PRELIS 2, PRELIS 2, Model Pengukuran dan Model Pengukuran dan Model Struktural Model Struktural Dr. Setyo Hari Wijanto <[email protected]>
PRELIS 2,PRELIS 2,Model Pengukuran Model Pengukuran dan Model Strukturaldan Model Struktural
Dr. Setyo Hari Wijanto<[email protected]>
Bab 5Bab 5Data Input dan Data Input dan PRELIS2PRELIS2
Dr. Setyo Hari Wijanto<[email protected]>
April 2009Bab 6 Model Pengukuran3
Jenis Raw Data
Salah satu data input adalah raw data yang merupakan data dari variabel-variabel teramati (observed variables) yang ada di dalam model.
Raw data ini bisa diperoleh dari survei (yang dikenal sebagai data primer) atau dari berbagai sumber data (data sekunder).
LISREL 8 mempunyai 3 jenis data input yang dapat digunakan, yaitu: Ordinal (ordinal) Continue (kontinu) Censored (sensor)
April 2009Bab 6 Model Pengukuran4
ORDINAL VARIABLE Ketika data dikumpulkan melalui wawancara atau
kuesioner observed variables adalah ordinal, yaitu, respon-respon diklasifikasikan ke dalam kategori-kategori yang berurutan.
Sebuah ordinal variabel z bisa dianggap sebagai ukuran mentah/kasar dari unobserved atau unobservable continuous variable z* yang mendasarinya.
Misalkan 4 point ordinal scale dapat dituliskan: If z* <= T1 , z is scored 1
If τ1 < z* <= τ2, z is scored 2
If τ2 < z* <= τ3, z is scored 3
If τ3 > z* , z is scored 4
Dimana τ1 < τ2 < τ3 adalah threshold values untuk z* . Sering diasumsikan bahwa z* mempunyai distribusi normal.
April 2009Bab 6 Model Pengukuran5
ORDINAL VARIABLE Dengan mengasumsikan setiap pasang z*
mempunyai bivariate normal distribution, maka jenis koefisien korelasi yang dapat dihitung adalah sebagai berikut:
Polychoric – Kedua z variables mempunyai sebuah skala ordinal
Tetrachoric - Kedua z variables mempunyai sebuah skala dikotomi
Polyserial – Satu z variable mempunyai skala ordinal dan yang lainnya mempunyai skala interval
Biserial - Satu z variable mempunyai skala interval dan yang lainnya mempunyai skala dikotomi
April 2009Bab 6 Model Pengukuran6
ORDINAL VARIABLE
April 2009Bab 6 Model Pengukuran7
Variabel Kontinu dan Skor Normal
Perlunya Skor Normal Untuk variabel kontinu yang tidak normal. Jika
metode estimasi Maximum Likelihood (ML) yang digunakan, kesalahan standar dan chi-squares bisa agak tidak tepat. Secara teoritis, Weighted Least Square (WLS atau ADF) dengan weight matrix yang tepat akan menghasilkan kesalahan standar dan chi-squares tepat, tetapi hal ini memerlukan sampel yang besar.
Salah satu solusi atas ketidak-normalan dari variabel kontinu jika sampel tidak terlalu besar adalah melakukan normalisasi variabel sebelum melakukan analisis (Joreskog et.al. 1999). Salah satu fitur dari LISREL 8.8 adalah Normal Scores yang menawarkan cara efektif untuk menormalisasikan variabel kontinu.
April 2009Bab 6 Model Pengukuran8
Variabel Kontinu dan Skor Normal
April 2009Bab 6 Model Pengukuran9
Variabel Kontinu dan Skor Normal
April 2009Bab 6 Model Pengukuran10
Variabel Kontinu dan Skor Normal
April 2009Bab 6 Model Pengukuran11
Variabel Sensor
Suatu variabel sensor adalah variabel yang mempunyai bagian dari observasi yang cukup banyak pada nilai minimum dan maksimum.
Sensor di bawah (censored below)
y = c jika y* ≤ c = y* selain itu
April 2009Bab 6 Model Pengukuran12
Variabel Sensor MAINTNCE = Rata-rata banyaknya waktu (dalam menit) setiap
hari yang dihabiskan untuk pemeliharaan (sebagai variabel dependen)
AGE = dalam tahun HOUSE = 1, jika responden tinggal di sebuah rumah, = 0 selain
itu RECHOUSE = 1, jika responden mempunyai sebuah rumah
rekreasi, = 0 selain itu CAR =1, jika responden mempunyai mobil, = 0 selain itu SCHOOLYR = lamanya responden bersekolah (dalam tahun) INCOME = disposable income dari responden (dalam mata
uang Swedia SEK) MARGTAX = Respondent’s marginal tax rate dalam %
SY= MAINTENANCE.PSFCR MAINTNCE on AGE – MARGTAXOU
April 2009Bab 6 Model Pengukuran13
Variabel Sensor
Variable MAINTNCE is censored below It has 1329 (65.76%) values = 0.000 Estimated Mean and Standard Deviation based on 2021 complete cases. Mean = -61.951 (0.028) Standard Deviation = 143.009 (0.000) Estimated Censored Regression based on 2021 complete cases. MAINTNCE = - 202.106 + 1.124*AGE + 50.517*HOUSE + 10.932*RECHOUSE Standerr (25.958) (0.308) (8.380) (8.875) Z-values -7.786 3.653 6.028 1.232 P-values 0.000 0.000 0.000 0.218 + 42.790*CAR - 4.130*SCHOOLYR + 0.357*INCOME + 0.838*MARGTAX (11.703) (1.276) (0.167) (0.407) 3.656 -3.236 2.137 2.060 0.000 0.001 0.033 0.039 + Error, R² = 0.0820
April 2009Bab 6 Model Pengukuran14
PRELIS2 PRELIS merupakan singkatan dari
preprocessor for LISREL telah tersedia sejak tahun 1986 yaitu pada LISREL 7.
Fungsi utama dari PRELIS adalah perhitungan berbagai statistik untuk digunakan sebagai input data bagi program LISREL.
Pada LISREL 7 dan LISREL 8 versi awal kita harus membuat sintaks PRELIS sebelum menjalankannya, maka pada LISREL 8.8, penggunaan PRELIS dapat dilakukan secara:
• interaktif • melalui pembuatan sintak
April 2009Bab 6 Model Pengukuran15
PRELIS2 Interaktif
April 2009Bab 6 Model Pengukuran16
PRELIS2 Interaktif
April 2009Bab 6 Model Pengukuran17
PRELIS2 Interaktif
April 2009Bab 6 Model Pengukuran18
PRELIS2 Sintak PRELIS untuk Input PSF
TI
<string>
SY=<psfname>
<commands>
OU <options>
April 2009Bab 6 Model Pengukuran19
PRELIS2 Sintak PRELIS untuk Input PSF Contoh
TI
Contoh Censored Regression
SY = MAINTENANCE.PSF
CR MAINTNCE on AGE – MARGTAX
OU
April 2009Bab 6 Model Pengukuran20
PRELIS2 Sintak PRELIS untuk Input Text File
TI
<string>
DA=<data specifications>
LA
<labels>
RA=<filename>
<commands>
OU <options>
April 2009Bab 6 Model Pengukuran21
PRELIS2 Sintak PRELIS untuk Input Text File Contoh
TI
Creating a merge data file
DA NI=3 NO=350,259
LA
ASC MSC ESC
RA=ACADSCM.DAT, ACADSCF.DAT FO
(3F1.0)
(3F1.0)
OU RA=ACADSC.DAT WI=1 ND=0
Bab 6Bab 6Model PengukuranModel Pengukuran
Dr. Setyo Hari Wijanto<[email protected]>
April 2009Bab 6 Model Pengukuran23
Confirmatory Factor Analysis
Iktisar Prosedur Spesifikasi Model Pengumpulan Data Pembuatan Program SIMPLIS Menjalankan Program SIMPLIS dan Analisis
Keluarannya• Offending Estimate Negative Error
Variance• Analisis Validitas Standardized Loading
≥0.50 atau ≥ 0.70• Uji Kecocokan Keseluruhan Model• Analisis Reliabilitas
Respesifikasi Modification Index
April 2009Bab 6 Model Pengukuran24
Confirmatory Factor Analysis
Spesifikasi
April 2009Bab 6 Model Pengukuran25
Confirmatory Factor Analysis
Pengumpulan Data dikonversikan ke Normal Score
Pembuatan Program SIMPLIS
System File from File TTFINPUT.DSF Latent Variable: Ttf Utility Performance Relationship: LEVEL= 1 * Ttf ACCURACY = Ttf LOCATABI = Ttf ACCESSIB = Ttf MEANING = Ttf ASSISTAN = Ttf EASE = Ttf CURRENCY = Ttf PRESENTA = Ttf P1= 1 * Performance P2 = Performance UT1= 1 * Utility UT2= Utility UT3= Utility Method: Weighted Least Square Options: SC Path Diagram End of Problem
April 2009Bab 6 Model Pengukuran26
Confirmatory Factor Analysis
Menjalankan Program SIMPLIS dan analisis keluaran
April 2009Bab 6 Model Pengukuran27
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan analisis
keluaran
ACCURACY = 0.63*Ttf, Errorvar.= -0.021 , R² = 1.04 (0.019) (0.051) 33.45 -0.41 W_A_R_N_I_N_G : Error variance is negative. . . P1 = 1.00*Performa, Errorvar.= -0.016, R² = 1.01 (0.26) -0.060 W_A_R_N_I_N_G : Error variance is negative.
April 2009Bab 6 Model Pengukuran28
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan analisis
keluaran
Completely Standardized Solution LAMBDA-X Ttf Utility Performa -------- -------- -------- LEVEL 0.87 - - - - ACCURACY 1.02 - - - - LOCATABI 0.81 - - - - ACCESSIB 0.89 - - - - MEANING 0.93 - - - - ASSISTAN 0.92 - - - - EASE 0.86 - - - - CURRENCY 0.97 - - - - PRESENTA 0.87 - - - - UT1 - - 0.96 - - UT2 - - 0.37 - - UT3 - - 0.88 - - P1 - - - - 1.00 P2 - - - - 0.85
April 2009Bab 6 Model Pengukuran29
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaranGoodness of Fit Statistics
Degrees of Freedom = 63
Minimum Fit Function Chi-Square = 156.74 (P = 0.00) Estimated Non-centrality Parameter (NCP) = 93.74
90 Percent Confidence Interval for NCP = (60.71 ; 134.47)
Minimum Fit Function Value = 1.02 Population Discrepancy Function Value (F0) = 0.61
90 Percent Confidence Interval for F0 = (0.40 ; 0.88) Root Mean Square Error of Approximation (RMSEA) = 0.099
90 Percent Confidence Interval for RMSEA = (0.079 ; 0.12) P-Value for Test of Close Fit (RMSEA < 0.05) = 0.00
Expected Cross-Validation Index (ECVI) = 1.39
90 Percent Confidence Interval for ECVI = (1.17 ; 1.66) ECVI for Saturated Model = 1.19
ECVI for Independence Model = 22.06
Chi-Square for Independence Model with 78 Degrees of Freedom = 3349.04
Independence AIC = 3375.04 Model AIC = 212.74
Saturated AIC = 182.00 Independence CAIC = 3427.52
Model CAIC = 325.78 Saturated CAIC = 549.36
April 2009Bab 6 Model Pengukuran30
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaran
Normed Fit Index (NFI) = 0.95 Non-Normed Fit Index (NNFI) = 0.96
Parsimony Normed Fit Index (PNFI) = 0.77 Comparative Fit Index (CFI) = 0.97 Incremental Fit Index (IFI) = 0.97 Relative Fit Index (RFI) = 0.94
Critical N (CN) = 90.81
Root Mean Square Residual (RMR) = 2.24
Standardized RMR = 0.35 Goodness of Fit Index (GFI) = 0.97
Adjusted Goodness of Fit Index (AGFI) = 0.96 Parsimony Goodness of Fit Index (PGFI) = 0.67
April 2009Bab 6 Model Pengukuran31
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaranThe Modification Indices Suggest to Add the
Path to from Decrease in Chi-Square New Estimate LEVEL Utility 20.2 1.06 MEANING Performa 8.7 0.24 ASSISTAN Utility 23.9 -2.96 CURRENCY Utility 12.0 0.48
The Modification Indices Suggest to Add an Error
Covariance Between and Decrease in Chi-Square New Estimate LOCATABI ACCURACY 9.1 0.73 ACCESSIB ACCURACY 12.1 -0.13 ACCESSIB LOCATABI 17.8 1.87 CURRENCY LEVEL 9.5 -0.12 CURRENCY ASSISTAN 10.6 0.39 PRESENTA ACCURACY 9.9 0.10 UT1 LEVEL 14.6 0.15 UT3 ACCESSIB 9.0 0.14 P1 LOCATABI 14.4 -1.40 P2 ACCESSIB 9.4 -0.33
April 2009Bab 6 Model Pengukuran32
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaran
System File from File TTFINPUT.DSF Latent Variable: Ttf Utility Performance Relationship: LEVEL= 1 * Ttf ACCURACY = Ttf LOCATABI = Ttf ACCESSIB = Ttf MEANING = Ttf ASSISTAN = Ttf EASE = Ttf CURRENCY = Ttf PRESENTA = Ttf UT1= 1 * Utility UT3 = Utility P1= 1 * Performance P2 = Performance
April 2009Bab 6 Model Pengukuran33
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaran
Set Error Variance of ACCURACY to 0.01 Set Error Variance of P1 to 0.01 Let Error Covariance of ACCESSIB and LOCATABI Free Let Error Covariance of ACCESSIB and ACCURACY Free Let Error Covariance of CURRENCY and ASSISTAN Free Let Error Covariance of PRESENTA and ACCURACY Free Let Error Covariance of CURRENCY and LEVEL Free Let Error Covariance of LOCATABI and ACCURACY Free Method: Weighted Least Square Options: SC Path Diagram End of Problem
April 2009Bab 6 Model Pengukuran34
Confirmatory Factor Analysis
Menjalankan Program SIMPLIS dan analisis keluaran
April 2009Bab 6 Model Pengukuran35
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaranGoodness of Fit Statistics Degrees of Freedom = 58
Minimum Fit Function Chi-Square = 113.50 (P = 0.00) Estimated Non-centrality Parameter (NCP) = 55.50
90 Percent Confidence Interval for NCP = (29.08 ; 89.71) Minimum Fit Function Value = 0.74
Population Discrepancy Function Value (F0) = 0.36 90 Percent Confidence Interval for F0 = (0.19 ; 0.58)
Root Mean Square Error of Approximation (RMSEA) = 0.079 90 Percent Confidence Interval for RMSEA = (0.057 ; 0.10)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.017 .
Normed Fit Index (NFI) = 0.97 Non-Normed Fit Index (NNFI) = 0.98
Parsimony Normed Fit Index (PNFI) = 0.72 Comparative Fit Index (CFI) = 0.98 Incremental Fit Index (IFI) = 0.98 Relative Fit Index (RFI) = 0.96
Critical N (CN) = 117.62
Root Mean Square Residual (RMR) = 1.76 Standardized RMR = 0.30
Goodness of Fit Index (GFI) = 0.98 Adjusted Goodness of Fit Index (AGFI) = 0.97 Parsimony Goodness of Fit Index (PGFI) = 0.62
April 2009Bab 6 Model Pengukuran36
Confirmatory Factor Analysis Menjalankan Program SIMPLIS dan
analisis keluaran Ttf
(ΣSLF)2 = (0.84+0.99+0.74+0.86+0.91+0.88+0.87+0.90+0.81)2
= (7.8)2 = 60.84
ΣSLF2 = 0.842+0.992+0.742+0.862+0.912+0.882+0.872+0.902+0.812 = 6.80
Σerrors = 0.30+0.02+0.45+0.26+0.18+0.22+0.25+0.18+0.35 = 2.21
Costruct Reliability (CR) = 60.84/(60.84+2.21) = 0.96
Variance Extracted (VE) = 6.8/(6.8+2.21) = 0.75
Utility
(ΣSLF)2 = (0.92+0.98)2 = 3.61 ΣSLF2 = 0.922+0.982 = 1.81
Σerrors = 0.14+0.04 = 0.18
CR = 3.61/(3.61+0.18) =0.95 VE = 1.81/(1.81+0.18) = 0.91
Performance
(ΣSLF)2 = (1.00+0.84)2 = 3.39 ΣSLF2 = 1.002+0.842 = 1.71
Σerrors = 0.00+0.30 = 0.30
CR = 3.39/(3.39+0.30) =0.92 VE = 1.71/(1.71+0.30) = 0.85
Catatan: SLF = Standardized Loading Factors
April 2009Bab 6 Model Pengukuran37
Second Order Confirmatory Factor Analysis
April 2009Bab 6 Model Pengukuran38
Second Order Confirmatory Factor AnalysisSystem File from File DATAKBR.DSF Latent Variables: Social External Combina Internal Kbr Relationships KBRS1 = 1* Social KBRS2 KBRS3 = Social KBRE1 = 1* External KBRE2 KBRE3 = External KBRC1 = 1* Combina KBRC2 KBRC3 = Combina KBRI1 = 1* Internal KBRI2 KBRI3 = Internal Social External Combina Internal = Kbr Set Error Variance of Social to 0.01 Let Error Between KBRE3 and KBRE1 Correlate Let Error Between KBRC1 and KBRE1 Correlate Let Error Between KBRC1 and KBRE3 Correlate Let Error Between KBRC3 and KBRC2 Correlate Let Error Between KBRI1 and KBRS3 Correlate Let Error Between KBRI2 and KBRE1 Correlate Let Error Between KBRS2 and KBRS1 Correlate Let Error Between KBRC3 and KBRE2 Correlate Method: Weighted Least Square Options: SC Path Diagram End Of Problem
April 2009Bab 6 Model Pengukuran39
Second Order Confirmatory Factor Analysis
April 2009Bab 6 Model Pengukuran40
Second Order Confirmatory Factor Analysis
Variabel Standardized Loading Factors
Errors Reliabilitas Keterangan
≥ 0.50 CR≥0.70 VE≥0.50 1stCFA Social
KBRS1 KBRS2 KBRS3
External
KBRE1 KBRE2 KBRE3
Combina
KBRC1 KBRC2 KBRC3
Internal
KBRI1 KBRI2 KBRI3
0.80 0.75 0.65
0.81 0.93 0.64
0.78 0.87 0.82
0.69 0.90 0.80
0.36 0.43 0.58
0.34 0.13 0.58
0.40 0.24 0.33
0.53 0.19 0.36
0.78
0.84
0.86
0.84
0.54
0.65
0.68
0.64
Reliabilitas baik Validitas baik Validitas baik Validitas baik Reliabilitas baik Validitas baik Validitas baik Validitas baik Reliabilitas baik Validitas baik Validitas baik Validitas baik Reliabilitas baik Validitas baik Validitas baik Validitas baik
2ndCFA Kbr
Social External
Combina Internal
0.99 0.79 0.95 0.97
0.02 0.37 0.10 0.06
0.96
0.86
Reliabilitas baik Validitas baik Validitas baik Validitas baik Validitas baik
April 2009Bab 6 Model Pengukuran41
Latent Variable Score (LVS)
April 2009Bab 6 Model Pengukuran42
Latent Variable Score (LVS)
April 2009Bab 6 Model Pengukuran43
Latent Variable Score (LVS)
Program SIMPLIS untuk menghitung Latent Variable Score (LVS)
System File from File Self.DSF
Latent Variables Self
Relationships
SELF1 - SELF5 = Self
Let Error Covariance of SELF5 and SELF1 Free
Let Error Covariance of SELF4 and SELF1 Free
!Statemen untuk menghitung Latent Variable Score
PSFFile Self.PSF
Path Diagram End of Problem
April 2009Bab 6 Model Pengukuran44
Latent Variable Score (LVS)
April 2009Bab 6 Model Pengukuran45
Latent Variable Score (LVS)
April 2009Bab 6 Model Pengukuran46
Latent Variable Score (LVS)
Self
Gambar 6.22.a.
Self Selfest
Gambar 6.22.b.
1
0
April 2009Bab 6 Model Pengukuran47
Latent Variable Score (LVS) Penyederhanaan Model
Infoacq Infodis Sharint Decmem Procmem
IT_Comp
Org_Learn
Firm_Perf
ITKnow ITOps ITObj
IA1 IA6 ID1 ID6 SI1 SI5 DM1 DM7 PM1
FP1
FP4
IK1 IK4 IO1 IO6 IT1 IT5
PM5
April 2009Bab 6 Model Pengukuran48
Latent Variable Score (LVS) Penyederhanaan Model
Infoacq Infodis Sharint Decmem
ITknow ITops ITobj
Procmem
FirmLVS
Org_learn
It_comp
Firm_Perf
10
Bab 7Bab 7Model StrukturalModel Struktural
Dr. Setyo Hari Wijanto<[email protected]>
April 2009Bab 6 Model Pengukuran50
Model Struktural Rekursif
April 2009Bab 6 Model Pengukuran51
Model Struktural RekursifSystem File from File TTFINPUT.DSF Latent Variable: Ttf Utility Performance Relationship: LEVEL= 1 * Ttf ACCURACY = Ttf LOCATABI = Ttf ACCESSIB = Ttf MEANING = Ttf ASSISTAN = Ttf EASE = Ttf CURRENCY = Ttf PRESENTA = Ttf UT1= 1 * Utility UT3 = Utility P1= 1 * Performance P2 = Performance Performance = Ttf Utility Utility = Ttf Set Error Variance of ACCURACY to 0.01 Set Error Variance of P1 to 0.01 Let Error Covariance of ACCESSIB and LOCATABI Free Let Error Covariance of ACCESSIB and ACCURACY Free Let Error Covariance of CURRENCY and ASSISTAN Free Let Error Covariance of PRESENTA and ACCURACY Free Let Error Covariance of CURRENCY and LEVEL Free Let Error Covariance of LOCATABI and ACCURACY Free Method: Weighted Least Square Options: SC Path Diagram End of Problem
April 2009Bab 6 Model Pengukuran52
Model Struktural Rekursif
April 2009Bab 6 Model Pengukuran53
Model Struktural Rekursif
Goodness of Fit Statistics
Degrees of Freedom = 58 Minimum Fit Function Chi-Square = 113.50 (P = 0.00)
Estimated Non-centrality Parameter (NCP) = 55.50 90 Percent Confidence Interval for NCP = (29.08 ; 89.71)
Minimum Fit Function Value = 0.74
Population Discrepancy Function Value (F0) = 0.36 90 Percent Confidence Interval for F0 = (0.19 ; 0.58)
Root Mean Square Error of Approximation (RMSEA) = 0.079 90 Percent Confidence Interval for RMSEA = (0.057 ; 0.10)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.017
Normed Fit Index (NFI) = 0.97 Non-Normed Fit Index (NNFI) = 0.98
Parsimony Normed Fit Index (PNFI) = 0.72 Comparative Fit Index (CFI) = 0.98 Incremental Fit Index (IFI) = 0.98
Relative Fit Index (RFI) = 0.96
Critical N (CN) = 117.62 Root Mean Square Residual (RMR) = 1.76
Standardized RMR = 0.30 Goodness of Fit Index (GFI) = 0.98
Adjusted Goodness of Fit Index (AGFI) = 0.97 Parsimony Goodness of Fit Index (PGFI) = 0.62
April 2009Bab 6 Model Pengukuran54
Model Struktural Rekursif
Structural Equations Utility = 0.46*Ttf, Errorvar.= 0.23 , R² = 0.55 (0.024) (0.031) 19.09 7.32 Performa = - 0.33*Utility + 0.96*Ttf, Errorvar.= 1.53 , R² = 0.37 (0.25) (0.14) (0.22) -1.34 6.72 7.02 Reduced Form Equations Utility = 0.46*Ttf, Errorvar.= 0.23, R² = 0.55 (0.024) 19.09 Performa = 0.81*Ttf, Errorvar.= 1.56, R² = 0.36 (0.071) 11.32
April 2009Bab 6 Model Pengukuran55
Model Struktural Rekursif
Path Estimasi Nilai - t Kesimpulan
1 Ttf Utility 0.46 19.09 Signifikan
2 Utility Performance -0.33 -1.34 Tdk Signifikan
3 Ttf Performance 0.96 6.72 Signifikan