A new approach to quantify isotopic (co)variation across levels of biological organization using double‐hierarchical generalized linear models

Abstract Double‐hierarchical generalized linear models (DHGLMs) offer a powerful yet overlooked quantitative approach for analysing ecological data with hierarchical structures. DHGLMs account for unit‐specific residual variances, allowing the estimation of variance components that expand our ability to quantify biodiversity patterns across levels of biological organization—from individuals to communities. A particularly promising application of these models lies in isotopic ecology, where stable isotopes of carbon (𝛿 13 C) and nitrogen (𝛿 15 N) are used to quantify continuous trophic (co)variation across individuals, populations and communities. Here, we show how DHGLMs can be a powerful tool for ecologists interested in trophic patterns and demonstrate the application of these models to isotopic data, showcasing study systems across three levels of biological organization. These examples illustrate how DHGLMs decompose (co)variance components within and across biological units of interest, uncover patterns of diet variation and ecological interactions. We also discuss the advantages, limitations and potential of DHGLMs for advancing research on different questions related to variation in diet and specialization at different biological levels.