<%BANNER%>

Using geographic decision information systems to support political endeavors at the University of Florida

University of Florida Institutional Repository

PAGE 1

USING GEOGRAPHIC DECISION INFORMATION SYSTEMS TO SUPPORT POLITICAL ENDEAVORS AT THE UNIVERSITY OF FLORIDA By NOELLE A. MECOLI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Noelle Mecoli

PAGE 3

This document is dedicated to Cephid.

PAGE 4

TABLE OF CONTENTS page LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................viii CHAPTER 1 INTRODUCTION........................................................................................................1 2 LITERATURE REVIEW.............................................................................................6 Beginning of Geodemographics...................................................................................6 Use of Geographic Information Systems......................................................................8 Geodemographics in Direct Marketing........................................................................9 Student Voters............................................................................................................12 Party Affiliation..........................................................................................................16 3 PILOT STUDY...........................................................................................................19 Background.................................................................................................................19 Data and Methods.......................................................................................................22 Results and Discussion...............................................................................................27 Pilot study............................................................................................................27 Geographic Patterns.............................................................................................30 Targeting Students and Their Parents With LSPs......................................................42 4 STUDENT VOTERS..................................................................................................45 Background.................................................................................................................45 Research......................................................................................................................48 Data.............................................................................................................................50 Methods......................................................................................................................51 Results and Discussion...............................................................................................54 Students Party Affiliation and Parents LSP......................................................58 Cross tabulations..........................................................................................58 T-Tests..........................................................................................................59 Typical Party Demographic.................................................................................59 iv

PAGE 5

Conclusions.................................................................................................................60 5 CONCLUSIONS....65 APPENDIX A CACI ACORN LIFESTYLE SEGMENTATION......................................................68 B CACI REPORT........................................................................................................109 LIST OF REFERENCES.................................................................................................111 BIOGRAPHICAL SKETCH...........................................................................................113 v

PAGE 6

LIST OF TABLES Table page 3-1. Number of studnets per District.................................................................................28 3-2. Student per LSP category.........................................................................................31 3-3. Students by County..................................................................................................43 4-1. Student registered voter demographics......................................................................55 4-3. Cross tabulations between party affiliation and LSP...............................................61 4-5. T-Test Republican and Democrat............................................................................62 4-6. T-Test Republican and Independent..........................................................................62 4-7. T-Test Democrat and Independent...........................................................................62 vi

PAGE 7

LIST OF FIGURES Figure page 2-1. Campaign for young voters finding young voters......................................................15 3-1 Florida Legislative District boundaries House Plan H406H020..............................25 3-2 District boundaries with students permanent addresses.........................................26 3-3. Database numbers for students and legislative districts............................................30 3-4. LSP Types 1A and 1B..............................................................................................32 3-5. Thriving Immigrants and East Coast Immigrants....................................................33 3-6. Young Immigrant Families......................................................................................34 3-7. Working Class Families...........................................................................................35 3-8. Twentysomething.....................................................................................................36 3-9. Wealthiest Seniors....................................................................................................37 3-10. African American Populations.................................................................................38 3-11. Hispanic and Asian Population................................................................................39 3-12. Age 18-29 Population...............................................................................................40 3-13. Age 65 and Up Population.......................................................................................41 3-14. Database numbers for permanent and student addresses.........................................43 4-2. 18 to 29 registered voters per House district............................................................63 4-3. Numbers for student registered voters.......................................................................64 vii

PAGE 8

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts USING GEOGRAPHIC DECISION INFORMATION SYSTEMS TO SUPPORT POLITICAL ENDEAVORS AT THE UNIVERSITY OF FLORIDA By Noelle A. Mecoli May 2004 Chair: Grant Ian Thrall Major Department: Geography The Registrars Office of the University of Florida (on behalf of the University Presidents Office) commissioned a proof-of-concept study to examine the political implications of funding for this public university, and to establish a method for reaching those significantly affected by state appropriations to the University: the parents of the students. Using Geographic Information Systems, a pilot study was conducted to assign the correct Florida State House Representative district to a students data record. The students declared permanent address is used to make the geographic assignment. In most instances, University of Florida students permanent address is their parents address. This allows enumeration of students permanent addresses by representative district, and the creation of databases for mailings to inform university constituents about legislative activities. Further segmentation of mailings is possible by differentiating between how the message would be presented based on measurements of the viii

PAGE 9

constituents geodemographic. This information is useful documentation to State Representatives as a measure of the importance of the University to their legislative district. This information is also useful for garnering support for higher education in Florida in general, and specifically higher education at the University of Florida. The study was extended to analyze student voting propensity. Voting propensity is frequency of voting and party affiliation. Voter registration data were compared and contrasted with students geodemographics (permanent address) and nationwide patterns of student voters. A procedure was thereby established for segmenting the population by propensity to vote, and by political affiliation. In combination with the students geodemographic, predictions of likely voting behavior can be made using this procedure. ix

PAGE 10

CHAPTER 1 INTRODUCTION The study was based on a pilot study authorized by the Presidents Office of the University of Florida (UF) to assign a Florida State House representative to a students permanent address within UFs current student database. The University would like to have the ability to carry out a letter campaign to the students parents when an issue concerning the university arose (such as funding for the University), and inform those parents to contact their legislator or vote on that issue. This pilot study will provide a method for the University of Florida administration to document the contribution of a state university to each individual state legislator in addition to mustering political support from those affected by state appropriations to universities: University of Florida students and the parents of students attending the university. To achieve this objective, this pilot study assigns to a students data record, the correct Florida State House Representative district. The students declared permanent address is used to make the geographic assignment. In most instances, University of Florida students permanent address is their parents address. This allows enumeration of students permanent addresses by representative district, and the creation of databases for mailings to inform university constituents about legislative activities. Further segmentation of mailings is possible by differentiating how the message would be presented based on measurements of the constituents geodemographic or lifestyle segmentation profiles (LSPs). 1

PAGE 11

2 The procedures adopted to execute this pilot study are those from the large field of Geographic Information Systems (GIS). GIS technology has been widely adopted across the university, from urban planning, to forestry, to natural resources analysis and business geography. This study applies GIS procedures and technologies from business geography. Among the various types of calculations performed by business geographers is deriving of the subject populations geodemographic. This is normally executed in the context of evaluating a retail stores customer profile. In this pilot study, the Universitys customers are its students and their parents. So, central to this study is the analysis of students geodemographics attending University of Florida. Jon Goss (1995b page 24) gave a comprehensive definition for geodemographics: Geodemographic works by collecting spatially referenced data on consumers, constructing statistical models of identity, and mapping distributions of consumer characteristics or types. A geodemographic system combines three essential components: massive electronic data bases composed of public and private, individual and aggregate records on consumer identity and behavior; Geographic Information Systems (GIS) that provide the tools to analyze, locate, and graphically represent the spatial distribution of consumer characteristics; and segmentation schemes that identify consumer types through factor and cluster analysis of spatially referenced demographic and psychographic data. The basic principle underlying geodemographics can be summed in the phrase birds of a feather flock together or you are where you live. People with similar characteristics or backgrounds tend to gravitate to the same neighborhoods. The term geodemographic is also referred to as lifestyle segmentation profile (LSP), clusters.

PAGE 12

3 or GD. Assigning a geodemographic to a students permanent address may give an indication of the type of lifestyle category the parents is in. Instead of sending letters to over 36,000 addresses (number of students who have a permanent address based in Florida), the University can cater letters depending on the geodemographic of the permanent address. Geodemographics are used in political campaigns as a way of soliciting voters so perhaps the same principle can be applied to this project. Weiss (1988) said knowing a neighborhoods political attitudes allows strategists to craft appeals using different voices to reach the distinct clusters and political surveys confirm that people living in the same cluster make similar choices. This study also reviewed the geographic patterns of the geodemographic of the permanent address in Florida noting the differences and similarities of the different cluster types Geodemographic clustering was originally designed to support firms in the direct marketing industry (Curry 1993) Geodemographics is primarily used to uncover and exploit new markets (Goss 1995), but it is also used for a wide range of application such as fund raising, mail campaigns, and site location analysis. Those who have taken advantage of using geodemographics find that it has the potential to eliminate redundancy, generate savings, and reduce unsolicited promotional communications (Goss 1995). Geodemographic systems were provided by a number of companies such as ACORN, by CACI Inc.; Cluster PLUS by Donnelley Marketing Information Services; PRIZM, by Claritas Corporation; and MicroVision by Equifax. This study also used the voter registration information from Alachua County Florida. The University of Florida is located in Alachua County, and there is no information regarding the number of registered student voters. If the number of

PAGE 13

4 registered student voters is obtained, the university may find this data useful in their letter campaigns. There are a number of students who have both their student address and their permanent address located in Alachua County, so the university can also solicit their students as a means of political support. The voter registration data contains information on party identification. This may prove to be useful in catering letters for increased political support. Another aspect of this study was a brief statistical analysis of the number of students who are registered to vote (Alachua County) yet have a different student address than permanent address (i.e., student address in Gainesville, permanent address in Ft. Lauderdale). This was be included to ascertain if there is a relationship between the students party identification and the parents (permanent address) geodemographic, and to establish a method for future research in this area given a greater population. A comparison of University of Florida registered student voters and research conducted by The Campaign for Young Voters ( www.campaginyoungvotesr.org 2003) was also examined to show any similarities within the two studies Chapter 2 presents a review of literature relating to geodemographics, GIS used in geodemographics, direct marketing campaigns that geodemographics provides benefits to, and finally a brief look at information regarding registered student voters. Chapter 3 presents and discusses the data, methods, and results from the project for the University of Florida. Chapter 4 presents and discusses the data, methods, and results from incorporating the voter registration data. Finally, Chapter 5 summarizes the finding and draws final conclusions.

PAGE 14

5 Chapter 3 is based on an article published in Geospatial Solutions (July 2003) titled Spatial Analysis, Political Support, and Higher Education Funding. Additional information was then added to this article for the purposes of this thesis for the UF graduate school and also as a report for the Presidents Office of the University of Florida. Chapter 4 is a draft for a future article to be published in Geospatial Solutions.

PAGE 15

CHAPTER 2 LITERATURE REVIEW Beginning of Geodemographics The term geodemographics came into light around the 1970s credited to Jonathan Robbin, a social scientist who brought together computer power with marketing methodology (Weiss 1988). He believed that consumers could be marketed directly by grouping them in neighborhood-sized markets. For geodemographics to be complied, two vital sources of information had to come to fruition: the ZIP (Zone Improvement Plan) Code, and computerized files from the United States Census. The ZIP Code allowed households to be divided into areas served by a single post office. Each zip had a numbering system that broke up the divided areas into about 800 households (Phillips & Curry 2002). In addition to the ZIP Code, the U.S. Census contained a wealth of geographic information. The use of the GBF-DIME (now the Tiger Line files) and the Census Block allowed for computerized maps to be produced (Phillips & Curry 2002). Merging the ZIP Code files with the block values from the Census allowed geographic coordinates for almost every mailing address in the United States to be determined. After the accomplishments that these resources could provide, companies such as Claritas Corporation and CACI Marketing Systems put the computerized maps (that contained a great deal of information) to use. But in the 1970s slow mainframe computers were used to process information. In the 1990s, with the advent of greater, 6

PAGE 16

7 cheaper, and faster computing power, the companies who produced geodemographics could now offer their services to a larger number of potential clients. These companies used public data (U.S. Census) with private data sources such as credit-reporting bureaus and intercensus-year data updates (Thrall, 2001) to provide geodemographic analysis to a number of companies in need of improving their target marketing strategies and for site-location analysis. The companies who provide geodemographics start with U.S. census data. The Census Bureau does not report the data collected per household, but rather by the smallest unit, the census-block group. 1 These units are analyzed for different characteristics, and people living in similar block groups are categorized together. Next is the statistical analysis. In this process, the redundancy is removed, and then the different groups are clustered together. The results divide the 250,000-plus census-block groups into a smaller number of geodemographic clusters (Curry 1993). The company used to provide geodemographics in this research is CACI Marketing Systems. Their system is called ACORN, an acronym for A Classification of Residential Neighborhoods. The company was founded in 1962, and they have since sold their ACORN system to Environmental Systems Research Institute (ESRI) ( www.esri.com 2003), in January 2002, yet for now retain the name. There are 43 segments and a descriptive name assigned to each segment (Appendix A). There are supergroups and then subgroups. The supergroup describes a broad range of characteristics shared by all 1 1 Census blocks are areas bounded on all sides by visible features, such as streets, roads, streams, and railroad tracks, and by invisible boundaries, such as city, town, township, and county limits, property lines, and short, imaginary extensions of streets and roads. Generally, census blocks are small in area; for example, a block bounded by city streets. However, census blocks in remote areas may be large and irregular and contain many square miles.

PAGE 17

8 the subgroups in a particular category that are then separated by specific characteristics (Curry 1993). Use of Geographic Information Systems Geographic Information Systems (GIS) is a collection of computer hardware, software, geographic data, and personnel designed to efficiently capture, store, update, manipulate analyze, and display all forms of geographically referenced information (ESRI, 2003). GIS is fundamental the development and applications of geodemographics. Approximately 85% of corporate databases have geographic content (Goss 1995). This, in turn, could lead to widespread use of GIS in many applications, particularly in marketing and marketers have, in turn, taken advantage of the benefits this tool provides. Using GIS makes geodemographics easier to analyze. GIS provides the spatial analytical tools used to aggregate and assign spatially referenced consumer information and to represent the results of analysis in visual form (Goss 1995). When GIS is used in developing and analyzing geodemographics, it is called Geographic Decision Information Systems (GDIS). Six benefits in using GDIS are discussed in a chapter of Ground Truth written by Jon Goss (1995). The first benefit Goss describes is accurately locating an address. This is accomplished with geocoding. Geocoding of a customer address is the first step in determining where a particular address is located in a neighborhood. To geocode an address simply means assigning a latitude and longitude coordinate to an address that then could be shown spatially. This process matches records in two databases: an address database and another database with know geographic positions such as a street map. After the databases are matched the coordinates are assigned to each record.

PAGE 18

9 Other benefits include the aggregation of individual records into spatial units like zip codes. This allows for the creation of new units such as trade areas, provides for the reconciliation of data collected at differing scales, examine the geographical ties between disparate data sets, as well as visual representation of data. Another key component of this research and in compiling geodemographics is relational databases. A relational database is the joining of several databases based on a particular attribute. This makes the analysis of many data sources at one time simpler. For this particular research, the student database is combined with a voter registration database based on a name attribute and then an address attribute. The databases and cluster analysis that generate geodemographics can be used by marketers independently of a Geographic Information Systems (GIS), but the ability to use spatial analysis within the database can increase the power and representation of the market analysis (Goss 1995). Other benefits provided by Baker & Baker (1993) in Market Mapping by using GIS is understanding spatial relationships between data are locating customers quickly, the ability for maps to make statistical marketing data easier to understand, and the ability to plan and track marketing campaigns is improved upon. Using GIS functions such as geocoding, joining databases, and mapping, increase the functionality of analyzing geodemographics. Geodemographics in Direct Marketing Geodemographics was developed with the intention of expanding a marketers possibilities in the area of direct marketing. With direct marketing, a market analyst could find new customers and keep remaining customers. Direct marketing can also locate new store locations and discover which locations were not doing well, as well as discover trade areas. Direct marketing can be less expensive if the marketer only targets

PAGE 19

10 what is needed and can influence the right segment of customers (Baker & Baker 1993). For this research, the direct marketing approach of mail campaigns will be the focus. This can be successful especially in the political arena. In a chapter of Marketing The New Marketing by Jon Goss (1995), he explains that using direct marketing, the marketer knows the characteristics and addresses then matches these with clusters to identify potential customers to increase sales (market penetration). From this, custom mailing lists can be complied to keep those customers. Goss also illustrates that marketing depends upon defining difference between products that resonate with the differences between consumers. Effective target marketing has only been limited by the development of the quality and quantity of information on consumers and the means in which to analyze it. This is why geodemographics can play an important role in developing a profitable market strategy. In The New Marketing Research Systems by Michael J. Curry (1993), he explains how the beginnings of geodemographics in mail campaigns could be successful. Mail order giant L.L. Bean used geodemographics to send their catalogs to a certain base of customers. By developing specific clusters based on those who responded to their catalog, a mailing list was created to target these groups of people. This proved to be cost effective and increase their sales by targeting a certain group of consumers. In Market Mapping by Baker & Baker (1993), a scenario is given for steps in targeting a direct mail program. In this scenario, letters need to be catered to target a direct mail program for a department store. A mail list is created to limit waste and reduce costs, and only target specific groups who would more likely be potential customers. The steps in developing this list include defining a need, acquiring the

PAGE 20

11 specific data, perform the relevant applications in a GIS, define relationships within the data, and finally make decisions based on the finding. This approach can be used in a wide variety of applications to develop a mailing list depending on the needs of the application. Using direct marketing becomes less expensive because the shotgun approach common in mailing campaigns is eliminated. Using geodemographics in mailing campaigns can be successful by influencing the right segment of customers. Political campaigns also use geodemographics as a means for targeting potential voters through direct mail. In Clustering of America by Michael J. Weiss (1988), a chapter is devoted to the use of geodemographics in political campaigns. Weiss states that clusters offered a way to reach nontraditional voters. Marketing issues to neighborhoods rather than groups as a whole could possibly generate increased support. Weiss also states that knowing a neighborhoods political attitudes allows strategists to craft appeals using different voices to reach the distinct clusters. A successful use of a highly targeted direct mail campaign using clusters is discussed in the book. In 1983, the governor of Louisiana was losing support and his political advisors sought the use of geodemographics provided by Claritas. They discovered three distinct groups the governors campaign could target through a mailing campaign. Writing one letter (message) and sending to all three groups would not muster the support that they needed to win the campaign. Three different letters were written in such a manor as to appeal to each cluster. In the end the governor won the election with 63% of the vote. These letters targeted a message in such a way as to fit the perspective of the clusters and thus gain a winning edge.

PAGE 21

12 Weiss discusses how neighborhood lifestyles are a natural way for predicting political behavior. Since the clusters are based on multiple socioeconomic indicators, they can also mirror voters. Weiss states that these clusters can explain why some issues cross party lines, but these same issues remain true to the neighborhood lifestyle. He also claims that geodemographics may prove to be practical rather then traditional voting blocks. Two other cases of geodemographics used in direct mailing campaigns are sited in Clustering of America. When geodemographics first entered the political arena in 1978 in the state of Missouri, a random sample of the population was surveyed to determine the clusters. The addresses were then linked to census block groups and the voters were identified. Instead of using a mass media campaign the strategists opted for a mail campaign as not to disturb the voters who were already on board. Different letters were crafted for each of the clusters found. While still expensive, this proved to be a cost effective method and generated the support that was needed. The other case involved targeting clusters for the renovation of the Statue of Liberty. Letters were written to cater to clusters such as first generation Americans where the Statue of Liberty is a symbol of their new freedom and independence. A letter had to be crafted differently however to those clusters where nationalism was strong and placed blame on immigrants for taking jobs away. Student Voters Young voters are the largest block of unclaimed voters in the country and candidates usually ignore this group due to their low voter turnout (Skaggs & Anthony 2002). Traditionally, it has been difficult to target this group of voters especially on the national level, but targeting this group on the state or local level may prove to be easier.

PAGE 22

13 Targeting this group can give opportunities to candidates in close races to gain a winning edge and to build a base among young voters for future elections (Skaggs & Anthony 2002). An article from Alternatives (1997), a newsletter from the Center for Policy Alternatives, suggests that businesses and market analysts are slowly coming to the realization that young consumers are a large group that needed to be mass targeted if they are to become future consumers of certain products. They propose that the politicians due the same to appeal to the young voters so they can become loyal supporters in the future. The Campaign for Young Voter (CYV) (2003) is a project of the Center for Democracy and Citizenship at the Council for Excellence in Government funded by a grant from The Pew Charitable Trusts. They are a non-partisan, non-profit organization. The Campaign for Young Voters (CYV), formerly known as the Young Voter Initiative, assists candidates for public office in their efforts to reach out and engage younger voters. Drawing on extensive field research concerning young adults' views about politics, elections and government, CYV publishes a Toolkit and suggested campaign practices and materials to assist candidates at all levels in dealing with young adults about political participation and voting. The CYV provides a wealth of information on polls, studies, and news links as well as links to data on voting and registration provided by the U.S. census (www.census.gov, 2003) and the Federal Election Commission ( www.fec.gov 2003). The information found on these web pages will be used throughout this research for comparison purposes to the data complied about University of Florida students and voter registration as well as general information about the overall data on young voters.

PAGE 23

14 Geodemographics are not widely used in the political community as a means of generating support (Weiss 1988). General demographics are more often used to target certain communities where support is known to be. The CYV complied there own demographic on young voters categorizing them as likely young voters, potential young voter, and unlikely young voters (Figure 2-1). The CYV has also complied data on likely demographics by party affiliation. They have found that young adults identify themselves almost equally as Republicans, Democrats, or Independents/third party. In comparing Republican vs. Democrats, they have found that they are divided almost equally as the general registered voting population is. The demographics by party are shown in Figure 2-2, and this information will be used to show any comparisons between the UF student registered voter populations. Research from various sources including the U.S. Census and the Federal Election Commission was complied by CIRCLE (The Center for Information and Research on Civic Learning and Engagement) to show facts on young voters on the national level. CIRCLE promotes research on the civic and political engagement of Americans between the ages of 15 and 25. In comparing registered voter information on the state level (Florida) with the national information, Florida has 74.3% (U.S. Census) of eligible citizens to vote registered to vote which follows closely to the national level of 78.1%. The number of 18-25 year olds registered to vote on the national level is 59.9%, but in Florida this number is lower at 39.9%. Males registered to vote on the national level between ages 18-25 is 40.3% and in Florida the percent is about the same at 40.0% while the female percents show a difference at the national level 44.3% and in Florida at 39.9%.

PAGE 24

15 Figure 2-1. Campaign for young voters finding young voters This group is the obvious place to start in developing a base of support among young voters. They typically are already more politically active and have stronger partisan leanings. They are presidential election voters, but in other elections their participation drops dramatically. If you can connect with them, they will turn out for you on Election Day, and some may volunteer on your campaign. They are likely to be knowledgeable about issues facing their communities, and have the motivation and time to work on them. Attitudes: Likely Young Voters are more politically oriented than their peers and trust government far more. They expect candidates to pay attention to their concerns and to want their vote. They view elections as the way for people to have a voice in their communities. They understand the power of votingthey are almost all registered to vote, say voting is important, and think their vote counts as much as anyone else's. They also see themselves as able to make a difference in solving problems in tcommunities. Demographics: Likely Young Voters have parents who took them to the polls as children, who voted themselves, and who discussed political issues at home far more than average. These young adults are also more educated, and the more education they have, the more likely they are to vote. They are likely to be full time students and to identify with Democrats or Republicans more than with third parties. Many attend religious services regularly, and many of those identify as born-again Christians. heir This group is essentially unclaimed and available. In many ways, Potential Young Voters are similar to Likely Young Voters, except they are typically community volunteers rather than political activists. They are not likely to be campaign workers for you, but will register and turn out if you and your party work to reach and persuade them. Their hot button is community involvementand the more you can make the case that you are a community problem solver and have been a community volunteer, the better chance you have of getting their votes. They are much more independent and less partisan than Likely Young Voters, making them a harder sell, but still workable. Attitudes: Potential Young Voters are frequent volunteers in their community. They recognize that elections are a way for people to have a voice, but are not likely to be interested in volunteering for a political campaign. While they are as trusting of government and candidates as likely voters, they are more ambivalent about voting. These potential voters are slightly more likely than average to believe that voting is important, to be registered to vote, and to believe their vote counts. They also have lower expectations about receiving attention Demographics: Compared to Likely Young Voters, Potential Young Voters are far less likely to come from politically oriented families, especially measured by their discussion of political issues. Somewhat higher percentages of these voters are women and attend religious services regularly. from candidates. This group is hard to reach, and then only with extraordinary time and resources. Polar opposites of Likely and Potential Voters, these young adults are difficult to engage in politics or voting. They are distrustful when candidates do reach out to them, and tend to come from families with little political orientation or involvement. Attitudes: Unlikely voters are disengaged from and distrustful of the electoral process. They are often not registered to vote and don't believe voting is important. A large majority of this group never volunteers, and is unlikely to believe they can make a difference in solving their community's problems. They're distrustfulof government, of other people, of candidates' motives, and of the importance of their vote. They feel candidates do not pay attention to them and do not want their vote Demographics: Unlikely voters come from families whose parents didn't vote, didn't take them to vote, and didn't discuss politics with them. They are also much less likely to have a significant affiliation with an institutiona college, a church, or a political party. This group is generally politically independent and non-ideological, and less likely to be conservative. s.

PAGE 25

16 Figure 2-2. Likely demographics by party for young adults Young Adult Republicans Younger: 18-22 years old, especially men Full time students especially non-working students Self-described conservatives White men Attend religious services regularly From politically-oriented families Concerned with higher education costs Occasional volunteers Young Adult Democrats Older 23-25 years old College degree or beyond Work part time Self-described liberals African-American or Hispanic, especially African American women and college graduates Examining the various methods in which geodemographics, GIS, and information on young voters are utilized can prove to be beneficial for this research. The university can take this information and apply it to their direct mail campaign to generate support for the university on numerous issues. Party Affiliation In this research, a correlation may be established between students party affiliation to parents (students permanent address) geodemographic. The geodemographics used for this study currently do not make political assumptions. Instead, assumptions were made about the different geodemographic types and what party the group might be more statically associated with based on several studies. Also recall the political assumptions that can be drawn from geodemographics based on Michael J. Weisss work in Clustering of America (1988). Statistical comparisons shall be made in this study to examine the students party identification and the permanent address geodemographic of that student. The

PAGE 26

17 assumption is made that a students party identification shall reveal a correlation to a particular geodemographic type based on a students party affiliation corresponding to that of the parents. Several studies have examined this issue. In The American Voter (1960), it is believed that party affiliation is obtained early in adolescence and remains stable throughout the life of the child. Another study suggest that partisanship does not change significantly after adolescence and party affiliation is nearly complete after leaving the parental household (Hess and Torney, 1967). In more recent studies conducted on the subject it was found that party affiliation remained stable over time and variation in party was small on a year to year basis (Green and Palmquist, 1994). Currently no party identification information exists with the geodemographics from CACI. Only assumptions can be generated on which party would most likely fit to a particular geodemographic. Since geodemographics are based on socio-economic indicators, it may be assumed that a party can be assigned to a geodemographic based on studies conducted by demographic groups (Weiss, 1988). The study of party identification and groups is dominated by Robert Axelrod. In a quadrennial report in the American Political Science Review titled Where the Vote Comes From, Axelrod examines party affiliation by several bivariate factors such as race and religion. More extensive multivariate studies have been conducted to determine party support by differing groups. One such multivariate study by Erikson, Lancaster & Romero (1989), examines this type of analysis of group influences on party identification in presidential voting. Among their results they find that Urban voters show pro-Republican influence although large urban population tend to vote for Democrats and for youth to vote Democratic and college educated to vote more Republican are muted. They

PAGE 27

18 also find that region plays a significant role in party identification. Another study examines group support for political parties by multivariate techniques. This study titled Partisanship and Group Support Over Time: A Multivariate Analysis (1986) finds groups that tend to support the Democratic Party and those who have moved away from the party.

PAGE 28

CHAPTER 3 PILOT STUDY Background Public education in the United States is highly political. It is political because expenditures on higher education comprise one of the largest categories of discretionary expenditures that the state has to choose how to spend tax dollars. Most expenditures made by a State government are comprised of entitlements, the amount of which are difficult, if not impossible, to change from year to year. State legislatures choose the amount of funds that go into higher education, versus say kindergarten through 12th grade education, community college education, special education programs, prison systems, and a host of special projects. When state tax revenue declines, State Representatives choose which discretionary programs will have a reduction in revenue. When state tax revenue increases, State Representatives choose which discretionary programs will be beneficiaries. Of course, higher education in general, and University of Florida in particular, lobby to minimize reductions, and increase appropriations. Numbers support lobbying efforts. It is believed that the greater is University of Florida constituency within a Representatives district, the more attention will be paid to that constituency. That attention hopefully translates into revenue to the University. Most colleges and universities in the State University System are commuter campuses. Students commute from their permanent address to the campus. A State Representative has a fairly accurate perception of the importance of the campus to his or her district. University of Florida 19

PAGE 29

20 however is instead a residence campus. Students for the large part come from other places reside on or near campus while attending classes, and after they complete their education at University of Florida they leave the campus vicinity. The residence campus provides an opportunity for a student to become completely immersed in higher education, as well as an intermediate step toward independence from parents. However, this same remoteness comes at a price. Because the campus is remote from most urban areas in the State, and therefore remote from most State Representatives districts, it is easy for the State Representative to not consider the importance the University of Florida is to their district. The numbers that this research generates are used to document the importance of University of Floridas campus in Gainesville to Representative districts throughout the state. Geographic Information Systems technology allows for more than the placement of pins on a map representing the locations of students permanent addresses, and more than the enumeration of counts of those pins by Representative district. GIS allows for the evaluation of which types of people live where. This is done by way of assigning lifestyle segmentation profiles (LSPs) to those individuals who comprise our database. Instead of sending over 36,000 addresses (number of students who have a permanent address based in Florida, both undergraduate and graduate students), with a single message, the University can send letters with a message targeted to specific geodemographic characteristics, including legislative district and LSP. Geographers are familiar with applying this same technology to political campaigns as a means to sway a voting population. Weiss (1988, p.24) writes in The Clustering Of America, that knowing a neighborhoods political attitudes allows strategists to craft appeals using

PAGE 30

21 different voices to reach the distinct clusters, and political surveys confirm that people living in the same cluster make similar choices. The basic principle that underlies the creation and use of an LSP is Toblers first law of geography: all things are related, but near things are more related than distant things. That applies to characteristics of people. People with similar demographic and consumption characteristics, generally reside in close vicinity to one another. LSPs are derived by analyzing hundreds of characteristics of neighborhoods and individuals, to create a single composite measurement that summarizes the population by location. Statistical methods used to derive LSPs include principal components analysis, correlation matrices, and a variety of graphical methods, as well as various cluster techniques. The derivation and use of LSPs in geography date to the pioneering research of Brian J. L. Barry in the 1960s (Thrall, 2002 p. 29). Today, an analyst need not calculate their own LSP database. Instead, LSPs are part of commercial GIS technology. Not only are commercial LSP databases a time saver for the analyst, but there are analytic and communicative advantages to using standardized LSP categories. Also, contemporary LSP databases are improved by the use of proprietary databases not available to most analysts, such as individual expenditure and credit information. If the business geographer knows an address, that address will hopefully have a ZIP+4 code. The GIS technology can assign a ZIP+4 code to a data record with only a fragmentary address. That ZIP+4 code can be matched with one in a database of ZIP+4 codes and LSPs. Using relational database management, the appropriate LSP is then assigned to every address record in the database. At the same time, the address will be used in GIS address matching procedures to calculate a latitude-longitude coordinate.

PAGE 31

22 The smaller is the geographic area, the more likely will the LSP code accurately correlate with the individual household, and the more accurate will be the geographic positional assignment. Therefore, the database to be analyzed can be populated with LSP codes as well as geographic coordinates. SQL operations can then be used to select records by LSP, and those student permanent addresses can then be used for specific mailings. The pilot study data analysis begins with assigning to a students data record, the correct Florida State House Representative district. The students declared permanent address would be used to make the geographic assignment. In most instances, University of Florida students permanent address is their parents address. This allows enumeration of students permanent addresses by representative district, and the creation of databases for mailings to inform university constituents about legislative activities. Further segmentation of mailings can be made by differentiating between how the message would be presented based upon measurements of the constituents lifestyle segmentation profiles (LSPs). This information could then be used to document to State Representatives one very clear measure of the importance of the University to their legislative district, and an effective method garnering support for higher education in Florida in general, and specifically higher education at University of Florida. In addition to this pilot study, supplementary information was provided about the students geographical patterns based on their permanent address geodemographic and that data was compared to 2000 U.S. census data. Other mean to target parents and students based on their LSP is also provided. Data and Methods The University of Florida student database is a private database, provided to us by the Registrars Office of the University of Florida. It is not available to the general

PAGE 32

23 public. The student database for enrollment of fall 2002 included a variety of data fields, including permanent address. ESRI BIS (formerly CACI/Coder Plus using A Classification of Residential Neighborhoods (ACORN) Software) was chosen to assign ZIP+4 codes, geographic coordinates and LSPs to student records based upon permanent address. ESRIs BIS product assigns LSP codes (Appendix A) and geographic location coordinates to individual data records at the geographic scale of a ZIP+4 code, five digit ZIP code, or census tract. The company protects the process involved in providing geodemographics, but they do provide some general methodology. Over 220,000 neighborhoods are assigned to one of 43 market segments based on similar characteristics from census block groups such as household type and occupation. Other determinants are also analyzed including consumer behavior. CACI then uses statistical methods such as principal components analysis, correlation matrices, and graphical methods in addition to a variety of cluster techniques to create the segments. A K-means algorithm followed by Wards hierarchical minimum-variance method is used to group the clusters which are then validated by using external variables to check the market segments. The ACORN system is also linked to updated consumer survey data to ensure validity. In this proof of concept study, data records were assigned geographic coordinates using only the more accurate street address matching procedure. ESRI Business feature was used which allows assigning geographic coordinates based upon best of street address, ZIP+4, and ZIP code. The University Registrars office was concerned that an address might be assigned to the wrong legislative district if geographic positioning with less accuracy than street address matching was used. A student whose permanent address

PAGE 33

24 was outside the State of Florida was dropped from the database used for the analysis. Also, students permanent addresses that were not assigned latitude and longitude coordinates, such as those with post office boxes, were deleted from the database. The students were then mapped using ArcMap ArcGIS 8.2 by ESRI (www.esri.com). The Florida Department of State, Division of Elections (2003) provided contact information for each state legislator. Individual legislator information was updated to include the winning candidate in the 2002 general election. The State of Florida Legislative District boundaries House Plan H406H020 was approved in July 2002. That boundary file was obtained from the Legislature of the State of Florida. The boundary file was in ESRIs shapefile format, decimal degrees North American Datum 1983 (NAD 83). There are 120 House districts in the State of Florida. A map of the districts is shown in Figure 3-1. The boundary file included district number as an attribute. The legislator contact information database included the legislators name, title, district number, and legislative office address. The common district number allowed the contact information to be joined to the district polygon. Next, the appropriate legislative district and contact information was assigned to each student record using the point-in-polygon operation (Spatial Join). First the students permanent addresses are mapped, as shown in Figure 3-2. Figure 3-2 also includes the legislative district polygon layer. The point layer (student addresses) and polygon layers (legislative districts) can be spatially joined, with attributes of objects from one layer assigned to objects in the other layer.

PAGE 34

25Figure 3-1 Florida Legislative District boundaries House Plan H406H020

PAGE 35

26 26 Figure 3-2 District boundaries with students permanent addresses

PAGE 36

27 Maps were also included for comparison purposes using the students permanent address geodemographic and the 2000 U.S. Census data on the track level to show the general distribution of certain population in the State of Florida including that of African American, Hispanic and Asian, 18 to 29 years olds, and the 65-and-over age groups. The process used to discover the number of students who have the same student address and permanent address was to first copy the original file of students whose permanent address is located in Florida. These two files can then be merged on the student address and the students permanent address, only keeping the records that matched. The number records that were discarded are the students with differing addresses. This is information is useful as a means to target parents and students in differing ways; based on LSP geographic patterns and location of permanent address if it differed from the student address. Results and Discussion Pilot study There are a total of 46,923 in the student database, and 42,567 of the students list an address in Florida as their permanent address. There were 6373 or 15% that were deleted because they did not geocode to street address, leaving 36,194 students to be assigned to a legislative district (Figure 3-3). Table 3-1 shows the 120 legislative districts in the State of Florida, and the number of University of Florida students whose permanent address is in that legislative district. The districts with the highest numbers of students (permanent address) are District 23 (Gainesville/Ocala), 22(Gainesville/Ocala), 96(Ft. Lauderdale), 19(Jacksonville), 97(Ft. Lauderdale), and 37(Orlando).

PAGE 37

28 Table 3-1. Number of studnets per District District # of Students District # of Students District # of Students 1 128 41 345 81 257 2 132 42 130 82 300 3 176 43 160 83 414 4 297 44 124 84 96 5 60 45 192 85 375 6 165 46 109 86 138 7 61 47 405 87 303 8 53 48 500 88 128 9 437 49 126 89 83 10 35 50 277 90 257 11 389 51 271 91 283 12 71 52 223 92 115 13 260 53 122 93 158 14 119 54 227 94 135 15 89 55 58 95 258 16 303 56 221 96 658 17 255 57 308 97 627 18 392 58 85 98 416 19 632 59 42 99 211 20 225 60 352 100 322 21 103 61 153 101 241 22 5075 62 254 102 96 23 5358 63 186 103 153 24 425 64 77 104 135 25 153 65 126 105 137 26 300 66 125 106 197 27 114 67 179 107 122 28 223 68 192 108 149 29 156 69 261 109 61 30 322 70 277 110 105 31 356 71 103 111 135 32 239 72 97 112 140 33 321 73 149 113 97 34 294 74 222 114 340 35 311 75 154 115 385 36 114 76 243 116 221 37 598 77 105 117 321 38 193 78 248 118 100 39 105 79 125 119 244 40 316 80 238 120 135

PAGE 38

29 The Registrars Office of the University of Florida on behalf of the University Presidents Office commissioned this proof of concept study. The purpose of the study to provide a tool whereby the University could bring to the attention of State legislators the importance of the university in serving the educational needs of students of their district and state region. Public education in the United States is highly political, becauseit is a large part of a states discretionary budget, and many constituencies are vying for the same discretionary dollar. While the University registrar has not at this time applied the methods and procedures demonstrated in this work, the work serves as a demonstration of concept. Higher education is important to the success of the next generation to be edHigher education requires considerable and sustained funding if it is going to serve itsmission. G is ucated. IS has been demonstrated here as one means to generate the information necessary to receive the political,373 records not able to geocode, due to bad addresses or a Post Office Box, the Re correct legislative district. The accuracy in wh support to further the mission of higher education With 6 egistrars Office loses the ability to reach these students and/or their parents, but there is a considerable amount of records left to send letters to. A concern of the Registrars Office was how accurate is the spatial join and how would they know if thstudents permanent address was placed in the ich the software geocodes the address should place the address within the correct district, with only potential problems arising with those addresses that are on the border of two districts. A buffer could be included in the process of a desired distance to eliminate this problem, yet that would exclude a number of records.

PAGE 39

30 G eographic Patterns A LifeStyle Segmentation Profile was assigned to each student record for their perment addres out o44 pble LSPs, tudentsit int categorihe numr of studenr LSP shown Table 3atternsmerge when looking at the students permanent address just in the State of Florida per LSP type. Most of the LSPs are distributed rather evenly throughout the state, but significant geographic patterns can be s with specypes oLSPSP type 1p OneerceB Wealthy Seaboard Suburbs, are highly concentrated in South Florida with some in the m cities Florsuch as, Jacksonville, Orlando, and Tampa ure 3-4). e 1A an 1B he wealthiest of the LSP types. Their residential classification puts these types in the suburbs of maarehe other tconcenratedouth Flornd oth maimmant categTypeC Thng Immigrants, and 6A East Coast Immigrants are highly concentrated in South Florida with some found in other maities throut the state (Figure 3-5 ). CACI describes these types as living in urban areas with a large percent being of Hispanic origin. Another immi categoy, 8Aung ImmiFams, is almoely coentr in South da (Figre 3Figure 3-3. Database numbers for students and legislative districts an s and f ossi UF s f o 31 es. T be ts pe is n i -2. P e een ific t f s. L A To P nt, 1 ajor in ida (Fig Typ d are t jor urban as. T ypes t in S ida a er jor urban areas are the igr ories. 2 rivi jor c ghou grant r Yo grant ilie st sol nc ated Fi lor u -6).

PAGE 40

31 Table 3-2. Student per LSP category Acorn Description # % 1A Top One Percent 446 1.2 1C Upper Income Empty Nesters 1614 4.5 1E Prosperous Baby Boomers 4876 13.5 2A Urban Professional Couples 2428 6.7 2C Thriving Immigrants 275 0.8 3B Enterprising Young Singles 9534B 1B Wealthy Seaboard Suburbs 332 0.9 1D Successful Suburbanites 2353 6.5 1F Semirural Lifestyles 2187 6.0 2B Baby Boomers With Children 1160 3.2 2E Older, Settled Married Couples 552 1.5 2.6 4A Retirement Communities 840 2.3 Active Senior Singles 797 2.2 4C Prosperous Older Couples 1091 3.0 4F Senior Sun Seekers 1101 3.0 0.8 0.5 6C Newly Formed Households 843 2.3 7B Young Frequent Movers 1456 4.0 7E Small Town Working Families 130 0.4 1.0 8A Young Immigrant Families 744 2.1 8D Hard Times 133 0.4 9C Unpopulated Areas 13 0.0 4D Wealthiest Seniors 2618 7.2 5A Twentysomethings 983 2.75B College Campuses 5830 16.1 6A East Coast Immigrants 285 6B Working Class Families 176 7A Middle America 637 1.8 7C Rural Industrial Workers 87 0.27G Heartland Communities 346 8C Distressed Neighborhoods 24 0.1 8E Urban Working Families 284 0.8

PAGE 41

32 Figure 3-4. LSP Types 1A and 1B

PAGE 42

33Figure 3-5. Thriving Immigrant s a nd Eas t C oa st I mm igr ant s

PAGE 43

34Figure 3-6. Young Immigrant Families

PAGE 44

35Figure 3-7. Working Class Families

PAGE 45

36 36 Figure 3-8. Twentysomething

PAGE 46

37 Figure 3-9. Wealthiest Seniors

PAGE 47

38Figure 3-10. African American Populations

PAGE 48

39Figure 3-11. Hispanic and Asian Population

PAGE 49

40 40 Figure 3-12. Age 18-29 Population

PAGE 50

41 Figure 3-13. Age 65 and Up Population

PAGE 51

42 Targeting Students and Their Parents With LSPs Knowing how many students permanent addresses are in each district and learning the different LSP types that make up these addresses can lead the University to cater their direct mail campaign to whatever group of addresses that fit their needs. After merging the student address with the students permanent address (State of Florida only), it is discovered that 18,850 students have the same student address and permanent address. This leaves 17,344 records that have different addresses and in these cases, a letter would have a better chance to go to a parent or grandparent of the student if that is indeed the focus of the University. The LSPs of 17,344 records could be analyzed and then catered to depending on the issue the University is trying to raise (Figure 3-14). Another way to limit or cater the number of mailings is to target areas throughout the state with a high concentration of permanent addresses. According to the CACI Geographic Summary (Appendix B), the top 8 counties where students have their permanent addresses are located in the major cities of Florida. The South Florida counties (Palm Beach, Broward, and Dade) have the largest concentration of addresses next to Alachua, then Hillsborough, Pinellas, Orange, and Duval (Table 3-3, Figure 3-15) With the completion of this research, the University will also have the ability to target just the students who are registered to vote in Alachua County (Chapter 4). Since a large percentage of students put their permanent address in the county of Alachua (30%), and then further reducing this number by finding the students registered to vote, will permit the University to target students who may be more inclined to vote for issues of concern to the student and to UF in the districts that represent Alachua County. Recall, the largest number of students (permanent address) are located in Districts 23 and 22.

PAGE 52

43 Figure 3-14. Database numbers for permanent and student addresses Table 3-3. Students by County COUNTY COUNT% Alachua 12,424 26.1 Broward 3,752 00.8 Dade 3,057 6.5 Palm Beach 2,171 4.6 Hillsborough 1,899 4.0 Pinellas 1,811 3.9 Orange 1,612 3.4 Duval 1,530 3.3

PAGE 53

Figure 3-15. Florida counties and major cities

PAGE 54

CHAPTER 4 STUDENT VOTERS Background Young voters are the largest block of unclaimed voters in the country. Candidates for political office generally ignore young voters group due to their low voter turnout (Skaggs & Anthony 2002). It has been difficult to target this group of voters at the national level. It is believed that this group of voters is hard to target because they have a low voter turnout (Highton & Wolfinger, 1998). Targeting this group at the state and local levels is feasible. Targeting the young voter group can give opportunities to candidates in close races to gain a winning edge, can serve to build an ongoing consistency of young voters in future elections, and can contribute to building a base of voters that will stay with the candidate or political party as the young voters age (Skaggs & Anthony 2002). An article from Alternatives (1997), a newsletter from the Center for Policy Alternatives, presents a case that businesses and market analysts are slowly coming to the realization that young consumers are a large group that need to be mass targeted if they are to be relied upon to become future consumers of a businesss products. Alternatives proposes that politicians for the same reasons appeal to the young voters so that thoyoung voters become loyal supporters in the future. In practice today, geodemographics are not widely used in the political community for planning or strategy, including actively gaining political support (Weiss 1988). se 45

PAGE 55

46 The Campaign for Young Voter (CYV) (2003) is a project of the Center for Democracy and Citizenship at the Council for Excellence in Government funded by a grant from The Pew Charitable Trusts. Both The Pew Charitable Trusts and CYV are non-partisan, non-profit organizations. The Campaign for Young Voters (CYV), formerly known as the Young Voter Initiative, assists candidates for public office in their efforts to reach out and engage younger voters. Drawing on extensive field research concerning young adults' views about politics, elections and government, CYV publishes a Toolkit and suggested campaign practices and materials to assist candidates at all levels in dealing with young adults about political participation and voting. The CYV provides a wealth of information on polls, studies, and news links as well as links to data on voting and registration provided by the U.S. census (2003) and the Federal Election Commission (2003). The information found on these web pages will be drawn upon throughout this research. The information is valuable for comparison purposes with the data complied in this research about University of Florida students and voter registration in Alachua County, Florida. The information is valuable for establishing a benchmark about young voters in general. Research on the topic of young voters has been executed by the U.S. Census and the Federal Election Commission. The results of this research has been compiled and disseminated by CIRCLE (The Center for Information and Research on Civic Learning and Engagement). This research includes data on young voters at the national level. CIRCLE promotes research on the civic and political engagement of Americans between the ages of 15 and 25. A comparison of registered voter information at the state level

PAGE 56

47 (Florida) with the national information, reveals that Florida has 74.3% (U.S. Census) of citizens eligible to vote are registered to vote. This characteristic of Florida voters closely mirrors voting registration at the national level, which overall is 78.1%. A focus on the young voter population reveals a different trend between the State of Florida versus the national level for the same age group. At the national level, the 59.9% of 18-25 year olds registered to vote. In contrast, only 39.9% of eligible voters in the same age group registered in Florida. Male registration the national level for the 18-25 age group is 40.3%. Males in the same age group in Florida have the same propensity to vote as their counterparts at the national level, at 40.0%. However, young females at the national level have a higher propensity to vote at 44.3%, while in Florida have a slightly lower propensity to vote than their male counterparts at 39.9%. Propensity to vote among young eligible voters can be viewed in a wider context than those voters showing up at the voting booth on Election Day. Propensity to vote is also a measure of social commitment, social involvement, and political participation. These characteristics are important to more than just politicians running for office. They are important to any politicized organization where opinions of voters might sway. One mple of such a politicized organization is the state public university (Thrall and coli, 2003). What makes a state public university politicized? Admissions in the state premier ersities are much less than the number of qualified students that might desire to nd the university. Funding from the state, and student paid tuition, is not based upon sideration of expense of maintaining the university and providing a constant quality of exaMeunivattecon

PAGE 57

48 service. Instead, revenue to support the university is more often based upon political considerations. The most affected population groups are university students themselves. How may this affected population be drawn upon to become politically active in those issues that they have a self interest in supporting? The answer must take into consideration the characteristics of the affected population subgroup, and their propensity to be politically active, and for this political action to be manifested by voting behavior. The characteristics of the affected population subgroup can be evaluated using contemporary geographic information systems technology. The results of the study can be put to use to contact the affected population subgroup and thereby have an impact at the voting booth. Research The previous chapter concluded with how a university can be proactive in reaching out to its constituency its students so that they acting in their clear self interest knowledgeably act on that self interest at the polls and voting booth. To accomplish this, it was argued, the university must achieve a better understanding of who their students are, what motivates them to become politically active, and what are the odds of their students becoming sufficiently aware of how their self interest is affected by political decisions that those students become politically active in the cause of the university. Geographic information systems technology when properly applied can help to addreand answer these questions. The below demonstrates how geographic information systems technology, in particular that technology developed in support of business geography, can answer these questions and also become a valuable instrument for improved university administrative judgment. ss

PAGE 58

49 This demonstration draws upon the registered voter database for Alachua County, Florida, the home of University of Florida. It also uses the database of students enrolled at University of Florida. No other study has integrated university student enrollment data and voting registration data. Benefits of this study include documenting the number of students that register to vote in the county. The fewer the number, the less the local political importance. The greater the number, the larger is the importance of the student population to local politics. Because state representatives are elected by voters in their jurisdiction, the more important are the student voters to the overall voting population of the jurisdiction, the more politicians will consider the students preferences. The University can use this information in documentation to politicians whose legislative districts overlap where University of Florida students reside. The University might encourage its students who are registered to vote in to engage in a letter writing campaign to those legislators within whose districts they reside. What type of student is registering to vote, and voting? What are their geodemographic characteristics? The discovery of their geodemographic patterns might be useful to groups that organize voter registration drives, and useful to legislators that need to get the student vote to win election. But getting students to register to vote, and to vote, is of little practical value to a special interest group unless there is some knowledge about how those young voters will vote. So, this research also addresses of the methods and procedures for discovering the relationship between propensity to register and vote, and party affiliation. This research drills deeper into the geodemographic characteristics of students by evaluating the relationship between the lifestyle segmentation profiles of the parents of students that register to vote, and vote. In this way, because LSP of parents of students can easily be

PAGE 59

50 determined, the university can more effectively target its students who are susceptible to registering and voting. Data Two data bases were of primary importance to this study: voter registration for Alachua County, Florida, which includes name, address and voter affiliation; and, University of Florida student records, which includes local and permanent addresses for its students, race, grade level, and gender. This data is not available to the general public. This data is made available on a case by case basis, and then only if it is judged by the officials responsible for the data that individuals listed in the database will not be subjected to any inconvenience and harm, and if there are sufficient benefits that will arise from making the data available. It was easy to make the case to the University to make its student records available for this study because the University Registrar and the Office of the President requested that this study be undertaken as a proof of concept demonstration. That the university originated this study is evidence of its importance to the university, and evidence of how state universities are becoming proactive in this politicized environment to protect its student constituency. For nonpartisan reasons, the Alachua County Registrar of Voters was also interested in the results of this study. wing how many students register to vote, and actually vote, has long been of concern lachua County Registrar. One of the goals of the Registrar of Voters is to maximize voter registration and voter turn out. The appropriate approvals were obtained. The stered voter database for Alachua County was acquired in September 2002; the database contained a total of 124,515 names, party affiliation, and addresses. At the same ent database for Fall (September) 2002 was obtained from the versity of Florida Registrar. The student record database included students name, Knoin Athe regitime, the student enrollmUni

PAGE 60

51 local address, age, and permanent address. Recall from the previous chapter that 30% of students list their permanent address in Alachua County. Methods To join the voter database to the student database, a data field must exist between the two data files that contain the same information and that uniquely identifies each record (student, voter) in each database. Last, first, and middle names are included in each database. However, the name in each database might not be the same. The most common reason that correct records could not be joined was revealed to be different treatments of the last name. The university maintained full middle name, while the voter database often abbreviated the middle name. This problem was avoided by using only the first and middle name. But some students have the same first and last name. To insure that the correct records were joined, an address join was also performed. So, for records of the two databases to be joined, the first name, last name and address must be the same for each of the databases. Not only does the use of address assure that names are correctly matched between databases, but according to Florida State Election Laws, a person must be registered to a precinct in which they vote (Florida Division of Elections Election Code 101.045). So the students who have not changed their address with the Supervisor of Elections yet changed their address with the University would legally be ineligible to vote in their previous precinct if the current address falls outside. A summary of demographic information from the UF registered student voters is presented in Table 4-2. The demographic information was created by processing students permanent addresses through ESRI BIS Coder (formerly CACI Coder/Plus). The demographic characteristic of University of Florida students can be then compared to those published by the Campaign for Young Voters (CYV). CYV links students

PAGE 61

52 demographic characteristics with party affiliation, and other information obtained from surveys. Student voters with a different local and permanent address were identified. These data records were used to establish a possible relationship between the students party affiliation and demographic characteristic as revealed via their LSP obtained from their permanent address. A cross tabulation was created to discover possible relationships between party affiliation and LSP. SPSS (a statistical software package) was used to create the cross tabulation. Party affiliations were limited to the big three: Democrats, Republicans, and Independents. Only a trivial percentage of registered voters in Alachua County fall outside the three largest party affiliations. When addresses are processed through ESRI BIS Coder, ZIP+4s are assigned to each address, and then ACORN (a classification of residential neighborhoods) LSP measurements are assigned on the basis of ZIP+4. The product assigns both LSP description, and LSP index number (Appendix A). Various ACORN LSP index numbers were then agglomerated together, treating categories 1A, 1B and 1C as the same. Because the purpose of this part of the study was to predict which students were likely to be Republicans or Democrats, those categories higher propensities to be one party or the other were then agglomerated together. example, ACORN LSP categories 1 and 2 are both highly likely to be Republican, so were combined. At the same time, category 6 was combined with 7 because both gories have dominated by Democrats. These assumptions are based on surveys and ies (Erikson, Lancaster, &Romero, 1989), (Stanley & Niemi, 1994), but no research been conducted to indicate party type by LSP (from CACI). In Clustering of America ichael J. Weiss (1988), how neighborhood lifestyles (LSPs) are a natural way for withFortheycatestudhasby M

PAGE 62

53 predicting politica l behavior is discussed. Although no political informat ion is collected whenfestyle. He also claimH1: There is dependence between party affiliation and LSP In the above hypothesis test setup, if H1 is accepted, it can be reasonably be concluded that high correlation exists between the lower indexed LSP categories 1 and 2, and the Republican party, while the higher number LSP index categories 6 and 8 are highly correlated with the Democratic party. Independent voters are expected to fall in the middle to high LSP indexed categories. An independent sample t-Test was also performed to determine whether the mean LSP scores were different for each political party. The t-test requires that two groups be tested at a time. T-tests were calculated on the difference in LSP scores between Republicans and Democrats, Republicans and Independents, and Democrats and Independents. compiling LSPs from CACI, the clusters are on multiple socioeconomic indicators, which can mirror voters. Weiss states that these clusters can explain why some issues cross party lines, but these same issues remain true to the neighborhood li s that geodemographics may prove to be practical rather then traditional voting blocks. The Pearson chi-square coefficient was calculated using SPSS to test for dependence between these two categorical variables (party affiliation and LSP). The hypotheses for the cross tabulations are: H0: There is no dependence between party affiliation and LSP

PAGE 63

54 The hypotheses for the t-Test are: H0: The mean LSP type is the same for the categories party affiliation H1: The mean LSP type is different for the categories of party affiliation If H1 is accepted, and the means differ between parties and the p value (sig. 2-tailed) is not close to 1 then the assumption can be made that the two parties being tested most likely appear in different LSP categories. Results and Discussion After joining the registered voter data with the student data as described above, it was determined that 7,621 (Figure 4-1) students enrolled at University of Florida in Fall 2002 were also registered to vote at that same time in Alachua County. This is a very small population. The results on this small a population can not extend to the entire student body, but instead provide a method to examine larger populations. The registered student voters by race, grade level, gender, party affiliation is shown in Table 4-1 in ddition to the breakdown of these categories by party affiliation. The LSP types of se not shown thrity of students by student address are cized in type 5Asomngs or type 5B College Campuses. gure 4-ent registered voter numbers a tudents ar because e majo ategor Twenty ethi Fi 1. Stud and

PAGE 64

Total 7,621 Student Registered Voters 55Table 4-1. Student registered voter demographics PARAMETER COUNT % Undergrad 5261 69.0 Graduate 2365 31.0 Female 3885 51.0 Male 3324 43.6 Hispanic 602 07.9 Black 583 07.7 White 6005 78.8 Asian 326 04.3 Other 58 0.76 Democrat 3119 41.0 Republican 2101 27.6 Independent 1781 23.4 Other 619 08.1 Female Democrat 1769 45.5 Female Republican 947 24.4 Female Independent 899 23.1 Female Other 270 00.7 Male Democrat 1220 36.7 Male Republican 1083 32.6 Male Independent 785 23.6 Male Other 236 07.1

PAGE 65

56 A comparison of national level data from the Campaign for Young Voters (www.campaginforyoungvoters.com) to registered student voters at the University of Florida is valuable for both corroborative purposes and for determining if and how University of Florida students difnd. The sample from UF represents those registered to vote andl percentage of total UF students. The cats rats tend ublicans. In the national survey, whites (15-25 fer from the national tre is only a smal omparison reveals that full time students on the national level tend to be more Republican (38%) then Democrat (22%), but Independents were similar to the Democr(24%). However, at University of Florida, the majority of student voters are Democ(40%) then Republicans (27.5%) followed by Independents (23.4%). On the national level as a whole, it would seem that a majority of young voters would be Democrat, especially full time college students, but a parents political affiliation may have aninfluence with the way the student identifies to a particular party and votes (www.stateofthevote.org 2003). Although comparing results to the national level may not be accurate due to the small population being sampled (UF), these results allow general comparisons When comparing gender within the registered students voters at UF, females to be more Democrat while the males have similar percentages for Democrat and Republican. In comparing race, the black registered student voters are predominantly Democrat followed by Independents, Others, and finally Republicans. It is interesting to note that the Others category has a higher percentage than that of Republicans. Democrats rank highest with the Hispanic and White registered student voters, but differ when comparing Republicans and Independents. The Independents have a higher percentage among Hispanics than do Rep

PAGE 66

57 yearst nt ts t records, leaving the possibility of survey error. At the same time, the Univenty the to be a viable voting block, yet college aged students are the fasted growi t old) have a higher percentage of Republicans than Democrats, but the highespercentage lies with the Independents. The Black young adult percentages are consistewith that UF, but the Hispanic young adults have the highest percentage as Independenfollowed by Democrats. The differences between the national study and the University of Florida study may arise from data collection procedures. The national study interviewed college and non-college students and at a much younger age (15-25). Influences of the parents (Beck & Jennings, 1991; Sears & Valentino, 1997) have a considerable impact on how theserespondents answer to the survey questions by the CYV. There was no reliable source ofinformation on just college student registered voter surveys that could be used in comparison. The CYV did not rigorously compare voter registration records with student enrollmen rsity of Florida study evaluated only students registered to vote in Alachua Coubecause only Alachua County voter registration records were available. This might createa bias, but in an unknown direction, because some students might register to vote at address of their parents outside of Alachua County. Politicians might not consider university students ng generation surpassing that of the baby boomers (Skaggs & Anthony 2002), andover 7,000 were found to be registered in Alachua County a very considerable votingblock. The Florida State Legislature does keep information regarding the number of registered voters by age groups per district. In Figure 4-2 the number of registered voters between the ages of 18-29 is shown. This age group would most closely identify to tha

PAGE 67

58 of college aged students and the highest concentration of registered voters this age are found where the major universities are located (Tallahassee Florida State, and Gainesville University of Florida). Students Party Affiliation and Parents LSP Cross tabulations The number of registered student voters who have a different student addresspermanent address was determined to be 1,074 (Figure 4-3). Again, this is an extremely small population to test from the overall student population. This analysis illustrates the process that can then be used in future research. The results of the cross tabulation is shown in Table 4-3. This reveals if there is a significant relationship between paraffiliation and LSP. The adjusted residual compares the d and ty ifference between expected count ports at more Republicans would be in this category than either Democrats or Indept. be affiliated with the Independents then Democrats. Type 8 shows the strongest evidence supporting H1 that this category is more likely to be Democrat. Performing the Pearsons and the count. If the null hypothesis (H0) were true (no dependence between partyaffiliation and LSP) the expected count would be the number that is most likely to occurWhen the adjusted residual is around 2 (meaning 2 standard deviation away from the mean), then it can be statistically concluded that the difference is significant. In analyzing the different categories, LSP type 1 and 2 have more Republicans. Therefore, H1 (dependence between party affiliation and LSP) is accepted. This supthe hypothesis th endents. Type 3 shows no significant results due to the fact that there is a low counType 4 is approaching significance in the fact that this category is more likely to be Republican. Type 5 shows no significance in establishing dependence between party affiliation and LSP. Type 6 (combined with 7) shows that this category is more likely to

PAGE 68

59 Chi-squared test (p value of .025) on this data (Table 4-4), confidently reject the null hypothesis (H0). T-Teest jects the dependents (Table 4-6) results in H0 also being rejected. The last test between Democrats and Independents (Table 4-7 ) accepts H0 because the p value is .75 (closes a ly udies s into ne before, and the results are not based sts The t-Tests compare the differences in the means of the data analyzed. This twas used to show which LSP categories the different parties were likely to fall within. The first T-Tests performed was between Democrats and Republicans (Table 4-5). This test resulted in a significant difference between means and the P value of .007 renull hypothesis H0 (LSP type is the same for the categories of party affiliation). This supports evidence the Democrats will be in the higher LSP types (6,8). The test betweenRepublicans and In to 1). Typical Party Demographic It has been hypothesized elsewhere that political affiliation and demographic characteristics are highly correlated (Erikson, Lancaster, & Romero, 1989), (Stanley & Niemi, 1994). For example, one is statistically more likely to be a Republican if one iman, a college graduate, in the top income, and living in a rural area. One is more liketo be a Democrat if one is a woman, a senior citizen, nonwhite, and living in an urbanarea. These hypotheses arise from various polls and surveys such as The Harris Poll, a 2000 exiting poll for CNN (Harris Poll, January 5, 2001), U.S. Census, and other stconducted on issues such as the gender gap (Trevor, 1999) and the party of the rich (Sander, 1988). This University of Florida study breaks down the geodemographicfiner and more accurate detail than has been do

PAGE 69

60 upon le r n behind and ill served. It has been demonstrated here that geographic information systems a survey but actual cross tabulation of student enrollment and voter registration data. Conclusions Understanding student propensity to vote gives the University of Florida an understanding of how much its student constituents may be relied upon to support political actions that might be necessary to maintain the present education environment. Moreover, this study demonstrates methods that can be followed that can identify particular kinds of students and how those students might be approached to support a given political position. Geodemographics of students can be used to project specific issues to susceptibsubgroups of students and their parents. A reasonable conjecture is that if a student votes, then his or her parent is also likely to vote (Beck & Jennings, 1991). The University can encourage subgroups of its students to go to the polls and vote in support of issues aligned with university needs. A statistical relationship has been demonstrated between party affiliation and LSP. This correlation is made on a small population and demonstrates a process for furthestudies. It has been shown that Republicans are more likely to be within LSP types 1 and 2, therefore, letters addressed to these students and their parents might take on a more Republican tone. At the same time, students and their parents shown to fall within those LSP categories more likely to be of Democratic affiliation might receive informatiospecifically directed to that bias. Universities have become politicized. In response, universities must engage actively in the political process else its constituents university students will be left

PAGE 70

61 technology, specifically the technology and analysis of the business geographer, can be applied to generate inform ation that can be used to improve University administration make decisions. Table 4-3. Cross tabulations between party affiliation and LSP Table 4-4. Chi-Square tests

PAGE 71

62 Table 4-5. T-Test Republican and Democrat Table 4-6. T-Test Republican and Independent Table 4-7. T-Test Democrat and Independent

PAGE 72

63 Figur-2.to reereotpeouist e 4 18 29 gistd vers r Hse dtric

PAGE 73

64 Figure 4-3. Numbers for student registered voters

PAGE 74

CHAPTER 5 CONCLUSIONS s e the lot study, ersity a better understanding of the political aspects of the student population. After the successful completion of populating the student database with the correct Florida State Legislator and geodemographic for the students permanent address using GIS, the University will now have the ability to develop a mail campaign. The University now knows which district the parent or student may reside in and vote in so issues central to the well being of the school can be addressed. Further targeting through mail can be performed based on the geodemographic of the parent or student. This will The Registrars Office of the University of Florida on behalf of the University Presidents Office commissioned this proof of concept study. The purpose of the study is to provide a tool whereby the University could bring to the attention of State legislatorthe importance of the university in serving the educational needs of students of their district and state region. GIS and the geodemographics (Geographic Decision Information Systems) havbeen demonstrated here as one means to generate the information necessary to receivepolitical support to further the mission of higher education. In addition to the piinformation regarding registered student voters is included to allow the Univ 65

PAGE 75

66 allow the University to craft letters in such a manner that will appeal to the different LSP types and may generate increased support. The University also knows where trends in the geodemographics types have emerged. If the University was looking for more support in additional state funding to aid students with financial support, they may looks to LSP type 8A Young Immigrant Families. Although there may be portions of this group located throughout the state, they are highly concentrated in South Florida. Only sending letters to this group in South Florida as opposed to just everyone in the database with type 8A can cut costs and still havLearning which students are registered to vote may give the University insight into generating increased support from the student who is more likely to vote. Dem students can help target issues to the students and there parents. There ote.org), (Beck & Jennings, 1991). The University can specifically cater to this group eople urging them to go to the polls and vote in support of university issues. istically discovering if there is any significance between party affiliation and LSP from the sample of student registered voters with a different student address and address, allows assumptions to be made on the entire student database. Since ps letter the mail campaign can be more conservative in tone and the letter to LSP type 8 that is more likely to be Democrats can be crafted in such a way to appeal to a more liberal outlook. e en ou gh sup por t fo r th e is sue ographics is a heseliho od tha t if a s tude nt vot es, his or her par ent ma y a lso vot e (w w w.s tate of man sho ent wn th at R epu bli can s a re m or e li kely to be in LSP ty pes 1 a nd 2 p erha the of tlikethevof pStattypeperit isfor

PAGE 76

67 Utilizing GIS, geodemographics, and voter registration data, can necessitatedifferent data sources to be used in conjunction with each other. This can allow for in depth analysis to make wiser decisions about ones business, in this research, the University of Florida.

PAGE 77

APPENDIX A n of s, home value of over $400,000 is more than four times the national average. Well over 8orders. They drive luxury cars and enjoy taking domestic and foreign excursions. They CACI ACORN LIFESTYLE SEGMENTATION Consumer Type 1A: Top One Percent Demographic: Top One Percent households are usually older married couples in their peak earning years. Almost half are between the ages of 35 and 64 years; their mediaage is 43.0 years. Most are empty nesters, although many have older children attending college. This predominantly white segment also has an above-average percentage Asians and Pacific Islanders. Socioeconomic: Their median household income is $115,900. For most householdsources of income include salaries, interest, dividends, or rental property income for most households and self-employment income. Almost 60 percent of the labor force work in executive or professional positions. They are highly educated; over 60 percent of the adult population aged 25 years or older hold bachelors or graduate degrees. Residential: Most own their single-family homes found in the older suburbs of large metropolitan fringe areas scattered throughout the U.S. High concentrations of Top OnePercent are found in California, New York, New Jersey, Illinois and Texas. Their median 0 percent of these homes are valued at $250,000 or more. Preferences: Top One Percent households rank high for many consumer purchases including most clothing and apparel categories, frequently buying via mail or phone 68

PAGE 78

69 are active volunteers and fundraisers. They also enjoy reading newspapers and playing racquet sports. Consumer Typ e 1B: Wealthy Seaboard Suburbs Demographic: These, but a re of their adult children still live at home: theyre empty-nester wanna Resid9. ty of d older, married couples have no younger children disproportionate sha abes. Their median age is 40.6 years. The proportion of householders between 45 and 64 is over 30 percent higher than the national average. This population is predominantly white but has an above-average percentage of Asians and Pacific Islanders. Socioeconomic: Their median household income is $72,900; almost half of the households earn $75,000 or more. Most households receive income from dividends, interest, or rental properties; almost 20 percent are receiving retirement income or pensions. More than half of the households have a net worth of $150,000 or greater, three times the national average. They are well educated; over 40 percent of the adults hold college degree and/or work in professional or managerial positions. ential: They own their single-family homes that were built between 1950 and 196The median home value reflects the markets where these neighborhoods are located, $261,900; more than two and one half times the national average of $99,800. Wealthy Seaboard Suburbs are found in the suburban areas of East and West Coast urban metros. Preferences: Wealthy Seaboard Suburbs pursue physical fitness by playing a varieracquet sports, golf, working out at the gym, and taking vitamins regularly. They spenfreely on home furnishings and improvements, and also travel extensively, preferring

PAGE 79

70 foreign to domestic trips. They read two or more daily newspapers and metropolitan anbusiness-related m d agazines and listen to classical, jazz and alternative radio. Consumer Type 1C: Upper Income Empty Nesters Demoat ies. r finished high school; more than 40 percent hold a college or graduate Residearly bovegraphic: They are usually empty nesters: married couples with no children living home. They also are predominantly white and middle-aged. Almost 50 percent of the householders are between the ages of 45 and 64; their median age is 42.4 years. Socioeconomic: These couples are prosperous, with a median household income of $68,400. Two-thirds of them receive income from dividends, interest, or rental propertThey are usually business owners or managers. With approximately 20 percent receiving retirement or pension income, early retirement is common. They are also highly educated, which is atypical of their generation. Over 90 percent of the adults aged 25 years and olde degree. ential: They own their single-family homes with a median value of $157,300, n60 percent above the national average. The These neighborhoods are primarily suburbanbut range from urban areas to smaller, non-metropolitan communities scattered throughout the U.S. High concentrations are found in Pennsylvania, New York, Texas, Illinois and Ohio. Preferences: They play a variety of sports including golf, racquet, and sidewalk sports. They also enjoy traveling, preferring foreign travel to domestic trips; visiting museums, and attending concerts. They own PCs, pianos, swimming pools and luxury cars at aaverage rates. They are avid readers, usually two or more daily newspapers, along with metropolitan and business-related magazines.

PAGE 80

71 Consumer Type 1D: Successful Suburbanites Demo graphic: This family market has an average household size of 3.1 persons, 19 percent above the national average. They are between the ages of 35 and 54 years, with schoovalue ers nites ent rates are high for men and women. The work force is profeome were ence for lifestyle and moved out to newer, suburg ith ss, finance and metropolitan magazines. l-aged children. Their median age is 37.1 years, slightly higher than the U.S. of 35.5 years. The population is predominantly white, but Asians and Pacific Islandcomprise a disproportionate share of almost 8 percent. Socioeconomic: At $87,200, the median household income of Successful Suburbais more than twice that of the U.S. figure. Dual incomes and investments contribute totheir affluence. Employm ssional, well educated, and mobile. Residential: Successful Suburbanites own homes in newer suburbs with a median hvalue of $210,500, twice the national average. Most of their single-family housesrecently built between 1980 and 1990. Located predominantly in the urban areas of metropolitan centers, they traded conveni ban developments; over 30 percent commute across county or state lines to work. Preferences: Successful Suburbanites drive minivans and luxury cars, own swimminpools and PCs, and play golf and tennis. They splurge on home furnishings and improvements, apparel, electronic "toys", and travel, and fund much of this spending wcredit cards and loans. In fact, Successful Suburbanites are more likely than most of theother consumer markets to have loans. They attend concerts and movies; read two or more daily newspapers; and busine

PAGE 81

72 Consumer Type 1E: Prosperous Baby Boomers Demographic: Th e age profile of this market is singular: baby boomers who were born de school age children; over 40 percent greater than the national average. Nearly 20 percent of the population is under 10 years of age as compared with less than 15 percent of the U.S. population. Typical of their generation, these families are very mobile. Over 35 percent of the population have moved in the past 5 years, double the national mobility rates. Socioeconomic: Prosperous Baby Boomers have a median household income of $60,100, more than 50 percent higher than the U.S. figure. Their higher income is primarily the result of two salaries. Over 70 percent of the households have 2 or more workers. Also typical of this cohort is educational attainment: two-thirds have attended college or completed a degree. Residential: Over 74 percent Prosperous Baby Boomers tend to be homeowners in new housing developments, 75 percent of their houses were built since 1980. Their median ome value of $124,700 is approximately 25 percent higher than the U.S. median. These primarily suburban neighborhoods are located within commuting distance of metropolitan job centers in the Sunbelt and West. Preferences: Prosperous Baby Boomers are active sports enthusiasts: playing golf, tennis or racquetball; lifting weights; and doing aerobics. Much of their time is spent on the PC playing games, doing their taxes and finances, or just browsing the web. They are likely to visit theme parks or gambling casinos. They buy electronics, camcorders, PCs, between 1949 and 1964 with young, primarily preschool and gra h

PAGE 82

73 sports equipment, and home furnishings. They do not invest or save in proportion to their Consumer Type 1F: Semirural Lifestyles Demographic: Married couples aged 35 to 54 years with and without children living at home dominate this market. The median age is 36.8 years as compared to 35.5 years for e U.S. About 35 percent of the households are empty nesters; 40 percent have school-his population is white. Socioeconomic: With a population of over 15 million, or over 5 percent of the total U.S. population, this is the largest Affluent market with a median household income of $61,500 but not the wealthiest one. Unemployment is low, but their income sources are ore varied in this affluent market, including self-employment income from both farm and non-farm jobs with the usual wages and salaries, plus interest and dividend income. Residential: Semirural Lifestyles live in semi-rural areas or on the urban fringe of metropolitan areas. Many of their homes are newer, built since 1970, with a median home value above the average of $131,500. Most of these homes are valued between $75,000 and $200,000. These neighborhoods are scattered throughout the U.S., but compose over 10 percent of neighborhoods in Connecticut, New Hampshire and Alaska. Preferences: Semirural Lifestyles spend a significant amount of time at home swimming in their pools, doing home projects, gardening, or using the PC. Outside the home, they join political and civic groups, the play golf and attend movies. They buy home furnishings, tools and the latest electronic gadgets, PCs, camcorders, and cameras. Semirural Lifestyles avidly read books; two or more newspapers per day; and science, news and lifestyle magazines. income, but they do carry a lot of insurance. th age children living at home. Over 92 percent of t m

PAGE 83

74 Consumer Type 2A: Urban Professional Couples Demographic: With a median age of 37.8 years, Urban Professional Couples are older than the U.S. overall (35.5 years). They have above-average indexes for each adult age group from 25-29 years and older and below-average indexes for the age group 20-24 years and younger. They are predominantly married-couple families, with few or no children, but the mix also includes single-person and shared households, the results of high divorce rates through the 1980's. Socioeconomic: Their median household income of $49,700, is 25 percent above the U.S. figure. Employment levels are high. The labor force is well educated and employed mainly in professional or managerial positions. Over 35 percent of adults aged 25+ hold a college degree. Residential: They live in a high-density mix of single-family homes and townhouses with smaller two-to-fiveunit rentals. Most homes are owner-occupied, including condominiums. The median home value is above average of $142,700. These neighborhoods are located inside urban areas scattered throughout the U.S., but especially in Massachusetts, California and New York. Preferences: Urban Professional Couples rank near the top in their consumption of imported wines, listening to classical music, traveling overseas and domestically and visiting museums. Physically active, they play a variety of sports, including tennis, golf, bicycling, and jogging. They join AAA, frequent flier programs, and health clubs. They read two or more daily newspapers and gourmet, business, travel and lifestyle magazines.

PAGE 84

75 Consumer Type 2B: Baby Boomers with Children Demographic: Approximately two-thirds of the households are married couples, most with children; over 50 percent more than the national average. Their median age is 31.2 years; more than 35 percent of the population is under the age of 20; 34 percent of the householders are between the ages of 25 and 44 years. Typical of the cohort, many are mobile, still moving to find the best jobs or location. Socioeconomic: With a population of 11.8 million, Baby Boomers with Children represent 4.3 percent of the U.S. population. Like the Prosperous Baby Boomers, this young market has very high employment rates including a high proportion of dual income families, but they are not as affluent. Their median household income is $48,000. Some have graduate schooling, but this group is less likely to have completed a graduate or professional degree. Residential: They live in single-family homes built during the 1970s. Most of their homes are owner occupied and valued with a median home value of $95,700, just below the U.S. median of $99,800. Neighborhoods are found within the metropolitan urban fringe, smaller towns and in rural non-farm communities. Western states, especially Utah, Alaska and Wyoming have a lot of Baby Boomers with Children, harking back to the westward migration of baby boomers in the 1970s. Preferences: Baby Boomers with Children represent familyand home-oriented consumerism. They enjoy do-it-yourself home projects, go camping, fishing, and hunting. They rent videos; grow vegetables; use their PCs; and visit museums, zoos, and theparks. They own pets and spend their discretionary income on swimming pools, outdoor grills, campers. They own multiple cars, minivans, and sports utility vehicles. in me

PAGE 85

76 Consumer Type 2C: Thriving Immigrants Demographic: Although fewer than 30 percent of the population are foreign-born, 40 percent speak a language other than English at home. Thriving Immigrants are immigrants and their children. Most households are families, typically married couples with either very young or adult children. The population is diverse, including Asian and Pacific Islander at 4.5 times higher than the national average and Hispanic origin at three times the national average. The age distribution is slightly younger; the median age is 32.9 years. Socioeconomic: Although this is a small market, only 1.9 percent of the U.S. population, e households in this market earn moderate to high incomes, primarily from wage and salary employment. Their median household income is $48,900. Employment is high; unemployment is low. Despite their relative affluence, these consumers are a weak financial market. They neither invest nor save their income. Loans also rank low for this group. Residential: Home values and rents are especially high for Thriving Immigrants due to their location. These neighborhoods are most likely to be found in Californias urban areas. The median value of owner occupied units is $184,900, almost twice the national median. Home ownership is slightly below just over 59 percent, comparable to the U.S. rate. Most of their homes are single family, built between 1950 and 1969. Preferences: Because of ties to their birth countries, Thriving Immigrants are likely to have passports and spend a sizable portion of their income on foreign trips and long-distance phone calls. They also enjoy day or weekend excursions to casinos and theme th

PAGE 86

77 parks. They watch television; use the PC; listen to adult alternative, rock and contemporary radio; and read news, fashion and lifestyle magazines. Consumer Type 2D: Pacific Heights Demographic: Pacific Heights are predominantly Asians and Pacific Islanders but with nearly 14 percent, there is a significant Hispanic minority. Almost 35 percent are foreign-born and half speak a language other than English at home. The segments age profile is almost identical to that of the U.S. overall; indexes for all age categories are within 10 percent of their U.S. counterparts. The household distribution parallels the United States dian household income of $48,400. One of every two house Residiums. with a slightly higher proportion of adult children living with their parents 35 percent above the national average. Socioeconomic: This small about one percent of the U.S. population but upscale consumer market has a me holds earn $50,000 or more annually. Their sources of income include self-employment, interest and dividends, wages and salaries. Education is a priority: the labor force is well educated. Half of the adult population over age 25 attended college or received a bachelor's or graduate degree. Another 10 percent are presently enrolled incollege. ential: Pacific Heights neighborhoods are high-rent districts, located primarily in Hawaii and California. With median home values around $250,000, ownership is below average (50 percent). Twelve percent of the owner-occupied units are condominMost of the housing is in multi-unit complexes with five or more units. Population density is high, over 15,000 persons per square mile.

PAGE 87

78 Preferences: True urbanites, Pacific Heights own few cars, and when they do, theyll buy a foreign, subcompact model. To keep in touch with relatives, they spend a dispros. o their age profile is slightly older, with a median of 37.2 yian net r-e d on home improvements, furnishings, and outdoor equipment. They watch portionate share of their income on foreign trips and long-distance phone callThey watch TV, listen to the radio, gamble at a casino, play the lottery, go skating and play billiards. They are inclined to invest rather than to save their money. Consumer Type 2E: Older, Settled Married Couples Demographic: Older, Settled Married Couples are middle-aged married couples whhave settled into their neighborhoods and surroundings. Although many households include school age or adult children ears. The population is predominately white. Socioeconomic: This market includes 4.4 percent of U.S. households and 4.6 percent of the U.S. population. The income distribution reflects a degree of comfort earned by working couples aged 45-64 years. Almost half earn $50,000 or m ore. The me dworth of this segment is $69,000, 75 percent higher than the U.S. median of $39,000. Almost 20 percent of these households are drawing retirement income. Residential: Older, Settled Married Couples live in older, single-family homes owneoccupied with a median value of $95,100, compared to the U.S. median of $99,800. Thgeographic distribution is national, but higher concentrations are in Michigan, Illinois andOhio, along with New York and California. Preferences: Older, Settled Married Couples are likely to own stock, savings accounts, bonds, IRAs, and other securities, but may also have debt from loans and credit cards. They spen

PAGE 88

79 TV; le single with a median age of 37.9 years that belies the concentration of younger households. Half are between 20 and 44 years of age, compared with only 37 percent of the U.S. population. Fifty percent of these households are single-person or shared -14 percent. The population includes foreign-born residents, non-Mexican Hispanic, Asian and Pacific Islander. Socioeconomic: Contrary to the usually lower incomes for single households, High Rise Renters are affluent. They have a median household income of $48,900 versus the U.S. median of $39,800. Well educated and gainfully employed, half of them hold a bachelor's or graduate degree and work in professional or managerial positions. They prefer investing over saving. Residential: They are urban, mobile, apartment dwellers in densely populated, central city locales in the largest metropolitan areas. Renters comprise six of ten households versus three of ten U.S. households. Although most of the neighborhoods are located in the largest metropolitan areas of New York, Illinois and California, more than 30 percent of the neighborhoods in the District of Columbia are High Rise Renters. Preferences: Typical of more affluent consumers, High Rise Renters join health clubs and environmental groups and pursue racquet sports, bicycling, and jogging. Their leisure time is singles-oriented with dancing, going to bars/nightclubs, the theater, or museums. They buy expensive clothing. They do not take out loans but use their credit cards isten to jazz and golden oldies; read newspapers and lifestyle, health and hommagazines. Consumer Type 3A: High Rise Renters Demographic: They are

PAGE 89

80 extensively. They often read metropolitan, fashion or gourmet magazines, and two or more daily newspapers. Consumer Type 3B: Enterprising Young Singles Demographic: With a median age of 30.1 years, this population is young and mobile. More than three of every five persons are under the age of 35. Approximately half of the households are sd households -er ompleted a two-year associate degree. Employment is high. t Sunbelt. If Many read health, computer ingle-person or single roommates or share compared with less than 30 percent of U.S. households. Socioeconomic: Despite their youth, Enterprising Young Singles have a median household income of $40,200 that is slightly above the U.S. median. Approximately onethird of the population aged 25+ years has a four-year college degree or higher; anoththird has attended college or c Enterprising Young Singles have the highest rate for female employment; over 75 percent. Residential: Most Enterprising Young Singles rent, nearly 64 percent versus 30 percenfor the U.S. overall. They prefer living in newer apartment complexes built in the 80sand s in urban areas throughout the U.S., but concentrated in the West and they own a home, its median value is $113,200. Preferences: Enterprising Young Singles enjoy an especially young, active lifestyle. Fitness is a priority; many belong to gyms, lift weights, jog and play racquet sports. own only one car, usually a foreign, compact model. They go dancing in nightclubs, visitmuseums, theme parks and zoos, and play computer games. They and photography magazines, and listen to jazz, rock and R&B music on the radio.

PAGE 90

81 Consumer Type 4A: Retirement Communities Demographic: They age of 40 years, d 75 years ers me of heir income sources are varied because of the large number of retirees who salaries. the f ds include singlebe g re older, but not exclusively elderly. With a median a they are actually the youngest of Retirement Communities. Householders ageand older represent over 15 percent of the household distribution, but so do householdaged 35-44 years. Single-person households are 35 percent of the total, followed by 27 percent of married couples without children. Socioeconomic: This is a small, prosperous market with a median household inco$42,600. T receive their income from pensions, Social Security, interest and dividends, and The labor force is also small, but well educated and professionally employed. Nearly three out of ten hold a bachelors degree or higher, compared with two out of ten inU.S. overall. Residential: Congregate housing with meals are included in the rent, accounts for most othe housing in Retirement Communities. Surrounding neighborhoo family homes with a median value of $124,400. Most of the housing is relatively new with almost 40 percent being built in the last decade. Retirement Communities mayfound in any state, but Florida and California locations are especially popular, accountinfor the higher mobility of this group. Preferences: The lifestyles and spending of Retirement Communities are eclectic as diverse as their neighborhoods. They play golf; train with weights; go bicycling; camping; listen to classical music; and read consumer, science and nature magazines.

PAGE 91

82 They enjoy traveling overseas more than taking domestic trips. They invest, probably owning certificates of deposit, IRAs, stocks and other investments. Consumer Type 4B: Active Senior Singles Demographic: This mature market has a median age of 43 years. Nearly 25 percent are aged 65 or older; many are widowed. Single-person households make up over 40 perceof these households. Although younger families live in these neighborhoods, ther nt e are erage at ore inclined to save than to invest. eas with e Senior Singles enjoy traveling, usually taking foreign trips, RAs, bonds and stock. few children. With over 85 percent of the population white, this market ranks below average in diversity. Socioeconomic: Representing 2.8 percent of U.S. households, Active Senior Singles are one of the largest senior markets. Their median household income is below av$36,200, but many of them are retired. Poverty and unemployment rates are low. Theyare fairly well educated; nearly 50 percent have attended college or higher. They are slightly m Residential: Most Active Senior Singles live in apartments. Two-thirds of these housing are multi-unit, built before 1970. At $117,400, the median home value is above the U.S. median. Neighborhoods are naturally higher density, located in urban ara disproportionate share in the Mid-Atlantic and Northeast. Preferences: Activ gambling at the casino, playing the lottery, attending concerts, visiting museums, and going to their health clubs. They are active volunteers and involved in political campaigns. Their heavy expenditures for imported wine, champagne and bottled water reflect their taste for premium products. They own a variety of investments such as longterm certificates of deposit, savings accounts, I

PAGE 92

83 Consumer Type 4C: Prosperous Older Couples mfortable living; their median st, ber other nd Demographic: Their median age is 43.2 years, but nearly half are 55 years of age or older. Most are married; few have younger children, although some families still have adult children living at home. With its population that is more than 90 percent white, this market lacks diversity. Socioeconomic: Prosperous Older Couples enjoy a co household income is $45,200 and their median net worth is $88,900. Retirement is prevalent, but not dominant. Many householders are evidently planning for their retirement with investments and savings. More than half receive income from interedividends, or rental properties. Residential: They own suburban, single-family houses with a median value of $99,000, slightly lower than the national average. These homes are in older, established neighborhoods. The geographic distribution is somewhat more urban, but otherwise similar to the U.S. population distribution by region. The states with the highest numof these households are New York, Pennsylvania, Illinois and Ohio. Preferences: Prosperous Older Couples are enthusiastic investors with diverse portfolios that include certificates of deposit, IRAs, stocks, savings accounts andinvestments. Typically, they own luxury cars, use coupons, make contributions and recycle. They enjoy diverse interests that include golf, needlework, outdoor gardening and gambling in casinos and playing the lottery. They watch TV; read senior, health ahome magazines; and listen to news/talk radio.

PAGE 93

84 Consumer Type 4D: Wealthiest Seniors Demographic: Their mile of of rcent in California. They usually own their single-family orhood type stments, especially long-term edian age of 53.5 summarizes the demographic prof Wealthiest Seniors: over half of them are over 50. Over 30 percent are 65 or older, more than twice the national average for this age group. Half are married couples with nochildren at home, and approximately one fourth are single-person. Socioeconomic: Wealthiest Seniors is a small market, slightly more than 1 percenthouseholds, but very affluent. Their median household income of $44,700 hides thedisproportionate share of upper income households. For example, one in seven households had an income of $100,000 or more compared with one in 11 U.S. households. This is the top financial market for investments and savings. Their median net worth is $125,900, more than three times the U.S. median. Residential: They live in small cities and towns in sunny climates. Nearly half of them are located in Florida and ten pe homes or condominiums. Their housing is newer with a median value of $144,500. Approximately 30 percent of the housing inventory is seasonal. Preferences: The expendable income available to Wealthiest Seniors retirees distinguishes this small market. The purchase potential indices for this neighbrank near the top among ACORN markets for inve certificates of deposit and stock; travel; luxury cars; credit card use; and jewelry purchases. To keep in shape, they go bicycling, play golf, and walk for fitness. They alsoenjoy reading books; newspapers; and senior, lifestyle and business magazines; and watching news programs and dramas on TV.

PAGE 94

85 Consumer Type 4E: Rural Resort Dwellers Demographic: Rural Resort Dwellers are us ually older. Their median age is 41.7 years but they are concentrated in the 45-plus age groups. More than four of ten are aged 45 e t earn a modest e. r all t and e years and older. Families are predominantly married couples, many recently retired to tharea. Over 95 percent of local residents are almost exclusively white. Socioeconomic: Over one-quarter of the households earns less than $20,000 annually, butheir poverty rate is the same as the national rate. Rural Resort Dwellers living from seasonal employment and farming. Over 20 percent receive retirement income and 35 percent receive Social Security. Most of the labor force works part-timAbout 15 percent have a bachelors degree or higher compared with 20 percent of the U.S. population. Residential: They own single-family or mobile homes. Over sixty percent of the housing in these areas is vacant for seasonal use. The median home value of $92,900 is low foowner-occupied units, but comparatively high for rural areas. Over 90 percent of the Rural Resort Dwellers are in rural villages. These areas can be found in most states, bumost are concentrated in the North, around the Great Lakes or in New England, especially Maine, New Hampshire, and Vermont. Preferences: Rural Resort Dwellers love the great outdoors and often go huntingfishing. They own four-wheel drive vehicles, powerboats and campers. Health and nutrition are important as they shop at health food stores, follow low-fat diets, and takvitamins. They are likely to buy tools and outdoor equipment, satellite dishes, household appliances and cameras.

PAGE 95

86 Consumer Type 4F: Senior Sun Seekers Demographic: The oldest in the Retirement Styles summary group, more than half of the Senior Sun Seekers are 55 years or older. Their median age of 59.2 years is nearly 24 years older than the U.S. median. Over 45 percent are married couples without childrenbut there is 32 percent of single-person households in this segment. Socioeconomic: Senior Sun Seekers is a modest market in numbers, totaling 1.4 percent of U.S. households and 1.1 per cent of the U.S. population. Although most y are ng lottery tickets. $86,200. percent of Arizona neighborhoods. end households earn less than $30,000 annually, much of their income is disposable. Their primary sources of income are Social Security, interest, dividends and pensions. Theone of the top-ranked markets for investments and savings. They are also one of the top markets for buyi Residential: Mobile and single-family homes account for most of the housing of Senior Sun Seekers. They own homes in newer areas, with a median home value of Their neighborhoods also include seasonal housing for visiting snowbirds, pluscongregate housing and nursing homes. Almost 90 percent are in the South or West, including about 12 percent of Florida and ten Preferences: Senior Sun Seekers have income and time at their disposal, and they spit golfing, traveling overseas and domestically, playing cards, doing needlework, gambling in casinos and playing the lottery. They are health conscious, following low-fator other diets, and taking vitamins. They enjoy reading senior, health and science magazines and watching news programs, dramas and situation comedies on TV.

PAGE 96

87 Consumer Type 5A: Twentysomethings Demographic: The median age of Twentysomethings is 30 years, 5.5 years younger than the U.S. median. Over 27 percent of residents are in their 20s that is double that of the U.S. percentage. They are mobile and in transition, completing college or starting their careers. Nearly 60 percent live in single-person or shared households compared to 30 percent for the U.S. overall. Socioeconomic: Only 1.7 percent of the U.S. population, Twentysomethings is a relatively small market. Their median household income is $26,700. Employment is average, although half are employed only part time. Education is the key to the futuTwentysomethings; over 35 percent have an associate degree or higher, 20 percent are attending college. Residential: Twentysom re of ethings live in city apartments. Nearly 20 percent of the go hiking, jogging and train with d neighborhoods are in California or Texas. Rent is generally below average. Approximately 22 percent of the housing in these neighborhoods are single-family, owner-occupied homes. The median home value is $88,800, 10 percent lower than the U.S. median. Preferences: The Twentysomethings young, active and urban lifestyle reflects their ageand environment. They play racquet and team sports weights. They frequently visit museums, go to the movies, attend concerts, bars, annightclubs. They buy CDs, books, shoes, sportswear, soft drinks and alcohol, and vitamins. They watch sports and late-night shows on TV; read newspapers; and mens photography, science and computer magazines.

PAGE 97

88 Consumer Type 5B: College Campuses Demographic: Not surprisingly, almost three-fourths of College Campuses are collestudents. Their median age is 21.7 years compared to the U.S. median of 35.5 yea45 percent live in dormitories; the rest, in nearby neighborhoods that are primarily student housing. Forty percent are under the age of 25; 70 percent are living in single-person or shared households. Socioeconomic: This market is either well educated or on its way to becoming well educated. O ge rs. Over ver fifty percent have a bachelors degree or higher, 2.5 times the U.S. is single-family, owner-occupied. Their rent is below the U.S. average gross rent, ng. They also percentage. Another quarter of the population has attended some college or has an associate degree. Their median household income is $21,000. Over half is employed, with most working in part-time, low-paying, service-sector jobs. Residential: Apartment rentals dominate off-campus housing. Over 50 percent of thehousing inventory are multi-unit buildings, with five or more units. One sixth of the housing but not by much. Most College Campuses are located in cities; however, these neighborhoods also represent the smaller college towns. Preferences: Aside from college expenses, most of their budget is spent on travel or theiractive social lives. College Campuses spend much of their free time playing racquet sports, training with weights, hiking, playing billiards, jogging, and bicycli visit museums, go to the movies, bars, and nightclubs. They buy CDs and books, PCs, computer software, beer and wine, and vitamins. They watch a lot of sitcoms, dramaslate-night shows and sports on TV.

PAGE 98

89 Consumer Type 5C: Military Proximity Demographic: This young, mobile population has a median age of 23. 6 years. Over 40 percent have recently moved. Half of these households are young couples, many with t me is $32,500. The incidence of poverty is just above the arracks. n home es: The expenditures and leisure activities of Military Proximity reflect this and preschool children, although there are some single-parent families. These racially-mixed neighborhoods are predominantly black and white with an aboveaverage percentage of Asians or Pacific Islanders. Socioeconomic: Over 30 percent of the labor force is in the Armed Forces. Employmenis high for civilians, too, although unemployment is slightly above average. Most workers have postsecondary schooling, but only 17 percent hold a bachelor's or graduate degreeTheir median household inco U.S. average. Residential: Almost ten percent of Military Proximity live on-base in military bThe housing units in the residential areas are mainly apartment rentals. The mediavalue of $82,000 is 18 percent lower than the U.S. figure. Military Proximity neighborhoods are located inside the urban areas of Virginia, the Carolinas, GeorgiaFlorida, Texas and California. Preferenc markets composition of young families. These households rank high in their purchases for their youngest: childrens clothing, shoes, toys, videos, and vitamins. They enjoy jogging and weight training, going to bars and nightclubs, eating fast food, and visitingtheme parks. They watch dramas and situation comedies on TV and listening to jazzR&B on the radio.

PAGE 99

90 Consumer Type 6A: East Coast Immigrants Demographic: East Coast Immigrants are the new Americans. Almost 40 perceforeign-born and most speak a language other than English at home. Their demographiprofile is a rich mix of ethnic and racial groups and household types such as Hispanic, Asian, and black. Sixty percent are married-couple or single-parent familie nt are c s; forty percent are single-prs, with ol diploma are 40 percent higher -hese sehold expenditures include foreign travel, gambling casinos. erson or shared households. Their median age is 33.8 yea above-average percentages of younger householders aged 20-34 years. Socioeconomic: East Coast Immigrants represent 2.5 percent of all U.S. households and 2.4 percent of the U.S. population. Their median household income is $33,100. Unemployment and the lack of a high scho among this segment but employment is comparable to the U.S. figure. These workers commute, over 30 percent cross county or state lines to work. Residential: Most rent in the urban canyons of large cities in high-density, high-rise pre1950s apartment buildings. The average rent is slightly above the national average. Homeless shelters and other non institutional group quarters are also in t neighborhoods. Almost half of East Coast Immigrants live in New York; 70 percent livealong the Eastern Seaboard. Preferences: Living in urban areas with good public transportation, East Coast Immigrants dont often own cars. Their hou long distance phone calls, buying sportswear, and VCRs. They watch situation comedies, dramas and sports on primetime TV; listen to jazz and R&B on the radio, and read newspapers and fashion magazines. They take day trips to museums, theme parks, and

PAGE 100

91 Consumer Type 6B: Working Class Families Demographi c: Working Class Families are approximately 90 percent black and 75 percent family. Although slightly older than people in the U.S. overall, many households se slightly 37,600. Despite chronic high ge ns per square mile, but single-unit, owner-occupied homes are cent es own few have grade-school age children or teenagers. Single-parent households compriless than one of five households at 18.5 percent compared to one of eleven U.S. households at 9.3 percent. Most are 35 to 74 years old. Their median age is 36.8 years. Socioeconomic: This small, middle-income market totaling 1.2 percent of all U.S. households, has a median household income is $ unemployment, the employment rate is above average for women and just below average for men. Most work in service or government jobs. At $41,800, their median net worth is above the U.S. median of $39,100. Residential: Working Class Families live in urban neighborhoods, with an averadensity of 10,000 perso prominent. Almost 15 percent of the housing inventory are row houses, twice the national average. Built decades ago, these homes have a median value of $71,900. These communities are located in urbanized areas scattered throughout the U.S. Over 20 perof the neighborhoods in the District of Columbia are Working Class Families. Preferences: Consistent with their urban lifestyle, Working Class Famili cars, but when they buy a car, it is a luxury model. They also buy TVs and clothing. Theywatch time watching sitcoms, dramas, and sports on TV, gambling at casinos, playing the lottery, playing team sports like softball, dancing at nightclubs, or going to the movies. Consumer Type 6C: Newly Formed Households

PAGE 101

92 Demographic: These young, newly formed households are characteristic of this markeWhile their median age i t. s 33.9 years, many of them are between the ages of 20 and 34 years. The mix of household types includes single parents, single-person and shared housed e. is a S. population. These households nd t monly in the smaller, metropolitan areas of Iowa, Kansas, Maine, Nebraska, ng. Most of their TV, radio, and print d in the U.S. for generations. Most speak Spanish at home. Demographically, Southwestern holds, in addition to below national level, but still sizable percentage of marriecouple households with and without children. Eighty-five percent of this market is whitSocioeconomic: With a population of over 13 million, Newly Formed Householdslarge market that comprises almost five percent of the U. are moderate-income; their median household income is $35,300. Unemployment apoverty are low. Half work in the manufacturing and service industries. Residential: Newly Formed Households offer affordable housing in older, single-family homes, and duplexes. Their median home value is $73,500, 26 percent below the national figure. These neighborhoods can be found in the urban areas of any state, bumost com New Hampshire, Oregon, Rhode Island, Vermont, Washington, South Dakota, and Wisconsin. Preferences: Newly Formed Households visit zoos, use tobacco products, play cards, gamble at casinos, play the lottery, and go campi preferences are near the U.S. average, except for the popularity of certain TV sports events and sitcoms, and several womens and science magazines. Consumer Type 6D: Southwestern Families Demographic: These families portray the Hispanic culture characteristic of the Southwest. While about 20 percent is foreign-born, most of these families have live

PAGE 102

93 Families are large; the average family size of 3.7 persons is 20 percent higher than the U.S. average. Their meildren and mparable to the U.S. figure but opportunity y aller cities, Southwestern Families also ramas, science fiction programs, and movies on ten households are families with children, either married couples or single parents. Their dian age of 28.4 years reflects their emphasis on ch family. Socioeconomic: The Southwestern Families market is midsize at 2.7 percent of the U.S. population with a median household income of $25,600. Half of the adult population has not completed high school. Unemployment and poverty rates are about twice the national average. Employment for males is co for female employment is well below average. Jobs are primarily part-time employment with almost seven percent of the labor force working in farming, 2.7 times the U.S. average. Residential: They live in predominantly urban, older areas with relatively high vacancrates. Most of their homes are single-family or mobile, with a median value of $56,800. Located primarily in urbanized areas and sm live in rural, non-farm neighborhoods. Naturally, most of this market can be found in New Mexico, Texas, Arizona, Colorado, and California. Preferences: Not surprisingly, Southwestern Families lead a family-oriented lifestyleMost households have at least one car and many own a dog. They usually buy family necessities such as baby products, children's clothing, over-the-counter medications, vitamins, and groceries. They watch d television, and frequently rent videos. Consumer Type 6E: West Coast Immigrants Demographic: West Coast Immigrants are young and family-oriented. Almost six of

PAGE 103

94 median age is 24 years compared to the U.S. median of 35.5 years. The average familysize is 3.9 persons versus with 3.1 persons per family for the U.S. Ei ghty percent of the population is Hispanic ann is immigrants. -ce y t . Most households own at s on alf of the d speak Spanish at home. Half of the populatio Socioeconomic: This is a small, working class market with a median household income of $28,400. Despite their language problems and low education 60 percent have no high school diploma employment is above average. Most work in skilled or unskilled bluecollar jobs: many, in manufacturing. Unemployment and poverty rates are almost twithe national rates. Residential: Almost 75 percent of the West Coast Immigrants neighborhoods are in California, with 13 percent in Texas and seven percent in Iowa. Most residents 63 percent rent, primarily in older apartment buildings. Their average gross rent is slightlabove the U.S. average. Their median home value of $115,600 is more than 15 percenabove the U.S. median. Preferences: West Coast Immigrants neighborhoods are naturally family-orientedMost of their budgets include heavy expenditures on groceries, childrens apparel and shoes, baby products, and long-distance phone calls. They enjoy family trips to theme parks and going to the movies, but they also frequent casinos least one car, usually a two-door sedan. They watch dramas, science fiction, sitcomTV, listen to jazz, alternative adult and R&B radio, and read car/motorcycle, mens andscience, weekly and news-oriented magazines. Consumer Type 6F: Low Income, Young & Old Demographic: The Low Income, Young & Old are the very young and the elderly, whoare supported by a relatively young working-age population. Almost h

PAGE 104

95 households are single-ars represents ods households earn less than $30,000. Their unemployment and ty, er 13 nt for most of the housing in these neighborhoods. Half he ildrens vitamins, nts. seholders aged 45-64 and fewer under 35 years. Seventy percent are married couples, compared to 55 percent for the U.S. The parent or single-person. Their median age of 31.8 ye the gap between those under 35 and over 64 years. These racially diverse neighborhoalso include whites, blacks, Hispanics, and American Indians. Socioeconomic: Their median household income is $22,500. Over 60 percent of Low Income, Young & Old poverty rates are twice the national averages. Thirty percent are receiving Social Securiand 20 percent are on public assistance. Employment is below the national average.Residential: Most of their housing is older, built before 1950. Vacancy rates of ovpercent are relatively high, and the median home value of $44,500 is low. Single-familyunits, duplexes and quads accou are renter-occupied. These neighborhoods are urban, located in metropolitan areas and smaller cities throughout the country, with an especially high concentration in tMidwest. Preferences: Not surprisingly, Low Income, Young & Old is a dual consumer market with expenditures for young children and senior citizens. They buy ch baby products, denture cleaners, over the-counter medications and home permaneThey watch daytime dramas, movies and sitcoms on TV and read soap opera magazinesConsumer Type 7A: Middle America Demographic: The demographic profile of these communities is similar to that of the U.S. population; theyre just a little older, more family-oriented, and white. Their medianage of 36.8 years is slightly older with more hou

PAGE 105

96 distribution of children is s3.1 persons per et, ercent of these cated he 20 percent lower than sic; y families own as dogs f the to move frequently. imilar; the average size is at the U.S. level, family. Socioeconomic: With over 22 million in this group, Middle America is a huge markrepresenting almost eight percent of the total U.S. population. Their median household income is $40,400 compared with $39,800 for the U.S. Almost 40 p households earn less than $30,000, but few are below the poverty level. Employment is average, unemployment below average. Most work in manufacturing or farming Residential: Middle America households are non-farm rural neighborhoods, loprimarily in the Midwest or South. Single-family and mobile homes dominate tlandscape. Over 15 percent are mobile homes, twice the national proportion. Their homes are owner-occupied and valued at an average of $82,700, nearly the national average. Most were built after 1970. Preferences: Middle America lives are busy and centered around home and the outdoors. They hunt and fish; do needlework; grow vegetables; listen to country muand read country, family, or hunting and fishing magazines. Man powerboats, satellite dishes, campers, chain saws and other tools, and pets such and cats. Consumer Type 7B: Young Frequent Movers Demographic: Young families dominate in this market. Children are 30 percent opopulation, lowering the median age to 33 years. The population is 85 percent Anglo (non-Hispanic white), but also includes blacks, American Indians, and Hispanics. Characteristic of their youth, this group tends

PAGE 106

97 Socioeconomic: Young Frequent Movers include over 3.4 percent of the U.S. ble to national ge. h the states with the highest concentrations of these neighborhoods are or t for pet e ng at home. Their median age is 36.6 years. This neighborhood population and U.S. households. These families earn low-to-moderate incomes. Their median household income is $35,900. Employment is above average with most of the workers working in skilled or unskilled blue-collar jobs in farming, manufacturing, mining or construction. Unemployment and poverty rates are compara averages. Residential: Mobile homes outnumber single-family houses in these neighborhoods. Over45 percent of the housing are mobile homes, six times higher than the national averaMost of their homes are newer, owner-occupied, and valued below the U.S. median of $86,600. Most of the Young Frequent Movers are in non-farm rural areas. Half are inthe South, althoug Nevada, New Mexico, and Wyoming. Preferences: Young Frequent Movers are more likely to have loans than investments or savings. Most of their loans are personal or automobile loans, usually for trucks, vanssports utility vehicles. They hunt, fish, watch rented videos, listen to country music, and read automobile or hunting/fishing magazines. This is also the top-ranked markeownership such as cats and dogs. Consumer Type 7C: Rural Industrial Workers Demographic: Primarily composed of older families with older children, the RuralIndustrial Workers population is fairly stable. Born and raised in the same state, they arnot even inclined to migrate to a different county. Most are married couples with school-aged or adult children livi

PAGE 107

98 type is mostly white, but also has a disproportionate share of blacks and American t of the adult than l Workers represent 15 to 35 percent of the tances to s, or o move away. Most households are families, married couples with or without children. Few householders are under the age of 35; almost 40 percent are 45 or older. The younger Indians. Socioeconomic: Rural Industrial Workers live in one-horse economies that are dependent upon manufacturing and sustained by farming. Employment is below average, especially for women, and unemployment is high. Almost 45 percen population do not have a high school degree. The median household income is less $28,000, with over 50 percent of the households earning less than $30,000. Residential: Single-family houses are over 70 percent of the housing; many are mobile homes. More than seventy percent are owner-occupied; 13 percent are vacant. Their median home value of $52,800 is just over the national average. Most neighborhoods arerural and Southern. Rural Industria neighborhoods in Alabama, Arkansas, Georgia, Kentucky, Louisiana, Mississippi, the Carolinas, Tennessee, Virginia, and West Virginia. Preferences: This markets lifestyle is rural, but not remote. Commuting long diswork is a way of life. Most households own two or more vehicles, usually trucks, vansports utility vehicles and prefer American-made cars. They hunt, fish, and listen t country or gospel music. They also enjoy urban conveniences like eating fast food and renting videos. They own pets, satellite dishes, powerboats, and chain saws, have personal loans, and watch movies, dramas and sitcoms on TV. Consumer Type 7D: Prairie Farmers Demographic: Prairie Farmers are aging and stable as their children mature and

PAGE 108

99 householders have school-f the population e c income are self employment in farming and non-ed ner-occupied and one-fifth is vacant, including about five percent of s e Dakotas. and apportion little of their budget for luxury items. Active outdoors, they ilies in these neighborhoods. Their median age children, which lowers the median age o somewhat to 37.7 years. The population is over 95 percent white. Socioeconomic: Farmers dominate the Prairie Farmers work force; about 30 percent ofhouseholds are self-employed farmers. Their median household income is $32,100. Thincidence of poverty is just above the national average, but few households receive publiassistance. Sources of household farming occupations, interest and dividends, and Social Security. Most have graduatfrom high school, though few have college degrees. Residential: Prairie Farmers usually live in older, single-family homes. The housing is supplemented by a large number of mobile homes. There are few renters. Over 64 percentof the housing is ow the units held for seasonal use. The median home value is $54,000. Prairie Farmerrepresents ten to twenty percent of the households in Iowa, Nebraska, and thPreferences: Prairie Farmers are more inclined to have checking accounts, savingsaccounts, and loans than to invest their money. They are well insured with health and home policies hunt, fish, hike and go camping. They also enjoy dancing, needlework, watching television, and spending time with their pets. Consumer Type 7E: Small Town Working Families Demographic: The age and household distributions for Small Town Working Famalmost parallel the U.S. profile. More grandparents, aged 75 or older, and more familiesespecially couples with school-age children live

PAGE 109

100 age is 36.1 years comparedre predominantly ofessional to laborer, but most are employed in blue-collar but e U.S., about half of eel amas and sitcoms on TV, rent er .6 f senior householders. Nearly 20 percent are 65 years or older compared with approximately 14 percent for the overall U.S. These with the U.S. median of 35.5 years. They a white and born in the U.S. Socioeconomic: Small Town Working Families is a relatively small market, representing less than two percent of the U.S. population. Their median household income is $35,000. Employment and unemployment rates are average. They work in a variety of occupations from pr jobs in the main industries, manufacturing and farming. Residential: Small Town Working Families homes are usually modest, owner occupied, single-family structures in older neighborhoods. Some live in apartmentsthe rest of the housing are mobile homes. At $71,800, the median home value is more than 25 percent below the national figure. Scattered throughout th these neighborhoods are rural, in non-farm areas, the balance being urban in non metropolitan areas. Preferences: Small Town Working Families enjoy small-town family life. Their lifestyle also includes a bit of country living; hunting and fishing, driving four-whdrive vehicles and pickups. They watch cable, movies, dr videos, read womens magazines, and listen to country music. They spend more than most ACORN markets on groceries, but also are more likely to diet. Consumer Type 7F: Rustbelt Neighborhoods Demographic: The population of Rustbelt Neighborhoods is stable, but aging. Youngpeople are leaving these areas while the older residents remain. Their median age of 39years belies the large concentration o

PAGE 110

101 households are also typicame with adult 40 ages and Social Security produce a median household y munities can be found in urban areas throughout the United States, lework, agazines. They are conservative investors, more likely er rtland Communities are older, with the median population age of 41 years versus 35.5 years for the overall U.S. Few younger householders or children are in l of an older population: married couples, so children still living at home, and single-person households. Socioeconomic: The economic vitality of these industrialized areas has waned. Almostpercent of the population have retired so employment is below average. The remaining jobs are mainly in blue-collar, skilled, unskilled and service occupations. Manufacturing remains the chief industry. W income of $33,100. Residential: Rustbelt Neighborhoods are located on older urban streets in metropolitan cities and smaller towns. Most homes are owner-occupied, single-family houses, but townand row houses are also common. The median home value of $61,600 is relativellow. These com although the highest concentrations are around the Great Lakes in Pennsylvania, Ohio, and New York. Preferences: Rustbelt Neighborhoods enjoy more sedate pastimes such as needwatching movies, mysteries, sitcoms, and daytime dramas on TV, and reading soap opera, family, senior, and home m to have savings accounts, U.S. savings bonds, and certificates of deposit, than to own common or preferred stock. Their purchases of denture cleaners, bifocals, over-the-counter and prescription medications, lottery tickets, and news tabloids indicate an oldpopulation. Consumer Type 7G: Heartland Communities Demographic: Hea

PAGE 111

102 this market. As the population ages, the dependency ratio of those under 15 years and old r, wages and salariployed t are owner-occupied and he ting, at 65 years, to the working age population of 15-64 years is increasing. Households arestill predominantly families, but married couples with no children at home and singles arebecoming increasingly common in the Heartland Communities. Socioeconomic: Their median household income is $27,000. Employment is low, especially for women, due in part to retirement or the lack of opportunity. Over 40 percent of the households are receiving Social Security income. Howeve es are not the only source of income because many in this market are self-emin farming or other business. Heartland Communities are active financially, with investments, savings and loans. Nearly one-third of adults did not finish high school. Residential: Their homes are usually older, single-family and owner-occupied. Single-family houses account for most of the housing. Over 60 percen 13 percent is vacant. The median home value of $54,900 is 45 percent lower than the national average. These neighborhoods are in rural communities and towns outside metropolitan areas. Heartland Communities represents ten to 22 percent of thouseholds in Idaho, Iowa, Kansas, Wyoming, Nebraska, Montana, the Dakotas, Oklahoma, and West Virginia Preferences: Heartland Communities enjoy their surroundings by camping, hunfishing, and growing vegetables and. They also do needlework; read country, home, and outdoor magazines; listen to country music and spend time with their pets. Their big-ticket purchases include American-made, large model cars, campers, satellite dishesoutboard motors, chains saws and other tools.

PAGE 112

103 Consumer Type 8A: Young Immigrant Families Demographic: These communities of Young Immigrant Families are family-oriented and nearly 78 percent are Hispanic. The population is young with a median age of 30.8. Their families are large, with an average family size of 3.4. Over 45 percent of thhouseholds are either single-parent or married-couples with children. Those are the ese ns in in the d education impede employment, which is below t of 5 often al products, children's aproceries. They similarities; the differences are ethnic, including Puerto Ricans in New York, CubaFlorida, and other Hispanics in the Southwest. Most of the population was bornU.S. and most speak Spanish at home. Socioeconomic: The median household income is $25,300 versus the U.S. median of $39,800. Lack of language skills an average. Unemployment and poverty are twice the national average. Nearly 25 percenthese households receive some form of public assistance. Over half of the adults aged 2and over did not finish high school. Young Immigrant Families account for only one percent of the total U.S. households and 1.2 percent of the U.S. population. Residential: Home states for Young Immigrant Families include New York 34 percent, Florida -33 percent, and New Jersey -13 percent. They live in apartments in high-rise buildings. Nearly two-thirds of householders are renters, double the nationrate. The median home value is $92,100. Preferences: Typically urban, most Young Immigrant Families use public transportation, not automobiles. They usually buy family necessities such as baby parel, over-the-counter medications, vitamins, and g

PAGE 113

104 splurge on videos, long-distance phone calls, cable TV, and theme park and casino visitsThey watch dramas, sitcoms, and awards shows on TV. Consumer Type 8B: Social Security Dependents Demographic: The profile of the Social Security Dependents market is straightforwaelderly, living alone. Their median age is 52.6 years. Nearly half of the householders are rd: ependents hese sus r the U.S. They live in low-rent, high-rise apartment buildings. These paying for necessities, little is left for saving ornd dramas on tainment, 65 years or older; almost 65 percent live alone. Higher proportions of adult children live in these households, possibly caring for their parents. These neighborhoods also incorporate group quarters, institutional and other quarters. Socioeconomic: With a population of 1.6 million, the Social Security D segment is not a large market, representing only 0.6 percent of the U.S. population. Thouseholds depend on supplemental income from government sources. Over half receive Social Security, and almost 20 percent on some form of public assistance. Their medianhousehold income is $14,600 compared with $39,800 for the overall U.S. Residential: Social Security Dependents are usually renters, almost 68 percent ver30 percent fo metropolitan neighborhoods are geographically distributed throughout the U.S., with large numbers in Pennsylvania, New York, and Florida. The median home value of $69,500 is 30 percent lower than the national figure. Preferences: Social Security Dependents do not own cars. After investing. They watch sitcoms, movies, cable, a TV; listen to jazz and gospel radio; read newspapers; and read specifically enterhome, and cooking magazines. Their purchases of denture cleaners, bifocals, vitamins, home permanents, and news tabloids indicate an older population.

PAGE 114

105 Consumer Type 8C: Distressed Neighborhoods Demographic: In Distressed Neighborhoods, single-parent households outnumber aother ho ny usehold type. Single-person households are second. The population is very f e e mployed, most work in part-time jobs, primarily in the service industry. f nd rk, rel, n g young, with a median age of 23.2 years. Over 50 percent are children under 18 years oage. The average family size of 3.5 persons is relatively large. Most residents are black. Socioeconomic: Representing one percent of US households, this small market is at thlow end of the socioeconomic scale. Approximately half of the households in this market are below the poverty level. Their median household income of $12,200 is 70 percent below the national level. Unemployment is well above the national rate. Among thoswho are e Residential: Distressed Neighborhoods are 95 percent urban, in the central cities of large metropolitan areas. Over 40 percent of the housing are multi-unit buildings. Most othese homes are rentals. Homeless shelters, street locations, and juvenile homes are fouin these neighborhoods. The largest numbers of these households are found in New YoIllinois, and Michigan. Preferences: Most Distressed Neighborhoods use public transportation. They splurge on fast food, cable TV and lottery tickets. They also buy childrens and womens appababy products, vitamins, and athletic shoes. They watch sitcoms, dramas, and sports oTV; listen to R&B and jazz music; and read soap opera, womens, fashion and parentinmagazines. Consumer Type 8D: Hard Times Demographic: Hard Times is a population of extremes; the very young and the elderlyThe dependency ratio of young (<15 years) and old (65 years) to the working age (15-64

PAGE 115

106 years) population is much higher than the U.S. ratio. Although nearly one-third of the households are occupied by singles, many are single parent or multigenerational, shared households. Their median age of 34.4 years is comparable to the U.S. figure. The vast majority of the population is black. Socioeconomic: The median household income of $16,800 is 60 percent below the national level. About 40 percent of the households are below poverty; one-fourth receiveincome from public assistance; 40 percent, from Social Security. Unemployment is consistently well ab ove the national level. Of those who do have jobs, over half work 75,000; the median value is $39,800. The vacancy rate of over 17 s. In their free time, they also enjoy listening to gospel and part-time, mainly in the service industry. Over half of the adult population has not completed high school. Residential: A 50-50 mix of renters and owners live in a 60-40 mix of single-family and multi-unit dwellings. Home value is low. Over 80 percent of owner-occupied housing units are valued below $ percent is high. Located in central cities, Hard Times can be found in most states. Alabama, Louisiana, and Mississippi have a disproportionate share of these neighborhoods. Preferences: Economic reality limits the purchasing power of Hard Times. Most of their budgets go for lifes basics, like rent and groceries. They splurge on lottery tickets and cable TV. They watch a lot of dramas, sitcom, and sports on television, but have oneof the lowest consumption rates for video rental R&B music and reading soap opera and tabloid magazines. Consumer Type 8E: Urban Working Families

PAGE 116

107 Demographic: Almost 40 percent of the Urban Working Families population is under the age of 20; their median age is 29.4 years. This is a family market with a high percentage of single-parent households. Also, economic necess ity forces a significant ork s, g units at less than $75,000. These neighborhoods are primarily located in urban areas r cable TV, baby products, an dramas and sitcoms number of adult children to live at home. Over 80 percent of the population in these neighborhoods are black. Socioeconomic: This market is the working poor. Employment is well above average, but unemployment and poverty rates are double those for the United States. Most of the w force is full-time, but over half of the households earn less than $30,000. The median household income of $28,500 is 25 percent lower than the U.S. median. Residential: Urban Working Families neighborhoods are older, pre-war residential developments of townhouses and small, multi-unit buildings. Renters outnumber owner45.5 percent to 41.4 percent. About 70 percent of specified owner-occupied housinare valued scattered across the U.S. Over 15 percent of the neighborhoods in the District of Columbia are Urban Working Families. Preferences: Urban Working Families have more discretionary income than the otheDowntown markets. They buy lottery tickets, take-out food, hair and skin care products, d children's clothing. They watch sports, on television; listen to jazz and R&B radio; and read weekly news, mens, women's, tabloids, and soap opera magazines. Group 9. Nonresidential Neighborhoods Consumer Type 9A: Business Districts

PAGE 117

108 Business districts with a residential population of fewer than 100 and a significandaytime t business population are included in this segment. Consumer Type 9B: Institutional Populations These neighborhoods are home to institutional group quarters, including prisons, juvenile detention homes, and mental hospitals. The residential population is small, except for the institution's staff and employees. Consumer Type 9C: Unpopulated Areas These areas are unpopulated and include parks, cemeteries, golf courses, or undeveloped land.

PAGE 118

APPENDIX B Number of Records: 46,923 Match Description CACI REPORT Geocoding Match Level Summary Street Match S 42,127 89.8% One Tract in ZIP Code T 0 0.0% 46,923 100.0% Match Description CACI assigns your customer records to one of four match levels. The best the Street Address or ZIP+4 level, the Geo-Fix option can assign a census with the largest population in the ZIP Code is appended to the record. the ZIP Code is built from only one census tract, then the match code (T) No Geocode, no match was possible below the county level. Post office boxes, rural routes or drops, or incomplete addresses can nnot be matched either. ZIP Code to Census Tract Correspondence Summary Census tracts, and in nontracted areas, block numbering areas (BNAs), are component blocks on ZIP Code boundaries. Expressed as centers of blocks. If the centroid of a block falls within the ZIP Code, it is included. Blocks are then aggregated, and the ratio of the block totals to tracts/BNAs is used to create the ZIP Code to census tract correspondence. Match Code Number Percent Dominant Tract in ZIP CodeZ 0 0.0% No Geocode X 4,796 10.2% match level is Street Address or ZIP+4. If a record cannot be matched at tract through a ZIP Code to census tract correspondence. The census tract This match is referred to as a Dominant Tract in ZIP Code (Z) match. If denotes only one tract in the ZIP Code. If an address is assigned a preclude a match. Addresses outside the U.S. ca assigned to residential ZIP Codes by overlaying the centroids of the latitude/longitude coordinates, centroids approximate the geographic 109

PAGE 119

110 Customer Geographic Summary Record Count Number of Records: 46,923 Top 20 States Georgia (13) 263 0.6% Texas (48) 177 0.4% California (06) 140 0.3% Maryland (24) 125 0.3% Ohio (39) 118 0.3% Alabama (01) 77 0.2% Tennessee (47) 59 0.1% Indiana (18) 53 0.1% Broward, FL (12011) 3,752 8.0% Palm Beach, FL (12099) 2,171 4.6% Orange, FL (12095) 1,612 3.4% Brevard, FL (12009) 1,034 2.2% Sarasota, FL (12115) 662 1.4% Polk, FL (12105) 522 1.1% Leon, FL (12073) Clay, FL (12019) St. Johns, FL (12109) 436 0.9% Okaloosa, FL (12091) 356 0.8% Record Count Florida (12) 39,293 83.7% New York (36) 244 0.5% New Jersey (34) 166 0.4% Virginia (51) 162 0.3% Top 20 ZIP Codes 32608 Gainesville, FL 3,844 32601 Gainesville, FL 1,905 4.1% 32603 Gainesville, FL 32606 Gainesville, FL Percent North Carolina (37) 135 0.3% Pennsylvania (42) 131 0.3% Illinois (17) 122 0.3% Michigan (26) 99 0.2% South Carolina (45) 87 0.2% Louisiana (22) 72 0.2% Connecticut (09) 61 0.1% Massachusetts (25) 55 0.1% Other States 488 1.0% Non-geocoded Records 4,796 10.2% 32653 Gainesville, FL 369 0.8 33071 Pompano Beach, FL 364 0.8 33176 Miami, FL 277 33594 Valrico, FL 221 0.5% 33624 Tampa, FL 200 0.4% 32308 Tallahassee, FL 19 32312 Tallahassee, FL 19 32223 Jacksonville, FL 183 0. Other Recor Top 20 Counties Alachua, FL (12001) 12,242 26.1% Dade, FL (12025) 3,057 6.5% Hillsborough, FL (12057) 1,889 4.0% Pinellas, FL (12103) 1,811 3.9% Duval, FL (12031) 1,530 3.3% Seminole, FL (12117) 1,205 2.6% Volusia, FL (12127) 679 1.4% Marion, FL (12083) 641 1.4% Lee, FL (12071) 563 1.2% Top 20 Metropolitan Areas Gainesville, FL (2900) 12,242 26.1% Fort Lauderdale, FL (2680) 3,752 8.0% Miami, FL (5000) 3,057 Jacksonville, FL (3600) 2,473 Melbrn-Titusvl-P.Bay (4900) 1,034 2.2 Sarasota-Bradenton (7510) 954 2.0 Ocala, FL (5790) 641 Ft.Pierce-P.St.Luc Tallahassee, FL (8 508 1.1% 444 0.9% Pensacola, FL (6080) Naples, FL (5345) Collier, FL (12021) 357 0.8% Other Counties 6,656 14.2%Non-geocoded Records 4,796 10.2% Panama City, FL (6015) 199 0.4% Other MAs 2,512 Records not in an MA 6,173 Copyright 2002 CACI East: (800) 292-2224 FAX: (703) 243-6272 11/12/2002 West: (800) 394-3690 FAX: (858) 677-5420 Internet: http://demographics.caci.com Percent 8.2% 32607 Gainesville, FL 2,311 4.9% 32605 Gainesville, FL 1,180 2.5% 1,122 2.4% 780 1.7% 32609 Gainesville, FL 403 0.9% % % 32779 Longwood, FL 288 0.6% 0.6% 33414 West Palm Beach, FL 228 0.5% 33067 Pompano Beach, FL 196 0.4% 5 0.4% 4 0.4% 33186 Miami, FL 190 0.4% 4% 33076 Pompano Beach, FL 181 0.4% ZIP Codes 30,691 65.4% ds with no ZIP Code 1,601 3.4% Tampa-St.Pete-Clearw (8280) 4,154 8.9% Orlando, FL (5960) 3,274 7.0% 6.5% 5.3% W.Palm Beach-B.Raton (8960) 2,171 4.6% % % Daytona Beach, FL (2020) 758 1.6% 1.4% Ft. Myers-Cape Coral (2700) 563 1.2% ie (2710) 537 1.1% 240) 534 1.1% Lakelnd-Winter Haven (3980) 522 1.1% 481 1.0% 357 0.8% Fort Walton Beach,FL (2750) 356 0.8% Atlanta, GA (0520) 179 0.4% 5.4% 13.2%

PAGE 120

ew tions, Political Periods, and The Wiley. Center for Information and Research on Civic Learning (2000) Fast Facts on Young 97. Erikson, R. S., Lancaster, T. D., & Romero, D. W. (1989). Group Components of the f Politics, 51(2), 337-346. Florida Department s.state.fl.us. Geodemographic Information Systems. Ground Truth. New York, NY: The Instrumental Rationality of Geodemographics Systems. Economic Geography Vol 71 No.2 p. 171-198. LIST OF REFERENCES Axelrod, Robert. (1972). Where the Votes Comes From: An Analysis of Electoral Coalitions. American Political Science Review 66 p. 11-20. Baker, Sunny,& Baker Kim (1993) Market Mapping: How to Use Revolutionary NSoftware to Find, Analyze, and Keep Customers. New York, NY: McGraw-Hill, Inc. Beck, P. A., & Jennings, M. K. (1991). Family Tradi Development of Partisan Orientations. Journal of Politics, 53(3), 742-763. Campbell, Angus, Converse, Miller, & Stokes. (1960). The American Voter. New York; Voters. November 2000. Center for Policy Alternatives (1997) Generation X in the News Themes from CPAs Youth Voices Echoed around the Country. Alternatives. Vol 5, No. 6 July-August 19Conservative Politics U.S. (2003) Political Affiliation and Demographics. www.usconservative.about.com. Curry, David J. (1993) The New Marketing Research Systems. New York, NY: John Wiley & Sons, Inc. Presidential Vote, 1952-1984. Journal o of State Division of Elections (2003) www.election.do Goss, Jon (1995) Marketing the New Marketing: The Strategic Discourse of Guilford Press. Goss, Jon (1995) We Know Who You Are and We Know Where You Live: The 111

PAGE 121

112 Green Donald, P., & Palmquist, Brad. (1994). How Stable Is Party Identification?. Political Behavior 16(December) p. 437-66. Hess, Robert D., & Torney, Judith, V. (1967). The Development of Political Attitudes in r Registration Act of 1993. Political Behavior, 20(2), 79-104. The Western Political Quarterly s & Elections August 2002. Stanley, Harold W., Bianco, William T., Niemi, Richard G. (1986) Partisanship and Science Review Vol. 80 No. 3 September 1986 p. 969-976. Stanley, H. W., & Niemi, R. G. (1994). Partisanship and Group Support over Time. In L. eader. New York: Garland Press. Thrall, Grant Ian (2001) Geopsychographics LSPs. Geospatial Solutions April 2001. Thrall, Grant Ian, Mecoli, Noelle. (2003) Spatial Analysis, Political Support, and Higher r Gap. Public Opinion Quarterly Vol. 63 No. 1 p. 62-89. s. Children. Chicago: Aldine. Highton, B., & Wolfinger, R. E. (1998). Estimating the Effects of the National VotePhillips, David J., Curry, Michael R (2002) Privacy and The Phenetic Urge: Geodemographics and The Changing Spatiality of Local Practice. Sanders, Arthur. (1988). The Meaning of Party Images. Vol. 41 No. 3 p. 583-599. Skaggs, David., Anthony, Adam (2002) Winning With Young Voters. CampaignGroup Support Over Time: A Multivariate Analysis. The American Political S. Maisel & W. Shade (Eds.), Parties and Politics in American History: A R Education Funding. Geospatial Solutions. July 2003. Trevor, Margaret, C. (1999). Political Socialization, party Identification, and the GendeWeiss, Michael J. (1988) The Clustering of America. New York, NY: The Tilden Pres

PAGE 122

the raphy in December 2000. She specializes in Geographic Information th BIOGRAPHICAL SKETCH Noelle A. Mecoli is seeking her Master of Arts in the College of Liberal Arts and Sciences from the Department of Geography. Noelle acquired a Bachelor of Arts in Department of Geog Systems, and has several years of experience in the professional work environment, including precinct redistricting for Alachua County. Noelle is currently employed wiJones, Edmunds, and Associates, Inc. as a GIS Analyst II. 113


Permanent Link: http://ufdc.ufl.edu/UFE0004480/00001

Material Information

Title: Using geographic decision information systems to support political endeavors at the University of Florida
Physical Description: ix, 113 p. ; ill.
Language: English
Creator: Mecoli, Noelle A. ( Dissertant )
Thrall, Grant Ian ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2004
Copyright Date: 2004

Subjects

Subjects / Keywords: Geography thesis, M.A
Dissertations, Academic -- UF -- Geography

Notes

Abstract: The Registrar’s Office of the University of Florida (on behalf of the University President’s Office) commissioned a proof-of-concept study to examine the political implications of funding for this public university, and to establish a method for reaching those significantly affected by state appropriations to the University: the parent’s of the students. Using Geographic Information Systems, a pilot study was conducted to assign the correct Florida State House Representative district to a student’s data record. The student’s declared permanent address is used to make the geographic assignment. In most instances, University of Florida students’ permanent address is their parents’ address. This allows enumeration of students’ permanent addresses by representative district, and the creation of databases for mailings to inform university constituents about legislative activities. Further segmentation of mailings is possible by differentiating between how the message would be presented based on measurements of the constituents’ geodemographic. This information is useful documentation to State Representatives as a measure of the importance of the University to their legislative district. This information is also useful for garnering support for higher education in Florida in general, and specifically higher education at the University of Florida. The study was extended to analyze student’ voting propensity. Voting propensity is frequency of voting and party affiliation. Voter registration data were compared and contrasted with students’ geodemographics (permanent address) and nationwide patterns of student voters. A procedure was thereby established for segmenting the population by propensity to vote, and by political affiliation. In combination with the students’ geodemographic, predictions of likely voting behavior can be made using this procedure.
Subject: Florida, geodemographics, higher, student
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 122 pages.
General Note: Includes vita.
Thesis: Thesis (M.A.)--University of Florida, 2004.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0004480:00001

Permanent Link: http://ufdc.ufl.edu/UFE0004480/00001

Material Information

Title: Using geographic decision information systems to support political endeavors at the University of Florida
Physical Description: ix, 113 p. ; ill.
Language: English
Creator: Mecoli, Noelle A. ( Dissertant )
Thrall, Grant Ian ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2004
Copyright Date: 2004

Subjects

Subjects / Keywords: Geography thesis, M.A
Dissertations, Academic -- UF -- Geography

Notes

Abstract: The Registrar’s Office of the University of Florida (on behalf of the University President’s Office) commissioned a proof-of-concept study to examine the political implications of funding for this public university, and to establish a method for reaching those significantly affected by state appropriations to the University: the parent’s of the students. Using Geographic Information Systems, a pilot study was conducted to assign the correct Florida State House Representative district to a student’s data record. The student’s declared permanent address is used to make the geographic assignment. In most instances, University of Florida students’ permanent address is their parents’ address. This allows enumeration of students’ permanent addresses by representative district, and the creation of databases for mailings to inform university constituents about legislative activities. Further segmentation of mailings is possible by differentiating between how the message would be presented based on measurements of the constituents’ geodemographic. This information is useful documentation to State Representatives as a measure of the importance of the University to their legislative district. This information is also useful for garnering support for higher education in Florida in general, and specifically higher education at the University of Florida. The study was extended to analyze student’ voting propensity. Voting propensity is frequency of voting and party affiliation. Voter registration data were compared and contrasted with students’ geodemographics (permanent address) and nationwide patterns of student voters. A procedure was thereby established for segmenting the population by propensity to vote, and by political affiliation. In combination with the students’ geodemographic, predictions of likely voting behavior can be made using this procedure.
Subject: Florida, geodemographics, higher, student
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 122 pages.
General Note: Includes vita.
Thesis: Thesis (M.A.)--University of Florida, 2004.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0004480:00001


This item has the following downloads:


Full Text












USING GEOGRAPHIC DECISION INFORMATION SYSTEMS
TO SUPPORT POLITICAL ENDEAVORS
AT THE UNIVERSITY OF FLORIDA
















By

NOELLE A. MECOLI


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ARTS

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Noelle Mecoli

































This document is dedicated to Cephid.

















TABLE OF CONTENTS

page

LIST OF TABLES ................ .. .................. .................. .........vi

LIST OF FIGURES ....................................... ........... vii

ABSTRACT ........................................... ............ viii

CHAPTER

1 INTRODUCTION ................... .......... .. ................... .......... .. .......... 1

2 LITERATURE REVIEW .................................................. ...............6

Beginning of Geodem graphics ............................................. ...............6
Use of Geographic Information Systems.................. .............................................8
Geodemographics in Direct Marketing ............................ ...............9
Student Voters .................................................12
P arty A affiliation ............... .................................................................. ...............16

3 PILOT STUDY ................ ....................... ..................... ...... .. ........ 19

Background...................................... .................. .............. .........19
Data and Methods..................................................22
Results and Discussion ................................................... ........27
Pilot study ............................................. ................ ........27
G geographic Patterns ......................................................................... ............ ..... 30
Targeting Students and Their Parents With LSPs ....................................... ...42

4 STUDENT VOTERS................ .....................45

Background ......................................................... ....................45
Research...................................... .................. ................ ........48
Data...................................................... ........50
Methods ........................................ .............................. ........51
Results and Discussion .......................................................................54
Student's Party Affiliation and Parent's LSP ....................................................58
Cross tabulations .................................... ........... .........58
T-Tests............... ....................... .............................................................59
Typical Party Demographic....................... ..........59










C conclusions ............................................ 60

5 CONCLUSIONS........ ......... ...................... ..........65

APPENDIX

A CACI ACORN LIFESTYLE SEGMENTATION....................... ........... 68

B C A C I R E P O R T ....................................................... 109

L IST O F R E F E R E N C E S ................................................................................................. 111

B IO G R A PH IC A L SK E T C H ...................................................................................... 113












































v
















LIST OF TABLES

Table page

3-1. Number of studnets per District............................................................ ........ 28

3-2. Student per LSP category ............. .... ......... .... ........31

3-3. Students by County ......................................... ...... ..... ...43

4-1. Student registered voter demographics........................................ ...............55

4-3. Cross tabulations between party affiliation and LSP ......................................... 61

4-5. T-Test Republican and Democrat ................................ ............... 62

4-6. T-Test Republican and Independent............................. .................. 62

4-7. T-Test Democrat and Independent................ .................... 62
















LIST OF FIGURES

Figure page

2-1. Campaign for young voters finding young voters......... .......................................15

3-1 Florida Legislative District boundaries House Plan H406H020...........................25

3-2 District boundaries with students' permanent addresses............... ...26

3-3. Database numbers for students and legislative districts................... ................. 30

3-4. LSP Types 1A and IB ............. ............ ........ 32

3-5. Thriving Immigrants and East Coast Immigrants .......................... ..........33

3-6. Y oung Im m igrant Fam ilies .............................................. ............... 34

3-7. W working C lass Fam ilies ............................................................ 35

3-8. Twentysomething .............. .. .... ............................... 36

3-9. W wealthiest Seniors ..................... .... ... ...... ................ 37

3-10. African American Populations.......... ......... ......... ................38

3-11. H ispanic and A sian Population ........................................................ 39

3-12. Age 18-29 Population.................................... ......... 40

3-13. A ge 65 and U p Population .............................................. ............... 41

3-14. Database numbers for permanent and student addresses .......................................43

4-2. 18 to 29 registered voters per House district.................................... ......................63

4-3. Numbers for student registered voters...................................... 64















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Arts

USING GEOGRAPHIC DECISION INFORMATION SYSTEMS TO SUPPORT
POLITICAL ENDEAVORS AT THE UNIVERSITY OF FLORIDA

By

Noelle A. Mecoli

May 2004

Chair: Grant Ian Thrall
Major Department: Geography

The Registrar's Office of the University of Florida (on behalf of the University

President's Office) commissioned a proof-of-concept study to examine the political

implications of funding for this public university, and to establish a method for reaching

those significantly affected by state appropriations to the University: the parent's of the

students.

Using Geographic Information Systems, a pilot study was conducted to assign the

correct Florida State House Representative district to a student's data record. The

student's declared permanent address is used to make the geographic assignment. In

most instances, University of Florida students' permanent address is their parents'

address. This allows enumeration of students' permanent addresses by representative

district, and the creation of databases for mailings to inform university constituents about

legislative activities. Further segmentation of mailings is possible by differentiating

between how the message would be presented based on measurements of the









constituents' geodemographic. This information is useful documentation to State

Representatives as a measure of the importance of the University to their legislative

district. This information is also useful for garnering support for higher education in

Florida in general, and specifically higher education at the University of Florida.

The study was extended to analyze student' voting propensity. Voting propensity

is frequency of voting and party affiliation. Voter registration data were compared and

contrasted with students' geodemographics (permanent address) and nationwide patterns

of student voters. A procedure was thereby established for segmenting the population by

propensity to vote, and by political affiliation. In combination with the students'

geodemographic, predictions of likely voting behavior can be made using this procedure.














CHAPTER 1
INTRODUCTION

The study was based on a pilot study authorized by the President's Office of the

University of Florida (UF) to assign a Florida State House representative to a student's

permanent address within UF's current student database. The University would like to

have the ability to carry out a letter campaign to the student's parents when an issue

concerning the university arose (such as funding for the University), and inform those

parents to contact their legislator or vote on that issue. This pilot study will provide a

method for the University of Florida administration to document the contribution of a

state university to each individual state legislator in addition to mustering political

support from those affected by state appropriations to universities: University of Florida

students and the parents of students attending the university.

To achieve this objective, this pilot study assigns to a student's data record, the

correct Florida State House Representative district. The student's declared permanent

address is used to make the geographic assignment. In most instances, University of

Florida students' permanent address is their parents' address. This allows enumeration of

students' permanent addresses by representative district, and the creation of databases for

mailings to inform university constituents about legislative activities. Further

segmentation of mailings is possible by differentiating how the message would be

presented based on measurements of the constituents' geodemographic or lifestyle

segmentation profiles (LSPs).









The procedures adopted to execute this pilot study are those from the large field of

Geographic Information Systems (GIS). GIS technology has been widely adopted across

the university, from urban planning, to forestry, to natural resources analysis and business

geography. This study applies GIS procedures and technologies from business

geography. Among the various types of calculations performed by business geographers

is deriving of the subject population's geodemographic. This is normally executed in the

context of evaluating a retail store's customer profile. In this pilot study, the University's

customers are its students and their parents. So, central to this study is the analysis of

students' geodemographics attending University of Florida.

Jon Goss (1995b page 24) gave a comprehensive definition for geodemographics:

Geodemographic works by collecting spatially referenced data on consumers,

constructing statistical models of identity, and mapping distributions of consumer

characteristics or types. A geodemographic system combines three essential components:

massive electronic data bases composed of public and private, individual and aggregate

records on consumer identity and behavior; Geographic Information Systems (GIS) that

provide the tools to analyze, locate, and graphically represent the spatial distribution of

consumer characteristics; and segmentation schemes that identify consumer types

through factor and cluster analysis of spatially referenced demographic and

psychographic data.

The basic principle underlying geodemographics can be summed in the phrase

"birds of a feather flock together" or "you are where you live." People with similar

characteristics or backgrounds tend to gravitate to the same neighborhoods. The term

geodemographic is also referred to as "lifestyle segmentation profile" (LSP), "clusters."









or "GD." Assigning a geodemographic to a student's permanent address may give an

indication of the type of lifestyle category the parents is in. Instead of sending letters to

over 36,000 addresses (number of students who have a permanent address based in

Florida), the University can cater letters depending on the geodemographic of the

permanent address. Geodemographics are used in political campaigns as a way of

soliciting voters so perhaps the same principle can be applied to this project. Weiss

(1988) said "knowing a neighborhood's political attitudes allows strategists to craft

appeals using different voices to reach the distinct clusters and "political surveys confirm

that people living in the same cluster make similar choices. This study also reviewed the

geographic patterns of the geodemographic of the permanent address in Florida noting

the differences and similarities of the different cluster types

Geodemographic clustering was originally designed to support firms in the direct

marketing industry (Curry 1993) Geodemographics is primarily used to uncover and

exploit new markets (Goss 1995), but it is also used for a wide range of application such

as fund raising, mail campaigns, and site location analysis. Those who have taken

advantage of using geodemographics find that it has the potential to eliminate

redundancy, generate savings, and reduce unsolicited promotional communications (Goss

1995). Geodemographic systems were provided by a number of companies such as

ACORN, by CACI Inc.; Cluster PLUS by Donnelley Marketing Information Services;

PRIZM, by Claritas Corporation; and MicroVision by Equifax.

This study also used the voter registration information from Alachua County

Florida. The University of Florida is located in Alachua County, and there is no

information regarding the number of registered student voters. If the number of









registered student voters is obtained, the university may find this data useful in their letter

campaigns. There are a number of students who have both their student address and their

permanent address located in Alachua County, so the university can also solicit their

students as a means of political support. The voter registration data contains information

on party identification. This may prove to be useful in catering letters for increased

political support.

Another aspect of this study was a brief statistical analysis of the number of

students who are registered to vote (Alachua County) yet have a different student address

than permanent address (i.e., student address in Gainesville, permanent address in Ft.

Lauderdale). This was be included to ascertain if there is a relationship between the

student's party identification and the parents (permanent address) geodemographic, and

to establish a method for future research in this area given a greater population. A

comparison of University of Florida registered student voters and research conducted by

The Campaign for Young Voters (www.campaginyoungvotesr.org, 2003) was also

examined to show any similarities within the two studies

Chapter 2 presents a review of literature relating to geodemographics, GIS used in

geodemographics, direct marketing campaigns that geodemographics provides benefits

to, and finally a brief look at information regarding registered student voters. Chapter 3

presents and discusses the data, methods, and results from the project for the University

of Florida. Chapter 4 presents and discusses the data, methods, and results from

incorporating the voter registration data. Finally, Chapter 5 summarizes the finding and

draws final conclusions.






5


Chapter 3 is based on an article published in Geospatial Solutions (July 2003) titled

Spatial Analysis, Political Support, and Higher Education Funding. Additional

information was then added to this article for the purposes of this thesis for the UF

graduate school and also as a report for the President's Office of the University of

Florida. Chapter 4 is a draft for a future article to be published in Geospatial Solutions.














CHAPTER 2
LITERATURE REVIEW

Beginning of Geodemographics

The term geodemographics came into light around the 1970s credited to Jonathan

Robbin, a social scientist who brought together computer power with marketing

methodology (Weiss 1988). He believed that consumers could be marketed directly by

grouping them in neighborhood-sized markets.

For geodemographics to be complied, two vital sources of information had to come

to fruition: the ZIP (Zone Improvement Plan) Code, and computerized files from the

United States Census. The ZIP Code allowed households to be divided into areas served

by a single post office. Each zip had a numbering system that broke up the divided areas

into about 800 households (Phillips & Curry 2002). In addition to the ZIP Code, the U.S.

Census contained a wealth of geographic information. The use of the GBF-DIME (now

the Tiger Line files) and the Census Block allowed for computerized maps to be

produced (Phillips & Curry 2002). Merging the ZIP Code files with the block values

from the Census allowed geographic coordinates for almost every mailing address in the

United States to be determined.

After the accomplishments that these resources could provide, companies such as

Claritas Corporation and CACI Marketing Systems put the computerized maps (that

contained a great deal of information) to use. But in the 1970s slow mainframe

computers were used to process information. In the 1990s, with the advent of greater,









cheaper, and faster computing power, the companies who produced geodemographics

could now offer their services to a larger number of potential clients.

These companies used public data (U.S. Census) with private data sources such as

credit-reporting bureaus and intercensus-year data updates (Thrall, 2001) to provide

geodemographic analysis to a number of companies in need of improving their target

marketing strategies and for site-location analysis. The companies who provide

geodemographics start with U.S. census data. The Census Bureau does not report the

data collected per household, but rather by the smallest unit, the census-block group.

These units are analyzed for different characteristics, and people living in similar block

groups are categorized together. Next is the statistical analysis. In this process, the

redundancy is removed, and then the different groups are clustered together. The results

divide the 250,000-plus census-block groups into a smaller number of geodemographic

clusters (Curry 1993).

The company used to provide geodemographics in this research is CACI Marketing

Systems. Their system is called ACORN, an acronym for A Classification of Residential

Neighborhoods. The company was founded in 1962, and they have since sold their

ACORN system to Environmental Systems Research Institute (ESRI) (www.esri.com,

2003), in January 2002, yet for now retain the name. There are 43 segments and a

descriptive name assigned to each segment (Appendix A). There are supergroups and

then subgroups. The supergroup describes a broad range of characteristics shared by all



1 1 Census blocks are areas bounded on all sides by visible features, such as streets, roads, streams, and
railroad tracks, and by invisible boundaries, such as city, town, township, and county limits, property lines,
and short, imaginary extensions of streets and roads. Generally, census blocks are small in area; for
example, a block bounded by city streets. However, census blocks in remote areas may be large and
irregular and contain many square miles.









the subgroups in a particular category that are then separated by specific characteristics

(Curry 1993).

Use of Geographic Information Systems

Geographic Information Systems (GIS) is a collection of computer hardware,

software, geographic data, and personnel designed to efficiently capture, store, update,

manipulate analyze, and display all forms of geographically referenced information

(ESRI, 2003). GIS is fundamental the development and applications of geodemographics.

Approximately 85% of corporate databases have geographic content (Goss 1995). This,

in turn, could lead to widespread use of GIS in many applications, particularly in

marketing and marketers have, in turn, taken advantage of the benefits this tool provides.

Using GIS makes geodemographics easier to analyze. GIS provides the spatial

analytical tools used to aggregate and assign spatially referenced consumer information

and to represent the results of analysis in visual form (Goss 1995). When GIS is used in

developing and analyzing geodemographics, it is called Geographic Decision Information

Systems (GDIS). Six benefits in using GDIS are discussed in a chapter of Ground Truth

written by Jon Goss (1995). The first benefit Goss describes is accurately locating an

address. This is accomplished with geocoding.

Geocoding of a customer address is the first step in determining where a particular

address is located in a neighborhood. To geocode an address simply means assigning a

latitude and longitude coordinate to an address that then could be shown spatially. This

process matches records in two databases: an address database and another database with

know geographic positions such as a street map. After the databases are "matched" the

coordinates are assigned to each record.









Other benefits include the aggregation of individual records into spatial units like

zip codes. This allows for the creation of new units such as trade areas, provides for the

reconciliation of data collected at differing scales, examine the geographical ties between

disparate data sets, as well as visual representation of data.

Another key component of this research and in compiling geodemographics is

relational databases. A relational database is the joining of several databases based on a

particular attribute. This makes the analysis of many data sources at one time simpler.

For this particular research, the student database is combined with a voter registration

database based on a name attribute and then an address attribute.

The databases and cluster analysis that generate geodemographics can be used by

marketers independently of a Geographic Information Systems (GIS), but the ability to

use spatial analysis within the database can increase the power and representation of the

market analysis (Goss 1995). Other benefits provided by Baker & Baker (1993) in

Market Mapping by using GIS is understanding spatial relationships between data are

locating customers quickly, the ability for maps to make statistical marketing data easier

to understand, and the ability to plan and track marketing campaigns is improved upon.

Using GIS functions such as geocoding, joining databases, and mapping, increase

the functionality of analyzing geodemographics.

Geodemographics in Direct Marketing

Geodemographics was developed with the intention of expanding a marketer's

possibilities in the area of direct marketing. With direct marketing, a market analyst

could find new customers and keep remaining customers. Direct marketing can also

locate new store locations and discover which locations were not doing well, as well as

discover trade areas. Direct marketing can be less expensive if the marketer only targets









what is needed and can influence the right segment of customers (Baker & Baker 1993).

For this research, the direct marketing approach of mail campaigns will be the focus. This

can be successful especially in the political arena.

In a chapter of Marketing The New Marketing by Jon Goss (1995), he explains that

using direct marketing, the marketer knows the characteristics and addresses then

matches these with clusters to identify potential customers to increase sales (market

penetration). From this, custom mailing lists can be complied to keep those customers.

Goss also illustrates that marketing depends upon defining difference between products

that resonate with the differences between consumers. Effective target marketing has

only been limited by the development of the quality and quantity of information on

consumers and the means in which to analyze it. This is why geodemographics can play

an important role in developing a profitable market strategy.

In The New Marketing Research Systems by Michael J. Curry (1993), he explains

how the beginnings of geodemographics in mail campaigns could be successful. Mail

order giant L.L. Bean used geodemographics to send their catalogs to a certain base of

customers. By developing specific clusters based on those who responded to their

catalog, a mailing list was created to target these groups of people. This proved to be

cost effective and increase their sales by targeting a certain group of consumers.

In Market Mapping by Baker & Baker (1993), a scenario is given for steps in

targeting a direct mail program. In this scenario, letters need to be catered to target a

direct mail program for a department store. A mail list is created to limit waste and

reduce costs, and only target specific groups who would more likely be potential

customers. The steps in developing this list include defining a need, acquiring the









specific data, perform the relevant applications in a GIS, define relationships within the

data, and finally make decisions based on the finding. This approach can be used in a

wide variety of applications to develop a mailing list depending on the needs of the

application. Using direct marketing becomes less expensive because the "shotgun"

approach common in mailing campaigns is eliminated. Using geodemographics in

mailing campaigns can be successful by influencing the right segment of customers.

Political campaigns also use geodemographics as a means for targeting potential

voters through direct mail. In Clustering of America by Michael J. Weiss (1988), a

chapter is devoted to the use of geodemographics in political campaigns. Weiss states

that clusters offered a way to reach nontraditional voters. Marketing issues to

neighborhoods rather than groups as a whole could possibly generate increased support.

Weiss also states that knowing a neighborhood's political attitudes allows strategists to

craft appeals using different voices to reach the distinct clusters. A successful use of a

highly targeted direct mail campaign using clusters is discussed in the book.

In 1983, the governor of Louisiana was losing support and his political advisors

sought the use of geodemographics provided by Claritas. They discovered three distinct

groups the governor's campaign could target through a mailing campaign. Writing one

letter (message) and sending to all three groups would not muster the support that they

needed to win the campaign. Three different letters were written in such a manor as to

appeal to each cluster. In the end the governor won the election with 63% of the vote.

These letters targeted a message in such a way as to fit the perspective of the clusters and

thus gain a winning edge.









Weiss discusses how neighborhood lifestyles are a natural way for predicting

political behavior. Since the clusters are based on multiple socioeconomic indicators,

they can also mirror voters. Weiss states that these clusters can explain why some issues

cross party lines, but these same issues remain true to the neighborhood lifestyle. He also

claims that geodemographics may prove to be practical rather then traditional voting

blocks.

Two other cases of geodemographics used in direct mailing campaigns are sited in

Clustering ofAmerica. When geodemographics first entered the political arena in 1978

in the state of Missouri, a random sample of the population was surveyed to determine

the clusters. The addresses were then linked to census block groups and the voters were

identified. Instead of using a mass media campaign the strategists opted for a mail

campaign as not to disturb the voters who were already on board. Different letters were

crafted for each of the clusters found. While still expensive, this proved to be a cost

effective method and generated the support that was needed. The other case involved

targeting clusters for the renovation of the Statue of Liberty. Letters were written to cater

to clusters such as first generation Americans where the Statue of Liberty is a symbol of

their new freedom and independence. A letter had to be crafted differently however to

those clusters where nationalism was strong and placed blame on immigrants for taking

jobs away.

Student Voters

Young voters are the largest block of unclaimed voters in the country and

candidates usually ignore this group due to their low voter turnout (Skaggs & Anthony

2002). Traditionally, it has been difficult to target this group of voters especially on the

national level, but targeting this group on the state or local level may prove to be easier.









Targeting this group can give opportunities to candidates in close races to gain a winning

edge and to build a base among young voters for future elections (Skaggs & Anthony

2002). An article from Alternatives (1997), a newsletter from the Center for Policy

Alternatives, suggests that businesses and market analysts are slowly coming to the

realization that young consumers are a large group that needed to be mass targeted if they

are to become future consumers of certain products. They propose that the politicians

due the same to appeal to the young voters so they can become loyal supporters in the

future.

The Campaign for Young Voter (CYV) (2003) is a project of the Center for

Democracy and Citizenship at the Council for Excellence in Government funded by a

grant from The Pew Charitable Trusts. They are a non-partisan, non-profit organization.

The Campaign for Young Voters (CYV), formerly known as the Young Voter Initiative,

assists candidates for public office in their efforts to reach out and engage younger voters.

Drawing on extensive field research concerning young adults' views about politics,

elections and government, CYV publishes a Toolkit and suggested campaign practices

and materials to assist candidates at all levels in dealing with young adults about political

participation and voting. The CYV provides a wealth of information on polls, studies,

and news links as well as links to data on voting and registration provided by the U.S.

census (www.census.gov, 2003) and the Federal Election Commission (www.fec.gov

2003).

The information found on these web pages will be used throughout this research for

comparison purposes to the data complied about University of Florida students and voter

registration as well as general information about the overall data on young voters.









Geodemographics are not widely used in the political community as a means of

generating support (Weiss 1988). General demographics are more often used to target

certain communities where support is known to be. The CYV complied there own

"demographic" on young voters categorizing them as likely young voters, potential

young voter, and unlikely young voters (Figure 2-1).

The CYV has also complied data on likely demographics by party affiliation. They

have found that young adults identify themselves almost equally as Republicans,

Democrats, or Independents/third party. In comparing Republican vs. Democrats, they

have found that they are divided almost equally as the general registered voting

population is. The demographics by party are shown in Figure 2-2, and this information

will be used to show any comparisons between the UF student registered voter

populations.

Research from various sources including the U.S. Census and the Federal Election

Commission was complied by CIRCLE (The Center for Information and Research on

Civic Learning and Engagement) to show facts on young voters on the national level.

CIRCLE promotes research on the civic and political engagement of Americans between

the ages of 15 and 25. In comparing registered voter information on the state level

(Florida) with the national information, Florida has 74.3% (U.S. Census) of eligible

citizens to vote registered to vote which follows closely to the national level of 78.1%.

The number of 18-25 year olds registered to vote on the national level is 59.9%, but in

Florida this number is lower at 39.9%. Males registered to vote on the national level

between ages 18-25 is 40.3% and in Florida the percent is about the same at 40.0% while

the female percent show a difference at the national level 44.3% and in Florida at 39.9%.













Likely young voters
This group is the obvious place to start in developing a base of support among young voters.

They typically are already more politically active and have stronger partisan leanings. They are presidential election voters, but in
other elections their participation drops dramatically. If you can connect with them, they will turn out for you on
Election Day, and some may volunteer on your campaign. They are likely to be knowledgeable about issues facing
their communities, and have the motivation and time to work on them.

Attitudes: Likely Young Voters are more politically oriented than their peers and trust government far more. They expect
candidates to pay attention to their concerns and to want their vote. They view elections as the way for people to have a voice in
their communities. They understand the power of voting-they are almost all registered to vote, say voting is important, and think
their vote counts as much as anyone else's. They also see themselves as able to make a difference in solving problems in their
communities.

Demographics: Likely Young Voters have parents who took them to the polls as children, who voted themselves, and who
discussed political issues at home far more than average. These young adults are also more educated, and the more education they
have, the more likely they are to vote. They are likely to be full time students and to identify with Democrats or Republicans more
than with third parties. Many attend religious services regularly, and many of those identify as bom-again Christians.

Potential young voters
This group is essentially unclaimed and available.

In many ways, Potential Young Voters are similar to Likely Young Voters, except they are typically community volunteers rather
than political activists. They are not likely to be campaign workers for you, but will register and turn out if you and your party work
to reach and persuade them. Their hot button is community involvement-and the more you can make the case that you
are a community problem solver and have been a community volunteer, the better chance you have of getting their votes. They are
much more independent and less partisan than Likely Young Voters, making them a harder sell, but still workable.

Attitudes: Potential Young Voters are frequent volunteers in their community. They recognize that elections are a way for
people to have a voice, but are not likely to be interested in volunteering for a political campaign. While they are as trusting of
government and candidates as likely voters, they are more ambivalent about voting. These potential voters are slightly more likely
than average to believe that voting is important, to be registered to vote, and to believe their vote counts. They also have lower
expectations about receiving attention from candidates.

Demographics: Compared to Likely Young Voters, Potential Young Voters are far less likely to come from politically oriented
families, especially measured by their discussion of political issues. Somewhat higher percentages of these voters are women and
attend religious services regularly.

Unlikely young voters
This group is hard to reach, and then only with extraordinary time and resources. Polar opposites of Likely and Potential Voters,
these young adults are difficult to engage in politics or voting. They are distrustful when candidates do reach out to
them, and tend to come from families with little political orientation or involvement.

Attitudes: Unlikely voters are disengaged from and distrustful of the electoral process. They are often not registered to vote and
don't believe voting is important. A large majority of this group never volunteers, and is unlikely to believe they can make a
difference in solving their community's problems. They're distrustful-of government, of other people, of candidates' motives, and
of the importance of their vote. They feel candidates do not pay attention to them and do not want their votes.

Demographics: Unlikely voters come from families whose parents didn't vote, didn't take them to vote, and didn't discuss
politics with them. They are also much less likely to have a significant affiliation with an institution-a college, a church, or a
political party. This group is generally politically independent and non-ideological, and less likely to be conservative.


Figure 2-1. Campaign for young voters finding young voters







Young Adult Republicans
Younger: 18-22 years old, especially men
Full time students especially non-working students
Self-described conservatives
White men
Attend religious services regularly
From politically-oriented families
Concerned with higher education costs
Occasional volunteers

Young Adult Democrats
Older 23-25 years old
College degree or beyond
Work part time
Self-described liberals
African-American or Hispanic, especially African American women and
college graduates
Figure 2-2. Likely demographics by party for young adults


Examining the various methods in which geodemographics, GIS, and information on

young voters are utilized can prove to be beneficial for this research. The university can

take this information and apply it to their direct mail campaign to generate support for the

university on numerous issues.

Party Affiliation

In this research, a correlation may be established between students party affiliation

to parents (student's permanent address) geodemographic. The geodemographics used

for this study currently do not make political assumptions. Instead, assumptions were

made about the different geodemographic types and what party the group might be more

statically associated with based on several studies. Also recall the political assumptions

that can be drawn from geodemographics based on Michael J. Weiss's work in Clustering

ofAmerica (1988).

Statistical comparisons shall be made in this study to examine the student's party

identification and the permanent address geodemographic of that student. The









assumption is made that a student's party identification shall reveal a correlation to a

particular geodemographic type based on a student's party affiliation corresponding to

that of the parents. Several studies have examined this issue. In The American Voter

(1960), it is believed that party affiliation is obtained early in adolescence and remains

stable throughout the life of the child. Another study suggest that partisanship does not

change significantly after adolescence and party affiliation is nearly complete after

leaving the parental household (Hess and Torney, 1967). In more recent studies

conducted on the subject it was found that party affiliation remained stable over time and

variation in party was small on a year to year basis (Green and Palmquist, 1994).

Currently no party identification information exists with the geodemographics from

CACI. Only assumptions can be generated on which party would most likely fit to a

particular geodemographic. Since geodemographics are based on socio-economic

indicators, it may be assumed that a party can be assigned to a geodemographic based on

studies conducted by demographic groups (Weiss, 1988).

The study of party identification and groups is dominated by Robert Axelrod. In a

quadrennial report in the American Political Science Review titled Where the Vote Comes

From, Axelrod examines party affiliation by several bivariate factors such as race and

religion. More extensive multivariate studies have been conducted to determine party

support by differing groups. One such multivariate study by Erikson, Lancaster &

Romero (1989), examines this type of analysis of group influences on party identification

in presidential voting. Among their results they find that "Urban" voters show pro-

Republican influence although large urban population tend to vote for Democrats and for

youth to vote Democratic and college educated to vote more Republican are muted. They






18


also find that region plays a significant role in party identification. Another study

examines group support for political parties by multivariate techniques. This study titled

Partisanship and Group Support Over Time: A Multivariate Analysis (1986) finds groups

that tend to support the Democratic Party and those who have moved away from the

party.














CHAPTER 3
PILOT STUDY

Background

Public education in the United States is highly political. It is political because

expenditures on higher education comprise one of the largest categories of discretionary

expenditures that the state has to choose how to spend tax dollars. Most expenditures

made by a State government are comprised of entitlements, the amount of which are

difficult, if not impossible, to change from year to year. State legislatures choose the

amount of funds that go into higher education, versus say kindergarten through 12th grade

education, community college education, special education programs, prison systems,

and a host of special projects. When state tax revenue declines, State Representatives

choose which discretionary programs will have a reduction in revenue. When state tax

revenue increases, State Representatives choose which discretionary programs will be

beneficiaries.

Of course, higher education in general, and University of Florida in particular,

lobby to minimize reductions, and increase appropriations. Numbers support lobbying

efforts. It is believed that the greater is University of Florida constituency within a

Representative's district, the more attention will be paid to that constituency. That

attention hopefully translates into revenue to the University. Most colleges and

universities in the State University System are commuter campuses. Students commute

from their permanent address to the campus. A State Representative has a fairly accurate

perception of the importance of the campus to his or her district. University of Florida









however is instead a residence campus. Students for the large part come from other

places reside on or near campus while attending classes, and after they complete their

education at University of Florida they leave the campus vicinity. The residence campus

provides an opportunity for a student to become completely immersed in higher

education, as well as an intermediate step toward independence from parents. However,

this same remoteness comes at a price. Because the campus is remote from most urban

areas in the State, and therefore remote from most State Representatives districts, it is

easy for the State Representative to not consider the importance the University of Florida

is to their district. The numbers that this research generates are used to document the

importance of University of Florida's campus in Gainesville to Representative districts

throughout the state.

Geographic Information Systems technology allows for more than the placement of

pins on a map representing the locations of students' permanent addresses, and more than

the enumeration of counts of those pins by Representative district. GIS allows for the

evaluation of which types of people live where. This is done by way of assigning lifestyle

segmentation profiles (LSPs) to those individuals who comprise our database.

Instead of sending over 36,000 addresses (number of students who have a

permanent address based in Florida, both undergraduate and graduate students), with a

single message, the University can send letters with a message targeted to specific

geodemographic characteristics, including legislative district and LSP. Geographers are

familiar with applying this same technology to political campaigns as a means to sway a

voting population. Weiss (1988, p.24) writes in The Clustering OfAmerica, that

"knowing a neighborhood's political attitudes allows strategists to craft appeals using









different voices to reach the distinct clusters," and "political surveys confirm that people

living in the same cluster make similar choices."

The basic principle that underlies the creation and use of an LSP is Tobler's first

law of geography: all things are related, but near things are more related than distant

things. That applies to characteristics of people. People with similar demographic and

consumption characteristics, generally reside in close vicinity to one another. LSPs are

derived by analyzing hundreds of characteristics of neighborhoods and individuals, to

create a single composite measurement that summarizes the population by location.

Statistical methods used to derive LSPs include principal components analysis,

correlation matrices, and a variety of graphical methods, as well as various cluster

techniques. The derivation and use of LSPs in geography date to the pioneering research

of Brian J. L. Barry in the 1960s (Thrall, 2002 p. 29). Today, an analyst need not

calculate their own LSP database. Instead, LSPs are part of commercial GIS technology.

Not only are commercial LSP databases a time saver for the analyst, but there are analytic

and communicative advantages to using standardized LSP categories. Also,

contemporary LSP databases are improved by the use of proprietary databases not

available to most analysts, such as individual expenditure and credit information.

If the business geographer knows an address, that address will hopefully have a

ZIP+4 code. The GIS technology can assign a ZIP+4 code to a data record with only a

fragmentary address. That ZIP+4 code can be matched with one in a database of ZIP+4

codes and LSPs. Using relational database management, the appropriate LSP is then

assigned to every address record in the database. At the same time, the address will be

used in GIS "address matching procedures" to calculate a latitude-longitude coordinate.









The smaller is the geographic area, the more likely will the LSP code accurately correlate

with the individual household, and the more accurate will be the geographic positional

assignment. Therefore, the database to be analyzed can be populated with LSP codes as

well as geographic coordinates. SQL operations can then be used to select records by

LSP, and those student permanent addresses can then be used for specific mailings.

The pilot study data analysis begins with assigning to a student's data record, the

correct Florida State House Representative district. The student's declared permanent

address would be used to make the geographic assignment. In most instances, University

of Florida students' permanent address is their parents' address. This allows enumeration

of students' permanent addresses by representative district, and the creation of databases

for mailings to inform university constituents about legislative activities. Further

segmentation of mailings can be made by differentiating between how the message would

be presented based upon measurements of the constituents' lifestyle segmentation

profiles (LSPs). This information could then be used to document to State

Representatives one very clear measure of the importance of the University to their

legislative district, and an effective method garnering support for higher education in

Florida in general, and specifically higher education at University of Florida.

In addition to this pilot study, supplementary information was provided about the

students geographical patterns based on their permanent address geodemographic and

that data was compared to 2000 U.S. census data. Other mean to target parents and

students based on their LSP is also provided.

Data and Methods

The University of Florida student database is a private database, provided to us by

the Registrar's Office of the University of Florida. It is not available to the general









public. The student database for enrollment of fall 2002 included a variety of data fields,

including permanent address.

ESRI BIS (formerly CACI/Coder Plus using A Classification of Residential

Neighborhoods (ACORN) Software) was chosen to assign ZIP+4 codes, geographic

coordinates and LSPs to student records based upon permanent address. ESRI's BIS

product assigns LSP codes (Appendix A) and geographic location coordinates to

individual data records at the geographic scale of a ZIP+4 code, five digit ZIP code, or

census tract. The company protects the process involved in providing geodemographics,

but they do provide some general methodology. Over 220,000 neighborhoods are

assigned to one of 43 market segments based on similar characteristics from census block

groups such as household type and occupation. Other determinants are also analyzed

including consumer behavior. CACI then uses statistical methods such as "principal

components analysis, correlation matrices, and graphical methods" in addition to a

variety of cluster techniques to create the segments. A K-means algorithm followed by

Ward's hierarchical minimum-variance method is used to group the clusters which are

then validated by using external variables to check the market segments. The ACORN

system is also linked to updated consumer survey data to ensure validity.

In this proof of concept study, data records were assigned geographic coordinates

using only the more accurate "street address matching procedure." ESRI Business feature

was used which allows assigning geographic coordinates based upon best of street

address, ZIP+4, and ZIP code. The University Registrar's office was concerned that an

address might be assigned to the wrong legislative district if geographic positioning with

less accuracy than street address matching was used. A student whose permanent address









was outside the State of Florida was dropped from the database used for the analysis.

Also, students' permanent addresses that were not assigned latitude and longitude

coordinates, such as those with post office boxes, were deleted from the database. The

students were then mapped using ArcMap ArcGIS 8.2 by ESRI (www.esri.com).

The Florida Department of State, Division of Elections (2003) provided contact

information for each state legislator. Individual legislator information was updated to

include the winning candidate in the 2002 general election.

The State of Florida Legislative District boundaries House Plan H406H020 was

approved in July 2002. That boundary file was obtained from the Legislature of the State

of Florida. The boundary file was in ESRI's shapefile format, decimal degrees North

American Datum 1983 (NAD 83). There are 120 House districts in the State of Florida.

A map of the districts is shown in Figure 3-1.

The boundary file included district number as an attribute. The legislator contact

information database included the legislator's name, title, district number, and legislative

office address. The common district number allowed the contact information to be joined

to the district polygon. Next, the appropriate legislative district and contact information

was assigned to each student record using the point-in-polygon operation (Spatial Join).

First the students' permanent addresses are mapped, as shown in Figure 3-2.

Figure 3-2 also includes the legislative district polygon layer. The point layer (student

addresses) and polygon layers (legislative districts) can be spatially joined, with attributes

of objects from one layer assigned to objects in the other layer.


















JACKSONVILLE


Jacksonville



101





Tampa Orlando Area



118



20
23


TAMPA

ST. PETERSBURG


State of Florida
Legislative House Districts


NAPLES


FORT LAUDERDALE
SMI AMI

)MIAMI BEACH


KEY WEST


Source: Florida State Legislature


Figure 3-1 Florida Legislative District boundaries House Plan H406H020


South Florida






























TAMPA


ST. PETERSBURG


State House Districts with
Student's Permanent Address







'Only Students who place their permanent
address in Florida are shown
Source. University of Florida Registrar's Office Fall 2002
Florida Legislature July 2002


LAUDERDALE


Figure 3-2 District boundaries with students' permanent addresses









Maps were also included for comparison purposes using the student's permanent

address geodemographic and the 2000 U.S. Census data on the track level to show the

general distribution of certain population in the State of Florida including that of African

American, Hispanic and Asian, 18 to 29 years olds, and the 65-and-over age groups. The

process used to discover the number of students who have the same student address and

permanent address was to first copy the original file of students whose permanent address

is located in Florida. These two files can then be merged on the student address and the

student's permanent address, only keeping the records that matched. The number records

that were discarded are the students with differing addresses. This is information is

useful as a means to target parents and students in differing ways; based on LSP

geographic patterns and location of permanent address if it differed from the student

address.

Results and Discussion

Pilot study

There are a total of 46,923 in the student database, and 42,567 of the students list

an address in Florida as their permanent address. There were 6373 or 15% that were

deleted because they did not geocode to street address, leaving 36,194 students to be

assigned to a legislative district (Figure 3-3).

Table 3-1 shows the 120 legislative districts in the State of Florida, and the number

of University of Florida students whose permanent address is in that legislative district.

The districts with the highest numbers of students (permanent address) are District 23

(Gaine'7'il/l' Ocala), 22(G/tine''//i'e Ocala), 96(Ft. Lauderdale), 19(Jacksonville), 97(Ft.

Lauderdale), and 37(Orlando).











Table 3-1. Number of studnets per District


District
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40


# of Students
128
132
176
297
60
165
61
53
437
35
389
71
260
119
89
303
255
392
632
225
103
5075
5358
425
153
300
114
223
156
322
356
239
321
294
311
114
598
193
105
316


District
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80


# of Students
345
130
160
124
192
109
405
500
126
277
271
223
122
227
58
221
308
85
42
352
153
254
186
77
126
125
179
192
261
277
103
97
149
222
154
243
105
248
125
238


District
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120


# of Students
257
300
414
96
375
138
303
128
83
257
283
115
158
135
258
658
627
416
211
322
241
96
153
135
137
197
122
149
61
105
135
140
97
340
385
221
321
100
244
135









The Registrar's Office of the University of Florida on behalf of the University

President's Office commissioned this proof of concept study. The purpose of the study is

to provide a tool whereby the University could bring to the attention of State legislators

the importance of the university in serving the educational needs of students of their

district and state region. Public education in the United States is highly political, because

it is a large part of a state's discretionary budget, and many constituencies are vying for

the same discretionary dollar. While the University registrar has not at this time applied

the methods and procedures demonstrated in this work, the work serves as a

demonstration of concept.

Higher education is important to the success of the next generation to be educated.

Higher education requires considerable and sustained funding if it is going to serve its

mission. GIS has been demonstrated here as one means to generate the information

necessary to receive the political support to further the mission of higher education

With 6,373 records not able to geocode, due to bad addresses or a Post Office Box,

the Registrar's Office loses the ability to reach these students and/or their parents, but

there is a considerable amount of records left to send letters to. A concern of the

Registrar's Office was how accurate is the spatial join and how would they know if the

student's permanent address was placed in the correct legislative district. The accuracy

in which the software geocodes the address should place the address within the correct

district, with only potential problems arising with those addresses that are on the border

of two districts. A buffer could be included in the process of a desired distance to

eliminate this problem, yet that would exclude a number of records.









Geographic Patterns

A LifeStyle Segmentation Profile was assigned to each student record for their

permanent address and out of 44 possible LSPs, UF students fit into 31 categories. The

number of students per LSP is shown in Table 3-2. Patterns emerge when looking at the

student's permanent address just in the State of Florida per LSP type. Most of the LSPs

are distributed rather evenly throughout the state, but significant geographic patterns can

be seen with specific types ofLSPs.

LSP type lA Top One Percent, lB Wealthy Seaboard Suburbs, are highly

concentrated in South Florida with some in the major cities in Florida such as,

Jacksonville, Orlando, and Tampa (Figure 3-4). Type lA and IB are the wealthiest of the

LSP types. Their residential classification puts these types in the suburbs of major urban

areas.

The other types concentrated in South Florida and other major urban areas are the

immigrant categories. Type 2C Thriving Immigrants, and 6A East Coast Immigrants are

highly concentrated in South Florida with some found in other major cities throughout

the state (Figure 3-5 ). CACI describes these types as living in urban areas with a large

percent being of Hispanic origin. Another immigrant category, 8A Young Immigrant

Families, is almost solely concentrated in South Florida (Figure 3-6).


Figure 3-3. Database numbers for students and legislative districts









Table 3-2. Student per LSP category

Acorn Description # %
lA Top One Percent 446 1.2
IB Wealthy Seaboard Suburbs 332 0.9
1C Upper Income Empty Nesters 1614 4.5
ID Successful Suburbanites 2353 6.5
IE Prosperous Baby Boomers 4876 13.5
IF Semirural Lifestyles 2187 6.0
2A Urban Professional Couples 2428 6.7
2B Baby Boomers With Children 1160 3.2
2C Thriving Immigrants 275 0.8
2E Older, Settled Married Couples 552 1.5
3B Enterprising Young Singles 953 2.6
4A Retirement Communities 840 2.3
4B Active Senior Singles 797 2.2
4C Prosperous Older Couples 1091 3.0
4D Wealthiest Seniors 2618 7.2
4F Senior Sun Seekers 1101 3.0
5A Twentysomethings 983 2.7
5B College Campuses 5830 16.1
6A East Coast Immigrants 285 0.8
6B Working Class Families 176 0.5
6C Newly Formed Households 843 2.3
7A Middle America 637 1.8
7B Young Frequent Movers 1456 4.0
7C Rural Industrial Workers 87 0.2
7E Small Town Working Families 130 0.4
7G Heartland Communities 346 1.0
8A Young Immigrant Families 744 2.1
8C Distressed Neighborhoods 24 0.1
8D Hard Times 133 0.4
8E Urban Working Families 284 0.8
9C Unpopulated Areas 13 0.0

















.JACKSONVILLE


PENSACOLA












Geographic Patterns
Type 1A and 1B








A 1 =1 Student (Permanent Address)


TAMPA
ST PETERSBURG


NAPLES


r ORT LAUDERDALE
SMIAMI
IMIAMI BEACH


Source. University of Flonda Registrar's Office Fall 2002
CACI GeodemographicData


Figure 3-4.LSP Types 1A and 1B


















.JACKSONVILLE


PENSACOLA













Geographic Patterns
Type 2C and 6A









A 1 =1 Student (Permanent Address)


TAMPA
ST PETERSBURG


NAPLES


FORT LAUDERDALE

MIAMI
NMIAMI BEACH


SoLice. University of Flonda Registrar's Office Fall 2002
CACI GeodemographicData


Figure 3-5.Thriving Immigrants and East Coast Immigrants


















.JACKSONVILLE


PENSACOLA













Geographic Patterns
Type 8A









A 1 =1 Student (Permanent Address)


TAMPA
ST PETERSBURG


NAPLES


FORT LAUDERDALE

MIAMI
,MIAMI BEACH


SoLice. University of Flonda Registrar's Office Fall 2002
CACI GeodemographicData


Figure 3-6.Young Immigrant Families


















.JACKSONVILLE


PENSACOLA













Geographic Patterns
Type 6B









A 1 =1 Student (Permanent Address)


TAMPA
ST PETERSBURG


NAPLES


FORT LAUDERDALE

MIAMI
MIAMI BEACH


SoLice. University of Flonda Registrar's Office Fall 2002
CACI GeodemographicData


Figure 3-7.Working Class Families


















.JACKSONVILLE


PENSACOLA













Geographic Patterns
Type 5A









A 1 =1 Student (Permanent Address)


TAMPA
ST PETERSBURG


NAPLES


FORT LAUDERDALE

MIAMI
MIAMI BEACH


Source. University of Flonda Registrar's Office Fall 2002
CACI GeodemographicData


Figure 3-8.Twentysomething
















.JACKSONVILLE


PENSACOLA -A











Geographic Patterns ST
Type 4D








A 1 =1 Student (Permanent Address)


r ORT LAUDERDALE
l MIAMI
MIAMI BEACH


Soiuce. University of Flonda Registrar's Office -Fall 2002
CACI GeodemographicData


Figure 3-9.Wealthiest Seniors


















JACKSONVILLE


African American
Population












1 Dot= 750 People
Source U S Census Bureau 2000


TAMPA

ST PETERSBURG


NAPLES


00




FORT LAUDERDALE

MIAMI
MIAMI BEACH


KEY WEST
s^y^"~


Figure 3-10. African American Populations

















JACKSONVILLE


PENSACOLA









Hispanic and Asian
Population












1 Dot= 750 People
Source. U.S. Census Bureau 2000


TAMPA
ST PETERSBURG


FORT LAUDERDALE

MIAMI
MIAMI BEACH


KEY WEST *-


Figure 3-11. Hispanic and Asian Population


















JACKSONVILLE


AGE 18-29













1 Dot= 2,500 People
Sources U S Census Bureau 2000


TAMPA

ST PETERSBURG


NAPLES


0




FORT LAUDERDALE

MIAMI
MIAMI BEACH


KEY WEST
o


Figure 3-12. Age 18-29 Population


















JACKSONVILLE


PENSACOLA










AGE 65 and UP













1 Dot= 1,500 People
Sources U S Census Bureau 2000


TAMPA

ST PETERSBURG


NAPLES


[FORT LAUDERDALE
MIAMI
MIAMI BEACH


KEY WEST
acPf^*


Figure 3-13. Age 65 and Up Population









Targeting Students and Their Parents With LSPs

Knowing how many students' permanent addresses are in each district and learning

the different LSP types that make up these addresses can lead the University to cater their

direct mail campaign to whatever group of addresses that fit their needs.

After merging the student address with the student's permanent address (State of

Florida only), it is discovered that 18,850 students have the same student address and

permanent address. This leaves 17,344 records that have different addresses and in these

cases, a letter would have a better chance to go to a parent or grandparent of the student if

that is indeed the focus of the University. The LSPs of 17,344 records could be analyzed

and then catered to depending on the issue the University is trying to raise (Figure 3-14).

Another way to limit or cater the number of mailings is to target areas throughout

the state with a high concentration of permanent addresses. According to the CACI

Geographic Summary (Appendix B), the top 8 counties where students have their

permanent addresses are located in the major cities of Florida. The South Florida

counties (Palm Beach, Broward, and Dade) have the largest concentration of addresses

next to Alachua, then Hillsborough, Pinellas, Orange, and Duval (Table 3-3, Figure 3-15)

With the completion of this research, the University will also have the ability to

target just the students who are registered to vote in Alachua County (Chapter 4). Since a

large percentage of students put their permanent address in the county of Alachua (30%),

and then further reducing this number by finding the students registered to vote, will

permit the University to target students who may be more inclined to vote for issues of

concern to the student and to UF in the districts that represent Alachua County. Recall,

the largest number of students (permanent address) are located in Districts 23 and 22.






































Figure 3-14. Database numbers for permanent and student addresses


Table 3-3. Students by County

COUNTY COUNT %
Alachua 12,424 26.1
Broward 3,752 00.8
Dade 3,057 6.5
Palm Beach 2,171 4.6
Hillsborough 1,899 4.0
Pinellas 1,811 3.9
Orange 1,612 3.4
Duval 1,530 3.3


46,923 Total Records



42,567 Florida Permanent Addresses

36,194 Assigned
Legislative Districts
(15%NoGeoode)













JACKSONVILLE


PENSACOLA


ST PETERSBURG


FORT LAUDERDALE

MIAMI
MIAMI BEACH


Figure 3-15. Florida counties and major cities














CHAPTER 4
STUDENT VOTERS

Background

Young voters are the largest block of unclaimed voters in the country. Candidates

for political office generally ignore young voters group due to their low voter turnout

(Skaggs & Anthony 2002). It has been difficult to target this group of voters at the

national level. It is believed that this group of voters is hard to target because they have a

low voter turnout (Highton & Wolfinger, 1998). Targeting this group at the state and

local levels is feasible. Targeting the young voter group can give opportunities to

candidates in close races to gain a winning edge, can serve to build an ongoing

consistency of young voters in future elections, and can contribute to building a base of

voters that will stay with the candidate or political party as the young voters age (Skaggs

& Anthony 2002).

An article from Alternatives (1997), a newsletter from the Center for Policy

Alternatives, presents a case that businesses and market analysts are slowly coming to the

realization that young consumers are a large group that need to be mass targeted if they

are to be relied upon to become future consumers of a business's products. Alternatives

proposes that politicians for the same reasons appeal to the young voters so that those

young voters become loyal supporters in the future. In practice today, geodemographics

are not widely used in the political community for planning or strategy, including actively

gaining political support (Weiss 1988).









The Campaign for Young Voter (CYV) (2003) is a project of the Center for

Democracy and Citizenship at the Council for Excellence in Government funded by a

grant from The Pew Charitable Trusts. Both The Pew Charitable Trusts and CYV are

non-partisan, non-profit organizations. The Campaign for Young Voters (CYV), formerly

known as the Young Voter Initiative, assists candidates for public office in their efforts to

reach out and engage younger voters. Drawing on extensive field research concerning

young adults' views about politics, elections and government, CYV publishes a Toolkit

and suggested campaign practices and materials to assist candidates at all levels in

dealing with young adults about political participation and voting. The CYV provides a

wealth of information on polls, studies, and news links as well as links to data on voting

and registration provided by the U.S. census (2003) and the Federal Election Commission

(2003).

The information found on these web pages will be drawn upon throughout this

research. The information is valuable for comparison purposes with the data complied in

this research about University of Florida students and voter registration in Alachua

County, Florida. The information is valuable for establishing a benchmark about young

voters in general.

Research on the topic of young voters has been executed by the U.S. Census and

the Federal Election Commission. The results of this research has been compiled and

disseminated by CIRCLE (The Center for Information and Research on Civic Learning

and Engagement). This research includes data on young voters at the national level.

CIRCLE promotes research on the civic and political engagement of Americans between

the ages of 15 and 25. A comparison of registered voter information at the state level









(Florida) with the national information, reveals that Florida has 74.3% (U.S. Census) of

citizens eligible to vote are registered to vote. This characteristic of Florida voters closely

mirrors voting registration at the national level, which overall is 78.1%. A focus on the

young voter population reveals a different trend between the State of Florida versus the

national level for the same age group. At the national level, the 59.9% of 18-25 year olds

registered to vote. In contrast, only 39.9% of eligible voters in the same age group

registered in Florida. Male registration the national level for the 18-25 age group is

40.3%. Males in the same age group in Florida have the same propensity to vote as their

counterparts at the national level, at 40.0%. However, young females at the national level

have a higher propensity to vote at 44.3%, while in Florida have a slightly lower

propensity to vote than their male counterparts at 39.9%.

Propensity to vote among young eligible voters can be viewed in a wider context

than those voters showing up at the voting booth on Election Day. Propensity to vote is

also a measure of social commitment, social involvement, and political participation.

These characteristics are important to more than just politicians running for office. They

are important to any politicized organization where opinions of voters might sway. One

example of such a politicized organization is the state public university (Thrall and

Mecoli, 2003).

What makes a state public university politicized? Admissions in the state premier

universities are much less than the number of qualified students that might desire to

attend the university. Funding from the state, and student paid tuition, is not based upon

consideration of expense of maintaining the university and providing a constant quality of









service. Instead, revenue to support the university is more often based upon political

considerations. The most affected population groups are university students themselves.

How may this affected population be drawn upon to become politically active in

those issues that they have a self interest in supporting? The answer must take into

consideration the characteristics of the affected population subgroup, and their propensity

to be politically active, and for this political action to be manifested by voting behavior.

The characteristics of the affected population subgroup can be evaluated using

contemporary geographic information systems technology. The results of the study can

be put to use to contact the affected population subgroup and thereby have an impact at

the voting booth.

Research

The previous chapter concluded with how a university can be proactive in reaching

out to its constituency its students so that they acting in their clear self interest

knowledgeably act on that self interest at the polls and voting booth. To accomplish this,

it was argued, the university must achieve a better understanding of who their students

are, what motivates them to become politically active, and what are the odds of their

students becoming sufficiently aware of how their self interest is affected by political

decisions that those students become politically active in the cause of the university.

Geographic information systems technology when properly applied can help to address

and answer these questions. The below demonstrates how geographic information

systems technology, in particular that technology developed in support of business

geography, can answer these questions and also become a valuable instrument for

improved university administrative judgment.









This demonstration draws upon the registered voter database for Alachua County,

Florida, the home of University of Florida. It also uses the database of students enrolled

at University of Florida. No other study has integrated university student enrollment data

and voting registration data. Benefits of this study include documenting the number of

students that register to vote in the county. The fewer the number, the less the local

political importance. The greater the number, the larger is the importance of the student

population to local politics. Because state representatives are elected by voters in their

jurisdiction, the more important are the student voters to the overall voting population of

the jurisdiction, the more politicians will consider the students' preferences.

The University can use this information in documentation to politicians whose

legislative districts overlap where University of Florida students reside. The University

might encourage its students who are registered to vote in to engage in a letter writing

campaign to those legislators within whose districts they reside. What type of student is

registering to vote, and voting? What are their geodemographic characteristics? The

discovery of their geodemographic patterns might be useful to groups that organize voter

registration drives, and useful to legislators that need to get the student vote to win

election. But getting students to register to vote, and to vote, is of little practical value to

a special interest group unless there is some knowledge about how those young voters

will vote. So, this research also addresses of the methods and procedures for discovering

the relationship between propensity to register and vote, and party affiliation. This

research drills deeper into the geodemographic characteristics of students by evaluating

the relationship between the lifestyle segmentation profiles of the parents of students that

register to vote, and vote. In this way, because LSP of parents of students can easily be









determined, the university can more effectively target its students who are susceptible to

registering and voting.

Data

Two data bases were of primary importance to this study: voter registration for

Alachua County, Florida, which includes name, address and voter affiliation; and,

University of Florida student records, which includes local and permanent addresses for

its students, race, grade level, and gender. This data is not available to the general public.

This data is made available on a case by case basis, and then only if it is judged by the

officials responsible for the data that individuals listed in the database will not be

subjected to any inconvenience and harm, and if there are sufficient benefits that will

arise from making the data available. It was easy to make the case to the University to

make its student records available for this study because the University Registrar and the

Office of the President requested that this study be undertaken as a proof of concept

demonstration. That the university originated this study is evidence of its importance to

the university, and evidence of how state universities are becoming proactive in this

politicized environment to protect its student constituency. For nonpartisan reasons, the

Alachua County Registrar of Voters was also interested in the results of this study.

Knowing how many students register to vote, and actually vote, has long been of concern

in Alachua County Registrar. One of the goals of the Registrar of Voters is to maximize

the voter registration and voter turn out. The appropriate approvals were obtained. The

registered voter database for Alachua County was acquired in September 2002; the

database contained a total of 124,515 names, party affiliation, and addresses. At the same

time, the student enrollment database for Fall (September) 2002 was obtained from the

University of Florida Registrar. The student record database included students' name,









local address, age, and permanent address. Recall from the previous chapter that 30% of

students list their permanent address in Alachua County.

Methods

To join the voter database to the student database, a data field must exist between

the two data files that contain the same information and that uniquely identifies each

record (student, voter) in each database. Last, first, and middle names are included in

each database. However, the name in each database might not be the same. The most

common reason that correct records could not be joined was revealed to be different

treatments of the last name. The university maintained full middle name, while the voter

database often abbreviated the middle name. This problem was avoided by using only the

first and middle name. But some students have the same first and last name. To insure

that the correct records were joined, an address join was also performed. So, for records

of the two databases to be joined, the first name, last name and address must be the same

for each of the databases. Not only does the use of address assure that names are correctly

matched between databases, but according to Florida State Election Laws, a person must

be registered to a precinct in which they vote (Florida Division of Elections Election

Code 101.045). So the students who have not changed their address with the Supervisor

of Elections yet changed their address with the University would legally be ineligible to

vote in their previous precinct if the current address falls outside.

A summary of demographic information from the UF registered student voters is

presented in Table 4-2. The demographic information was created by processing

students' permanent addresses through ESRI BIS Coder (formerly CACI Coder/Plus).

The demographic characteristic of University of Florida students can be then compared to

those published by the Campaign for Young Voters (CYV). CYV links students'









demographic characteristics with party affiliation, and other information obtained from

surveys.

Student voters with a different local and permanent address were identified. These

data records were used to establish a possible relationship between the student's party

affiliation and demographic characteristic as revealed via their LSP obtained from their

permanent address. A cross tabulation was created to discover possible relationships

between party affiliation and LSP. SPSS (a statistical software package) was used to

create the cross tabulation. Party affiliations were limited to the big three: Democrats,

Republicans, and Independents. Only a trivial percentage of registered voters in Alachua

County fall outside the three largest party affiliations. When addresses are processed

through ESRI BIS Coder, ZIP+4s are assigned to each address, and then ACORN (a

classification of residential neighborhoods) LSP measurements are assigned on the basis

of ZIP+4. The product assigns both LSP description, and LSP index number (Appendix

A). Various ACORN LSP index numbers were then agglomerated together, treating

categories lA, lB and IC as the same. Because the purpose of this part of the study was

to predict which students were likely to be Republicans or Democrats, those categories

with higher propensities to be one party or the other were then agglomerated together.

For example, ACORN LSP categories 1 and 2 are both highly likely to be Republican, so

they were combined. At the same time, category 6 was combined with 7 because both

categories have dominated by Democrats. These assumptions are based on surveys and

studies (Erikson, Lancaster, &Romero, 1989), (Stanley & Niemi, 1994), but no research

has been conducted to indicate party type by LSP (from CACI). In Clustering ofAmerica

by Michael J. Weiss (1988), how neighborhood lifestyles (LSPs) are a natural way for









predicting political behavior is discussed. Although no political information is collected

when compiling LSPs from CACI, the clusters are on multiple socioeconomic indicators,

which can mirror voters. Weiss states that these clusters can explain why some issues

cross party lines, but these same issues remain true to the neighborhood lifestyle. He also

claims that geodemographics may prove to be practical rather then traditional voting

blocks.

The Pearson chi-square coefficient was calculated using SPSS to test for

dependence between these two categorical variables (party affiliation and LSP). The

hypotheses for the cross tabulations are:



* Ho: There is no dependence between party affiliation and LSP
* H1: There is dependence between party affiliation and LSP

In the above hypothesis test setup, if Hi is accepted, it can be reasonably be

concluded that high correlation exists between the lower indexed LSP categories 1 and 2,

and the Republican party, while the higher number LSP index categories 6 and 8 are

highly correlated with the Democratic party. Independent voters are expected to fall in

the middle to high LSP indexed categories.

An independent sample t-Test was also performed to determine whether the mean

LSP scores were different for each political party. The t-test requires that two groups be

tested at a time. T-tests were calculated on the difference in LSP scores between

Republicans and Democrats, Republicans and Independents, and Democrats and

Independents.









The hypotheses for the t-Test are:



* HO: The mean LSP type is the same for the categories party affiliation
* HI: The mean LSP type is different for the categories of party affiliation

If Hi is accepted, and the means differ between parties and the p value (sig. 2-

tailed) is not close to 1 then the assumption can be made that the two parties being tested

most likely appear in different LSP categories.

Results and Discussion

After joining the registered voter data with the student data as described above, it

was determined that 7,621 (Figure 4-1) students enrolled at University of Florida in Fall

2002 were also registered to vote at that same time in Alachua County. This is a very

small population. The results on this small a population can not extend to the entire

student body, but instead provide a method to examine larger populations. The registered

student voters by race, grade level, gender, party affiliation is shown in Table 4-1 in

addition to the breakdown of these categories by party affiliation. The LSP types of

students are not shown because the majority of students by student address are

categorized in type 5A Twentysomethings or type 5B College Campuses.

Figure 4-1. Student and registered voter numbers


;II(Uan rLnnlill












Table 4-1. Student registered voter demographics
PARAMETER COUNT %
Undergrad 5261 69.0
Graduate 2365 31.0
Female 3885 51.0
Male 3324 43.6
Hispanic 602 07.9
Black 583 07.7
White 6005 78.8
Asian 326 04.3
Other 58 0.76
Democrat 3119 41.0
Republican 2101 27.6
Independent 1781 23.4
Other 619 08.1
Female Democrat 1769 45.5
Female Republican 947 24.4
Female Independent 899 23.1
Female Other 270 00.7
Male Democrat 1220 36.7
Male Republican 1083 32.6
Male Independent 785 23.6
Male Other 236 07.1


7,621 Student Registered Voters


Total









A comparison of national level data from the Campaign for Young Voters

(www.campaginforyoungvoters.com) to registered student voters at the University of

Florida is valuable for both corroborative purposes and for determining if and how

University of Florida students differ from the national trend. The sample from UF

represents those registered to vote and is only a small percentage of total UF students.

The comparison reveals that full time students on the national level tend to be more

Republican (38%) then Democrat (22%), but Independents were similar to the Democrats

(24%). However, at University of Florida, the majority of student voters are Democrats

(40%) then Republicans (27.5%) followed by Independents (23.4%). On the national

level as a whole, it would seem that a majority of young voters would be Democrat,

especially full time college students, but a parent's political affiliation may have an

influence with the way the student identifies to a particular party and votes

(www.stateofthevote.org 2003). Although comparing results to the national level may

not be accurate due to the small population being sampled (UF), these results allow

general comparisons.

When comparing gender within the registered students voters at UF, females tend

to be more Democrat while the males have similar percentages for Democrat and

Republican. In comparing race, the black registered student voters are predominantly

Democrat followed by Independents, Others, and finally Republicans. It is interesting to

note that the "Others" category has a higher percentage than that of Republicans.

Democrats rank highest with the Hispanic and White registered student voters, but differ

when comparing Republicans and Independents. The Independents have a higher

percentage among Hispanics than do Republicans. In the national survey, whites (15-25









years old) have a higher percentage of Republicans than Democrats, but the highest

percentage lies with the Independents. The Black young adult percentages are consistent

with that UF, but the Hispanic young adults have the highest percentage as Independents

followed by Democrats.

The differences between the national study and the University of Florida study may

arise from data collection procedures. The national study interviewed college and non-

college students and at a much younger age (15-25). Influences of the parents (Beck &

Jennings, 1991; Sears & Valentino, 1997) have a considerable impact on how these

respondents answer to the survey questions by the CYV. There was no reliable source of

information on just college student registered voter surveys that could be used in

comparison. The CYV did not rigorously compare voter registration records with student

enrollment records, leaving the possibility of survey error. At the same time, the

University of Florida study evaluated only students registered to vote in Alachua County

because only Alachua County voter registration records were available. This might create

a bias, but in an unknown direction, because some students might register to vote at the

address of their parents' outside of Alachua County. Politicians might not consider

university students to be a viable voting block, yet college aged students are the fasted

growing generation surpassing that of the baby boomers (Skaggs & Anthony 2002), and

over 7,000 were found to be registered in Alachua County a very considerable voting

block.

The Florida State Legislature does keep information regarding the number of

registered voters by age groups per district. In Figure 4-2 the number of registered voters

between the ages of 18-29 is shown. This age group would most closely identify to that









of college aged students and the highest concentration of registered voters this age are

found where the major universities are located (Tallahassee Florida State, and

Gainesville University of Florida).

Student's Party Affiliation and Parent's LSP

Cross tabulations

The number of registered student voters who have a different student address and

permanent address was determined to be 1,074 (Figure 4-3). Again, this is an extremely

small population to test from the overall student population. This analysis illustrates the

process that can then be used in future research. The results of the cross tabulation is

shown in Table 4-3. This reveals if there is a significant relationship between party

affiliation and LSP. The adjusted residual compares the difference between expected

count and the count. If the null hypothesis (Ho) were true (no dependence between party

affiliation and LSP) the expected count would be the number that is most likely to occur.

When the adjusted residual is around 2 (meaning 2 standard deviation away from the

mean), then it can be statistically concluded that the difference is significant.

In analyzing the different categories, LSP type 1 and 2 have more Republicans.

Therefore, Hi (dependence between party affiliation and LSP) is accepted. This supports

the hypothesis that more Republicans would be in this category than either Democrats or

Independents. Type 3 shows no significant results due to the fact that there is a low count.

Type 4 is approaching significance in the fact that this category is more likely to be

Republican. Type 5 shows no significance in establishing dependence between party

affiliation and LSP. Type 6 (combined with 7) shows that this category is more likely to

be affiliated with the Independents then Democrats. Type 8 shows the strongest evidence

supporting Hi that this category is more likely to be Democrat. Performing the Pearson 's









Chi-squared test (p value of .025) on this data (Table 4-4), confidently reject the null

hypothesis (Ho).

T-Tests

The t-Tests compare the differences in the means of the data analyzed. This test

was used to show which LSP categories the different parties were likely to fall within.

The first T-Tests performed was between Democrats and Republicans (Table 4-5). This

test resulted in a significant difference between means and the P value of .007 rejects the

null hypothesis Ho (LSP type is the same for the categories of party affiliation). This

supports evidence the Democrats will be in the higher LSP types (6,8). The test between

Republicans and Independents (Table 4-6) results in Ho also being rejected. The last test

between Democrats and Independents (Table 4-7 ) accepts Ho because the p value is .75

(close to 1).

Typical Party Demographic

It has been hypothesized elsewhere that political affiliation and demographic

characteristics are highly correlated (Erikson, Lancaster, & Romero, 1989), (Stanley &

Niemi, 1994). For example, one is statistically more likely to be a Republican if one is a

man, a college graduate, in the top income, and living in a rural area. One is more likely

to be a Democrat if one is a woman, a senior citizen, nonwhite, and living in an urban

area. These hypotheses arise from various polls and surveys such as The Harris Poll, a

2000 exiting poll for CNN (Harris Poll, January 5, 2001), U.S. Census, and other studies

conducted on issues such as the gender gap (Trevor, 1999) and the party of the rich

(Sander, 1988). This University of Florida study breaks down the geodemographics into

finer and more accurate detail than has been done before, and the results are not based









upon a survey but actual cross tabulation of student enrollment and voter registration

data.

Conclusions

Understanding student propensity to vote gives the University of Florida an

understanding of how much its student constituents may be relied upon to support

political actions that might be necessary to maintain the present education environment.

Moreover, this study demonstrates methods that can be followed that can identify

particular kinds of students and how those students might be approached to support a

given political position.

Geodemographics of students can be used to project specific issues to susceptible

subgroups of students and their parents. A reasonable conjecture is that if a student

votes, then his or her parent is also likely to vote (Beck & Jennings, 1991). The

University can encourage subgroups of its students to go to the polls and vote in support

of issues aligned with university needs.

A statistical relationship has been demonstrated between party affiliation and LSP.

This correlation is made on a small population and demonstrates a process for further

studies. It has been shown that Republicans are more likely to be within LSP types 1 and

2, therefore, letters addressed to these students and their parents might take on a more

Republican tone. At the same time, students and their parents shown to fall within those

LSP categories more likely to be of Democratic affiliation might receive information

specifically directed to that bias.

Universities have become politicized. In response, universities must engage

actively in the political process else its constituents university students will be left

behind and ill served. It has been demonstrated here that geographic information systems










technology, specifically the technology and analysis of the business geographer, can be

applied to generate information that can be used to improve University administration

make decisions.

Table 4-3. Cross tabulations between party affiliation and LSP


PARTY
DEM IND REP Total
life 1 Count 116 66 108 290
style Expected Count 124.9 69.5 95.7 290.0
seg Adjusted Residual -1.3 -.6 1.8
profile -
3 Count 10 3 3 16
Expected Count 6.9 3.8 5.3 16.0
Adjusted Residual 1.6 -.5 -1.2
4 Count 29 13 29 71
Expected Count 30.6 17.0 23.4 71.0
Adjusted Residual -.4 -1.2 1.5
5 Count 190 111 151 452
Expected Count 194.7 108.3 149.1 452.0
Adjusted Residual -.6 .4 .3
6 Count 38 24 16 78
Expected Count 33.6 18.7 25.7 78.0
Adjusted Residual 1.1 1.5 -2.4
8 Count 27 11 7 45
Expected Count 19.4 10.8 14.8 45.0
Adjusted Residual 2.4 .1 -2.5
Total Count 410 228 314 952
Expected Count 410.0 228.0 314.0 952.0


Table 4-4. Chi-Square tests













Table 4-5. T-Test Republican and Democrat

t-test for Equality of Means


Mean Std. Error
t df Sig. (2-tailed) Difference Difference
life style seg profile Equal variances
assumed



Table 4-6. T-Test Republican and Independent

t-test for Equality of Means


Mean Std. Error
t df Sig. (2-tailed) Difference Difference
life style seg profile Equal variances -2.057 544 .040 -.37 .18
assumed


Table 4-7. T-Test Democrat and Independent

t-test for Equality of Means


Mean Std. Error
t df Sig. (2-tailed) Difference Difference
life style seg profile Equal variances 318 639 .751 5.53E02 .1
assumed _______

















JACKSONVILLE


-. I-T L


7]


TAMPA

ST. PETERSBURG

Number of Registered Voters
Per State House Districts
18-29 Years Old


-jI






r'
I L.


% 18-29
I I I
S11-16
S17-22
S23-28


Source Florida State Legislature


Si



IJ '"


I -1-1I -


i'ORT LAUDERDALE

MIAMI
MIAMI BEACH


I I- I-


Figure 4-2.18 to 29 registered voters per House district


I


-J J'



















124515 Alachua County Registered Voters





42,667 Florida Permanent Addresses


7,621 UF Students Registered To Vote


Figure 4-3. Numbers for student registered voters



















CHAPTER 5
CONCLUSIONS




The Registrar's Office of the University of Florida on behalf of the University

President's Office commissioned this proof of concept study. The purpose of the study is

to provide a tool whereby the University could bring to the attention of State legislators

the importance of the university in serving the educational needs of students of their

district and state region.

GIS and the geodemographics (Geographic Decision Information Systems) have

been demonstrated here as one means to generate the information necessary to receive the

political support to further the mission of higher education. In addition to the pilot study,

information regarding registered student voters is included to allow the University a

better understanding of the political aspects of the student population.

After the successful completion of populating the student database with the correct

Florida State Legislator and geodemographic for the student's permanent address using

GIS, the University will now have the ability to develop a mail campaign. The

University now knows which district the parent or student may reside in and vote in so

issues central to the well being of the school can be addressed. Further targeting through

mail can be performed based on the geodemographic of the parent or student. This will









allow the University to craft letters in such a manner that will appeal to the different LSP

types and may generate increased support.

The University also knows where trends in the geodemographics types have

emerged. If the University was looking for more support in additional state funding to

aid students with financial support, they may looks to LSP type 8A Young Immigrant

Families. Although there may be portions of this group located throughout the state, they

are highly concentrated in South Florida. Only sending letters to this group in South

Florida as opposed to just everyone in the database with type 8A can cut costs and still

have enough support for the issue.

Learning which students are registered to vote may give the University insight into

generating increased support from the student who is more likely to vote. Demographics

of these students can help target issues to the students and there parents. There is a

likelihood that if a student votes, his or her parent may also vote (www.stateof

thevote.org), (Beck & Jennings, 1991). The University can specifically cater to this group

of people urging them to go to the polls and vote in support of university issues.

Statistically discovering if there is any significance between party affiliation and LSP

type from the sample of student registered voters with a different student address and

permanent address, allows assumptions to be made on the entire student database. Since

it is shown that Republicans are more likely to be in LSP types 1 and 2 perhaps the letter

for the mail campaign can be more conservative in tone and the letter to LSP type 8 that

is more likely to be Democrats can be crafted in such a way to appeal to a more liberal

outlook.






67


Utilizing GIS, geodemographics, and voter registration data, can necessitate

different data sources to be used in conjunction with each other. This can allow for in

depth analysis to make wiser decisions about one's business, in this research, the

University of Florida.

















APPENDIX A
CACI ACORN LIFESTYLE SEGMENTATION

Consumer Type 1A: Top One Percent

Demographic: "Top One Percent" households are usually older married couples in their

peak earning years. Almost half are between the ages of 35 and 64 years; their median

age is 43.0 years. Most are empty nesters, although many have older children attending

college. This predominantly white segment also has an above-average percentage of

Asians and Pacific Islanders.

Socioeconomic: Their median household income is $115,900. For most households,

sources of income include salaries, interest, dividends, or rental property income for most

households and self-employment income. Almost 60 percent of the labor force work in

executive or professional positions. They are highly educated; over 60 percent of the

adult population aged 25 years or older hold bachelor's or graduate degrees.

Residential: Most own their single-family homes found in the older suburbs of large

metropolitan fringe areas scattered throughout the U.S. High concentrations of "Top One

Percent" are found in California, New York, New Jersey, Illinois and Texas. Their

median home value of over $400,000 is more than four times the national average. Well

over 80 percent of these homes are valued at $250,000 or more.

Preferences: "Top One Percent" households rank high for many consumer purchases

including most clothing and apparel categories, frequently buying via mail or phone

orders. They drive luxury cars and enjoy taking domestic and foreign excursions. They









are active volunteers and fundraisers. They also enjoy reading newspapers and playing

racquet sports.

Consumer Type 1B: Wealthy Seaboard Suburbs

Demographic: These older, married couples have no younger children, but a

disproportionate share of their adult children still live at home: they're empty-nester

wannabes. Their median age is 40.6 years. The proportion of householders between 45

and 64 is over 30 percent higher than the national average. This population is

predominantly white but has an above-average percentage of Asians and Pacific

Islanders.

Socioeconomic: Their median household income is $72,900; almost half of the

households earn $75,000 or more. Most households receive income from dividends,

interest, or rental properties; almost 20 percent are receiving retirement income or

pensions. More than half of the households have a net worth of $150,000 or greater, three

times the national average. They are well educated; over 40 percent of the adults hold a

college degree and/or work in professional or managerial positions.

Residential: They own their single-family homes that were built between 1950 and 1969.

The median home value reflects the markets where these neighborhoods are located,

$261,900; more than two and one half times the national average of $99,800. "Wealthy

Seaboard Suburbs" are found in the suburban areas of East and West Coast urban metros.

Preferences: "Wealthy Seaboard Suburbs" pursue physical fitness by playing a variety of

racquet sports, golf, working out at the gym, and taking vitamins regularly. They spend

freely on home furnishings and improvements, and also travel extensively, preferring









foreign to domestic trips. They read two or more daily newspapers and metropolitan and

business-related magazines and listen to classical, jazz and alternative radio.

Consumer Type 1C: Upper Income Empty Nesters

Demographic: They are usually empty nesters: married couples with no children living at

home. They also are predominantly white and middle-aged. Almost 50 percent of the

householders are between the ages of 45 and 64; their median age is 42.4 years.

Socioeconomic: These couples are prosperous, with a median household income of

$68,400. Two-thirds of them receive income from dividends, interest, or rental properties.

They are usually business owners or managers. With approximately 20 percent receiving

retirement or pension income, early retirement is common. They are also highly

educated, which is atypical of their generation. Over 90 percent of the adults aged 25

years and older finished high school; more than 40 percent hold a college or graduate

degree.

Residential: They own their single-family homes with a median value of $157,300, nearly

60 percent above the national average. The These neighborhoods are primarily suburban,

but range from urban areas to smaller, non-metropolitan communities scattered

throughout the U.S. High concentrations are found in Pennsylvania, New York, Texas,

Illinois and Ohio.

Preferences: They play a variety of sports including golf, racquet, and sidewalk sports.

They also enjoy traveling, preferring foreign travel to domestic trips; visiting museums,

and attending concerts. They own PCs, pianos, swimming pools and luxury cars at above-

average rates. They are avid readers, usually two or more daily newspapers, along with

metropolitan and business-related magazines.









Consumer Type ID: Successful Suburbanites

Demographic: This family market has an average household size of 3.1 persons, 19

percent above the national average. They are between the ages of 35 and 54 years, with

school-aged children. Their median age is 37.1 years, slightly higher than the U.S. value

of 35.5 years. The population is predominantly white, but Asians and Pacific Islanders

comprise a disproportionate share of almost 8 percent.

Socioeconomic: At $87,200, the median household income of "Successful Suburbanites"

is more than twice that of the U.S. figure. Dual incomes and investments contribute to

their affluence. Employment rates are high for men and women. The work force is

professional, well educated, and mobile.

Residential: "Successful Suburbanites" own homes in newer suburbs with a median home

value of $210,500, twice the national average. Most of their single-family houses were

recently built between 1980 and 1990. Located predominantly in the urban areas of

metropolitan centers, they traded convenience for lifestyle and moved out to newer,

suburban developments; over 30 percent commute across county or state lines to work.

Preferences: "Successful Suburbanites" drive minivans and luxury cars, own swimming

pools and PCs, and play golf and tennis. They splurge on home furnishings and

improvements, apparel, electronic "toys", and travel, and fund much of this spending with

credit cards and loans. In fact, "Successful Suburbanites" are more likely than most of the

other consumer markets to have loans. They attend concerts and movies; read two or

more daily newspapers; and business, finance and metropolitan magazines.











Consumer Type IE: Prosperous Baby Boomers

Demographic: The age profile of this market is singular: baby boomers who were born

between 1949 and 1964 with young, primarily preschool and grade school age children;

over 40 percent greater than the national average. Nearly 20 percent of the population is

under 10 years of age as compared with less than 15 percent of the U.S. population.

Typical of their generation, these families are very mobile. Over 35 percent of the

population have moved in the past 5 years, double the national mobility rates.

Socioeconomic: "Prosperous Baby Boomers" have a median household income of

$60,100, more than 50 percent higher than the U.S. figure. Their higher income is

primarily the result of two salaries. Over 70 percent of the households have 2 or more

workers. Also typical of this cohort is educational attainment: two-thirds have attended

college or completed a degree.

Residential: Over 74 percent "Prosperous Baby Boomers" tend to be homeowners in new

housing developments, 75 percent of their houses were built since 1980. Their median

home value of $124,700 is approximately 25 percent higher than the U.S. median. These

primarily suburban neighborhoods are located within commuting distance of

metropolitan job centers in the Sunbelt and West.

Preferences: "Prosperous Baby Boomers" are active sports enthusiasts: playing golf,

tennis or racquetball; lifting weights; and doing aerobics. Much of their time is spent on

the PC playing games, doing their taxes and finances, or just browsing the web. They are

likely to visit theme parks or gambling casinos. They buy electronics, camcorders, PCs,









sports equipment, and home furnishings. They do not invest or save in proportion to their

income, but they do carry a lot of insurance.

Consumer Type 1F: Semirural Lifestyles

Demographic: Married couples aged 35 to 54 years with and without children living at

home dominate this market. The median age is 36.8 years as compared to 35.5 years for

the U.S. About 35 percent of the households are empty nesters; 40 percent have school-

age children living at home. Over 92 percent of this population is white.

Socioeconomic: With a population of over 15 million, or over 5 percent of the total U.S.

population, this is the largest "Affluent" market with a median household income of

$61,500 but not the wealthiest one. Unemployment is low, but their income sources are

more varied in this affluent market, including self-employment income from both farm

and non-farm jobs with the usual wages and salaries, plus interest and dividend income.

Residential: "Semirural Lifestyles" live in semi-rural areas or on the urban fringe of

metropolitan areas. Many of their homes are newer, built since 1970, with a median home

value above the average of $131,500. Most of these homes are valued between $75,000

and $200,000. These neighborhoods are scattered throughout the U.S., but compose over

10 percent of neighborhoods in Connecticut, New Hampshire and Alaska.

Preferences: "Semirural Lifestyles" spend a significant amount of time at home -

swimming in their pools, doing home projects, gardening, or using the PC. Outside the

home, they join political and civic groups, the play golf and attend movies. They buy

home furnishings, tools and the latest electronic gadgets, PCs, camcorders, and cameras.

"Semirural Lifestyles" avidly read books; two or more newspapers per day; and science,

news and lifestyle magazines.









Consumer Type 2A: Urban Professional Couples

Demographic: With a median age of 37.8 years, "Urban Professional Couples" are older

than the U.S. overall (35.5 years). They have above-average indexes for each adult age

group from 25-29 years and older and below-average indexes for the age group 20-24

years and younger. They are predominantly married-couple families, with few or no

children, but the mix also includes single-person and shared households, the results of

high divorce rates through the 1980's.

Socioeconomic: Their median household income of $49,700, is 25 percent above the U.S.

figure. Employment levels are high. The labor force is well educated and employed

mainly in professional or managerial positions. Over 35 percent of adults aged 25+ hold a

college degree.

Residential: They live in a high-density mix of single-family homes and townhouses with

smaller two-to-five- unit rentals. Most homes are owner-occupied, including

condominiums. The median home value is above average of $142,700. These

neighborhoods are located inside urban areas scattered throughout the U.S., but especially

in Massachusetts, California and New York.

Preferences: "Urban Professional Couples" rank near the top in their consumption of

imported wines, listening to classical music, traveling overseas and domestically and

visiting museums. Physically active, they play a variety of sports, including tennis, golf,

bicycling, and jogging. They join AAA, frequent flier programs, and health clubs. They

read two or more daily newspapers and gourmet, business, travel and lifestyle magazines.









Consumer Type 2B: Baby Boomers with Children

Demographic: Approximately two-thirds of the households are married couples, most

with children; over 50 percent more than the national average. Their median age is 31.2

years; more than 35 percent of the population is under the age of 20; 34 percent of the

householders are between the ages of 25 and 44 years. Typical of the cohort, many are

mobile, still moving to find the best jobs or location.

Socioeconomic: With a population of 11.8 million, "Baby Boomers with Children"

represent 4.3 percent of the U.S. population. Like the "Prosperous Baby Boomers", this

young market has very high employment rates including a high proportion of dual

income families, but they are not as affluent. Their median household income is $48,000.

Some have graduate schooling, but this group is less likely to have completed a graduate

or professional degree.

Residential: They live in single-family homes built during the 1970s. Most of their homes

are owner occupied and valued with a median home value of $95,700, just below the U.S.

median of $99,800. Neighborhoods are found within the metropolitan urban fringe, in

smaller towns and in rural non-farm communities. Western states, especially Utah,

Alaska and Wyoming have a lot of "Baby Boomers with Children," harking back to the

westward migration of baby boomers in the 1970s.

Preferences: "Baby Boomers with Children" represent family- and home-oriented

consumerism. They enjoy do-it-yourself home projects, go camping, fishing, and hunting.

They rent videos; grow vegetables; use their PCs; and visit museums, zoos, and theme

parks. They own pets and spend their discretionary income on swimming pools, outdoor

grills, campers. They own multiple cars, minivans, and sports utility vehicles.









Consumer Type 2C: Thriving Immigrants

Demographic: Although fewer than 30 percent of the population are foreign-born, 40

percent speak a language other than English at home. "Thriving Immigrants" are

immigrants and their children. Most households are families, typically married couples

with either very young or adult children. The population is diverse, including Asian and

Pacific Islander at 4.5 times higher than the national average and Hispanic origin at three

times the national average. The age distribution is slightly younger; the median age is

32.9 years.

Socioeconomic: Although this is a small market, only 1.9 percent of the U.S. population,

the households in this market earn moderate to high incomes, primarily from wage and

salary employment. Their median household income is $48,900. Employment is high;

unemployment is low. Despite their relative affluence, these consumers are a weak

financial market. They neither invest nor save their income. Loans also rank low for this

group.

Residential: Home values and rents are especially high for "Thriving Immigrants" due to

their location. These neighborhoods are most likely to be found in California's urban

areas. The median value of owner occupied units is $184,900, almost twice the national

median. Home ownership is slightly below just over 59 percent, comparable to the U.S.

rate. Most of their homes are single family, built between 1950 and 1969.

Preferences: Because of ties to their birth countries, "Thriving Immigrants" are likely to

have passports and spend a sizable portion of their income on foreign trips and long-

distance phone calls. They also enjoy day or weekend excursions to casinos and theme









parks. They watch television; use the PC; listen to adult alternative, rock and

contemporary radio; and read news, fashion and lifestyle magazines.

Consumer Type 2D: Pacific Heights

Demographic: "Pacific Heights" are predominantly Asians and Pacific Islanders but with

nearly 14 percent, there is a significant Hispanic minority. Almost 35 percent are foreign-

born and half speak a language other than English at home. The segment's age profile is

almost identical to that of the U.S. overall; indexes for all age categories are within 10

percent of their U.S. counterparts. The household distribution parallels the United States -

with a slightly higher proportion of adult children living with their parents 35 percent

above the national average.

Socioeconomic: This small about one percent of the U.S. population but upscale

consumer market has a median household income of $48,400. One of every two

households earn $50,000 or more annually. Their sources of income include self-

employment, interest and dividends, wages and salaries. Education is a priority: the labor

force is well educated. Half of the adult population over age 25 attended college or

received a bachelor's or graduate degree. Another 10 percent are presently enrolled in

college.

Residential: "Pacific Heights" neighborhoods are high-rent districts, located primarily in

Hawaii and California. With median home values around $250,000, ownership is below

average (50 percent). Twelve percent of the owner-occupied units are condominiums.

Most of the housing is in multi-unit complexes with five or more units. Population

density is high, over 15,000 persons per square mile.









Preferences: True urbanites, "Pacific Heights" own few cars, and when they do, they'll

buy a foreign, subcompact model. To keep in touch with relatives, they spend a

disproportionate share of their income on foreign trips and long-distance phone calls.

They watch TV, listen to the radio, gamble at a casino, play the lottery, go skating and

play billiards. They are inclined to invest rather than to save their money.

Consumer Type 2E: Older, Settled Married Couples

Demographic: "Older, Settled Married Couples" are middle-aged married couples who

have settled into their neighborhoods and surroundings. Although many households

include school age or adult children, their age profile is slightly older, with a median of

37.2 years. The population is predominately white.

Socioeconomic: This market includes 4.4 percent of U.S. households and 4.6 percent of

the U.S. population. The income distribution reflects a degree of comfort earned by

working couples aged 45-64 years. Almost half earn $50,000 or more. The median net

worth of this segment is $69,000, 75 percent higher than the U.S. median of $39,000.

Almost 20 percent of these households are drawing retirement income.

Residential: "Older, Settled Married Couples" live in older, single-family homes owner-

occupied with a median value of $95,100, compared to the U.S. median of $99,800. The

geographic distribution is national, but higher concentrations are in Michigan, Illinois and

Ohio, along with New York and California.

Preferences: "Older, Settled Married Couples" are likely to own stock, savings accounts,

bonds, IRAs, and other securities, but may also have debt from loans and credit cards.

They spend on home improvements, furnishings, and outdoor equipment. They watch









TV; listen to jazz and golden oldies; read newspapers and lifestyle, health and home

magazines.

Consumer Type 3A: High Rise Renters

Demographic: They are single with a median age of 37.9 years that belies the

concentration of younger households. Half are between 20 and 44 years of age, compared

with only 37 percent of the U.S. population. Fifty percent of these households are single-

person or shared -14 percent. The population includes foreign-born residents, non-

Mexican Hispanic, Asian and Pacific Islander.

Socioeconomic: Contrary to the usually lower incomes for single households, "High Rise

Renters" are affluent. They have a median household income of $48,900 versus the U.S.

median of $39,800. Well educated and gainfully employed, half of them hold a bachelor's

or graduate degree and work in professional or managerial positions. They prefer

investing over saving.

Residential: They are urban, mobile, apartment dwellers in densely populated, central city

locales in the largest metropolitan areas. Renters comprise six of ten households versus

three of ten U.S. households. Although most of the neighborhoods are located in the

largest metropolitan areas of New York, Illinois and California, more than 30 percent of

the neighborhoods in the District of Columbia are "High Rise Renters."

Preferences: Typical of more affluent consumers, "High Rise Renters" join health clubs

and environmental groups and pursue racquet sports, bicycling, and jogging. Their leisure

time is singles-oriented with dancing, going to bars/nightclubs, the theater, or museums.

They buy expensive clothing. They do not take out loans but use their credit cards









extensively. They often read metropolitan, fashion or gourmet magazines, and two or

more daily newspapers.

Consumer Type 3B: Enterprising Young Singles

Demographic: With a median age of 30.1 years, this population is young and

mobile. More than three of every five persons are under the age of 35. Approximately

half of the households are single-person or single roommates or shared households

compared with less than 30 percent of U.S. households.

Socioeconomic: Despite their youth, "Enterprising Young Singles" have a median

household income of $40,200 that is slightly above the U.S. median. Approximately one-

third of the population aged 25+ years has a four-year college degree or higher; another

third has attended college or completed a two-year associate degree. Employment is high.

"Enterprising Young Singles" have the highest rate for female employment; over 75

percent.

Residential: Most "Enterprising Young Singles" rent, nearly 64 percent versus 30 percent

for the U.S. overall. They prefer living in newer apartment complexes built in the '80s

and '90s in urban areas throughout the U.S., but concentrated in the West and Sunbelt. If

they own a home, its median value is $113,200.

Preferences: "Enterprising Young Singles" enjoy an especially young, active lifestyle.

Fitness is a priority; many belong to gyms, lift weights, jog and play racquet sports. Many

own only one car, usually a foreign, compact model. They go dancing in nightclubs, visit

museums, theme parks and zoos, and play computer games. They read health, computer

and photography magazines, and listen to jazz, rock and R&B music on the radio.











Consumer Type 4A: Retirement Communities

Demographic: They are older, but not exclusively elderly. With a median age of 40 years,

they are actually the youngest of Retirement Communities. Householders aged 75 years

and older represent over 15 percent of the household distribution, but so do householders

aged 35-44 years. Single-person households are 35 percent of the total, followed by 27

percent of married couples without children.

Socioeconomic: This is a small, prosperous market with a median household income of

$42,600. Their income sources are varied because of the large number of retirees who

receive their income from pensions, Social Security, interest and dividends, and salaries.

The labor force is also small, but well educated and professionally employed. Nearly

three out of ten hold a bachelor's degree or higher, compared with two out of ten in the

U.S. overall.

Residential: Congregate housing with meals are included in the rent, accounts for most of

the housing in "Retirement Communities." Surrounding neighborhoods include single-

family homes with a median value of $124,400. Most of the housing is relatively new

with almost 40 percent being built in the last decade. "Retirement Communities" may be

found in any state, but Florida and California locations are especially popular, accounting

for the higher mobility of this group.

Preferences: The lifestyles and spending of "Retirement Communities" are eclectic as

diverse as their neighborhoods. They play golf; train with weights; go bicycling;

camping; listen to classical music; and read consumer, science and nature magazines.









They enjoy traveling overseas more than taking domestic trips. They invest, probably

owning certificates of deposit, IRAs, stocks and other investments.

Consumer Type 4B: Active Senior Singles

Demographic: This mature market has a median age of 43 years. Nearly 25 percent are

aged 65 or older; many are widowed. Single-person households make up over 40 percent

of these households. Although younger families live in these neighborhoods, there are

few children. With over 85 percent of the population white, this market ranks below

average in diversity.

Socioeconomic: Representing 2.8 percent of U.S. households, "Active Senior Singles"

are one of the largest senior markets. Their median household income is below average at

$36,200, but many of them are retired. Poverty and unemployment rates are low. They

are fairly well educated; nearly 50 percent have attended college or higher. They are

slightly more inclined to save than to invest.

Residential: Most "Active Senior Singles" live in apartments. Two-thirds of these

housing are multi-unit, built before 1970. At $117,400, the median home value is above

the U.S. median. Neighborhoods are naturally higher density, located in urban areas with

a disproportionate share in the Mid-Atlantic and Northeast.

Preferences: "Active Senior Singles" enjoy traveling, usually taking foreign trips,

gambling at the casino, playing the lottery, attending concerts, visiting museums, and

going to their health clubs. They are active volunteers and involved in political

campaigns. Their heavy expenditures for imported wine, champagne and bottled water

reflect their taste for premium products. They own a variety of investments such as long

term certificates of deposit, savings accounts, IRAs, bonds and stock.









Consumer Type 4C: Prosperous Older Couples

Demographic: Their median age is 43.2 years, but nearly half are 55 years of age or

older. Most are married; few have younger children, although some families still have

adult children living at home. With its population that is more than 90 percent white, this

market lacks diversity.

Socioeconomic: "Prosperous Older Couples" enjoy a comfortable living; their median

household income is $45,200 and their median net worth is $88,900. Retirement is

prevalent, but not dominant. Many householders are evidently planning for their

retirement with investments and savings. More than half receive income from interest,

dividends, or rental properties.

Residential: They own suburban, single-family houses with a median value of $99,000,

slightly lower than the national average. These homes are in older, established

neighborhoods. The geographic distribution is somewhat more urban, but otherwise

similar to the U.S. population distribution by region. The states with the highest number

of these households are New York, Pennsylvania, Illinois and Ohio.

Preferences: "Prosperous Older Couples" are enthusiastic investors with diverse

portfolios that include certificates of deposit, IRAs, stocks, savings accounts and other

investments. Typically, they own luxury cars, use coupons, make contributions and

recycle. They enjoy diverse interests that include golf, needlework, outdoor gardening

and gambling in casinos and playing the lottery. They watch TV; read senior, health and

home magazines; and listen to news/talk radio.









Consumer Type 4D: Wealthiest Seniors

Demographic: Their median age of 53.5 summarizes the demographic profile of

"Wealthiest Seniors": over half of them are over 50. Over 30 percent are 65 or older,

more than twice the national average for this age group. Half are married couples with no

children at home, and approximately one fourth are single-person.

Socioeconomic: "Wealthiest Seniors" is a small market, slightly more than 1 percent of

households, but very affluent. Their median household income of $44,700 hides the

disproportionate share of upper income households. For example, one in seven

households had an income of $100,000 or more compared with one in 11 U.S.

households. This is the top financial market for investments and savings. Their median

net worth is $125,900, more than three times the U.S. median.

Residential: They live in small cities and towns in sunny climates. Nearly half of them

are located in Florida and ten percent in California. They usually own their single-family

homes or condominiums. Their housing is newer with a median value of $144,500.

Approximately 30 percent of the housing inventory is seasonal.

Preferences: The expendable income available to "Wealthiest Seniors" retirees

distinguishes this small market. The purchase potential indices for this neighborhood type

rank near the top among ACORN markets for investments, especially long-term

certificates of deposit and stock; travel; luxury cars; credit card use; and jewelry

purchases. To keep in shape, they go bicycling, play golf, and walk for fitness. They also

enjoy reading books; newspapers; and senior, lifestyle and business magazines; and

watching news programs and dramas on TV.









Consumer Type 4E: Rural Resort Dwellers

Demographic: "Rural Resort Dwellers" are usually older. Their median age is 41.7 years

but they are concentrated in the 45-plus age groups. More than four of ten are aged 45

years and older. Families are predominantly married couples, many recently retired to the

area. Over 95 percent of local residents are almost exclusively white.

Socioeconomic: Over one-quarter of the households earns less than $20,000 annually, but

their poverty rate is the same as the national rate. "Rural Resort Dwellers" earn a modest

living from seasonal employment and farming. Over 20 percent receive retirement

income and 35 percent receive Social Security. Most of the labor force works part-time.

About 15 percent have a bachelor's degree or higher compared with 20 percent of the

U.S. population.

Residential: They own single-family or mobile homes. Over sixty percent of the housing

in these areas is vacant for seasonal use. The median home value of $92,900 is low for all

owner-occupied units, but comparatively high for rural areas. Over 90 percent of the

"Rural Resort Dwellers" are in rural villages. These areas can be found in most states, but

most are concentrated in the North, around the Great Lakes or in New England,

especially Maine, New Hampshire, and Vermont.

Preferences: "Rural Resort Dwellers" love the great outdoors and often go hunting and

fishing. They own four-wheel drive vehicles, powerboats and campers. Health and

nutrition are important as they shop at health food stores, follow low-fat diets, and take

vitamins. They are likely to buy tools and outdoor equipment, satellite dishes, household

appliances and cameras.









Consumer Type 4F: Senior Sun Seekers

Demographic: The oldest in the Retirement Styles summary group, more than half of the

"Senior Sun Seekers" are 55 years or older. Their median age of 59.2 years is nearly 24

years older than the U.S. median. Over 45 percent are married couples without children,

but there is 32 percent of single-person households in this segment.

Socioeconomic: "Senior Sun Seekers" is a modest market in numbers, totaling 1.4

percent of U.S. households and 1.1 percent of the U.S. population. Although most

households earn less than $30,000 annually, much of their income is disposable. Their

primary sources of income are Social Security, interest, dividends and pensions. They are

one of the top-ranked markets for investments and savings. They are also one of the top

markets for buying lottery tickets.

Residential: Mobile and single-family homes account for most of the housing of "Senior

Sun Seekers." They own homes in newer areas, with a median home value of $86,200.

Their neighborhoods also include seasonal housing for visiting snowbirds, plus

congregate housing and nursing homes. Almost 90 percent are in the South or West,

including about 12 percent of Florida and ten percent of Arizona neighborhoods.

Preferences: "Senior Sun Seekers" have income and time at their disposal, and they spend

it golfing, traveling overseas and domestically, playing cards, doing needlework,

gambling in casinos and playing the lottery. They are health conscious, following low-fat

or other diets, and taking vitamins. They enjoy reading senior, health and science

magazines and watching news programs, dramas and situation comedies on TV.









Consumer Type 5A: Twentysomethings

Demographic: The median age of "Twentysomethings" is 30 years, 5.5 years younger

than the U.S. median. Over 27 percent of residents are in their 20s that is double that of

the U.S. percentage. They are mobile and in transition, completing college or starting

their careers. Nearly 60 percent live in single-person or shared households compared to

30 percent for the U.S. overall.

Socioeconomic: Only 1.7 percent of the U.S. population, "Twentysomethings" is a

relatively small market. Their median household income is $26,700. Employment is

average, although half are employed only part time. Education is the key to the future of

"Twentysomethings"; over 35 percent have an associate degree or higher, 20 percent are

attending college.

Residential: "Twentysomethings" live in city apartments. Nearly 20 percent of the

neighborhoods are in California or Texas. Rent is generally below average.

Approximately 22 percent of the housing in these neighborhoods are single-family,

owner-occupied homes. The median home value is $88,800, 10 percent lower than the

U.S. median.

Preferences: The "Twentysomethings" young, active and urban lifestyle reflects their age

and environment. They play racquet and team sports, go hiking, jogging and train with

weights. They frequently visit museums, go to the movies, attend concerts, bars, and

nightclubs. They buy CDs, books, shoes, sportswear, soft drinks and alcohol, and

vitamins. They watch sports and late-night shows on TV; read newspapers; and men's

photography, science and computer magazines.









Consumer Type 5B: College Campuses

Demographic: Not surprisingly, almost three-fourths of "College Campuses" are college

students. Their median age is 21.7 years compared to the U.S. median of 35.5 years. Over

45 percent live in dormitories; the rest, in nearby neighborhoods that are primarily

student housing. Forty percent are under the age of 25; 70 percent are living in single-

person or shared households.

Socioeconomic: This market is either well educated or on its way to becoming well

educated. Over fifty percent have a bachelor's degree or higher, 2.5 times the U.S.

percentage. Another quarter of the population has attended some college or has an

associate degree. Their median household income is $21,000. Over half is employed,

with most working in part-time, low-paying, service-sector jobs.

Residential: Apartment rentals dominate off-campus housing. Over 50 percent of the

housing inventory are multi-unit buildings, with five or more units. One sixth of the

housing is single-family, owner-occupied. Their rent is below the U.S. average gross rent,

but not by much. Most "College Campuses" are located in cities; however, these

neighborhoods also represent the smaller college towns.

Preferences: Aside from college expenses, most of their budget is spent on travel or their

active social lives. "College Campuses" spend much of their free time playing racquet

sports, training with weights, hiking, playing billiards, jogging, and bicycling. They also

visit museums, go to the movies, bars, and nightclubs. They buy CDs and books, PCs,

computer software, beer and wine, and vitamins. They watch a lot of sitcoms, dramas,

late-night shows and sports on TV.









Consumer Type 5C: Military Proximity

Demographic: This young, mobile population has a median age of 23.6 years. Over 40

percent have recently moved. Half of these households are young couples, many with

preschool children, although there are some single-parent families. These racially-mixed

neighborhoods are predominantly black and white with an above- average percentage of

Asians or Pacific Islanders.

Socioeconomic: Over 30 percent of the labor force is in the Armed Forces. Employment

is high for civilians, too, although unemployment is slightly above average. Most workers

have post- secondary schooling, but only 17 percent hold a bachelor's or graduate degree.

Their median household income is $32,500. The incidence of poverty is just above the

U.S. average.

Residential: Almost ten percent of "Military Proximity" live on-base in military barracks.

The housing units in the residential areas are mainly apartment rentals. The median home

value of $82,000 is 18 percent lower than the U.S. figure. "Military Proximity"

neighborhoods are located inside the urban areas of Virginia, the Carolinas, Georgia,

Florida, Texas and California.

Preferences: The expenditures and leisure activities of "Military Proximity" reflect this

market's composition of young families. These households rank high in their purchases

for their youngest: children's clothing, shoes, toys, videos, and vitamins. They enjoy

jogging and weight training, going to bars and nightclubs, eating fast food, and visiting

theme parks. They watch dramas and situation comedies on TV and listening to jazz and

R&B on the radio.









Consumer Type 6A: East Coast Immigrants

Demographic: "East Coast Immigrants" are the new Americans. Almost 40 percent are

foreign-born and most speak a language other than English at home. Their demographic

profile is a rich mix of ethnic and racial groups and household types such as Hispanic,

Asian, and black. Sixty percent are married-couple or single-parent families; forty

percent are single-person or shared households. Their median age is 33.8 years, with

above-average percentages of younger householders aged 20-34 years.

Socioeconomic: "East Coast Immigrants" represent 2.5 percent of all U.S.

households and 2.4 percent of the U.S. population. Their median household income is

$33,100. Unemployment and the lack of a high school diploma are 40 percent higher

among this segment but employment is comparable to the U.S. figure. These workers

commute, over 30 percent cross county or state lines to work.

Residential: Most rent in the urban canyons of large cities in high-density, high-rise pre-

1950s apartment buildings. The average rent is slightly above the national average.

Homeless shelters and other non institutional group quarters are also in these

neighborhoods. Almost half of "East Coast Immigrants live in New York; 70 percent live

along the Eastern Seaboard.

Preferences: Living in urban areas with good public transportation, "East Coast

Immigrants" don't often own cars. Their household expenditures include foreign travel,

long distance phone calls, buying sportswear, and VCRs. They watch situation comedies,

dramas and sports on primetime TV; listen to jazz and R&B on the radio, and read

newspapers and fashion magazines. They take day trips to museums, theme parks, and

gambling casinos.









Consumer Type 6B: Working Class Families

Demographic: "Working Class Families" are approximately 90 percent black and 75

percent family. Although slightly older than people in the U.S. overall, many households

have grade-school age children or teenagers. Single-parent households comprise slightly

less than one of five households at 18.5 percent compared to one of eleven U.S.

households at 9.3 percent. Most are 35 to 74 years old. Their median age is 36.8 years.

Socioeconomic: This small, middle-income market totaling 1.2 percent of all U.S.

households, has a median household income is $37,600. Despite chronic high

unemployment, the employment rate is above average for women and just below average

for men. Most work in service or government jobs. At $41,800, their median net worth is

above the U.S. median of $39,100.

Residential: "Working Class Families" live in urban neighborhoods, with an average

density of 10,000 persons per square mile, but single-unit, owner-occupied homes are

prominent. Almost 15 percent of the housing inventory are row houses, twice the national

average. Built decades ago, these homes have a median value of $71,900. These

communities are located in urbanized areas scattered throughout the U.S. Over 20 percent

of the neighborhoods in the District of Columbia are "Working Class Families."

Preferences: Consistent with their urban lifestyle, "Working Class Families" own few

cars, but when they buy a car, it is a luxury model. They also buy TVs and clothing. They

watch time watching sitcoms, dramas, and sports on TV, gambling at casinos, playing the

lottery, playing team sports like softball, dancing at nightclubs, or going to the movies.

Consumer Type 6C: Newly Formed Households