DDI and *.do do not accurately reflect the data

Posting an issue raised previously in support but should probably go here in case someone else comes across it. (Case #: 4402)

I am using Nesstar to compile the responses from the survey and have found it difficult to import the data because the DDI does not include the Interview Key variable.
I can use the *.tab data and copy/paste without the Interview Key, but it would be much nicer to be able to directly load the *.sav files

In addition, the ordering of the data for rosters is mixed up, with the variable (list/multiselect list) that the roster is based on being placed after the roster ID, but in the DDI and the *.do it is placed last, just
before the interview_id and the interview key.

List of names - variable: names
*Roster - occupants
*Age - age
*Gender - gender

DDI indicates it should be:
occupants__id | age | gender | names | interview_id

The data comes out as
occupants__id | names | age | gender | interview_id | interview_key

Hello, Scott,

your software may

  1. read the tab data as a tab-delimited dataset with variable names in the first line.
  2. inquire about the properties of a particular column (variable) in the corresponding DDI data file, e.g. codes of categories for a particular variable.
    If you follow this sequence, the order of the variables in the DDI file doesn’t matter.

The interview__key variable is present in the DDI files, refer e.g. to the demo server:
(requires regular authentication).

For example:

<var ID="V2" name="interview__key" files="F1" intrvl="discrete">
  <location width="255" RecSegNo="1" />
  <labl><![CDATA[Uinque 8-digit long identifier of the interview]]></labl>
  <qstn />
  <varFormat type="character" schema="other" />

Note both interview__key and interview__id are spelled with two underscores (different from your illustration).

If you can demonstrate the issue on a particular questionnaire (real or minimal example) please upload to the demo server and provide links in your message.

Also, you wrote “I am using Nesstar to compile the responses from the survey” what specifically are you doing in Nesstar? (and which one?), please refer me to the Nesstar documentation for this procedure.

Best, Sergiy Radyakin

Hi Sergiy,
I apologise if it was unclear in the example, but it was referring to the roster data file.
Eg a household questionnaire with an occupants roster.
In the occupants roster data file (tab or spss), there is an interview__key column, but this is not present in the DDI nor the *.do

An example can be found here:
and the *.do and tab:

The interview__key is present in the DDI and *.do for the base questionnaire but not for the roster.

The DDI doesn’t contain the interview__status metadata either.

Hello Scott,

thank you for reporting the issue. We will add the metadata for the interview__key variable into the DDI and *.do files for the rosters.

Best, Sergiy Radyakin