Multiple Randomization for Risk Tasking Task

liznik — Tue, 01 Nov 2016 12:14:22 GMT

New to programming an entire task on inquisit. Not sure where to begin. I need participants to select an one of two images, one being a 'risky' choice and one being a 'safe' choice. I also want to randomize which of the two images end up being risky and safe.

If they chose the safe option they have a 10% chance of getting an image that represents a lose of 20 points, a 10% of getting an image that represents a gain of 20 points, and the other two images are a 40% of gaining 10 points or 40% of losing 10 points.

The risky option has the ratios flipped, so 40% of gaining/losing 20 points, 10% of losing/gaining 10 points.

We also DO NOT want total points to show anywhere on the screen.

They have a learning phase that I have already programmed where they will be learning the values of the images so they will already know how many points they will have as a result

RE: Multiple Randomization for Risk Tasking Task

Dave — Tue, 01 Nov 2016 12:14:22 GMT

The basic mechanics go like this (using <text> elements for the sake of example; it works the same with <picture> elements):

<block myblock>
/ trials = [1-10 = choicetrial]
</block>

<trial choicetrial>
/ inputdevice = mouse
/ stimulusframes = [1=safe, risky]
/ validresponse = (safe, risky)
/ branch = [if (trial.choicetrial.response == "safe") list.safechoice.nextvalue]
/ branch = [if (trial.choicetrial.response == "risky") list.riskychoice.nextvalue]
</trial>

<text safe>
/ items = ("SAFE")
/ position = (40%, 50%)
</text>

<text risky>
/ items = ("RISKY")
/ position = (60%, 50%)
</text>

<list safechoice>
/ items = (trial.20loss, trial.20gain, trial.10loss, trial.10gain)
/ itemprobabilities = (.10, .10, .40, .40)
/ poolsize = 10
/ replace = true
</list>

<list riskychoice>
/ items = (trial.20loss, trial.20gain, trial.10loss, trial.10gain)
/ itemprobabilities = (.40, .40, .10, .10)
/ poolsize = 10
/ replace = true
</list>

<trial 20loss>
/ stimulusframes = [1=20losstxt]
/ validresponse = (57)
</trial>

<trial 20gain>
/ stimulusframes = [1=20gaintxt]
/ validresponse = (57)
</trial>

<trial 10loss>
/ stimulusframes = [1=10losstxt]
/ validresponse = (57)
</trial>

<trial 10gain>
/ stimulusframes = [1=10gaintxt]
/ validresponse = (57)
</trial>

<text 20losstxt>
/ items = ("-20 points")
</text>

<text 20gaintxt>
/ items = ("+20 points")
</text>

<text 10losstxt>
/ items = ("-10 points")
</text>

<text 10gaintxt>
/ items = ("+10 points")
</text>

The two <list> elements are responsible for the varying probabilistic outcomes based on "safe" vs. "risyk" choice.

The 20gain, 20loss, etc. <trial> elements would be used to present the respective images instead of <text> elements; you would hanclle any scoring (if so desired) via /ontrialbegin or -end logic in those <trial> elements.

Millisecond Forums » Millisecond Forums » Inquisit 5 » Multiple Randomization for Risk Tasking Task

Multiple Randomization for Risk Tasking Task

RE: Multiple Randomization for Risk Tasking Task