Validation rules are an essential mechanism of improving the quality of data in CAPI surveys. However, not in all situations a check may point to a definitive problem with the data. In some cases the check may only request attention of the interviewer. In this case we are talking about warnings, rather than errors.
Consider, for example, the following validation rules:
In this case we are specifying that the wage must be positive, otherwise we should show an error to the interviewer. However, we are less sure about how high the possible wages are, so we will show a warning for the interviewer if it is above 100.0 USD per hour. To achieve this, we set the “is warning” flag in the Designer for this validation rule (first).
As with errors, several warnings may appear simultaneously as a result of the data check.
The warnings are rendered differently from errors (appear in orange color rather than red) and not counted in the interview errors counter. They are also not recorded in paradata.