Millisecond Forums

Hi there and first of all happy new year to all of you.

Now, I'm currently re-building an experiment in which subjects can only react to stimuli by pressing a or 6 to indicate a 'yes' or 'no' answer.

For some reason subjects were allowed to choose whether '6' was 'no' and 'a' was 'yes' or vice versa in the initial experiment. I want to recreate this feature too, because I'm likely going to use this preference more often in the future.

Now, I have created the only surveypage for the experiment. Subjects will enter gender, age and whether they want that "a" or "6" to be the yes-key using radiobuttons.

I want to use subjects' preference in the course of the entire experiment, since I intent to use errormessage and correctresponse attributes. I therefore created two versions of the experiment, two versions of each block and two versions of each trial. Yet, in the branch attribute I can only link my subjects' preference to a specific trial or block, not to an entire experiment element.

My first question is this: how can I get subjects' preference all the way through to the level of the experiment?

My second question would be if there is an easier way to use the two responsemanners than to create two versions of every trial, block and experiment like I did? (since it's time-consuming and errors slipped in every now and then)

Any answer to any of these questions would surely help me a lot.

Regards,

SWQ

As far as I know, as of version 3.0.3.2 Inquisit allows branching between <expt> elements, so this should be perfectly possible.

~Dave

I do use that version of Inquisit. After your response I checked again and still couldn't link the preference to an entire experiment element.

A solution was to change the name of the expt element. First it was called version_if_6_is_yes. This didn't work. Only when I changed both the expt element and the branch attribute to 'version1' it seemed to work. I suppose underscores don't work that well in branches then?

Anyway, thanks for your information and having me take a look at it again. It helps me a lot since at our faculty, we often use participants' preferences and I kept having this particular problem.

This sounds like a (potential) bug. You might want to additionally report this via email so it gets fixed in upcoming releases.

Cheers,

~Dave

After checking before sending a notice, I found out that it was no such much the underscores but the 'if' that caused the problem. Probably because 'if' already has meaning in the branch attribute and thus clouded up the branch.

Regards,

SWQ

So yesterday after Dave's reply and some new insight I was confident I could apply my subjects' preference.

Now, I can run the experiment but after the surveypage (containing the question of whether 6 or a correspond with yes or no) Inquisit quits functioning and tells me that the branches contains - I quote -  wrong element types.

The branches are as follows:

/branch = [if (radiobuttons.preference.response = 1) expt.version6]
/branch = [if (radiobuttons.preference.response = 2) expt.versiona]

The expt. elements are as follows:

<expt version6>
/blocks = [1-2=sequence(practice, target)]
</expt>

<expt versiona>
/blocks = [1-2=sequence(practicea, targeta)]
</expt>

In my humble estimation, that should fit perfectly. Yet it doesn't. The error 'wrong element type' makes me suspect expt elements are not applicable in branches. Yet, considering Dave's comment, I doubt that.

Another question on this topic is the following: since the choice of an experiment depends on the subjects' preference, I decided to make a third experiment just for the surveypage. Since if I build the surveypage into either/both expt version6 and versiona, participants are already in an experiment where the yes and no key have been determinated. Therefore, when I press ctrl +F5, I want Inquisit to run the expt.survey element before choosing between expt.version6 and expt.versiona. The problem here is that I don't know how to adress this order of expt. elements in Inquisit. As far as I know, there is no 'higher' level in which you can arrange the expt elements (like you can arrange block elements via the blocks attribute in the expt element).

So these are my two further questions:
1) What can possibly be wrong with the branching?
2) How can I get Inquisit to start with the expt.survey element?

Regards,

SWQ27

I agree that there are some oddities related to branching to / between <expt> elements. However, the good news is that the desired functionality can still be achieved by a slightly modified approach. I've attached a small demo script to help along the way.

Best wishes from a fellow Inquisit user,

~Dave

P.S.: Note that when a script contains multiple <expt> elements without assignment rules (/ subjects), they appear to be run in alphabetical order.

Wauw, that is exactly what I needed.

Thank you so much again for the effort and thoroughness of your reply. Because Inquisit takes everything in your syntax literal - and because I am not familiar with any form of programming besides Inquisit - it can be extremely difficult to even find out what the problem is. Let alone solve it. With English not being my first language, all the information I can get requires twice the concentration it'd normally cost me. Your help amazes me because this would have cost me hours and hours to solve by myself, if I'd even get to the solution that is.I'll apply it immediately & again: your help is much appreciated.

Again, you're very welcome and I'm glad the solution did the trick for you.

~Dave