I'm writing a paper on investment firms and their relationship with a sustainable finance initiative. I'm using a panel dataset with 307 investors, 125 of them signed this sustainable initiative.
I would like to add in a section in which I test which variables might be driving them to sign this initiative.
I believe I should use logit regression for this, but having not used these extensively, I'm looking for some guidance.
Currently the data looks like this:
| investor | year | activity | country | region | strategy | signatory |
|---|---|---|---|---|---|---|
| 123 IM | 2002 | 4.45 | France | europe | VC | 1 |
| 123 IM | 2003 | 3.2 | France | europe | VC | 1 |
| 123 IM | 2004 | 7.8 | France | europe | VC | 1 |
| Aegon | 2005 | 5.4 | Netherlands | europe | BY | 0 |
| Aegon | 2006 | 4.2 | Netherlands | europe | BY | 0 |
| Aegon | 2007 | 1.3 | Netherlands | europe | BY | 0 |
As you can see the signatory variable is a binary, and I would be looking to test variables such as country or region against it.
Any tips would be appreciated!
Rory
You can use the glm function in R. Following is an example with country and activity variables as independent variables:
# Assuming that your dataframe name is df
my_logit <- glm(signatory ~ activity + country, family = 'binomial', data=df)
# Check the output summary
summary(my_logit)