2x2 grid layout feature

[jwright8]

Recently I discussed with FindingFive the possibility of using a 2x2 grid for responses in an incidental learning paradigm. This would enable mouse tracking across four quadrants of the screen, which would also permit a confusion matrix analysis. FindingFive added this possibility via "layout": [2,2]. In that discussion there was a mention of a demo containing this feature. Can you share that demo with us? Thank you.

[Ting]

That demo study is actually one of my personal test studies for FF, so it requires a specific setup (which is not available on findingfive.com) for it to work. Is there a specific question that we can answer? We are more than happy to do that.

In case you haven't discovered it already, the documentation on "layout" is already up here (https://help.findingfive.com/api/responses.html#choice), and we have a very detailed mousetracking tutorial here: https://news.findingfive.com/2020/07/17/...y-example/. The "confusion matrix analysis" you referred to is maybe what we call "clickable images"? If so, how to use them can also be found in the documentation of the choice response.

[jwright8]

[quote pid="16" dateline="1604104358"]
(I will have several questions as I work through this, so I am not sure if I should keep them all under this thread. Please let me know if there is a better method.)

With the incidental paradigm I am measuring RT. So, I want them to be correct every time, meaning that I want them to only be able to choose one choice out of the four choices. The choice I want them to choose will have a target on it. I do not want them to be able to advance if they click on the wrong target. Using choice, can we do this? In the help file here (https://help.findingfive.com/api/responses.html#choice) I do not see this option. Thank you.
[/quote]

[Ting]

Posting under the same thread is absolutely fine!

I think we can add a "target_match" feature that is already available for the text response, which makes it impossible for participants to advance unless their response matches the target. How does that sound?

Thanks for this suggestion. It's actually really great!

[jwright8]

[quote pid="20" dateline="1604440996"]
Thank you. That should permit studies that only use mouse tracking and response time as assessments to function. I look forward to hearing when that is ready.
[/quote]

---

I am attempting to create the incidental paradigm using the example from the mouse tracking tutorial. However, it is not working. Can you help troubleshoot code here? If so, here is my code:

The stimuli in the trial template are audio files and the stimuli in the response are images.

Responses (an example response)
____________________

name: X_top_left

{
"type": "choice",
"choices": ["%s:Image-Box_With_X", "%s:Image-Box", "%s:Image-Box", "%s:Image-Box"],
"target": ["%s:Image-Box_With_X"],
"layout": [2,2],
"locations": "fixed",
"instruction": ""
}
____________________
Trial Template

{
"Instr": {
"type": "instruction",
"stimuli": ["TrainingInst1"]
},

"Trials": {
"type": "AFC",
"stimuli": ["Min_ma_low_01","Min_ma_low_02","Min_ma_high_01", "Min_ma_high_02", "Min_ma_fall_01", "Min_ma_fall_02", "Min_ma_rise_01", "Min_ma_rise_02"],
"delay": 1,
"responses": [["X_bottom_right","Mousetracking"],
["X_bottom_right","Mousetracking"],
["X_top_left","Mousetracking"],
["X_top_left","Mousetracking"],
["X_top_right","Mousetracking"],
["X_top_right","Mousetracking"],
["X_bottom_left","Mousetracking"],
["X_bottom_left","Mousetracking"]]
},

"ResetTemp": {
"type": "mouse_reset",
"stimuli": [{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]},
{"which": ["MouseInstructions"], "location": [5]}],
"responses": ["MouseReset"]
}
}

_______________
Procedure

{
"type": "blocking",
"blocks": {
"TestBlock": {
"trial_templates": ["ResetTemp", "Trials"],
"pattern": {"order": "alternate_random", "repeat": 1},
"cover_trials": ["Instr"]
}
},
"block_sequence": ["TestBlock"]
}

[Ting]

Everything *looks* quite alright. Are there any specific error messages? There may be a trivial typo somewhere that's hard to spot this way.

[jwright8]

Thank you. When I looked again, I saw that one of my responses had a double bracket [[]], like this:
{
"type": "choice",
"choices": [["%s:Image-Box_With_X", "%s:Image-Box", "%s:Image-Box", "%s:Image-Box"]],
"target": ["%s:Image-Box_With_X"],
"layout": [2,2],
"locations": "fixed",
"instruction": ""
}

I removed the extract pair of brackets and then I received the error:
Duplicate choices are found in the choice response X_top_left!
So, it seems that we cannot have the same stimuli choice occur more than once in our responses. Is that correct? If so, that is not a big deal. I can just upload the same image but with three different names, I think.

[Lucy - FF Researcher Support]

Hi, that is intended behavior. The same choice should not occur multiple times in the same response. In your setup, it will be best to create three different stimuli (Image-Box-A, Image-Box-B, Image-Box-C) that have the same image. That should solve your error!

[jwright8]

With that change it is basically working now. Although, I may have more questions coming up.

[Ting]

Posting under the same thread is absolutely fine!

I think we can add a "target_match" feature that is already available for the text response, which makes it impossible for participants to advance unless their response matches the target. How does that sound?

Thanks for this suggestion. It's actually really great!

This feature will be available on the testing server tomorrow (11/10/2020)! All you need to do is add:

Code:

"target_match": true

to your choice response!