Measurement on Radiographs

¹ Department of Orthodontics, Royal Dental College, Copenhagen, Denmark

A study was made to determine to what extent the manner of performing measurements on radiographs could affect the correlation between them. In Part I of the study the effect on the correlation coefficient of marking the reference points and lines on the film was investigated. It was found that the marking introduced a systematic error that increased the correlation coefficient. It was therefore concluded that all measurements to be used in correlation analyses should be made without marking reference points or lines. Such a method was described. In Part II the effect of using measurements obtained by the addition or subtraction of other measurements made directly on the film was studied. It was found that the use of such indirect measurements could, under certain conditions, cause an increase, and under other conditions a decrease, in the correlation coefficient. The conclusion was drawn, therefore, that all measurements used in correlation analyses should be measured directly and should not be obtained by calculation from other values.