Marking scheme

USMLE Foundation & Exam Strategy

Topic: Marking Scheme

Audience: Medical students preparing for the USMLE, exam‑strategy coaches, and educators who want a rigorous, MIT‑level understanding of how the USMLE scoring system works.
Goal: Provide a deep, mathematically grounded overview of the USMLE marking scheme, including its statistical underpinnings, edge‑case handling, and a challenging problem that tests mastery of the concepts.

1. Introduction

The United States Medical Licensing Examination (USMLE) is a three‑step computer‑based test that evaluates a candidate’s readiness to practice medicine in the U.S.
While the content of the exam is the most obvious focus, understanding the marking scheme is indispensable for:

• Strategic time‑management (knowing when to guess, when to skip).
• Predicting outcomes (estimating scaled scores from practice results).
• Interpreting the score report (e.g., why a raw score of 240 may translate to a scaled score of 230).

2. Overview of USMLE Step Exams

Key point: All steps use a multiple‑choice format; no partial credit is awarded for wrong answers.

3. Scoring System

3.1 Raw Score

• Definition: Number of correctly answered items.
• Range:
  • Step 1, Step 2 CK: 0–360.
  • Step 3: 0–240.

3.2 Standardized (Scaled) Score

• Purpose: Correct for differences in difficulty across exam “versions” (different item pools).
• Process:
  1. Item difficulty is estimated for each item using Item Response Theory (IRT).
  2. A raw score is converted to a theta (θ) value (candidate’s estimated ability).
  3. The theta is mapped to a scaled score using a lookup table derived from the entire candidate population.
• Scale Ranges:
  • Step 1, Step 2 CK: 1–300 (step size ≈ 1).
  • Step 3: 1–300 (step size ≈ 1).

3.3 Passing Thresholds (2024–2025)

Note: Thresholds are updated at the start of each year; they are determined using standard‑setting methods (see §5).

4. Scoring Methodology