You have done the hard part. Three choices are gone and you are down to two. Then you sit there, eyes moving between them, and the clock keeps running. Multiply that moment across a few hundred questions and it becomes one of the biggest sources of lost points on the exam.

Here is the order I go through when it happens.

1. Lock your first instinct

Your first read is right more often than your second guess. So do not switch unless you can point at an actual line in the stem that you missed the first time. Not a feeling. Not the sense that the other one could work too. A fact you can put your finger on.

Going back to hunt for reasons to justify the other choice is a trap. You are not finding new information at that point, you are building a case. The moment you notice yourself doing it, stop.

2. Read the last line again, by itself

Most switching errors happen because you are answering a slightly different question than the one on the screen. Before you commit, read the final sentence alone and ask what it is actually requesting. Next best step? Most likely diagnosis? Underlying mechanism? The answer is usually hiding in how specific that sentence is.

3. One contradiction kills the choice

If a choice conflicts with a single thing you are sure about, even something small, drop it. NBME choices are not written to be half right. One wrong detail makes the whole option wrong.

4. Watch for the findings that are missing

If your answer only works when the patient also has symptoms that are not in the stem, that is a warning. These cases are written tightly. When something you would expect to see is not there, that is usually on purpose.

5. Do not be clever

One choice is direct. The other needs a chain of reasoning that makes you feel smart. Take the direct one. The exam rewards clean clinical thinking, not creativity. If getting to an answer requires you to be unusually brilliant, it is probably not the answer.

If you are still stuck after all five, go with the one you felt first. Answer it, flag it if you want, and move on. Getting a question wrong is fine. Losing three minutes on it is not.

Want this as a plan instead of advice? If you want a day-by-day schedule built around your own exam date and your own weak systems, that is what Dumirieh mentorship does. You can read my full story on the about page.