| Reference | Downloads | Github

Precautionary question on editing experiment whilst running


I will be launching my experiment later this week. Currently, it picks a condition at random, using the following JS code:

function random_character() {
    var chars = "ABCD";
    return chars.substr( Math.floor(Math.random() * 4), 1);

condition = random_character();

Where my condition file is specified by:

condition: $'Order' + condition + '.xlsx'
(Files: OrderA, OrderB, OrderC, OrderD)

Since the condition is chosen at random, I will occasionally need to download the data files and tally which conditions have been run. In the case where one condition is being run more than others, I will need to remove the condition file from the experiment and let the experiment continue with the remaining condition files (side question: I assume that means that Math.floor becomes Math.floor(Math.random() * 3), 1) if I have 3 conditions?).

My question is whether this is “safe” and possible to do while the experiment is running (assuming I should temporarily put it in Piloting/Inactive during editing).

Thanks in advance.

Removing files sounds dangerous (what if you do it just when a session is trying to access it?), and will still be very difficult to achieve balancing with.

If you need to do exact balancing, I guess this starts with the service you are using to handle your recruiting. i.e. in the page here, it shows how you can send parameters in the URL for each session:

e.g. in your case:

and then you will have access to the order value in code via expInfo['order']