I'm pretty newbie in econometrics, I'm writing a research proposal, and I was wondering if you guys can give your opinion or some hints regarding an idea. My dependent variable is a measure of scientific output(number of papers), impact (number of citations). One of the problems of aggregating the citations and publications, is that this patterns change across disciplines. In other words, one need to account for the disciplinary composition (or specialization) of a country to compare. For that reason I would like to decompose the scientific production (papers) of a nation by 22 core disciplines. Using a production function approach, my explanatory variables are R&D investment, human capital, and measurements of co-authorship networks as a proxy of social capital. I'll have 22 observations per country in each year.I have the following two questions
1)Which is the minimum amount of countries that I should include to be feasible for a regression study? (I'm thinking on 5 countries)?
2)Gross Expenditure on Research and Development (GERD), can be decompose in three subdivisions (Goverment(GOVERD), University(HERD), Private Sector(BERD)). Since the objective of this study is to understand scientific output and impact, studies like this have considered using HERD only, assuming that universities, laboratories and research institutions are the ones interested more on scientific publications. However, I'm interested in Latin America, and is really difficult to find HERD, but must economies have GERD (which include HERD in some proportion). Question: Is possible to include GERD as a explanatory variable?
Thank you very Much,