关于标定

On Scaling

标定是使考试分数更加科学和有意义的一种统计过程。我国高考中曾经在一些省份进行的分数转换试验就是一种标定。本文从理论上介绍了标定的基本概念、教育和心理测验行业标准对标定的要求和在教育考试中常用的一些标定方法。指出高考分数转换失败的重要原因在于没有处理好不同科目的权重问题，提出基于累积的原始分数分布进行高考标定，并把重点放在年度间等值上的主张。

Scaling is a statistical process which makes test scores meaningful. An example of scaling process was the score transformation pilot study of the higher education entrance exam in some provinces. This paper introduced some basic concepts related with scaling, the requirements given by the standards for education and psychology tests, and some practical applicable methods. It is indicated that one critical reason that the score transformation pilot study had failed was the weights of different subjects had not been handled appropriately and a suggestion to conduct scaling based on the cumulative raw score distributions and put focus on equating instead of scaling was given.