SAMPLE 1 20 * Data set is TABLE 13.1 from * Griffiths, Hill and Judge, Learning and Practicing Econometrics READ C W P A 58.2 39.21 17.73 4.39 62.2 42.31 20.29 4.60 58.6 40.37 18.83 3.25 56.6 39.15 17.44 2.61 51.6 34.00 14.76 1.67 51.1 33.59 13.39 2.44 54.0 36.88 13.93 2.39 57.2 39.27 14.67 5.00 62.8 45.51 17.20 3.93 65.0 46.06 17.15 5.48 63.9 44.16 15.92 4.37 67.5 47.68 17.59 4.51 71.3 50.79 18.49 4.90 76.6 57.78 19.18 6.37 86.3 78.97 19.12 8.42 95.7 73.54 19.76 9.27 98.3 71.92 17.55 8.87 100.3 74.01 19.17 9.30 103.2 75.51 20.20 6.95 108.9 80.97 22.12 7.15 * Compute the Variance Inflation Factors as an indicator of the severity * of multicollinearity. * * REFERENCE: * Judge, Hill, Griffiths, Lutkepohl and Lee, Introduction to the Theory * and Practice and Econometrics, second edition, 1988, page 869. * * Get the correlation matrix of the X variables STAT W P A / PCOR COR=S MATRIX VIF=DIAG(INV(S)) PRINT VIF * Now do the computations a different way * * REFERENCE: * Neter, Wasserman and Kutner, Applied Linear Statistical Models, * 3rd edition, 1990, page 409. * SET NOOUTPUT OLS W P A GEN1 VIF:1=1/(1-$R2) OLS P W A GEN1 VIF:2=1/(1-$R2) OLS A W P GEN1 VIF:3=1/(1-$R2) SET OUTPUT PRINT VIF * * Run an OLS regression OLS C W P A * Impose some restrictions and run a restricted regression GENR X = W + 0.75*P + 0.625*A OLS C X * Get a t-test statistic for the coefficient on P TEST 0.75*X=0 * Get a t-test statistic for the coefficient on A TEST 0.625*X=0 STOP