The ORTHOREG Procedure

The labor statistics data set of Longley (1967) is noted for being ill-conditioned. Both the ORTHOREG and GLM procedures are applied for comparison (only portions of the PROC GLM results are shown).

Note: The results from this example vary from machine to machine, depending on floating-point configuration.

The following statements read the data into the SAS data set Longley:

title 'PROC ORTHOREG used with Longley data';
data Longley;
   input Employment Prices GNP Jobless Military PopSize Year;
   datalines;
60323  83.0 234289 2356 1590 107608 1947
61122  88.5 259426 2325 1456 108632 1948
60171  88.2 258054 3682 1616 109773 1949
61187  89.5 284599 3351 1650 110929 1950
63221  96.2 328975 2099 3099 112075 1951
63639  98.1 346999 1932 3594 113270 1952
64989  99.0 365385 1870 3547 115094 1953
63761 100.0 363112 3578 3350 116219 1954
66019 101.2 397469 2904 3048 117388 1955
67857 104.6 419180 2822 2857 118734 1956
68169 108.4 442769 2936 2798 120445 1957
66513 110.8 444546 4681 2637 121950 1958
68655 112.6 482704 3813 2552 123366 1959
69564 114.2 502601 3931 2514 125368 1960
69331 115.7 518173 4806 2572 127852 1961
70551 116.9 554894 4007 2827 130081 1962
;

The data set contains one dependent variable, Employment (total derived employment), and six independent variables: Prices (GNP implicit price deflator normalized to the value 100 in 1954), GNP (gross national product), Jobless (unemployment), Military (size of armed forces), PopSize (noninstitutional population aged 14 and over), and Year (year).

The following statements use the ORTHOREG procedure to model the Longley data by using a quadratic model in each independent variable, without interaction:

proc orthoreg data=Longley;
   model Employment = Prices   Prices*Prices
                      GNP      GNP*GNP
                      Jobless  Jobless*Jobless
                      Military Military*Military
                      PopSize  PopSize*PopSize
                      Year     Year*Year;
run;

Figure 72.1 shows the resulting analysis.

The estimates in Figure 72.1 compare very well with the best estimates available; for additional information, see Longley (1967) and Beaton, Rubin, and Barone (1976).

The following statements request the same analysis from the GLM procedure:

proc glm data=Longley;
   model Employment = Prices   Prices*Prices
                      GNP      GNP*GNP
                      Jobless  Jobless*Jobless
                      Military Military*Military
                      PopSize  PopSize*PopSize
                      Year     Year*Year;
   ods select OverallANOVA
              FitStatistics
              ParameterEstimates
              Notes;
run;

Figure 72.2 contains the overall ANOVA table and the parameter estimates produced by PROC GLM. Notice that the PROC ORTHOREG fit achieves a somewhat smaller root mean square error (RMSE) and also that the GLM procedure detects spurious singularities.