Suppose that you want to run a regression model and to test the statistical significance of a group of variables. For example, let’s say that you want to predict students’ writing score from their reading, math and science scores. The data set with these variables in it can be downloaded by following this link: hsb2.sav .
The SPSS syntax for this would be:
regression /dependent = write /method = enter read math science.
Variables Entered/Removed(b) Model Variables Entered Variables Removed Method 1 science score, reading score, math score(a) . Enter a All requested variables entered. b Dependent Variable: writing score
Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .684(a) .467 .459 6.97111 a Predictors: (Constant), science score, reading score, math score
ANOVA(b) Model Sum of Squares df Mean Square F Sig. 1 Regression 8353.990 3 2784.663 57.302 .000(a) Residual 9524.885 196 48.596 Total 17878.875 199 a Predictors: (Constant), science score, reading score, math score b Dependent Variable: writing score
Coefficients(a) Unstandardized Coefficients Standardized Coefficients t Sig. Model B Std. Error Beta
1 (Constant) 13.192 3.069 4.299 .000 reading score .236 .069 .255 3.410 .001 math score .319 .076 .316 4.222 .000 science score .202 .069 .211 2.918 .004 a Dependent Variable: writing score
Now let’s suppose that you wanted to test the combined effect of math and science on writing. The SPSS syntax for doing that is below. Note that the variables listed in the method = test() subcommand are not listed on the method = enter subcommand. In other words, the independent variables are listed only once. Also note that, unlike other SPSS subcommands, you can have multiple method = subcommands within the regression command.
regression /dependent = write /method = enter read /method = test(math science).
Variables Entered/Removed(b) Model Variables Entered Variables Removed Method 1 reading score(a) . Enter 2 science score, math score . Test a All requested variables entered. b Dependent Variable: writing score
Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .597(a) .356 .353 7.62487 2 .684(b) .467 .459 6.97111 a Predictors: (Constant), reading score b Predictors: (Constant), reading score, science score, math score
ANOVA(d) Model Sum of Squares df Mean Square F Sig. R Square Change 1 Regression 6367.421 1 6367.421 109.521 .000(a) Residual 11511.454 198 58.139 Total 17878.875 199 2 Subset Tests math score, science score 1986.569 2 993.284 20.439 .000(b) .111 Regression 8353.990 3 2784.663 57.302 .000(c) Residual 9524.885 196 48.596 Total 17878.875 199 a Predictors: (Constant), reading score b Tested against the full model. c Predictors in the Full Model: (Constant), reading score, science score, math score. d Dependent Variable: writing score
Coefficients(a) Unstandardized Coefficients Standardized Coefficients t Sig. Model B Std. Error Beta
1 (Constant) 23.959 2.806 8.539 .000 reading score .552 .053 .597 10.465 .000 2 (Constant) 13.192 3.069 4.299 .000 reading score .236 .069 .255 3.410 .001 math score .319 .076 .316 4.222 .000 science score .202 .069 .211 2.918 .004 a Dependent Variable: writing score
Excluded Variables(b) Beta In t Sig. Partial Correlation Collinearity Statistics Model
Tolerance 1 math score .396(a) 5.583 .000 .370 .561 science score .322(a) 4.609 .000 .312 .603 a Predictors in the Model: (Constant), reading score b Dependent Variable: writing score
If you wanted to test all three variables together, the syntax would be:
regression /dependent = write /method = test(read math science).
Variables Entered/Removed(a) Model Variables Entered Variables Removed Method 1 science score, reading score, math score . Test a Dependent Variable: writing score
Model Summary Model R R Square Adjusted R Square Std. Error of the Estimate 1 .684(a) .467 .459 6.97111 a Predictors: (Constant), science score, reading score, math score
ANOVA(c) Model Sum of Squares df Mean Square F Sig. R Square Change 1 Subset Tests reading score, math score, science score 8353.990 3 2784.663 57.302 .000(a) .467 Regression 8353.990 3 2784.663 57.302 .000(b) Residual 9524.885 196 48.596 Total 17878.875 199 a Tested against the full model. b Predictors in the Full Model: (Constant), science score, reading score, math score. c Dependent Variable: writing score
Coefficients(a) Unstandardized Coefficients Standardized Coefficients t Sig. Model B Std. Error Beta
1 (Constant) 13.192 3.069 4.299 .000 reading score .236 .069 .255 3.410 .001 math score .319 .076 .316 4.222 .000 science score .202 .069 .211 2.918 .004 a Dependent Variable: writing score
You will notice that the output from the first example with the three
independent variables on the method = enter subcommand and the output
from this example with the three independent variables on the method =
test() subcommand are virtually identical. The only difference between
them is the line in the ANOVA table that gives the test of the subset, which in
this case is all of the variables. The point of this example is that you
can put all of the independent variables in the regression on the method =
test() subcommand and not use a method = enter subcommand if you like.