This article describes how to rebase Binary - Multi, and Binary - Grid variable sets that do not have missing values recorded for respondents who skipped the question in the questionnaire, leading to a NET that is less than 100%. Any respondents who do not have at least one selection in the question will be removed from the base and so the NET will be 100%. This example shows a SUMMARY table of an Unaided awareness variable set. As indicated by the NET, some respondents have not given any answers, but they have also not been recorded as missing data.
If those respondents skipped the question in the questionnaire then we may wish to rebase the to only respondents with at least one answer. The table to the right shows the result of rebasing using this QScript.
Requirements
- A data file that contains multiple response variable sets.
- The QScript is intended to be used for multiple response data that has no missing data code. This is to correct data files that have been created poorly.
- OPTIONAL: It is usually better to ask the data provider to make sure that respondents who skip a question are marked as missing, according to the recommended specifications for SPSS data files.
Method
- Select the variable(s) set in the Data Sets tree.
- Next go to Anything > Data > Variables > New > Ready-Made New Variable(s) > Rebase Multiple Response Data in Variable(s) to NET.
- You will be prompted as to whether to or not you want to hide the original question, click Yes or No depending on your requirements.
- A new variable will be created that is labeled rebased to NET and this will be based on respondents who gave at least one answer.
If the script is applied to questions that already have missing data then respondents with no responses will still be removed from the base, but the base will not necessarily be the same as the original NET.
Next
How to Rebase One Question based on Another Question