My go-to resources: Recommended readings, tools, and more.
As a behavioral researcher using data science, I often field questions about the tools I use for programming experiments and analyzing data. Colleagues and students, especially those new to behavioral science and data-driven research, are curious about my recommendations. Below is a personal, non-exhaustive list of resources I find valuable - from academic papers to software tools. Whether you're experienced or just starting, you might find something useful here for your research.
Recent articles I enjoyed reading
What is a cell type, really? The quest to categorize life’s myriad forms
News Feature by Amber Dance. Nature, Sept 2024. (2024).
Defining cell types exemplifies the challenge of categorizing abstract scientific concepts. This complexity isn’t unique to biology; social sciences face similar hurdles when classifying, for example, strategic interactions. Notably, this article highlights two crucial points: even imprecisely defined concepts can prove useful, and taxonomies should be evaluated based on their ability to answer questions rather than as standalone structures. These insights are surely valuable for researchers across disciplines.
Bacteria use ‘crazy molecular mechanism’ to fight viruses
Commentary on the fascinating Science paper by Wilkinson et al. (2024).
The strategic methods bacteria employ to combat viruses offer valuable insights. This demonstrates that gaining even a one-step advantage is often sufficient for success, a principle that extends to human interactions as well.
AI models collapse when trained on recursively generated data
Shumailov, I., Shumaylov, Z., Zhao, Y., Papernot, N., Anderson, R., & Gal, Y. (2024). Nature, 631(8022), 755-759.
Language is primarily a tool for communication rather than thought
Fedorenko, E., Piantadosi, S. T., & Gibson, E. A. (2024). Nature, 630(8017), 575-586.
It’s unfortunate that the authors of this nice paper missed the opportunity to quote the work of the great pshychologist B. F. Skinner. He was the first to propose that language developed as a tool for communication; he wrote “The human species took a unique evolutionary step when its vocal musculature came under operant control and language was born. People could then tell, as well as show, each other what to do. Extraordinarily complex social environments or cultures evolved, and they gave the species its extraordinary power.“ (American Psychologist, 1986, 41(5), p. 568)
Bumblebees socially learn behaviour too complex to innovate alone
Bridges, A. D., Royka, A., Wilson, T., Lockwood, C., Richter, J., Juusola, M., & Chittka, L. (2024). Nature, 627(8004), 572-578.
This is a fascinating example of social learning in non-human animals shows that we’re not that special, after all... Also chimps have been shown to have a similar ability to learn from others (see Van Leeuwen, E. J., DeTroy, S. E., Haun, D. B., & Call, J. (2024),Nature Human Behaviour, 1-12.).
Data science
Improving Your Statistical Inferences
A blog by Professor Daniel Lakens (University of Technology Eindhoven) to statistical analysis in behavioral research.
Experience Statistics
This website collects useful or experimental statistical widgets. They can be used for teaching, exploration, or for performing specialized calculations without the need to go through a complete R installation.
R Psychologist - Visualizations
This website by Dr. Kristoffer Magnusson provides a collection of interactive visualizations to help you understand statistical concepts. It is a great resource for students and researchers alike.
Project Tier Protocol
The TIER Protocol specifies the contents and organization of reproduction documentation for a project involving computations with statistical data.
Development tools and Frameworks
PartyKit
PartyKit is an open source deployment platform for AI agents, multiplayer and local-first apps, games and websites.
D3 on Observable
D3.js is a JavaScript library for producing dynamic, interactive data visualizations in web browsers.
LangChain
LangChain is a framework to build with LLMs by chaining interoperable components.