Search

Before embarking on data collection, researchers typically compute how many individual observations they should do. This is vital for doing studies with sufficient statistical power, and often a cornerstone in study pre-registrations and grant applications. For traditional statistical tests, one would typically determine an acceptable level of statistical power, (gu)estimate effect size, and then use both values to compute the required sample size. However, for analyses that identify subgroups, statistical power is harder to establish. Once sample size reaches a sufficient threshold, effect size is primarily determined by the number of measured features and the underlying subgroup separation. As a consequence, a priory computations of statistical power are notoriously complex. In this tutorial, I will provide a roadmap to determining sample size and effect size for analyses that identify subgroups. First, I introduce a procedure that allows researchers to formalise their expectations about effect sizes in their domain of choice, and use this to compute the minimally required number of measured variables. Next, I outline how to establish the minimum sample size in subgroup analyses. Finally, I use simulations to provide a reference table for the most popular subgroup analyses: k-means, Ward agglomerative hierarchical clustering, c-means fuzzy clustering, latent class analysis, latent profile analysis, and Gaussian mixture modelling. The table shows the minimum numbers of observations per expected subgroup (sample size) and features (measured variables) to achieve acceptable statistical power, and can be readily used in study design., Comment: Code and data: https://github.com/esdalmaijer/cluster_power_tutorial

Cumulative cultural evolution occurs when adaptive innovations are passed down to consecutive generations through social learning. This process has shaped human technological innovation, but also occurs in non-human species. While it is traditionally argued that cumulative culture relies on high-fidelity social transmission and advanced cognitive skills, here I show that a much simpler system suffices. Cumulative culture spontaneously emerged in artificial agents who navigate with a minimal cognitive architecture of goal-direction, social proximity, and route memory. Within each generation, naive individuals benefitted from being paired with experienced navigators because they could follow previously established routes. Crucially, experienced navigators also benefitted from the presence of naive individuals through regression to the goal. As experienced agents followed their memorised path, their naive counterparts (unhindered by route memory) were more likely to err towards than away from the goal, and thus biased the pair in that direction. This improved route efficiency within each generation. In control experiments, cumulative culture was attenuated when agents' social proximity or route memory were lesioned, whereas eliminating goal-direction only reduced efficiency. These results demonstrate that cumulative cultural evolution occurs even in the absence of sophisticated communication or thought. One interpretation of this finding is that current definitions are too loose, and should be narrowed. An alternative conclusion is that rudimentary cumulative culture is an emergent property of systems that seek social proximity and have an imprecise memory capacity, providing a flexible complement to traditional evolutionary mechanisms., Comment: Code: https://github.com/esdalmaijer/artificial_navigators Data: https://doi.org/10.5281/zenodo.6944184

Many academics use yearly publication numbers to quantify academic interest for their research topic. While such visualisations are ubiquitous in grant applications, manuscript introductions, and review articles, they fail to account for the rapid growth in scientific publications. As a result, any search term will likely show an increase in supposed "academic interest". One proposed solution is to normalise yearly publication rates by field size, but this is arduous and difficult. Here, we propose an simpler index that normalises keywords of interest by a ubiquitous and innocuous keyword, such as "banana". Alternatively, one could opt for field-specific keywords or hierarchical structures (e.g. PubMed's Medical Subject Headings, MeSH) to compute "interest market share". Using this approach, we uncovered plausible trends in academic interest in examples from the medical literature. In neuroimaging, we found that not the supplementary motor area (as was previously claimed), but the prefrontal cortex is the most interesting part of the brain. In cancer research, we found a contemporary preference for cancers with high prevalence and clinical severity, and notable declines in interest for more treatable or likely benign neoplasms. Finally, we found that interest in respiratory viral infections spiked when strains showed potential for pandemic involvement, with SARS-CoV-2 and the COVID-19 pandemic being the most extreme example. In sum, the time is ripe for a quick and easy method to quantify trends in academic interest for anecdotal purposes. We provide such a method, along with software for researchers looking to implement it in their own writing., Comment: The software described in this manuscript can be found on https://github.com/esdalmaijer/bibliobanana

Disgust is a basic emotion that serves to avoid contaminants, and is central to the behavioural immune system. While disgust-motivated avoidance occurs in bonobos and chimpanzees, humans show uniquely high levels of contamination sensitivity. Current theory postulates that human disgust is primarily a genetic adaptation, side-lining social transmission and denying parental modelling. Here, we test whether this strong view is warranted by simulating 100000 years of cultural and biological evolution in various scenarios. We modelled disgust as a trait that governed the extent to which individuals forewent potentially contaminated nutrition. This lowered their risk of gastrointestinal illness, at the cost of increasing starvation risk and birth-interval. As environmental levels and costs of contamination rose, so did relative fitness of high-disgust individuals. Evolutionary shifts in disgust as polygenic trait occurred as a consequence of germ-cell mutations, but were most prominent in populations with high initial genetic variance. Crucially, cultural transmission between generations operated at a higher rate, even if parental modelling was eliminated. This study serves not only as evidence of cultural evolution shaping the behavioural immune system, but also as an illustration of emerging theories that paint affective and cognitive mechanisms as socially transmitted rather than biologically determined functions., Comment: 18 pages, 2 figures