learning-experiment

To analyze A/B testing for educational interventions, use randomized assignment and pre/post designs to compute effect sizes and apply ANCOVA adjustment. This accurately quantifies the causal impact of educational interventions on student outcomes.

IRT modeling estimates item discrimination and difficulty using 1PL/2PL/3PL concepts to evaluate assessment quality. It computes item and test information to infer measurement precision across different student ability levels.

To fit learning and forgetting curves from trial data, apply power law fitting for learning trajectories and Ebbinghaus-style retention modeling for forgetting. This estimates retention and mastery scenarios from post-learning measurements.

Yes, you can use Python with pandas, statsmodels, numpy, scipy, and scikit-learn for educational experimental design. These dependencies support statistical modeling, curve fitting, and effect size computation for learning experiments.

To estimate student ability and learning trajectories over time, model student responses using IRT concepts and fit learning curves. This produces parameter estimates that quantify ability, growth, and retention from response and timing data.