Toward a theory of bias and equivalence

To begin with, this article moves to two notions namely bias and equivalence. Firstly, these two terms are explained. The writer says bias is related to validity. In fact, it is said that something is biased when its scores do not have the same psychological meaning across languages. Equivalence is defined as the difference between measurement levels of within- and between- group comparisons. If the measure is biased against some cultural groups, individual difference within a cultural population and across cultural populations is not measured. At the same scale. There are three types of bias. They are construct bias, method bias, and item bias or differential item functioning.

Construct bias is the dissimilarity of construct across cultures. Method bias is a general word that refers to all sources of bias resulting from methodological- procedural aspects of a study. Sample bias is subsuming all differences in scores that are related to specific aspects of a sample. Parallel to three types of bias, there are three types of equivalence. They are construct equivalence, method equivalence, and measurement unit equivalence. It should be said that these concepts namely equivalence and bias are integral part of any empirical study.

Therefore, in any step of quantitative studies, there is a bias. All in all, the writer says that bias, holistically, refers to the absence or presence of nuisance factors. On the other hand, equivalence is the implication of bias on the cross cultural score comparisons to be made. As writer mentioned in the article, in order to translate better the existence of good equivalences are urgent. Consequently, in order to get the highest possible level of equivalence, bias should be scrutinized in each and all steps of quantitative study. To put it in a nutshell, the most important part of translation is equivalence and the most important part of cultural issues in translation is the bias. Consequently, in order to translate better and more adequate, we have to know the bias, then, we will choose the best equivalence across languages.