# Calculating marginal effect of logit model by hand

For an assignment I have to calculate the marginal effect of 'age' by hand. But I am dealing with a logit model, which makes it difficult for me. I have 4 variables, which are age, education, income and the price of cigarettes. The dependent variable is a dummy variable that indicates whether someone is a smoker, yes or no. I have been running regressions in Stata, since this is the main part of the assignment. Hence, I already have quite some information, such as the marginal effects at the mean and the average marginal effects.

Now the issue starts at the question where I cannot use Stata. I need to calculate the marginal effect of age by hand for a person with age = 28, education = 15, income = 12,500 and price of cigarettes = 60. Now these numbers do not really matter, but for completeness I have added them nonetheless. Where do I start with this? I think I need some of the already obtained data to perform my calculations. To be honest I am just clueless at this point.

• I am not sure if this question is on topic. However, at the least you should add the add the self-study tag and read the linked tag wiki. Then update the question to reflect what you currently understand and where you are stuck. You might add the regression output to your question, and then check here to do this. – GeoMatt22 Feb 26 '17 at 14:58
• This question was already answered here: stats.stackexchange.com/questions/19764/… – Jeremias K Feb 26 '17 at 16:34

A logit model is $y = (1+e^{-X\beta})^{-1}$. You have estimated beta and you are given X. Plug and chug.