Millisecond Forums

Subjects timing out in Heats and Flowers Task

https://forums.millisecond.com/Topic20070.aspx

By ramzadragoon - 11/8/2016

Hello,

I have programmed a Hearts and Flowers task, and I am concerned about participant inaccuracy due to lack of response stemming from my programming. Literature says that accuracy on these trials should be about 80% when the response window is 750ms, and we are seeing accuracies of about 60% instead.


There should be a 500ms pre-trial pause, a 500ms fixation cross, a 500ms blank screen, and then 750ms stimulus display ("Right")/response window.
values.respwindow is set to 750.


Here is my code. 

<trial CongRight>
/ontrialbegin = [picture.Right.setitem(picture.heart.currentitem, 1); if (expt.heartsflowers.currentblocknumber != 5) values.score = 0]
/stimulustimes = [1 = fixation; 500 = nofix; 1000 = Right]
/ beginresponsetime = 1000
/ inputdevice = keyboard
/ validresponse = ("q", "p", noresponse)
/correctresponse = ("p")
/timeout = 1500 + values.respwindow
/pretrialpause = 500
/ontrialend = [if (trial.congright.correct) values.score = (values.score + 1)]
</trial>

The data reflect too many instances of individuals hitting the cap of 750ms response time and missing the trial. Did I do something like use beginresponsetime or timeout incorrectly?

Thank you for your help.
By Dave - 11/9/2016

ramzadragoon - Wednesday, November 09, 2016
Hello,

I have programmed a Hearts and Flowers task, and I am concerned about participant inaccuracy due to lack of response stemming from my programming. Literature says that accuracy on these trials should be about 80% when the response window is 750ms, and we are seeing accuracies of about 60% instead.


There should be a 500ms pre-trial pause, a 500ms fixation cross, a 500ms blank screen, and then 750ms stimulus display ("Right")/response window.
values.respwindow is set to 750.


Here is my code. 

<trial CongRight>
/ontrialbegin = [picture.Right.setitem(picture.heart.currentitem, 1); if (expt.heartsflowers.currentblocknumber != 5) values.score = 0]
/stimulustimes = [1 = fixation; 500 = nofix; 1000 = Right]
/ beginresponsetime = 1000
/ inputdevice = keyboard
/ validresponse = ("q", "p", noresponse)
/correctresponse = ("p")
/timeout = 1500 + values.respwindow
/pretrialpause = 500
/ontrialend = [if (trial.congright.correct) values.score = (values.score + 1)]
</trial>

The data reflect too many instances of individuals hitting the cap of 750ms response time and missing the trial. Did I do something like use beginresponsetime or timeout incorrectly?

Thank you for your help.

Your code is correct. Participants have a total of 750ms to respond once the "Right" stimulus appears on-screen. Assuming the other relevant <trial> elements are set up identically, the data you are seeing reflects actual participant (under)performance and is not an artifact of some programming mistake.
By ramzadragoon - 11/9/2016

Dave - Wednesday, November 09, 2016
ramzadragoon - Wednesday, November 09, 2016
Hello,

I have programmed a Hearts and Flowers task, and I am concerned about participant inaccuracy due to lack of response stemming from my programming. Literature says that accuracy on these trials should be about 80% when the response window is 750ms, and we are seeing accuracies of about 60% instead.


There should be a 500ms pre-trial pause, a 500ms fixation cross, a 500ms blank screen, and then 750ms stimulus display ("Right")/response window.
values.respwindow is set to 750.


Here is my code. 

<trial CongRight>
/ontrialbegin = [picture.Right.setitem(picture.heart.currentitem, 1); if (expt.heartsflowers.currentblocknumber != 5) values.score = 0]
/stimulustimes = [1 = fixation; 500 = nofix; 1000 = Right]
/ beginresponsetime = 1000
/ inputdevice = keyboard
/ validresponse = ("q", "p", noresponse)
/correctresponse = ("p")
/timeout = 1500 + values.respwindow
/pretrialpause = 500
/ontrialend = [if (trial.congright.correct) values.score = (values.score + 1)]
</trial>

The data reflect too many instances of individuals hitting the cap of 750ms response time and missing the trial. Did I do something like use beginresponsetime or timeout incorrectly?

Thank you for your help.

Your code is correct. Participants have a total of 750ms to respond once the "Right" stimulus appears on-screen. Assuming the other relevant <trial> elements are set up identically, the data you are seeing reflects actual participant (under)performance and is not an artifact of some programming mistake.

What a relief! Thank you so much for your quick response!