ANALYSIS OF SKULL MEASUREMENTS AND MEASURERS: CAN WE USE DATA OBTAINED BY VARIOUS OBSERVERS?
ANALYSIS OF SKULL MEASUREMENTS AND MEASURERS: CAN WE USE DATA OBTAINED BY VARIOUS OBSERVERS?
复制标题
头骨测量和测量者的分析：我们可以使用不同观察者获得的数据吗？
DOI：
--
复制
发表时间：
1998
期刊：
影响因子：
--
通讯作者：
J. Palmeirim
中科院分区：
文献类型：
--
作者：
J. Palmeirim
为了研究颅骨测量数据的误差来源和使用不同观察者测量的相关问题，用卡尺测量了10个蝙蝠头骨样本，每个人由7个不同的人测量了3次。观察者内错误和观察者间错误都对某些性状的总方差有显著贡献。观察者内变异占总变异的17%(中位数=10%)，而观察者间变异达到30.7%(中位数=3.9%)。具有高观察者内变异性的特征是那些表现出高观察者间变异性的特征。在涉及由不同观察者进行的测量的比较中，存在相当大的风险，即仅由于测量伪影而可能产生统计上显著的结果，特别是在使用大样本时。然而，如果将潜在危险因素排除在考虑范围之外，则可以在一次分析中使用由不同观察者进行的测量。这些特征包括观察者内部变异性高、均值小或终点形态不清的特征。
To study sources of error in craniometric data and problems associated with using measurements taken by different observers, a sample of 10 bat skulls was measured with calipers three times by each of seven different individuals. Both intra- and inter-observer errors significantly contributed to the total variance of some characters. Intra-observer variability was responsible for -<17% of the total variance (median = 10%) whereas interobserver variability reached 30.7% (median = 3.9%). Characters with high intra-observer variability were those that exhibited high inter-observer variability. In comparisons involving measurements taken by different observers, a considerable risk exists that statistically significant results may arise due to measuring artifacts alone, especially when using large samples. However, measurements taken by different observers can be used in a single analysis if potentially risky characters are excluded from consideration. These include characters with high intra-observer variability, small means, or morphologically unclear endpoints.