Sources of goniometric error at the elbow.

We assessed accuracy and potential sources of error in goniometry by using a photographic reference standard. Forty-six physical therapy students measured elbow positions using plastic or steel goniometers following three protocols: ALIGN, in which the investigator's elbow was splinted and bony landmarks were prelabeled; ASSIGN, in which the elbow remained splinted but labels were removed; and PROM, in which raters measured full passive flexion of the elbow. F ratios of variances indicated that alignment of goniometer, identification of landmarks, and variations in manual force during PROM contributed to goniometric error and that accuracy of joint angle measurement by photography (s +/- 0.7-1.1 degrees) was greater than by standard goniometry (s +/- 2.4-3.4 degrees). Analysis of variance and post-hoc test results unexpectedly indicated that all but one goniometric mean differed statistically (p less than .05) from associated photographic means. Small systematic errors in alignment of goniometers and identification of reference landmarks may have accounted for these differences. The results indicate that relatively inexperienced raters should be able to use goniometers accurately to measure elbow position when given standardized methods to follow.