Hi, it's common to denote causes of nonresponse in survey analysis using specific codes. SPSS uses negative numbers (-9, -8, -7...) and Stata uses variants of its missing value code, "." for user-defined missing values, ".a", ".b", ".c", ....
Could Decipher support one of these standards?
With questions in checkbox format, adding an exclusive option for "none" or an open-end for "other" is very handy. Yet, for numeric and text questions, there's no comparable way to extend the responses to reflect, for example, "doesn't apply", "refused", "don't know". It should be possible to distinguish whether the question was asked or not --- which can be deduced from conditions variables, but the onus is on the programmer to know that the condition applies to the column.
Thanks, have a good day.
Please sign in to leave a comment.