Poulsen S, Bille J, Rugg-Gunn A J
Community Dent Oral Epidemiol. 1980 Jun;8(3):135-8. doi: 10.1111/j.1600-0528.1980.tb01274.x.
Eighteen dentists working in the same School Dental Service participated in a calibration trial with the purpose of reducing interexaminer disagreement on radiographic diagnosis of approximal caries. The calibration program was in two stages: firstly, discussion in small groups and, secondly, the use of a reference standard. The effect of the training programs on interexaminer disagreement was limited, but a statistically significant decrease in the number of D-surfaces recorded per child was observed. The finding that the major part of the error variance was due to random error, rather than systematic error, may explain why the calibration program was unsuccessful at reducing error.
18名在同一学校牙科服务机构工作的牙医参与了一项校准试验，目的是减少检查者之间在邻面龋放射诊断上的分歧。校准程序分两个阶段：首先，小组讨论；其次，使用参考标准。培训计划对检查者之间分歧的影响有限，但观察到每个儿童记录的D面数量有统计学意义的减少。误差方差的主要部分是由于随机误差而非系统误差这一发现，或许可以解释校准程序为何未能成功减少误差。
