About shuffling

Added by Gregório R. de Miranda over 2 years ago

Hi all!
I always shuffle my exams and everything works great! But recently I come across something peculiar: I noticed that some copies have the exact same order than my source file.
That wouldn't be a problem if I didn't put all correct answers on the first place. That way, some students have a very strange answer key: A, A, A, A...

So, (1) there's some setting I could do to avoid it?
(2) if not, how do you handle with it? I guess I'll kinda sort oddly directly in the source file from now on.


Replies (3)

RE: About shuffling - Added by Alexis Bienvenüe over 2 years ago

Maybe the default random seed is not very good (the random generator is not good), and you can use another one, like \AMCrandomseed{1237893}?

RE: About shuffling - Added by Gregório R. de Miranda over 2 years ago

Yeah, looks like it helped indeed.
An example:

Before:

After:

Doing this test, however, I found that maybe I wasn't completly fair. I didn't found copies sorted the exact same way than my source file as I said (I guess I was looking the catalog hehe), but I did found some homogeneous answer keys like the example above. On this source file, every correct answers were the first one.

On this another one, however, correct answers in the source file are sorted as A, A, A, D. The questions were shuffled, but the choices, not so much.

After changing the random seed as you suggested, it looks like it:

Not a big change, but I guess it's a bad example, as it's very small.

Changing the seed for 1587893 produced:

RE: About shuffling - Added by Gregório R. de Miranda over 2 years ago

Gregório R. de Miranda wrote:

Yeah, looks like it helped indeed.
An example:

Before:

After:

Doing this test, however, I found that maybe I wasn't completly fair. I didn't found copies sorted the exact same way than my source file as I said (I guess I was looking the catalog hehe), but I did found some homogeneous answer keys like the example above. On this source file, every correct answers were the first one.

On this another one, however, correct answers in the source file are sorted as A, A, A, D. The questions were shuffled, but the choices, not so much.

After changing the random seed as you suggested, it looks like it:

Not a big change, but I guess it's a bad example, as it's very small.

Changing the seed for 1587893 produced:

\setdefaultgroupmode is set as {withoutreplacement}

(1-3/3)