Table 3.
Results of analyses of measurement error on data automatically acquired using GPSA with and without dimensionality reduction
| df | SS | MS | Rsq | F | Z | p | Repeatability | |
|---|---|---|---|---|---|---|---|---|
| Procrustes ANOVA, full set of nonzero principal coordinates | ||||||||
| Species | 23 | 11394.1 | 495.4 | 0.72365 | 5.1235 | 2.1345 | .001 | 0.58 |
| Residuals | 45 | 4351.1 | 96.69 | |||||
| Total | 68 | 15745.3 | ||||||
| Procrustes ANOVA, first five principal coordinates | ||||||||
| Species | 23 | 7061.6 | 307.024 | 0.96809 | 59.364 | 2.8411 | .001 | 0.95 |
| Residuals | 45 | 232.7 | 5.172 | |||||
| Total | 68 | 7294.3 | ||||||
| Solutionix | NextEngine | Photogrammetry | ||||||
|---|---|---|---|---|---|---|---|---|
| p‐values for the pairwise tests of bias | ||||||||
| Solutionix | – | <0.001 | <0.001 | |||||
| NextEngine | 0.02 | – | 0.006 | |||||
| Photogrammetry | <0.001 | 0.172 | – | |||||
df, degrees of freedom; SS, sum of squares; MS, mean squares; Rsq, r squared; p, p‐value; in the pairwise test for bias, above the diagonal test based on the full set of nonzero principal coordinates and below the diagonal test based on the first five principal coordinates.