Classic Assumption Testing
Classic Assumption Testing
STATISTICS
B = BEST
L = LINIER
U = UNBIASED
E = ESTIMATOR
OBJECTIVE:
To test if in regression model have correlation between independent. variables
X1 2 3 5 4 6 2 3 4 5 6
X2 3 4 6 5 7 6 4 5 4 3
Y 5 8 8 9 9 13 6 9 4 3
Y = consumption
X1 = price
X2 = income
5
SPSS OUTPUT
Coefficient Correlationsa
Coefficientsa
Unstandardized Standardized
Coefficients Coefficients Collinearity Statistics
Model B Std. Error Beta t Sig. Tolerance VIF
1 (Constant) 2.553 1.626 1.570 .160
harga -1.092 .271 -.552 -4.029 .005 .950 1.052
pendapatan 1.961 .302 .889 6.490 .000 .950 1.052
a. Dependent Variable: keperluan konsumsi
Tolerance >0,1
VIF <10
We can make summary if no multicolonierity
between independent variables.
OBJECTIVE:
to test if in linier regression model have a corellation between
error in t period with error in before periode (t-1).
Requirement:
“Have intercept in regression model.”
Equation:
SPSS output:
Model Summaryb
Ho Hypothesis Decision If
No positive autocorellation Reject 0< d <dl
No positive autocorellation No decision* dl≤ d ≤ du
No negative autocorellation Reject 4-dl< d <4
No negative autocorellation No decision* 4-du ≤d ≤ 4-dl
d= 3.386
Keputusan:
dl = 0,95
du = 1,54
1. Open file
2. Choice Plot
3. Addition SRESID variable in Y box
4. Addition ZPRED variable in X box
5. Choice continue
6. Ok
^
From example below Y 2,5529 1,0921X 1 1,9608 X 2 .
Y = Consumption
X1 = Price
X2 = Income
X1 X2 Y
2 3 5
3 4 8
5 6 8
4 5 9
6 7 9
2 6 13
3 4 6
4 5 9
5 4 4
6 3 3
1.5
0.5
0.0
-0.5
-1.0
-1.5
-2 -1 0 1 2
Regression Standardized Predicted Value
To detect NORMALITy
1. Grafik analysis (normal P-P plot)
2. Statistic analysis (analisis Z skor skewness dan
kurtosis) one sample Kolmogorov-Smirnov Test.
Hipotesis:
H0: Residual data have normal distribution
H1: Residual data have not normal distribution
1. Analisis Grafik
1. Buka file
2. Tekan tombol Plot
3. Aktifkan standardized residual plot pada
Histogram dan Normal Probability Plot.
4. Tekan tombol Continue dan abaikan lainnya,lalu
tekan Ok.
^
Dari contoh sebelumnya di dapat Y 2,5529 1,0921X 1 1,9608 X 2 .
Y = Pengeluaran konsumsi
X1 = harga
X2 = pendapatan
X1 X2 Y
2 3 5
3 4 8
5 6 8
4 5 9
6 7 9
2 6 13
3 4 6
4 5 9
5 4 4
6 3 3
Buatlah uji normalitas dari data di atas.
19
Analisis.
Model Regresi memenuhi asumsi normalitas,karna:
1.Grafik Histogram memberikan pola distribusi normal.
2. Grafik normal plot terlihat data menyebar di sekitar garis diagonal dan tidak
menjauh dari garis diagonal.
2. Analisis Statistik
1. Buka file
2. Pilih menu Analyze
3. Pilih Non-parametric test
4. Pilihsub menu 1-sample k-S
5. Pada kotak test variabel list,isikan unstandardized
residual(RES_1),caranya lht pada Uji Gletjer.
6. Aktifkan test distribution pada kotak Normal.
Output SPSS
One-Sample Kolmogorov-Smirnov Test
Unstandardiz
ed Res idual
N 10
Normal Parameters a,b Mean .0000000
Std. Deviation 1.04224494
Most Extreme Absolute .297
Differences Positive .257
Negative -.297
Kolmogorov-Smirnov Z .940
As ymp. Sig. (2-tailed) .340
a. Test distribution is Normal.
b. Calculated from data.
ANALISIS
Karena Sig >0.05,maka H0 diterima,maka disimpulkan
data residual berdistribusi Normal.
Linearity is condition if relationship
between dependent and independent
variables have linear (strick line)in
variable range
23
Uji Linearitas Data
Contoh : Uji asumsi linearitas variabel usia – income
Buka file homoskedastisitas
1. Menu Graph Legacy Dialogs Scatter/dot
2. Pada kotak SCATTER/DOT, pilih SIMPLE SCATTER, kemudian tekan tombol
DEFINE.
3. Masukkan variabel usia pada sumbu X dan variabel olahraga pada sumbu Y.
4. Abaikan yang lain lalu klik OK.
24
Uji Linearitas Data
Output: Variabel Usia - Olahraga
25
Uji Linearitas Data
Contoh : Uji asumsi linearitas variabel Olahraga - Berat
Buka file homoskedastisitas
1. Menu Graph Legacy Dialogs Scatter/dot
2. Pada kotak SCATTER/DOT, pilih SIMPLE SCATTER, kemudian tekan tombol
DEFINE.
3. Masukkan variabel berat pada sumbu X dan variabel olahraga pada sumbu Y.
4. Abaikan yang lain lalu klik OK.
26
Uji Linearitas Data
Output: Variabel Berat - Olahraga
27
Uji Linearitas Data
Kesimpulan:
1. For age and sport variables have
linearity relationship
2. For weight and sport variables
have not linearity relationship
28
SELAMAT BELAJAR
Thank You