Please help us help you and fill where relevant:
Your LimeSurvey version: 6.4.12+240311
Own server or LimeSurvey hosting: Own Server
Survey theme/template: FRUITY (customized)
==================
Since our recent update from V3.6 to 6.4.12+240311, we’ve encountered some issues that we hadn’t seen before. While LimeSurvey now correctly creates a new answer option for single-choice questions with “other, namely:” options, which I’ve been eagerly awaiting, the export to SPSS (and .csv) has unfortunately worsened. I hope this is just a matter of my misunderstanding, so let’s look at the specific problems:

Using the “Export response data to SPSS” function (downloading .sys and .dat files):
Problem 1: For multiple-choice questions, only the “1” responses are exported. Zeroes, “N”s, or however they are labeled, are not included. This results in incorrect totals for all multiple-choice questions. Additionally, the variables cannot be used for any calculations since every variable appears to be selected by 100% of respondents. There should be a distinction between respondents who were not eligible for an answer option (missing) and those who were eligible but did not select the option (preferably 0, otherwise "N").
Problem 2: No extra variable is created for the “other” answer options in multiple-choice questions. This issue occurs with both export functions, so it might be expected behavior.

Using the “Export responses” function:

• Export format: SPSS (.sav)
• Export responses as: “Convert Y to: 1” and “Convert N to: 0” (Without conversion, N’s are not exported. You only get Y’s and missing values)
• All other settings: default

Problem 3: Single-choice questions with “Other, namely” answers are recoded to string variables by LimeSurvey, filling the respondent's cell with “-oth-” instead of a number.
Problem 4: Again, no extra variable is created for MC “other” answer options. Since this happens with both export functions, it might be standard behavior.

Is there any way to resolve at least problems 1 or 3? I would really like to export zeroes for multiple-choice questions and utilize the new functionality for “other” answers in single-choice questions. I look forward to hearing your suggestions!

Kind regards,
Pascal

Hi,

we’ve encountered some issues that we hadn’t seen before

As far as I know, all your mentioned "issues" are standard behaviour at least since version 2.06

While LimeSurvey now correctly creates a new answer option for single-choice questions with “other, namely:” options, which I’ve been eagerly awaiting

This option was always available, even in version 2.06
 

And also - for decades - the values are stored like this.
Single
 
Either the code is stored, or the text "-oth-" and the verbatim in column Q1a_other

Multiple
 
In each column (subquestion) "Y" is saved, if selected, else blank.
And there is only  a text column for the verbatim of "Other".
That's it.

No extra variable is created for the “other” answer options in multiple-choice questions.

Maybe it is a pity that there is no specific column also with "Y" or "", but this also is the way LimeSurvey works.
And as you know this you can easily add this to LimeSurvey with a hidden subquestion (Q1_OTH) and an equation like {Q1_OTH=if(!is_empty(Q1_other),"Y","")}

This results in incorrect totals for all multiple-choice questions

Why? In SPSS you have to create a "variable set" for "multiple questions". And here you are asked which value to calculate.Meaning, SPSS has no idea, what a different value means.
 

There should be a distinction between respondents who were not eligible for an answer option (missing) and those who were eligible but did not select the option

But this you do with a filter in SPSS. (or you add an instruction to the syntax file to add your desired values). Especially if you count each column separately without defining a set.

Without conversion, N’s are not exported. You only get Y’s and missing values)

Therefore there is this option (to convert to whatever you like)

Single-choice questions with “Other, namely” answers are recoded to string variables by LimeSurvey, filling the respondent's cell with “-oth-” instead of a number.

Yes, it is.
But, what? Again you may use an equation to change, like {Q1=if(Q1=="-oth-",99,Q1)}
 

To come to an end: I tell you, if LimeSurvey changes the behaviour this way, at once there will be a lot of users who cry "What did you do? Why did you change this?"
There always will be people who are not fully satisfied and would like to have changed this or that. But you can't satisfy anybody.

Joffm

Hi Joffm,

Thank you for the workarounds for my issues. I'll give them a try.

Regarding your question ("Why?") about my statement ("This results in incorrect totals for all multiple-choice questions"): In my case, it indeed leads to incorrect totals. I calculate the percentage of the sample that selected each answer option in a multiple-choice question. To do this accurately, I need to know 1) who selected the answer option, 2) who did not select the answer option, and 3) who did not answer the question (due to a different route, drop-out, etc.). Those who were not eligible to respond to the multiple-choice question (group 3) do not count towards the total N for my calculations. Only those who either selected an answer option or could have but didn’t are included in the N. The percentage is then calculated as P = count(Y) / sum(Y;N) * 100.

I understand that a company can’t please everyone, and I wasn’t expecting them to. But I’m glad there are options for customization.

Kind regards,
Pascal