Standardizing a Test

[RobbingReality]

How does the USMLE or COMLEX standardize a test that is given multiple times a year? Do they only develop a bell shape curve for that day/week of test takers or do they have a plethora of questions that have been used in the pass with good score values to base a grading scale from? Hypothetically, say an average med. student took the test on the same day as a bunch of freakishly smart med students, his score would be lower than normal if they based it on just one day of testing. However if they had a Q bank of 5,000 qestions which had been answered on a regular bases they would be able to judge if somone correctly identified 80% of the questions, and the average student would not have to worry about being compared to the freaks who identified 100% of the questions (they would be outliers) on that one day. Would someone with some insight set me straight.

---

Members don't see this ad.

 

[osli]

(All guessing, but very logical...) Statistical analysis based on the average score on each question, thus generating a difficulty rating for each question. Your score would be based on which particular questions you got correct and which you missed... someone gave a pretty good explanation here as well if you search for it.