Exercise

# Average marginal effects

The logistic response function is essentially nonlinear. Therefore, it is not immediately clear what is the effect of a unit change in the price ratio on the probability that a customer purchases **Hoppiness**. A solution is to interpret the effect of a unit change averaged over all customers.

This average marginal effect can be derived by using the function `margins()`

. The function is loaded from the add-on package `margins`

. Finally, you will compare the average marginal effect for `price.ratio`

of the `logistic.model`

to the `price.ratio`

coefficient of the `probability.model`

.

Instructions

**100 XP**

- Load the add-on package
`margins`

by using the function`library()`

. - Obtain the
`price.ratio`

coefficient for the`probability.model`

by using the function`coef()`

. - Obtain the marginal effect for
`price.ratio`

of the`logistic.model`

by using the function`margins()`

.