Data entry errors are common in clinical research, especially when workflows are busy, forms are unclear, source documents are incomplete, or users are insufficiently trained. Understanding common error types helps data managers design prevention and detection strategies.
Transcription errors occur when a value is copied incorrectly from a source document into the database. A weight of 12.5maybe entered as125. A date of 02-May maybe entered as 02 Mar. A participant ID may be mistyped. These errors are more common in paper-first workflows and can be reduced through validation rules, doublechecks, barcode scanning where available, and careful source verification.
Unit errors occur when values are entered using the wrong unit. Temperature may be entered in Fahrenheit instead of Celsius. Weight may be recorded in pounds instead of kilograms.
Height may be entered in meters rather than centimeters. Unit errors are often preventable through clear field labels, validation ranges, completion guidance, and training.
Record mix-ups occur when data are entered into the wrong participant record. This can happen when participants have similar names, when study IDs are misread, or when users keep multiple records open. Record mix-ups are serious because they may corrupt multiple participant histories. Prevention requires clear participant identifiers, careful verification before entry, and procedures that discourage copying between records.
Missing data may occur because a procedure was not done, a value was not documented, a field was skipped, or a user did not understand the form. The database should distinguish these situations where possible. A blank field alone rarely tells the full story. Coded reasons such as not done, refused, unknown, not applicable, or unable to assess may be necessary.
Copy-paste errors occur when users copy information from one record, form, or visit into another without verifying accuracy. This may save time but can duplicate incorrect information or create false consistency. Copy-paste should be discouraged for participant-specific data unless the workflow explicitly supports it and includes verification.
Late entry can also affect quality. If data are entered weeks after a visit, source documents may be harder to locate, staff may not remember details, and queries may be harder to resolve. Data entry timelines should be defined and monitored. For example, a study may require entry within 72 hours of a visit and query response within seven days.