Millisecond Forums

See the "How to erase stimuli" topic in the Inquisit documentation.

Thanks. I think I'm sticking with <surveypage> for now. Seems to be doing the trick.

On a different note...

My vignettes seem to be leaving remnants of the previously presented vignette on each randomized trial.

In other words, when a new vignette is presented, I can see some of the words from the vignette shown just before. It appears behind the new vignette's text. It looks like jumbled letters. I've tested it on multiple computer screens. Any idea why this might be happening? Is it something in my script?

Sample of one block:

<block vignetteorder1>
/ bgstim = (vignetteinstructions1)
/ trials = [1-4 = noreplace(vignettepage1, vignettepage2, vignettepage3, vignettepage4)]
/ recorddata = true
/ preinstructions = (instructionspage1)
</block>

Let me know if you want to see more of my script.

Greater flexibility, more concise and elegant code. It does, however, require a somewhat advanced understanding of Inquisit's programming model (i.e., some experience and lots of practice). For examples, which -- abstractly -- cover such things, see e.g. the Tower of London script or this short example:

http://www.millisecond.com/forums/Topic6989.aspx#6989

Nevertheless, if the <surveypage> approach works fine for you here, stick with it.

I got the gifs to play on the right side using the <video> element -- thank you, Dave!

Right now, the <surveypage> element seems to be accomplishing what I need. I was able to get everything I wanted displaying on the left side, with the vignettes randomized. I think I am forgoing randomization of the order of the likert questions for now.

But what are the advantages of using the <trial> element to build the likert scale questions and display them with the randomized vignettes?

Could you point me to a sample script? If there are particular advantages to using <trial> over <surveypage>, I'd love to see an example of how to use <trial> to create:

• a series of randomized vignettes, with...


• two 7-point likert scale questions displayed one on top of the other on the same page under each vignette, and...


• a "next" button at the bottom of each page.

Thanks again for all of your speedy, helpful replies.

I should also add that in order to display animated GIFs, you ought to use the <video> element, not <picture> or <image>.

#1: As I've said before, I would strongly advise to use *<trial>* elements. Do not use <surveypage> or <likert>.

#2: If, for some reason, you must use <surveypage>s, use their /stimulusframes to display <text> or <picture> elements.

#3: You can create a separate <surveypage> for each vignette, then randomize the order of the pages at the <block> level.

Here's where I am at:

I have decided to use a .gif as the blinking dot. There are 4 different .gifs that demonstrate different randomly timed blinking sequences. Participants will be instructed to respond to the blinking dot, making them think that their responses to the blinking dot matter. But the only responses that matter are their responses to the vignettes on the likert scales that appear on the same screen.

I've been able to develop custom likert scales using the <slider> element and get the two to be displayed at the same time using the <surveypage> element. But I don't know how to get the vignettes to show up and be randomized. Do I have to use something different than the <surveypage> element to do this? I was hoping to be able to use /stimulusframes to get the vignettes to display, but that doesn't seem to work.

Additionally, I cannot get the blinking dot to show up on the same screen and be randomized. Is the <surveypage> element blocking the dot <picture> element from appearing alongside the sliders?

A revised breakdown of what I need to happen:

• present vignettes one at a time to participants with two likert scales (sliders) that remain on the screen


• randomize the order of the vignettes between participants (different order for each participant)


• clear likert responses between vignettes (or position slider back to neutral each time)


• randomize the order that the likert scales are presented when each new vignette is presented


• calculate response time/reaction time on each vignette


• get the .gifs to present in random order on the same page as each vignette/likert scale presentation

Thanks for your help thus far!

Yes, that would be possible as far as I can tell. However, note that this only reduces complexity by a modest amount. Most of the caveats outlined previously still apply.

Thank you for your response, Dave!

What if I forfeited recording accuracy of the dot responses? So instead, the participant thinks data will be collected on their dot response accuracy, but it is just a dot flashing on the screen at random times without anything being recorded.

Could I then have the vignettes/likert scales be the only thing on the screen collecting responses while the dot continued to blink at random?

It might be possible.

(1) You'll want to set the elements representing the vignette, likert questions etc. to /erase=false so the remain on-screen across trials.

(2) You'll need at least two different <trial> elements. One to collect the likert responses (you cannot use the <likert> element here; you'll have to use standard <text, <picture> etc. elements to build them), one to display and collect response to the dot.

(3) You'll have to set up various <values> and <expressions> as well as make use of various timing properties (elapsedtime, latency) to calculate response times.

(4) During "dot"-trials, you'll may not be able to respond to the likert question (and possibly vice versa).

(5) You'll have to work out a relatively complex timing scheme for the dot flashes and use /timeout and/or /trialduration to implement the scheduling (alternating between vignette response and dot response trials).

All in all this requires a somewhat advanced understanding of Inquisit's programming model. I would thus recommend you first get down the basics by completing the various tutorials in the documentation (if you haven't done so already). Then move on to study a selection of scripts from the Task Library which illustrate more advanced programming approaches and techniques.

The overall limitations you'll have to cope with are

(1) You cannot run multiple <trial> elements concurrently (as would be required for a true "concurrent" dual-task), only one at a given time.

(2) A given <trial> element only allows input from a single input device, i.e., you cannot have it collect mouse (clicking a likert scale) AND keyboard responses (space bar press).

Good luck!