PREREQUISITE TEST (ANALYSIS OF NORMALITY,

HOMOGENEITY, AND LINEARITY USING SPSS PROGRAM)

Lecturer : Dr. Linda Rosalina. S.Ked.,M.Biomed

Course : Statistic

ARRANGED BY :
CLAUDIA MAILANDA SALSABILA
21078035

STUDY PROGRAM :
COSMETOLOGY AND BEAUTY EDUCATION
FACULTY OF TOURISM AND HOSPITALITY
 ANALYSIS OF REQUIREMENTS TEST DATA

Nonparametric test is used if the assumptions in the parametric test are not met. The
most common assumption in parametric tests is that a random sample comes from a
population that is normally distributed, the variance is homogeneous, and is linear. If
these assumptions are met, or at least the deviation from the assumptions is small, then
the parametric test can still be relied on. However, if the assumptions are not met, then
nonparametric tests are an alternative. There are three assumptions of parametric
statistical test as described above, namely normality, data linearity and homogeneity
of variance.
1. Normality Test
Normality testing is done to determine whether a data distribution is normal or not. It
is important to know that it is related to the accuracy of the selection of statistical tests
to be used. Parametric tests, for example, require the data to be normally distributed. If
the data distribution is not normal, it is recommended to use a nonparametric test.
This normality test must be carried out if there is no theory stating that the variables
studied are normal. In other words, if there is a theory which states that a variable
being studied is normal, then there is no need to test the normality of the data. There
are several techniques that can be used to test the normality of the data, including: Chi
Square test, Lilifors test, Kolmogorov-Smirnov test, etc.
To test the normality of the data with SPSS, the steps are as follows:
 Input the data below into SPSS

Variabel 1 2 3 4 5 6 7 8 9 10

IQ (X1) 40 45 30 35 30 40 45 45 35 35

EQ (X2) 18 20 14 12 14 16 10 15 18 16

Hasil Belajar (Y) 80 85 60 75 55 65 80 90 75 60

 In the Name field, type X1, X2 and Y.

 In the Decimals column, change the number to 0 for all variables.
 In the Label column, enter IQ on X1, EQ on X2 and Learning Outcomes on Y.
 In the Align column, fill in Center.
 In the Measure column, enter Scale on X1, Ordinal on X2 and Scale on Y.
 For other columns leave it alone (default field).
 Klik tab sheet [Variable View] pada SPSS data editor dan ketik/copy data sebagai
berikut:

 Selanjutanya klik [Analyze] > [Descriptives Statistics] > [Explore].

 Akan terbuka kotak dialog Explore.

 Masukan variabel Hasil Belajar (Y) ke dalam kotak Dependent List.
 Sorot variabel IQ (X1) dan EQ (X2) dan masukan ke dalam kotak Factor List.
 Klik tab Plots, centang pilihan Normality plots with tests.
 Klik Continue. Klik [OK].

 Uji normalitas menghasilkan beberapa output, untuk keperluan penelitian pada

umumnya hanya diperlukan output berupa Tes of Normality seperti berikut:

Tests of Normality
Shapiro-
Kolmogorov-Smirnova
Wilk
Statistic df Sig. Statistic df Sig.

Hasil belajar ,184 10 ,200* ,939 10 ,543

 *. This is a lower bound of the true
significance. a. Lilliefors Significance
Correction
 Menafsirkan hasil uji normalitas, pada gambar di atas menunjukkan uji normalitas
data Hasil Belajar. Pengujian normalitas dengan SPSS berdasarkan pada uji
Kolmogorov-Smirnov dan Shapiro Wilk. Pilih salah satu, misalnya
KolmogorovSmirnov. Hipotesis yang diuji adalah:
Ho : Data berdistribusi normal
H1 : Data berdistribusi tidak normal
 Cara mengetahui signifikan atau tidak hasil uji normalitas adalah dengan
memperhatikan angka pada kolom signifikansi (Sig). Kriteria uji normalitas sebagai
berikut:
 Tetapkan taraf nyata (signifikansi) misalnya α = 0,05.
 Bandingkan angak pada kolom Sig dengan α.
 Jika Sig > α maka data berdistribusi normal.
 Jika Sig < α maka data berdistribus tidak normal.
 Pada contoh di atas Sig > α atau 0,200 > 0,05 maka dapat disimpulkan data variabel
hasil belajar berdistribusi normal.

2. Uji Linearitas

The linearity test aims to determine whether two variables have a linear relationship or not
significantly. This test is used as a prerequisite for parametric statistics, especially in
correlation analysis or linear regression which is included in the associative hypothesis. In the
SPSS program, the linearity test uses the Test for Linearity at a significant level of 0.05. Two
variables are said to have a linear relationship if the significance (Deviation from Linearity) >
0.05.

Linearity test steps on SPSS.

 Input the above data into SPSS
 Click [Variable View] in the SPSS data editor.
 In the Name column, type in a row: X1, X2 and Y.
 For the label column, type in a row: IQ, EQ and Learning Outcomes.
 Other fields adjust (leave default).
 Open [Data View] and enter data for the three variables.
 Click [Analyze] > [Compare Means] > [Means].
 Masukkan variabel Hasil Belajar ke kotak Dependent List, sorot variabel IQ dan
EQ kemudian masukkan ke dalam kotak Independent List.
 Klik Options, pada Statistic for First Layer centang pilihan Test for Linearity
kemudian klik Continue.
 Klik [OK].

 Output pada kolom Anova Table sebagai berikut:

ANOVA Table
Sum of df Mean F Sig.
Squares Square
(Combined) 937,500 3 312,500 5,76 ,034
9
Between 16,4 ,007
889,113 1 889,113
Linearity 1
 Hasil belajar * Groups 4 ,659
IQ Deviation from
Linearity 48,387 2 24,194 ,447
Within
Groups 325,000 6 54,167

Total 1262,500 9

 Dari output di atas diketahu nilai Sig pada baris Deviation from Linearity sebesar
0,659. Karena nilai Sig > α atau 0,659 > 0,05 maka dapat disimpulkan bahwa antara
variabel Hasil Belajar dan IQ terdapat hubungan linear.

3. Homogeneity Test

Homogeneity test is used to determine whether two or more population variances are the
same or not. This test is carried out as a prerequisite for the independent sample t-test
and ANOVA analysis or for researchers who use more than one sample group. It is
generally used to prove comparative hypotheses. The basic assumption in the analysis of
variance (Anova) is that the variance of the population is the same. As a test criterion, if
the value of Sig > (by default in SPSS is 0.05) then it can be said that the variance of two
or more data groups is the same.

Example:

A student named Atun Bangkonawati conducted a study to find out: “Are there any
differences in learning outcomes of mathematics using the Number Head Together
(NHT) cooperative learning model with conventional learning models. Broadly speaking,
the procedure for collecting data is to take two groups of samples. The first sample group
of the learning process used the Make a Match type of cooperative learning model and
the second group used the conventional learning model. After learning in each sample
group is complete, then a test is given. The results of this test will be tested to determine
whether the treatment given to the two sample groups is different or not. To test the
hypothesis can use the independent sample t-test. Prior to the t-test as a prerequisite, the
homogeneity must be tested first. If the variance of the two samples is homogeneous,
then a t-test can be performed and if it is not homogeneous, use nonparametric statistics.
The research data obtained are as follows:

Data Model Pembelajaran NHT dan Konvensional

No. X1 X2
Responden (NHT) (Konvensional)

1 35 65
2 37 34
3 27 44
 4 29 36
5 31 31
6 30 33
7 27 37
8 36 28
9 37 33
10 24 34
11 26 35
12 29 28
13 31 29
14 36
15 33

The steps of homogeneity test with SPSS are as follows:

▪ Input data into SPSS program
▪ Click the sheet tab [variable view] in the SPSS data editor.
▪ In the Name column, in the first row, type X and in the second row, type k.
▪ In the Label column, in the first row, type Learning Outcomes and in the second row,
type Learning Models.
▪ For the Value column, fill in 1 and type NHT then click add. In the same way, enter 2
and type Conventional then click add.

Klik [OK].

 Klik tab sheet [Data View], masukan data di atas.

Klik [Analyze] > [Descriptives Statistic] > [Explore].
 Pada kotak dialog Explore, masukan Variabel Hasil Belajar ke kotak Dependent
List dan Model Pembelajaran ke kotak Factor List.
 Klik Plots, centenag Unstransformed kemudian klik Continue.
 Klik [OK].

Pada Output lihat bagian Test of Homogenity of Variances dan pada baris Based on
Mean lihat kolom Sig.

Test of Homogeneity of Variance

Levene df1 df2 Sig.
Statistic

Based on Mean ,680 1 26 ,417

Based on Median ,353 1 26 ,558

Hasil
Belajar Based on Median and
with adjusted df ,353 1 16,523 ,561

Based on trimmed mean ,403 1 26 ,531

 Dari tabel output di atas dapat diketahui nilai Sig sebesar 0,417. Karena nilai Sig >
0,05 maka dapat disimpulkan bahwa kedua kelompok data mempunyai varian sama
atau homogen. Angka Levene Statistic menunjukkan semakin kecil nilainya maka
semakin besar homogenitasnya.