Probit model for panel data

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

brandon.ellse
Posts: 3
Joined: Thu Dec 10, 2015 6:50 am

Probit model for panel data

Postby brandon.ellse » Fri Jul 21, 2017 5:05 am

I'm attempting to model the probability of a property broker closing a deal (the dependent variable, 1=deal closed, 0=no deal closed)) based on a few explanatory variables, namely, the number of adverts placed (adverts), the number of leads canvassed (leadscan), the number of advertising boards placed (adboards) and the number of new leads generated (leadsnew) and the number of property viewings conducted (views).

To my understanding, I'm dealing with an unbalanced panel data set taking the form (for example):

broker______week________adverts_____leadscan ...............closed_deals
1____________1___________2__________1____________________ 0
1____________2___________1__________0____________________ 0
1____________3___________5__________1____________________1
2____________1___________9__________2____________________1
2____________2___________5__________2____________________0
3____________3___________7__________1____________________1


The objective is to be able to estimate the marginal effect, in terms of the likelihood of closing a deal, when say, the broker places an additional advertising board.

I'm guessing that this would require some sort of mixed effects Probit model, but I'm really at a loss as to what the most appropriate modeling methodology and specification would be in eviews. Any necessary data transformations required as a result of the relevant model would also be very useful indeed. I'm using EViews v8.1.

Any assistance would be greatly appreciated.

Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 24 guests