Stat Computing > SPSS > Textbook Examples > Regression Analysis by Example, by Chatterjee, Hadi and Price

SPSS Textbook Examples
Regression Analysis by Example, Third Edition
Chapter 10:  Biased Estimation of Regression Coefficients

Table 10.1, Page 265. Regression results of fitting model (10.2) to the Import Data (1949-1959).

get file 'd:\p233.sav'.
select if (year <= 59).
exe.

The analysis requires the dependent variable be converted into its standardized value. This is done using the descriptive command using the save option.

descriptives variable = import
 /save.

[output omitted]

regression
/dependent = zimport
/method enter = doprod stock consum.

Table 10.2, page 265. Regression results of fitting model (10.5) to the Import Data (1949-1959).
Analysis requires calculation of principal component scores. SPSS does not calculate PC scores, rather it uses factor scores. There is a direct linear relationship between the two scores and the standardized regression coefficients are equivalent.

factor variables = doprod stock consum 
 /criteria factors(3) iterate(25) 
 /extraction pc
 /rotation norate
 /save reg( 3, fs).

regression
 /dependent = zimport
 /method enter = fs1 fs2 fs3.

Table 10.4, page 270. A response variable U and a set of principal components of four predictor variables.

get file 'd:\p270.sav'.
list.

        u        c1        c2        c3        c4

      .95      1.47      1.90      -.53       .04
     -.75      2.14       .24      -.29      -.03
    -2.32     -1.13       .18      -.01      -.09
     -.82       .66      1.58       .18      -.03
      .47      -.36       .48      -.74       .02
     -.30      -.97       .17       .09      -.01
      .21      -.93     -2.13      -.17       .01
      .56      2.23      -.69       .46       .02
    -1.12       .35     -1.43      -.03      -.05
      .50     -1.66      1.83       .85       .02
      .78      1.64     -1.29       .49       .03
      .92     -1.69      -.39      -.02       .04
      .92     -1.75      -.44      -.28       .04

Number of cases read:  13    Number of cases listed:  13

Table 10.5, page 270. Regression results using all four PCs of Hald's data.
NOTE:  The value for the regression coefficients and t-test for c4 are slightly off.

regression 
 /dependent = u
 /method enter = c1 c2 c3 c4. 

Table 10.6, page 270. Regression results using the first three PCs of Hald's data.

regression 
 /dependent = u
 /method enter = c1 c2 c3. 

Figure 10.1, page 271. Scatter plots of U versus each of the PCs of the Hald's data.

graph
 /scatterplot = c1 with u.

graph
 /scatterplot = c2 with u.

graph
 /scatterplot = c3 with u.

graph
 /scatterplot = c4 with u.

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services