Number of observations in byquestionnaire does not match byvariable (case 6026)

[jsimons]

Hi Jordan,

The first thing to come to mind is that it is something to do with the way the .csv file is getting parsed into Excel. This could be being driven by the participant's response in a free-response type item (e.g., thought listing, essay, fill-in-the-blank), especially if they have put a comma or new line in their response.

Is there a whole extra participant in the byvariable file, or just an additional blank line? If the former, is there anything notable about their responses (or maybe the responses of the participant before them)?

If you right-click the file, you can open it up in Notepad, and see whether the problematic participant is showing a notably different formatting to the others (e.g., extra commas or spreading responses over multiple lines). The file will be a list of numbers separated by commas. It can be a little hard to read but each participant will have the file directory for the experiment listed, which is a useful landmark for orientating yourself.

Hope this helps. Let me know if anything isn't clear.

[jsilberman]

Many thanks for your helpful response, which helped me figure out that the subject missing from the byquestionnaire folder wasn't actually missing, but rather just had data formatted differently from all other subjects.

Rather than giving each subj its own row, in the byquestionnaire folder, ML gave all but one subjects its own row. One of the subjects (the one I initially thought was missing) was added to the row of another subject, after the variables for the 1st subj end. So one of the rows contains 2 subjects and is twice as long as all of the other rows. Very strange. The .csv file that ML produced is attached (see columns starting at DW).

Any idea what could have caused this and how to prevent recurrence?

Thanks,
Jordan

[jsimons]

Hi Jordan,

I'm glad you managed to track down the responses. It seems like all the data has been written correctly, and the issue is just to do with the parsing of the file into .csv format. Usually when I have encountered problems like this, it is due to participants putting odd open-ended responses, but this does not seem like a plausible explanation in the current case.

Everything is correct in the byvariable file, right? Do you get the same issue with the .sav versions of the files?

Thanks,
Joe Simons

[jsilberman]

Yes, I think everything is OK in the byvar directory. I don't use/haven't looked at the .sav files (I just import the .csv files into Stata). Would still be good to determine the cause of the problem if possible to ensure that this glitch isn't causing unidentified data errors and to ensure that I can use byquestionnaire files in the future without having to compare numbers of Ss in the 2 folders. Thanks, Jordan