ANOVA TESTING

Understanding the concept and practice

Tuesday, May 21ST 2024
 +
The concept of
ANOVA test
+

+
ANOVA test
Statistic test that used to assess the difference means
of more than two (>2) groups.
+
This is categorized in the parametric statistic,
so need to know whether the data should be normally
distributed or not. +
Two types of ANOVA:
One Way Anova
Two Way Anova

+
 C
ONE WAY O TWO WAY
ANOVA N ANOVA
Should test the different mean between the Similar with One Way Anova, but TWO WAY
completely three different groups.
Only has ONE independent variable (that T ANOVA has TWO independent variable (that
contains >2 groups).
contains >2 groups).
Example: Assurance Company (Manulife, AIA, R Example: Brand (Toyota, Lexus, Honda, BMW)
and Type of Car (SUV, MPV, Sedan, City Car)
Allianz, BCALife) and Customer satisfaction.
A toward Customer Prestige.

S
T
ONE WAY ANOVA
ASSURANCE COMPANY CUSTOMER SATISFACTION
BCALife 75
AIA 77
ManuLife 89
Allianz 83
ManuLife 67
Allianz 82
BCALife 82
AIA 80
TWO WAY ANOVA
CAR BRAND CAR TYPE CUSTOMER PRESTIGE
Honda SUV 80
Lexus Sedan 92
Toyota City Car 79
BMW MPV 85
Lexus MPV 83
Honda City Car 88
BMW SUV 91
Toyota Sedan 86
 •Open SPSS
•Input Data
One Way •Click Analyze
Anova
Steps in •Click Compare Means...
SPSS •Click One Way Anova
•Send the variables to variables box.
•Click OK
 •Open SPSS
•Input Data
Two Way •Click Analyze
Anova
Steps in •Click General Linier Model...
SPSS •Click Univariate
•Send the variables to variables box.
•Click OK
 +

ONE WAY ANOVA

SPSS Simulation +
Use ‘Data OWA.sav’ +
 +
SPSS
+ Simulation Steps !
• Normal distribution test
•One Way Anova
+
Normality Test
•Click Analyze
•Click Descriptive Statistic
•Click Explore
•Wait until the new window
popping-up.
•Put ‘Self Compassion; to
dependent list.
•Put ‘Kategori ANX’ to
Factor List.

Note:
Interval/rasio data only.
Normality Test
•Click Plot
•Click normality plots with
test
•Click Continue
•Clik OK.
+ Output of Normality

+ Look at:
Sig. Value in kolmogorov smirnov (if samples >10)
Nilai shapiro-Wilk (if sample <10)
Data would be normal when sig > 0.05
One Way ANOVA

•Click Analyze
•Click Compare Means
•Click One Way Anova...
•Wait until the new window
popping-up.
•Put ‘Self Compassion’
into Test Variable(s)
box.
•Put ‘Kategori ANX’ to
Factor.
•Clik Options.
•Wait until the new
window popping-up.
•Check Homogenity of
variance test’
•Clik Continue
•Click OK
+ Output of One Way ANOVA

+
Look at:
Sig. Value.
the should have p < 0,05 à there is significant difference means of ‘self compassion’
between three groups of anxiety (no, low, and mild anxiety).
In this case, the p-value < 0,05 à there is significant mean difference three groups of
anxiety (no, low, and mild anxiety).
 +

TWO WAY ANOVA

SPSS Simulation +
Use ‘Data TWA.sav’ +
Two Way ANOVA

•Click Analyze
•Click General Linier Model
•Click Univariate...
•Wait until the new window
popping-up.
•Put ‘Customer Prestige’ into
Dependent Variable box.
•Put ‘Car Brand’ and ‘Car
Type’ to Fixed Factor(s)
box.
•Clik Options.
•Wait until the new window
popping-up.
•Check Descriptive
statistic
•Check Homogenity test
•Clik Continue
•Wait until window back
to main menu
•Click Post-hoc
•Put ‘CB’ and ‘CT’ into Post
Hoc Test for box.
•Check Benferroni.
•Click Continue.
•Click OK
 Output of Two Way ANOVA

+
Look at:
Sig. Value.
the should have p > 0,05 à The variance data are homogeny.
In this case, the p-value > 0,05 à The variance data are homogeny.
 Output of Two Way ANOVA

+
Look at:
Sig. Value.
the should have p < 0,05 à There is a sig difference of means.

Car Brand à 0,691 (> 0,05) à there is no sig difference of customer prestige based on car
brand. (Which means any kind of car brand involved in this research have similar customer
prestige).
Car type à 0,638 (> 0,05) à there is no sig difference of customer prestige based on car type.
(Which means any kind of car type involved in this research have similar customer prestige).
 Output of Two Way ANOVA

+
Look at:
Sig. Value.
the should have p < 0,05 à There is a sig difference of means.

Car type and car brand à 0,148 (> 0,05) à there is no sig difference of customer prestige
based on car type and brand. (Which means any kind of car brand and type involved in this
research have similar customer prestige).
+ Group Task!
Description of Task:
• Do it in the group (3 students/group).
• Please kindly to download the data set titled ”ANOVA Task” in
e-learning to complete this task.
• Please analyze the data and interpret the result of analysis.
• Every group need to present the result in the next meeting by power
point.
• The points that every group need to present are:
+ • Is there any sig different of awareness means based on gender?
• Is there any sig different of awareness means based on Shot of
coffee?
• Is there any sig different of awareness means between
+
coffeedrink and shot?
“ It’s easy to lie with statistic.
It’s hard to tell the truth
without statistic.

Thank you!