Yes, but they are wrong.
I’ve checked the data and identified two possibilities regarding the issue:
1, Some keypresses might be missing in every keyboard online in running mode even though I used ‘constant’.
2, The key_resp_press.keys column might actually be recording key_resp_press_tr.keys.
Of those two options I think that only the first is possible.
Do your participants press the keys in quick succession or hold them down? I think that your missing keypresses are being recorded but by increasing the duration of the recorded key presses.
Here I have highlighted cells where the RT of a key press is less than the previous one (and unusual first key presses)
These key presses always occur when there is a switch from q to l or back again.
and correspond to longer durations
I am not knowledgeable about the inner workings of recording key releases, but I know it is a relatively new feature online and there is different behaviour in PsychoPy when you hold down a key online compared with locally. Why are you recording keys on release rather than press? Does it make sense for periods of your data to be recorded as q and l together?
I haven’t seen any pilot data but I suspect that pilot behaviour is different from participant behaviour (fewer frequent presses/holding down).
Yes, they sometimes need to hold them down. I understand the point. Apart from the columns, all the keypress records are listed in a disorganized manner. Unlike the piloting mode, the numbers can become quite jumbled. 
Still, I’m also quite certain that some keys are missing. Is it possible to recover them from the log file?
If you open the log file and filter by DATA in the second column then you should be able to identify all the key presses individually. I’ve added times relative to the space key press in the if_ready routine in this screenshot. I can’t see any evidence in the log file of key releases being recorded so I suspect that keys presses within a certain time frame are counted as the same press. I recommend that you analyse the log file if you want all the key presses separately.
In context
Here’s the actual start of the trial
and the end
1 Like
I will review the contents of the log file again. Thank you once more for your patient response! 
I’ve reviewed all the files and noticed that the ‘q’ keypress data is missing for the last period of the trial, approximately from 300 seconds to 480 seconds. These data points are not shown in the log file but became evident during my personal test.
Could you please advise on how we might track or recover these data?
I’m sorry but if the data isn’t the the data file and isn’t in the log file then I don’t think it was saved.
Thx Wake