data.…”

A longitudinal study of dominance and aggression in greylag geese (Anser anser)

“…We sequentially deleted fixed terms in order of decreasing significance, starting with any interaction terms in the model (Galwey 2006;Garamszegi et al 2009). Only terms with P , 0.1 (main term) or P , 0.05 (interaction term) remained in the final model.…”

“…To obtain P values for main terms involved in an interaction, we calculated the same model without the interaction term but took all other statistical information from the interaction model. To evaluate the relative importance of terms in the final model, we compared their effect sizes (Garamszegi et al 2009), whereby the effect size of a factor was considered to be the range of effect sizes (minimum to maximum) across the factor levels.…”