Understanding Science
Unknown
A free UC Berkeley Museum of Paleontology site explaining how science actually works: testing ideas against evidence, designing controlled experiments, peer review, and common misconceptions. Readers learn to distinguish scientific explanations from untested claims and judge how studies use controls.
More resources on Control Groups
seeing-theory.brown.edu
Seeing Theory is an interactive, browser-based resource from Brown University that uses visual simulations to explain fundamental probability and statistics concepts. It offers interactive demos and explanations on topics like probability, distributions, sampling, the central limit theorem, confidence intervals, hypothesis testing, and Bayesian reasoning.
Designing Studies
Learn how to design effective studies with control groups! Created by Sal Khan and team.
NIST/SEMATECH e-Handbook of Statistical Methods
A free online statistics reference produced by NIST with SEMATECH, covering process characterization, control charts, capability analysis, reliability, and design of experiments. Includes worked case studies and Dataplot code, so you can apply each method to real measurement data.
Experimental and Quasi-Experimental Designs for Generalized Causal Inference
The standard reference on field experimentation, cataloguing threats to internal, external, construct and statistical-conclusion validity, and mapping which design (randomized, regression-discontinuity, interrupted time series, non-equivalent control group) rules out which threat.
Understanding and Misunderstanding Randomized Controlled Trials
Working-paper version of the Social Science and Medicine article arguing that randomization does not automatically balance covariates in finite samples, that trial results need not transfer to other populations, and where control groups mislead.
Randomization (J-PAL Research Resources)
The operational half of the topic that textbooks skip: how you physically construct the control group, what to do when strata do not divide evenly, whether to randomize individuals or clusters, and how to interpret a failed balance test. Practitioner guide to assigning units between treatment and control: simple, permuted and stratified randomization, choosing the unit of assignment, handling misfits and spillovers, checking baseline balance, and reproducible Stata code for each step.