Quick answer
UWSA 2 correlates strongly with the real Step 2 CK score (r = 0.89) but over-predicts by about 3–5 points on average; UWSA 1 over-predicts slightly more. This converter applies a bias correction derived from 5,039 verified student outcomes instead of taking the raw score at face value.
Last updated: July 2026 · How we calculate this
UWSA Overprediction & How We Correct It
UWorld's self-assessments (UWSA 1, 2, 3) are widely used as Step 2 CK predictors, but they have a documented over-prediction bias. UWSA 2 — the most-cited of the three — over-predicts actual Step 2 CK by ~3-5 points on average. UWSA 1 over-predicts more (often 5-8 pts), and both forms exhibit a stronger ceiling effect above 255, where the real-exam gap widens to 10-15 points.
Our correction model applies a base bias offset specific to each UWSA form, plus a progressive ceiling adjustment for scores above 255. The output is a bias-corrected point estimate with an 80% prediction interval anchored on the published UWSA-to-Step-2-CK correlation (r ≈ 0.85-0.90 for UWSA 2).
What this tool doesn't do:we don't reverse-engineer UWorld's proprietary scoring algorithm. We start from the 3-digit score UWSA already reports and apply empirical post-hoc bias correction. We are not affiliated with UWorld.
UWSA Converter FAQs
Why does UWSA over-predict?
UWSAs are calibrated on UWorld user populations who tend to score higher than the full Step 2 CK population. The form difficulty doesn't quite reach operational exam difficulty.
Which UWSA is most accurate?
UWSA 2, taken within 1-2 weeks of the real exam. UWSA 1 over-predicts more. UWSA 3 has limited published validity.
What is the ceiling effect?
Above UWSA 255, the prediction gap widens. Top scorers tend to lose 10-15 points relative to UWSA, even when other practice scores agree. We apply a progressive adjustment in this band.