FP220

Your assignment is to analyze a real problem using a data set. This problem will have a political science/public policy flavor, but I interpret those boundaries liberally. You are expected to locate these data, clean them up so that they are compatible with R, explore them, analyze them, and interpret your results. It is not a traditional research paper; your focus is on the mechanics of data analysis.

Do a Google search on your topic, or look in JSTOR for a relevant journal article that might reference a study with a data set. There are also major data libraries available online through ICPSR at the University of Michigan, and at government sites such as the World Bank or the Census Bureau. The point is to find data on your chosen variables (driven by theory and your hypothesis) out in the world and bring it together in one unique dataset of your own creation. For example, you might find data for your DV on a website at a think tank; other data for your main IV on the World Bank website; and data for various confounding variables in other locations on the internet. You will then bring it all together in a single Excel spreadsheet and import it into R/Rcmdr for analysis.

Your data must meet the following requirements:

• You must have at least 50 observations; more is better.
• Your DV must be a continuous (or approximately continuous) interval variable.
• Your key IV can be continuous or discrete.
• You must have at least three alternative explanation IV (“Z”) variables.
• At least one of your Z variables should be an interval variable.
• At least one of your Z variables should be a categorical (nominal or ordinal) variable.

** Note the use of the word “should” here. Should does not equal “must.” The most important thing is appropriate variable choice for your project based on your question and your specific project data. You may end up having IV/Z variables that are mostly one type.