2012 Mean Comparison
2012 Mean Comparison
Means
Anova
F-test can be used to determine whether the
expected responses at the t levels of an
experimental factor differ from each other
When the null hypothesis is rejected, it may be
desirable to find which mean(s) is (are) different,
and at what ranking order.
Means comparison
Three categories:
1. Pair-wise comparisons (Post-Hoc Comparison)
2. Comparison specified prior to performing the experiment
(Planned comparison)
3. Comparison specified after observing the outcome of the
experiment (Un-planned comparison)
Pair Comparison
Suppose there are t means
x1 , x2 , , x t
An F-test has revealed that there are
significant differences amongst the t
means
Performing an analysis to determine
precisely where the differences exist.
Pair Comparison
This method builds on the equal variances ttest of the difference between two means.
The test statistic is improved by using MSE
rather than sp2.
It is concluded that i and j differ (at %
significance level if |i - j| > LSD, where
LSD t 2,dfe
1 1
MSE ( )
ni n j
Example:
Cassava yields (ton/ha)
Source
of
variati
on
Degrees Sum of
of
Square
Freedo
m
Mena
Square
F
calculat
ed
Treatm
ent
136
45,333
34
Block
Error
3
9
40
12
13,333
1,33
10
F-table:
3,86
R p r , p , MSE n
where r,p, is the Duncans Significant Range Value with
parameters p
(= range-value) and (= MSE degree-of-freedom), and
experiment-wise alpha level (= joint).
2
18.00
6.09
4.50
3.93
3.64
3.46
3.35
3.26
3.20
3.15
3.11
3.08
3.06
3.03
3.01
3.00
2.98
2.97
2.98
2.95
2.89
2.86
2.83
2.80
2.77
3
18.00
6.09
4.50
4.01
3.74
3.58
3.47
3.39
3.34
3.30
3.27
3.23
3.21
3.18
3.16
3.15
3.13
3.12
3.11
3.10
3.04
3.01
2.98
2.95
2.92
4
18.00
6.09
4.50
4.02
3.79
3.64
3.54
3.47
3.41
3.37
3.35
3.33
3.30
3.27
3.25
3.23
3.22
3.21
3.19
3.18
3.12
3.10
3.08
3.05
3.02
5
18.00
6.09
4.50
4.02
3.83
3.68
3.58
3.52
3.47
3.43
3.39
3.36
3.35
3.33
3.31
3.30
3.28
3.27
3.26
3.25
3.20
3.17
3.14
3.12
3.09
6
18.00
6.09
4.50
4.02
3.83
3.68
3.60
3.55
3.50
3.46
3.43
3.40
3.38
3.37
3.36
3.34
3.33
3.32
3.31
3.30
3.25
3.22
3.20
3.18
3.15
p
7
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.47
3.44
3.42
3.41
3.39
3.38
3.37
3.36
3.35
3.35
3.34
3.29
3.27
3.24
3.22
3.19
8
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.47
3.45
3.44
3.42
3.41
3.40
3.39
3.38
3.37
3.37
3.36
3.32
3.30
3.28
3.26
3.23
9
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.47
3.46
3.44
3.44
3.42
3.42
3.41
3.40
3.39
3.39
3.38
3.35
3.33
3.31
3.29
3.26
10
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.47
3.46
3.46
3.45
3.44
3.43
3.43
3.42
3.41
3.41
3.40
3.37
3.35
3.33
3.32
3.29
20
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.48
3.48
3.48
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
50
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.48
3.48
3.48
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.48
3.53
3.61
100
18.00
6.09
4.50
4.02
3.83
3.68
3.61
3.56
3.52
3.48
3.48
3.48
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.48
3.53
3.67
Student-Newman-Keuls Test
Similar to the Duncan Multiple Range test, the StudentNewman-Keuls Test uses different Significant Difference
values for means next to each other, and those with 1, 2,
, a means in between the two means being compared.
The Significant Difference or the range value for this
test is
K p q , p , MSE n
Student-Newman-Keuls Test
MSE
q (dft , dfe)
ng
Studentized Range
Statistic
Scheffe
There are many multiple (post hoc)
comparison procedures
Considerable controversy:
I have not included the multiple
comparison methods of Duncan
because I have been unable to
understand their justification
y1 y3 y4 y5 0
C ci yi
i 1
and it is required:
c
i 1
ssC
c yi
i 1 i
t
n c
2
i 1 i
Planned Comparisons
1 dfe
Un-planned Comparisons or
Contrasts
c d
i 1
Contrast orthogonal
examples
Treatm
ent
Adira-4
GH-6
GH-7
Local
Yields (ton/ha)
19
25
18
18
Orthogonal Polynomials
Special sets of coefficients that test for
bends but manage to remain uncorrelated
with one another.
Sometimes orthogonal polynomials can be
used to analyze experimental data to test
for curves.
Restrictive
assumptions:
Require quantitative factors
Equal spacing of factor levels (d)
Equal numbers of observations at each cell
(rj)
Usually, only the linear and quadratic
Polynomial Examples
Treatment
(Urea
dosage)
kg/ha
50
100
150
200
Yields (ton/ha)
19
25
18
18
Orthogonal Polynomial
26
A second-order model
(quadratic model)
Polynomial Models
Polynomial
models
are
useful
in
situations where the analyst knows that
curvilinear effects are present in the
true response function.
Polynomial models are also useful as
approximating functions to unknown and
possible
very
complex
nonlinear
relationship.
Polynomial model is the Taylor series
expansion of the unknown function.
Theoretical background
Scatter diagram
Orthogonal polynomial test
30
Theoretical background
1.
2.
3.
Scatter Diagram
Scatter Diagram