Statistical Methods for Comparison of Forms - An Invariant Approach to Statistical Analysis of Shapes - page 175

LANDMARK 7 with LANDMARK 6 1.019 1.056 1.093*

LANDMARK 8 with LANDMARK 1 1.112 1.151 1.182*

LANDMARK 8 with LANDMARK 2 1.042 1.067 1.093*

LANDMARK 8 with LANDMARK 3 1.015 1.036 1.057*

LANDMARK 8 with LANDMARK 4 1.012 1.035 1.056*

LANDMARK 8 with LANDMARK 5 0.968 1.031 1.077

LANDMARK 8 with LANDMARK 6 1.031 1.068 1.099*

LANDMARK 8 with LANDMARK 7 1.011 1.054 1.109*

LANDMARK 9 with LANDMARK 1 1.130 1.170 1.197*

LANDMARK 9 with LANDMARK 2 1.050 1.070 1.085*

LANDMARK 9 with LANDMARK 3 1.009 1.025 1.037*

LANDMARK 9 with LANDMARK 4 1.003 1.025 1.038*

LANDMARK 9 with LANDMARK 5 0.920 1.014 1.094

LANDMARK 9 with LANDMARK 6 1.069 1.108 1.141*

LANDMARK 9 with LANDMARK 7 0.980 1.020 1.068

LANDMARK 9 with LANDMARK 8 0.989 1.038 1.087

LANDMARK 10 with LANDMARK 1 1.163 1.196 1.217*

LANDMARK 10 with LANDMARK 2 1.049 1.068 1.080*

LANDMARK 10 with LANDMARK 3 1.007 1.018 1.026*

LANDMARK 10 with LANDMARK 4 1.002 1.019 1.028*

LANDMARK 10 with LANDMARK 5 0.918 0.995 1.053

LANDMARK 10 with LANDMARK 6 1.019 1.053 1.079*

LANDMARK 10 with LANDMARK 7 1.001 1.041 1.078*

LANDMARK 10 with LANDMARK 8 1.025 1.077 1.132*

LANDMARK 10 with LANDMARK 9 0.941 1.000 1.062

LANDMARK 11 with LANDMARK 1 1.190 1.259 1.313*

LANDMARK 11 with LANDMARK 2 1.059 1.084 1.104*

LANDMARK 11 with LANDMARK 3 0.992 1.012 1.030

LANDMARK 11 with LANDMARK 4 0.994 1.015 1.031

LANDMARK 11 with LANDMARK 5 0.927 1.006 1.057

LANDMARK 11 with LANDMARK 6 1.021 1.054 1.079*

LANDMARK 11 with LANDMARK 7 1.007 1.041 1.068*

LANDMARK 11 with LANDMARK 8 1.016 1.060 1.099*

LANDMARK 11 with LANDMARK 9 1.012 1.050 1.089*

LANDMARK 11 with LANDMARK 10 0.967 1.031 1.109

Observe the first entry in the vector written above. The estimate 1.090
is greater than 1, indicating that the linear distance from landmark 1
to 2 is larger in the mean normal mandible (numerator) than in the
mean Ts65Dn mandible (denominator). Notice that the confidence
interval for this linear distance does not contain the value 1. A confi-
dence interval that does not contain the value 1 indicates that the two
forms are different for that specific linear distance. Confidence inter-
vals for distances between landmarks 3 and 4, and 3 and 5 contain the
value 1 and indicate that these distances are not significantly different
in the two samples using the 0.10 confidence limits.

α