Mixed data types in API export

Question

Hi All, I have a survey which uses a csv file to populate embedded values used in a series of drilldown questions. The survey responses are extracted via a python script with “useLabels : True” set in the download request. Recently the survey was updated to add a new response option. Since that change, the API extract populates several columns with both answer text (STR) and recode values (INT). So I get text values and numerical values--which likely correspond to the csv file which populates the answer option. Is there a reason why our other surveys using the same structure (but have not changed) continue to export text labels without any problem, while the recently changed survey has mixed data types in some columns?Many thanks for your thoughts.

Lpena · Accepted Answer

Hello,The difference in your API data extract—showing mixed data types (both answer text and recode values) for the recently updated survey, versus other unchanged surveys exporting text labels only—is likely due to how Qualtrics handles changes to Drill Down questions, specifically regarding recode values and answer options.Fixed Recode Values: Drill Down questions save recode values in a fixed manner.Disruption from Changes: Adding new response options after data collection can affect how these values export.Editing the Drill Down CSV: Recode values are not editable after collecting data. Changing the Drill Down CSV file (e.g., adding a new row in the middle) will disrupt the recode mapping for options beneath it.Deleted/Replaced Choices: Deleted or replaced answer choices retain their internal IDs in the data export. These IDs may appear as numeric codes when their associated text labels no longer exist or align correctly.Mapping Old and New Choices: Recent changes to the response options can cause the exported data to contain both text labels and numeric recode values as Qualtrics attempts to map old and new choices.Surveys without such changes consistently export text labels because their recode mappings have remained stable.The reason for mixed data types in your updated survey export is tied to the addition of new response options in Drill Down questions and how Qualtrics manages recode values in these scenarios.Avoid Editing Post-Collection: For best practices, Qualtrics documentation advises against editing Drill Down files after collecting data due to the disruption of recode values.New Question Workaround: Consider creating a new Drill Down question with the updated options and using display logic to manage data integrity (i.e., routing new responses to the new question).Verify Documentation: The appearance of numeric IDs for deleted choices is consistent with documentation. Even when exporting data with useLabels: True, deleted or mismatched recodes may still show numeric values.To mitigate this issue, you should verify the integrity of your Drill Down CSV file and consider the recommended workarounds outlined in the Qualtrics documentation: https://www.qualtrics.com/support/survey-platform/survey-module/editing-questions/question-types-guide/specialty-questions/drill-down/If you are still experiencing issues with your API extract or need a deeper review of your account settings and specific survey data, please submit a ticket through your Customer Success Hub for a personalized resolution.➡️ https://support-portal.qualtrics.com/

Lpena · Answer

:D

Sign up

Login to the Community

Scanning file for viruses.

This file cannot be downloaded