Analysis of Variance

Dr. Muhamed Wael Farouq

Statistics and Mathematics Department

Course: Inferential Statistics

College of Management & Technology - AASTMT
1
Copyright Notice
This presentation is protected by copyright laws.
Reproduction and distribution of the presentation without
written permission of the author is prohibited.
Sharing or copying this material for distribution (e.g. to a
third-party website) can lead to a violation of copyright
law.

Copyright © 2016 - 2023, Muhamed Wael Farouq. All

Rights Reserved.

2
 One Way
Analysis of Variance
(ANOVA)

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 3
Analysis of Variance (ANOVA)
The F-distribution is used in testing of hypothesis of the equality of more
than two population means.

The ANOVA to test the equality of more than two population means
requires the following assumptions:

1- The populations follow the normal distribution.

2- The populations have equal standard deviations or equal variances.
3- The populations are independent.

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 4
One-Way ANOVA
Step1. State the null hypothesis and the alternative hypothesis.

H 0 : 1 = 2 = 3 = ... There is no difference among means

H1 : 1  2  3  ... Not all  ' s are equal

Step 2. Construct the calculation table

Step 3. Calculate SST, SS total and SSE

SST ---------- Sum of squares of treatments
SSE---------- Sum of squares of errors
SS total------ Sum of squares of total

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 5
One-Way ANOVA

 Tc2  T 2
SST =   −
 nc  n

2
T
SS total =  x 2 −
n

SS total = SST + SSE

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 6
One-Way ANOVA
Step 4. Construct the ANOVA table

k number of samples, columns or treatments

n total number of observations

Dr. Muhamed Wael Farouq
College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 7
One-Way ANOVA
Step 5. Take the decision Ftabulated = Fnk−−k1 ( )

Fcalculated  Ftabulated Fcalculated  Ftabulated

Accept H 0 Reject H 0

acceptance rejection
region region
Ftabulated (Treatments) = Fnk−−k1 ( )
Dr. Muhamed Wael Farouq
College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 8
One-Way ANOVA
Exercise
A firm wishes to compare 4 programs for training workers to perform a
certain manual task. Twenty new employees are randomly assigned to the
train programs with 5 in each program. At the end of the training period, a
test is conducted to see how quickly trainees can perform the task, the
number of times the task is performed per minute is recorded for each
trainee with the following results.

Construct the ANOVA table; use the 0.10 level of significance to determine
whether treatments differ in effectiveness.
Dr. Muhamed Wael Farouq
College of Management & Technology - AASTMT 9
Course : Inferential Statistics
One-Way ANOVA
Solution
Step 1. State H0 and H1

H0 :  prog 1 =  prog 2 =  prog 3 =  prog 4

H1 : not all s areequal

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 10
One-Way ANOVA
Step 2. Construct the following table

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 11
One-Way ANOVA
Step 3. Calculate SST, SS total and SSE
**SST

 Tc2  T 2
SST =   −
 nc  n

 T12 T22 T32 T42  T 2

SST =  + + + −
 n1 n2 n3 n4  n

 (59) 2 (44) 2 (61) 2 (43) 2  (207) 2

SST =  + + +  −
 5 5 5 5  20

SST = 54.95
Dr. Muhamed Wael Farouq
College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 12
One-Way ANOVA
**SS total
2
T
SS total =  x 2 −
n

(9) 2 + (12) 2 + (14) 2 + (11) 2 + (13) 2 + 

 
2
+ 2
+ 2
+ 2
+ 2
+
SS total =  −
2
(10) (6) (9) (9) (10) (207)
(12) 2 + (14) 2 + (11) 2 + (13) 2 + (11) 2 +  20
 
(9) 2 + (8) 2 + (11) 2 + (7) 2 + (8) 2 

SS total = 96.55

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 13
One-Way ANOVA
**SSE

SS total = SST + SSE

SSE = SS total − SST

SSE = 96.55 − 54.95

SSE = 41.6

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 14
One-Way ANOVA
Step 4. Construct the ANOVA table

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 15
One-Way ANOVA
Step 5. Take the decision

Ftabulated = Fnk−−k1 ( )

Ftabulated = F163 ( 0.10 )

Ftabulated = 2.462

Fcalculated = 7.04 acceptance rejection

region region
Ftabulated = 2.462 Ftabulated = 2.46
Since Fcalculated  Ftabulated then reject H0

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 16
One-Way ANOVA
Exercise

a- How many populations are in the analysis?

b- Find the missing values.
c- State the null & alternative hypothesis.
d- Test using  = 0.05

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 17
One-Way ANOVA
How many populations are in the analysis?

df treatment + df error = df total

(k-1) + 12 = 15

k-1 + 12 = 15

k=4

There are 4 populations (treatments) in this analysis.

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 18
One-Way ANOVA
Find the missing values.
1- SSE

SST + SSE = SS total

23.37 + SSE = 33.09

SSE = 9.72

2- df treatments
df treatments = k-1
df treatments = 4-1= 3

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 19
One-Way ANOVA
3- MST 4- MSE 5- Fcalculated
SST SSE
MST = MSE = MST
k −1 n−k Fcalculated =
MSE

23.37 9.72 7.79

MST = MSE = Fcalculated =
4 −1 12 0.81

MST = 7.79 MSE = 0.81 Fcalculated = 9.61

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 20
One-Way ANOVA

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 21
One-Way ANOVA
State the null & alternative hypothesis.

H 0 : 1 = 2 = 3 = 4

H1 : not all s areequal

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 22
One-Way ANOVA
Test using  = 0.05
Ftabulated = Fnk−−k1 ( )

Ftabulated = F123 ( 0.05)

Ftabulated = 3.49

Fcalculated = 9.61 acceptance rejection

region region
Ftabulated = 3.49 Ftabulated = 3.49
Since Fcalculated  Ftabulated then reject H0

Dr. Muhamed Wael Farouq

College of Management & Technology - AASTMT Copyright © 2016 - 2023, Muhamed Wael Farouq
Course : Inferential Statistics All Rights Reserved 23