14.6 According to the paper “Assessing the Validity of the Post-Materialism Index” (American Political Science Review [1999]: 649–664), it is possible to predict an individual’s level of support for ecology based on demographic and ideological characteristics. The multiple regression model proposed by the authors was y 5 3.60 2 .01x1 1 .01x2 2 .07x3 1 .12x4 1 .02x5 2 .04x6 2 .01x7 2 .04x8 2 .02x9 1 e where the variables are defined as follows:

y 5 ecology score shigher values indicate a greater concern for ecologyd x1 5 age times 10 x2 5 income sin thousands of dollarsd x3 5 gender s1 5 male, 0 5 femaled x4 5 race s1 5 white, 0 5 nonwhited x5 5 education (in years)

x6 5 ideology s4 5 conservative, 3 5 right of center, 2 5 middle of the road, 1 5 left of center, and 0 5 liberald x7 5 social class s4 5 upper, 3 5 upper middle, 2 5 middle, 1 5 lower middle, and 0 5 lowerd x8 5 postmaterialist s1 if postmaterialist, 0 otherwised x9 5 materialist s1 if materialist, 0 otherwised

a. Suppose you knew a person with the following characteristics: a 25-year-old, white female with a college degree (16 years of education), who has a $32,000-per-year job, is from the upper middle class, and considers herself left of center, but who is neither a materialist nor a postmaterialist. Predict her ecology score.

b. If the woman described in Part

(a) were Hispanic rather than white, how would the prediction change? (Hint: See Example 14.5.)

c. Given that the other variables are the same, what is the estimated mean difference in ecology score for men and women?

d. How would you interpret the coefficient of x2?

e. Comment on the numerical coding of the ideology and social class variables. Can you suggest a better way of incorporating these two variables into the model? (Hint: See Example 14.5.)