Is there evidence of non linearity for age or for height? 

There was some previous literature that the effect of age and height on FEV might not be linear. Hence, we ran the regression model in Table 11.35.

Table 11.35 Quadratic regression of FEV on age, height, gender, and smoking

 

Transcribed Image Text:

Regression Analysis: FEV vs. Age, Hgt,. The regression equation is FEV = - 4.84 + 0.0633 Age + 0.110 Hgt + 0.0952 Sex + 0.00177 (Age-10)2 + 0.00284 (Hgt-61)2 - 0.140 Smoke Predictor SE Coef coef -20.22 -4.8360 0.2392 0.000 Constant 0.000 0.06334 0.01095 5.78 Age 0.000 0.109594 0.005239 20.92 Hgt 0.09515 0.03287 2.90 0.004 Sex (Age-10) 42 0.001767 0.001754 1.01 0.314 (Hgt-61) ^2 0.0028438 0.0004949 5.75 0.000 Smoke -0.13959 0.05746 -2.43 0.015 s = 0.395080 R-Sq = 79.4% R-Sq(adj) = 79.2% Analysis of Variance Source DF MS Regression 389.931 64.988 416.36 0.000 Residual 647 100.989 0.156 Error Total 653 490.920