Errors can creep into a data file from many different sources from simple typos to misunderstanding the question being asked. Data processing errors can also come into play when data are sent from many different locations (such as cancer registries) to a central area in similar but not identical formats. Discuss the ways to identify data process errors.

Question

What will be an ideal response?

Answer 1

One way to pick up on this type of error is to spot check a dozen or so records and compare the electronic information to the original document.  Another way to identify data process errors is to pay careful attention to the out of range values.  While it might be easy to pick up large errors (e.g., ages and dates that are out of range), subtle errors may go unnoticed.  Computer assisted data entry has been useful in reducing the number of errors that make their way into a completed data set.  This approach involves programming a computer to verify each entry before it can be saved.  If this is done while the person is being interviewed, conflicting information can be corrected immediately.  If it is done afterwards, the investigator can make decisions as to how to handle conflicting information.