The generic formulation of a multiple regression model relating an outcome (continuous, binary, number of events, or time-to-event) to a three-category predictor, health insurance type (none, public, private), is as follows:

LHS=intercept+

^

1

x

1

+

^

2

x

2

LHS=intercept+^1x1+^2x2, where

LHS = "left hand side", and intercept =

^

o

^oorln(

^

o

(t))

ln(^o(t))

x1= 1 if public insurance, 0 if not public insurance

x2= 1 if private insurance, 0 if not private insurance

- What does the slope,

^

1

^1, estimate?