Dear forum,

I have an array(numbers) question with restrictions on the numerical fields
Minimum value: - 1
Step value: 1
Maximum value: 3
(i.e. including 0)

The results are different between my statistical software (stata via csv-import or alternatively xml-import) and the LS-statistics tool. I have already found out via the manual that the array(numbers) field is numerical and that the string {no answer} will be stored as NULL.

My question: is there a way to control that NULL will not be exported as the value 0?


Background: I have different frequencies for the answer "0" in the LS-tool and the exported data matrices but I cannot see that the mistake is systematic because the exported data has both zeros (0) and missing values (not answered/not displayed) but the distribution between the zeros and missing values is different in the exported data and LS-statistics.

If this cannot solved in the export, is the coding in the LS-statistics tool reliable?

I work with LS-Version 3.15.9.

Thank you in advance!

Cheers,
Christine

We export null value as an empty string in CSV

Did you use STATA export plugin ?

Thank you!
So, just to spell it out and know that I understood correctly: the csv shows the correct matrix, not the LS-statistics tool. Correct?

I used the Stata-plugin but I had some troubles with it or more generally, troubles with exporting the data (also in csv).
Survey respondents used line breaks in their free-text answers. LS did not export this as a continued string but started a new line. The dataset is rather large and formatting it by hand in wordpad was not an option. But Stata had a solution for that.

I don't know if LimeSurvey stat have an issue : need a simple LSS showing this (and best in bugs.limesurvey.org/my_view_page.php

About CSV : in a real CSV newline can be used.
About STATA export : if export with new line are not read in STATA : must report the issue.