Smith D W, Bluhm G B
Eur J Rheumatol Inflamm. 1982;5(2):198-205.
Three experiments were conducted to evaluate the rater reliability for scoring osseous defects (articular erosions) and joint space narrowing (cartilage destruction) in patients with rheumatoid arthritis. The experiments show that two raters, using different techniques, obtain similar scores; the scores differing primarily by a multiplicative factor. A single rater achieves reproducible results under similar conditions. But with a change in reading technique, such as using a magnifying lens and intense light, the same rater obtains a change in scores. The problems of measurement must be carefully examined for any proposed measurement technique. The superficial objectivity and precision of a measurement scale are less important than its objectively and precision of a measurement scale are less important than its objectively evaluated reliability. Comparing alternative procedures is a statistical problem, and can be subjected to suitable statistical analysis to evaluate alternatives.
进行了三项实验，以评估类风湿性关节炎患者骨缺损（关节侵蚀）评分和关节间隙变窄（软骨破坏）评分的评估者信度。实验表明，两名评估者采用不同技术得出的分数相似；分数的差异主要在于一个乘法因子。单一评估者在相似条件下可获得可重复的结果。但是，当阅读技术发生变化，例如使用放大镜和强光时，同一名评估者的分数会发生变化。对于任何提议的测量技术，都必须仔细检查测量问题。测量量表表面上的客观性和精确性不如其经过客观评估的信度重要。比较替代程序是一个统计学问题，可以进行适当的统计分析以评估替代方案。
