Analysis of errors in cephalometric measurements of three-dimensional distances on the mandible.

Abstract Reliability (reproducibility) of measurements of five mandibular dimensions was analysed. Sixteen replicate determinations per mandibular dimension were made on cephalograms of ten boys 7 years of age by two trained locators (each independently making two tracings), and two measurers each making two separate measurements on every tracing; total sample of 160 distances per dimension was thus obtained. By analysis of variance technique (nested design), variability in distances observed was partitioned into four factors: between-and within-locators, and between-and within-measurers. Standard deviation component estimates for each factor showed amounts of contribution by each source. Landmark location variability was about 5 times that due to measurement. Comparison of within-system and among-child standard deviation component estimates showed within-system variabilities to be about 15 per cent to 37 per cent of among-child variabilities.