backtranslation

Backtranslation is a method in translation quality assessment in which a text translated into a target language is translated back into the original source language. The process aims to evaluate semantic equivalence and to identify translation or cultural issues that may alter meaning.

Typically, forward translation produces the target-language version, which is then back-translated by a second translator who has not seen the original text. The back-translation is compared with the original to identify deviations, ambiguous terms, or cultural mismatches. Findings guide revisions of the target-language version. Some workflows employ independent back-translation and blind or double translation to reduce bias.

Common applications include translating patient-reported outcome measures in clinical research, where preserving clinical meaning is essential; software and product localization; and evaluation or benchmarking of machine translation systems.

Advantages include detecting mistranslations, loss of nuance, and cultural mismatches, supporting quality assurance and regulatory compliance in medical contexts.

Limitations include that back-translation cannot guarantee conceptual equivalence, and may reflect back-translator choices or introduce new ambiguities. It is resource-intensive and should be supplemented by other quality checks, such as expert review and contextual testing.

Related approaches include forward translation with reconciliation, back-translation with multiple rounds, and independent reviews.