How are the MCAT practice exams accurate if they are based on a curve?

[studentdoctor08]

---

Members do not see this ad.

Different people take the MCAT each time, so I'd assume that the curve would be different each time. Given this, how can these MCAT practice exams be that accurate if you're not going to be taking the test with the same people who took those exams? What if when you take the MCAT, your competition is extra smart or extra dumb?

 

[chiddler]

they're not. they're just rough indicators of how well you are doing. they are not based on scores of students that have taken them recently. they are based on, presumably, the scores of the first group of premeds that took the exam.

 

[studentdoctor08]

Because I know someone who consistently scored around a 37 on the MCAT practice tests, but when he took the real thing he got a 26...

 

[kasho11]

He must have choked under pressure hard, there is no way he was scoring 37 consistently and strictly due to the difficulty of the test hit 26.

 

[SoulinNeed]

I believe the curve actually has nothing to do with the people you take the MCAT with. The curve is predetermined based on how difficult the test makers feel the exam is.

 

[SH3]

*facepalm*

 

[xroc]

I believe the curve actually has nothing to do with the people you take the MCAT with. The curve is predetermined based on how difficult the test makers feel the exam is.

this is true. each question has a predetermined difficulty (i.e. they know what % of students should get this right/wrong). this explains why on the same date, no two mcats are exactly the same (the types of passages vary, and even for the same passage, the types of questions vary).

 

[UrshumMurshum]

There was a poll on here somewhere that showed a bell curve. Roughly 60% of people scored within +/- 2 points of their AAMC average. Then like 80% were within +/- 3 points.

 

[Midifelder10]

These tests are neither standardized or accurate. All you can state with certainty is that if you are scoring high/low you will be within +/-4 variance. for a real test that is bad.