Analysis of variance in assessing registrations of natural head position

Recordings of natural head position may be performed directly during radiographic registration or indirectly by a photographic transfer procedure. The aim of this investigation was to examine and quantify the different sources of error in one photographic method using an analysis of variance. The material consisted of 15 senior dental students between the ages of 24-34 years. Intra-observer results show that the error associated with transfer of the plumb line to the tracing and digital recording of the reference points is small in comparison with the tracing error and intra-individual postural variation as registered photographically. By relating the total analysis error to the inter-individual postural variation it was possible to perform a preliminary evaluation of the precision of the entire registration method prior to continued collection of raw data. In this case it was found that duplication of tracings and photographs was the most rational approach resulting in an analysis error equivalent to 9.5 percent of the estimated postural variation between individuals.