Tool for pseudo-randomisation

December 1st, 2006

Whilst writing psychology experiments, one thing that I’ve come across time and time again is the ability to randomize data based on multiple constraints.  The first couple of times I came across this I simply hacked out a piece of Java using a brute force method that generated an order and then checked to make sure that all the constraints were valid.  This worked pretty well, but clearly wasn’t that scalable.  One other solution I could have used was to use a clever language like Prolog and do some constraint based programming.  That takes me back to the constraint based programming I did as an undergrad with Dr Andy King.  Anyway, after a couple of chats with different people I came across this rather neat tool called “Mix”.  It basically is a small program that allows you to specify exactly how you want your information randomized.  Furthermore the method in which you provide the specific is very short and to the point meaning that to make a specification that can generate 1000 sets takes a matter of seconds.  I’ve used this tool for a number of different experiments and it works well.  Although I’ve only just started to look into what else it can do.  So if you are in need of the ability to randomize information using multiple constraints, possibly fixed block position and anything else then you may want to give this tool a try.


Leave a Reply

You must be logged in to post a comment.