Model 1 pgm

ods html file='c:\mvar\semmdl1.html';

proc calis data=marsem nobs=1200 all;
lineqs
econ = lambda1 F1 + E1,
pay = lambda2 F1 + E2,
infla = F1 + E3,
invest = lambda3 F1 + E4,
finance= lambda4 F1 + E5,
ptrust = lambda5 F2 + E6,
graft = F2 + E7,
crim = lambda6 F3 + E8,
peace = F3 + E9,
terror = lambda7 F3 + E10,
law = lambda8 F4 + E11,
poverty = F4 + E12,
envi = F5 + E13,
power =lambda9 F5 + E14,
rtnoli = F6 + E15,
rtsen = lambda10 F6 + E16,
rthouse= lambda11 F6 + E17,
rtsc = lambda12 F6 + E18,
ppqol = lambda13 F7 + E19,
fpqol = lambda14 F7 + E20,
pnqol = F7 + E21,
fnqol = lambda15 F7 + E22,
rtgma = lambda16 F8 + E23,
trustgma = lambda17 F8 + E24,

F7 = beta1 F8 + gamma1 F1 + gamma2 F2 + beta2 F3

+ beta3 F4 + beta4 F5 + beta5 F6 + D1,
F8 = gamma3 F1 + gamma4 F2 + D2,
F3 = gamma5 F1 + gamma6 F2 + beta6 F8 + D3,
F4 = gamma7 F1 + gamma8 F2 + beta7 F8 + D4,
F5 = gamma9 F1 + gamma10 F2 + beta8 F8 + D5,
F6 = gamma11 F1 + gamma12 F2 + beta9 F8 + D6;
STD
E1 - E24 = 24*U:,
D1 - D6 = 6*V:,
F1 - F2 = 2*W:;
run;
ods html close;

71
 The SAS System
The CALIS Procedure
Covariance Structure Analysis: Maximum Likelihood Estimation
Fit Function 1.4378

Goodness of Fit Index (GFI) 0.9039

GFI Adjusted for Degrees of Freedom (AGFI) 0.8747

Root Mean Square Residual (RMR) 0.1332

Parsimonious GFI (Mulaik, 1989) 0.7533

Chi-Square 1723.9504

Chi-Square DF 230

Pr > Chi-Square <.0001

Independence Model Chi-Square 10267

Independence Model Chi-Square DF 276

RMSEA Estimate 0.0736

RMSEA 90% Lower Confidence Limit 0.0704

RMSEA 90% Upper Confidence Limit 0.0769

ECVI Estimate 1.5571

ECVI 90% Lower Confidence Limit 1.4491

ECVI 90% Upper Confidence Limit 1.6714

Probability of Close Fit 0.0000

Bentler's Comparative Fit Index 0.8505

Elliptic Corrected Chi-Square 1401.3354

Pr > Elliptic Corrected Chi-Square <.0001

72
Normal Theory Reweighted LS Chi-Square 1529.0877

Akaike's Information Criterion 1263.9504

Bozdogan's (1987) CAIC -136.7673

Schwarz's Bayesian Criterion 93.2327

McDonald's (1989) Centrality 0.5366

Bentler & Bonett's (1980) Non-normed Index 0.8206

Bentler & Bonett's (1980) NFI 0.8321

James, Mulaik, & Brett (1982) Parsimonious NFI 0.6934

Z-Test of Wilson & Hilferty (1931) 30.8203

Bollen (1986) Normed Index Rho1 0.7985

Bollen (1988) Non-normed Index Delta2 0.8512

Hoelter's (1983) Critical N 187

73
The CALIS Procedure
Covariance Structure Analysis: Maximum Likelihood Estimation

Manifest Variable Equations with Estimates

ppqol = 1.2770*F7 + 1.0000 E19

Std Err 0.0732 lambda13
t Value 17.4489
fpqol = 1.3729*F7 + 1.0000 E20
Std Err 0.0796 lambda14
t Value 17.2579
pnqol = 1.0000 F7 + 1.0000 E21
fnqol = 1.3251*F7 + 1.0000 E22
Std Err 0.0762 lambda15
t Value 17.3851
rtsen = 1.9674*F6 + 1.0000 E16
Std Err 0.1068 lambda10
t Value 18.4262
rthouse = 1.9058*F6 + 1.0000 E17
Std Err 0.1022 lambda11
t Value 18.6459
rtsc = 1.8250*F6 + 1.0000 E18
Std Err 0.0967 lambda12
t Value 18.8755
rtgma = 0.6047*F8 + 1.0000 E23
Std Err 0.0888 lambda16
t Value 6.8127
rtnoli = 1.0000 F6 + 1.0000 E15
econ = 0.9279*F1 + 1.0000 E1
Std Err 0.0528 lambda1
t Value 17.5631
crim = 0.8326*F3 + 1.0000 E8
Std Err 0.0503 lambda6
t Value 16.5445
law = 0.8585*F4 + 1.0000 E11
Std Err 0.0514 lambda8
t Value 16.7101
pay = 0.9943*F1 + 1.0000 E2
Std Err 0.0556 lambda2
t Value 17.8769
infla = 1.0000 F1 + 1.0000 E3
poverty = 1.0000 F4 + 1.0000 E12
ptrust = 1.0437*F2 + 1.0000 E6
Std Err 0.0731 lambda5
t Value 14.2689
envi = 1.0000 F5 + 1.0000 E13
peace = 1.0000 F3 + 1.0000 E9
graft = 1.0000 F2 + 1.0000 E7
invest = 0.8666*F1 + 1.0000 E4
Std Err 0.0507 lambda3
t Value 17.0850
power = 0.9378*F5 + 1.0000 E14
Std Err 0.0540 lambda9
t Value 17.3718
finance = 0.9103*F1 + 1.0000 E5
Std Err 0.0522 lambda4
t Value 17.4436

74
 The SAS System

The CALIS Procedure

Covariance Structure Analysis: Maximum Likelihood Estimation

terror = 0.8861*F3 + 1.0000 E10

Std Err 0.0521 lambda7
t Value 16.9940
trustgma = 0.7045*F8 + 1.0000 E24
Std Err 0 lambda17
t Value .

The SAS System

The CALIS Procedure

Covariance Structure Analysis: Maximum Likelihood Estimation

Latent Variable Equations with Estimates

F7 = 0.00555*F8 + 0.1315*F3
Std Err 0.4543 beta1 0.2216 beta2
t Value 0.0122 0.5937

+ 0.0400*F4 + 2.4975*F5 + -0.1251*F6

0.3333 beta3 8.5947 beta4 0.0479 beta5
0.1199 0.2906 -2.6125

+ -1.7396*F1 + -1.2993*F2 + 1.0000 D1

7.0316 gamma1 4.7850 gamma2
-0.2474 -0.2715

F8 = 0.6822*F1 + 0.2456*F2
Std Err 0.0458 gamma3 0.0434 gamma4
t Value 14.8987 5.6539

+ 1.0000 D2

F3 = 0.0121*F8 + 0.6499*F1
Std Err 0.0505 beta6 0.0649 gamma5
t Value 0.2405 10.0133

+ 0.4368*F2 + 1.0000 D3
0.0541 gamma6
8.0666

F4 = -0.00456*F8 + 0.7560*F1
Std Err 0.0502 beta7 0.0669 gamma7
t Value -0.0908 11.2979

+ 0.3782*F2 + 1.0000 D4
0.0504 gamma8
7.5082

75
 F5 = -0.0223*F8 + 0.6291*F1
Std Err 0.1436 beta8 0.1789 gamma9
t Value -0.1550 3.5158

The SAS System

The CALIS Procedure

Covariance Structure Analysis: Maximum Likelihood Estimation

+ 0.4764*F2 + 1.0000 D5
0.1558 gamma10
3.0574

F6 = 0.0649*F8 + 0.0704*F1
Std Err 0.0558 beta9 0.0496 gamma11
t Value 1.1631 1.4189

+ 0.0747*F2 + 1.0000 D6
0.0453 gamma12
1.6504

The SAS System

The CALIS Procedure

Covariance Structure Analysis: Maximum Likelihood Estimation

Stepwise Multivariate Wald Test

----Cumulative Statistics--- -Univariate Increment-

Parameter Chi-Square DF Pr > ChiSq Chi-Square Pr > ChiSq

V5 5.29031E-7 1 0.9994 5.29031E-7 0.9994

beta1 0.0001655 2 0.9999 0.0001650 0.9898
beta8 0.0004284 3 1.0000 0.0002628 0.9871
beta7 0.00126 4 1.0000 0.0008293 0.9770
beta3 0.00803 5 1.0000 0.00677 0.9344
beta6 0.01826 6 1.0000 0.01023 0.9194
V6 0.02851 7 1.0000 0.01025 0.9193
V4 0.08964 8 1.0000 0.06113 0.8047
beta9 0.15398 9 1.0000 0.06434 0.7998
V1 0.27218 10 1.0000 0.11820 0.7310
V3 0.42930 11 1.0000 0.15712 0.6918
beta2 0.79374 12 1.0000 0.36444 0.5460
gamma11 1.34169 13 1.0000 0.54794 0.4592
gamma9 1.92587 14 0.9999 0.58419 0.4447
gamma10 2.40026 15 0.9999 0.47439 0.4910
gamma12 2.93919 16 0.9999 0.53894 0.4629
U24 3.91519 17 0.9996 0.97600 0.3232
gamma1 5.75917 18 0.9971 1.84398 0.1745

76
ods html file='c:\mvar\semmdl2.html';
proc calis data=marsem nobs=1200 all;
lineqs
econ = lambda1 F1 + E1,
pay = lambda2 F1 + E2,
infla = F1 + E3,
invest = lambda3 F1 + E4,
finance= lambda4 F1 + E5,
ptrust = lambda5 F2 + E6,
graft = F2 + E7,
crim = lambda6 F3 + E8,
peace = F3 + E9,
terror = lambda7 F3 + E10,
law = lambda8 F4 + E11,
poverty = F4 + E12,
envi = F5 + E13,
power =lambda9 F5 + E14,
rtnoli = F6 + E15,
rtsen = lambda10 F6 + E16,
rthouse= lambda11 F6 + E17,
rtsc = lambda12 F6 + E18,
ppqol = lambda13 F7 + E19,
fpqol = lambda14 F7 + E20,
pnqol = F7 + E21,
fnqol = lambda15 F7 + E22,
rtgma = lambda16 F8 + E23,
trustgma = lambda17 F8 + E24,
F7 = beta1 F8 + D1,
F8 = gamma3 F1 + gamma4 F2 + D2,
F3 = gamma5 F1 + gamma6 F2 + D3,
F4 = gamma7 F1 + gamma8 F2 + D4,
F5 = gamma9 F1 + gamma10 F2 + D5,
F6 = beta2 F8 + D6;
STD
E1 - E24 = 24*U:,
D1 - D6 = 6*V:,
F1 - F2 = 2*W:;
run;
ods html close;

77
 The SAS System
The CALIS Procedure
Covariance Structure Analysis: Maximum Likelihood Estimation
Fit Function 1.4899

Goodness of Fit Index (GFI) 0.8986

GFI Adjusted for Degrees of Freedom (AGFI) 0.8738

Root Mean Square Residual (RMR) 0.1388

Parsimonious GFI (Mulaik, 1989) 0.7847

Chi-Square 1786.4371

Chi-Square DF 241

Pr > Chi-Square <.0001

Independence Model Chi-Square 10267

Independence Model Chi-Square DF 276

RMSEA Estimate 0.0731

RMSEA 90% Lower Confidence Limit 0.0700

RMSEA 90% Upper Confidence Limit 0.0763

ECVI Estimate 1.5905

ECVI 90% Lower Confidence Limit 1.4805

ECVI 90% Upper Confidence Limit 1.7067

Probability of Close Fit 0.0000

Bentler's Comparative Fit Index 0.8453

Elliptic Corrected Chi-Square 1452.1285

Pr > Elliptic Corrected Chi-Square <.0001

78
Normal Theory Reweighted LS Chi-Square 1622.6963

Akaike's Information Criterion 1304.4371

Bozdogan's (1987) CAIC -163.2714

Schwarz's Bayesian Criterion 77.7286

McDonald's (1989) Centrality 0.5252

Bentler & Bonett's (1980) Non-normed Index 0.8229

Bentler & Bonett's (1980) NFI 0.8260

James, Mulaik, & Brett (1982) Parsimonious NFI 0.7213

Z-Test of Wilson & Hilferty (1931) 31.3090

Bollen (1986) Normed Index Rho1 0.8007

Bollen (1988) Non-normed Index Delta2 0.8459

Hoelter's (1983) Critical N 188

79
Stepwise Multivariate Wald Test

----Cumulative Statistics--- -Univariate Increment-

Parameter Chi-Square DF Pr > ChiSq Chi-Square Pr > ChiSq

V5 0.27649 1 0.5990 0.27649 0.5990

SEM Model 3:

proc calis data=c.mar06sem nobs=1200 pdeterm pestim toteff mod;

lineqs
econ = lambda1 F1 + E1,
pay = lambda2 F1 + E2,
infla = F1 + E3,
invest = lambda3 F1 + E4,
finance= lambda4 F1 + E5,
ptrust = lambda5 F2 + E6,
graft = F2 + E7,
crim = lambda6 F3 + E8,
peace = F3 + E9,
terror = lambda7 F3 + E10,
law = lambda8 F4 + E11,
poverty = F4 + E12,
envi = F5 + E13,
power =lambda9 F5 + E14,
rtnoli = F6 + E15,
rtsen = lambda10 F6 + E16,
rthouse= lambda11 F6 + E17,
rtsc = lambda12 F6 + E18,
ppqol = lambda13 F7 + E19,
fpqol = lambda14 F7 + E20,
pnqol = F7 + E21,
fnqol = lambda15 F7 + E22,
rtgma = lambda16 F8 + E23,
trustgma = lambda17 F8 + E24,

F2 = beta1 F1 + beta2 F8 + D1,

F3 = beta2 F1 + beta3 F8 + D2,
F4 = beta4 F1 + beta5 F8 + D3,
F5 = beta6 F1 + beta7 F8 + D4,
F6 = beta8 F1 + gamma1 F2 + beta9 F8 + D5,
F7 = beta10 F1 + beta11 F8 + D6,
resign = beta12 F8 + beta13 cheat
+ beta14 F1
+ D7;
STD
E1 - E24 = 24*U:,
D1 - D7 = 7*V:,
F1 = W1,
F8 = W2;
run;

80
 The CALIS Procedure
Covariance Structure Analysis: Maximum Likelihood Estimation

Fit Function 1.1641

Goodness of Fit Index (GFI) 0.9112
GFI Adjusted for Degrees of Freedom (AGFI) 0.8906
Root Mean Square Residual (RMR) 0.1008
Parsimonious GFI (Mulaik, 1989) 0.7990
Chi-Square 1395.7340
Chi-Square DF 285
Pr > Chi-Square <.0001
Independence Model Chi-Square 10890
Independence Model Chi-Square DF 325
RMSEA Estimate 0.0570
RMSEA 90% Lower Confidence Limit 0.0540
RMSEA 90% Upper Confidence Limit 0.0600
ECVI Estimate 1.2750
ECVI 90% Lower Confidence Limit 1.1809
ECVI 90% Upper Confidence Limit 1.3772
Probability of Close Fit 0.0001
Bentler's Comparative Fit Index 0.8949
Normal Theory Reweighted LS Chi-Square 1519.2247
Akaike's Information Criterion 825.7340
Bozdogan's (1987) CAIC -909.9379
Schwarz's Bayesian Criterion -624.9379
McDonald's (1989) Centrality 0.6295
Bentler & Bonett's (1980) Non-normed Index 0.8801
Bentler & Bonett's (1980) NFI 0.8718
James, Mulaik, & Brett (1982) Parsimonious NFI 0.7645
Z-Test of Wilson & Hilferty (1931) 25.0315
Bollen (1986) Normed Index Rho1 0.8538
Bollen (1988) Non-normed Index Delta2 0.8953
Hoelter's (1983) Critical N 281

Output of pdeterm

Squared Multiple Correlations

Error Total
Variable Variance Variance R-Square

1 ppqol 0.66575 0.99442 0.3305

2 fpqol 0.61425 0.99356 0.3818
3 pnqol 0.79581 0.99658 0.2015
4 fnqol 0.64457 0.99406 0.3516
5 rtsen 0.56481 0.97933 0.4233
6 rthouse 0.59088 0.98057 0.3974
7 rtsc 0.61813 0.98187 0.3704
8 rtgma 0.33922 1.00000 0.6608
9 rtnoli 0.88283 0.99443 0.1122
10 econ 0.58202 1.00000 0.4180
11 crim 0.62369 0.97677 0.3615
12 law 0.65064 0.97723 0.3342
13 pay 0.50025 1.00000 0.4998
14 infla 0.48122 0.99999 0.5188

81
 15 poverty 0.52098 0.96878 0.4622
16 ptrust 0.47513 0.97670 0.5135
17 envi 0.52962 0.98563 0.4627
18 peace 0.45784 0.96653 0.5263
19 graft 0.50208 0.97789 0.4866
20 invest 0.59161 1.00000 0.4084
21 power 0.58565 0.98735 0.4068
22 finance 0.56961 1.00000 0.4304
23 terror 0.57286 0.97363 0.4116
24 trustgma 0.27704 1.00000 0.7230
25 resign 0.75705 0.93168 0.1874
26 F2 0.01745 0.47581 0.9633
27 F3 0.06808 0.50869 0.8662
28 F4 -0.04059 0.44780 1.0906
29 F5 0.01402 0.45601 0.9693
30 F6 0.09186 0.11160 0.1769
31 F7 0.18951 0.20077 0.0561

WARNING: Total determination of all equations cannot be

computed due to singular matrix. You may be
successful by using a smaller value for the SINGULAR=
option.

Total Determination of All Equations .

WARNING: Total determination of structural equations cannot be

computed due to singular matrix. You may be
successful by using a smaller value for the SINGULAR=
option.

Total Determination of Structural Equations .

Structural Equations Corresponding to the Dependent Variables

ppqol fpqol pnqol fnqol rtsen rthouse rtsc rtnoli crim

law poverty ptrust envi peace graft power terror F2 F3

F4 F5 F6 F7

Total Determination of Manifest Variables 0.995

82
 Stepwise Multivariate Wald Test

----Cumulative Statistics--- -Univariate Increment-

Parameter Chi-Square DF Pr > ChiSq Chi-Square Pr > ChiSq

beta8 0.21495 1 0.6429 0.21495 0.6429

V4 0.75196 2 0.6866 0.53701 0.4637
W2 2.68168 3 0.4433 1.92972 0.1648
beta14 5.24649 4 0.2629 2.56480 0.1093
V1 8.69404 5 0.1219 3.44755 0.0633

resign = 0.1811*cheat + -0.0583*F1

Std Err 0.0255 beta13 0.0371 beta14
t Value 7.0917 -1.5680

+ -6.6200*F8 + 1.0000 D7
2.4463 beta12
-2.7061

F2 = 0.9372*F1 + 0.9148*F8
Std Err 0.0412 beta1 0.0407 beta2
t Value 22.7511 22.4880

+ 1.0000 D1

F3 = 0.9148*F1 + 1.4190*F8
Std Err 0.0407 beta2 0.5473 beta3
t Value 22.4880 2.5927

+ 1.0000 D2

F4 = 0.9653*F1 + 1.2545*F8
Std Err 0.0409 beta4 0.4999 beta5
t Value 23.5990 2.5096

+ 1.0000 D3

F5 = 0.9218*F1 + 0.6045*F8
Std Err 0.0414 beta6 0.3691 beta7
t Value 22.2783 1.6375

+ 1.0000 D4

F6 = 0.3511*F2 + -0.2015*F1
Std Err 0.4617 gamma1 0.4346 beta8

83
 t Value 0.7604 -0.4636

+ 1.3705*F8 + 1.0000 D5
0.7630 beta9
1.7963

F7 = -0.0878*F1 + -1.5071*F8
Std Err 0.0238 beta10 0.6285 beta11
t Value -3.6898 -2.3981

+ 1.0000 D6

Rerunning without beta8 and beta14:

One major advantage of SEM is that in addition to direct structural effects, indirect
effects through intervening variables can be estimated.

Total Effects

cheat F1 F8 F2 F3

resign 0.18594 0.0000 -6.614 0.0000 0.0000

F2 0.00000 0.9351 0.914 0.0000 0.0000
F3 0.00000 0.9135 1.384 0.0000 0.0000
F4 0.00000 0.9643 1.218 0.0000 0.0000
F5 0.00000 0.9206 0.599 0.0000 0.0000
F6 0.00000 0.1285 1.615 0.1375 0.0000
F7 0.00000 -0.0872 -1.452 0.0000 0.0000

Indirect Effects

cheat F1 F8 F2 F3

resign 0 0.00000 0.0000 0.00000 0

F2 0 0.00000 0.0000 0.00000 0
F3 0 0.00000 0.0000 0.00000 0
F4 0 0.00000 0.0000 0.00000 0
F5 0 0.00000 0.0000 0.00000 0
F6 0 0.12854 0.1256 0.00000 0
F7 0 0.00000 0.0000 0.00000 0

84