Stacked data can be imported to Dapresy Pro which makes it easier to, for instance, handle brand tracker questionnaires where the same question, like brand statements or campaign follow up questions, are present multiple times. This occurs when the same respondent evaluates multiple brands or different campaigns. As an example a brand tracker which contains 20 statements for 100 brands usually is represented as 2000 variables (20x100) in the data file which makes it complicated to report on. The solution for creating a friendlier data format for easier reporting is to pull out these “looped” variables into a separate stacked data file which results in a, sometimes vastly, reduced number of variables.
So instead of only importing one data file containing all data for all respondents the data can now be split up into multiple files (a “normal” file + one or more “stacked” files) without losing any filter capabilities across variables located in different imported data sets. The system will automatically connect these files via the Respondent ID for achieving the cross-file calculation/filtering.
The images below show an example of how the data structure shall look like when importing stacked data to Dapresy Pro. The table below shows how the original data structure can look like when it is exported from the survey tool. As shown, the file contains some background variables and a general awareness question and then a set of looped statements, Statement 1-3 is asked for each car brand.
The tables below shows how a more report friendly data structure can look like which now can be imported to Dapresy Pro. The data has been split up in two files, the “normal” file and a stacked file for all the statements. In the stacked statement file you see that each respondent is present in multiple rows as the respondents answered the statements for multiple brands.
One using Stacked function it's mandatory to load the unstacked data first (flat file, left example in the picture above).
- The normal data file will follow the same formatting rules as always
- The stacked file will only contain the stacked (looped) variables as the system will use response date, weight and any other data from the normal file when reporting on the stacked file. Due to the “stacked” format the Respondent ID column contains duplicates which usually not allowed but when importing a stacked file it is required as this information is needed as the mapping between the files are be based on the respondent IDs.
As the calculations are done across the data sets the same variable cannot be imported in both the stacked file and in the flat file as conflicts in base sizes etc. will occur during result calculation. So, if a variable has been imported in a normal file you will not be able to import the same variable in a stacked file and vice versa. This validation is made during the file upload.
- stacked data cannot be exported
- stacked data cannot be used in an hierarchical filter
- weights can only be applied on the respondent level, not on a stacked level
- a compute variable cannot use stacked variables from 2 different sources*.
- the data cannot be filtered simultaneously by variables from 2 different sources*.
* Sources; you can upload multiple stacked files in Dapresy via:
- new respondents in a stacked file, in this case the questions are the same across the stacked files and they belong to the same source (Example below Stacked Source A loaded for month May and June)
- a new stacked data set with other questions. In this case you loaded a new source (Example below Stacked source A versus Stacked Source B).