Logistic regression: Development of a hypothesis question

#1
Logistic regression: Development of a hypothesis question



Hello guys,

I am working on my master thesis right now. I am examining human capital/labor factors that are relevant for making investments in global cities with a (conditional) logistic regression model

I developed the following hypothesis:
Hypothesis 1: The odds of a global city being chosen for Foreign Service Investments are greater, the ampler the labor supply of the global city.

Now I am experiencing the difficulty that I am unsure how to measure ample labor supply in the best way, as I do not have the same range of observations for the all cities (see below). Due to the fact that I will run regressions for Europe/OECD/Worldwide, I would like to develop an independent variable that I could use throughout all regressions. Normally I would use labor force/population, however as you see below labor force is only available for OECD countries. However, does it make sense in your opinion to use :


GDP per capita/city population as a measure of ample labor supply? As GDP per capita is the country's entire economic output per person, the ratio would give an indication what the economic power of a city is.

2.
This are relevant independent variables:
GDP (worldwide)
LN GDP per capita (worldwide)
Number of highereducation students (Only in Europe)
Labor force (Only of OECD countries)
Labor productivity(Only of OECD countries)
City population (worldwide)

Many thanks for any help in advance!

Max
 

maartenbuis

TS Contributor
#2
I would look further, in particular I suspect it would be possible to get the working age population for different cities. I would consider that a much more convincing indicator. You would still need to decide whether you want that in absolute terms or relative terms. You could also consider unemployment rate: lots of workers is no use to a firm if they are all being used already by a competitor. This all rests on your exact definition of "ample labor supply". Finding the right data and building a convincing argument that your opperationalization is meaningful is all part the "art of doing research".