Differing Data Sheets for the Exact Same Task


I have recently uploaded a task for participants on MTurk to complete. My team and I collected 37 participants’ data using this task. When downloading the data and trying to combine data sheets, we found that in 9 of the 37 participants, the data sheets were different and the columns did not line up. For example, the reaction time column for our first routine was placed in a different column across two datasheets, so it was difficult to conduct analysis. The task was not edited across these participants, they were the exact same. Is there a way to prevent this from happening? Please let me know if I need to provide more information.

Thanks in advance!

This happens when some data only gets saved under certain circumstances. The solution is to ensure that if you have conditional saving of data you save a blank value when you don’t want something saved.

Alternatively, it might be possible to save a blank value for all custom columns in a Begin Routine code component at some point before the first custom data is saved.

Unfortunately, I don’t think this will work if the issue is related to randomisation of the order of your routines.

You might be interested in trying out my new summarise data tool (see my post in the blog category) or using some Python code (to select and resave the data files with the columns you want.

Hi there @aboye , as Wake said, you can use a python script
extract_combine.py (716 Bytes) to select specific columns from your CSVs and collate them into a separate spreadsheet.

The data files here are from the stroop demo and if you save these and the python script above in the same folder and run it through coder view, it would only save the relevant columns
data_000_stroop_2022-03-25_08h08.39.785.csv (3.4 KB)
data_1_stroop_2022-02-01_13h46.45.514.csv (3.5 KB)
data_000_stroop_2022-03-25_08h08.39.785 (1).csv (3.4 KB)

You would only need to replace the name of the Stroop column headers with your own and it should do the trick.

1 Like

Hi. I think I need to specify that across every data sheet, there are the same number of columns, it is just that some data sheets have a different order. There are no columns on these odd 9 data sheets that are not on the regular 28, it is just that they vary in their order on the data sheets.


Is there a reaction time for every trial. If not, then the order of columns will depend on whether the participant responded on the first trial.


Yes, we noticed that for our out-of-order timesheets, they did not respond to the first trial. Is there a way to change this so all of our columns end up in the same order?


Is there an easy way to keep the same order of columns even if they do not respond to the first trial?

Thank you :slight_smile:

Not that I know of, unless you forced a response on the first trial. I often use the same routines for practice and main trials so could have a forced response on the practice trials but end in code on the main trials.