GREVIS (Groupe de recherche en révision humaine) aimed to set up an accelerated method of revising while improving the quality of the operation. The project had a three fold objective: to strengthen the place of revision in the field of translation studies, to increase revisers' satisfaction and to help the translation industry. The hypothesis of this study was that monolingual revision was just as effective as bilingual revision, and could be done at a lower cost, because it is less time-consuming. However, the results of the study disproved this hypothesis: bilingual revision was more than twice as effective as monolingual revision. The 19,407-word corpus comprised translations from the E?F pair (translated and revised in Canada) and from the F?E pair (translated and revised in the United States). Each sub-corpus (E?F and F?E) was analyzed by a team of scholars and/or revisers, according to Louise Brunette's (1997) revision criteria: accuracy, readability, appropriateness and linguistic coding. The study looked at the number of corrections, omissions and revisor-injected errors, in relation to these four criteria.

No abstract