ECON 2223 - Assignment III Charles Saunders, Western

ECON 2223 - Assignment III
Charles Saunders, Western University
Due date : Thursday March 10, 2016 at 4 :00 pm
Assignments must be submitted as a hard copy, in class or my office.
You must submit your written results (tables and discussion) as well as your script
file. Do NOT give me an annotated “copy and paste” of your log file.
1. Start a new script file.
2. Bring in the Ghana Demographic and Health Survey (DHS) data : GhanaDHS.csv
3. Tabulate age, education and daily earnings, and include : the number of observations, mean, minimum and maximum each series.
4. Tabulate the observation counts for : region, electricity, radio, television, fridge,
bicycle, motorcycle, car, marital, and slocal.
5. (Model 1) : Regress daily earnings on age, education, region, electricity, radio,
television, fridge, bicycle, motorcycle, car, marital, and slocal, and present and
discuss the results. Discuss your results, which variables are significant ? How are
the coefficients on electricity, region, age and education interpreted ?
6. Provide a histogram of the Model 1 residuals.
7. Generate a new series : the log of daily earnings.
8. (Model 2) : Regress the log of daily earnings on age, education, region, electricity,
radio, television, fridge, bicycle, motorcycle, car, marital, and slocal, and present
and discuss the results. Discuss your results, which variables are significant ? How
are the coefficients on electricity, region, age and education interpreted ?
9. Provide a histogram of the Model 2 residuals.
10. For Model 2, which variables could be dropped without any significant effect on
the overall regression ? Justify your decision(s). Estimate this model (2b) with your
restrictions imposed, and present the results.
11. Run an F-test that compares Models (2) and (2a). (Hint : anova function in R).
12. Generate a new series called experience (exp) that is equal to age minus education,
also generate experience squared (exp2). Re-estimate model 2 and include exp and
1

exp2 in the regression. Explain what happens ? And present the coefficients and
t-stats on exp and exp2.

2