introduction to randomized block designs
TRANSCRIPT
1
Reduced slides
Introduction to Randomized
block designs
Accounting for predicted but random
variance
A
B C
D B
C A
D C
B D
A B
D C
A
Block 1 Block 2 Block 3 Block 4
2
Blocking
• Aim:
– Reduce unexplained variation, without increasing size of experiment.
Approach:
– Group experimental units (“replicates”) into blocks.
– Blocks usually spatial units, one experimental unit from each treatment in each block.
Walter & O’Dowd (1992)
• Effects of domatia (cavities
on leaves) on number of
mites
• Two treatments (Factor A):
– shaving domatia removes
domatia from leaves
– normal domatia as control
• Required 14 leaves for
each treatment
3
Control leaves Shaved domatia leaves
Completely randomized design: - 28 leaves randomly allocated to each of 2
treatments
Completely randomized ANOVA • Factor A with p groups (p = 2 for domatia)
• n replicates within each group (n = 14 pairs of
leaves)
Source general df example df
Factor A p-1 1
Residual p(n-1) 26
Total pn-1 27
4
Walter & O’Dowd (1992)
• Required 14 leaves for each treatment
• Set up as blocked design
– paired leaves (14 pairs) chosen
– 1 leaf in each pair shaved, 1 leaf in each
pair control
1 block
Control leaves Shaved domatia leaves
5
Rationale for blocking
• Micro-temperature, humidity, leaf age,
etc. more similar within block than
between blocks
• Variation in response variable (mite
number) between leaves within block
(leaf pair) < variation between leaves
between blocks
Rationale for blocking
• Some of unexplained (residual)
variation in response variable from
completely randomized design now
explained by differences between
blocks
• More precise estimate of treatment
effects than if leaves were chosen
completely randomly
6
Null hypotheses
• No main effect of Factor A
– H0: m1 = m2 = … = mi = ... = m
– H0: a1 = a2 = … = ai = ... = 0 (ai = mi - m)
– no effect of shaving domatia, pooling
blocks
• Factor A usually fixed
Null hypotheses
• No effect of factor B (blocks):
– no difference between blocks (leaf pairs),
pooling treatments
• Blocks usually random factor:
– sample of blocks from population of blocks
– H0: 2 = 0
7
• Factor A with p groups (p = 2 treatments for
domatia)
• Factor B with q blocks (q = 14 pairs of leaves)
Source general example
Factor A p-1 1
Factor B (blocks) q-1 13
Residual (p-1)(q-1) 13
Total pq-1 27
Randomized blocks ANOVA
Notice
that this
is not
pq(n-1)
Randomized block ANOVA
• Randomized block ANOVA is 2 factor
factorial design
– BUT no replicates (n) within each cell
(treatment-block combination), i.e.
unreplicated 2 factor design
– No measure of within-cell variation
– No test for treatment by block interaction
8
If factor A fixed and factor B (Blocks)
random:
MSwatering 2 + a2 + n (ai)
2/p-1
MSBlocks 2 + n2
MSResidual 2 + a2
Expected mean squares
General Randomized Block
Mean Square calcualtion
9
Testing null hypotheses
• Factor A fixed and blocks random
• If H0 no effects of factor A is true:
– then F-ratio MSA / MSResidual 1
• If H0 no variance among blocks is true:
– no F-ratio for test unless no interaction
assumed
– if blocks fixed, then F-ratio MSB / MSResidual
1
Walter & O’Dowd (1992)
• Factor A (treatment - shaved and
unshaved domatia) - fixed
• Blocks (14 pairs of leaves) - random
Source df MS F P
Treatment 1 31.34 11.32 0.005
Block 13 1.77 0.64 0.784 ??
Residual 13 2.77
Should this be reported??
10
Randomized Block vs
Completely Randomized designs
• Total number of experimental units
same in both designs
– 28 leaves in total for domatia experiment
• Test of factor A (treatments) has fewer
df in block design:
– reduced power of test
RCB vs CR designs
• MSResidual smaller in block design if blocks explain some of variation in Y:
– increased power of test
• If decrease in MSResidual (unexplained variation) outweighs loss of df, then block design is better:
– when blocks explain much of variation in Y
11
Assumptions
• Normality of response variable
– boxplots etc.
• No interaction between blocks and
factor A, otherwise
– MSResidual increase proportionally more than
MSA with reduced power of F-ratio test for
A (treatments)
– interpretation of main effects may be
difficult, just like replicated factorial ANOVA
Interaction plots
Y
Block
Y
No interaction
Interaction
12
Worked Example – seastar colors
• Comparison of numbers of purple vs orange
seastars along the CA coast
• Data number of purple and orange seastars
collected at 7 random locations
• Compare models (block vs completely
random vs paired t test)
sea star colors all sites two sample
Number of Seastars as a function
of color and site
Any obvious problem with the data??
13
Diagnostics: Log transform helps normality and
homogeneity of variance assumptions
Model 1: One factor ANOVA
Why the difference?
SE=0.184
SE=0.181
14
Model 2: Paired t test
• Accounts for site specific (block) differences
• But no way to assess site (block) differences
LORANGE LPURPLE
Index of Case
1.0
1.5
2.0
2.5
3.0
3.5
Valu
e
Model 3: Randomized Block Design -
using least squares
• Accounts for and assesses (with a
caveat) site specific effects
1) Compare to paired t (same p value for Color) but no Site effect
2) Compare to single factor ANOVA (look at p-value for Color). Here
tradeoff between df and partitioning of variance makes for a more
powerful test
Be careful
15
Any hint of Interaction (site*color)? If not then how does this change our interpretation of results?
Govpt
BoatSta
ir
Hazard
s
Shell Beach
Cayuco
sPSN
SITE
1.0
1.5
2.0
2.5
3.0
3.5
LN
UM
BE
R
PurpleOrange
COLOR
If factor A fixed and factor B (Blocks) random:
MSA 2 + a2 + n (ai)
2/p-1
MSBlocks 2 + n2
MSResidual 2 + a2
Model 3: Randomized Block Design - using
(restricted) Maximum Likelihood Estimation
• Accounts for site specific effects
1) Variance component used to calculate percent of variance
associated with the random effect
2) P-value for Color is identical to that from the Least Squares
Estimation (this will always be true for balanced designs)
Identical to least squares
solution
16
Model 3: Mixed Model Solution
• Also accounts for site specific
effects
Identical to least squares
solution and REML
Examples of randomized block
designs • Effect of feeding time (pre, post) on metabolic rate in otters. Each
otter is measured twice (pre post). Hence otter ID is the random effect
unless????
• Effect of Health Care reform on percentage of insured people in
counties of CA. Each county is measured twice (pre post). Hence
county is the random effect unless??
• Effect of watering regime (0,1,2,4,6 times weekly in replicate plots).
Each treatment (ttt) is in each of 10 plots. Plots are random effect.
• Effect of gender on grades in replicated classrooms. Grades for males
and females are measured in each of 20 classrooms. Classrooms
(teachers) are a random effect unless??
17
Full set of slides
18
Introduction to Randomized
block designs
Accounting for predicted but random
variance
A
B C
D B
C A
D C
B D
A B
D C
A
Block 1 Block 2 Block 3 Block 4
Blocking
• Aim:
– Reduce unexplained variation, without increasing size of experiment.
Approach:
– Group experimental units (“replicates”) into blocks.
– Blocks usually spatial units, one experimental unit from each treatment in each block.
19
Walter & O’Dowd (1992)
• Effects of domatia (cavities
on leaves) on number of
mites
• Two treatments (Factor A):
– shaving domatia removes
domatia from leaves
– normal domatia as control
• Required 14 leaves for
each treatment
Control leaves Shaved domatia leaves
Completely randomized design: - 28 leaves randomly allocated to each of 2
treatments
20
Completely randomized ANOVA • Factor A with p groups (p = 2 for domatia)
• n replicates within each group (n = 14 pairs of
leaves)
Source general df example df
Factor A p-1 1
Residual p(n-1) 26
Total pn-1 27
Walter & O’Dowd (1992)
• Required 14 leaves for each treatment
• Set up as blocked design
– paired leaves (14 pairs) chosen
– 1 leaf in each pair shaved, 1 leaf in each
pair control
21
1 block
Control leaves Shaved domatia leaves
Rationale for blocking
• Micro-temperature, humidity, leaf age,
etc. more similar within block than
between blocks
• Variation in response variable (mite
number) between leaves within block
(leaf pair) < variation between leaves
between blocks
22
Rationale for blocking
• Some of unexplained (residual)
variation in response variable from
completely randomized design now
explained by differences between
blocks
• More precise estimate of treatment
effects than if leaves were chosen
completely randomly
Null hypotheses
• No main effect of Factor A
– H0: m1 = m2 = … = mi = ... = m
– H0: a1 = a2 = … = ai = ... = 0 (ai = mi - m)
– no effect of shaving domatia, pooling
blocks
• Factor A usually fixed
23
Null hypotheses
• No effect of factor B (blocks):
– no difference between blocks (leaf pairs),
pooling treatments
• Blocks usually random factor:
– sample of blocks from population of blocks
– H0: 2 = 0
• Factor A with p groups (p = 2 treatments for
domatia)
• Factor B with q blocks (q = 14 pairs of leaves)
Source general example
Factor A p-1 1
Factor B (blocks) q-1 13
Residual (p-1)(q-1) 13
Total pq-1 27
Randomized blocks ANOVA
Notice
that this
is not
pq(n-1)
24
Randomized block ANOVA
• Randomized block ANOVA is 2 factor
factorial design
– BUT no replicates (n) within each cell
(treatment-block combination), i.e.
unreplicated 2 factor design
– No measure of within-cell variation
– No test for treatment by block interaction
Example – effect of watering on
plant growth • Factor 1: Watering, no watering
• Factor 2: Blocks (1-4). One replicate of each treatment
(watering, no watering) in each of 4 plots
• Replication: 1 plant for each watering/black combination
(8 total)
No water
Water
25
Treatment
No Water Water
G r o
w t h
0
5
10
15
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
Results
Block
1.0 2.0 3.0 4.0
G r o
w t h
0.0
5.0
10.0
15.0
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
Results
26
Treatment
No Water
Water
Block
1.0 2.0 3.0 4.0
G r o
w t h
0
5
10
15
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
Results
If factor A fixed and factor B (Blocks)
random:
MSwatering
MSBlocks
MSResidual
Expected mean squares
27
Treatment
No Water
Water
Block
1.0 2.0 3.0 4.0
G r o
w t h
0
5
10
15
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
EMSResidual = 2 + a2 : Why is this not simply 2
Residual
• Cannot separately estimate 2 and a2:
– no replicates within each block-treatment
combination
• MSResidual estimates 2 + a2
28
If factor A fixed and factor B (Blocks)
random:
MSwatering
MSBlocks
MSResidual 2 + a2
Expected mean squares
Block
1.0 2.0 3.0 4.0
G r o
w t h
0.0
5.0
10.0
15.0
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
EMSBlocks =2 + n2
29
If factor A fixed and factor B (Blocks)
random:
MSwatering
MSBlocks 2 + n2
MSResidual 2 + a2
Expected mean squares
Treatment
No Water Water
G r o
w t h
0
5
10
15
Block Treatment Growth
1 No Water 6
1 Water 10
2 No Water 4
2 Water 6
3 No Water 11
3 Water 15
4 No Water 5
4 Water 8
EMSwatering= 2 + a2 + n (ai)
2/p-1
Why does the EMS for watering include a2 , which is the effect of the interaction?
Block is a random effect, hence there are unsampled combinations of block and
watering that could affect the estimates of EMSwatering
30
General Randomized Block
Mean Square calcualtion
If factor A fixed and factor B (Blocks)
random:
MSwatering 2 + a2 + n (ai)
2/p-1
MSBlocks 2 + n2
MSResidual 2 + a2
Expected mean squares
31
Testing null hypotheses
• Factor A fixed and blocks random
• If H0 no effects of factor A is true:
– then F-ratio MSA / MSResidual 1
• If H0 no variance among blocks is true:
– no F-ratio for test unless no interaction
assumed
– if blocks fixed, then F-ratio MSB / MSResidual
1
Walter & O’Dowd (1992)
• Factor A (treatment - shaved and
unshaved domatia) - fixed
• Blocks (14 pairs of leaves) - random
Source df MS F P
Treatment 1 31.34 11.32 0.005
Block 13 1.77 0.64 0.784 ??
Residual 13 2.77
Should this be reported??
32
Explanation Blocks
Treatment 1 2 3 4 5 6 7 8 9 10 11 12 13
Shaved 1 1 1 1 1 1 1 1 1 1 1 1 1
Control 1 1 1 1 1 1 1 1 1 1 1 1 1
Cells represent the possible effect of the Block by Treatment
interaction but:
1) There is only one replicate per cell, therefore
2) No way to estimate variance term for each cell, therefore
3) No way to estimate the variance associated with the interaction,
therefore
4) The residual term estimates 2 + a2
Randomized Block vs
Completely Randomized designs
• Total number of experimental units
same in both designs
– 28 leaves in total for domatia experiment
• Test of factor A (treatments) has fewer
df in block design:
– reduced power of test
33
RCB vs CR designs
• MSResidual smaller in block design if blocks explain some of variation in Y:
– increased power of test
• If decrease in MSResidual (unexplained variation) outweighs loss of df, then block design is better:
– when blocks explain much of variation in Y
Assumptions
• Normality of response variable
– boxplots etc.
• No interaction between blocks and
factor A, otherwise
– MSResidual increase proportionally more than
MSA with reduced power of F-ratio test for
A (treatments)
– interpretation of main effects may be
difficult, just like replicated factorial ANOVA
34
Checks for interaction
• No real test because no within-cell
variation measured
• Tukey’s test for non-additivity:
– detect some forms of interaction
• Plot treatment values against block
(“interaction plot”)
Interaction plots
Y
Block
Y
No interaction
Interaction
35
Growth of Plantago
Growth of Plantago
• Growth of five genotypes (3 fast, 2 slow) of Plantago major (ribwort)
• Poorter et al. (1990)
• One replicate seedling of each genotype placed in each of 7 plastic containers in growth chamber – Genotypes (1, 2, 3, 4, 5) are factor A
– Containers (1 to 7) are blocks
– Response variable is total plant weight (g) after 12 days
36
Poorter et al. (1990)
1
2 3 4
5
1 2
3 4 5
Container 1 Container 2
Similarly for containers 3, 4, 5, 6 and 7
Source df MS F P
Genotype 4 0.125 3.81 0.016
Block 6 0.118
Residual 24 0.033
Total 34
Conclusions:
• Large variation between containers (= blocks) so
block design probably better than completely
randomized design
• Significant difference in growth between
genotypes
37
Mussel recruitment and seastars
Mussel recruitment and seastars
• Effect of increased mussel (Mytilus spp.) recruitment on seastar numbers
• Robles et al. (1995) – Two treatments: 30-40L of Mytilus (0.5-3.5cm
long) added, no Mytilus added
– Four matched pairs (blocks) of mussel beds chosen
– Treatments randomly assigned to mussel beds within pair
– Response variable % change in seastar numbers
38
mussel bed with added mussels
mussel bed without added mussels
+ -
- +
- +
+ -
1 block (pair of mussel beds)
+
-
Source df MS F P
Blocks 3 62.82
Treatment 1 5237.21 45.50 0.007
Residual 3 115.09
Conclusions:
• Relatively little variation between blocks so a
completely randomized design probably better
because treatments would have 1,6 df
• Significant treatment effect - more seastars
where mussels added
39
Worked Example – seastar colors
• Comparison of numbers of purple vs orange
seastars along the CA coast
• Data number of purple and orange seastars
collected at 7 random locations
• Compare models (block vs completely
random vs paired t test)
sea star colors all sites two sample
Number of Seastars as a function
of color and site
Any obvious problem with the data??
40
Diagnostics: Log transform helps normality and
homogeneity of variance assumptions
Model 1: One factor ANOVA
Why the difference?
SE=0.184
SE=0.181
41
Model 2: Paired t test
• Accounts for site specific (block) differences
• But no way to assess site (block) differences
LORANGE LPURPLE
Index of Case
1.0
1.5
2.0
2.5
3.0
3.5
Valu
e
Model 3: Randomized Block Design -
using least squares
• Accounts for and assesses (with a
caveat) site specific effects
1) Compare to paired t (same p value for Color) but no Site effect
2) Compare to single factor ANOVA (look at p-value for Color). Here
tradeoff between df and partitioning of variance makes for a more
powerful test
Be careful
42
Any hint of Interaction (site*color)? If not then how does this change our interpretation of results?
Govpt
BoatSta
ir
Hazard
s
Shell Beach
Cayuco
sPSN
SITE
1.0
1.5
2.0
2.5
3.0
3.5
LN
UM
BE
R
PurpleOrange
COLOR
If factor A fixed and factor B (Blocks) random:
MSA 2 + a2 + n (ai)
2/p-1
MSBlocks 2 + n2
MSResidual 2 + a2
Model 3: Randomized Block Design - using
(restricted) Maximum Likelihood Estimation
• Accounts for site specific effects
1) Variance component used to calculate percent of variance
associated with the random effect
2) P-value for Color is identical to that from the Least Squares
Estimation (this will always be true for balanced designs)
Identical to least squares
solution
43
Model 3: Mixed Model Solution
• Also accounts for site specific
effects
Identical to least squares
solution and REML
Examples of randomized block
designs • Effect of feeding time (pre, post) on metabolic rate in otters. Each
otter is measured twice (pre post). Hence otter ID is the random effect
unless????
• Effect of Health Care reform on percentage of insured people in
counties of CA. Each county is measured twice (pre post). Hence
county is the random effect unless??
• Effect of watering regime (0,1,2,4,6 times weekly in replicate plots).
Each treatment (ttt) is in each of 10 plots. Plots are random effect.
• Effect of gender on grades in replicated classrooms. Grades for males
and females are measured in each of 20 classrooms. Classrooms
(teachers) are a random effect unless??
44
Sphericity assumption
This is for reference – much more
important for repeated measures
Block Treat 1 Treat 2 Treat 3 etc.
1 y11 y21 y31
2 y12 y22 y32
3 y13 y23 y33
etc.
45
Block T1 - T2 T2 - T3 T1 - T3 etc.
1 y11-y21 y21-y31 y11-y31
2 y12-y22 y22-y32 y12-y32
3 y13-y23 y23-y33 y13-y33
etc.
Sphericity assumption
• Pattern of variances and covariances within
and between treatments:
– sphericity of variance-covariance matrix
• Equal variances of differences between all
pairs of treatments :
– variance of (T1 - T2)’s = variance of (T2 - T3)’s =
variance of (T1 - T3)’s etc.
• If assumption not met:
– F-ratio test produces too many Type I errors
46
Sphericity assumption
• Applies to randomized block
– also repeated measures designs
• Epsilon (e) statistic indicates degree to which
sphericity is not met
– further e is from 1, more variances of treatment
differences are different
• Two versions of e
– Greenhouse-Geisser e
– Huyhn-Feldt e
Dealing with non-sphericity
If e not close to 1 and sphericity not met,
there are 2 approaches:
– Adjusted ANOVA F-tests
• df for F-ratio tests from ANOVA adjusted
downwards (made more conservative)
depending on value e
– Multivariate ANOVA (MANOVA)
• treatments considered as multiple
response variables in MANOVA
47
Sphericity assumption
• Assumption of sphericity probably OK for randomized block designs:
– treatments randomly applied to experimental units within blocks
• Assumption of sphericity probably also OK for repeated measures designs:
– if order each “subject” receives each treatment is randomized (eg. rats and drugs)
Sphericity assumption
• Assumption of sphericity probably not OK for repeated measures designs involving time:
– because response variable for times closer together more correlated than for times further apart
– sphericity unlikely to be met
– use Greenhouse-Geisser adjusted tests or MANOVA