Changes were made to both the Wave 1 and Wave 2 datasets during the production of Wave 3. Wave 1 and Wave 2 cleaning of the data has been on-going since the previous release of the data. We paid special attention to key variables, such date of birth, gender, population group and highest education, as well as the birth history section and questions about parent vital status (are they alive or dead) in Wave 3. Where there was a discrepancy across waves we did callbacks to confirm what the correct answer is. Were we have learned that either Wave 1 or Wave 2 was incorrect we updated the data in those datasets to reflect the correct information. In some cases this has resulted in questionnaire type changes due to changes in dates of birth. In other cases we have had to update the CSM/TSM tag for new children as we learned either their real date of birth or where we learned that their mother was interviewed as a TSM in Wave 2.
During Wave 3 fieldwork we also learned about more cases were we had duplicate households from Wave 1 or where people were interviewed in more than one household and incorrectly presented as two separate records. The second scenario was especially prevalent in Wave 2. In extreme cases we also learned of people being incorrectly recorded as deceased in Wave 2 when they are still alive.
Back to Wave 2 Overview Documentation.