Translation and post-editing performance of translation students – a cross-sectional analysis

Abstract

This study presents partial results of a comprehensive study to reveal what role PACTE’s translation sub-competences play in human translation and in the post-editing of machine translated texts. In the PACTE model, both language competence and background knowledge related to the source text are given a prominent role. The present research explores how these factors are associated with MA students' translation performance. 20 first-year and 27 second-year master’s students of translation (University of Szeged, Hungary) translated or post-edited the abstract of a study on bilingualism from English to Hungarian and completed a test measuring their relevant thematic knowledge and language tests assessing their source language competence. Our analysis focuses on comparing the quality of translated and post-edited texts, and on the time needed to complete the target text. The correlations between test scores, on the one hand, and error types and error numbers in the translated and post-edited Hungarian target language texts, on the other hand, are also examined. Our results indicate that both at the beginning and at the end of the training, post-editing was faster than human translation and post-edited texts contained fewer errors than human translations. In the second-year sample, thematic knowledge and time on task showed significant correlations with performance indices.