Why are four eyes better than two? Effects of collaboration on the detection of errors in proofreading

Abstract

We investigated why collaboration has positive effects on performance in detecting errors. Forty-eight undergraduate students worked on a proofreading task individually or in pairs. In the first group, 16 subjects performed the task individually (the individual group). In a second group, 16 pairs of subjects performed the task in dyads, discussing the errors detected (the collaborating group). The results show that the individual group detected significantly more "contextual errors," which are found by tracking the context of a passage, than "surface errors," which are based on the meaning of a word. In contrast, the collaborating group detected significantly more surface errors than contextual errors. The surface error detection rate of the dyads was high enough that it could not be explained by a simple interpretation such as two additional eyes found errors missed by the first two eyes. The results indicate that proofreading in dyads is beneficial because it enables the subjects to free themselves from a default style of reading. Moreover, for the collaborating group, the rate of error detection tended to increase in the latter half of the work, suggesting that interaction between the two individuals gave them an opportunity to learn about the task by cognitive restructuring.