data Analysis project

Econ 3100 Writing Project: Results section checklist Writing Project: Checklist for the results section-table of regression results _____ The table of regression results has a title. _____ The table of regression results is laid out in the conventional format and self-explanatory. _____ The table of regression results contains the coefficients, t-stats or p-values for each, the R-

squared and adjusted R-squared for each model, and the sample size (if it changes from model to model).

_____ Coefficients and p-values in the regression results table are rounded consistently. _____ The text and table(s) are well integrated; the text refers to the table of results where appropriate. Writing Project: Checklist for the results section-written portion

DEVELOPING THE MODEL:

_____ The results section clearly states the methodology being used. _____ Dummy variables are used correctly to incorporate zip code in the model _____ Multi-colinearity is discussed and addressed; correlation coefficients between independent

variables are calculated to identify potential multicollinearity. _____ Outliers and observations with unusually high leverage are identified and addressed. _____The statistical significance of coefficients is discussed and addressed; the possibility of eliminating

variables with low statistical significance is considered. _____A partial F-test is conducted for the joint statistical significance of the dummy variables for zip

code. _____ There is a clear discussion of each model specification presented and why it is preferred or not

preferred to other possible specifications. INTERPRETING THE MODEL: _____ Both the magnitude and the statistical significance of the coefficients in the preferred

specification are discussed. _____ The discussion of the regression coefficients uses the appropriate units and relevant magnitudes.

APPLYING THE MODEL: _____ The paper gives real-world examples of how the preferred model could be used to predict the

prices of homes in the area. _____ The paper discussion the implications of the model for area home prices. Example of a regression results table (from sample paper): Note: While the sample paper gives the standard error of the coefficient in parentheses, it is also acceptable—even preferable--to give the p-value associated with the coefficient as an indication of its statistical significance, rather than the standard error.

Here is another example of a regression results table:

Notice a few things: 1. The model starts by including all the variables, then eliminates those with weaker relationships to Y. 2. The adjusted r-squared improves (grows) as variables are eliminated. 3. The table gives the value of the coefficient and some indication of statistical significance (here, the tstat. Could also be the p-value). 4. An asterisk is used to indicate variables that are statistically significant at the 5% level.