Biology Reference

In-Depth Information

Monte Carlo-based EDMA-II testing procedure

The difference between EDMA-I and EDMA-II is described in detail in

Part 1
of this chapter. We do not repeat the discussion here.

STEP 1: Calculate the shape difference matrix using the esti-

mated mean forms for the two populations. Calculate the

extreme value in the shape difference matrix, either negative

or positive.

STEP 2: Assuming that the first population is the base sam-

ple, generate a new set of samples
X
1
*
,
X
2
*
,…,
X
n
*

from

ˆ

,
ˆ

ˆ

,
ˆ

and
Y
1
*
,
Y
2
*
,…,
Y
m
*
from .

STEP 3: Calculate the shape difference matrix based on these

samples and calculate the most extreme value, negative or

positive and call it
T
*
.

STEP 4: Repeat Steps 2 and 3
B
times to obtain
T
1
*
,
T
2
*
,…,
T
B
*
.

STEP 5: Arrange
T
1
*
,
T
2
*
,…,
T
B
*
in an increasing order. Assume

that the size of the test is 0.10, then calculate the 5-th per-

centile and the 95-th percentile values of these
T
*
's. If zero

lies in between these values, then accept the null hypothesis

of equality of shapes, otherwise reject it.

NM

(

,)

I

NM

(

,)

I

1

K

,

1

1

K

,

1

We want to reemphasize the fact that if the distribution of the test

statistic is bimodal, one should not conduct the above test. In general,

we prefer the confidence intervals for the shape or form difference

matrices over simplistic null hypothesis testing.

Search WWH ::

Custom Search