Question 1 Type the following command in Stata to load the MURDER dataset: use http://fmwww.bc.edu/ec-p/data/wooldridge/MURDER The dataset contains three years of data (1987, 1990, and 1993) for all...

1 answer below »

Question 1

Type the following command in Stata to load the MURDER dataset:
use http://fmwww.bc.edu/ec-p/data/wooldridge/MURDER
The dataset contains three years of data (1987, 1990, and 1993) for all fifty states plus the District of Columbia, where:

mrdrte
is the number of murders per 100,000 people

exec
is the total number of executions in the past three years

unem
is the annual unemployment rate
For this question, use the following command to keep only the observations from 1990:

keep if year==90

a) Using OLS, regress the murder rate (mrdrte) on the number of executions (exec). Do executions have a statistically significant impact on the murder rate?
b) Generate a scatter plot with the murder rate on the vertical axis and the number of executions on the horizontal axis. Do you see any outlying observations?
c) The observation with the highest murder rate is the District of Columbia (DC). Try running the regression from part (a) using a least absolute deviation model. Also, rerun it using OLS, but excluding DC (HINT: run your regression on the observations where state~="DC"). Does DC have a large impact on the regression results?


Question 2


Reload the MURDER dataset using the following command (since now we want to use the data from all three years):
use http://fmwww.bc.edu/ec-p/data/wooldridge/MURDER, clear

Hint:
For this question, you will need to generate a “first difference” for each variable. The commands to do this are:
bys id (year): generate fd_mrdrte=mrdrte-mrdrte[_n-1]
bys id (year): generate fd_exec=exec-exec[_n-1]
bys id (year): generate fd_unem=unem-unem[_n-1]
Consider the following model:
mrdrteit = _0 + _1execit + _2unemit + ai + uit
(a) If executions deter crime, what should the sign of ß1 be?
(b) Estimate the model using (pooled) OLS. Is there evidence of a deterrent effect?
(c) Estimate the model in first differences. Is there evidence of a deterrent effect?
(d) Estimate the model using fixed effects. Is there evidence of a deterrent effect?
(e) Estimate the model with random effects. Is there evidence of a deterrent effect?
(f) Suppose you have reason to believe there is a time-invariant characteristic of some states (e.g, a taste for violence) that leads to high murder rates and large numbers of executions. In this case, which of the above models do you prefer? Explain your reasoning.


Question 3


Type the following command in Stata to load the MROZ dataset:
use http://fmwww.bc.edu/ec-p/data/wooldridge/MROZ, clear
This dataset contains information on married working women. Use the following
command to keep only those women who are in the labor force:
keep if inlf==1
(a) Run an OLS regression of log wage (lwage) on years of schooling (educ). What are the returns to education? In order to address the concern that education may be correlated with unobserved factors such as ability, you would like to use mother’s years of schooling (motheduc) as an instrument for education.
(b) Theoretically speaking, is mother’s education a good instrument for education?
Explain your reasoning. [There is no wrong answer here, as long as your response
is well reasoned.]
(c) Empirically, test whether mother’s education is a relevant instrument?
(d) Reestimate the model from part (a), but this time implement 2SLS using mother’s
education as your instrument for education. Compare the economic and statistical
significance of your estimated returns to education with your results from part (a).
Are the results similar?
(e) Which are higher – your standard errors in part (a) or your standard errors in part
(d)? Is this as you would expect?
Answered Same DayDec 20, 2021

Answer To: Question 1 Type the following command in Stata to load the MURDER dataset: use...

Robert answered on Dec 20 2021
102 Votes
Answer as succinctly as possible while still fully answering the question. Keep in mind that a
simple “yes” or “no” answer is never sufficient – you are always expected to explain your
reasoning. To receive full credit, you must include your S
tata output.
Question 1
Type the following command in Stata to load the MURDER dataset:
use http://fmwww.bc.edu/ec-p/data/wooldridge/MURDER
The dataset contains three years of data (1987, 1990, and 1993) for all fifty states plus the
District of Columbia, where:
mrdrte is the number of murders per 100,000 people
exec is the total number of executions in the past three years
unem is the annual unemployment rate
For this question, use the following command to keep only the observations from 1990:
keep if year==90
a) Using OLS, regress the murder rate (mrdrte) on the number of executions (exec). Do
executions have a statistically significant impact on the murder rate?
Below is the summary of OLS results –


The numbers in the bracket are values of t-test statistic. As we can see, exec do not have
any statistically significant impact on the murder rate. The t-test statistic value of the exec
coefficient is less than what it should be it to be significant at 5% level of significance.
b) Generate a scatter plot with the murder rate on the vertical axis and the number of
executions on the horizontal axis. Do you see any outlying observations?
http://fmwww.bc.edu/ec-p/data/wooldridge/MURDER


Yes, there are outlier observations which mainly belong to the District of Columbia.
c) The observation with the highest murder rate is the District of Columbia (DC). Try
running the regression from part (a) using a least absolute deviation model. Also, rerun it
using OLS, but excluding DC (HINT: run your regression on the observations where
state~="DC"). Does DC have a large impact on the regression results?

Quintile Regression is one type of least absolute deviation model where we take
a deviation from the median. Below are the results of a Quintile regression –


We notice that executions have a significantly positive effect on murder rate
which is counter-intuitive since we expected that executions will have a negative
effect on murder rate.

Now, we may try another alternative OLS model where District Columbia is
excluded. Below are the results –


We find a significant positive relationship between murder rate and...
SOLUTION.PDF

Answer To This Question Is Available To Download

Related Questions & Answers

More Questions »

Submit New Assignment

Copy and Paste Your Assignment Here