Predict Your USMLE Step 3 Score

Step 3 Score Prediction: How It Actually Works

Step 3 is the only USMLE that still returns a 3-digit score, and predicting it well means combining several signals rather than reading a single practice test. The model blends your Step 2 CK score, UWorld and UWSA performance, NBME Forms 6/7, and a CCS adjustment into one estimate with a pass probability.

Step 2 CK: the strongest anchor (r = 0.68)

Your Step 2 CK score is the single best predictor of Step 3 — it correlates at roughly r = 0.68 across 27,118 examinees (PMC8368809) and carries the most weight in the model. Most candidates in the mid-240s and above show a very high Step 3 pass probability. Because everyone taking Step 3 has already taken Step 2 CK, it is the one input we always anchor on.

UWorld and UWSA — with a calibration correction

UWorld cumulative percentage and the UWSAs refine the estimate. The key adjustment: UWSA tends to under-predict real Step 3 by about 10 points, so the model recalibrates it rather than taking the raw number at face value. UWSA 2 (late-prep) is weighted more heavily than the noisier, often-early UWSA 1.

NBME Forms 6/7 and Free 137

The official NBME CCSSA Forms 6 and 7 are read as a percent-correct signal anchored to the passing standard — roughly 55% correct ≈ 200 (passing), with each point above adding about 0.8. They are a secondary confirmation input alongside the official Free 137.

CCS: the wildcard no MCQ score captures

Computer-based Case Simulations are a substantial, separately-scored part of Step 3 and are graded separately from every multiple-choice form — so the model applies them as an adjustment (detailed below) rather than a core input.

How CCS Cases Affect Your Predicted Step 3 Score

Computer-based Case Simulations (CCS) are what make Step 3 different from every other USMLE exam — interactive patient-management cases on Day 2, scored separately from the multiple-choice sections. No NBME form or UWSA grades them, which is why a multiple-choice-only prediction is incomplete.

How our model handles CCS: you enter either your average percent-correct on CCS practice cases (UWorld CCS or ccscases.com) or a simple self-rating. The model then applies a calibrated adjustment on top of your multiple-choice estimate: strong CCS performance adds roughly +6 points, while self-rated “struggled” performance subtracts about −12 points — enough to pull an otherwise-passing MCQ score below the 200 passing line.

Why CCS is an adjustment, not a matching input: in holdout testing, self-reported CCS percentages were too noisy to improve neighbor matching, so the model deliberately applies CCS as a post-hoc modifier instead of letting it distort the core prediction. That is also our honest advice for test day: aim for ~70%+ on high-yield CCS practice cases so CCS builds your buffer rather than eroding it.

What Is a Good Step 3 Score?

Passing standard and national mean

The Step 3 passing standard is 200 (raised from 198 on January 1, 2024), and the national mean is about 227 with a standard deviation near 15. For most residents a comfortable pass is the goal — the majority of programs never see or ask about the exact number.

Percentiles and score bands

Roughly, ~220 is a solid pass, ~230 sits near the mean, and 240+ places you toward the top of the distribution; below 200 is a fail. Because the exam spans two days and includes CCS, treat any single estimate as a range rather than a precise number.

Does the Step 3 score matter for fellowship?

For most applicants, no — a pass is what matters. A minority of competitive fellowships (some Cardiology or GI programs, for example) may glance at a strong Step 3 as a supporting data point, but it rarely outweighs your Step 2 CK and residency performance.

When to Take Step 3 and How to Prepare

PGY-1 vs PGY-2 timing

Most US graduates sit Step 3 during PGY-1 or early PGY-2, frequently to clear visa or credentialing requirements. Taking it while Step 2 CK knowledge is still fresh generally makes preparation lighter.

A workable dedicated plan

The exam runs across two days — Foundations of Independent Practice (FIP), which leans on biostatistics and drug mechanisms, and Advanced Clinical Medicine (ACM) plus CCS. A practical sequence: take UWSA 1 about three weeks out to surface Day-1 gaps, sit NBME Form 7 roughly one week out to fix your margin, and spend the final 48 hours on high-yield CCS cases to protect your buffer.

The single biggest mistake

Ignoring CCS. Residents comfortable with multiple-choice questions routinely under-prepare for the case simulations and lose the margin their MCQ score earned them. Aim for ~70%+ on CCS practice before test day.