*For Suraya Only*

The Digital Divide and Internet Voting Acceptance France Bélanger Accounting and Information Systems Virginia Tech 3007 Pamplin Hall Blacksburg, USA [email protected] Lemuria Carter School of Business and Economics name of North Carolina A & T State University 1601 East Market Street Greensboro, North Carolina 27411, USA [email protected] Abstract - Governments and corporations are increasingly considering the use of the Internet for individuals to cast votes.

Yet, not everyone has access to and is comfortable with the use of technology. This is the problem of the digital divide. This study explores the impact of the digital divide on Internet voting (I-voting). A model of I-voting and the digital divide is proposed. The proposed model suggests that age, income, education and frequency of Internet use have an impact on I- voting utilization. To test the model both an online and paper- based version of the survey was administered to a large sample of citizens. The results of multiple linear regressions indicate that age and income (access and skills) have a significant impact on Internet voting. These findings indicate that, like other e-government services, I-voting is subject to the barriers associated with the digital divide, and this digital divide introduces several challenges to government agencies. Keywords: Internet voting; technology adoption; digital divide; e-government I. INTRODUCTION Voter turnout is vital to the health of all democracies. A key element of a democracy is the continuing responsiveness of the government to the preferences of its citizens. Turnout rates in U.S. presidential elections (which are the most popular) vary between 50 and 60 percent with winners never receiving more than 60 percent of the turnout. Hence, presidents are selected by the votes of 25 to 30 percent of the electorate [1]. In fact, the United States ranks at the bottom, or just above last place, in voter involvement when compared to other democratic nations [2]. Research suggests Internet voting could increase voter participation [3]. Internet voting, or I-vot ing, is defined as “an election system that uses encryption to allow a voter to transmit his or her secure and secret ballot over the Internet [4, p. 2].” Researchers suggest that I-voting has the potential to increase “turnout” among individuals between the ages of 18-25 since they have experience in surfing the Net and like the idea of using the latest technology [5]. Morris [6] agrees that the Internet has the potential to mobilize the otherwise disenfranchised voters under the age of thirty-five. Many countries have conducted research on or experimented with Internet voting [7]. In the Netherlands, 62% of the people with access to the Internet would prefer to vote online [8]. In New Zealand, a taskforce concluded that Internet technology might boost the number of voters, speed the count, and reduce costs. In Japan, the Center for Political Public Relations experimented with poll site Internet voting in the 2001 gubernatorial election in Hiroshima. In 2005, Estonia was the first country to offer Internet-voting as an option nationwide for mayors and city councilors [9]. In the United States, the 2000 Arizona Democratic primary offered the first binding Internet election for public office [3, 10-11]. In 2008, Okaloosa County in Florida allowed hundreds of military personnel in Germany, Japan and the United Kingdom to cast their votes in the presidential election [12]. Despite the gradual implementation of I-voting and its potential to increase participation, some citizens may not benefit from this innovation due to the digital divide. The paper is organized as follows. First, we discuss the issues of the digital divide to provide background for the development of the research model. The methodology section describes the research conducted. The next section presents the results and their implications for research and practice. II. THE ISSUE OF THE DIGITAL DIVIDE As governments worldwide begin to implement more technology-based voting systems, in particular Internet voting, concerns about the potential impacts of the digital divide continue to grow. The digital divide refers to the distinction between the information haves and have-nots; the gap between the computer literate and the computer illiterate. The digital divide is composed of two major barriers: access to technology and comfort with technology [13]. Both of these barriers may play a role in limiting the use and convenience of Internet voting. Not surprisingly, researchers have found that demographically, citizens who use the Internet for political purposes differ from the rest of the population, particularly in terms of income and education [14]. It could be because education and income increase the likelihood of openness toward Internet voting [3]; it could also be due to the digital divide barriers of access and skills. 2010 Fourth International Conference on Digital Society 978-0-7695-3953-9/10 $26.00 © 2010 IEEEDOI 10.1109/ICDS.2010.54 317 2010 Fourth International Conference on Digital Society 978-0-7695-3953-9/10 $26.00 © 2010 IEEEDOI 10.1109/ICDS.2010.54 307 A. The Access Divide The access divide refers to factors that may limit an individual’ access to technology that can be used, in this case for Internet voting. Prior research has identified ethnicity, income, age and education as significant predictors of access to technology [15-16]. A more recent study finds that income, education and age significantly impact who is willing to use e-government services such as electronic tax filing or license renewals [13]. This is not surprising since other researchers have found that approximately 78 percent of households with income between $50,000 and $75,000 had Internet access compared to only 40 percent of those with household incomes between $20,000 and $25,000. Others find that young citizens (18- 24) and their parents (45-54) report the highest levels of home Internet access, reaching better than 61 percent[17]. Research also shows that more younger Americans have an Internet connection than older Americans [5]. Thomas and Streib [16] suggest that among Internet users, ethnicity and education are important predictors of government Web sites utilization, with white and better educated users more likely to be uses such sites[16]. Interestingly, gender differences in access and use of computers has narrowed over the years, with recent research suggesting that it does not impact use of e-government services[13]. This is consistent with findings from the Pew Internet Project report which suggests that although men and women have different attitudes toward technology, the surge in the number of women online has eliminated some of the disparity in access between genders [18]. B. The Skills Divide In addition to Internet access, comfort with Internet technology is also a major element of the digital divide. The skills divide refers to a disparity in skills necessary to effectively interact with online systems. Mossenburg, Tolbert, and Stansbury [15] identify two components of this skill divide: technical competence and information literacy [15]. Technical competencies are “the skills needed to operate hardware and software, such as typing, using a mouse, and giving instructions to the computer to sort records a certain way”. Information literacy is “the ability to recognize when information can solve a problem or fill a need and to effectively employ information resources.” Researchers have found that the old, less-educated, poor and minority individuals (African Americans and Latinos) were more likely to need computer assistance (such as help using the mouse and keyboard, using e-mail, or using word processing and spreadsheet programs), although recent studies show some of the differences disappearing after a year or two of use [19]. In this study, we use frequency of Internet use in general as a proxy measure of technical competence and information literacy. The use of this proxy is consistent with Belanger and Carter [20]. Citizens who use the Internet frequently should possess a level of technical and information literacy. C. The Research Model In summary, differentials in age, income, education, and computer skills seem to create a digital divide that should affect which individuals will choose to use Internet voting as a means of performing their constitutional right. Figure 1 summarizes the access and skills divide factors that are expected to affect one’s intention to use I-voting. Intention to Use Internet Voting Income Education Age Internet Use Figure 1. I-Voting Digital Divide Factors III. METHODOLOGY To identify the salient I-voting divide factors, we surveyed a diverse pool of citizens. Both online and paper- based versions of the resulting instrument were administered to participants. There were various sources of data collection for each version. The paper version of the survey was administered to members of a church choir, students in a religious seminary class, attendees of a symphony concert, and employees in a county agency. The online version was posted on a local website, disseminated through a graduate student listserv at a university, and sent to the listserv of a community fitness group. A total of 464 surveys were received, 187 on paper and 277 online. Eighty-seven (87) incomplete surveys were eliminated along with five surveys from participants who were not old enough to vote. Hence, 372 surveys were used for data analysis: 133 paper responses and 239 online responses. An independent samples t-test was used to identify any differences between online and paper responses. Since the two groups did not exhibit differences for the dependent variable - intention to use an I-voting system - a combined sample was used in the data analyses. Regarding sample demographics, the age range of participants is 18 to 75 years with an average of 33 years. Most participants (78%) have a college degree and the reported income range is well distributed. 44% of the sample makes $50,000 (or more) a year. In addition to the demographics mentioned above, general information about the participants was collected. The sample was 63% female.

A majority of the subjects were Caucasian (64%). African- Americans accounted for 26% of the sample and Hispanic, Asian and Native Americans accounted for seven percent of 318308 the sample. The remaining three percent of the subjects did not report ethnicity. In terms of access to and experience with the Internet, most participants reported high levels, with the exception of having used e-government services, where only 70% of respondents indicated having done so, as can be seen in Table 1. TABLE 1. INTERNET AND WEB EXPERIENCE. PERCENTAGE OF RESPONDENTS WHO… have access to the Web at home 91% used the Web to make a purchase 90% had used the Web to complete a government transaction. 70% voted in the 2004 presidential election 82% IV. RESULTS Multiple regression analysis was used to test the model after, assumptions of multivariate normal distribution, independence of errors, and equality of variance were verified. The USE variable was slightly skewed with a mean of 4.78. Correlations were low correlations among variables, except for age and income with a correlation of 0.48. There were no violations of the other assumptions. A. Model Testing The regression analysis results in a model with an F- value of 8.053, resulting in a p-value of p< 0.0001, which indicates that at least one of the coefficients corresponding to an independent variable is not equal to zero. The r-square value was 8 %, indicating that digital divide factors identified in this research ac count for eight percent of the variance in intentions to use I-voting. This is important because this is variance explained on top of what typical adoption factors from theories such as the Technology Acceptance Model (TAM) should account for. Since the model itself is significant, the individual beta coefficient t- tests can be used to test which digital divide factors are significant. Results indicate that age and income are significant predictors of I-voting intentions at the p <0.0001 level. Figure 2 shows the significant results. Figure 2. I-Voting Significant Factors B. Significant Results: Age and Income The results of regression analysis indicate that age and income are significant elements of the I-voting divide.

Younger citizens and individuals with higher income levels are more likely to use Internet voting. These findings indicate that, like other e-government services, I-voting is also subject to the barriers associated with the digital divide. Regarding I-voting, both access (impacted by income and age) and skills (impacted by age) effect I-voting intentions. Citizens with higher income-levels are more likely to have access to the technology necessary to take advantage of Internet voting. Older citizens are not only less likely to have access to the Internet, but also less likely to posses the computer skills necessary to take advantage of Internet voting. As municipalities begin to make I-voting a viable option for civic participation it is imperative that whole sectors of the population are not “left behind.” This digital divide introduces several challenges to government agencies: 1) the sectors in danger of exclusion are already disenfranchised and 2) as long as there is a divide, the government will need to maintain traditional voting methods in addition to Internet options. Older, lower income citizens will need an advocate to ensure that they are not disregarded as I-voting initiatives become more commonplace. The existence of this divide means that I-voting should be used as an accompaniment to, not a replacement of, existing voting procedures. C. Factors Not Affecting the I-voting Divide Interestingly, education and frequency of Internet use did not have a significant impact on one’s intention to use an Internet voting system. With regards to education, this finding could be a result of our sample, which did not have a large variance in education. Seventy eight percent of our survey respondents have a college degree. This percentage is far greater than the population at large. Future studies should continue to explore the effects of education on the digital divide. In addition, frequency of Internet use did not have a significant impact on intention. Regular use of the Internet does not translate into an affinity towards Internet voting. Undoubtedly, online voting will introduce unique concerns, even among frequent Internet users. Future studies should explore the impact of concepts such as Internet trust and Internet self-efficacy on I-voting acceptance. Future studies should also explore the impact of technology adoption variables on intention to use Internet voting. Perhaps, constructs such as, compatibility and social influence, would have a significant impact on I-voting intentions. D. Implications for I-voting Diffusion Government agencies need to discover ways to make online services more appealing to older citizens. The results of this study indicate that younger voters are more inclined to use Internet voting than older citizens. Perhaps government agencies could work with community and/or non-profit organizations designed to help senior citizens, such as the American Association for Retired Persons (AARP) (www.aarp.org), to increase adoption among older users. As senior citizens often become increasingly less mobile, having an easy way to cast their vote could improve 319309 the level of participation of this group of citizens in the democratic process. As I-voting becomes more popular, municipalities also need to make I-voting options available to low-income citizens that may not have Internet access at home. For instance, the government may be able to make voting kiosk available in public places such as libraries, supermarkets and post offices to increase citizens’ access to this innovation. E. Social Impact of Increased Voter Participation The impact of I-voting on political participation cannot be fully ascertained until Internet voting actually becomes a common option for voting in major elections. Recent studies suggest that its diffusion is steadily approaching.

Researchers at the Georgia Tech Research Institute (GTRI) predict that kiosk I-voting will be available at post offices, malls, and automated teller machines. By 2012, they predict that some states, especially Oregon, which only uses mail-in ballots, will be the first to adopt Internet voting [21]. In light of the potential for Internet voting to increase voter participation, it is important to consider the potential impact of increased voter turnout on the nation’s political system. V. CONCLUSION As local and state governments begin to experiment with Internet voting, now is the time to identify the characteristics that distinguish I-voters from non-I-voters. This study identifies digital divide factors that impact one’s intention to use an I-voting system. The results indicate that age and income have a significant impact on I-voting utilization. In particular, younger citizens with higher levels of income are more likely to support online voting initiatives. The factors identified here can serve as a foundation for future studies of the digital divide and I- voting adoption. REFERENCES [1] J. Petrocik, and D. Shaw, Nonvoting in America: Attitudes in Context , New York: Greenwood Press, 1991. [2] R. E. Wolfinger, and S. J. Rosenston, Who Votes? , New Haven: Yale University Press, 1980. [3] R. S. Done, Internet Voting: Bringing Elections to the Desktop , The PricewaterhouseCoopers Endowment for the Business of Government, 2002. [4] A.-M. Oostveen, and P. V. D. Besselaar, “Internet Voting Technologies and Civic Participation: The Users' Perspective,” The Public, vol. 11, pp. 1-18, 2004. [5] M. R. Alverez, and T. E. Hall, Point , Click, and Vote: The future of Internet Voting , Washington, D.C.: Brookings Institution Press, 2004. [6] D. Morris, Vote.com , Los Angeles: Renaissance Books, 1999. [7] C. Eliasson, and A. Zuquete, “An electronic voting system supporting vote weights,” Internet Research, vol. 16, no. 5, pp. 507-518, 2006. [8] W. Pieters, and M. J. Becker, "Ethics of e-voting: An essay on requirements and values in Internet elections," Institute for computing and information sciences 2006. [9] Report of the National Workshop on Internet Voting , Internet Policy Institute, 2001. [10] M. R. Alverez, and J. Nagler, “Internet Voting and Political Representation,” Loyola of Los Angeles Law Review vol. 34, pp. 1115-1152, 2001. [11] F. I. Soloop, “Digital Demo cracy Comes of Age: Internet Voting and the 2000 Arizona Primary Election,” Political Science and Politics, vol. 34, no. 2, pp. 289- 293, 2001. [12] E. Sofge, "Internet Voting in Florida Raises Security Concerns: Geek the Vote," 2008. [13] F. Belanger, and L. Carter, “The Impact of the Digital Divide on E-government Use,” Communications of the ACM, vol. 52, no. 4, pp. forthcoming, 2009. [14] B. Bimber, “Information and Political Engagement in America: The Search for the Effects of Information Technology at the Individual Level,” Political Research Quarterly, vol. 54, no. 1, pp. 53-67, 2001. [15] K. Mossenburg, C. Tolbert, and M. Stansbury, Virtual Inequality: Beyond the Digital Divide , Washington, D.C., 2003. [16] J. C. Thomas, and G. Streib, “The new face of government: Citizen-initiated contacts in the era of E- government,” Journal of Public Administration Research and Theory, vol. 13, no. 1, pp. 83, January, 2003. [17] P. G. Harwood, and W. V. McIntosh, Virtual Distance and America's Changing Sense of Community , New York, NY: Routledge, 2004. [18] R. Pew Internet Project . "Tracking Online Life: How Women Use the Internet to Cultivate Family and Friends," Accessed June 1, 2005; www.pewinternet.org/reports . [19] L. A. Jackson, A. v. Eye, G. Barbatsis et al. , “The impact of Internet use on the other side of the digital divide ” Communication of the ACM vol. 47 no. 7, pp. 43-47 2004 [20] F. Bélanger, and L. Carter, “Trust and Risk in eGovernment Adoption,” Journal of Strategic Information Systems, vol. forthcoming, 2008. [21] J. Sanders. "The Future of Voting: Researchers Explore the Social and Technical Issues of Voting Via the Internet," Accessed March 12, 2006; www.gtresearchnews.gatech.edu/newsrelease/VOTING.

html . [22] J. Citrin, E. Schickler, and J. Sides, “What if Everyone Voted? Simulating the Impact of Increased Turnout in Senate Elections ” American Journal of Political Science, vol. 47 no. 1, pp. 75-90, 2003. 320310