Neophyte question: time series regression

[Natalie Platts]

First post from a lurker. Be gentle

I have a dataset where I have annual observations for my explanatory variables (1980-2000) for 180 countries but only two observations for my dependent variable (1998 and 2000) for the same number of countries. I am trying to understand the effect of the explanatory variables (socioeconomic and political) on the dependent variable (% of people who have received a specific vaccination in each country) but I'm not quite sure how to start. I will likely only use the data from 1990-2000 because that's what makes sense from a theoretical perspective.

So my question is: is possible to run a time series regression on a dataset with this structure? And if so, what model would you recommend?

(I know this is probably a very basic question for most of the people here. Self-taught and still a lot to learn!)

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

With only two observations on the dv and them not in sequential years, then I'd be tempted to run two separate cross-sectional regressions. You have a time series-cross sectional data set (see Stata pdf documentation for [xt] Longitudinal Data/Panel Data). However, you only have two observations within each panel. It is also likely that most of your variance is cross-sectional rather than longitudinal which may make the standard fixed effects panel estimators problematic. If the cross-sectional influence is different than the influence within country over time, a fixed effects estimator that throws out cross-sectional variation may miss the real action.