*Official June 2016 MCAT Thread*

This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.

aalamruad

Full Member
7+ Year Member
Joined
Oct 30, 2014
Messages
2,046
Reaction score
3,371
Hey there everyone! Figured I'd start this thread since I recently started preparing for the June 18th MCAT. For those of you who are taking the June 2nd MCAT, June 18th MCAT, or still considering one of those two dates, feel free to post here with any questions, comments, concerns, or support you have to offer!

Good luck to you all! Let's crush it.
 
Anyone know if experimental questions tend to be individual questions or all questions associated with an entire passage?


Sent from my iPhone using SDN mobile

That's a good question. If it were questions associated with an entire passage that would mean there is always a throwaway passage built in. I feel like that's less likely than individual experimental questions. Still possible, especially for the new MCAT, since they could have built the test around 1 less passage for each section, and 1 for experimenting for producing future tests.

But who knows, they are so super-secret-squirrel.


Sent from my iPhone using SDN mobile
 
I do understand how sample sizes work. I create scaled tests, but methods vary greatly. The sample size you are referring to as 100,000 would be the total sample size of that year. The sample size I am referring to is that of a single test, on a single administration date. That 100,000 could be comprised of hundreds of versions of the test, thus doesn't not explain the point I was trying to make. On a single test, that maybe 1000 people take, if it were curved, the gradient could be moved. So yes, you seem to misunderstand the point I was making. No one knows the amount of people that takes a particular version of the test in a given year. If one test has a small amount of test takers, curving on that test, from the specific administration date, would produce a different result than the scale that is produced from the history of both real and experimental test takers who took it.


Sent from my iPhone using SDN mobile
Aside from nonsensical hypothetical scenarios where super rich pre-meds pay thousands of people to bomb the test, none of this scaled vs. curved debate really matters. The bottom line is that particularly hard tests and particularly easy tests will be adjusted accordingly, such that your score is representative of how you stack up to other applicants. Again, all I'm saying is that there's no need to correct people who say it's curved. Can you think of a single realistic example of when it could matter that it's scaled rather than curved?
 
No worries I'll explain to the best of my knowledge. Scaling inherently means you have predictive value on the score associated with each question. So a question likely doesn't go into a test officially to be scored until after many, many, tests in which it was included as an experimental question. Statistical analysis will then give you the confidence interval for how the typical student will perform. This would then go into creating a test, in which this is done for each question. Because each test has a combination of different questions, their raw %correct is different but they are all balanced so that they produce a scaled score around the same center. Which is why 500 is deigned to be the the middle point, 125 for sub sections.

The post test modifications serve a multitude of reasons, most of which I'm sure are well beyond me. But they are not for curving. They probably serve to test-retest validity (aka accuracy), reliability (aka predictability), and fairness. They probably include a series of human raters to check interrater reliability and confirm the scale that was used for said test.

Just because you came up with a scale doesn't mean you can just implement it without checking and confirming you have predicted what you said you would.
Furthermore, a given test may include questions that just came out of the experimental phase and going into their first "real test". All these things require post-test analysis. And it's a good thing, because it means they are not just pumping out algorithms and not check for errors they made, etc.


Sent from my iPhone using SDN mobile
Also, just to clarify, what exactly do you mean by the bolded? It seems to me like you're saying that since scales aren't perfectly predictive, post-test analysis, potentially including minor post-test modifications to scoring, could be performed if the raters do in fact find that there are some slight errors in the scales. I might just be misunderstanding, but isn't that just curving?
 
Aside from nonsensical hypothetical scenarios where super rich pre-meds pay thousands of people to bomb the test, none of this scaled vs. curved debate really matters. The bottom line is that particularly hard tests and particularly easy tests will be adjusted accordingly, such that your score is representative of how you stack up to other applicants. Again, all I'm saying is that there's no need to correct people who say it's curved. Can you think of a single realistic example of when it could matter that it's scaled rather than curved?

Agreed. I think the reason the aamc stays away from the word "curve" is because it seems to connote that scores are being adjusted based on other test-takers on your particular test day, which is apparently not the case. Who knows though. There's so much mystery surrounding the scoring system, and I'm sure they have their reasons for that.


Sent from my iPhone using SDN mobile
 
Agreed. I think the reason the aamc stays away from the word "curve" is because it seems to connote that scores are being adjusted based on other test-takers on your particular test day, which is apparently not the case. Who knows though. There's so much mystery surrounding the scoring system, and I'm sure they have their reasons for that.


Sent from my iPhone using SDN mobile
Yeah, could just be AAMC not wanting us to be competitive when studying, for all we know. It would be nice if they were a bit more open about the whole process haha
 
Also, just to clarify, what exactly do you mean by the bolded? It seems to me like you're saying that since scales aren't perfectly predictive, post-test analysis, potentially including minor post-test modifications to scoring, could be performed if the raters do in fact find that there are some slight errors in the scales. I might just be misunderstanding, but isn't that just curving?

My point is that even when you create a scaled test. You need to collect the data on the test to check your tests validity and reliability. This is done in the experimental phase to make it so you don't HAVE to do this. But at least in psychology, I have never seen anyone not use interrater post test anyway, redundant or not, to make sure it predicts what it said it would. The potential change to said test as a result of fluctuations could happen after thousands of tests or hundreds of tests, depending on how dramatic the change is. So it wouldn't be called curved because if it took 5 years for the change to be significant, the person from 5 years ago who took one administration isn't getting their score adjusted because the scale didn't hold up as it was predicted over time. The scale would just be getting adjusted over time, which would impact future takers.




Sent from my iPhone using SDN mobile
 
Aside from nonsensical hypothetical scenarios where super rich pre-meds pay thousands of people to bomb the test, none of this scaled vs. curved debate really matters. The bottom line is that particularly hard tests and particularly easy tests will be adjusted accordingly, such that your score is representative of how you stack up to other applicants. Again, all I'm saying is that there's no need to correct people who say it's curved. Can you think of a single realistic example of when it could matter that it's scaled rather than curved?

Well even without the the ridiculous scenario, which I stated was ridiculous to begin with. Like someone just mentioned above, believing you are stacking up in one administration, as if it were one test, at that given time, would be a good assumption if the test is curved. It is a misleading one when it is actually scaled.




Sent from my iPhone using SDN mobile
 
I was in my school library today to chill and read some Lehninger's, and I saw some girl doing the AAMC section bank. It all hit me: I'm done with my MCAT, I finally finished studying and took the freaking test. It's over. Good luck random woman, my MCAT days are over and I wish you luck. I actually got a little giddy inside and wanted to just sit and watch her like a creep and try to coach her about it like an old man who likes to tell random stories, but I just did my own thing.

I REALLY hope this is the last time I take this exam. 7 days until I find out.
 
Guys, let's talk about what REALLY matters... what's your trainer level in Pokemon Go

All of my interns were talking about this today. I have not a clue what this is. iPhone app?

I've never felt so old not being up to date and they are 2 years younger than me lol


Sent from my iPhone using SDN mobile
 
Long time lurker but I thought I'd post to ease some exam anxiety for those who have practice exams scoring lower than their goal. Trust your knowledge and your abilities!

June 2nd - first MCAT
523
CP 131/ CARS 129/ BB 132/ PS 131

Practice exams:
Kaplan diagnostic: 496
Kaplan 1: 500
Kaplan 2: 503
Kaplan 3: 506
Kaplan 4: 506
Kaplan 5: 507
Kaplan 6: 508
Kaplan 7: 507
AAMC unscored: 80%/92%/81%/85%
AAMC Scored: 514 (one week prior) 129/128/128/129
 
Me: after first week: 510+,
2nd week: 508
3rd week: 503
....
and keep going down till this time next week.
I plan to submit tonight but still really scared abt MCAT
510+ the first week? Haha lucky. I felt like a 505 walking out of the exam. I don't even remember CARS. I swear I was on autopilot. Either I completely bombed that section (121) because I wasn't comprehending, or I killed it (129-130) because I was do locked in. I marked ~20 in the P/S section and the B/B section was hella easy. I'm predicting 42/59 will be a 125. FML.


06/18 taker.

Sent from my SAMSUNG-SM-N910A using Tapatalk
 
510+ the first week? Haha lucky. I felt like a 505 walking out of the exam. I don't even remember CARS. I swear I was on autopilot. Either I completely bombed that section (121) because I wasn't comprehending, or I killed it (129-130) because I was do locked in. I marked ~20 in the P/S section and the B/B section was hella easy. I'm predicting 42/59 will be a 125. FML.


06/18 taker.

Sent from my SAMSUNG-SM-N910A using Tapatalk

I marked 10 in the B/B section, but I think i'm the only one that just thought it was okay, but not easy. Looks like I did bad in that section.
 
I marked 10 in the B/B section, but I think i'm the only one that just thought it was okay, but not easy. Looks like I did bad in that section.
Just because you marked 10 doesn't mean you didn't do well. You didn't miss every question you marked. I just felt it was easy compared to the section bank and Kaplan FL's. You'll score around where you scored in the sample AAMC FL scored. I got a 127 and I'll be lucky to get that. Feeling like a 125.

How did you do on the scored ?

Sent from my SAMSUNG-SM-N910A using Tapatalk
 
Just because you marked 10 doesn't mean you didn't do well. You didn't miss every question you marked. I just felt it was easy compared to the section bank and Kaplan FL's. You'll score around where you scored in the sample AAMC FL scored. I got a 127 and I'll be lucky to get that. Feeling like a 125.

How did you do on the scored ?

Sent from my SAMSUNG-SM-N910A using Tapatalk

509, but at this point I think I made maybe a 500 lol
 
I was planning to take the MCAT in January, and I was thinking I would go through the EK books (hopefully before November), then do many practice tests in the months after, and look at kaplan for reference to questions I missed if ek didnt have the content. (I have both ek and kaplan). Is this a good idea or would another strategy be better?
 
I submitted to 13 schools. If my MCAT comes back good I have another 8 I'll submit to.

I've submitted secondaries to 3 schools.


Sent from my iPhone using SDN mobile app
How can you get the secondaries without the MCAT? I thought they would look at the GPA and MCAT and your overall app for the secondary screening...
 
Some schools automatically send secondaries to everyone who applies
1. So is there a deadline for the secondaries?
2. And if I submit the secondaries early, and when the MCAT comes out is low, then I would retake it afterwards, will the schools look at that MCAT for the interview selection or will they wait for the new MCAT to come out?
3. How to know which schools do not screen for the secondaries?
Thanks,
 
1. So is there a deadline for the secondaries?
2. And if I submit the secondaries early, and when the MCAT comes out is low, then I would retake it afterwards, will the schools look at that MCAT for the interview selection or will they wait for the new MCAT to come out?
3. How to know which schools do not screen for the secondaries?
Thanks,

1.) Yes. Depending on the school.

2.) Not sure.

3.) AZCOM, MSU, CCOM, A.T. Still Arizona & Nova let me submit my secondaries. So more than 3. Might be 1 or 2 more. Can't quite remember at the moment.. There are schools I applied to that haven't yet released secondaries to anyone. RVUCOM and LECOM won't let you submit secondaries without MCAT


Sent from my iPhone using SDN mobile app
 
510+ the first week? Haha lucky. I felt like a 505 walking out of the exam. I don't even remember CARS. I swear I was on autopilot. Either I completely bombed that section (121) because I wasn't comprehending, or I killed it (129-130) because I was do locked in. I marked ~20 in the P/S section and the B/B section was hella easy. I'm predicting 42/59 will be a 125. FML.


06/18 taker.

Sent from my SAMSUNG-SM-N910A using Tapatalk

I'm in the same boat. I marked a lot but that's only because during my practice tests I made a habit to mark every single question that I even slightly struggled with, so that when I finish I can double check them.

For cars- like you said, I was either so dialed in that I did better than I thought, or I totally bombed it.

Bio was easy as hell for me but that's only because I practiced a lot so I got really good at scientific reasoning. I was aware that bio was really difficult, but I felt totally in control and held up well.

Psych/soc is one of those things where you're down to two, and you either picked the right one and did good, or you picked the wrong one all the time. I noticed that for psych and soc you really have to follow your gut, because the concepts are very similar and it's easy to reason your way into choosing the wrong answer.


Sent from my iPhone using Tapatalk
 
I dont know at this point whether to wait until Tuesday to submit or not... I mean, Im pretty much done but just in case I get a 490 or something...
 
Ok I admit it. I'm stressing out a lot in this limbo thing. Want to just hibernate for the next week


Sent from my iPhone using Tapatalk
 
There's no way it's that many, who told you that?
I agree. With 10 passages in a given section, that is way too little, in my opinion, of tested questions.


Sent from my iPhone using SDN mobile
Actually it does make sense since in each test they have to accumulate statistical data and validate the same number of passages (10) for the next year tests. Even if they make 3-4 version of each test, there must be at least 3 out of 10 passages that are experimental. And don't forget, some of these passages may not survive the validation process and will be wasted.
 
Actually it does make sense since in each test they have to accumulate statistical data and validate the same number of passages (10) for the next year tests. Even if they make 3-4 version of each test, there must be at least 3 out of 10 passages that are experimental. And don't forget, some of these passages may not survive the validation process and will be wasted.

We understand all about the experimental and validation process. My point is that there are definitely more than 4 versions of the test. I think the number is far greater. The point was having only 2/3 of the exam actually "count" is not as robust a measurement. I think the test itself would be too short if 1/3 was experimental. Seems unlikely. I think it's more likely there are a greater number of tests, with a smaller number of experimental questions in each.


Sent from my iPhone using SDN mobile
 
Last edited:
We understand all about the experimental and validation process. My point is that there are definitely more than 4 versions of the test. There I think the number is far greater. The point was having only 2/3 of the exam actually "count" is not as robust a measurement. I think the test itself would be too short if 1/3 was experimental. Seems unlikely. I think it's more likely there are a greater number of tests, with a smaller number of experimental questions in each.


Sent from my iPhone using SDN mobile
You could be right! But considering that for each test day there are approximately 4000 (100,000/26 test days) test takers, then if number of versions far exceed 4 versions then the number of test takers (N) for each version would be too low to make it statistically significant.
 
People that came out of the test thinking they got a 510+ are probably going to do fantastic. I came out thinking I did 505+, and the thought of seeing a number between 500-507 scares me so bad. I had a nightmare about it.
 
You could be right! But considering that for each test day there are approximately 4000 (100,000/26 test day) test takers, then if number of versions far exceed 4 versions then the number of test takers (N) for each version would be too low to make it statistically significant.

But the score isn't based on the number of people who took it that day, nor are the percentile ranks, which is why they don't really change. They are based on the historical performance, which you get from years of experimental questions, as well as whatever other experimental measures they perform for rating questions. So the statistical significance is based off of (big) N, not (little) n.

It's likely pretty balanced in the amount each test is administered but if one particular version was circulated less it wouldn't matter. They probably use a randomizer, which is what my lab does. And some weeks we have a fraction of one test actually administered compared to another. Doesn't change anything.


Sent from my iPhone using SDN mobile
 
Top