Sorry Barron - my mistake. You need to alter the jump scripts a little, since the mod command will return a value from zero to four - i.e. five values, but starting at zero. In the scripts I wrote out here I began the jump statements from 1. My bad.
The mod command is important since it completes the random number generation and matches the number of questions that you have. If you don't want this randomised then it isn't necessary at all.
I have now set up a simple project for you to look at. It definitely only allows me to do five questions, but the randomiser will repeat - I haven't yet written the scripts to track what has already played. You can get it from my .mac space:
http://homepage.mac.com/halgernon
You are looking for a .zip file named 'quiz.zip'. In there is a disc image and the project file (in PAL format) which I used. The way it works is that if the user gets the answer wrong they return to the same question. If they get it right they get a chance at a new question. After five questions the end is reached, but they can choose to try again. You will see that I have re-named the GPRMs to show what I was doing, and you may notice one unused one is named for scoring - but like you I took it out... I couldn't decide if the user should score when they get it wrong then repeat it and get it right. That's something for you to decide, I think!
True random will require bit-wise maths, or GPRM partitioning. Essentially, what you do is move a value into a partition each time a question is asked. before a new question gets displayed, you first check to see if it has already been asked (i.e. check to see if the value is in place). If it is, you can either generate another random number, or simply subtract 1 from what you had and check that slot.
The more questions in the quiz the more likely it is that you'll not get repeats. Three to five questions
will show repeats, almost certainly, but 80 or more will make it unlikely.