Method errors and reliabilities were estimated for seven sonographic measurements in pregnancies of 106 women examined between January and July 1989. Teams of two experienced sonographers replicated the following measurements: biparietal diameter (BPD), occipital-frontal diameter (OFD), anterior-posterior diameter (APD), transabdominal distance (TAD), and femur diaphysis length (FDL). Multilevel modeling procedures were used to estimate the variance components. Significant (p < 0.01) covariates in the fixed part of the model included an increase in error with greater parity, estimated menstrual age (EMA), and maternal abdominal wall thickness (taken at the umbilicus). Intraobserver reliability ranged from 85.2% (AC) to 99.3% (FDL); interobserver reliability ranged from 80.8% (TAD) to 92.4% (FDL). Method errors, describing the expected error for 68% of the measurements taken, ranged from 0.8 mm to 7.7 mm (intraobserver) and from 1.2 mm to 7.8 mm (interobserver). These results suggest that large error components should be considered in the interpretation of the reliability of ultrasonographically obtained measurements.
