Problems with the scrambled sentence task


Author
Message
Orchid
Orchid
Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)
Group: Awaiting Activation
Posts: 6, Visits: 38
Dear all,

I downloaded the scramled sentece Task from Inquisit library and adapted it, so that it works for Inquisit 4.
When I look into the data file, it is not possible to see whether the subject was too slow to unscramble the sentence.
As far as I understand, the subject has 10s to unscramble the sentence. In case the subject was too slow to unscramble the sentence, the Computer will "fill up" the missing words from the left (1) to the right (6)

Example:
Scrambled sentence:
Weather (1) good (2) was (3) today (4) bad (5) the (6)

The right order would be for example:
Today (4) the (6) weather (1) was (3) good (2)

But when the subject is too slow to unscramble the sentence, (just clicked at "Today (4) the (6) weather (1)...")
the Computer will fill up the empty spaces with word 4 and 5. So you get "Today(4) the(6) weather(1) today (4) bad (5)"
In this case it is obvious, that the subject was too slow and the Computer filled up the sentence, because of the repeated word (Although when you look at reaction times, they are under 10s).
But in some cases the sentence still makes sense even when the Computer fills up the missing words, for example

Scrambled sentence:
ate(1) I(2) today(3) a(4) Sandwich(5) banana(6)

When the subject was too slow to Response (only clicked at "Today(3) I(2) ate(1)"), the Computer will automatically fill up the empty spaced with word 4(a) and 5(Sandwich), so that the Sentence makes perfectly sense "Today I ate a sandwich"). 
In this case you cannot see, whether the subject was too slow or did it right, since the reaction times are still under 10s.

Is there any other possibility to see in the data file, whether the subject's response was too slow?

Best wishes


 

Dave
Dave
Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)
Group: Administrators
Posts: 13K, Visits: 104K
Orchid - Friday, June 30, 2017
Dear all,

I downloaded the scramled sentece Task from Inquisit library and adapted it, so that it works for Inquisit 4.
When I look into the data file, it is not possible to see whether the subject was too slow to unscramble the sentence.
As far as I understand, the subject has 10s to unscramble the sentence. In case the subject was too slow to unscramble the sentence, the Computer will "fill up" the missing words from the left (1) to the right (6)

Example:
Scrambled sentence:
Weather (1) good (2) was (3) today (4) bad (5) the (6)

The right order would be for example:
Today (4) the (6) weather (1) was (3) good (2)

But when the subject is too slow to unscramble the sentence, (just clicked at "Today (4) the (6) weather (1)...")
the Computer will fill up the empty spaces with word 4 and 5. So you get "Today(4) the(6) weather(1) today (4) bad (5)"
In this case it is obvious, that the subject was too slow and the Computer filled up the sentence, because of the repeated word (Although when you look at reaction times, they are under 10s).
But in some cases the sentence still makes sense even when the Computer fills up the missing words, for example

Scrambled sentence:
ate(1) I(2) today(3) a(4) Sandwich(5) banana(6)

When the subject was too slow to Response (only clicked at "Today(3) I(2) ate(1)"), the Computer will automatically fill up the empty spaced with word 4(a) and 5(Sandwich), so that the Sentence makes perfectly sense "Today I ate a sandwich"). 
In this case you cannot see, whether the subject was too slow or did it right, since the reaction times are still under 10s.

Is there any other possibility to see in the data file, whether the subject's response was too slow?

Best wishes


 

You can do something like this:
(1) Introduce a new variable (a <values> entry) that reflects whether the participant was too slow or not:

<values>
...
/ CLdigits = 0
/ CLcorrect = 0
/ tooslow = false
</values>

At the start (/ontrialbegin) of the various "end" trials, check whether the last response to the unscrambling trial was "no response", i.e. a value of zero, which means the trial timed out / the subject did not submit a response:

<trial scrambledSentenceC_end>
/ ontrialbegin = [
    if(trial.scrambledSentenceC.response == 0) values.tooslow = true else values.tooslow = false;
]

/ stimulustimes = [0 = frame1, word1, frame2, word2, frame3, word3, frame4, word4, frame5, word5, frame6, word6, number1, number2, number3, number4, number5, number6; 50 = clearscreen]
/ timeout = 50
...
</trial>

<trial scrambledSentenceE_end>
/ ontrialbegin = [
    if(trial.scrambledSentenceE.response == 0) values.tooslow = true else values.tooslow = false;
]

/ stimulustimes = [0 = frame1, word1, frame2, word2, frame3, word3, frame4, word4, frame5, word5, frame6, word6, number1, number2, number3, number4, number5, number6; 50 = clearscreen]
/ timeout = 50
...
</trial>

Finally, log values.tooslow to the data file by adding it to the /columns:

<data>
/ separatefiles = true
/ columns = (build, computer.platform, date, time, subject, group, blockcode, blocknum, trialcode, trialnum,
values.sentenceCondition, values.CL_Condition, values.index, values.scrambledSentence, values.positiveSolution, values.negativeSolution,
values.numericSentence, values.sentenceResponse, values.responseCategory, values.numericResponse, values.tooslow, correct, values.cumulativeRT)
</data>

A logged value of 1 ("true") indicates that the participant was too slow; a logged value of 0 ("false") indicates that the participant completed the unscrambling in time (was not too slow).

Hope this helps.

Orchid
Orchid
Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)Partner Member (784 reputation)
Group: Awaiting Activation
Posts: 6, Visits: 38
Dear Dave,

thank you for your help!

Unfortunately this does not work. In the data file the value of "tooslow" is always set as 1.
I also tried: trial.scrambledSentenceC.response == false, again in the data, there was always a 1
then I tried: trial.scrambledSentenceC.latency >= 10000, but in this case, it always presents a 0 in the data file.
(the same for scrambledSentenceE)

Do you have any other idea?







Dave
Dave
Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)
Group: Administrators
Posts: 13K, Visits: 104K
Orchid - Thursday, July 13, 2017
Dear Dave,

thank you for your help!

Unfortunately this does not work. In the data file the value of "tooslow" is always set as 1.
I also tried: trial.scrambledSentenceC.response == false, again in the data, there was always a 1
then I tried: trial.scrambledSentenceC.latency >= 10000, but in this case, it always presents a 0 in the data file.
(the same for scrambledSentenceE)

Do you have any other idea?







No, I don't have any other idea, but I am certain that what I detailed works.

> Unfortunately this does not work. In the data file the value of "tooslow" is always set as 1.

Maybe you've made some mistake when implementing the approach?

GO

Merge Selected

Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...




Reading This Topic

Explore
Messages
Mentions
Search