Statistics and Visualization for Data Analysis and Inference
MIT OpenCourseWare offers this course on statistical methods for behavioral science research, built as an alternative to standard null-hypothesis significance testing. Topics include data visualization techniques, constructing null distributions through permutation methods, parametric distributions, the generalized linear model, and model-based analysis approaches more broadly. The course assumes some familiarity with MATLAB, Octave, or R, and prior statistics exposure is helpful though not required. Materials come from MIT's Brain and Cognitive Sciences department and follow OCW's standard format of lecture notes, assignments, and readings released under a Creative Commons license. As with other OCW offerings, there is no instructor interaction or certificate, just the raw course materials for self-study. The focus stays practical throughout, aimed at researchers who need to analyze real behavioral data rather than survey statistical theory in the abstract.