Project

In this course project you are required to replicate the results in McCarthy (1994) ("the paper" hereafter). The paper and the data can be downloaded from the course web page. You are required to submit a report for the project. Note that you may not be able to get exactly the same numerical results for unknown reasons.

In the project you are asked to replicate some of the results in the paper. You can extend the paper if you wish, but extensions are not required.

Speci?c requirements about the project include:

1. Read the paper and understand the author's arguments.

2. Inspect the data and familiarize yourself with the data.

3. Present the summary statistics (mean, standard deviation, minimum, and maximum) for all the variables in Table 0 in your report.

4. Try to replicate columns (1), (2), (3) and (4) in Table 1 of the paper and report your results in Table 1 in your report. Remember in this table only the coef?cient for the variable spdlaw is reported in each regression. You should also report the t-statistic and the adjusted R-squared, ¯ R2, like the paper did. Notice that the robust standard errors were used in the paper. You can use either the usual or the robust standard errors in your report. If you want to use the (White) robust standard errors, in your regression command add an option "robust" at the end so that the command looks like regress y x1 x2, robust. There is a typo in the footnote in Table 1 of the paper: the independent variables should include time trend t. Your numbers in columns (1), (3), and (4) should be very close to those in the paper. Slight differences can be expected due to numerical issues. You will notice that your results in column (2) are different from those in the paper (for unknown reasons).

5. No need to replicate Table 2.

6. Start writing your report. You should have the following sections in your report:

• Introduction. Tell us what you want to do in this report (replicating the paper).
• Data. Acknowledge the data source (i.e., McCarthy (1994)). Present the summary statistics (Table 0). Brie?y explain what a variable means if it is not obvious.
• Model and Results. Present the model in equation format. Present Table 1 and discuss your results. You should also present the OLS result for "All roads" and "Total accidents" in equation format. Interpret each coef?cient in this equation.
• Conclusion. Wrap up your report in this section.
• References. List the paper and any other references you use for the report.
• Appendix: print out your Stata program (the commands you use to replicate the paper).

7. The report should be double spaced on letter size paper, There is no required minimum number of pages. But it should not be more than 10 pages (tables and references included). The report should be typeset in 12 point fonts.

8. You are strongly encouraged to ?nish the report as early as possible, though the due date is much later. In fact, you should be able to ?nish the project now. There is no reason to wait given the typical end-of-semester rush.

9. Description of the variables:

totacc                    statewide total accidents

fatacc                    statewide fatal accidents

injacc                     statewide injury accidents

pdoacc                   property damage only accidents

ntotacc                   noninterstate total acc.

nfatacc                   noninterstate fatal acc.

ninjacc                   noninterstate injur acc.

npdoacc                 noninterstate property acc.

rtotacc                   total acc. on rural 65 mph roads

rfatacc                   fatal acc. on rural 65 mph roads

rinjacc                   injury acc. on rural 65 mph roads

rpdoacc                 property acc. on rural 65 mph roads

t time                    trend

tsq                        t^2

unem                    state unemployment rate

spdlaw                  =1 after 65 mph in effect

beltlaw                  =1 after seatbelt law

wkends                 # weekends in month

feb                       =1 if month is Feb.

mar

apr

may

jun

jul

aug

sep

oct

nov

dec

Some comments regarding help:

1. If you want to use a subset of a sample to run a regression in Stata, you can use if to limit the observations in your regressions. For example, if you only want to include those whose age is above or equal to 65 in a regression, you can say regress medical age female bmi if age >= 65

2. If you need clari?cations on the requirements, I will be happy to answer your questions. Otherwise I do not offer "help" on the project. For example, following questions will not be enter- tained: "Which command should I use for ...?" "My numbers are different from the author's. Should I keep trying or just present what I have?"

References - McCarthy, P.S. (1994), Relaxed speed limits and highway safety: New evidence from California, Economics Letters, 46: 173-179.

Attachment:- Assignment Files.rar