Research Project-
Goal: Conduct a short, independent research project using stata and present the results orally and in writing.
Specifics: Students are to conduct a study on the topic relating to political science or politics in general and must be specific not a board topic. This project must be quantitative in nature and use the statistical tools learned in class. The report should be at least 7 pages of text (double spaced, 12pt font) The text should highlight the research question motivating the project, briefly discuss some previous political science research (at least 5-6 scholarly journal articles or books cited, no websites or newspapers count as citations), present your theory offering an answer to the research question, and have a cause and effect, provide at least one hypothesis, present your empirical analysis, and draw a conclusion about your theory from your empirical analysis. The text should be written as a research paper, not as a list of bullet points.
Data: Use a unique dataset that you have collected (i.e. do not use Pollock's datasets or a dataset used for a previous assignment). These data can take a number of forms, like public opinion polls (many are available for free download from various polling firms), government data (turnout, crime, economic data), demographics (Census data), etc. Data should either be obtained from a random sample (all professionally conducted surveys will use random sampling) or contain data on all available units in the population.
Presentation of results: Your statistical tests should be presented in appropriate tabular and/or graphical forms. I expect these to be professionally formatted.
- Introduction
- Background includes your theory and hypothesis
- Methodology / Data
- Results
- Conclusion
- A copy of the do file from stata
Analysis: Use a full range of statistical tools available to you. Each project should include a table of descriptive statistics for all variables used in your analysis, a simple measure of bivariate comparison (cross-tab, mean comparison, or scatterplot), significance tests (difference of means or proportions, χ2, or a significance test on a measure of association), a measure of association (λ, Cramer's V, Somer's d, or Pearson's r correlation coefficient), a bivariate regression, and multiple multivariate regression controlling for rival explanations. Each test should be appropriate for the levels of measurement of your independent and dependent variables.