* PS6.9 using DATA4-4 for estimating demand elasticities -- Section 6.11 SAMPLE 1 40 READ(data4-4) BUSTRAVL FARE GASPRICE INCOME POP DENSITY LANDAREA * Generate variables GENR LBUSTRAV=LOG(BUSTRAVL) GENR LFARE=LOG(FARE) GENR LGASP=LOG(GASPRICE) GENR LINC=LOG(INCOME) GENR LPOP=LOG(POP) GENR LDEN=LOG(DENSITY) GENR LLANDA=LOG(LANDAREA) * * The first model to be estimated includes all the explanatory (independent) * variables. Examine the coefficients to determine which are significant to * the model. The LOGLOG option is specified since the dependent and * independent variables are in logged form. If the LOGLOG option is not * specified then the ELASTICITY AT MEANS in the OLS regression output * will be incorrect. * OLS LBUSTRAV LFARE LGASP LINC LPOP LDEN LLANDA / LOGLOG * * In the above regression, the logarithm of the variables FARE, GASPRICE, * POP, DENSITY, and LANDAREA had insignificant coefficients. The variable * DENSITY had the least significant coefficient. Now re-estimate the above * model with the variable logged DENSITY omitted. * OLS LBUSTRAV LFARE LGASP LINC LPOP LLANDA / LOGLOG * omit LGASP OLS LBUSTRAV LFARE LINC LPOP LLANDA / LOGLOG * omit LFARE OLS LBUSTRAV LINC LPOP LLANDA DELETE / ALL STOP