SAE Technical Standards – J2450 scoring metric

SAE Technical Standards – J2450 scoring metric

Prior to the proposal of the J2450 scoring metric, created by the SAE Technical Standards, quality measurement of language translation in the automotive industry was subjective. In an industry where poor quality translation services could result in lower consumer confidence, additional warranty costs as well as injury and death, it was determined that a specified method for quality assurance was necessary. In response, the SAE developed the J2450 scoring metric to allow for tighter control of translation quality while also making sure that the source language, target language and method of translation did not bias the measure of quality.

More specifically, the actual metric consists of four elements, each of which further classify errors. The first element of the standard classifies errors into seven primary error categories, ranging from the use of the wrong term, grammatical errors as well miscellaneous errors. Secondly, errors are then classified as either serious or minor, as determined by the evaluator. Next, each error is given a numerical score, which is defined in the standard. The fourth and final element results in the normalized score which is computed by dividing the sum of the individual numeric weights by the total number of words in the source language text.

It is important to note that while the standard results in a normalized score, the SAE does not specify or suggest an acceptable threshold for the score.

For clients that require definite quality assurance measures to be part of their workflows, we are able to implement the J2450 scoring metric, among other procedures.