Transcription Is Where Errors Are Born

Part of Research Assistant (ASSIST) Training: Copying Without Introducing an Error.

About 1 minutes. Free, in your language, no account needed.

Copying is not a low-skill task

Most data errors in clinical trials are not people inventing values. They are people copying a correct value into the wrong place, or copying it slightly wrong.

A 3 that becomes an 8. Two digits swapped. A value entered against yesterday's visit. A reading put in the row above the right one.

None of these feel like mistakes as you make them, which is why they need a method rather than care alone.

A method that works

Copy from the source, never from another copy. Each hop multiplies the chance of an error and makes it impossible to find where it entered.

Do one participant at a time, start to finish. Batching by field across several people is how values land in the wrong row.

Read back what you entered against the source, out loud if you are alone. Your eye will confirm what it expects to see; your ear will not.

Stop when you are interrupted. Finish the field, then handle the interruption. Resuming mid-field is where a digit goes missing.

Units and decimals

Check the unit every time. A weight in pounds entered into a field expecting kilograms is a huge error that looks like an ordinary number.

Decimal points are worth a deliberate second look. Ten times a dose is not a typo with a small consequence.

If you are not sure what you copied

Go back to the source and check. Do not reason about what it probably was. This takes a minute now and saves a query, a correction and an explanation later.