Jack and John regarding your question about joining two tables. On the left side of the equation, we'll have research.id, research.year, research.country, research.author, research.journal, research.org. On the right side, we'll have aid.year, aid.country (coded either as iso2, iso3 depending on digits), aid.amount (use both disbursement and commitment in constant/real $). On the right side, we'll also have policy.id, policy.year, policy.country, policy.author, policy.org. Common variable against which we join the table is research.year and aid.year. It needs to be the same calendar year. We can try a dynamic model, taking into account a lagged response (time gap between research and policy response).