Millisecond Forums

How to structure

https://forums.millisecond.com/Topic10945.aspx

By nashby - 7/17/2013

I am trying to figure out how I would structure an experiment where I have participants who need to make 30 choices from each of 20 different choice sets with the presentation of choice set being random.


So for example choice set 3 might be presented first with a participant making 30 choices in it with feedback about each choice, and then they might be presented with choice set 20 again making 30 choices, so on and so forth.


I though maybe defining blocks then randomly picking blocks but certainly there is a way to build 1 trial that does all of this in some kind of loop?

By Dave - 7/17/2013

There are many ways to do this. However, the information you provide is to unspecific and vague to suggest a particular approach or strategy. If you can re-phrase your question in more concrete terms, I'll gladly add my $ 0.02.

By nashby - 7/17/2013

So I have 20 choice sets. Each choice set contains two different gambles which are made up of two different outcomes which are probabilistically determined.


What I need is for each of the 20 choice sets to be presented randomly but 30 times once they are presented. So, if first gamble choice chosen at random was PAIR 1 I would chose between the options present in PAIR 1 30 times with feedback about the monetary outcome that occurred showing up after each choice. This monetary outcome should be determined by the probabilities present in the gamble selected. So, if I choose a gamble with 10% chance of 2 and 90% chance of 3 a random number between 1 and 100 should be drawn, if the number is 10 or less then feedback would be 2 else feedback would be 3.


By Dave - 7/17/2013

So I have 20 choice sets. Each choice set contains two different gambles which are made up of two different outcomes which are probabilistically determined.


Meaning -- within a given set -- the gamble choice is always the same?

By nashby - 7/17/2013

Correct.

By Dave - 7/17/2013

Then e.g.


<expt>
/ blocks = [1-20=myblock]
</expt>

<block myblock>
/ trials = [1-30=mytrial]
</block>

<trial mytrial>
/ posttrialpause = 500
/ stimulusframes = [1=A,B, debug]
/ validresponse = (A,B)
/ inputdevice = mouse
</trial>

<text debug>
/ items = ("Block: <%script.currentblocknumber%>| Trial: <%block.myblock.currenttrialnumber%>")
/ position = (50%, 10%)
/ erase = false
</text>


<text A>
/ items = choiceaitems
/ select = noreplace
/ selectionrate = block
/ position = (40%, 50%)
</text>

<text B>
/ items = choicebitems
/ select = current(A)
/ position = (60%, 50%)
</text>

<item choiceaitems>
/ 1 = "Set1_ChoiceA"
/ 2 = "Set2_ChoiceA"
/ 3 = "Set3_ChoiceA"
/ 4 = "Set4_ChoiceA"
/ 5 = "Set5_ChoiceA"
/ 6 = "Set6_ChoiceA"
/ 7 = "Set7_ChoiceA"
/ 8 = "Set8_ChoiceA"
/ 9 = "Set9_ChoiceA"
/ 10 = "Set10_ChoiceA"
/ 11 = "Set11_ChoiceA"
/ 12 = "Set12_ChoiceA"
/ 13 = "Set13_ChoiceA"
/ 14 = "Set14_ChoiceA"
/ 15 = "Set15_ChoiceA"
/ 16 = "Set16_ChoiceA"
/ 17 = "Set17_ChoiceA"
/ 18 = "Set18_ChoiceA"
/ 19 = "Set19_ChoiceA"
/ 20 = "Set20_ChoiceA"
</item>

<item choicebitems>
/ 1 = "Set1_ChoiceB"
/ 2 = "Set2_ChoiceB"
/ 3 = "Set3_ChoiceB"
/ 4 = "Set4_ChoiceB"
/ 5 = "Set5_ChoiceB"
/ 6 = "Set6_ChoiceB"
/ 7 = "Set7_ChoiceB"
/ 8 = "Set8_ChoiceB"
/ 9 = "Set9_ChoiceB"
/ 10 = "Set10_ChoiceB"
/ 11 = "Set11_ChoiceB"
/ 12 = "Set12_ChoiceB"
/ 13 = "Set13_ChoiceB"
/ 14 = "Set14_ChoiceB"
/ 15 = "Set15_ChoiceB"
/ 16 = "Set16_ChoiceB"
/ 17 = "Set17_ChoiceB"
/ 18 = "Set18_ChoiceB"
/ 19 = "Set19_ChoiceB"
/ 20 = "Set20_ChoiceB"
</item>


For advanced stuff such as scoring, see scripts like the Information Sampling Task or Columbia Card Task (available from the Task Library at millisecond.com).

By nashby - 7/17/2013

Thanks but how would I make this random in terms of what choice set is presented? here it looks like it would go through in the order they are on the item list?

By Dave - 7/17/2013

here it looks like it would go through in the order they are on the item list?


No. The <text> elements are set to 'noreplace', i.e. random selection without replacement. Run it and see for yourself.