Hi everyone,

I am trying
to expand the source code for the IOWA-Gambling task available at the Inquisit
website. In this version the loss-values are selected randomly as follows:

 <expressions>

/ getlossdeck1
= noreplace(0, 0, 0, 0, 0, 150, 200, 250, 300, 350) 

</expressions>

By contrast,
I want to define a fixed sequence of loss values for loss values for each trail.
I want to choose the current value depending on how many cards were already
taken from the deck. Speaking more technically - I want to get the nth element
out of an array, vector or list with numerical values.

Therefore I
am trying to define something like a list or a vector with loss values (integer
values) for each trail.

 / loss_deck1
= (0, 0, 0, 0, 0, 150, 200, 250, 300, 350) 

 Since that
did not work I have defined a counter

 <counter
deck1_loss>

/ items = (0,
0, 0, 0, 0, 150, 200, 250, 300, 350)

/ select = sequence(1-10)

</counter>

 Additionally,
I have introduced a variable "values.cards_deck1" that counts how
often deck1 was chosen. This counter is updated in each trial.

In the last
step I try to define the loss value for the current trial as:

/ getlossdeck1
= getitem(counter.deck1_loss.items, values.cards_deck1)

But each
time I get the error message "Expression is not valid. Item not found".

Does
someone have an idea whether it is possible to create a list or vector with
numerical values & get back a specific element out of this list?

Thanks a
lot for your help in advance.

Best regards, 

Olga Kunina

Hmm, glancing over the code you posted, I think

/ getlossdeck1
= getitem(counter.deck1_loss.items, values.cards_deck1)

should be changed to a /ontrialend or /ontrialbegin expression:

<values>
/ getlossdeck1 = 0
[...]
</values>

and then

<trial sometrial>
/ ontrialbegin = [values.getlossdeck1=getitem(counter.deck1_loss, values.cards_deck1]
[...]
</trial>

If that doesn't do the trick, you might want to attach your script to this thread (-> see 'Options' tab when posting a reply).

~Dave

Hi Dave,

 thank you
very much for your quick answer.  I have
followed your suggestions and changed my code as follows:

 <trial
igtfeedback>

   / ontrialbegin
= [ if (trial.igt.response == "deck1") { [..]

    values.loss
= getitem(counter.deck1_loss, values.cards_deck1); [..] }]

 </trial>

 Unfortunately
it does not work. It seems to be a problem with the expression "getitem(counter.deck1_loss,
values.cards_deck1)".

 I have
uploaded the script. Thanks a lot for taking time to answer the questions,

Best regards,

Olga

What exactly isn't working?

Okay, I had a quick look at your script and compared it to the original IGT script available from millisecond.com. Well, if all you want is a fixed sequence of losses for each deck, there's a much easier way to pull this off. In the original script, losses are randomly determined by the following expressions:

<expressions>
/ getlossdeck1 = noreplace(0, 0, 0, 0, 0, 0, 0, 0, 0, 1250)
/ getlossdeck2 = noreplace(0, 0, 0, 0, 0, 150, 200, 250, 300, 350)
/ getlossdeck3 = noreplace(0, 0, 0, 0, 0, 0, 50, 50, 50, 50)
/ getlossdeck4 = noreplace(0, 0, 0, 0, 0, 0, 0, 0, 0, 250)
</expressions>

Each time the subject selects e.g. deck1, a random value is returned from expressions.getlossdeck1 (without replacement). Now, if you want to switch that behavior to a predictable, fixed sequence in which certain amounts of losses should occur, you may simply define:

<expressions>
/ getlossdeck1 = sequence(0, 0, 0, 0, 0, 0, 0, 0, 0, 1250)
/ getlossdeck2 = sequence(0, 0, 0, 0, 0, 150, 200, 250, 300, 350)
/ getlossdeck3 = sequence(0, 0, 0, 0, 0, 0, 50, 50, 50, 50)
/ getlossdeck4 = sequence(0, 0, 0, 0, 0, 0, 0, 0, 0, 250)
</expressions>

I.e.only when the subject selects deck1 for the tenth time, a loss of 1250 will occur. No further changes are necessary as you can see from the attached script.

Let me know if that helps,

~Dave

Dear Dave,

thank you so much for your help. That is indeed a very simple soultion for this probelm. [:D]

However, we are developing a new card risking task and there it would be helpful to be able to get the n-th element of a counter. Unfortunatelly there are not many examples for the usage of the function getitem in the documentation .

getitem(counter.deck, 1)

getitem(text.presenteditems, text.presenteditems.itemcount)

getitem(item.useritems, item.useritems.itemcount - 1)

Can you tell me how one would have to define "counter.deck", "text.presenteditems", or "item.useritems" so that this function is executed regularely?

Kind regards,

Olga

Hmm, I don't think there's much more too it. Let's do this by way of an example. Suppose you have defined a counter element:

<counter fruit>
/ items = ("apple", "banana", "cherry", "damson", "elderberry")
[...]
</counter>

Now, suppose you want to retrieve the third item from the counter, i.e. "cherry", with the getitem function. You can use that function either by way of the <expressions> element

<expressions>
/ getcherry = getitem(counter.fruit, 3)
[...]
</expressions>

or by way of Inquisit's event attributes ('/ onexptbegin', '/ onexptend', '/ onblockbegin', '/ onblockend', '/ ontrialbegin', '/ ontrialend') available at the <expt>, <block> and <trial> level

<trial getcherry>
/ ontrialbegin = [text.favoritefruit.item.1 = getitem(counter.fruit, 3)]
/ stimulusframes = [favoritefruit]
[...]
</trial>

<text favoritefruit>
[...]
</text>

and do all sorts of things with it, e.g. write it to a <text> element or to a <values> entry. Note that the item index number, i.e. the second argument of the getitem function  ( 3 for "cherry") may also be a variable. So you might as well define another <values> entry (or use anything that returns a valid numerical value) and use it as the functions argument:

<values>
/ cherryindex = 3
[...]
</values>

<expressions>

/ getcherry = getitem(counter.fruit, values.cherryindex)

[...]

</expressions>

Best wishes from a fellow Inquisit user,

~Dave

Dear Dave,

thanks for these examples and your answers. They were very helpfull. [:)]

Have a good weekend. With best wishes,

Olga

You're welcome. BTW, if you're into risk-taking card tasks, definitely check out the Columbia Card Task script available from the Inquisit Task Library.

Cheers,

~Dave