(USMLE score calculation) hard questions correct, easy question wrong

[tarsuc]

I've tried getting answers to this question before but nobody seems to be sure about it.
Apologies for bringing this up again as a separate post.

If i answer an easy question wrong, will my score drop by a bigger margin than(compared to) the score drop while getting a hard question wrong?


or in other words

Is getting an easy question wrong that everybody else gets correct more dangerous than, getting a hard question wrong, which not many can answer correctly anyway?


Or is it all the same in the end?

---

Members don't see this ad.

 

[MedicalBeast]

i'd also like to know

 

[Transposony]

I don't think that anyone knows.
According to USMLE they are all scored equally.

But in the end it doesn't really matter since you are not going to attempt a question based on how it will be scored.

 

[tarsuc]

But in the end it doesn't really matter since you are not going to attempt a question based on how it will be scored.

That is so true, but atleast it will reduce the anxiety/time of laboriously rechecking every 'seemingly' simple question, and making sure that the simplest sounding answer is correct anyway.

Hope that makes sense.🙂

 

[Leonidis I]

a hard question would probably have a higher chance of being experimental IMO because it would not have been covered in fa/uw. whereas an easy question would have a lower chance of being experimental for that reason.

so missing more hard questions might yield you a similar score of missing fewer easier questions simply because the hard ones are experimental and don't count towards your scaled score.

this might explain why so many people especially for the july 10 score realease felt like crap after taking the exam but still score near their NBME averages. because many of the the questions they found challenging ended up having no bearing on their score.

[this is completely based on conjecture and has no proof whatsoever]

 

[paul411]

a hard question would probably have a higher chance of being experimental IMO because it would not have been covered in fa/uw. whereas an easy question would have a lower chance of being experimental for that reason.

Actually almost every question, regardless of difficulty, would have an equal chance of being experimental (unscored). The reason is that the test-makers need at least n amount of data for every question before they can start including it in the scored items.

 

[paul411]

If i answer an easy question wrong, will my score drop by a bigger margin than(compared to) the score drop while getting a hard question wrong?


or in other words

Is getting an easy question wrong that everybody else gets correct more dangerous than, getting a hard question wrong, which not many can answer correctly anyway?


Or is it all the same in the end?

Every question is worth the same. In the end, the raw score (# correct) for a particular form is converted to a scaled score based on each individual item's characteristics (called an Item Characteristic Curve). To learn more about how this works, read up on 'Item Response Theory'.

Here's some good resources to get started: 1, Wiki

 

[tarsuc]

thanks. makes some sense. a little too complicated to fully understand though 🙂