Millisecond Forums

Randomize blocks in Corsi Block Tapping Task

https://forums.millisecond.com/Topic9669.aspx

By Richard - 2/11/2013

I would like to randomize blocks in Corsi Block Tapping Task because i dont want the respondent to remember sequences of blocks. Another important thing is  that i dont want blocks to repeat. I dont know how to change source.If  I change / selectionmode  from sequence to random, blocks are repeating. I am new member and i dont have much experience with changing the source. Can you change it for me pls?


<list blockseq>
/ items =
    (8,5,
    6,4,
    4,7,2,
    8,1,5,
    3,4,1,7,
    6,1,5,8,
    5,2,1,8,6,
    4,2,7,3,1,
    3,9,2,4,8,7,
    3,7,8,2,9,4,
    5,9,1,7,4,2,8,
    5,7,9,2,8,4,6,
    5,8,1,9,2,6,4,7,
    5,9,3,6,7,2,4,3,
    5,3,8,7,1,2,4,6,9,
    4,2,6,8,1,7,9,3,5)
/ selectionmode = random
/ resetinterval = 0
</list>

By Dave - 2/11/2013

Richard,


randomizing the sequences defeats the purpose of the task. Have you read the article?


If you want to test subjects repeatedly, you should use different sequences based on pretesting to show equivalence with the original sequences.


Regards,


~Dave

By Richard - 2/11/2013

What article do you mean? Can you send me link?


I will try to explain in more details. I am doing research about sleep deprivation, and i want to figure out, how is short-term memory changed during 48 hours of sleep deprived person. Accordint to that I have to make my measuring every 2 or 3 hours and therefore i have to change sequences of blocks to prevent person  remember them.

By Dave - 2/11/2013

What article do you mean?


The reference is at the top of the script.


My point about random sequences is this: There is no guarantee whatsoever that any two or more random sequences of the same length are equivalent. For example, a sequence consisting of mostly adjacent blocks being highlighted may be vastly easier to memorize than a sequence that jumps around the grid a lot. If you want to compare results, you need to ensure the sequences are for all practical intents and purposes equally hard. Otherwise you've got a potential confound to deal with. Plus, any standardization goes out the window, because only the exact sequences included in the script (see the article) have been assessed.


Regardless, if you absolutely want to have random sequences, simply change <counter blockseq> to


<counter blockseq>
/ select = noreplace(1-9)
</counter>


and


/ ontrialend = [if(trial.recallseq.response=="done") values.totalscore=values.blockspan*values.ncorrecttotal]


to


/ ontrialend = [if(trial.recallseq.response=="done") {values.totalscore=values.blockspan*values.ncorrecttotal;
    reset(counter.blockseq)}]


in <trial recallseq>.


Finally, please work through the existing script until you fully understand how it works prior to making any modifications. You'll want to first complete the various tutorials included in the Inquisit documentation to familiarize yourself with the basic syntax constructs.


Regards,


~Dave

By Richard - 2/11/2013

thank you very much for your prompt answer and for your advice, you helped me a lot and  i really appreciate it.