First and foremost, hello to everyone. I have been lurking around for a while and I finally decided to join.
With no further ado, back to the question at hand. I am working on my MSC Dissertation and using a gravity type model. I am investigating UK Outward FDI (53 countries) vis-a-vis institutional quality and cultural proximity. I created a working model and upon presenting this to my supervisor he demands a principal factor component analysis. As an economist, I am looking in finance books and psychology books to comprehend how to interpret the eigenvalues and eigenvector loadings. The econometrics books I've used in my undergrad and master degree do not cover this. This technique seems to be avoided by economists. I have five criteria for institutional quality and the correlation coefficients between these variables are very high. Consequently, identifying their partial effects on the variable of interest in very difficult. I am running the principal component analysis on these 5 criteria and if I use the Kaiser criterion I end up with 1 factor and if I use Catell's scree test, I end up with two.
Can someone explain to me how I can interpret the eigenvector loadings in my case? The way I understand it, PC1 might be interpreted as a general index due to the fact that this part of the analysis shows the linear combination of the coefficients. I might be missing something because my Principal Component Analysis is a pre-OLS test to find the best model by avoiding the inclusion of all the variables and removing them sequentially by looking at the most insignificant.
Thanks for taking the time to read this,
Wish ...