Meta-analysis is a powerful tool for synthesizing behavioural research and identifying general patterns. However, we should consider if the conclusions we draw from these analyses are truly representative across animal groups, or whether our conclusions are shaped by taxonomic biases in the underlying research. For example, in animal behaviour, vertebrates are over-represented in the research we conduct. The occurrence of this taxonomic imbalance raises concerns about the validity of generalizations drawn in the field. To study this issue, we examined the meta-analyses published in Animal Behaviour, Behavioral Ecology and Behavioral Ecology and Sociobiology from 2000-2024. We then conducted a 'meta-meta-analysis' to calculate the degree to which overall effects in prior meta-analytical results may have been misestimated owing to taxonomic bias. We found that taxonomic biases in the primary research strongly influence effect size estimates in meta-analyses and may lead to improper inferences and generalizations. On average, meta-analytical averages are significantly misestimated and taxonomic bias also results in apparent changes in statistical significance. Because meta-analyses aggregate data, they propagate the biases present in an area of research, leading to potentially incorrect generalizations. Addressing this taxonomic bias is critical for generalizations that describe the true richness of animal behaviour. (c) 2026 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
原文链接
分享
引用
