Daily Assignments
November 29
1) Study Chapter 10.5
(mostly examples)
2) Assignment #6 due on
Thursday
November 22
1) Study Chapters 10.1 and
10.2 (skip distributed lags)
2) Do this practice exercise (Due in class
on Tuesday 11/29)
3) Solution
November 17
1) Study Chapter 17.1
2) Do this practice exercise (Due in class
on Tuesday 11/22)
3) Solution
Advanced Posting:
*Research idea*
This will be due (together with another short exercise) on
Tuesday November 22
Think of a small research project that you would like to
perform. This can be something like the analysis of the gender
gap in wage that we did in class, or estimating the impact of a
policy (like the many examples described in Chapter 13, etc).
Look at Chapter 19, there are lots of ideas too. There are also
lots of ideas on effects of policies in the news (NYTimes, the
Economist), and, as you saw from the assignments, lots more
ideas in the American Economic Journal: Applied Economics or
Economic Policy on the web at
http://www.aeaweb.org/aea_journals.php )
Write a (maximum 1 page) description including the following:
Why is the relationship/research interesting? What equation(s)
would you estimate? What data would you need to estimate the
relationship? Where could you find such data? What would you
expect to find? Will there be a problem in the estimation of
omitted variable bias? How would you deal with such a problem?
November 15
1) Study Chapter 7.5 and
Beginning of 17.1
2) Do this practice exercise
3) Solution
to first problem
4) Solution to second problem
November 10
1) Study Chapter on Impact
Analysis (Posted on bSpace)
2) Assignment #5 is due
next Tuesday
November 8
1) Study Chapter 13.3 and
13.4
2) Do this practice exercise
3) Assignment #5 is due
next Tuesday
4) Solution
November 3
1) Study Chapter 13.3 and
13.4
2) Do this practice exercise
3) Solution
to 2009 #5
4) Solution to 2010 #3
November 1
1) Study Chapter 13.1 and
13.2
2) Do this practice exercise
3) Solution
October 27
1) Study Chapter 7 (skip
7.5)
2) Assignment #4 due next
Tuesday
October 25
1) Study Chapter 7.1-7.4
2) Here is the practice exercise
3) Solution
October 20
:
1) Study Chapter 6
2) Do practice exercises
C6.5 and C6.8. Here are the data. DUE
NEXT TUESDAY IN CLASS.
3) For those of you that
don't have the textbook, here is a scanned version
4) Solution
to C6.5
5) Solution
to C6.8
October 13
:
1) Study Chapter 6 (Skip interaction terms)
October 11
:
1) Study Chapter 6.1-6.3
2) Practice Midterm
Questions
October 6
:
1) Study Chapter 4
2) Assignment 3 due next
Tuesday 10/11
October 4
:
1) Study Chapter 4.3 and 4.4
2) Practice exercises are
posted here.
3) Solution to practice
problem is in solutions to midterm practice exercises
4) Practice exercises for
midterm will be posted soon
September 29
:
1) Study Chapter 4.1 and 4.2
2) Do these practice exercises.
Solutions to all problems except 2 come from midterm practice
problems. Solutions to the second problem are here.
September 27
:
1) Study Appendix C5, C6, and handout
2) Do these practice
exercises.
3) For solutions see the
midterm practice problem solutions
September 22
:
1) Study Appendix C1, C2, and B5 (normal
distribution)
2) Problem Set 2 due next Tuesday at 9:45 AM
September 20
:
1) Study Chapter 3
2) Practice Exercise 3.1 (for your own
practice, not to be turned in)
3) Solution
September 15
:
1) Study Chapter 3.1 to 3.3 (Skip Partialing
out)
2) Do practice exercise C3.1. Here is a
scanned copy. Here is the data.
Be sure to save this file as BWGHT9.dta file on your hard drive.
When you open stata, change your directory using the cd command.
To read in the data, type use BWGHT9.dta, clear. After that all
you have to run is the reg command. NOTE: THIS EXERCISE IS TO BE
TURNED IN ON TUESDAY SEPTEMBER 20TH.
3) Solution
Kyle Comments:
The main goal of this problem was to get you started
thinking about ommitted variable bias and running some basic
regressions to evaluate how a parameter estimate changes when
you include an additional variable in the regression. Including
faminc should have resulted in your estimated coefficient on
cigs going from -0.6 to -0.55. A great answer to this question
would explain this difference by noting the sign of the
correlation between cigs and faminc (negative) and the sign of
the effect of faminc on birthweight (positive). Therefore, the
bias in the effect of cigs when we do not control for faminly
income is a downward bias. But since the correlation between
cigs and faminc is not that large, the change is only from -0.6
to -0.55.
September 13
:
1) Study Chapter 2 (except 2.6)
2) Do practice exercise 2.6 in Chapter 2 (for
your own practice, not to be turned in)
3) Solution
September 8
:
1) Study Chapter 2.1-2.3
2) No practice exercise today, Assignment 1
due Tuesday 9/13
September 6
:
1) Study Appendix B2-B3
2) No practice exercise today, Assignment 1
due Tuesday 9/13
September 1
:
1) Study Appendix B1
2) Do practice exercise 2.3 in Chapter 2.
Here is a scanned copy
(Due in class on Tuesday
9/6)
Kyle Comments: There were a few errors in calculations. It is a
good idea to use excel to do these calculations. One way to
check if you have made a mistake in calculating the regression
coefficients is when predicting the GPA for an ACT of 20.
If you get a very unreasonable value for your predicted
value of GPA, then something probably went wrong. Some people
also had trouble interpreting the R-squared of the regression.
We will talk more about how to interpret the R-squared in words
tomorrow. For those that would like to see, here is a
brief solution that I
prepared.
August
30
:
1) Study Appendix A4
2) Do practice exercise A7 (for your own
practice, not to be turned in)
August 25:
1) Study Chapter 1 and Appendix A2-A3
2) Create a scatter plot in excel using these data or any other data
(Due in class on Tuesday 8/30)
Kyle Comments: The
scatter plots were well done. When making a scatter plot, note
that "connecting the dots" has little meaning. Some of you
included a regression line, which is good. A line connecting the
dots however does not have a meaning. Some of you chose to label
some of the points with country names when possible. I think
this is a good way of displaying a scatter plot when the unit of
observation has a meaningful identification (i.e. countries,
states, etc.)