PSC 104 (10): Methods of Public Policy Analysis
Data Analysis Assignment 1
Due: September 30, 2002
I. The questions in part I require you to collect data on a variable that you are interested in
exploring over the course of the semester. “Collecting data” is broadly defined for the
purposes of this assignment. You can get data off the web that has been compiled by
others, you can type in data from a reference book, or you can collect original data of
your own. The data you collect for this assignment will be the dependent variable in your
analyses, so choose carefully. The data must be numeric, and you need to have a
minimum of 50 observations in your data set. Be sure to answer all questions posed.
(a) Describe your data in words. What phenomenon of interest will you be
analyzing? What, exactly, was the source for the data? You need to provide
enough information so that your reader could find the data, given your
description. Why would it be important to explain variation in the variable? For
example, assume you are collecting data on cheese consumption per capita in the
United States from 1945 to 1995. Why is it worth explaining variation in cheese
consumption over time?
(b) Describe the central tendency of your variable. What is the mean of the variable?
[Use summaries, means and SDs in the StataQuest menus] What is the median of
the variable? [Use summaries, median/percentiles in the StataQuest menus] Is the
mean similar to the median in your data? If not, what does that mean? If so, what
does that mean?
(c) Describe the dispersion of your variable. What is the standard deviation of your
variable? [Use summaries, means and SDs in the StataQuest menus] What is the
interquartile range of your variable? [Use summaries, median/percentiles in the
StataQuest menus] Does there seem to be much dispersion in your variable?
(d) Now describe your variable graphically. First make a box plot [Use graphs, one
variable, box plot in the StataQuest menus] Next make a histogram [Use graphs,
one variable, histogram, continuous variables in the StataQuest menus]. What, if
anything, do you learn from the graphs that you don’t learn from the numeric
summaries of (b) and (c) above?
II. The questions in part II must be answered using the Florida data available on my web
site. Go here: home.gwu.edu/~edl/florida.zip and you can download the file. Download
that file, and once you have it, unzip it. Once unzipped, there will be a Stata file called
FloridaVote.dta, which contains 2000 Presidential election results, by Florida County.
You may prefer to save the files to a floppy disk so that any changes you make to the data
will be preserved.
(a) Using the variable for the number of Nader votes, by county (nader) and the variable
for the number of Buchanan votes, by count (buchanan) repeat steps (b)-(d) from part
I for each variable.
(b) Based on the graphs, do the distributions of the two variables look similar? Why or
why not?