What is the resulting regression equation? In particular, how does the dummy variable parameter in the recoded model relate to 7? Describe how to interpret the parameters of the recoded model. Is there a reason to prefer one coding over the other?

Altering Dummy Variable Values

Consider the two variable regression problem with a dummy variable in the design matrix:

• g = /30 + thi+ 71+ e,

where 1 is a dummy-variable that is coded 1 if the observation satisfies a certain condition, and is 0 ti otherwise. Consider replacing the 0 and 1 values in d with -0.5 and 0.5, respectively. What is the resulting regression equation? In particular, how does the dummy variable parameter in the recoded model relate to 7? Describe how to interpret the parameters of the recoded model. Is there a reason to prefer one coding over the other?