Without valid data, the integrity of your outcomes is in question. To ensure that your systems contain accurate data, you must have mechanisms in place to manage data integrity. This should start with the software you select to training staff who work with data. Running reports and completing data audits will contribute to managing the integrity of your data.
Purchase software programs with coding that identifies data issues as they occur. Create program rules that allow auto-correction and “if then” coding so accurate data is input in the appropriate fields. For example, if someone answers “yes” to a question, another field opens up to enter additional required data. Additionally, software fields should be alphanumeric-appropriate and contain size limitations. For example, a date of birth field would be numeric and allow for eight digits.
Program your systems to kick out “bad” data that is missing or formatted incorrectly when loading, entering or merging data into the system. Run error reports that show data omitted from the system and complete a “cleanup” of any records that need manual intervention to fix.
Train staff on the importance of entering data correctly. Explain the consequences of data inaccuracy and instill a sense of urgency regarding quality data. Show case scenarios of how bad data can cost the company time and money, and decrease customer satisfaction.
Complete data integrity audits of your data-entry staff. Review random records to evaluate performance based upon their usage of the system and the accuracy of the data they enter. If they make certain critical errors or meets satisfactory levels on random audits, give them a performance warning and complete additional audits to monitor their progress. Employees who cannot meet or exceed satisfaction measures can be subject to termination.
Review system reports periodically to ensure they make sense. Know how to spot errors or when data that does not make sense in the context of what you are reviewing. For instance, if you see a list of employees that all have ages over 100, you would question the accuracy of the data.