Most appropriate statistical method for master thesis

#1
Dear all,

I'm currenctly doing my empirical part of my master thesis with title: Socio-Economic analysis of investment in Renewable energy.
As a part of my research I would like to test 2 of my hypotheses:

1. Higher share of Renwables in energy mix reflects in lower Energy import. (EU-28)
2. Higher production from Renewables would reflect in lower CO2 emission rate. (EU-28)
3. Higher production from Renewables would reflect in higher access to electricity in Sub Saharan Africa

I have following data available :
1. Share of renewables (% of total production) for EU-28, from 1990-2016
2. Net energy import (% of total energy consumed) for EU-28, from 1990-2016
3. Share of renewables (% of total consumption) for EU-28, from 1990-2016
4. CO2 emission (in Tonnes) for EU-28, from 1990-2016
5. GDP per capita (in $) for EU-28 and Sub Saharan Africa, from 1990-2016
6. Access to electricity (% of population) for Sub Saharan Africa, from 1990-2016
7. Share of fossil fuels (% of total production) for Sub Saharan Africa, from 1990-2016
8 Population in EU-28 (in millions) for 1990-2016

Probably I have too little sample size (N=27) for running multiple linear regression, haven't I?

What kind of statistical method do you suggest to sufficiently test and interpret my hypotheses?
Any correlation methods?

Thanks a lot!
 

Karabiner

TS Contributor
#2
How many data points do you have, 27 (years), or 27 (years) *28 (countries)?

Probably I have too little sample size (N=27) for running multiple linear regression, haven't I?
Why multiple regression? What would the model(s) look like?

With kind regards

Karabiner
 
#3
How many data points do you have, 27 (years), or 27 (years) *28 (countries)?


Why multiple regression? What would the model(s) look like?

With kind regards

Karabiner

Hi Karabiner,

I currently have data for 27 years for EU-28, but surely I can get data for every country seperately.

I just mentioned multiple regression, since I remember it the best from the courses at university, but I would like to get some other ideas which method to choose.
My main purpose of empirical analysis is to test if there is a positive impact of higher production of renewable energy on economy. I'm looking to test my hypotheses to see if there is a positive correlation between Renewable energy investment and other macroeconomic factors.... Hope this helps :)
 

Karabiner

TS Contributor
#4
If you do not aggregate over 28 countries, but include data from each country,
then a multilevel model could be used. You would (amongst others) have the
opportunity to observe cross-sectional effects between countries as well as
longitudinal effects, and their interaction.

With kind regards

Karabiner
 
#5
If you do not aggregate over 28 countries, but include data from each country,
then a multilevel model could be used. You would (amongst others) have the
opportunity to observe cross-sectional effects between countries as well as
longitudinal effects, and their interaction.

With kind regards

Karabiner
Can I use data from each country for each year (27*28) into the regression model using SPSS?
Sorry, but I'm really confused right now about doing my statistical research.

Is there any simple method to compare higher rate of renewable energy production to lower emission/higher access to electricity/net import?

thanks a lot!
 

Karabiner

TS Contributor
#6
Your hypotheses just seem to imply 3 simple regressions or just 3 correlations.
This can easily be done with n=28. But I am not sure whether some simple
regressions /correlations will provide much insight into the phenomenon.
Maybe a Time Series approach is possible, but that is beyond my expertise.
I‘d suggest that you discuss the possibilities with your supervisor.

With kind regards

Karabiner
 
Last edited:
#7
Your hypotheses just seem to imply 3 simple regressions or just 3 correlations.
This can easily be done with n=28. But I am not sure whether some simple
regressions /correlations will provide much insight into the phenomenon.
Maybe a Time Series approach is possible, but that is beyond my expertise.
I‘d suggestive that you discuss the possibilities with your supervisor.

With kind regards

Karabiner
Thanks you on your answers!

BR