- UID
- 223251
- 帖子
- 193
- 主题
- 90
- 注册时间
- 2011-7-11
- 最后登录
- 2013-8-16
|
10#
发表于 2011-7-11 18:44
| 只看该作者
My understanding of Angoff method is that they think about the minimum qualified guy. On each question they estimate "what is the probability that the minimum qualified guy is going to get this question correctly" (it's got to be somewhere between 0.33 and 1.00, since a blind guess will give you 0.33).
Then they average the probabilities for each question on the exam, multiply by 100, and that gives you minimum passing percentage. Multiply by 240 questions, and you get the minimum passing score. If the MPS is not an integer and you're on the integer just below the MPS, there's a possibility that they'll look at your ethics score and determine whether it justifies pushing you over.
The old method was to take 70% of the average of the top 1%. Presumably when they switched the method and had to start thinking about the minimum qualified candidate, they probably imagined someone who was getting about 70% of the average of the top 1%, so the two methods probably yielded similar results, at least initially. |
|