Numerical Example of the Metric Effect for a Polynomial


To help demonstrate the metric effect of a polynomial the previous example is utilized. The following represents the relationship of between age and attitudes toward military intervention. X1 represents age and X2 represents age squared. The general model estimated is:
Y= 58.3 + -4.1( X1) + .7 (X2 )
The metric effect of age (X1 ) on salary (Y) at different values of age:

Age Group 1: -4.1 + 2(.7)(1)= -2.7
Age Group 2: -4.1 + 2(.7)(2)= -1.3
Age Group 3: -4.1 + 2(.7)(3)= 0.1
Age Group 4: -4.1 + 2(.7)(4)= 1.5
Age Group 5: -4.1 + 2(.7)(5)= 2.9

The results provided above represent the estimate of the conditional effect of age on attitudes toward military intervention at specified values of age.

The predicted values of the response variable can also be obtained as follows:
Age Group 1: 58.3 + -4.1(1) + .7(1)(1) = 54.9
Age Group 2: 58.3 + -4.1(2) + .7(2)(2) = 52.9
Age Group 3: 58.3 + -4.1(3) + .7(3)(3) = 52.3
Age Group 4: 58.3 + -4.1(4) + .7(4)(4) = 53.1
Age Group 5: 58.3 + -4.1(5) + .7(5)(5) = 55.3

The predicted value of the response variable indicate that the evaluation of military intervention becomes decreasingly negative to the inflection point at which evaluations become increasingly positive.

Back to Interpretation

Revised: August 17, 1999.