When ever in any walk of life I see exams with multiple choice, I get amused, as you have 25% or 33% chance if 4 or 3 questions to get it right, without having a clue about the subject, how can that reflect the true knowledge of the students! With multiple choice the pass level should maybe be increased to a higher level...
truckflyer,
your amusement is clearly unjustified as a little piece of probability calculation quickly reveals:
The chance of a candidate who is merely guessing is following a binomial distribution:
Pn is the probability of k successes in n trials, p is the chance of success for each single trial, in our case 25%. Out of pure curiosity I calculated a few examples:
If there was only 1 question to answer, the chance of success is clearly 1:4 as you stated correctly. But don't get deceived by this apparently easy prey. If there were 9 out of 12 questions to be answered the chance decreases quickly to less than 1 : 2500. The shortest JAR-FCL exam subject (mass&balance) has 22 questions out of which 17 must be answered correctly to pass. Your chances of passing even this short test by pure luck are less than 1 : 2.5 million. A 40-question-test would give you a 1 : 21 billion probability, the longer exams are virtually impossible to "crack". It's just like playing roulette: to hit black or red once is easy. But to hit black 30 times in a row is almost impossible.
To sum this up: a pass level of 75% is more than enough to sort out the clueless. And if someone managed to pass all 14 subjects this reflects at least a profound knowledge of the question bank...
