## Economic Data and Theory

16 Oct

Data and Methodology for: “Can a higher minimum wage reduce poverty?”
1. A brief description of the data sources. You should mention the source (name of survey/dataset, agency, etc), time frame, geographic region, data structure, and the sample size. Also briefly describe the dataset. If you have multiple sources, you need to specify every data source. Also, you will have to put the correct citation in the reference section.
2. Justification. Explain why you use this data source. For example, why this year, why this region? Why do you look at states, metropolitan areas, counties, G7 countries, or certain cities? Why is this dataset particularly suitable to answer your research questions? This is especially important if you are only using part of the available data. For example, unemployment rate between 1980 to 2012. In this case, you will need to justify and be ready to defend your selection of years when you present.
3. If you combine different sources to build your dataset, you should also include how you actually construct the dataset. In this case, you may experience some attrition (losing some observations). If this is the case, explain how this may affect your research outcome if at all. Also state the final number of observations (sample size).
4. The variables. You need to talk about what variables you are going to use. a. Define each variable if it is not obvious like unemployment rate or GDP growth rate. For example: time spent playing outside, green house gas emission, or inequality. etc. b. Specify the unit of each variable. (percentage, dollar, years, etc) If it is unclear, you will be asked during your final presentation. c. For dummy variables, describe how they are set up. (ex. Female=1 and 0 otherwise, ethnicity variables, etc)
5. The type of regression. Example: OLS or probit or other.
6. The actual regression equation, dependent variable, independent variable(s) and other controls such as regional dummies or time fixed effects.
7. The expected (theoretical prediction) of coefficients to independent variables. You may not need to the expected sign for every variable. Rather, you should focus on the most important ones. For example, does being bilingual affect one’s wage. Be very careful because in some cases, the sign may be ambiguous due to contradicting theories. If this is the case, make sure you explain both sides of the theory.