Average reaction time

Lorenzo · February 2021

Hi to everyone!

In the cotext of the Navon Task experiment, where you have to detect the local and global stimuli in random letters, i'm trying to put at the end of the experiment, the average reaction time per stimuli category, like in that case, local or global average reaction time. On that line, how could i collect those data already calculated in average reaction time for each subject?. Thanks

lvanderlinden · February 2021

Hi Lorenzo,

Are you running your experiment online or locally? And is it possible to attach the experiment here? I'm sure we'll find a way to calculate the averages per category! :)

Cheers,

Lotje

Lorenzo · February 2021

Hi Lotje, for now i'm running it locally but i will have to export it on Jatos. Sure i can, i can send you a link maybe?

lvanderlinden · February 2021

Hi @Lorenzo ,

It sounds like we need a little bit of inline scripting to calculate the separate averages, and since you will be using OSWeb, the programming language that we are going to use in JavaScript. Could you upload your experiment by clicking the paperclip icon? Otherwise, a link should also work.

Cheers,

Lotje

Lorenzo · March 2021

navonp.osexp

Hi Lotte,

I've attached the experiment here. Thanks for your help, how we are going to do this?.

Cheers,

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

First of, I simplified the structure of your experiment by:

Adding a column to the loop table that indicates the PNG file corresponding to the current trial

Use the square-bracket notation in a sketchpad item to determine which PNG file to show
By doing so, you don't need to make separate sketchpads for each trial

Regarding the accuracy per condition, you could do the following:

Add another column to the loop item indicating the current condition. Please check whether I did this correctly

Append three inline_javascript items to the overview area (as in the screenshot above)
In the first inline_javascript item, put the following script in the Prepare tab to initialize the counters for the number of correct responses per category:

// Initiate counters for the two different categories:
vars.count_correct_global = 0
vars.count_correct_local = 0

In the second inline_javascript, update the counters depending on the current condition and participants' accuracy. This time, place the script in the Run phase:

// If response was correct:
if (vars.correct === 1){
    // If mapping is local:
    if (vars.mapping === "local"){
        // Increment local counter with 1
        vars.count_correct_local = vars.count_correct_local + 1
    }
    // If mapping is global:
    if (vars.mapping === "global"){
        // Increment global counter with 1
        vars.count_correct_global = vars.count_correct_global + 1
    }
}

console.log("correct local = " + vars.count_correct_local)
console.log("correct global = " + vars.count_correct_global)

In the last inline_script item, divide the number of correct trials per condition by the total number of trials in this category like so (again in the Run tab):

// Number of local trials
// NOTE: change this value if the number of trials 
// or the number of repetitions changes:
vars.n_local = 12 // 4 trials * 3 repetitions
vars.n_global = 12 // 4 trials * 3 repetitions

// Calculate the average in %
vars.avg_correct_local = (vars.count_correct_local / vars.n_local) * 100
vars.avg_correct_global = (vars.count_correct_global / vars.n_global) * 100

I attached the modified script. Hope this helps. :)

Cheers,

Lotje

6973.osexp

Lorenzo · March 2021

Many thanks Lotje!

I was trying to download the file that you've attached me back here.. but i don't why it opens a new experiment page.

Lorenzo · March 2021

Hi Lotje,

I was able to run the experiment with the new script you sent me. It works Great, but the Problem is that in the end, for what concerns the global and local average reaction time, it show only the percentage about the right or wrong answers in the 2 domains (global and local), not the average about the reaction time. How can we fix this?. I promise i'll Buy you a bar XD.

Cheers,

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

Oops, sorry! Here you go!

Cheers,

Lotje

6973.osexp

lvanderlinden · March 2021

Thanks so much, @Lorenzo ! ☕️

☕️

Lorenzo · March 2021

Hi Lotje,

It's me again! i'm trying to download the experiment on Jatos, i saved it both in zip and osx file and it worked. The Problem is that when i try to run it from Jatos, when i click Ok to start, an error message occur. I made a screenshot of that error, i attach it here.

Schermata 2021-03-05 alle 22.44.26.png

Cheers,

Lorenzo

eduard · March 2021

Hi Lorenzo,

Could you open the console of your browser when running the experiment and report the error message that is presented there? You can open the console on firefox by pressing ctrl+shift+c. (For safari, follow these instructions)

That error message will be a little more informative that the one in the screenshot you shared.

Thanks,

Eduard

Lorenzo · March 2021

Schermata 2021-03-08 alle 23.23.06.png

Schermata 2021-03-08 alle 23.28.23.png

Hi Eduard,

thanks for managing my request. I upload here 2 screenshot about the console during the experiment.

Cheers

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

I think the problem comes from the keyboard_response item where the allowed responses are defined. The value "None" is not recognized/understood, so you could simply remove it (and set the allowed responses to "b;n". Does that solve the issue?

Cheers,

Lotje

Lorenzo · March 2021

Hi Lotje,

Yes i tried to do that, but now when i upload the experiment on JATOS it says "study invalid"..don't know why exactly.

6973-2N.osexp

Lorenzo · March 2021

Hi Lotje,

I realized that the compatibility check is failing and is giving "item_form_text_display is not supported"

"item_form_text_display is not supported"

How to fix that?.

Cheers,

Lorenzo

Lorenzo · March 2021

Hi @Lotje,

I don't know why finally i uploaded the experiment on JATOS, and it works! but...at the end of the experiment it doesn't show the average response time about local and global stimuli, and when i download the results i can't convert them in CSV or excel, because they are just text i Guess.

lvanderlinden · March 2021

Hi @Lorenzo ,

I realized that the compatibility check is failing and is giving "item_form_text_display is not supported"

Form items are indeed not supported in OSWeb.

https://osdoc.cogsci.nl/3.3/manual/osweb/osweb/#supported-functionality

Did you end up using sketchpad items for the instructions and goodbye message?

But...at the end of the experiment it doesn't show the average response time about local and global stimuli

I tried to run the most recent experimental file that you uploaded here (called "6973-2N.osexp") and can't reproduce this. For me, it correctly shows the average RT. To avoid confusion, could you upload your most recent experimental file, if the feedback item at the end of the experiment is indeed still not shown?

When I download the results I can't convert them in CSV or excel, because they are just text I guess

I just noticed that your experiment does not contain a logger item. You should append one to the end of your trial_sequence like so. Otherwise, no output file is created.

Hope this already helps you a bit.

Cheers,

Lotje

Lorenzo · March 2021

Hi Lotje,

Sorry for my late response i had lesson this afternoon. Well, i've added the logger as you told me and now i can directly acces to the datas from OS. The experiment work well also if i run it from JATOS, but i can't convert the jatos results into a cvs/xlsx file through OS. When i search for the JATOS result file i downloaded from it, i can't select it to convert it. It remains as a text file that is really hard to read.

I was guessing, is there a way to get in the results files directly the two reaction times for global and local? like in two different columns?.

Funge.osexp

i upload here the experimental file i uploaded onto JATOS. (it would be amazing to set the experiment to be runnable also from a smartphone..in that case we should have to assign to b & n press for example, the down left Side for b and the down right Side for n, correct?).

Cheers,

Lorenzo

eduard · March 2021

Hi Lorenzo,

I have never used JATOS to run a study, so I don't know how the output looks like and how to fix it. Based on your description, I also don't know why Opensesame can't convert it. Could you share an output file, so I could try to convert it myself?

in that case we should have to assign to b & n press for example, the down left Side for b and the down right Side for n, correct?).

On a smartphone, you probably don't want to keyboard responses, but touch responses instead. For that purpose I think you could use a mouse response and two "buttons" on the screen. Lotje made a great tutorial on that: https://www.youtube.com/watch?v=-L9pzw4tpx8

Hope this helps,

Eduard

Lorenzo · March 2021

Hi @eduard,

Nope, i can't share the output files because those are document file, and the forum doesn't support this kind of format. The fact is that if i run the experiment from OS, i'm able to download and upload in excel the results, that will appear correctly divided per column. When i download the experiment results from JATOS, they are saved as documents and if try to convert the results, through the tool in OSWeb called convert JATOS results to csv/xlsx, i can't select the documents about the results.

I tried also to create a zip file of the results and in this way a can at least select it from OSweb, but when i click to convert it OS give me this message:

In that way i can't process the datas, they looks like exactly as a broken json file.

I hope it is not a Problem of MacOS.

lvanderlinden · March 2021

Hi @Lorenzo ,

Is there a way to get in the results files directly the two reaction times for global and local?

Yes, perhaps the following is easiest: Replace the inline_javascript that calculates the averages to the trial_sequence and make sure it appears right before the logger item:

This will add additional columns to the output file for the variables "avg_RT_local" and "avg_RT_global". Make sure to only interpret the very last rows of these columns, because these will include the final calculations (after all 48 trials):

I can't share the output files because those are document file, and the forum doesn't support this kind of format.

Could you perhaps zip it, then? Or email it to me: l.vanderlinden@cogsci.nl

And one final question: In the most recent OpenSesame file that you uploaded here, the form_text items are still there. These are not supported by OSWeb. Did you replace them by sketchpad items?

Cheers,

Lotje

lvanderlinden · March 2021

Hi @Lorenzo ,

Does it help if you manually add the extension ".txt" to your output file? For example, rename "jatos_results_20210310162303" (the one you sent me) to "jatos_results_20210310162303.txt" and try converting again.

You should not try to convert files with the extension ".zip". Then you indeed get the "utf-8" message.

Let me know if this helps.

Cheers,

Lotje

Lorenzo · March 2021

Hi Lotje,

Yes if i insert now the .txt it works, meaning that OS can convert the jatos results file..but and i don't know why, in the excel file i can see only one string with the variables name without actually the numerical datas of them. I'm going to try another time, as i will repeat the running though jatos and download the rispective results.

What you were meaning when you asked me that:

"In the most recent OpenSesame file that you uploaded here, the form_text items are still there. These are not supported by OSWeb. Did you replace them by sketchpad items?" ?.

Another thing Lotje, if i do modifications on the source experiment that i have already uploaded in a JATOS study, these modifications will be valid also on it or, i will have to create a new study with the modified one? Like in that case the fact that i've put the new logger after the inline javascript about the averages?.

Many thanks,

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

But and I don't know why, in the excel file I can see only one string with the variables name without actually the numerical datas of them.

The example output file that you sent me indeed only contained some variable names. No actual responses were recorded (yet).

In the most recent OpenSesame file that you uploaded here, the form_text items are still there. These are not supported by OSWeb. Did you replace them by sketchpad items?

Sorry for the misunderstanding. I found out that text_forms are (minimally) supported online after all (although not properly tested yet). If it works for your experiment, then perhaps it is fine and you can leave it like this.

Another thing Lotje, if i do modifications on the source experiment that i have already uploaded in a JATOS study, these modifications will be valid also on it

No, you have to export the modified OpenSesame script to a JATOS script, and import it as a new study on JATOS after every change you made. Perhaps this will also solve the issue with the almost-empty output file?

Good luck!

Cheers,

Lotje

Lorenzo · March 2021

Hi Lotje,

I Think i understand now when you was saying to replace the form_text items with sketchpad items. Infact i Think that the reason for this error:

it's exactly due to that, right?. Well, sorry but how can i do that?.

Cheers,

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

No, this error message tells you that OpenSeame could not find the PNG file called "6.png". Could you check whether this image is still in the filepool?

Cheers,

Lotje

Lorenzo · March 2021

Hi Lotje,

It's strange because the feedback response file disappeared also! the filepool you mean the one in the experimental loop?.

Funge.osexp

Cheers,

Lorenzo

lvanderlinden · March 2021

Hi @Lorenzo ,

For some reason all png files were gone from your file pool. (For what the file pool is, see here: https://osdoc.cogsci.nl/3.3/tutorials/beginner/#step-4-add-images-and-sound-files-to-the-file-pool.)

I put them back, and also moved the feedback item and logger (as I suggested earlier) and replaced the form_items by sketchpad items. Could you try uploading the new version to OSWeb, do a test run and look at the output file to see if all variables are there?

Cheers,

Lotje

Funge_LvdL.osexp

Howdy, Stranger!

Categories

Average reaction time

Comments

Howdy, Stranger!

Quick Links

Categories

Average reaction time

Comments