MCAT Scales?

[jpmiz]

Advertisement - Members don't see this ad

Hi, so I've been hearing after being on here that easier MCATs have a harsher scale and vice versa. How true is this really? I was going through the scales for all the AAMCs and most of them seem pretty similar in terms of how many questions you can miss to get a certain score. Is the real MCAT scaled differently than this, or is a more "brutal" scale really only a couple questions different than a more forgiving scale?


Thanks!

 

[texan2414]

Nobody knows for the most part.
Also, easy/difficult are subjective when it comes to MCAT prep. But when your taking practice exams, you will know where your scores lie.

The real MCAT might be scaled differently because it might have experimental questions that are discarded. But since you don't know which ones are experimental, you have to take each question at its face value.

 

[BerkReviewTeach]

They used to be a little more transparent in their grading mechanism and performance data before moving to the CBT format. The larger number of testing dates must have created a wider range of difficulties and smaller N values for each exam, so in order to normalize the scores they would have to have some curve adjustment. Exactly how much and which questions count is information they don't seem to share.

What I've found fascinating is that for a test with three scores in the 1-15 range that are supposed to fit a bell curve, the overall average is just over 25 rather than equal to 24. I'm wondering if for the new exam, being that it has more questions per section, whether there will be a truer bell curve. If it stays the same, then a 30 (3 x 10) on the current exam will equate to a 508 (4 x 127) with an average of just over 501 (maybe 501.3). If they recenter their bell curve, then a 507 will fall in the same spot as a 30 does now on the overall distribution.

The admissions folks are going to have an interesting time sorting it out this next application cycle. I guess they'll make 30 = either 507 or 508 and go from there.

 

[GreyF0X]

They used to be a little more transparent in their grading mechanism and performance data before moving to the CBT format. The larger number of testing dates must have created a wider range of difficulties and smaller N values for each exam, so in order to normalize the scores they would have to have some curve adjustment. Exactly how much and which questions count is information they don't seem to share.

What I've found fascinating is that for a test with three scores in the 1-15 range that are supposed to fit a bell curve, the overall average is just over 25 rather than equal to 24. I'm wondering if for the new exam, being that it has more questions per section, whether there will be a truer bell curve. If it stays the same, then a 30 (3 x 10) on the current exam will equate to a 508 (4 x 127) with an average of just over 501 (maybe 501.3). If they recenter their bell curve, then a 507 will fall in the same spot as a 30 does now on the overall distribution.

The admissions folks are going to have an interesting time sorting it out this next application cycle. I guess they'll make 30 = either 507 or 508 and go from there.

@Goro mentioned that they will be looking at percentile scores instead of the numerical score. By the way, when will berkeley review be releasing their prep material for the new MCAT?

 

[Gauss44]

I think I counted that on some AAMC practice tests, you can miss 10 questions and get a 10/15 on the science sections, and on other AAMC practice tests, you can miss 15 questions on the science sections and still get a 10/15. If I counted right, then missing 10 questions will get you a 10 or better, and missing 15 questions will get you a 10 or worse.

If someone somehow knows how many questions they missed on the real MCAT and what their score was, that could shed light on one of the current scales. The trouble will be figuring out if such a person is reliable and that discussing that could arguable be against AAMC's rules.

 

[stitchattack]

The way I see it is that the test's level of difficulty lies with the test taker..the test either plays to your weakness or strength. I don't think one test is "easier" than another. I think it relates more to the topics that are tested. Even if you give a supposedly easier test to someone with minimal prepping, they are probably not going to get 30+. Also, you need to take into account that SDN only represents a small population of actual pre-meds and many of the posters I have had the chance to associate with take prepping seriously and mostly try to devote prepping with minimal distractions which may or may not be the case for the rest of the pre-meds not on sdn. For scaling, we may never know how AAMC really operates. However, perhaps for "easier" tests maybe the scaling is more like AAMC 10 where you need to miss like 3 questions or less to get the same score like in AAMC 11. When I took those tests before my test, I only had 1 pt difference overall despite AAMC 10 having the infamous harsher scaling. I don't think the scoring can get any worse than AAMC 10 imo..hopefully not because even on sdn, people were talking about how harsh the scale was lol