By rnglab - 9/9/2015
I am using Inquisit 4's "Color Word Stroop with Speech Recognition" script. This script has congruent trials, (e.g. the word "yellow" printed in the color yellow) and incongruent trials (e.g. the word "yellow" printed in red). I would like to keep the incongruent trials, but I would like to replace the congruent trials with control trials (e.g. the character string "XXXXXX" printed in the color yellow). I would like all trials to be either incongruent or control trials, so, for example, there would be no trials that present the word "yellow" in yellow ink, and all possible stimuli printed in yellow ink would be: red, blue, black, green, XXXXXX. I need to do this because I am interested in isolating interference effects from incongruent trials without measuring facilitation effects from congruent trials.
Perhaps related, I would like the automatic score calculations to still work. The metric I need is (average reaction time for all correct responses to incongruent trails) - (average reaction time to all correct responses to control trials).
I've tried playing with the script quite a bit but I cannot get this to work. If necessary, I can post the attempts I've made, but I figure there is probably an easy answer that I'm missing. I have a small amount of scripting experience but not with this program.
Thank you.
|
By Dave - 9/9/2015
#1: I must assume you only want to run the *color naming* portion of the script. The *word naming* portion would not make any sense since the string "XXXXXX" is not pronounceable. Correct?
#2: If so (and if i understood your description correctly), all you need to do then is create separate <item> elements for each color. I.e. instead of
<item colors>
/1 =" red "
/2 =" green "
/3 =" blue "
/4 =" black "
/5 =" yellow "
</item>
<text red>
/items=colors
/color = (255,0,0)
</text>
<text blue>
/items=colors
/color = (0,0,255)
</text>
...
you set up
<item colors_red>
/1 =" XXXXX "
/2 =" green "
/3 =" blue "
/4 =" black "
/5 =" yellow "
</item>
<item colors_blue>
/1 =" red "
/2 =" green "
/3 =" XXXXX "
/4 =" black "
/5 =" yellow "
</item>
...
<text red>
/items=colors_red
/color = (255,0,0)
</text>
<text blue>
/items=colors_blue
/color = (0,0,255)
</text>
...
What were previously "congruent" trials are now "control" trials.
|
By rnglab - 9/11/2015
Thank you very much, Dave. If I knew where you lived, I'd order you a fruit basket.
1. Correct--I deleted the word naming portion because I only need color naming.
2. This worked perfectly. A part I didn't mention is that some recommend that the string length of control trials match the character length of the word, but this was easy to add. For example, yellow printed in yellow became ****** and red printed in red became ***.
I have to add more trials and increase the proportion of control trials, but I see other topics on that in the forum that will help.
|
By Dave - 9/11/2015
To increase the proportion of "control" (formerly "congruent") trials, it would be easiest to set up a bunch of <list> elements and use those for item selection. Let's look at <trial red> as an example:
<trial red>
/ontrialbegin = [values.color = "red"]
/stimulustimes = [0=red]
/posttrialpause = 250
/validresponse = ("yellow" "green" "blue" "black" "red")
/correctresponse = ("red")
/ontrialend = [if (text.red.currentitemnumber == 1) {
values.count_congruent += 1;
values.congruence = 1}
else {
values.count_incongruent += 1;
values.congruence = 2}]
/ontrialend = [if (text.red.currentitemnumber == 1 && trial.red.correct) {
values.countcorrect_congruent += 1;
values.sumrt_congruent += trial.red.latency}]
/ontrialend = [if (text.red.currentitemnumber != 1 && trial.red.correct) {
values.countcorrect_incongruent += 1;
values.sumrt_incongruent += trial.red.latency}]
</trial>
As you can see it displays the stimulus element <text red> with the below items. Whenever the 1st item is selected (the Xs), the trial is recorded as "control" (formerly "congruent") as per the /ontrialend logic highlighted above.
<item colors_red>
/1 =" XXXXX "
/2 =" green "
/3 =" blue "
/4 =" black "
/5 =" yellow "
</item>
...
<text red>
/items=colors_red
/color = (255,0,0)
</text>
What that effectively means is that 1 in every 5 instances (20%) of <trial red> will be "control" / "congruent".
Now, let's say you want to increase the proportion of "control" / "congruent" instances to 50% (or 4 out of 8). To do this, you could set up
<list red_items>
/ items = (1,1,1,1,2,3,4,5)
/ selectionmode = random
</list>
which contains item #1 four times, and use that list for item selection in <text red>:
<text red>
/items=colors_red
/color = (255,0,0)
/select = list.red_items.nextvalue
</text>
The changes for the other color <trial> and <text> elements (<trial green> / <text green>, <trial blue> / <text blue>) would be analogous. Only the item numbers differ in each case.
|
|