Two-Way Anova with m

UNIT 8 TWO-WAY ANOVA WITH m Observations Per Cell

OBSERVATIONS PER CELL

Structure
8.1 Introduction
Objectives
8.2 ANOVA Model for Two-way Classified Data with m Observations per
Cell
8.3 Basic Assumptions
8.4 Estimation of Parameters
8.5 Test of Hypothesis
8.6 Degrees of Freedom of Various Sum of Squares
8.7 Expectations of Various Sum of Squares
8.8 Summary
8.9 Solutions/Answers

8.1 INTRODUCTION

In the analysis of variance technique, if explanatory variable is only one and

different levels of independent variable is under consideration then it is called
one-way analysis of variance and a test of hypothesis is developed for the
equality of several mean of different levels of a factor/independent variable/
explanatory variable. But if we are interested to consider two independent
variables for analysis in place of one, and able to perform the two hypotheses
for the levels of these factors independently (there is no interaction between
these two factors). The above analysis has been given in the Units 6 and 7
respectively. But if we are interested to test the interaction between two
factors and we have repeated observations then the two-way analysis of
variance with m observation per cell is considered. If there are exactly same
numbers of observations in the cell then it is called balance.
In this unit, a mathematical model for two-way classified data with m-
observations per cell is given in Section 8.2. The basic assumptions are given
in Section 8.3 whereas the estimation of parameters is given in Section 8.4.
Test of hypothesis for two-way ANOVA is explained in Section 8.5 and
degrees of freedom of various sum of squares are described in Section 8.6.
The expected values of sum of squares for two factors and their interactions
are derived in Section 8.7.
Objectives
After studying this unit, you would be able to
 describe the ANOVA model for two-way classified data with m
observations per cell;
 describe the basic assumptions for the given model;
 obtain the estimates of the parameters of the given model;

61
Analysis of Variance
 describe the test of hypothesis for two-way classified data with m
observations per cell;
 derive the expectations of the various sum of squares; and
 perform to test the hypothesis for two-way classified data with m
observations per cell.

8.2 ANOVA MODEL FOR TWO-WAY CLASSIFIED

DATA WITH m OBSERVATIONS PER CELL

In Unit 7, it was seen that we cannot obtain an estimate of, or make a test for
the interaction effect in the case of two-way classified data with one
observation per cell. This is possible, however, if some or all of the cells
contain more than one observations. We shall assume that there is an equal
number of (m) observations in each cell. The m observations in the (i, j)th cell
will be denoted yij1, yij2, . . ., yijm. Thus, yijk is the kth observation for ith level of
factor A and jth level of factor B, i = 1, 2, …, p; j = 1, 2, …,q & k = 1, 2, …,
m.
The mathematical model
yijk = µij + eijk
where µij is the true value for the (i, j)th cell and eijk is the error. eijk are
assumed to be independently identical normally distributed, each with mean
zero and variance σe2. The table of observations can be displayed as follows:
A/B B1 B2 Bj … Bq Total Total
A1 y111 y121 y1j1 … y1q1 y1.1 y1..
y112 y122 y1j2 … y1q2 y1.2
. . . . .
. . . . .
. . . . .
y11m y12m y1jm … y1qm y1.m
A2 y211 y221 y2j1 … y2q1 y2.1 y2..
y212 y222 y2j2 … y2q2 y2.2
. . . . .
. . . . .
. . . . .
y21m y22m y2jm … Y2qm y2.m
. yi11 yi21 yij1 … yiq1 yi.1 yi..
. yi12 yi22 yij2 … yiq2 yi.2
. . . . … . .
Ai . . . … . .
. . . . … . .
. yi1m yi2m yijm … yiqm yi.m
Ap yp11 yp21 ypj1 … ypq1 yp.1 yp..
yp12 yp22 ypj2 … ypq2 yp.2
. . . . .
. . . . .
. . . . .
yp1m yp2m ypjm … ypqm yp.m
62 Total y.1. y.2. y.j. … y.q. y…
The model can be written as Two-Way Anova with m
Observations Per Cell
yijk = µ + (µi.-µ) + (µ.j-µ) + (µij-µi.-µ.j+µ) + eijk
= µ + αi + βj + (αβ)ij + eijk
where, µ is general mean effect, αi is the effect of ith level of the factor A,  j is
the effect of jth level of factor B, (αβ)ij is the interaction effect between ith level
of A factor and jth level of B factor.
p q p q

α i  0,  β j  0,  αβ ij  0,  αβ ij  0

i 1 j1 i 1 j1

where,
y... = Sum of all the observations.
yi.. = Total of all observations in the ith level of factor A
y.j. = Total of all observations in the jth level of factor B.

8.3 BASIC ASSUMPTIONS

Following assumptions should be followed for valid and reliable test

procedure for testing of hypothesis as well as for estimation of parameters
1. All the observations yijk are independent.
2. Different effects are additive in nature.
3. e ijk are independent and identicaly distributed as normal with mean zero
and constant variance  2e .

8.4 ESTIMATION OF PARAMETERS

The least square estimates for various effects, obtained by minimizing the
residual sum of squares

p q m 2

E  y
i1 j1 k 1
ijk     i   j   ij 
by partially differentiating E with respect to μ, αi (i=1, 2, ..., p), βj (i=1, 2, ...,
q) and (αβ)ij for all i = 1, 2, ..., p; j =1, 2, … , q and equating these equations
equal to zero. These equations are called normal equations. Solution of these
normal equations provide the estimates of these parameters [μ, αi, βj, (αβ)ij].

p q m
E


  2  y ijk     i   j   ij  0 
i1 j1 k 1

q m
E
 i

  2 y ijk     i   j   ij  0 
j1 k 1

63
Analysis of Variance
p m
E
 j

  2  y ijk     i   j   ij  0 
i 1 k 1

m
E
  ij

  2 y ijk     i   j   ij  0 
k 1

p q m


 y
i 1 j1 k 1
ijk

These equations give,    y...

pqm
q m


 y
j 1 k 1
. jk 
i     yi ..  y...
qm
p m

  y
i 1 k 1
i. k 
Similarly, j     y . j.  y ...
pm


( ) ij  y ij.  y i..  y. j.  y...

  
Substituting the values of i ,  i ,  j and ( α β̂ )ij, in the model and then
select the value of eijk such that both the sides are equal, so
yijk  y...   yi..  y...   y. j.  y...   y ij.  y i..  y. j.  y...   yijk  y ij. 

or

yijk  y...  yi..  y...   y. j.  y...   y ij.  yi..  y. j.  y...   yijk  yij. 

Squaring and summing both the sides over i, j & k, then we get
p q m p q

  y  y...   mq  yi..  y...   mp  y. j.  y... 

2 2 2
ijk
i 1 j1 k 1 i 1 j1
p q p q m
 m  yij.  y i..  y. j.  y...    y ijk  y ij. 
2 2

i 1 j1 i 1 j1 k 1

as usual product terms vanish.

Total Sum of Squares = Sum of Squares due to Factor A+ Sum of Squares due
to Factor B + Sum of Squares due to Interaction A and
B + Sum of Squares due to Error
or TSS = SSA + SSB + SSAB + SSE

8.5 TEST OF HYPOTHESIS

There are three hypotheses which are to be tested are as follows:
H0A: α1 = α2 = . . . = αp = 0
H1A: α1 ≠ α 2 ≠ . . . ≠ α p ≠ 0

64 H0B: β1 = β 2 = . . . = βq = 0
H1B: β1 ≠ β2 ≠ . . . ≠ βq ≠ 0
H0AB: (αβ)ij = 0 for all i and j or A and B are independent to each other Two-Way Anova with m
Observations Per Cell
H1AB: (αβ)ij ≠ 0
The appropriate test statistics for testing the above hypothesis is:
SSA p  1 MSSA
F 
SSE pq m  1 MSSE

If this value of F is greater than the tabulated value of F with [(p-1), pq (m-
1)] df at α level of significance so we reject the null hypothesis, otherwise we
may accept the null hypothesis.
Similarly, test statistics for second and third hypotheses are
SSB q  1 MSSB
F 
SSE pqm  1 MSSE

SS AB p  1q  1 MSSAB
F 
SSE pqm  1 MSSE

For practical point of view, first we should decide whether or not H0AB can be
rejected at an appropriate level of significance by using above F. If interaction
effects are not significant i.e. the factor A and factor B are independent then
we can find the best level of A and best level of B by multiple comparison
method using t-test. On the other hand, if they are found to be significant,
there may not be a single level of factor A and single level of factor B that
will be the best in all situations. In this case, one will have to compare for
each level of B at the different levels of A and for each level of A at the
different levels of B.
The above analysis can be shown in the following ANOVA table:
ANOVA Table for Two-way Classified Data
with m Observations per Cell
Sourses of DF SS MSS F
Variation

Between p-1 p MSSA = F=

SSA  mq  y i..  y ... 
2
the levels SSA / (p-1) MSSA / MSSE
i 1
of A
Between q-1 q MSSB = F=
SSB  mp y . j.  y ... 
2
the levels SSB / (q-1) MSSB / MSSE
j1
of B
Interaction (p-1) SSAB  MSSAB = F=
AB (q-1) p q SSAB / MSS(AB) /
m yij.  y i..  y. j.  y... 
2 (p-1)(q-1) MSSE
i1 j1

Error pq (m-1) p q m MSSE =

TSS   y ijk  y ij. 
2
SSE / pq
i 1 j1 k 1
(m-1)

Total mpq-1

65
Analysis of Variance
Steps for Calculating Various Sums of Squares
p q m
1. Calculate G = Grand Total = Total of all observations =  y ijk
i 1 j1 k 1

2. Determine N = Numebr of observations.

3. Find Correction Factor (CF) = G2/N
p q m
2
4. Raw Sum of Squares ( RSS) =  y
i 1 j1 k 1
ijk

5. Total Sum of Squares (TSS) = RSS - CF

6. Sum of Squares due to Factor A (SSA)
= {y1..2 /mq + y2..2/mq + ... + yi..2/mq + ... + yp..2/mq} – CF
7. Sum of Squares due to Factor B (SSB)
= {y.1.2 /mp+y.2.2/mp + … +y.j.2/mp + ... + y.q.2/mp} – CF
8. Sum of Squares due to Means (SSM)
= {y..12 /pq + y..22/pq +… + y..k2/pq + ... + y..m2/pq} – CF
9. Sum of Squares due to Interation AB(SSAB) = SSM – SSA - SSB
10. Sum of Squares due to Error (SSE) = TSS -SSA-SSB-SSAB
11. Calculate MSSA = SSA/df
12. Calculate MSSB = SSB/df
13. Calculate MSSAB = SS(AB)/df
14. Calculate MSSE = SSE/df
15. Calculate FA = MSSA/MSSE ~ F (p-1), pq (m-1)
16. Calculate FB = MSSB/MSSE ~ F(q-1), pq (m-1)
17. Calculate FAB = MSS(AB)/MSSE ~ F((p-1)(q-1), pq (m-1)

8.6 DEGREES OF FREEDOM OF VARIOUS SUM

OF SQUARES

Total sum of squares (TSS) considers the pqm observations so the degrees of
freedom for TSS are (pqm-1). One degree of freedom is lost due to the
p q m
restriction that  ( y
i 1 j1 k 1
ijk  y... )  0 .

The degrees of freedom for sum of squares due to factor A is (p-1) because it
has p levels. Similarly, the degrees of freedom for sum of squares due to
factor B is (q-1) because it has q levels, under consideration. Sum of squares
due to interaction of factors A and B is (p-1) (q-1) and the degrees of
freedom for sum of squares due to errors is pq (m-1). Thus partitioning of
degrees of freedom is as follows:

(mpq-1) = (p-1) + (q-1) + (p-1)(q-1) + pq (m-1)

66
which implies that the df are additive.
 Two-Way Anova with m
8.7 EXPECTATIONS OF VARIOUS SUM OF Observations Per Cell
SQUARES
8.7.1 Expected Value of Sum of Squares due to Factor A
2
 p 
E (SSA) = E mq y i..  y...   
 i 1 
Substituting the value of yi.. and y... from the model, we get
2
 p 
 E mq  i  ei..  e...  
 i1 
p
 
or 
E (SSA) = E mq  i2  ei..  e... 2  2 i ei..  e...   
 i 1 
p
 
= mq E α i2  ei ..  e...   2α i ei ..  e... 
2

i 1

p
p 2 
= mq αi2  mqE ei..  e..    0
i1  i1 
Because Eei..  e..   0
p
 p 
 E (SSA )  mq  α i2  mq E  ei..   p e...2 
2

i1  i1 
p
p 
= mq   i2  mq  Eei.. 2  p E e...2 
i 1  i 1 

 p 2
p
2  2e 
= mq   mq  e
p
i 
i 1  i1 mq mpq 
p
= mq  2i  p  e2   2e
i 1

p
 E SSA   mq  2i  (p  1)  2e
i 1

 SSA  mq p 2
 E      i   2e
 p  1  ( p  1) i 1

mq p 2
or E MSSA    2e   i
( p  1) i 1

Under H0A the MSSA is an unbiased estimate of e2 .

8.7.2 Expected Value of Sum of Squares due to Factor B

Proceeding similarly, or by symmetry, we have
2
 q 
E(SSB) = E mp y . j.  y...   
 j1 
67
Analysis of Variance
Substituting the value of y. j. and y... from the model

2
 q 
we get 
E(SSB) = E mp  j  e. j.  e...  
 i1 

q
  
E(SSB) = mp β 2j  e. j.  e...   2β j e. j.  e... 
2
or
 i1  
q
q 2
or E(SSB) = mp β 2j  mp E  e. j.  e...    0
j1  j1 
q
q 
or E(SSB) = mp  β 2j  mp E  e. j.  2  q e...2 
j1  j1 
q
q 
= mp β 2j  mp  E e. j.  2  qE e...2 
j1  j1 
q
 q 2 2 
= mp 2j  mp  e  q e 
j1  j1 mp mpq 
q
= mp 2j  q  2e   e2
j1

q
E(SSB) = q  1 e2  mp   2j
j1

 SSB  mp q 2
  e2 
q  1 
E j
 q 1 j1

mp q 2
E(MSSB) =  e2 
q  1 
or j
j1

Under H 0 B the MSSB is an unbiased estimate of e2 . Similarly you can

obtain the expected value of SSAB , which will be
p q
2
E SSAB  m     p  1q  1 
ij
2
e
i 1 j1

p q
 SSAB  m
E     2e     2

 p  1q  1  p  1q  1 ij
i1 j1

p q
m 2
or E (MSSAB)   e2 
p  1q  1   
i 1 j1
ij

Under H0 AB , the mean sum of squares due to interaction between Factor A

and B is an unbiased estimate of e2 .
68
8.7.3 Expected Value of Sum of Squares due to Error Two-Way Anova with m
Observations Per Cell
Proceeding similarly, or by symmetry, we have
 p q m 
E (SSE) = E   y ijk  y ij.  2 
 i 1 j1 k 1 
Substituting the value of y ijk and y ij. from the model, we have

 p q m 
E(SSE) = E  e ijk  eij.  2 
 i1 j1 k 1 
p q  m 
= E     e ijk  eij. 2 
 i 1 j1  k 1 
 p q  m 2 
= E    e ijk  m e ij2. 
 i 1 j1  k 1 
 2
p q m p q

= E   eijk  m  eij2. 
 i1 j1 k 1 i 1 j1 
p q m p q
2
=   E eijk  
 m E eij2.  
i 1 j1 k 1 i 1 j1

= mpq  2e  mpq 2e / m

= mpq  1  e2

 SSE 
or E     2e
 pq ( m  1) 
or E (MSSE) = 2e

Hence, mean sum of squares due to error is an unbiased estimate of 2e .

Example 1: A manufacturer wishes to determine the effectiveness of four
types of machines (A, B, C and D) in the production of bolts. To accumulate
this, the numbers of defective bolts produced for each of two shifts in the
results are shown in the following table:

First shift Second Shift

Machine
M T W Th F M T W Th F
A 6 4 5 5 4 5 7 4 6 8
B 10 8 7 7 9 7 9 12 8 8
C 7 5 6 5 9 9 7 5 4 6
D 8 4 6 5 5 5 7 9 7 10

Perform an analysis of variance to determine at 5% level of significance,

whether there is a difference (a) Between the machines and (b) Between the
shifts.

69
Analysis of Variance
Solution: There are two factors the machine and shift. The levels of machine
are four and levels of shift are two. The Computation results are as follows:
G = 6+ 4 + 5 + 5 + 4 + 5 + 7 + 4 + 6 + 8 + 10 + 8 + 7 + 7 + 9 + 7
+ 9 + 12 + 8 + 8 + 7 + 5 + 6 + 5 + 9 + 9 + 7 + 5 + 4 + 6 + 8 + 4
+ 6 + 5 + 5 + 7 + 9 + 7 + 10
= 268
N = 40

G 2 268  268
CF =  1795.6
N 40
Raw Sum of Squares (RSS) = 62 + 42 +……+ 10 2 = 1946
Total Sum of Squares (TSS) = RSS - CF = 1946-1795.6 =150.4
Sum of square due to machines and due to shifts can be calculated by
considering the following two-way table:
Machine Shift Total
I Shift II Shift

A 24 30 54
B 41 44 85
C 32 31 63
D 28 38 66
Total 125 143 268

542 852 632 662

Sum of Squares due to Machine SSM      CF
10 10 10 10
= 1846.6  1795.6 = 51.0

1252 1432
Sum of Squares due to Shifts SSS    CF
20 20
= 1803.7 - 1795.6 = 8.1
Sum of Squares due to Interaction (SSMS)


242  412  322  282  302  442
5 5 5 5 5 5


31  38  CF  SSM   SSS 
2 2

5 5
= 1861.2  1795.6  51.0  8.1 = 6.5
Finaly, the Sum of Squares due to error is founded by subtracting the SSM,
SSS and SSSM from TSS
SSE = TSS- SSM  SSS  SSMS
70
= 150.4 -51.0 -8.1 -6.5 = 84.8
 SSM 51.0 Two-Way Anova with m
MSSM    17 Observations Per Cell
df 3
SSS 8.1
MSSS    8.1
df 1
SSE 84.8
MSSE    2.65
df 31
SSMS 6.5
MSSMS    2.167
df 3
For testing H 0 A : Mean effect of Machine A= Machine B =

Machine C = Machine D, is
17
F=  6.42
2.65
For testing H0B : Mean effect of Shift A = Shift B, is

8.1
F  3.06
2.65
Similarly, for testing H0AB : Interaction effect of Machine and Shift, is

2.167
F  0.817
2.65
ANOVA Table for Two-way Classified Data m- Observation per Cell
Sources of Degrees of Sum of Mean Sum F-test or
Variation Freedom Squares of Squares Variance
(DF) (SS) (MSS) Ratio

Due to 3 51.0 17 17
 6.42
Machinery 2.65
Due to Shift 1 8.1 8.1 8.1
 3.06
2.65
Due to 3 6.5 2.167 2.167
 0.817
Interaction 2.65
Due to Error 32 84.8 2.65
Total 39 150.4

The tabulated value of F at 3 and 32 degrees of freedom at 5% level of

significance is 2.90. The computed value of F for interaction is 0.817 so
the average performances in different shifts are not significant. There is a
significant difference among machines, since the calculated value of F for
machines is 6.42 and the critical value (tabulated value) of F is 2.90. The
tabulated value for shifts is 4.15. The calculated value of F for shifts is
3.06. Hence, there is no difference due to shifts.

71
Analysis of Variance
E1) An experiment is performed to determine the effect of two
advertising campaigns on three kinds of cake mixes. Sales of each
mix were recorded after the first advertising campaigns and then
after the second advertising campaign. This experiment was
repeated three times for each advertising campaign and got the
following results:

Campaign I Campaign II
Mix1 574, 564, 550 1092, 1086, 1065
Mix2 524, 573, 551 1028, 1073, 998
Mix3 576, 540, 592 1066, 1045, 1055

Perform an analysis of variance to determine at 5% level of significance,

whether there is a difference (a) Between the cake mixes and (b)
Between the campaigns.

8.8 SUMMARY
In this unit, we have discussed:
1. The ANOVA model for two-way classified data with m observations per
cell;
2. The basic assumptions for the given model;
3. How to obtain the estimates of the parameters of the given model;
4. How to test the hypothesis for two-way classified data with m
observations per cell;
5. How to derive the expectations of the various sum of squares; and
6. Numerical problems to test the hypothesis for two-way classified data
with m observations per cell.

8.9 SOLUTIONS /ANSWERS

E1) For set up an ANOVA Table for this problem, the computation results
are as follows:
Grand Total G = 14552
N = 18
Correction Factor (CF) = (14552 14552) /18 = 11764483.55
RSS = 12882026
TSS = 1117542
SSA = 1107070
SSB = 2957
SSAB = 1126
72
SSE = 6389
 Two-Way Anova with m
Observations Per Cell
ANOVA Table

Sources of DF SS MSS F-Calculated F-Tabulated at

Variation 5% level of
significance

Advertising 1 1107070 1107070 1107070/532.4 F(1,12) = 243.9

campaign 2 = 2079.32

Cake Mix 2 2957 1478.5 1478.5/532.42 F(2,12) = 19.41

= 2.8

Interaction 2 1126 563 563/532.42 = F(2,12) = 19.41

1.06

Error 12 6389 532.42

Total 17 1117542

Since computed value of F for cake mix and interaction are 2.8 and
1.06 respectively which are less than corresponding tabulated value so
they are not significant. Whereas the calculated value of F for
advertising campaign is greater than corresponding tabulated value so
there is a significant difference among advertising campaign.

73
Analysis of Variance TABLE: The F Table

Value of F Corresponding to 5% (Normal Type) and 1% (Bold Type) of the Area in the Upper Tail
Degrees of
Freedom: Degrees of Freedom (Numerator)
(Denominator) 1 2 3 4 5 6 7 8 9 10 11 12 14 16 20 24 30 
____________________________________________________________________________________________________________________
1 161 200 216 225 230 234 237 239 241 242 243 244 245 246 248 249 250 254
4,052 4,999 5,403 5,625 5,764 5,859 5,928 5,981 6,022 6,056 6,082 6,106 6,142 6,169 6,208 6,234 6,258 6,366

2 18.51 19.00 19.16 19.25 19.30 19.33 19.36 19.37 19.38 19.39 19.40 19.41 19.42 19.43 19.44 19.45 19.46 19.50
98.49 99.00 99.17 99.25 99.30 99.33 99.34 99.36 99.38 99.40 99.41 99.42 99.43 99.44 99.45 99.46 99.47 99.50

3 10.13 9.55 9.28 9.12 9.01 8.94 8.88 8.84 8.81 8.78 8.76 8.74 8.71 8.69 8.66 8.64 8.62 8.53
34.12 30.82 29.46 28.71 28.24 27.91 27.67 27.49 27.34 27.23 27.13 27.05 26.92 26.83 26.69 26.60 26.50 26.12

4 7.71 6.94 6.59 6.39 6.26 6.16 6.09 6.04 6.00 5.96 5.93 5.91 5.87 5.84 5.80 5.77 5.74 5.63
22.20 18.00 16.69 15.98 15.52 15.21 14.98 14.80 14.66 14.54 14.45 14.37 14.24 14.15 14.02 13.93 13.83 13.46

5 6.61 5.79 5.41 5.19 5.05 4.95 4.88 4.82 4.78 4.74 4.70 4.68 4.64 4.60 4.56 4.53 4.50 4.36
16.26 13.27 12.06 11.39 10.97 10.67 10.45 10.27 10.15 10.05 9.96 9.89 9.77 9.68 9.55 9.47 9.38 9.02

6 5.99 5.14 4.76 4.53 4.39 4.28 4.21 4.15 4.10 4.06 4.03 4.00 3.96 3.92 3.87 3.84 3.81 3.67
13.74 10.92 9.78 9.15 8.75 8.47 8.26 8.10 7.98 7.87 7.79 7.72 7.60 7.52 7.39 7.31 7.23 6.88

7 5.59 4.47 4.35 4.12 3.97 3.87 3.79 3.73 3.68 3.63 3.60 3.57 3.52 3.49 3.44 3.41 3.38 3.23
12.25 9.55 8.45 7.85 7.46 7.19 7.00 6.84 6.71 6.62 6.54 6.47 6.35 6.27 6.15 6.07 5.98 5.65

8 5.32 4.46 4.07 3.84 3.69 3.58 3.50 3.44 3.39 3.34 3.31 3.28 3.23 3.20 3.15 3.12 3.08 2.93
11.26 8.65 7.59 7.01 6.63 6.37 6.19 6.03 5.91 5.82 5.74 5.67 5.56 5.48 5.36 5.28 5.20 4.86

9 5.12 4.26 3.86 3.63 3.48 3.37 3.29 3.23 3.18 3.13 3.10 3.07 3.02 2.98 2.93 2.90 2.86 2.71
10.56 8.02 6.99 6.42 6.06 5.80 5.62 5.47 5.35 5.26 5.18 5.11 5.00 4.92 4.80 4.73 4.64 4.31

10 4.96 4.10 3.71 3.48 3.33 3.22 3.14 3.07 3.02 2.97 2.94 2.91 2.86 2.82 2.77 2.74 2.70 2.54
10.04 7.56 6.55 5.99 5.64 5.39 5.21 5.06 4.95 4.85 4.78 4.71 4.60 4.52 4.41 4.33 4.25 3.91

11 4.84 3.98 3.59 3.36 3.20 3.09 3.01 2.95 2.90 2.86 2.82 2.79 2.74 2.70 2.65 2.61 2.57 2.40
9.65 7.20 6.22 5.67 5.32 5.07 4.88 4.74 4.63 4.54 4.46 4.40 4.29 4.21 4.10 4.02 3.94 3.60

12 4.75 3.88 3.49 3.26 3.11 3.00 2.92 2.85 2.80 2.76 2.72 2.69 2.64 2.60 2.54 2.50 2.46 2.30
9.33 6.93 5.95 5.41 5.06 4.82 4.65 4.50 4.39 4.30 4.22 4.16 4.05 3.98 3.86 3.78 3.70 3.36

13 4.67 3.80 3.41 3.18 3.02 2.92 2.84 2.77 2.72 2.67 2.63 2.60 2.55 2.51 2.46 2.42 2.38 2.21
9.07 6.70 5.74 5.20 4.86 4.62 4.44 4.30 4.19 4.10 4.02 3.96 3.85 3.78 3.67 3.59 3.51 3.16

14 4.60 3.74 3.34 3.11 2.96 2.85 2.77 2.70 2.65 2.60 2.56 2.53 2.48 2.44 2.39 2.35 2.31 2.13
8.86 6.51 5.56 5.03 4.69 4.46 4.28 4.14 4.03 3.94 3.86 3.80 3.70 3.62 3.51 3.43 3.34 3.00

15 4.54 3.68 3.29 3.06 2.90 2.79 2.70 2.64 2.59 2.55 2.51 2.48 2.43 2.39 2.33 2.29 2.25 2.07
8.68 6.36 5.42 4.89 4.56 4.32 4.14 4.00 3.89 3.80 3.73 3.67 3.56 3.48 3.36 3.29 3.20 2.87

16 4.49 3.63 3.24 3.01 2.85 2.74 2.66 2.59 2.54 2.49 2.45 2.42 2.37 2.33 2.28 2.24 2.20 2.01
8.53 6.23 5.29 4.77 4.44 4.20 4.03 3.89 3.78 3.69 3.61 3.55 3.45 3.37 3.25 3.18 3.10 2.75

17 4.45 3.59 3.20 2.96 2.81 2.70 2.62 2.55 2.50 2.45 2.41 2.38 2.33 2.29 2.23 2.19 2.15 1.96
8.40 6.11 5.18 4.67 4.34 4.10 3.93 3.79 3.68 3.95 3.52 3.45 3.35 3.27 3.16 3.08 3.00 2.65

18 4.41 3.55 3.16 2.93 2.77 2.66 2.58 2.51 2.46 2.41 2.37 2.34 2.29 2.25 2.19 2.15 2.11 1.92
8.28 6.01 5.09 4.58 4.25 4.01 3.85 3.71 3.60 3.51 3.44 3.37 3.27 3.19 3.07 3.00 2.91 2.57

19 4.38 3.52 3.13 2.90 2.74 2.63 2.55 2.48 2.43 2.38 2.34 2.31 2.26 2.21 2.15 2.11 2.07 1.88
8.18 5.93 5.01 4.50 4.17 3.94 3.77 3.63 3.52 3.43 3.36 3.30 3.19 3.12 3.00 2.92 2.84 2.49

20 4.35 3.49 3.10 2.87 2.71 2.60 2.52 2.45 2.40 2.35 2.31 2.28 2.23 2.18 2.12 2.08 2.04 1.84
8.10 5.85 4.94 4.43 4.10 3.87 3.71 3.56 3.45 3.37 3.30 3.23 3.13 3.05 2.94 2.86 2.77 2.42

21 4.32 3.47 3.07 2.84 2.68 2.57 2.49 2.42 2.37 2.32 2.28 2.25 2.20 2.15 2.09 2.05 2.00 1.81
74 8.02 5.78 4.87 4.37 4.04 3.81 3.65 3.51 3.40 3.31 3.24 3.17 3.07 2.99 2.88 2.80 2.72 2.36
 Two-Way Anova with m
TABLE (Continued) Observations Per Cell
Degrees of
Freedom: Degrees of Freedom: Numerator
Denominator 1 2 3 4 5 6 7 8 9 10 11 12 14 16 20 24 30 
22 4.30 3.44 3.05 2.82 2.66 2.55 2.47 2.40 2.35 2.30 2.23 2.23 2.18 2.13 2.07 2.03 1.98 1.78
7.94 5.72 4.82 4.31 3.99 3.76 3.59 3.45 3.35 3.26 3.18 3.12 3.02 2.94 2.83 2.75 2.67 2.31

23 4.28 3.42 3.03 2.80 2.64 2.53 2.45 2.38 3.32 2.28 2.24 2.20 2.14 2.10 2.04 2.00 1.96 1.76
7.88 5.66 4.76 4.26 3.94 3.71 3.54 3.41 3.30 3.21 3.14 3.07 2.97 2.89 2.78 2.70 2.62 2.26

24 4.26 3.40 3.01 2.78 2.62 2.51 2.43 2.36 2.30 2.26 2.22 2.18 2.13 2.09 2.02 1.98 1.94 1.73
7.82 5.61 4.72 4.22 3.90 3.67 3.50 3.36 3.25 3.17 3.09 3.03 2.93 2.85 2.74 2.66 2.58 2.21

25 4.24 3.38 2.99 2.76 2.60 2.49 2.41 2.34 2.28 2.24 2.20 2.16 2.11 2.06 2.00 1.96 1.92 1.71
7.77 5.57 4.68 4.18 3.86 3.63 3.46 3.32 3.21 3.13 3.05 2.99 2.89 2.81 2.70 2.62 2.54 2.17

26 4.22 3.37 2.98 2.74 2.59 2.47 2.39 2.32 2.27 2.22 2.18 2.15 2.10 2.05 1.99 1.95 1.90 1.69
7.72 5.53 4.64 4.14 3.82 3.59 3.42 3.29 3.17 3.09 3.02 2.96 2.86 2.77 2.66 2.58 2.50 2.13

27 4.21 3.35 2.96 2.73 2.57 2.46 2.37 2.30 2.25 2.20 2.16 2.13 2.08 2.03 1.97 1.93 1.88 1.67
7.68 5.49 4.60 4.11 3.79 3.56 3.39 3.26 3.14 3.06 2.98 2.93 2.83 2.74 2.63 2.55 2.47 2.10

28 4.20 3.34 2.95 2.71 2.56 2.44 2.36 2.29 2.24 2.19 2.15 2.12 2.06 2.02 1.96 1.91 1.87 1.65
7.64 5.45 4.57 4.07 3.76 3.53 3.36 3.23 3.11 3.03 2.95 2.90 2.80 2.71 2.60 2.52 2.44 2.06

29 4.18 3.33 2.93 2.70 2.54 2.43 2.35 2.28 2.22 2.18 2.14 2.10 2.05 2.00 1.94 1.90 1.85 1.64
7.60 5.42 4.54 4.04 3.73 3.50 3.33 3.20 3.08 3.00 2.92 2.87 2.77 2.68 2.57 2.49 2.41 2.03

30 4.17 3.32 2.92 2.69 2.53 2.42 2.34 2.27 2.21 2.16 2.12 2.09 2.04 1.99 1.93 1.89 1.84 1.62
7.56 5.39 4.51 4.02 3.70 3.47 3.30 3.17 3.06 2.98 2.90 2.84 2.74 2.66 2.55 2.47 2.38 2.01

 3.84 2.99 2.60 2.37 2.21 2.09 2.01 1.94 1.88 1.83 1.79 1.75 1.69 1.64 1.57 1.52 1.46 1.00
6.64 4.60 3.78 3.32 3.02 2.80 2.64 2.51 2.41 2.32 2.24 2.18 2.07 1.99 1.87 1.79 1.69 1.00

75