SYSTAT
Rectangular file D:\mydocs\ys209\gss89m16.syd,
created
Fri Apr 19, 2002 at 10:03:50, contains variables:
ID
PRESTIGE PAPRES16 AGE
ZODIAC EDUC
PAEDUC
MAEDUC SEX
RACE WORDSUM
>idvar=id
>print=long
>if
educ>90 then let educ=.
>if
paeduc>90 then let paeduc=.
>if
maeduc>90 then let maeduc=.
>if
wordsum>90 then let wordsum=.
>corr
>save
gss89imp.syd/data
>pearson
papres16 paeduc maeduc wordsum educ prestige/em
NOTE:
Case is an outlier. Mahalanobis D^2=22.676698 z=3.090202
NOTE:
Case is an outlier. Mahalanobis D^2=28.732751 z=3.754679
NOTE:
Case is an outlier. Mahalanobis D^2=19.906854 z=3.236595
NOTE:
Case is an outlier. Mahalanobis D^2=21.640463 z=3.197591
NOTE:
Case is an outlier. Mahalanobis D^2=24.257678 z=3.274090
NOTE:
Case is an outlier. Mahalanobis D^2=15.982172 z=3.015025
NOTE:
Case is an outlier. Mahalanobis D^2=24.746567 z=3.551029
NOTE:
Case is an outlier. Mahalanobis D^2=21.738339 z=3.452252
NOTE:
Case is an outlier. Mahalanobis D^2=18.530947 z=3.065711
NOTE:
Case is an outlier. Mahalanobis D^2=22.579376 z=3.078607
NOTE:
Case is an outlier. Mahalanobis D^2=23.856495 z=3.228203
NOTE:
Case is an outlier. Mahalanobis D^2=20.329374 z=3.038216
NOTE:
Case is an outlier. Mahalanobis D^2=27.096911 z=3.799247
NOTE:
Case is an outlier. Mahalanobis D^2=31.988545 z=4.073729
EM
Algorithm Iteration Maximum Error -2*log(likelihood)
--------- ------------- ------------------
1 1.077
339951.313
2 7.495
247586.041
3 8.858
195353.708
4 0.192
162506.910
5 0.035
145462.139
6 0.025
137537.631
7 0.014
133855.878
8 0.006
132230.165
9 0.003
131574.375
10 0.002
131328.433
11 0.001
131239.105
No.of Missing value patterns
Cases
(X=nonmissing; .=missing)
41 XXXXX.
35 X...XX
365 XXX.XX
100 ..XXXX
628 XXXXXX
13 XXX.X.
2 .XXXX.
38 X.XXXX
5 X.XXX.
47 XX.XXX
61 ..X.XX
27 XX..XX
31 ...XXX
4 ...XX.
55 X..XXX
3 .XXXXX
2 X...X.
8 ..XXX.
1 ..X..X
25 X.X.XX
17 ....XX
2 .....X
8 ..X.X.
3 XX..X.
2 .XX.XX
3 XXXX.X
2 ....X.
2 X.X.X.
1 ......
3 X..XX.
3 XX.XX.
Little MCAR test statistic: 256.741 df = 99 prob = 0.000
NOTE: the p-value is 0.000 so we reject the null hypothesis that the values for the 6 variables are MCAR (Missing Completely At Random). PS: It is "Little's MCAR test" (i.e., the MCAR test invented by Roderick Little) not "a little MCAR test" (as opposed to a "big" MCAR test), in case you wondered.
EM estimate of means
PAPRES16 PAEDUC
MAEDUC WORDSUM
EDUC PRESTIGE
40.992 10.093
10.397 5.889
12.727 41.046
EM estimated correlation matrix
PAPRES16 PAEDUC
MAEDUC WORDSUM
EDUC PRESTIGE
PAPRES16
1.000
PAEDUC
0.453 1.000
MAEDUC
0.334 0.701
1.000
WORDSUM
0.155 0.281
0.307 1.000
EDUC
0.242 0.462
0.469 0.505
1.000
PRESTIGE
0.171 0.242
0.209 0.338
0.521 1.000
Pairwise frequency table
PAPRES16 PAEDUC
MAEDUC WORDSUM
EDUC PRESTIGE
PAPRES16
1295
PAEDUC
1130 1137
MAEDUC
1120 1057
1305
WORDSUM
823 727
828 971
EDUC
1292 1134
1301 968
1530
PRESTIGE
1223 1075
1226 905
1434 1440
Matrix has been saved.
>USE "D:\mydocs\ys209\gss89imp.syd"
SYSTAT
Rectangular file D:\mydocs\ys209\gss89imp.syd,
created
Mon Apr 22, 2002 at 09:43:04, contains variables:
PAPRES16 PAEDUC MAEDUC WORDSUM EDUC PRESTIGE
>stats
>stat
PAPRES16 PAEDUC
MAEDUC WORDSUM
EDUC PRESTIGE
N of cases
1536 1536
1536 1536
1536 1536
Minimum
12.000 0.000
0.000 0.000
0.000 12.000
Maximum
82.000 20.000
20.000 10.000
20.000 82.000
Mean
40.992 10.093
10.397 5.889
12.727 41.046
Standard Dev 12.208
4.083 3.509
1.903 3.031
14.178
>regress
>print
>rem
say print to cancel out previous print=long
>model
wordsum=constant+paeduc+maeduc+papres16
>estimate
1 case(s) deleted due to missing data.
Dep
Var: WORDSUM N: 1536 Multiple R: 0.391
Squared multiple R: 0.153
Adjusted
squared multiple R: 0.151 Standard error of estimate: 1.753
Effect
Coefficient Std Error Std Coef
Tolerance t P(2 Tail)
CONSTANT
3.586 0.179
0.000 . 20.076
0.000
PAEDUC
0.071 0.018
0.152 0.386 4.022
0.000
MAEDUC
0.136 0.019
0.250 0.437 7.028
0.000
PAPRES16
0.004 0.004
0.028 0.764 1.032
0.302
Analysis of Variance
Source Sum-of-Squares df Mean-Square F-ratio P
Regression
849.701 3 283.234
92.141 0.000
Residual
4709.232 1532 3.074
----------------------------------------------------------------------------------------------------------------------------------
***
WARNING ***
Case
113 has large leverage (Leverage =
0.022)
Case
1077 has large leverage (Leverage =
0.019)
Case
1200 has large leverage (Leverage =
0.018)
Durbin-Watson
D Statistic 1.827
First
Order Autocorrelation 0.085
>print
>model
educ=constant+wordsum+paeduc+maeduc+papres16
>estimate
1 case(s) deleted due to missing data.
Dep
Var: EDUC N: 1536 Multiple R: 0.676
Squared multiple R: 0.457
Adjusted
squared multiple R: 0.456 Standard error of estimate: 2.236
Effect
Coefficient Std Error Std Coef
Tolerance t P(2 Tail)
CONSTANT
5.327 0.256
0.000 . 20.804
0.000
WORDSUM
0.723 0.033
0.454 0.847 22.196
0.000
PAEDUC
0.160 0.023
0.215 0.382 7.059
0.000
MAEDUC
0.138 0.025
0.159 0.423 5.502
0.000
PAPRES16
0.002 0.005
0.010 0.763 0.449
0.654
Analysis of Variance
Source
Sum-of-Squares df Mean-Square
F-ratio P
Regression
6450.367 4 1612.592
322.469 0.000
Residual
7656.167 1531 5.001
----------------------------------------------------------------------------------------------------------------------------------
***
WARNING ***
Case
113 has large leverage (Leverage =
0.023)
Case
1381 is an outlier (Studentized
Residual = -4.671)
Durbin-Watson
D Statistic 1.809
First
Order Autocorrelation 0.095
>model
prestige=constant+educ+wordsum+paeduc+maeduc+papres16
>estimate
1 case(s) deleted due to missing data.
Dep
Var: PRESTIGE N: 1536 Multiple R: 0.551
Squared multiple R: 0.304
Adjusted
squared multiple R: 0.302 Standard error of estimate: 11.847
Effect
Coefficient Std Error Std Coef
Tolerance t P(2 Tail)
CONSTANT
7.291 1.536
0.000 .
4.746 0.000
EDUC
2.196 0.135
0.470 0.543 16.223
0.000
WORDSUM
1.060 0.198
0.142 0.641 5.343
0.000
PAEDUC
0.118 0.122
0.034 0.370 0.973
0.331
MAEDUC
-0.438 0.134
-0.108 0.415 -3.272
0.001
PAPRES16
0.071 0.028
0.061 0.763 2.505
0.012
Analysis of Variance
Source
Sum-of-Squares df Mean-Square
F-ratio P
Regression
93831.517 5 18766.303
133.714 0.000
Residual
214729.575 1530 140.346
----------------------------------------------------------------------------------------------------------------------------------
***
WARNING ***
Case
113 has large leverage (Leverage =
0.023)
Case
1381 has large leverage (Leverage =
0.025)
Durbin-Watson
D Statistic 2.045
First
Order Autocorrelation -0.025
>rem
re-estimate last model without the 2 high leverage observations
>just
out of curiosity
>select
case<>113 and case<>1381
>model
prestige=constant+educ+wordsum+paeduc+maeduc+papres16
>estimate
Data for the following results were selected according to:
case<>113 and case<>1381
1 case(s) deleted due to missing data.
Dep
Var: PRESTIGE N: 1534 Multiple R: 0.552
Squared multiple R: 0.305
Adjusted
squared multiple R: 0.303 Standard error of estimate: 11.834
Effect
Coefficient Std Error Std Coef
Tolerance t P(2 Tail)
CONSTANT
7.213 1.541
0.000 .
4.680 0.000
EDUC
2.211 0.136
0.471 0.541 16.235
0.000
WORDSUM
1.042 0.199
0.140 0.638 5.243
0.000
PAEDUC
0.145 0.123
0.042 0.363 1.179
0.239
MAEDUC
-0.466 0.135
-0.115 0.408 -3.445
0.001
PAPRES16
0.071 0.028
0.061 0.764 2.505
0.012
Analysis of Variance
Source
Sum-of-Squares df Mean-Square
F-ratio P
Regression
93938.992 5 18787.798
134.150 0.000
Residual
213997.539 1528 140.051
----------------------------------------------------------------------------------------------------------------------------------
Durbin-Watson
D Statistic 2.044
First
Order Autocorrelation -0.024