Simple Analysis of Variance

xa xb xc 2
(x ¿¿ a) ¿
2
(x ¿¿ b) ¿
2
(x c )
Method A Method B Method C
Student 1 84 70 90 7,056 4,900 8,100
Student 2 90 75 95 8,100 5,625 9,025
Student 3 92 90 100 8,464 8,100 10,000
Student 4 96 80 98 9,216 6,400 9,604
Student 5 84 75 88 7,056 5,625 7,744
Student 6 88 75 90 7,744 5,625 8,100

∑ x =1,560
∑ x 2=136,484
(∑ x )
2

TSS = ∑ x 2−
N
2
−( 1,560 )
= 136,484
18
−2,433,600
= 136,484
18
= 136,484 −135,200
= 1,284

(∑ x )
2
1 2
SSb =
No . of rows
∑ ( ∑ of each column) −
N
(∑ x )
2
1
¿ ∑ (x jj ) −
2
r N
2
1 1,560
¿ ∑ (534 + 465 +561 ) −
2 2 2 ❑
6 18
1 2,433,600
¿ ∑ (285,156+216,255+ 314,721) −
❑
6 18

816,102 2,433,600
¿ −
6 18
¿ 136,017−135,200
¿ 817
SSw =TSS−SS b
¿ 1,284−817
¿ 467
ANOVA Table

Source of variation Sum of squares df MSS=SS/df

Between-column 817 2 408.5
Within-column 467 15 31.13
Total 1,284 17

MSS b=SS b /df b MSS b

F=
¿ 817 /2 MSS w
¿ 408.5 408.5
F=
31.13
MSS w =SS w /df w ¿ 13.12
¿ 467 /15
¿ 31.13

The tabular value at n1−2 and n2 =15, at 5% significance level is 3.68. at 1% significance level, the
tabular value is 6.36
ANALYSIS OF ENUMERATION DATA

Status Frequency
(f)
Single 18
Married 24
Widowed 5
Leg. Separated 3
Total
1. Civil Status of Employees.

Based on the data above, is the actual observed proportion significantly different from the expected proportion, if
the ideal or expected proportion is 30% single, 50% married, 10% widowed, and 10% legally separated? (use α =
5%)

Ho: The actual observed proportion is not significantly different from the expected proportion.

df = number of categories – 1 tabular value = 7.81 Expected Frequency

=c–1 fe = total frequency x proportion
=4–1 fe1 = 50 (.30) = 15
=3 fe2 = 50 (.50) = 25
fe3 = 50 (.10) = 5
fe4 = 50 (.10) = 5

Solution:
2
( fo−fe)
x 2= ∑
fe
2 2 2 2
(18−15) (24−25) (5−5) (3−5)
¿ + + +
15 25 5 5
9 1 0 4
¿ + + +
15 25 5 5
¿ .60 +.04 + 0 + .80
¿ 1.44

Conclusion: Since 1.44 < 7.81, we accept the null hypothesis. The actual observed proportion is not significantly
different from the expected proportion.

2. Classification on the IQ of Students.

High IQ Average IQ Low IQ Total

15 20 5

Based on the data above find out whether or not the given sample distribution agrees with the hypothetical normal
distribution whose ratio is 1:2:1 (use α = 5%).

Ho: The sample distribution agrees with the hypothetical normal distribution.

df = number of categories – 1 tabular value = 9.21 Expected Frequency

=c–1 fe = total frequency x proportion
=3–1 fe1 = 40 (1/4) = 10
 =2 fe2 = 40 (1/2) = 20
fe3 = 40 (1/4) = 10

Solution:
2
( fo−fe)
x 2= ∑
fe
2 2 2
(15−10) (20−20) (5−10)
¿ + +
10 20 10
25 0 25
¿ + +
10 20 10
¿ 2.5 +0 + 2.5
¿ 5.0

Conclusion: Since 5.0 < 9.21, the null hypothesis is accepted. The sample distribution agrees with the hypothetical
normal distribution.

3. Classification on the Production Quality.

Production Quality Number of Products

Non-defective 1000
Defective 200
Total
Using the data above, is the observed proportion really different from the ideal ratio of 2 non-defective goods to 1
defective good (2:1) if there are 200 defective goods out of 1,200 goods? (use α = 5%)

Ho: The proportion 1200:200 is not significantly different from the ideal proportion of 2:1.

df = number of categories – 1 tabular value = 3.84 Expected Frequency

=c–1 fe = total frequency x proportion
=2–1 fe1 = 1200 (2/3) = 800
=1 fe2 = 1200 (1/3) = 400
Solution:
2
(|fo−fe|−.5)
x =∑
2
fe
2 2
(|1000−800|−.5) (|200−400|−.5)
¿ +
1000 400
2 2
(199.5) (199.5)
¿ +
800 400
¿ 49.75 +99.50
¿ 149.25

Conclusion: Since 149.25 > 3.84, we reject the null hypothesis. The sample proportion is significantly different form
the ideal proportion.

4. Sex and Attitude Toward Household Chores of College Freshmen.

Sex Boys Girls Total

Attitude
Positive 9 21
Negative 9 11
Total
Does attitude toward household chores depend on sex for the children being considered in the data above? (use α =
5%).

Ho: Attitude toward household chores does not depend on sex.

 df = (r–1) (k-1) tabular value = 3.84 Expected Frequency
= (2-1) (2-1) ( subtotal A ) ( subtotal B)
fe =
=1 grand total
30 x 18
where fo1 = 9, fe1 = = 10.8
50
20 x 18
where fo2 = 9, fe2 = = 7.2
50
30 x 32
where fo3 = 21, fe3 = = 19.2
50
20 x 32
where fo4 = 11, fe4= = 12.8
50
Solution:
2
(|fo−fe|−.5)
x =∑
2
fe
2 2 2 2
(|9−10.8|.5) (|9−7.2|−.5) (|21−19.2|−.5) (|11−12.8|−.5)
¿ + + +
10.8 7.2 19.2 12.8
2 2 2 2
(1.3) (1.3) (1.3) (1.3)
¿ + + +
10.8 7.2 19.2 12.8
¿ .1565+.2347+.0880+ .1320
¿ .6112 or .61

Conclusion: Since .61 < 3.84, we accept the null hypothesis. Attitude toward household chores does not depend on
sex.

5. Academic Performance and IQ of Students

Academic IQ High Average Low Total

Performances

Passed 31 45 4
Failed 1 4 15

Total
Based on the above, test the hypothesis that academic performance does not depend on IQ (use α = 5%).

Ho: Academic performance does not depend on IQ.

df = (r–1) (k-1) tabular value = 9.21 Expected Frequency

= (2-1) (3-1) ( subtotal A ) (subtotal B)
fe =
=2 grand total
32 x 80
where fo1 = 31, fe1 = = 25.6
100
32 x 20
where fo2 = 1, fe2 = = 6.4
100
49 x 80
where fo3 = 45, fe3 = = 39.2
100
49 x 20
where fo4 = 4, fe4= = 9.8
100
19 x 80
where fo4 = 4, fe4= = 15.2
100
19 x 20
where fo4 = 15, fe4= = 3.8
100
Solution:
2
( fo−fe)
x 2= ∑
fe
2 2 2 2 2 2
(31−25.6) (1−6.4) ( 45−39.2) (4−9.8) ( 4−15.2) (15−3.8)
¿ + + + + +
25.6 6.4 39.2 9.8 6.4 3.8
29.16 29.16 33.64 33.64 125.44 125.44
¿ + + + + +
25.6 6.4 39.2 9.8 15.2 3.8
¿ 1.139+ 4.556+.858+3.433+8.253+ 33.011
¿ 51.25

Conclusion: Since 51.25 < 9.21, the null hypothesis is rejected. Academic performance depends on IQ.