Thread Rating:
  • 1 Vote(s) - 3 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Lexical decision with yoked pair stimuli
#1
I am creating a study with priming and a lexical decision task and I have arranged to use yoked pairs of audio primes and text targets. I also have 2 separate lexical decision blocks for 2 separate groups of participants, so that they only hear a subset (half of each of 2 types) of the primes. 

Problem 1: 
I am stuck on how to code the response options. In the tutorial on using yoked pairs, RESPONSES is set to "lexical decision": Is this a response I would have to create or is it built-in? I can't find any information on that in the RESPONSES section of the site. If it is a response I must define, how do you set which targets are YES/WORD and which are NO/NOT WORD without having separate responses for these two options? I know how to create separate responses for WORD and NOT WORD, but I don't see how to use these response options with the design I am working with, where I have 2 blocks for participant groups, but would also need 2 blocks for responses, for each participant group (?). 

I think I would have to figure out a way to create 4 blocks: WORD and NOT WORD targets for each group, but then I would need to keep my primes and targets yoked, while randomizing the order of the yoked trials that are WORD and NOT WORD, while also keeping my 2 participant blocks separate. The NOT WORD trials are fillers and are the same for both participant groups, but the WORD trials are different.

Do you have any advice or guidance, or other relevant tutorials? 
Reply
#2
Hello! FindingFive does not have a "lexical decision" response type. That's just the name of the choice response that was used in the tutorial here: https://news.findingfive.com/2019/07/17/...ed-trials/ You do need to create two choice responses for a lexical decision task-- one where "WORD" is the target and one where "NOT WORD" is the target. However, you do not necessarily need to have separate trial templates for the two responses. You can use response_pairing to pair the appropriate stimulus (test or filler) with the appropriate response (word or not word) within the same trial template: https://help.findingfive.com/api/trial-t...html#basic. Then, it's easy to keep the primes and targets yoked with the alternate_random stimulus pattern just as described in the tutorial.
With this set up, you will still have to create separate trial templates and blocks for your two participant groups, since they see different stimuli. You can see more about group assignment in this tutorial: https://news.findingfive.com/2019/05/06/...-grouping/ 
Please let us know if you have further questions!
Reply
#3
(11-16-2020, 12:57 PM)Thank you! After reading about the response_pairing option, I am still confused as to how to set this up. If I create two responses (WORD target and NOT WORD target), then how do I set which targets match each response? From what it sounds like, if I use the "partitioned" option, it would take the first 1/2 of my stimuli as WORD and the 2nd 1/2 as NOT WORD, is that correct? What do I do about stimuli that have no response (such as instructions etc.)? Does this mean that I have to have the stimuli uploaded in the proper order, or can I set this manually (ie list the stimuli names that should be the first choice response and then the second, etc.)? Also, do I code the response_pairing option in the trial template for my targets, something like this:"practice_primes": { Wrote:   "type": "basic",

  "stimuli": ["pracP01", "pracP02", "pracP03", "pracP04", "pracP05","pracP06", "pracP07", "pracP08", "pracP09", "pracP10"],

  "order":["random"],

  "delay":[1.5]

},



 

"practice_targets": {

  "type": "AFC",

  "stimuli": ["pracT01", "pracT02", "pracT03", "pracT04", "pracT05","pracT06", "pracT07", "pracT08", "pracT09", "pracT10"],

  "responses": ["word", "not word"],
  "response_pairing":["partitioned"],


Will it assume that pracT01-05 are WORD and 06-10 are NOT WORD?


Sorry for so many questions, and thank you for your help!




Lucy - FF Researcher SupportHello! FindingFive does not have a "lexical decision" response type. That's just the name of the choice response that was used in the tutorial here: https://news.findingfive.com/2019/07/17/...ed-trials/ You do need to create two choice responses for a lexical decision task-- one where "WORD" is the target and one where "NOT WORD" is the target. However, you do not necessarily need to have separate trial templates for the two responses. You can use response_pairing to pair the appropriate stimulus (test or filler) with the appropriate response (word or not word) within the same trial template: https://help.findingfive.com/api/trial-t...html#basic. Then, it's easy to keep the primes and targets yoked with the alternate_random stimulus pattern just as described in the tutorial.
With this set up, you will still have to create separate trial templates and blocks for your two participant groups, since they see different stimuli. You can see more about group assignment in this tutorial: https://news.findingfive.com/2019/05/06/...-grouping/ 
Please let us know if you have further questions!
Reply
#4
Hello! See my answers in-line in purple text. 

Thank you! After reading about the response_pairing option, I am still confused as to how to set this up. If I create two responses (WORD target and NOT WORD target), then how do I set which targets match each response? From what it sounds like, if I use the "partitioned" option, it would take the first 1/2 of my stimuli as WORD and the 2nd 1/2 as NOT WORD, is that correct?
Yes, correct. 
What do I do about stimuli that have no response (such as instructions etc.)?
These stimuli should be placed in a separate trial template, where you don't use response pairing. 
Does this mean that I have to have the stimuli uploaded in the proper order, or can I set this manually (ie list the stimuli names that should be the first choice response and then the second, etc.)?
See the documentation: https://help.findingfive.com/api/trial-t...html#basic response pairing goes off of the order of stimuli and responses in your trial template. 
Also, do I code the response_pairing option in the trial template for my targets, something like this:

"practice_primes": { 

  "type": "basic",

  "stimuli": ["pracP01", "pracP02", "pracP03", "pracP04", "pracP05","pracP06", "pracP07", "pracP08", "pracP09", "pracP10"],

  "order":["random"],
"order" is not a top-level property. If you are trying to randomize these stimuli, use a stimulus pattern: https://help.findingfive.com/api/trial-t...s-patterns

  "delay":[1.5]
Don't put 1.5 in brackets, as "delay" takes a numeric value, not a list. 
},



 

"practice_targets": {

  "type": "AFC",

  "stimuli": ["pracT01", "pracT02", "pracT03", "pracT04", "pracT05","pracT06", "pracT07", "pracT08", "pracT09", "pracT10"],

  "responses": ["word", "not word"],
  "response_pairing":["partitioned"],
Don't put "partitioned" in brackets, as "response_pairing" takes a string, not a list. 

Will it assume that pracT01-05 are WORD and 06-10 are NOT WORD?
Correct. 

Sorry for so many questions, and thank you for your help!

You're welcome!!
Reply
#5
Thank you! So far it all seems to be working properly now which is very exciting!   My next (last?) question is how to program in a few breaks. Currently, my main lexical decision blocks (2 blocks for 2 participant groups) have 240 trials each, which is too many. I have created trial templates and a block in the procedures section for breaks, but I am not sure how to get these breaks into the block sequence. Is there a way to code it so that the block called "break" is automatically triggered every n number of trials in each of my lex_dec blocks, or to say that I want 3 breaks evenly spaced, something like that? Right now, my block sequence is defined as:

{"block_sequence": [                   
"practice",
  {"GroupA":["lexdec_blockA Wrote:
,
                      "GroupB":["lexdec_blockB"]}, 
                     "survey"
                    ]}

Thanks again for your patience and help! I am new to Finding Five and relatively new to coding.

---------------------------------------------------------------------------------------------------------------
Lucy - FF Researcher SupportHello! See my answers in-line in purple text. 

Thank you! After reading about the response_pairing option, I am still confused as to how to set this up. If I create two responses (WORD target and NOT WORD target), then how do I set which targets match each response? From what it sounds like, if I use the "partitioned" option, it would take the first 1/2 of my stimuli as WORD and the 2nd 1/2 as NOT WORD, is that correct?
Yes, correct. 
What do I do about stimuli that have no response (such as instructions etc.)?
These stimuli should be placed in a separate trial template, where you don't use response pairing. 
Does this mean that I have to have the stimuli uploaded in the proper order, or can I set this manually (ie list the stimuli names that should be the first choice response and then the second, etc.)?
See the documentation: https://help.findingfive.com/api/trial-t...html#basic response pairing goes off of the order of stimuli and responses in your trial template. 
Also, do I code the response_pairing option in the trial template for my targets, something like this:

"practice_primes": { 

  "type": "basic",

  "stimuli": ["pracP01", "pracP02", "pracP03", "pracP04", "pracP05","pracP06", "pracP07", "pracP08", "pracP09", "pracP10"],

  "order":["random"],
"order" is not a top-level property. If you are trying to randomize these stimuli, use a stimulus pattern: https://help.findingfive.com/api/trial-t...s-patterns

  "delay":[1.5]
Don't put 1.5 in brackets, as "delay" takes a numeric value, not a list. 
},



 

"practice_targets": {

  "type": "AFC",

  "stimuli": ["pracT01", "pracT02", "pracT03", "pracT04", "pracT05","pracT06", "pracT07", "pracT08", "pracT09", "pracT10"],

  "responses": ["word", "not word"],
  "response_pairing":["partitioned"],
Don't put "partitioned" in brackets, as "response_pairing" takes a string, not a list. 

Will it assume that pracT01-05 are WORD and 06-10 are NOT WORD?
Correct. 

Sorry for so many questions, and thank you for your help!

You're welcome!!
Reply
#6
Hi! You can add break trials with "catch_trials" in your procedure code: https://help.findingfive.com/api/procedu...tch_trials
Reply
#7
Thank you!
I have tried doing that, and I am not sure what I've done wrong, as it isn't working. There are no reported issues with the code, but when I run the study no catch-trials appear. I have the following block (and 3 others like it) defined in the procedure with catch trials. Am I defining it in the wrong place?

  {
"lexdec_blockA1": {
      "trial_templates": ["primes_templateA1","targets_templateA1"],
      "pattern": {"order": "alternate_random", "repeat": 1} ,
      "catch_trials":[{"template":"Break", "frequency":10}]
}
Reply
#8
Hi, this code looks correct. What does the trial template for "Break" look like? Perhaps there is a problem there.
Reply
#9
(11-20-2020, 02:10 PM)Ive tested the trial template "Break" thinking that might be the problem, but it works as a regular block if I code it that way. I can leave it as its own block, but then there would be fewer breaks than I would like unless I break down the other blocks more, which I'd rather not do. The code I have is:{  "Break":{ Wrote:     "type":"basic",
    "cover_trials":["Break_start"],
    "end_trials":["Break_end"]
  },

"Break_start":{

    "type":"instruction",
    "stimuli":["Exp_break"],
    "duration":10,
    "countdown_visible":false
  },
 
  "Break_end":{
    "type": "instruction",
    "stimuli":["Exp_break_end"],
    "duration":5
}
}


Lucy - FF Researcher SupportHi, this code looks correct. What does the trial template for "Break" look like? Perhaps there is a problem there.
Reply
#10
This code is a mix of trial template and procedure code and will not work:
"Break":{
"type":"basic",
"cover_trials":["Break_start"],
"end_trials":["Break_end"]
}
Trial templates do not have cover trials and end trials. What is the goal with having "Break_start" and "Break_end"? The most common implementation is to have one trial template defined something like this:
"Break":{
"type":"basic",
"stimuli":["Exp_break"],
"duration":10
}
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)