When an NBME self-assessment and a UWorld self-assessment give you different answers about your readiness, both results are probably telling the truth about something. They are built by different organisations, from different question pools, with different conversions from raw performance to a predicted result, and they are taken at different points in your preparation. The useful question is not which one is right. It is which one to act on, and the answer is usually the NBME trend, read alongside the most recent results from both.
Different exams, so different answers are normal
NBME develops the USMLE, and its self-assessments are written by the same organisation in the same style. UWorld writes its own self-assessments to match its own question bank. Both are trying to estimate the same thing, but from different material, so the same student will rarely produce identical results on both.
That alone should take the alarm out of a gap. Two thermometers from different manufacturers rarely agree to the decimal. What matters is whether they agree about the direction and rough level.
The usual reasons UWorld reads higher
The most common pattern is a UWorld self-assessment that looks better than a nearby NBME form. Three explanations account for most of it.
Familiarity. If you have done most of the UWorld bank, the self-assessment’s style, phrasing and favourite topics will feel familiar. Familiarity raises performance without necessarily reflecting readiness for a different writer’s questions.
Conditions. Students often sit one assessment at home in one go and another in pieces, or one after a full study day and another fresh. Different conditions produce different results, and that difference is noise, not signal.
Form difficulty. Individual forms vary. A particularly easy or hard form on either side can open a gap that disappears on the next one.
The less common reason NBME reads higher
Sometimes the official form comes out higher. This is often because NBME items are shorter or more direct than some bank questions, or because the student’s strengths happen to line up with that form. It is reassuring, but it is still one result, and it deserves the same scrutiny as a disappointing one.
Check the comparison is fair before you interpret it
Before you treat a disagreement as information, check three things.
Were both sat under real conditions? Timed, in one sitting, without notes or pauses. If not, discount the one that was not.
Were they close together? Results a few weeks apart reflect the learning in between. Comparing an early NBME with a late UWorld measures your progress, not the forms.
Did you know any of the questions? Repeated or familiar items inflate results. Be honest about whether you recognised material.
If the comparison fails any of these checks, the gap is not evidence of anything.
Read the trend from each source, not one pair
A better method than comparing single results is to look at each source as its own series. Are your NBME results rising, flat or falling? Are your UWorld results doing the same? If both series point the same way, the disagreement in level matters much less than the agreement in direction.
Where you have several results from one source, that series is usually more informative than any cross-source comparison, because it controls for the writer, the conversion and the style.
Which one to act on
As a working rule:
For deciding whether you are ready, anchor on NBME. It is the closest available proxy for the exam itself, and the Step 1 self-assessment reports an estimated probability of passing, with NBME’s own caveat that it is not a guarantee.
For deciding what to study, use both. Each report highlights weak systems and disciplines. Where both point at the same weakness, that is your priority.
When you are close to the exam, be conservative. If the most recent results disagree and the lower one is below where you need to be, treat that as the safer basis for a decision. The cost of moving a date is usually smaller than the cost of sitting unprepared.
What a persistent gap does tell you
If NBME results stay consistently below UWorld results across several properly sat forms, that pattern is worth acting on. It often means you have learned the bank rather than the material, recognising familiar questions without being able to reason through unfamiliar ones. The fix is more varied practice, more attention to why each wrong answer is wrong, and fewer repeated questions.
The practical upshot
Sit both kinds of assessment under the same conditions, spaced deliberately, and log each result with the date and circumstances. Then judge readiness on the official trend and use everything else to decide where the remaining weeks go.
Frequently asked questions
Which is more accurate, NBME or UWorld self-assessments?
Neither is a perfect predictor, but NBME forms come from the organisation that develops the USMLE, so most students treat them as the anchor. UWorld self-assessments are a useful second source, particularly for their explanations. When they disagree, weight the NBME trend more.
Why is my UWorld self-assessment higher than my NBME?
Common reasons include having already seen similar UWorld questions in the bank, sitting the two under different conditions, and the forms differing in difficulty and style. Check conditions and timing before treating the gap as meaningful.
Should I postpone my exam if my practice scores disagree?
Not on a single disagreement. Look at the trend from each source and at the most recent results. If the lower and more recent results are below where you need to be, that is a reason to consider moving the date.