CLEP scoring

[Publius]

My test officer gave me a copy of the Prin of Management Clep guide which clearly spells out score equivalents and statistical methodology. (i.e. to get a 50 you need 52-54 right, to get a "B" which is a scaled score of 63, you need 67-69, to get an 80 you need 90 right).

That's likely still speculation. The REA has their own 'equivalence chart' but no one actually has the data. Since it's still scored with a norm referenced approach, every test is going to be different. Your score, depends not only how well you did, but how well past test takers have done. The last part is what makes it tricky. Nevertheless those are still fun to look at and probably are in the ballpark of somewhat close.

[clep3705]

It is not a fact that about half of the questions on an exam must be answered correctly to pass CLEP and DSST exams. From time to time I help develop standardized tests working as a part of a development team. When I did a project for the Educational Testing Service, I was very impressed with the professionalism of the ETS employees. Once, long before I ever did exam development, I helped with CLEP score setting. A CLEP exam was given at my school instead of the school's final exam. Our performance on the CLEP test was used in to set the passing score on the test given to the public.

Passing marks aren't really based on percentages per se. Questions can have difficulty ratings depending on the exam vendor's approach to testing. Some are worth more than others. Additionally, some exams are simply more difficult than others. It isn't possible to get the questions at a consistent difficulty level across all subjects so that passing percentages are about the same on all exams. I've seen percentage passing marks range widely. They definitely did not cluster around a certain percentage (this data is often considered a trade secret, which is why I'm being vague). Exams with a more difficult test bank end up with a lower passing percentage. I worked on a proposed exam where all answers were correct, but each answer had degrees of correctness. All would fulfill the requirements, but each had a different cost and complexity. The better (lower cost, better efficiency) the answer chosen, the more it was worth. I haven't worked on CLEP or DSST exams, but psychometric theory is generic. If questions are equally weighted, a test bank with a lot of harder questions will have a lower percentage correct passing score. As questions are rotated into and out of the test bank, percentage passing scores will change.

You can take your vast exam experience and your subjective impressions and estimate what you think the passing percentage for CLEP and DSST is, but that conflicts with psychometric theory.

From a psychometric theory perspective, the NCLEX that nurses take is particularly interesting. It is an adaptive test that adjusts itself so that it gets to the point where you are missing half of the questions. For that test, I would agree that you need to get about half correct. The NCLEX is an exception to the rule being discussed in these forums.

Here are the official DSST exam facts: http://getcollegecredit.com/assets/pdf/D...coring.pdf

Notice that all DSST questions are weighted equally. Also notice that Fundamentals of College Algenra has a lower percentage required for passing than Business Mathematics. Over all exams, passing percentages range from 36% to 72%.

[Prloko]

That's likely still speculation. The REA has their own 'equivalence chart' but no one actually has the data. Since it's still scored with a norm referenced approach, every test is going to be different. Your score, depends not only how well you did, but how well past test takers have done. The last part is what makes it tricky. Nevertheless those are still fun to look at and probably are in the ballpark of somewhat close.

The score methodology my proctor gave me came straight from the college board, not REA.

[Publius]

The score methodology my proctor gave me came straight from the college board, not REA.

I wasn't saying it was from the REA, just saying they have something similar.

That's interesting that it came from the college board. Would you be allowed to share it on here (granted you have it?)?

[Prloko]

I wasn't saying it was from the REA, just saying they have something similar.

That's interesting that it came from the college board. Would you be allowed to share it on here (granted you have it?)?

I tried to, but the pdf is too large to post. PM me with your email and I'll shoot it over.

[Publius]

I tried to, but the pdf is too large to post. PM me with your email and I'll shoot it over.

Well I found a copy of it on the web. Here: http://cths.umhb.edu/system/files/Princi...nt.pdf.pdf

I found a bunch more in the process and will probably make a new thread with all of what I find.

It appears to be published by the CollegeBoard, and the numbers all vary as to what's needed to pass. Pretty interesting to say the least.