Data set for Multinomial Logit Model

[Snehal Patel]

How to set up data in excel for Multinomial Logit Model?
I have two data sets in separate excel sheets:
First one has Firm level data (Company_ID, Company_Name, Date_of_Incorporation, Employees, State, Entry_Mode etc,.)
Second has Location specific data (Labour_cost, Tax, Population_density, Corruption, Air_pollution etc,.) which is organised state wise and year wise (2001-2013). So for each state all the location specific data is for 13 years.
How to combine these two excel sheets for Multinomial Logit?

Any help and guidance will be much appreciated. Thanks

[Joe Canner]

Snehal,

The first thing you need to figure out is how to merge these two data sets together. You don't indicate this, but I assume that there is a variable (State?) that links the two data sets; i.e., a location variable in the Firm level data set that tells where to find the correct entry in the Location specific data set. Then you would do something like:

import excel using Firm, first clear
save Firm, replace
import excel using Location, first clear
save Location, replace

use Firm
merge 1:1 State using Location

The is probably not quite what you want, however, because your Location data set has data for each year. Does your Firm data set also have data for each year? If so, you can do merge 1:1 State Year using Location. If not, you need to decide how to combine the data from the different years before merging.

Regards,
Joe

[Joe Canner]

P.S. The fact that you are doing Multinomial Logit is most irrelevant as far as preparing your data set is concerned. The issues I raised are applicable to any kind of regression analysis. If your data are in panel format (i.e., multiple years for each firm), that may be relevant.

[Snehal Patel]

Hi Joe,
Many thanks for your response I really appreciate it.
I am trying to see if Environmental Regulations have an impact on Foreign Direct Investments in Indian states. So I have firm level data (foreign firms in India) and location data (state related data for each year). First of all I am not at all sure how to merge the data to be very honest. Secondly, I think the data has to be merged in a way that can compare the location variables of one state with the location variables of the rest of the states.
In firm level data I have year of incorporation which is important as when we merge it can pick location variables for that particular year for each state (or all states I am not able to decide)
So the two datasets are linked by 'State' or by 'Year' is a confusion for me.
Please see attached the two datasets just to make it even more clear.
Thanks for your time and help. Thats very kind of you.

Best,
Snehal

[Snehal Patel]

I must also mention that this is the first time I am working on data and this statistical model so please excuse me if I have asked very silly questions.