Item-based translation evaluation : the evolution from RPM to CDI to PIE

Abstract

Translation evaluation is a key aspect of translator training and professional translation which poses quite a few challenges, particularly in terms of subjectivity. In the past two decades there has been a marked shift from subjective, evaluator-dependent translation evaluation methods to more objective, (partly) evaluator-independent methods with automation potential. One such method is Preselected Items Evaluation (PIE), a criterion- and norm-referenced item-based analytical translation evaluation method developed in 2014 (Kockaert & Segers, 2014). It can be used as a criterion-referenced method only or as a criterion- and normreferenced method, which includes a psychometric and a statistical component. The entire method is strongly rooted in docimology, which guarantees greater transparency and objectivity, as well as inter- and intra-rater reliability. The proposed poster illustrates the evolution of PIE, which was inspired by the Reference Points Method (RPM)1 and the latter’s successor, Calibration of Dichotomous Items (CDI), and is currently being refined. Tracing the development of PIE provides an insight into the advantages and disadvantages of item-based analytical methods. It also gives us a glimpse of the (r)evolutions that lie ahead.