Evolutionary Biology

Statistics may seem like an unavoidable burden to many empiricists. However, it not only represents an indispensable tool for understanding biological processes, enabling researchers to fine-tune experimental design and draw robust inferences from their data, but (in our opinion) can also be good fun too. We follow two integral lines to strengthen the statistical expertise in our research community. First, we track recent developments in different fields in order to make use of novel statistical approaches or established approaches in novel contexts to keep pace with best practice in biostatistical analysis. Second, we offer training in the form of a weekly Stats Club that aims to discuss our own research projects both before and after data collection from a statistical perspective.

During our daily work, we follow a philosophy that aims to match the statistical models we apply to the experimental design in order to estimate biologically relevant effect sizes. Naturally, the data analysis is intimately linked to the biological questions that we aim to answer. Most of our analyses are concerned with the analysis of animal behavior and evolution and our data sets are often complex with multiple levels of non-independence. Mixed effects models are therefore often the tool of choice and we have accumulated considerable knowledge in their application. While mixed effects models are widely used in ecology and evolution,interpretation is often limited to either the fixed or the random part of the model. We follow a philosophy that aims to interpret all components of statistical models in biological terms.

Our Stats Club offers a discussion forum for researchers at all levels, including Bachelor, Master and PhD students. Current research projects can be presented at any stage, but we encourage a presentation before data collection, because the discussion in the Club might result in suggestions on how to improve experimental design for example to maximise statistical power or avoid confounding effects. However, also in the post-data collection phase, the Stats Club might offer fresh insights and ideas fuelled by the combined expertise of the Club members. Besides discussing data analysis of ongoing projects, we regularly discuss key stats papers that provide insights into novel techniques and offer solutions to common problems.