Need input from the statistically inclined regarding Kaplan QBank

[RichieVallens]

---

Members don't see this ad.

Aside from consideration of a Gaussian Distribution, ch-square, ANOVA, and whatever else...Is it a fairly correct assumption that the average cumulative score of all students taking the Kaplan QBank should be close to the performance of the average score on the USMLE step 1? For example, if the average Kaplan QBank score is 60%, then scoring a 60% would be a decent predictor of scoring a 220?

 

[withrye]

Not necessarily. Even if the populations being sampled are the same (which isn't really true, since not all people are inclined to use Kaplan and there might be a bias with confounding effects present), the people taking Kaplan might be taking it at different points in time relative to their exam. So someone scoring 60% really early may go on to get a 270 on the exam.

 

[RichieVallens]

well yeah....but you know what I mean...... Thanks though 🙂

 

[medpack87]

I dont think there is much correlation. Mainly because many people are using the qbanks as a learning tool, so they might be for example: using their books and notes, doing the questions grouped by type and subject, doing small groups of questions, resting in between, learning how to time, testing different test-taking strategies etc. So some will get very high % and some will get very low.