Computerized Adaptive Testing: Item Response Engine

This simulator runs a real computerized adaptive test (CAT) loop over a randomly generated item bank, each item defined by a 3-parameter logistic (3PL) item response curve. Every step, the engine scans the remaining bank and administers whichever item currently carries the most Fisher information about the examinee's ability — visualized as the amber bar — simulates a correct or incorrect response using the true ability you set, and re-estimates ability with a Bayesian expected-a-posteriori (EAP) calculation over a normal prior, shown live along with its standard error. The back-row ribbon traces the bank's aggregate information curve so you can see where measurement precision is concentrated, and the test stops automatically once the standard error drops below your chosen threshold — the same targeting-and-stopping logic used by real adaptive exams like the GRE and NWEA MAP.