Question

You are considering developing a regression equation relating a dependent variable to two independent variables. One of the variables can be measured on a ratio scale, but the other is a categorical variable with two possible levels.

a. How many dummy variables are needed to represent the categorical variable?
b. Write the multiple regression equation relating the dependent variable to the independent variables.
c. Interpret the meaning of the coefficients in the regression equation.

Answer 1

Explanation:

a. How many dummy variables are needed to represent the categorical variable?

We need 1 less dummy variable than outcomes for the categorical variable.

categorical variable has 2 outcomes

So, we need 2-1 = 1 dummy variable.

Answer: 1

Note: If our categorical variable had 5 outcomes, we would need 5-1 = 4 dummy variables.

b. Write the multiple regression equation relating the dependent variable to the independent variables.

Note: Depending your your professor, this can be written one of the following ways...

yhat = a + b1 x1 + bx2

yhat = b0 + b1 x1 + b2 x2

yhat = beta0hat + beta1hat x1 + betahat2 x2

where x1 is the ratio variable

and x2 is the dummy variable

c. Interpret the meaning of the coefficients in the regression equation.

b0 = the predicted value of y when x1 and x2 both equal 0.

b1 = the change in y when x1 increases by 1 unit holding x2 constant.

b2 = the predicted difference in y between the two possible outcomes of the categorical variable holding x1 constant.