Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Burton et al. probe the question of moral contagion through out-of-sample prediction, model comparisons and specification curve analyses, demonstrating the limitations of conclusions based on large-scale, observational social media datasets alone.
Bhattacharjee and Schaeffer et al. map exclusive breastfeeding (EBF) in 94 low- and middle-income countries (LMICs), finding increased EBF practice and reduced subnational variation across the majority of LMICs from 2000 to 2018. However, only six LMICs will meet WHO’s target of ≥70% EBF by 2030 nationally, and only three will achieve this in all districts.
The implementation of COVID-19 stay-at-home policies was associated with a considerable drop in urban crime in 27 cities across 23 countries. More stringent restrictions over movement in public space were predictive of larger declines in crime.
Sandeford challenges the standard model of intensification using an ethnographic dataset that describes food production in 40 human societies ranging in complexity from small-scale foraging bands to large-scale agricultural states.
Petitet et al. demonstrate that while human decision-making typically follows a speed–efficiency trade-off in sampling information, it is possible to break this trade-off, sampling both faster and more efficiently. In a computational model, they show that incorporating the cost of cognitive effort provides an improved account of this sampling behaviour.
Existing measures of cognitive impulsivity have suboptimal reliability and validity. Here, the authors introduce a new online test battery featuring a gamified interface and show that it aids in the prediction of real-world, addiction-related problems.
Judd and Klingberg analysed data from more than 17,000 children who performed mathematical training together with randomly assigned training on spatial tasks. The type of cognitive training had a significant impact on mathematical learning.
Using administrative data from Denmark, Sønderskov et al. find that terrorism in the country of origin is associated with poorer mental health among refugees, as indicated by an increased use of psychotropic drugs.
How do we evaluate art? Here, Iigaya et al. show that aesthetic preferences for visual art can be predicted by a mixture of low- and high-level image features, and that a convolutional neural network trained only on object recognition naturally encodes many of these features.
Scientists fear that systemic incentives lead to poor science. Stewart and Plotkin use modelling to show how a scientific process emphasizing the use of theory to select hypotheses can allow good science to thrive in the face of pressure to publish.
In a pre-registered meta-analysis, Parry et al. find that, when self-reported media use is compared with digital logs of media use, subjective judgements are often inaccurate. This suggests caution when self-reports are used to test associations between media use and other outcomes.
Trust in science is important for vaccine confidence, and this is true for countries as well as individuals. Sturgis et al. find that confidence in vaccination is higher in countries where people agree that scientists are trustworthy.
Two eye-tracking experiments revealed that adolescents and older adults look less at social stimuli during a face-to-face conversation and while navigating the real world compared with young adults, which may affect the success of their social interactions.
Schmid et al. present a unified framework for direct and indirect reciprocity, exploring how people choose to cooperate on the basis of either their direct experience with others (direct reciprocity) or the others’ general reputation (indirect reciprocity).
Storrs et al. train unsupervised generative neural networks on glossy surfaces and show how gloss perception in humans may emerge in an unsupervised fashion from learning to model statistical structure.
Why does cooperation decline? Burton-Chellew and West compare data from 237 public goods games and find that cooperation declines faster when learning is easier.
The most parsimonious network of routes taken by the first people navigating Sahul emerge from landscape-based rules, which can also be applied to other peopling events, to quantify the likely patterns of the peopling of Earth.
Merkley and Loewen find that anti-intellectualism (distrust in experts and intellectuals) is linked to COVID-19 (mis)perceptions, compliance with public health directives and information search using survey and experimental data from Canada.
Ruggeri et al. tested perceptions of opposing political party members in 10,207 participants from 26 countries. Results show that beliefs about others are overly negative but could be more realistic with transparency about actual group beliefs.
Using relationship data for over 5,000 hospital employees inferred from 2 years of workplace cafeteria sales data, Levy et al. found significant associations in the healthfulness of food purchases between socially connected employees.