This preview shows pages 1–2. Sign up to view the full content.
1
Homework # 4
ECO 7427
Prof. Sarah Hamersma
Due Wednesday, March 16 (in class)
1.
Use DataFerrett to view the variables available from the 1990 U.S. Census.
The data set is called
“Decennial Public Use Microdata Samples” or PUMS.
Use the 1% Person sampl
e.
Download a
limited set of information for each person: age, hours usually worked per week last year, and wage
and salary income last year.
(a) Get the data into Stata format and
report the summary statistics for the variables you
decided to download (this is all you should submit for part (a)).
Please do (b)  (d) below for
each
of the following three variables:
(1) age
(2) hours usually worked per week last year
(3) wage and salary income last year.
For the last two, leave out people who did not work and/or did not earn any income last year
(you do not need to drop them from the data
–
just restrict your commands to positive values
using an “if” statem
ent at the end of each command  see Stata help for details).
In the commands below, I will refer to the variable as X.
You should save each graph you
generate (do this in your dofile directly
–
it will be much more efficient than separately saving
each graph as it pops up).
You can then open the graphs later and paste them into a
Microsoft Word document, along with any other tables out of your Stata log files, so that you
can give me a single document with your answers typed in between.
(b)
Generate a histogram of variable X, using Stata’s default
(i.e. do not indicate a particular bin size
or width).
Does this do a reasonable job of displaying the distribution?
If there seem to be strange
or interesting patterns, describe them and indicate any reason you think they may be there.
(c)
Report the mean and standard deviation of X (again, not including the zeros if there are any).
Do these do a good job of summarizing the distribution of the data?
(d)
Estimate the distribution of X using a kernel density estimator (kdensity; see Stata help for
details).
Allow Stata to choose your default kernel and bandwidth (which they refer to as a window
width).
Does this kernel estimator with the default bandwidth seem to do a nice job of smoothing
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview. Sign up
to
access the rest of the document.
 Spring '08
 HAMERSMA

Click to edit the document details