Millisecond Forums

All of these depend on the specifics of what you want to do, without those your questions are impossible to answer.

A few more thoughts up front:

This strikes me as a matter of participant managment rather than a problem that an Inquisit script could solve for you.

This can also be construed as a matter of participant management. You can randomly assign conditions up front (outside of any Inquisit script) and pass them to Inquisit Web via URL query parameters. Details however depend on specifics.

Inquisit allows for subject-entered (alphanumerical) ids. Alternatively you can always capture whatever additional id you want in an Inquisit script via <openended> or <textbox>/<surveypage>.

Hey Dave,

thanks for your helpful input. The experiment has further ripened and the specifics have become clearer. We have purchased an Inquisit 4 web license for one script and would like to run it on the Millisecond servers.

First off, we will gather participants for the online study per Mechanical Turk. This means that we will get their email addresses, to which we'll then send the URL with the query parameters that assign the subject to an experimental condition, as you suggested.

Half of the participants will be completing online surveys elsewhere on the web inbetween Inquisit experiment time 1 and time 2, meaning that we need them to enter a subject code, in order to identify them with their input on the other survey. Is it possible to prompt the user to enter an ID code/subject number while still sending them to a specific experimental condition per URL? Is it simpler just to gather the code per <textbox> and only use the URL query parameter for placing a subject in a condition?

Next question, though not unrelated: We only have a one script license, yet will be running the script at two different times in a slightly different form. The different form includes different blocks of the experiment and different start and end pages. We could simply create double the amount of experimental conditions in <expt> and then assign them to another one - only how could we program it in a way that we could identify the subject in the different conditions? Maybe per above-mentioned subject code?

Thanks a lot for your help!

Best,

Marton

Both can work, but using a simple <openended> or <textbox> to me seems like the simpler solution in this particular case.

Yes, I don't see why that wouldn't work based on the information you provided. Perhaps the best way to figure out the details would be to set up a realistic test scenario. I.e., put simplified surveys and scripts online and then bounce back and forth between them as needed.

Hey Dave,

you're a real life saver. Thanks very much for your help!

We're going to test the set-up but I agree it's looking good. I have a question regarding the personal code in <textbox> or <openended> to which I didn't find anything in the forums or the web, though.

I'd like to be able to check if the above-mentioned personal code was given in a correct form, for example per <correctresponse>. The code might consist of 3 letters, then a number and then another letter. Is there a way to do this?

Also, I was considering having them enter the code at the beginning and again at the end of the experiment, to make sure they get it right. I saw it's possible to do this per <correctresponse>. How would I integrate this with the other <correctresponse>-check and how would I then give them the chance to completely reenter the code, if they happen to have made the mistake upon first entering the code and not upon the second time?

Thanks very much!

Best,

Marton

The /mask attribute allows for this. It supports regular expressions.

You would use /iscorrectresponse and /branch attributes to do this.