Fall, 2001 (Roos, Soc. 311)

Assignment 2: Introduction to SAS: State Data (due September 28)

This assignment is intended to give you a quick immersion into doing statistical analysis using SAS in the IML (Tillett Rm. 119). I have provided a sample sas run and sas output to help you out). The assignment involves gathering data on the 50 states (and the District of Columbia) and doing statistical analyses on these data.

Before starting this assignment, you should read Ch. 14 in Babbie, and all the SAS reading noted in sections I and II on the course outline (and pp. 176-77 on PROC CORR). Pages given are for the second edition of the SAS manual (look up PROC CORR in the index to get pages #'s if you have the first edition). Most important: you will also need to read section 2.11 in the second edition of the SAS manual (click here).

Bring a 5 ¼" diskette, so you can save your work, since you may not finish in one sitting.

1) The first step involves gathering data on the internet. Your task is to gather interesting data on three variables for each of the 50 states and Washington, D.C. You can either choose from among the websites I've noted below, or from a website of your choice (just make sure the data are from a reputable source). You may need or wish to collect from two or more websites.

All the following websites provide state data. The first two give state-level census data for a variety of possible variables (for example, population counts, education, infant mortality rates, personal income, poverty rates, race percentages). For my example, I used data from the first census website and the Children's Defense League (3rd website).

http://www.census.gov/statab/www/ranks.html (Census: state rankings on basic data)
http://quickfacts.census.gov/qfd/index.html (Census Quick Facts on states)
http://www.childrensdefense.org/statesdata.htm (Children's Defense League data, by state)
http://teacher.deathpenaltyinfo.msu.edu/overviews/statesoverview.htm (Death Penalty Statistics, by state)
http://www.acf.dhhs.gov/programs/cb/publications/cwo98/Sec5/states.html (Child Welfare Outcomes, by state)
http://www.iwpr.org/states/index.html (Institute for Women's Policy Research, state data on women's rights)
http://www.edweek.org/sreports/tc99/states/usmap.htm (Education Week on the Web's Technology Counts data, by state)
http://www.iusb.edu/~jmcintos/USA98StatesTab.htm (Suicide Data, from the American Association of Suicidology, by state)

If you want to look for other variables of interest, start with a search engine, and type "state data". I found the above through google.com.

2) From these data, choose three (n=3) interval-level variables, in addition to the name of the state. Let your interests guide you. Choose variables that you think are associated with each other.

To make your study more interesting, it's best to avoid variables that are raw numbers; choose rates or percentages instead. State comparisons of raw numbers (i.e., the # of children in poverty) will simply reflect the fact that bigger states like California or New York have more of just about everything (e.g., crime, children in poverty, death penalty cases). Better to use rates or percentages, which control for population size (e.g., # of children in poverty per 1,000 people, or % of children in poverty).

Print out the pages you use for your three variables and turn them in with your assignment. (Note: this assumes that all your state data can be summarized on a page or two; if not, do not print out 51 pages! If you need to go to separate pages to get data on each state, provide 2-3 states to illustrate how the data are provided). Important: write the URL on any web page you turn in!!

3) Code the state data directly onto the SAS coding sheet (SAS Viewtable Window), as we talked about in class (get to Viewtable Window from within SAS by clicking Tools then Table Editor). Use a period for any missing data.

Print out a copy of your completed SAS Viewtable Window (your coding sheet) and turn it in with your assignment.

4) Write up and run a SAS program to evaluate the data you collected (see my sample run and sas output). Your program should do at least three things:
PROC PRINT (which will list the data you input)
PROC MEANS (which will give you basic statistics for your variables)
PROC CORR (which will give you correlations between your variables)
You can find descriptions of each of these procedures in the SAS Manual.

Print out your SAS log and SAS Output and turn it in with your assignment.

5) Interpret your results in no more than two double-spaced pages. Don't just describe your results, say something interesting and substantive. Try to be sociological. What are the average values of your variables? How do your variables correlate with one another?

In using data from any source, it's important to know the reliability of the data source. Please include a paragraph in your final write-up that describes where the data came from, and your assessment of the reliability of the data source (include the URL as a citation for the data).

Along with your write-up, please turn in your SAS Viewtable Window, SAS log, SAS output, and the printed web pages for your variables.

THIS ASSIGNMENT MUST BE TYPED.