Statistical notes for clinical researchers: Evaluation of measurement error 1: using intraclass correlation coefficients

Abstract

Evaluation of measurement error is a fundamental procedure in clinical studies, field surveys, or experimental researches to confirm the reliability of measurements. If we examine oral health status of a patient, the number of caries teeth or degree of periodontal pocket depth need to be similar when an examiner repeated the measuring procedure (intra-examiner reliability) or when two independent examiners repeated the measuring procedure, to guarantee the reliability of measurement. In experimental researches, confirming small measurement error of measuring machines should be a prerequisite to start the main measuring procedure for the study.