20 May 2018

Did you or your students get a nonstandard form for an AP Physics exam?

The vast majority of students who take an AP Physics exam get the “operational” version of the exam.  This is the one whose multiple choice questions are never released, but whose free response is available publicly 2 days after the exam date.

A small fraction of students take a different form of the exam.  In particular, students outside North America usually get this different form - otherwise, it’s possible for a particularly dirty player to reconstruct the exam and text details to a student in a different time zone.  Within the continental US, that’s difficult - the west coast has entered the exam before the east coast is finished.  In the fall the College Board releases the majority of this “international” version of the exam to teachers who have passed an official course audit.

For statistical purposes, occasional randomly-selected students stateside take the exam form that’s otherwise earmarked for the international audience.  Invariably, when a few students find out that they took a different test than their classmates did, they claim that (a) they took the harder version, and (b) no fair.

Claim (a) has a 50-50 shot of being true.  Claim (b) is utter, ignorant baloney.

Is the international version harder?
Sometimes, sometimes not.  Is this year’s exam harder than last year’s?  It depends on your perspective, what you’re personally good at, and random chance.  The development committee attempts to construct exams of similar difficulty in each year, in each form. If you have access to the released international exams, take a look through.  Some questions are harder, some easier, most about the same.  Once many years ago I was in the super secret room where the international and alternate forms of the exam are graded - on pain of pain, I’m not allowed to tell you any details about the exams I saw.  That’s just as well, because I don’t remember details.  Yet, I can tell you that the questions read exactly like the operational exams - I perceived no difference in difficulty, no difference in the range of student responses.

Are the AP scores lower for the international version?
No.  The difficulty of exam questions does not matter when AP scores are compiled!

On each exam, a subset of multiple choice questions are repeated from previous years.  These serve as a control on exam difficulty.  Even when an exam consists of demonstrably more difficult questions, the meaning of a 5, 4, 3 is identical.  This subset of multiple choice ensures that desired outcome.

Let’s say that all students taking the international version got 40% of the available points, while all students taking the operational version got 50% of the available points.  There are two possible explanations for this discrepancy:

(1) The international exam consists of more difficult questions than the operational exam.
(2) The population of students taking the international exam is weaker.

Performance on the subset of repeated multiple choice questions can differentiate between the two hypotheses.  

If there was no statistical difference in performance on the repeated subset, then explanation (1) applies.  The same level of student getting fewer points means the exam is harder.  Thus, the cutoff percentages for 5, 4, 3 would be dropped.

However, if the students on the international exam scored similarly worse on the subset of repeated questions, then hypothesis (2) is confirmed.  Even on the exact same questions, these students performed to a lower standard.  The cutoff percentages would reflect that lower level of performance.


The overarching goal is that a 5, 4, or 3 means the same level of performance from year to year.  Yes, it is true that for AP Physics 1, about 70% generally works out to a 5, 55% to a 4; those numbers are variable to serve the overarching goal.  Similarly, it is true that about 5% of the student population gets a 5 on AP Physics 1.  But that's also variable, dependent on the performance of this year's students.  There's nothing preventing teachers and students getting better year to year, such that suddenly half the country earns a 5.  I mean, that's unlikely... but the College Board would be utterly thrilled if that happened.

Scores are not manipulated for political reasons.  In the very first year the AP Physics 1 and 2 exams were administered, a diverse committee was assembled and teachers/professors were surveyed to determine the standard of performance appropriate to each score.  While statistics were used copiously, this original score setting included some subjective as well as objective input.  

Once those original scores were set, though, the standard of performance for each score was also set in stone.  The scores this year will be what they will be... if more people get high scores (as has been happening very gradually across the last few years), that represents authentic improvement in the student population taking the exam.  No one is pulling strings behind a curtain trying to depress or increase scores.  Those who promulgate conspiracy theories do so out of malicious ignorance.  The numerous ETS statisticians, the chief readers, the College Board representatives, the development committees... all these people would have to be part of a such a grand conspiracy.  I know many of these folks - I know them to be people of considerable integrity.  They wouldn't stand for fudging scores.  It's not happening.

Similarly, there's no conspiracy to give half of your school a harder exam with worse scores.  If you happened to get a nonstandard form, well, everything will come out in the wash.  Your exam score will reflect your knowledge of physics.  As it's supposed to.  :-)



No comments:

Post a Comment