Statistical Methods for Comparison of Forms - An Invariant Approach to Statistical Analysis of Shapes - page 191

Table 4.7. Distribution of Bootstrapped T Statistics
1.034 |****

1.044 |***********

1.054 |***************

1.064 |**********************

1.074 |*********************

1.085 |***************************

1.095 |********************

1.105 |*****************

1.115 |***********

1.125 |*****************

1.136 |*******

1.146 |*****

1.156 |**********

1.166 |**T obs

1.176 |****

1.186 |***

1.197 |**

1.207 |

1.217 |*

1.227 |*

When confidence intervals are estimated for the reduced set of
landmarks (see below), the exact values of the confidence intervals will
not be the same as when these are estimated with all eleven land-
marks. This occurs due to changes in the estimates for K * or the
boostrapping procedures when landmarks are added to or subtracted
from analysis (review Chapter 3 ).

90% confidence intervals for form differences between Ts65Dn
and normal mouse mandibles after deleting landmark1:

Low Estimate High

LND 3 LND 2 1.105 1.146 1.179

LND 4 LND 2 1.091 1.150 1.215

LND 4 LND 3 0.984 1.017 1.056

LND 5 LND 2 1.053 1.102 1.160

LND 5 LND 3 1.005 1.029 1.063

LND 5 LND 4 1.007 1.032 1.068