Table 3. Comparison of operator error in 3D scan and µCT scan datasets using Procrustes ANOVAs and repeatability scores.

The repeatability score (R) is a value that reflects the ease of digitizing in a repeated measure study design. It is calculated from the Procrustes ANOVA using formulas for the intra-class correlation coefficient. The Procrustes ANOVAs were found by subsetting the dataset by scan device and by landmark types and then performing separate generalized Procrustes and bilateral symmetry alignments. (A–C) Results from the 3D-only dataset. (D–F) Results from the µCT-only dataset. (A) and (D) show the repeatabilites from the entire landmark datasets of each scan device. (B) and (E) remove patch points. (C) and (F) contain only fixed landmarks.

(A) 3D scan all landmarks including patches (n = 289)
	Df	SS	MS	Rsq	F	Z	Pr(>F)	R
Ind	18	3.53E−02	1.96E−03	0.826	10.0	16.0	0.001	0.750
Residuals	38	7.46E−03	1.96E−04	0.174
Total	56	4.28E−02