page 269 The regressions at the bottom of the page.
GET FILE='D:davis.sav'.
For the uncorrected data:
compute measf = measwt*female. execute. regression /dep=reptwt /method=enter measwt female measf.
Model  Variables Entered  Variables Removed  Method 

1  MEASF, Measured Weight, Gender, 0=male, 1=female(a)  .  Enter 
a All requested variables entered.  
b Dependent Variable: Reported Weight 
Model  R  R Square  Adjusted R Square  Std. Error of the Estimate 

1  .942(a)  .887  .886  4.661 
a Predictors: (Constant), MEASF, Measured Weight, Gender, 0=male, 1=female 
Model  Sum of Squares  df  Mean Square  F  Sig.  

1  Regression  30654.729  3  10218.243  470.408  .000(a) 
Residual  3888.254  179  21.722  


Total  34542.984  182  



a Predictors: (Constant), MEASF, Measured Weight, Gender, 0=male, 1=female  
b Dependent Variable: Reported Weight 

Unstandardized Coefficients  Standardized Coefficients  t  Sig.  

Model  B  Std. Error  Beta  
1  (Constant)  1.359  3.277  
.415  .679 
Measured Weight  .990  .043  1.099  23.236  .000  
Gender, 0=male, 1=female  39.964  3.929  1.447  10.171  .000  
MEASF  .725  .056  1.611  12.957  .000  
a Dependent Variable: Reported Weight 
For the corrected data:
compute nmwt = measwt. if subject=12 nmwt=57. if subject=12 measht=166. compute measwtf = nmwt*female. execute. regression /dep=reptwt /method=enter nmwt female measwtf.
Model  Variables Entered  Variables Removed  Method 

1  MEASWTF, NMWT, Gender, 0=male, 1=female(a)  .  Enter 
a All requested variables entered.  
b Dependent Variable: Reported Weight 
Model  R  R Square  Adjusted R Square  Std. Error of the Estimate 

1  .987(a)  .974  .973  2.243 
a Predictors: (Constant), MEASWTF, NMWT, Gender, 0=male, 1=female 
Model  Sum of Squares  df  Mean Square  F  Sig.  

1  Regression  33642.345  3  11214.115  2228.780  .000(a) 
Residual  900.639  179  5.032  


Total  34542.984  182  



a Predictors: (Constant), MEASWTF, NMWT, Gender, 0=male, 1=female  
b Dependent Variable: Reported Weight 

Unstandardized Coefficients  Standardized Coefficients  t  Sig.  

Model  B  Std. Error  Beta  
1  (Constant)  1.359  1.577  
.861  .390 
NMWT  .990  .021  .962  48.279  .000  
Gender, 0=male, 1=female  1.983  2.450  .072  .809  .420  
MEASWTF  5.668E02  .038  .119  1.474  .142  
a Dependent Variable: Reported Weight 
page 270 Figure 11.2 Davis's data on reported and measured weight for women (F) and men (M), showing the leastsquares linear regression line for each group (the broken line for men, the solid line for women). The outlying observation has a substantial effect on the fitted line for women.
USE ALL. COMPUTE filter_$=(female=1). VARIABLE LABEL filter_$ 'female=1 (FILTER)'. VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected'. FORMAT filter_$ (f1.0). FILTER BY filter_$. EXECUTE. formats reptwt measwt (f4.0). GGRAPH /GRAPHDATASET NAME="GraphDataset" VARIABLES= reptwt measwt female /GRAPHSPEC SOURCE=INLINE. BEGIN GPL SOURCE: s=userSource( id( "GraphDataset" ) ) DATA: reptwt=col( source(s), name( "reptwt" ) ) DATA: measwt=col( source(s), name( "measwt" ) ) DATA: female = col(source(s), name("female"), unit.category()) GUIDE: axis( dim( 1 ), label( "Measured Weight (kg.)" ), start(0.0), delta(25) ) GUIDE: axis( dim( 2 ), label( "Reported Weight (kg.)" ), start(0.0), delta(40) ) SCALE: linear( dim( 1 ), min(25), max(175) ) SCALE: linear( dim( 2 ), min(40), max(160) ) ELEMENT: point( position(measwt * reptwt)) ELEMENT: line(position(smooth.linear(measwt * reptwt)), shape(female)) END GPL.
page 271 The largest hat value (middle of the page).
NOTE: Various statistics, including DFBETAs, studentized residuals and covariance ratios, are discussed in this section. Those statistics have been calculated and displayed by the regression command below.
GET FILE='D:davis.sav'. compute rptfem = reptwt*female. execute. regression /dep=measwt /method=enter reptwt female rptfem /save lev sres dfbeta covratio.
Model  Variables Entered  Variables Removed  Method 

1  RPTFEM, Reported Weight, Gender, 0=male, 1=female(a)  .  Enter 
a All requested variables entered.  
b Dependent Variable: Measured Weight 
Model  R  R Square  Adjusted R Square  Std. Error of the Estimate 

1  .837(a)  .700  .695  8.449 
a Predictors: (Constant), RPTFEM, Reported Weight, Gender, 0=male, 1=female  
b Dependent Variable: Measured Weight 
Model  Sum of Squares  df  Mean Square  F  Sig.  

1  Regression  29786.378  3  9928.793  139.071  .000(a) 
Residual  12779.436  179  71.393  


Total  42565.814  182  



a Predictors: (Constant), RPTFEM, Reported Weight, Gender, 0=male, 1=female  
b Dependent Variable: Measured Weight 

Unstandardized Coefficients  Standardized Coefficients  t  Sig.  

Model  B  Std. Error  Beta  
1  (Constant)  1.794  5.924  
.303  .762 
Reported Weight  .969  .076  .873  12.681  .000  
Gender, 0=male, 1=female  2.074  9.297  .068  .223  .824  
RPTFEM  9.525E03  .147  .018  .065  .948  
a Dependent Variable: Measured Weight 
Case Number  Std. Residual  Measured Weight 

12  12.830  166 
a Dependent Variable: Measured Weight 

Minimum  Maximum  Mean  Std. Deviation  N 

Predicted Value  43.20  121.94  66.22  12.793  183 
Std. Predicted Value  1.799  4.355  .000  1.000  183 
Standard Error of Predicted Value  .841  3.743  1.183  .402  183 
Adjusted Predicted Value  43.49  122.66  66.24  12.811  183 
Residual  7.66  108.41  .00  8.380  183 
Std. Residual  .906  12.830  .000  .992  183 
Stud. Residual  .935  12.895  .001  .997  183 
Deleted Residual  8.15  109.50  .01  8.475  183 
Stud. Deleted Residual  .935  48.221  .192  3.580  183 
Mahal. Distance  .810  34.720  2.984  3.619  183 
Cook's Distance  .000  .421  .003  .031  183 
Centered Leverage Value  .004  .191  .016  .020  183 
a Dependent Variable: Measured Weight 
NOTE: SPSS values for the leverage do not exactly match those obtained by Fox.
compute measf = measwt*female. execute. regression /dep=reptwt /method=enter measwt female measf /save lev.
Model  Variables Entered  Variables Removed  Method 

1  MEASF, Measured Weight, Gender, 0=male, 1=female(a)  .  Enter 
a All requested variables entered.  
b Dependent Variable: Reported Weight 
Model  R  R Square  Adjusted R Square  Std. Error of the Estimate 

1  .942(a)  .887  .886  4.661 
a Predictors: (Constant), MEASF, Measured Weight, Gender, 0=male, 1=female  
b Dependent Variable: Reported Weight 
Model  Sum of Squares  df  Mean Square  F  Sig.  

1  Regression  30654.729  3  10218.243  470.408  .000(a) 
Residual  3888.254  179  21.722  


Total  34542.984  182  



a Predictors: (Constant), MEASF, Measured Weight, Gender, 0=male, 1=female  
b Dependent Variable: Reported Weight 

Unstandardized Coefficients  Standardized Coefficients  t  Sig.  

Model  B  Std. Error  Beta  
1  (Constant)  1.359  3.277  
.415  .679 
Measured Weight  .990  .043  1.099  23.236  .000  
Gender, 0=male, 1=female  39.964  3.929  1.447  10.171  .000  
MEASF  .725  .056  1.611  12.957  .000  
a Dependent Variable: Reported Weight 
Case Number  Std. Residual  Reported Weight 

12  6.270  56 
115  3.342  77 
a Dependent Variable: Reported Weight 

Minimum  Maximum  Mean  Std. Deviation  N 

Predicted Value  51.64  119.15  65.62  12.978  183 
Std. Predicted Value  1.078  4.124  .000  1.000  183 
Standard Error of Predicted Value  .464  3.939  .618  .306  183 
Adjusted Predicted Value  51.99  158.24  66.01  14.583  183 
Residual  29.22  15.58  .00  4.622  183 
Std. Residual  6.270  3.342  .000  .992  183 
Stud. Residual  11.728  3.392  .029  1.241  183 
Deleted Residual  102.24  16.04  .39  8.643  183 
Stud. Deleted Residual  24.304  3.497  .096  2.008  183 
Mahal. Distance  .808  128.987  2.984  9.774  183 
Cook's Distance  .000  85.927  .474  6.352  183 
Centered Leverage Value  .004  .709  .016  .054  183 
a Dependent Variable: Reported Weight 
page 284 Figure 11.5 Partialregression plots for Duncan's regression of occupational prestige on the income (a) and educational levels (b) of 45 U.S. occupations. in 1950. Three potentially influential observations (ministers, railroad conductors, and railroad engineers) are identified on the plots. The partialregression plot for the intercept A is not shown.
NOTE: In order to get the regression line on the plot, you need to use the code below, doubleclick on the resulting graph, select chart from the menu at the top, select options, click on fit line total, select fit options, select regression, click on continue and OK. We do not know how to add the regression line using code.
GET FILE='D:duncan.sav'. regression /dep=prestige /method=enter income educ /partialplot all /sav sresid lev cook.
Model  Variables Entered  Variables Removed  Method 

1  Percent of males in occupation in 1950 who were highschool graduates, Percent of males in occupation earning $3500 or more in 1950(a)  .  Enter 
a All requested variables entered.  
b Dependent Variable: Percent of raters in NORC study rating occupation as excellent or good in presti 
Model  R  R Square  Adjusted R Square  Std. Error of the Estimate 

1  .910(a)  .828  .820  13.369 
a Predictors: (Constant), Percent of males in occupation in 1950 who were highschool graduates, Percent of males in occupation earning $3500 or more in 1950  
b Dependent Variable: Percent of raters in NORC study rating occupation as excellent or good in presti 
Model  Sum of Squares  df  Mean Square  F  Sig.  

1  Regression  36180.946  2  18090.473  101.216  .000(a) 
Residual  7506.699  42  178.731  


Total  43687.644  44  



a Predictors: (Constant), Percent of males in occupation in 1950 who were highschool graduates, Percent of males in occupation earning $3500 or more in 1950  
b Dependent Variable: Percent of raters in NORC study rating occupation as excellent or good in presti 

Unstandardized Coefficients  Standardized Coefficients  t  Sig.  

Model  B  Std. Error  Beta  
1  (Constant)  6.065  4.272  
1.420  .163 
Percent of males in occupation earning $3500 or more in 1950  .599  .120  .464  5.003  .000  
Percent of males in occupation in 1950 who were highschool graduates  .546  .098  .516  5.555  .000  
a Dependent Variable: Percent of raters in NORC study rating occupation as excellent or good in presti 

Minimum  Maximum  Mean  Std. Deviation  N 

Predicted Value  1.95  96.42  47.69  28.676  45 
Std. Predicted Value  1.595  1.699  .000  1.000  45 
Standard Error of Predicted Value  2.077  6.935  3.334  .903  45 
Adjusted Predicted Value  .81  97.04  47.59  28.701  45 
Residual  29.54  34.64  .00  13.062  45 
Std. Residual  2.209  2.591  .000  .977  45 
Stud. Residual  2.272  2.849  .003  1.019  45 
Deleted Residual  31.24  41.89  .10  14.249  45 
Stud. Deleted Residual  2.397  3.135  .007  1.056  45 
Mahal. Distance  .084  10.862  1.956  1.928  45 
Cook's Distance  .000  .566  .032  .090  45 
Centered Leverage Value  .002  .247  .044  .044  45 
a Dependent Variable: Percent of raters in NORC study rating occupation as excellent or good in presti 
page 285 Figure 11.6 "Bubble plot" of Cook's D, studentized residuals, and hat values, for Duncan's regression of occupational prestige on income and education. Each point is plotted as a circle with area proportional to D. Horizontal reference lines are drawn at studentized residuals of 0 and +2; vertical reference lines are drawn at values of 2h and 3h. Several observations are identified on the plot: Ministers and conductors have large hat values and relatively large residuals; reporters have a relatively large residual, but a small hat value; railroad engineers have a large hat value, but a small residual.
formats sre_1 (f4.1) lev_1 (f5.2) coo_1 (f7.5). GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=LEV_1 sre_1 coo_1 /GRAPHSPEC SOURCE=INLINE. BEGIN GPL SOURCE: s=userSource(id("graphdataset")) DATA: LEV_1=col(source(s), name("LEV_1")) DATA: sre_1=col(source(s), name("sre_1")) DATA: COO_1=col(source(s), name("coo_1")) GUIDE: axis(dim(1), delta(.05), label("HatValue")) GUIDE: axis(dim(2), delta(2.5), label("Studentized Residual")) GUIDE: legend(aesthetic(aesthetic.size), label("Cook's Distance")) GUIDE: form.line(position(.2), color(color.black)) GUIDE: form.line(position(*, 0), color(color.black)) GUIDE: form.line(position(.13), color(color.black)) GUIDE: form.line(position(*, 2.1), color(color.black)) GUIDE: form.line(position(*, 2.1), color(color.black)) SCALE: linear(dim(1), min(0), max(.3)) SCALE: linear(dim(2), min(2.5), max(5)) ELEMENT: point(position(LEV_1*sre_1), size(COO_1)) END GPL.
page 288 Table 11.1 Data on the 1907 Romanian Peasant Rebellion: I, intensity of the rebellion (corrected from the original); C, commercialization of agriculture; T, traditionalism; M, market forces; and G, inequality of land tenure.
GET FILE='D:chirot.sav'.list county rebel agric trad market inequal.
COUNTY REBEL AGRIC TRAD MARKET INEQUAL1 1.39 13.80 86.20 6.20 .60 2 .65 20.40 86.70 2.90 .72 3 1.89 27.60 79.30 16.90 .66 4 .15 18.60 90.10 3.40 .74 5 .86 17.20 84.50 9.00 .70 6 .11 21.50 81.50 5.20 .60 7 .51 11.60 82.60 5.10 .52 8 .86 20.40 82.40 6.30 .64 9 .24 19.50 87.50 4.80 .68 10 .77 8.90 85.60 9.50 .58 11 .24 25.80 82.20 10.90 .68 12 1.57 24.10 83.50 8.40 .74 13 .51 22.00 88.30 6.20 .70 14 1.57 24.20 84.90 6.10 .62 15 .51 30.60 76.10 1.30 .76 16 1.13 33.90 85.50 5.80 .70 17 1.22 28.60 84.20 2.90 .58 18 1.22 36.50 78.10 4.30 .72 19 .86 40.90 84.40 2.30 .64 20 1.39 6.80 76.30 3.60 .58 21 2.81 41.90 89.70 6.60 .66 22 1.04 25.40 83.20 2.50 .68 23 1.57 30.50 80.20 4.10 .76 24 4.32 48.20 91.00 4.20 .70 25 3.79 46.00 90.50 3.70 .68 26 3.79 45.10 85.50 5.10 .64 27 1.75 12.50 83.80 7.20 .50 28 .82 39.30 85.60 4.90 .60 29 2.59 47.70 87.60 5.20 .58 30 .86 15.20 87.30 10.80 .42 31 1.84 11.70 82.30 81.70 .42 32 1.84 25.60 80.10 68.40 .26
Number of cases read: 32 Number of cases listed: 32