Tuesday, November 10, 2009

PMI Examination Methodology.

I'm posting excerpts from an interesting thread in PMZilla (for the complete version visit PMZilla). The thread was originally posted by dbissonn and his question was: 
"How are PMP exam questions weighted?

After passing numerous prep exams, including Rita's, I failed my first real PMP exam.  I called PMI, and while they would not use the term, apparently some questions wieighted, are given more credit than others.  It used to be a straight 61%, but no more.  Here is a quote from an email they sent me:

"Scaled Scoring is not a percentage-based scoring system, that is, the assessment passing point is assigned to a point of difficulty instead of a number or percent of correct responses."

The problem is no one will say anything about what is more or less weighted.  I suspect that, for example, not being able to do EV, or Crit Path, or float, will cost you more than not knowing what a fixed fee contract implies.

 Does anyone know how PMI actually determines how one has passed or failed?
"


He continues...
"The PMP exam report does not give % at all.  There are three ratings for each of 5 process groups plus professional and social resp.  Results are not broken down for KNOWLEDGE AREA).  Proficient is highest, Moderately Proficient, and Below Perficient.  I got the lowest in three. I have no way of knowing what I missed or their weighted value, in fact I am sure PMI does not want that made public.  It is possible, and this is a rough guess, to score 74% right for the whole test, but miss several high weighted questions, or score below 61% in one area, and still fail.

 On the last 10 or so practice exams I was never below 70%, and consistently above 75.  I think the real issue here is not how PMI does things, they are trying to judge if someone is truly knowledgeable, and they are using sophisticated testing methods to do so.  I have no problem with that.  But I think we need realize that preparation exams like Rita Mulcahy's FASTrack (US$299!!!) use a much less sophisticated method...they only consider a straight percentage, which is NOT the way the real exam works.



 I suggest no one call himself prepared with less than 80% right on practice exams.
"



Here's a consolidation of my responses:
"We all would like to hear how the exams are evaluated, wouldn't we? But I guess from the way the exam and the grading is setup, PMI is trying to ensure that there is no way to dis-assemble or reverse-engineer the exam process. Why?

First a look at what we know: 

The exam would result in a "pass" or "fail" and does not annouce the percentage or grade scored. We know that of the 200 questions 175 are actually used for evaluating the candidate and one must have answered atleast 106 out of 175 correctly in order to pass. (A combination of "Modified Angoff Technique" and psychometric method adopted by PMI since Sept 2005.) That's 61% of the questions.

Now the exam transcript, after you finish the exam, actually breaks the exam into Process Groups and your profeciency in it. And according to PMI the spread of questions is thus:

Initiating - 23 Questions

Planning - 46

Executing - 53

M&C - 42

Closing - 18

Professional & Social Responsibility - 18

The transcript does not merely say, "Pass/Fail" for each of the knowledge areas. It rates it as "Proficient", "Moderately proficient" and "Below Proficient". And you need to have Moderately Proficient and above in 4 of the 6 knowledge areas to pass. (correct me here, if i'm wrong) (moderately proficient in 4 of 6 KAs will result in a MAX score of 100, which is 6 less than the requisite 106. Which means the scores have to be a either Moderately Profient in all the areas or a combination of Moderately proficient and Proficient in 4 of 6 Knowledge Areas.) 

One could try to reverse engineer this and say that a 61% pass on each of the knowledge area (14 in Initiating, 28 in Planning, 32 in Executing, 26 in M&C, 11 in Closing, 11 in Professional Conduct. Moderately Proficient on all, sums up to 122) could fetch us passing results, but it needs to be confirmed if it is just that or more. In reverse-engineering, we need to ensure that the number of correct answers sums up to 106 or more and then we could look at the areas that the correct answers come from and rate it on proficiency.

 Add to this the fact that the questions are rated using the Modified Angoff method, where volunteers for PMI adjudge the difficulty level of the questions. The tougher a question, lower are the points gained from getting the answer right, and easier the question, higher the points.One would therefore notice that the way the exam goes, we'll see a whole bunch of easy questions appearing together followed by a series of really tough ones.

There is no sure fire way of determining how the results would appear. But to be on the safer side, ensure that you are proficient on all the individual Knowledge Areas while taking mock exams. And by Proficient I'm presuming 80% and above. Then go for the complete exams.

There is another point to the way the exam is structured and evaluated. Well, like all the big kahunas of PMP training and preperation have been saying, no one person can get all of the answers right in the exam. That's because of the nature of the exam and each person's take on the subject. But end of the day, the reason some of us are PMPs is because of a common understanding of the practices of Project Management.

On a side note, I really don't understand how RMC's products got so much of a hype. They are good, but really not that great! Especially when you consider what they charge for it. There are free products out there that should be making a killing for the kind of quality they deliver (case in point Oliver Lehmann's 75 and 125 questions).
"



So... Your thoughts or comments?...
Your Ad Here