Exam score reliability

[medschoolplease123]

---

Members don't see this ad.

Hi guys i was wondering if you would mind sharing how accurate your actual mcat score from your practice exams. Specifically kaplan, aamc, and the princeton review exam. Thanks!

 

[clathrin413]

TPR will undershoot your score, Kaplan will severely undershoot your score.

It's hard to say if AAMC is accurate since nobody has taken an MCAT since the Scored Test was released. We'll know soon! It seems to be a better predictor than the other exams though, at least based on converting scores from the Sample Test using an excel doc, which was done on here last summer.

 

[medschoolplease123]

What do you mean by the sample excel doc on here last summer?

 

[clathrin413]

What do you mean by the sample excel doc on here last summer?

Over the summer AAMC had only released the Sample Test, not the Scored Test. The Sample Test only gave percentages, not scores.

See: http://forums.studentdoctor.net/threads/tentative-aamc-sample-fl-score-conversion-sheet.1149516/

 

[theonlytycrane]

also, any comparison with EK tests?

 

[medschoolplease123]

EK?

 

[medschoolplease123]

exam krackers I assume. my bad.

 

[darknecrosforte]

EK tests, when taken exactly as suggested therein, were very very accurate. However, EK biopsychosocial content review was severely lacking in key terms. Next Step Social Content Review is 100% on definitions and random theories you need to know

 

[efle]

DO NOT put any stock in this thread http://forums.studentdoctor.net/threads/tentative-aamc-sample-fl-score-conversion-sheet.1149516/

It's based on the flawed assumption that similar % correct will correspond to similar percentile between old and new MCAT. We have no idea if this is true. It may very well be that you need less raw number correct for a given percentile on old vs new.

This is what you're looking for.

 

[GrapesofRath]

k

 

[GrapesofRath]

DO NOT put any stock in this thread http://forums.studentdoctor.net/threads/tentative-aamc-sample-fl-score-conversion-sheet.1149516/

It's based on the flawed assumption that similar % correct will correspond to similar percentile between old and new MCAT. We have no idea if this is true. It may very well be that you need less raw number correct for a given percentile on old vs new.

This is what you're looking for.

By and large this does look like it has held true. The scored AAMC FL released last month has a very similar scoring scale/conversion as the old AAMC FL's in terms of % correct give you X score which is equivalent to X percentile. Only real noteworthy thing is the psych/soc curve on the Scored AAMC FL seems to be a bit harsher than other sections.

Ill also say for that chart, a) its self selecting to some extent, people who dont do well just arent likely to report and/or just dont exist that much on SDN b) we have no idea when these people are taking the AAMC exam. Obviously it matters if you take it 6 weeks before your test day vs 3 days before.

 

[efle]

By and large this does look like it has held true. The scored AAMC FL released last month has a very similar scoring scale/conversion as the old AAMC FL's in terms of % correct give you X score which is equivalent to X percentile. Only real noteworthy thing is the psych/soc curve on the Scored AAMC FL seems to be a bit harsher than other sections.

Ill also say for that chart, a) its self selecting to some extent, people who dont do well just arent likely to report and/or just dont exist that much on SDN b) we have no idea when these people are taking the AAMC exam. Obviously it matters if you take it 6 weeks before your test day vs 3 days before.

Is the score scale public/can someone post a screenshot? Would like to compare to the old MCAT score conversion curves

 

[GrapesofRath]

Is the score scale public/can someone post a screenshot? Would like to compare to the old MCAT score conversion curves

Unfortunately its not. Hell from my understanding they dont even release it to you if you buy the test; they just give you the score without showing you the scoring scale. But having seen enough data points from people self reporting their scores on here and reddit, it's pretty obvious to me the scales are very similar to the old AAMC material. The psych/soc is the one where the curve is a little harsher which makes sense honestly; it's high school level material that everybody is going to find easier.

 

[efle]

Unfortunately its not. Hell from my understanding they dont even release it to you if you buy the test; they just give you the score without showing you the scoring scale. But having seen enough data points from people self reporting their scores on here and reddit, it's pretty obvious to me the scales are very similar to the old AAMC material. The psych/soc is the one where the curve is a little harsher which makes sense honestly; it's high school level material that everybody is going to find easier.

Do they not even tell you how many you got wrong, just spit out your score and percentile?

 

[GrapesofRath]

Do they not even tell you how many you got wrong, just spit out your score and percentile?

No they tell you how many you got wrong and the score; that's how I'm able to compare the old AAMC scale to the new one. But there were old scales floating around for each AAMC test that would tell you say 46-48/52 is a 13 etc. There is nothing like that for the scored new AAMC FL. The only way you can get a feel for the scale is to see enough people self report their scores compared to the percent they got correct.

 

[efle]

I'm surprised they kept the scaling similar, on the old test the upper range was condensed to the point where 13/14/15 were sometimes differentiated by only one or two more correct. Would have thought they'd put in more challenging questions so that 130/131/132 covered larger bins.

Damn you AAMC making things difficult. Perhaps they don't wish for people to make direct comparisons like what I'd try to make.

 

[GrapesofRath]

The thing is there are many old tests where it wasnt even possible to get a 15 on a section. That's kind of why its so hard to differentiate between a 13, 14 and 15; it's just so variable by test.

The new AAMC FL's honestly arent any more dificult than AAMC 9-11. The differences in the new MCAT vs the old MCAT from my experience of taking the MCAT last year and comparing it to old AAMC FL's were also largely overstated. There was however a section bank the AAMC released in Nov that I took a look at which is pretty damn hard; much harder than their FLs. Who knows, maybe that's what the AAMC is trying to trend towards and making the MCAT more like in coming years. But itll be a gradual process, you cant jack up the difficulty of the MCAT considerably over night.