psychopy.org | Reference | Downloads | Github

Random Numbers In CSV Data Output

OS: Win10
PsychoPy version: 3.2.3
Standard Standalone?: Yes

Problem: Seemingly random numbers appear across hundreds of rows in the data output files during the testing block of my study. This happens fairly rarely (2/33 participants so far) and is not consistent with a particular condition or list in the experiment.

Attached is a copy of both what the CSV output typically looks like (CORRECT_DATAOUTPUT.csv) and a copy of when the random numbers appear (WRONG_DATAOUTPUT.csv).

CORRET_DATAOUTPUT.csv (168.0 KB) WRONG_DATAOUTPUT.csv (544.0 KB)

What did you try to make it work?: I tried looking through the psychopy output script to identify any issues that arise during these rare cases. No error codes are present.

Questions:
(1) Why do these seemingly random numbers appear?

(2) In the CVS data output files, column DJ indicates the participants’ answers in the test block. Can I still use the data from the participants where the random numbers appear (do the random numbers across rows 80-82 mess with the order of the participant answers? All 12 answers are there but the random numbers pushes one answer out of this DJ column it should be in.)?

Thank you in advance for any help and guidance! I can easily share more details of the experiment I created using the builder if needed as well.

Miranda

Hi, the good news is that there don’t appear to be any data quality issues with these files.

What you are seeing is simply the way that your spreadsheet (I’m assuming Excel) is displaying the file. i.e. the values aren’t in the wrong columns, Excel just makes them appear to be shunted over. You can verify this by selecting all the cells in the sheet and formatting their alignment text control setting to be “wrap text”. This will force all columns to be displayed with the same width.

You’ll then be able to see that some cells just contain long lists of values (which now makes them take up a lot more vertical rather than horizontal space). At some point in your experiment, you must end up tracking the mouse coordinates. This records values on every screen refresh, rather than just one value per trial. These multiple values are all packed into a single cell, and it is Excel not wrapping the display of those cells which is giving you the impression that something is wrong.

If you don’t want to be tracking mouse trajectories, then you ned to check the settings in your mouse component’s “Save mouse state” popup menu. “Every frame” will records values continuously, whereas “Final”, say, would just record the very last mouse coordinates and button values.

As above, they aren’t random.

Yes, assuming that there is nothing else strange going on.

As above, it is still in column DJ: setting the “wrap” formatting will show that this is just a display glitch in Excel, and that all data is in the column that it should be.

Hi Michael,

Thank you for your help! I feel a bit foolish since it was such a simple fix, but I am glad there were no issues.

I will make changes in the “Save mouse state” menu to inhibit excess output.

Thank you!!

Miranda