Post-Editing a Google Translated Output: Experienced Translators vs. Trainees

Abstract

The present empirical study reports on an experiment in which 20 participants (actual job applicants) were asked to post-edit a 394-word legal Google translated text (GTT) to investigate the type of post edits done in relation to the quality of the product as assessed by the recruitment test assessors in the translation service provider. For the purposes of the empirical research, participants were categorized in two groups; translators with practical experience between 3-5 years and trainees (recent translation graduates) with no practical experience. Assessors at the translation service provider used LISA QA model 3.1 version for quality assessment. The three factors investigated by assessors were time spent on the task, number and type of changes (post-edits) as well as the quality of the final post-edited text based on errors committed in the post-editing (PE). Results reveal a correlation between the type and number of edits done by participants and the quality of the final output and consequently a correlation between practical experience and the quality of the post-edited output. The research unveils some areas that need to be improved in the study plans at the translator training programs in Jordan, particularly in relation to PE efficiency. Results also imply that general experience in translation may not be enough to excel in post-editing specialized texts that require special knowledge in a given subject matter.