Missing values in exported data

Survey Solutions user Edith Felix has sent us the following question:

We are working on the cleaning of our Survey Solutions data and wanted to confirm our interpretation of the different missing codes from Survey Solutions. Can you please let us know if the following interpretation is correct and if there is updated documentation on the Survey Solutions website related to missing codes:
The missing codes .a or “##N/A##” (for string variables) are always used when a question was illogically not answered. This is their only use.
The missing codes . or “” (for string variables) are always used when a question was logically skipped. This is their only use.

Survey Solutions recognizes two situations of missingness as described at the following page:
https://docs.mysurvey.solutions/headquarters/export/missing-values/

Edith’s definitions are specific to Stata format. In tab-delimited and SPSS files one will not find the .a extended missing value specific to Stata, but rather will find a -999,999,999 value.

On the surface this is it. But there are some corner cases, such as whether to consider a single whitespace to be an answer to a text question? (no) or what do we mean by an answer to the text list question? Should there be a missing by omission in the 40th column corresponding to the text list (because, say, 40 members’ names could have been entered)? Or should there be missing by logic, because only 5 were entered and there are no more? What do we do with calculated variables A=B+C when B is missing because of the logic and C because of no response? We try to do rational choices with these cases. But your reasoning may differ.

Applying reverse logic is dangerous here. I mean statements like “This is their only use.” Survey Solutions reserves these values (see limits), but it will not actively prevent the interviewer from typing such a value as an answer, nor will it object the value to be a result of the calculated variable (such as s="##"+“N/A”+"##" or d=-333333333*3). So if you do see these values in your output it is our best interpretation (this explains why these values were so particularly chosen.)

In any case if there is a particular situation where the behavior contradicts the intent described in the article online, please provide the details here.

Best, Sergiy

Thank you very much for your response Sergiy!