Chapter 7: Analysis Of Variance

(ANOVA)
• Completely randomised design Test at 1% for the significant difference amongst the treatment
1 2 3 means
11 6, 5 
2
8,2 7,7 6,9  2
Y 920.65
ij
G
CF  
2
9 0 4, 8 1 6 7
8,7 8,4 5,8 N 15
SSTr  Rn  CF 917,3305  904,8167 12,5138
2
9,4 8,6 7,2
9,2 8,1 6,8 SST  Yij2  CF 920, 65  904,8167 15,8333
8,0 7,4
6,1
𝑆 𝑆 𝐸=𝑆𝑆𝑇 − 𝑆𝑆𝑇𝑟 =15.8333−12.5138=3.3195
SSE  Y   920, 65  917,3305 3,3195
2
ij
R2
n

T Observations ni Ri Ri2
n
1 8,2 8,7 9,4 9,2 4 35,5 315,0625
2 7,7 8,4 8,6 8,1 8,0 5 40,8 332,928
3 6,9 5,8 7,2 6,8 7,4 6,1 6 40,2 269,34
15 116,5 917,3305
ANOVA
Source SS df MS F Ftab (0.01)

Treatments 12,5138 2 6,2569 22,6208 6,93

Errors 3,3195 12 0,2766
Total 15,8333 14
 H 0 : 1 2 3 H a : at least two of the treatment means differs

Since 22,6208 ≥ 6,93, we reject the null hypothesis at 1% level of significance and conclude that at least two of the
treatment means differs.
1 1
1 1  CI 1  3  Y1  Y3  tne;1  MSE   
CI 1   2  Y1  Y2  tne;1  MSE   
 n1 n3 
2
2
 n1 n2 
1 1
8,875  8,16  t12;0.995
1 1
0, 2766    8,875  6, 7  t12;0.995 0, 2766   
 4 5  4 6
0, 715 3, 055 0,3528  2,175 3, 055 0,3395 
0, 715 1, 0778 2,175 1, 0372
  0,3628;1, 7928 1,1378;3, 2122 

Zero is included in the interval meaning mean 1 and Zero is not included in the interval meaning mean 1 and 3 do
2 don’t differ differ
 t distribution critical values
Lower tail probability (1  )
df 0.75 0.90 0.95 0.975 0.99 0.995 0.9975 0.999
1 1.000 3.078 6.314 12.706 31.821 63.657 127.321 318.309
2 .816 1.886 2.920 4.303 6.965 9.925 14.089 22.327
3 .765 1.638 2.353 3.182 4.541 5.841 7.453 10.214
4 .741 1.533 2.132 2.776 3.747 4.604 5.598 7.173
5 .727 1.476 2.015 2.571 3.365 4.032 4.773 5.893
6 .718 1.440 1.943 2.447 3.143 3.707 4.317 5.208
7 .711 1.415 1.895 2.365 2.998 3.499 4.029 4.785
8 .706 1.397 1.860 2.306 2.896 3.355 3.833 4.501
9 .703 1.383 1.833 2.262 2.821 3.250 3.690 4.297
10 .700 1.372 1.812 2.228 2.764 3.169 3.581 4.144
11 .697 1.363 1.796 2.201 2.718 3.106 3.497 4.025
12 .695 1.356 1.782 2.179 2.681 3.055 3.428 3.930
13 .694 1.350 1.771 2.160 2.650 3.012 3.372 3.852
14 .692 1.345 1.761 2.145 2.624 2.977 3.326 3.787
15 .691 1.341 1.753 2.131 2.602 2.947 3.286 3.733
16 .690 1.337 1.746 2.120 2.583 2.921 3.252 3.686
17 .689 1.333 1.740 2.110 2.567 2.898 3.223 3.646
18 .688 1.330 1.734 2.101 2.552 2.878 3.197 3.610
19 .688 1.328 1.729 2.093 2.539 2.861 3.174 3.579
20 .687 1.325 1.725 2.086 2.528 2.845 3.153 3.552
21 686 1.323 1.721 2.080 2.518 2.831 3.135 3.527
22 .686 1.321 1.717 2.074 2.508 2.819 3.119 3.505
23 .685 1.319 1.714 2.069 2.500 2.807 3.104 3.485
24 .685 1.318 1.711 2.064 2.492 2.797 3.090 3.467
25 .684 1.316 1.708 2.060 2.485 2.787 3.978 3.450
  1 1
CI 2  3  Y2  Y3  t ne;1  MSE   
 n2 n3 
2

1 1
8,16  6, 7  t12;0.995 0, 2766   
 5 6
1, 46 3, 055 0,3185 
1, 46 0,9730
 0, 487; 2, 433

Zero is not included in the interval meaning mean 2 and 3 do differ.

 Observations ni Ri Ri2
ni

1 6,9 5,8 6,2 6,8 6,1 5 31,8 202,248

2 7,8 8,4 8,7 9,6 9,2 8,7 6 52,4 457,6267
3 6,8 7,3 6,3 6,9 4 27,3 186,3225
Total 15 111,5 846,1972
111,5
2

Y 2
ij 849,55 CF 
15
828,8167

SSTr 846,1972  828,8167 17,3805

SST 849,55  828,8167 20, 7333

𝑆𝑆𝐸=20.7333 −17.3805=3.3528
ANOVA
Source SS df MS F Ftab (0.01)
Treatments 17,3805 2 8,6903 31,1034 6,93
Errors 3,3528 12 0,2794
Total 20,7333 14

H 0 : 1  2 3 H a : at least two of the treatment means differs

Since 31,1034 > 6,93, we reject the null hypothesis at 1% level of significance and conclude that at least two of the treatment means
differs.
 1 1 
CI 1   2  Y1  Y2  t ne;1  MSE   
2
 n1 n2 
1 1
6,36  8, 7333 t12;0.995 0, 2794   
 5 6
 2,3733 3, 055 0,3201
 2,3733 0,9779
  3,3512;  1,3954

Zero is not included in the interval, mean 1 and mean 2 differs.

1 1
CI 1  3  Y1  Y3  t ne;1  MSE   
 n1 n3 
2

1 1
6,36  6,825  t12;0.995 0, 2794   
 5 4
 0, 465 3, 055 0,3546 
 0, 465 1, 0833
  1,5483;0, 6183

Zero is included in the interval, mean 1 and mean 3 do not differ.

 t distribution critical values
Lower tail probability (1  )
df 0.75 0.90 0.95 0.975 0.99 0.995 0.9975 0.999
1 1.000 3.078 6.314 12.706 31.821 63.657 127.321 318.309
2 .816 1.886 2.920 4.303 6.965 9.925 14.089 22.327
3 .765 1.638 2.353 3.182 4.541 5.841 7.453 10.214
4 .741 1.533 2.132 2.776 3.747 4.604 5.598 7.173
5 .727 1.476 2.015 2.571 3.365 4.032 4.773 5.893
6 .718 1.440 1.943 2.447 3.143 3.707 4.317 5.208
7 .711 1.415 1.895 2.365 2.998 3.499 4.029 4.785
8 .706 1.397 1.860 2.306 2.896 3.355 3.833 4.501
9 .703 1.383 1.833 2.262 2.821 3.250 3.690 4.297
10 .700 1.372 1.812 2.228 2.764 3.169 3.581 4.144
11 .697 1.363 1.796 2.201 2.718 3.106 3.497 4.025
12 .695 1.356 1.782 2.179 2.681 3.055 3.428 3.930
13 .694 1.350 1.771 2.160 2.650 3.012 3.372 3.852
14 .692 1.345 1.761 2.145 2.624 2.977 3.326 3.787
15 .691 1.341 1.753 2.131 2.602 2.947 3.286 3.733
16 .690 1.337 1.746 2.120 2.583 2.921 3.252 3.686
17 .689 1.333 1.740 2.110 2.567 2.898 3.223 3.646
18 .688 1.330 1.734 2.101 2.552 2.878 3.197 3.610
19 .688 1.328 1.729 2.093 2.539 2.861 3.174 3.579
20 .687 1.325 1.725 2.086 2.528 2.845 3.153 3.552
21 686 1.323 1.721 2.080 2.518 2.831 3.135 3.527
22 .686 1.321 1.717 2.074 2.508 2.819 3.119 3.505
23 .685 1.319 1.714 2.069 2.500 2.807 3.104 3.485
24 .685 1.318 1.711 2.064 2.492 2.797 3.090 3.467
25 .684 1.316 1.708 2.060 2.485 2.787 3.978 3.450
  1 1
CI 2  3  Y2  Y3  tne;1  MSE   
 n2 n3 
2

1 1
8, 7333  6,825  t12;0.995 0, 2794   
 6 4
1,9083 3, 055 0,3412 
1,9083 1, 0424
 0,8659; 2,9507 

Zero is not included in the interval, mean 2 and mean 3 differ.

Block design
TRM 1 2 3 4 Ri2
b Ri
b

1 70 75 68 81 4 294 21 609
2 66 59 55 63 4 243 14 762.25
3 59 66 39 42 4 206 10 609
4 41 57 39 55 4 192 9 216
a 4 4 4 4 16 56 196,25
Cj
236 257 201 241 935
C 2j 13 924 16 512,25 10 100,25 14 520,25 55 056,75
a

Test at 5% level of significance for the difference between the treatment and block means.
935   Yij2  57 139
2
CF 
16 54 639,0625 all

Ri2
SSTr   CF  56 196,25 – 54 639,0625 = 1 557,1875
b
C 2j
SSB   CF = 55 056,25 – 54 639,25 = 417,6875
a

SST  Yij2  CF = 57 139 – 54 639,0625 = 2 499,9375

all

SSE  SST  SSTr  SSB = 2 499,9375 – 417,6875 – 1 557,875 = 525,0625

 ANOVA
SOURCE SS df MS F Ftab (0.05)

Treatments 1 557,1875 3 519,2917 8,9011 3.86

Blocks 417,6875 3 139,2292 2,3865 3.86
Errors 525,0625 9 58,3403
Total 2 499,9375 15

H0a: the treatment means are the same, Haa: not the null hypothesis.

H0b: the block means are the same, Hab: not the null hypothesis.
 TRM Ri2
1 2 3 b Ri
b

1 11,3 12,4 10,9 3 34,6 399,0533

2 14,2 14,3 13,7 3 42,2 593,6133
3 14,9 15,2 17,7 3 47,8 761,6133
4 12,2 11,3 12,5 3 36 432
a
4 4 4 12 2 186,2799
Cj

C 2j 52,6 53,2 54,8 160,6

a
691,69 707,56 750,76 2 150,01
Test at 5% level of significance for the difference between the treatment and block means

160,6 
2
Y 2
 2 193,2
CF 
12
 2 149,3633 all
ij

Ri2
SSTr   CF  2 186,2799 – 2 149,3633 = 36,9166
b

C 2j
SSB   CF = 2 150,01 – 2 149,3633 = 0,6467
a

SST  Yij2  CF
all
= 2 193,2 – 2 149,3633 = 43,8367
SSE  SST  SSTr  SSB = 36,9166 – 36,9166 – 0,6467 = 6,2734
 ANOVA
SOURCE SS df MS F Ftab (0.05)

Treatments 36,9166 3 12.3055 11,7688 4.76

Blocks 0,6467 2 0,3234 0,3093 5.14

Errors 6,2734 6 1,0456

Total 43,8367 11
 Factorial design
Factor B
Factor A 1 2 3
1 28,0 29,0 28,5 30,0 31,0 29,5 27,0 28,5 29,0

2 23,0 23,5 24,0 26,0 24,0 23,5 22,0 22,5 23,0

Factor B
472 
2
2
Factor A 1 2 3 Ri Ri CF   12 376.8889
nb 18

1 85.5 90.5 84.5 260.5 7 540.0278

Y 2
ij 
12 531,5
2 70.5 73.5 67.5 211.5 4 970.25 all

Cj 156 164 152 472

12 510.2778
2
C j
na 4 056 4 482.6667 3 850.6667 12 389.3334

1
SS sub 
3
 
85.82  90.52  84.52  70.52  73.52  67.52  CF = 12 523.1667 – 12 376.8889 = 146.2778 (ab-1)
Ri
SS a   CF = 12 510.2778 -12 376.8889 = 133,3889 (a-1)
nb

Ci
SSb   CF
na
 472 
2
Factor A 1 2 3
CF   12 376.8889
18
1 85.5 90.5 84.5 260.5 7 540.0278
2 70.5 73.5 67.5 211.5 4 970.25 Y
all
2
ij  12 531,5
156 164 152 472
12 510.2778
4 056 4 482.6667 3 850.6667 12 389.3334

1
SS sub 
3
 
85.82  90.52  84.52  70.52  73.5 2  67.5 2  CF = 12 523.1667 – 12 376.8889 = 146.2778 (ab-1)

Ri
SS a   CF
nb = 12 510.2778 -12 376.8889 = 133,3889 (a-1)

Ci
SSb   CF = 12 389.3334 – 12 376.8889 = 12,4445 (b-1)
na

SS ab SS sub  SS a  SSb

= 146.2778 – 133. 3889- 12.4445 = 0,4444 [(a-1)(b-1)]

SST  Yij2  CF = 12 531.5 – 12 376.8889 = 154.6111 (N-1)

all

SSE SST  SSa  SSb  =SS154.6111

ab – 133.3889 – 12.4445 – 0.4444 = 8.3333 [ab (n-1)]
 ANOVA
SOURCE SS df MS F Ftab (0.01)
SUB 146.2778 5 29.2556 42.1308 5.06
A 133.3889 1 133.3889 192.0923 9.33
B 12.4445 2 6.2223 8.9607 6.93
AB 0.4444 2 0.2222 0.3200 6.93
Errors 8.3333 12 0.6944
Total 154.6111 17

H0a: a1 = a2 H0a’ : not the null hypothesis

H0b: b1 = b2 = b3 H0b’ : not the null hypothesis

H0ab: there is no interaction between factor A and factor B

H0ab’: there is an interaction between factor A and factor B

 1 2 3 4 N = 24 n = 2;
1 139.3 148.9 150.2 143.0 132.1 124.4 129.7 118.3
Factor A 2 150.4 137.8 156.5 131.6 134.4 117.8 119.2 131.7 a=3 b=4
3 131.7 133.1 141.7 125.4 122.2 114.9 127.3 125.3
3186.9 
2

CF   423 180.4838
R 2 24
1 2 3 4 Rj j
nb

1 288.2 293.2 256.5 248 1 085.9 147 397.3513

Y 2
ij  426 145.81
Fac A 2 288.2 288.1 252.2 250.9 1 079.4 145 638.045 all

3 264.8 267.1 237.1 252.6 1 021.6 130 458.32

423 493.7163
Cj
841.2 848.4 745.8 751.5 3 186.9 (G)
C 2j 117 936.24 119 963.76 92 702.94 94 125.375 424 728.315
na

1
SS sub 
2
 
288.22  293.2 2  256.52  248 2  288.2 2  288.12  .....  252.6 2  CF = 425 211.425 – 423 180.4838 = 2 030.9412
R 2i
SS a   CF = 423 493.7163 -423 180.4838 = 313.2325 (a-1)
nb

C 2i
SSb   CF
na = 424 728.315 – 423 180.4838 = 1 547.8312 (b-1)
SS ab SS sub  SS a  SS b = 2 030.9412 – 313.2325 – 1 547.8312 = 169.8775 (a-1) (b-1)
SST  Yij2  CF = 426 145.81 – 423 180.4838 = 2 965.3262 [abn-1]
all

SSE SST  SSa  SSb  SSab = 2 965.3262 – 313.2325–1547.8312– 169.8775= 934.385 [ab (n-1)]
 ANOVA
SOURCE SS df MS F Ftab (0.01)

A 313.2325 2 156.6163 2.0114 6.93

B 1 547.8312 3 515.9437 6.6261 5.95
AB 169.8775 6 28.3129 0.3636 4.82
Errors 934.385 12 77.8654
Total 2 965.3262 23

H0a: a1 = a2 H0a’ : not the null hypothesis

H0b: b1 = b2 = b3 H0b’ : not the null hypothesis

H0ab: there is no interaction between factor A and factor B

H0ab’: there is an interaction between factor A and factor B