Proof-reading texts on screen and paper

This study examined the speed and accuracy of proof-reading a text presented on a CRT, relative to performance with print on paper. Two groups of 16 people each proof-read four published texts, roughly 1500 words per text For all readers, half the texts were presented as print on paper and half were presented on a 12 in. CRT screen. The two groups differed in whether the errors found in the screened text were recorded on the screen or on paper. The results suggested that the method of recording errors on the screen was quickly learned, but that both speed and accuracy were impaired when the text was presented on the screen. The implications of this for refereeing electronic journals is discussed.