Methods II: Research Methods / Economics 303
Profs. Chesley&Fenn

Applied Regression Assignment

(Due 3:00 p.m. Today )

For this exercise you will be given the choice of working with one of two data sets that have been developed and used by various members of the Economics and Business Department for use in their research.

The objectives of this exercise are to:

  1. Develop a regression model from a data set.
  2. Understand why some models explain the data better than others. IN THE ASSIGNMENT YOU ARE ALLOWED TO PRESENT ONLY ONE REGRESSION MODEL. ONE THAT YOU DEEM TO BE THE BEST.
  3. Understand whether or not assumptions about the data for ordinary least squares regression (OLS) are met.
  4. Explain the statistical significance of the model based on its descriptive statistics.
  5. Interpret the implications of the estimated coefficients.
  6. Demonstrate how to present regression results in a clear and succinct manner.

Follow the guidelines outlined in the handout, “Summary – Building a Regression Model.” These will lead you through the steps needed to understand the data and build a model. PENALTY FOR EXCEEDING LENGTH LIMITS IS 5 POINTS PER ANSWER.

The write-up. 4 pages, maximum, typed, double-spaced.

THIS SECTION SHOULD BE BASED UPON A FINAL REGRESSION EQUATION. YOU MAY RUN SEVERAL DIFFERENT REGRESSIONS BUT PICK THE BEST ONE AND ANSWER THE FOLLOWING QUESTIONS.

  1. Prepare a brief document of your results. This document should include the following:
  1. A brief statement of the objective – what are you trying to explain with this model (1 paragraph) 5 Points
  1. Your model and your predictions about the signs of the variables in the model. Justify these predictions based on your understanding of the relevant economic theory. (1-2 paragraphs) 10 Points
  1. The model, stated as an equation with the t-statistics of the estimated coefficients given in parentheses below the equation. 10 Points
  1. An assessment of this model including whole model tests (F-tests & R-Square), the statistical significance of the coefficients (t-tests) and tests regarding the satisfaction of the assumptions about the residuals (i.e. homoskedasticity, normality and no serial correlation) (1-2 pages) Apply fixes for an econometric problems that you may detect. 15 Points
  1. Interpret the economic meaning of the model i.e. discuss the impact of the significance and magnitude of the coefficients and fit of the model for your predictions from part b (1 paragraph) 10 Points