For each respondent who takes a survey, Discover automatically generates a 24-character unique identifier known as a Record ID. You can create your own identifier variable in the Variable manager. Similar to the Record ID, this new variable must be unique for each respondent. Only one variable can be designated as the unique ID.
Using a unique ID has several benefits:
- It enables respondents to resume the survey if they leave and return, provided they use the same link with the original unique ID value.
- It facilitates identifying respondents during data cleaning processes.
- It allows for tracking respondents across different platforms or sample providers, enabling data to be passed back and matched based on this variable.
While Record IDs and Unique IDs are functionally similar, unique IDs offer the flexibility of using custom values that can be defined before respondents take the survey. This feature allows respondents to continue their survey progress even if they leave and return.
Consider this data table. It is an example of a survey data export that includes three variables:
Record ID |
country |
name |
83f0d5 |
uk |
beverly |
be5bf2 |
uk |
chet |
ef93dc |
australia |
betty |
b4j2ms |
usa |
leon |
Record ID: automatically assigned to every respondent, ensuring each survey record is unique.
Country: variable created by the survey author; it has not been marked as unique, therefore multiple respondents can be assigned the same values (uk was given to two respondents).
Name: variable created by the survey author; it has been marked as unique. Every respondent will have a unique value assigned by the survey author or panel provider.
The first respondent would be given a link such as:
https://surveys.sawtoothsoftware.com/83f0d5?country=uk&name=skyler
This link allows the respondent to enter, exit, and return to the same survey. As survey authors, we can use the value "skyler" to locate their record in the dataset.
For a more in-depth tutorial on survey linking (passing respondents and data in or out of a survey), see our linking in and out of surveys article.