A guest post by Carsten F. Dormann, University of Freiburg In “biodiversity science”, species richness takes a very prominent role. It is the most analysed response, and an extremely common predictor, too. Of course everyone in the field (and lab) knows that “species richness” is a somewhat weird variable. When Cedar Creek and Jena experiment report results with “number of species sown” on the x-axis (e.g. Tilman & Downing 1994, Weisser et al. 2017), we knew that species are not comparable in traits and abundance and hence effects on plot-level aboveground biomass must be mediate by these ..read more

A guest post by Carsten Dormann. A recent talk in the colloquium of the Freiburg Centre for Data and Modelling (FDM) on a “Mystery in Machine Learning” (by Lena Kulla) brought a 2019 paper by Belkin et al. to my attention. In it, they showed a graph of a model’s error (squared loss) over model complexity (which they measure by the number of parameters/weights in a deep neural network). One would expect the typical bias-variance trade-off, i.e. a constant improvement on fitting to the data (training) and a U-shaped curve for the prediction to new data (test). Instead, they report a “double-desc ..read more

The latest update of the DHARMa R package (0.4.6) includes an option to perform simulated likelihood ratio tests (LRTs) for GLMMs based on a parametric bootstrap. I wanted to shortly comment on how this works and why this is useful. Why is this useful? A well-known issue with mixed models is that the df lost for a random effect (RE) depend on the fitted RE variance, which controls the freedom (or conversely, the shrinkage) of the RE estimates. As a consequence, the df lost by a RE a not fixed a priori, but have to be estimated after fitting the model (adaptive shrinkage). One could add that ev ..read more

This is a guest post by Carsten Dormann. It is his invited keynote lecture at the 2021 Annual Meeting of the Gesellschaft für Ökologie in Braunschweig (specifically on 1 Sept 2021 at 8:15). The topic of the conference was “From Science to Policy”, and he deemed it necessary to question whether thus step is not a bit early. Dear fellow ecologists, dear ladies and gentlemen, it’s an honour to be given the opportunity to talk to you about something I deeply care about. As you will see, it is not a scientific talk, but rather reflections on the current state of ecology as a science, and more speci ..read more

A guest post by Carsten F. Dormann. In October 2021, Paul Ehrlich wrote a “Correspondence” in Nature with which I seemed to sympathise to a surprisingly large part. But the last sentence wiped out all agreement, as it revealed a fundamentally different view of the role and “duty” of “scientists”. I wrote this blog at that time, and stashed it away. Now, a year later, with an actual war raging in a country close by, Ehrlich’s rhetoric seems more misplaced than ever. Here’s what I wrote then. “Scientists are in a war for the future of humanity: they must get off of their peacetime footing.” Paul ..read more

Note also the blog post by Daniel Lakens on the same topic here. Background Criticism of p-values has been rampant in recent years, as were predictions of their imminent demise. When looking closer, most of these critiques are actually not sparked by the definition of the p-value (Fisher, 1925) as such, but rather by the NHST framework (Neyman and Pearson, 1933; Neyman, 1950), which introduces a cut-off (significance level alpha) to transform the p-value into a binary decision (significant / n.s.) while trading off Type I/II error rates in the process. Just to be clear: I fully agree with many ..read more

In regression analysis, a common problem is to decide on the right functional form of the fitted model. On the one hand, we would like to make the model as flexible as possible so that it can adjust itself bias-free to the true data-generating process. On the other hand, the more freedom we give the model, the more uncertain parameter estimates get (= variance), which leads to total model error increasing with complexity after a certain point. This phenomenon, known as the bias-variance trade-off, leads to the insight that we should limit or penalise model complexity to get reasonable inferenc ..read more

A guest post by Oskar Hagen (iDiv) | www.hagen.bio | @hagen_oskar What is the cause and what are the processes that gave rise to Earth’s biodiversity patterns through space and time? Much research has been devoted to describing these patterns, and over the years, the fields of macroecology and macroevolution have slowly transitioned from a mainly correlational to a more mechanistic perspective (1, 2). The challenge with understanding the mechanisms of macroevolution is that, while evolution has in principle simple general rules, it operates across a complex dynamic world. As a result ..read more

Do you remember the notorious hurricane / himmicane study (Jung et al., PNAS, 2014)? At the time, there was a heavy backlash against the study, and probably rightly so, as the statistical analysis turns out to be highly unstable against a change of the regression formula. You can find some links here. Over the years, however, I have found that this study has at least one virtue: it is an excellent example for teaching students about the importance of selecting the right functional relationship when running an analysis, and that substantial “dark” uncertainty can arise from these researcher de ..read more

tl;dr: DHARMa tests will pick up on overdispersion before you see a rise of Type I error. Overdispersion is a common problem in GL(M)Ms with fixed dispersion, such as Poisson or binomial GLMs. Here an explanation from the DHARMa vignette: GL(M)Ms often display over/underdispersion, which means that residual variance is larger/smaller than expected under the fitted model. This phenomenon is most common for GLM families with constant (fixed) dispersion, in particular for Poisson and binomial models, but it can also occur in GLM families that adjust the variance (such as the beta or negative bin ..read more