- Joined
- Oct 11, 2006
- Messages
- 10,418
- Reaction score
- 14,645
I can't comment on anyone's chances of passing, but I can comment on the scoring.
As mentioned above, different questions are weighted differently. That's why it's possible to get a higher percentage correct, yet get a lower scaled score -- or why someone with a lower overall percentage can pass, where someone with a higher percentage can fail. However, in these cases it's usually just 1-2 questions making the difference.
Because of the nature of a scaled score, the "scale" becomes very steep at the ends of the curve. One question can have a large impact on your scaled score when you're down at the pass/fail margin, where when you're up near the mean, it will have a very small effect.
Contrary to what you might think, it's the "easier" questions that are weighted more heavily. Rather than try to decide which questions are easy and which are hard, the exam uses test-taker's answers to assess that. If there's a question that very few people get correct, then it's a "hard" question (or poorly written), and it counts for less. If enough people get it wrong, it's removed completely. If a question is answered correctly by everyone but you, that counts quite a bit.
Before any question is actually used to score, they are tested -- hence the experimental questions. They're building a database of questions and answer profiles, to see what questions will be used and how they will be scaled.
Anyway, good luck to all.
As mentioned above, different questions are weighted differently. That's why it's possible to get a higher percentage correct, yet get a lower scaled score -- or why someone with a lower overall percentage can pass, where someone with a higher percentage can fail. However, in these cases it's usually just 1-2 questions making the difference.
Because of the nature of a scaled score, the "scale" becomes very steep at the ends of the curve. One question can have a large impact on your scaled score when you're down at the pass/fail margin, where when you're up near the mean, it will have a very small effect.
Contrary to what you might think, it's the "easier" questions that are weighted more heavily. Rather than try to decide which questions are easy and which are hard, the exam uses test-taker's answers to assess that. If there's a question that very few people get correct, then it's a "hard" question (or poorly written), and it counts for less. If enough people get it wrong, it's removed completely. If a question is answered correctly by everyone but you, that counts quite a bit.
Before any question is actually used to score, they are tested -- hence the experimental questions. They're building a database of questions and answer profiles, to see what questions will be used and how they will be scaled.
Anyway, good luck to all.