Investigating Machine Translation Post-Editing Outcomes in Expert-Based Closed Crowdsourcing: Evidence from a Qualitative Study

Articles in Press

Document Type : Research Paper

Author

Foreign Languages Department, Faculty of Tourism, Higher Education Complex of Bam, Bam, Kerman, Iran

10.22111/ijals.2026.52165.2526

Abstract

Post-editing refers to activities aimed at improving machine translation (MT) output to achieve human-level quality. Crowdsourcing translation post-editing (CTPE) can provide a cost-effective approach, yet open crowdsourcing models often suffer from inconsistent quality due to varying participant expertise. This study examines a closed, expert-driven crowdsourcing model for Persian health translations, using a qualitative approach to analyze post-editing practices within a Telegram-based expert community. Thirty qualified translators performed post-editing on machine-generated translations of health-related texts, with three independent expert proofreaders evaluating multiple candidate outputs per instance. Problematic instances were identified and assessed according to linguistic criteria adapted from the Patient-Oriented and Culturally Appropriate (POCA) health translation framework, focusing on clarity, coherence, lexical appropriateness, and readability. Findings indicate that this closed model facilitates careful, context-sensitive post-editing, producing multiple acceptable solutions per instance, and may tentatively reduce the need for extensive proofreading. These results underscore the potential of structured, expert-driven CTPE for supporting translation quality in specialized domains while highlighting the interpretive complexity inherent in post-editing tasks.

Keywords

Iranian Journal of Applied Language Studies

Articles in Press, Accepted Manuscript
Available Online from 02 July 2026

Files

Share

How to cite

Statistics