Statistical Methods for Comparison of Forms - An Invariant Approach to Statistical Analysis of Shapes - page 183

LND 2 LND 5 0.05064

LND 2 LND 4 0.07016

LND 2 LND 3 0.08164

LND 1 LND 5 0.09063

LND 1 LND 6 0.11170

LND 1 LND 7 0.13085

LND 1 LND 11 0.13680

LND 1 LND 9 0.13751

LND 1 LND 10 0.13835

LND 1 LND 8 0.15318

Z statistic: 0.15318
In this example, the element at the positive end of the vector is far-
ther from 1 than the element at the negative end of the vector (the
shape difference between LND 1 and LND 8

0.15318 while the shape

.14184), and so the former
is chosen as the Z statistic. Using the parametric bootstrap procedure
described previously, a 90% confidence interval for the Z statistic is
estimated and shown below.

difference between LND 1 and LND 4

α

Table 4.4. Confidence Interval for Z with 0.100

Distribution of Bootstrapped Z Statistics
-0.179 |*

-0.158 |***********

-0.138 |*********

-0.118 |

-0.098 |

-0.078 |

-0.058 |

-0.037 |

-0.017 |

0.003 |

0.023 |

0.043 |

0.064 |

0.084 |

0.104 |

0.124 |****

0.144 |Z******************************

0.164 |********************************

0.185 |************

0.205 |*

Upper bound: 0.18138 Lower bound: -0.16500