17.871Spring 2002Group ProjectsAssignment summaryWorking with your assigned group, answer the question posed to you. You will give a 15 minutepresentation (with 5 minutes available for questions) on your work on March 5. Your group willalso turn in a five-page written report on your project on Friday, March 8. (Please submit a hardcopy and then e-mail Steve and me an electronic version—virtually any format will do, exceptpdf.) The report should be in the form of a memo. It should describe how you measured thevariables of interest to you, report where you gathered the data, and summarize your findingsusing the appropriate figures and tables. (The five-page limit includes tables and figures.)Statement about CollaborationYou are encouraged to seek and extend as much help as you can, both within and betweengroups.GradingI will assign a letter grade to each group’s project. That will be the grade you receive, plus orminus an adjustment that will be determined as follows: I will ask each member of the group toindicate the relative amount of effort each person contributed to the successful completion of theproject. If someone in the group stands out as being a conspicuous over- or under-contributor tothe group effort, that person’s letter grade will be adjusted upward or downward as appropriate.2Project 1: Ballot Design and Voting Machines in FloridaGroup: Johanson, Muniz, Priebe, TappanBackground. In the 2000 presidential election controversy arose over voting technologymalfunctions and ballot confusions. Policymakers and social scientists have attempted tounderstand the degree to which these problems affected the outcome of the election. Perhaps justas important, reformers have attempted to bring about changes in Florida so that problems won’thappen again. In order to do that, it is important to parcel out responsibility for the problems in2000 between, say, bad machines and bad ballot designs.Question: What was the bigger problem in Florida in 2000, bad voting machines or bad ballotdesigns?Possible explanationsVoting machines. Some machines may be more likely to fail to record a vote cast by avoter. A good example is problems with “dimpled chad,” which usually indicatesthat a voter intended to cast a vote for a candidate, but which will usually gouncounted because light doesn’t shine through the card. Punch cards areparticularly prone to under-voting.Ballot design. Some counties in Florida used ballot designs that many people thoughtwere confusing to some voters. The best-known was the “butterfly ballot” inPalm Beach County, but some other counties ran a “caterpillar ballot,” in whichcandidate names were displayed across two columns. Both ballot designs(compared to a simple intuitive design, with candidates in one column) mayencourage both more over-votes and more under-votes.County demographics. Even if voting machines and ballot designs may seem to haveaffected whether voters’ votes were counted, the real problem might be that theseproblems just happened to coincide with counties that had an especially largenumber of voters who didn’t have much experience in voting, such as the youngor highly mobile. Once properly accounted for, the previously-mentioned effectsmay have been spurious.Data sourcesFlorida Division of Elections Web site. URL: http://election.dos.state.fl.us/. Contains aton of downloadable data, including election returns from the 2000 generalelection in Florida. It also contains information about voting systems used inFlorida. (As of January 29, 2002 the table that’s on the web site appears to havebeen accurate for the 2000 general election.)National Opinion Research Center Florida Ballots Project Web site. A consortium ofnewspapers employed the National Opinion Research Center to review everyover- and under-voted ballot in Florida. The Washington Post has run a number ofarticles on the data generated by this effort, and they can be viewed at3http://www.washingtonpost.com/wp-srv/onpolitics/elections2000/recount/front.htm. NORC has made the data available to the public for independent analysis. Among the data they published is a spreadsheet that records the format of thepresidential ballot (if it was a “caterpillar ballot”) along with the votingtechnology used.Demographics. The 2000 Census continues to be released. Basic demographic data foreach county in Florida can be found on the census web site at www.census.gov.Bibliographic sourcesCaltech/MIT Voting Technology Project. “A Preliminary Assessment of the Reliabilityof Existing Voting Equipment.” Also Voting: What Is; What Could Be. Tworeports by the Caltech-MIT Voting Technology Project about the performance ofvoting technology nationwide. Among other things, the report introduce a way ofthinking about the problem of voting technology performance empirically and away of measuring performance.Greg Adams and Chris Fastnow, “A Note on the Voting Irregularities in Palm BeachCounty, FL.” url: http://madison.hss.cmu.edu/. This web site contains pointers toa host of papers that analyzed voting irregularities in Florida.Comments/hints. This is a typical quick-and-dirty analysis that political scientists are called on todo all the time. There has been some newspaper reportage on this subject, but very little socialscience. An important think to keep in mind is that voting machines are bought by local electionofficials and the ballots are designed by them. Local election officials vary considerably in howgood they are. How good they are may depend on how well-endowed their counties are. And,how good they are may affect other (unmeasured) administrative practices in a county that mightaffect the “residual vote rate.” Therefore, getting a clean estimate of the independent effects ofmachines and ballot design may depend on getting good variables to control for the quality oflocal election administration.4Project II: Improving Voter Registration PracticesGroup: Horst, Larson, TahkBackground. To vote in the United States you must be registered. Unlike most democraticnations, in the U.S. registration happens in the states, not at the national level, and each state isdifferent. Considerable research has shown that various requirements of state voter registrationlaws serve as barriers to people voting, and that these registration barriers account for asignificant amount of the reason why Americans vote at lower levels than citizens in