The Cliff and the Continuum:
Defining the Digital Divide

Davis Foulger

davis@foulger.info
http://davis.foulger.info

Visiting Associate Professor
Oswego State University
Oswego, NY
http://www.oswego.edu/~dfoulger

Evolutionary Media
http://www.evolutionarymedia.com

December 1, 2001
Abstract Added: March 13, 2002

Inspired by discussion at the IAMCR & ICA Symposium on Digital Divide: November, 2001.

This paper can be downloaded page formatted in MicroSoft Word.

Comments to this paper can be posted to Cliff and Continuum page on my Wiki.

Abstract

The digital divide is generally defined in terms of unequal access to digital and network resources. In practice, however, there is a fundamental confusion as to what the "digital divide" is. For the larger number of papers at the IAMCR-ICA Digital Divide Symposium, the divide was often a function of psycho-social and economic factors that affected the decision to make use of such resources. For a smaller number of papers, choice was not the issue: people in "digital have not countries" have little real prospect of using digital resources even when they have the means and inclination to do so. While papers in both groups defined the digital divide in terms of unequal access to internet resources, there were consistent and substantive differences in the concerns of the papers in each group that suggest that the there are systematic connotative differerences underneath their shared definition of the digital divide. This paper explores these differences using an 188 country dataset drawn from the U.N. and other data sources. It identifies a continuum of Internet use for people living in those countries in which Internet use is reasonably widespread, and is therefore to some extent a matter of choice that is informed by psycho-social issues. For people living in the countries for which Internet use is almost non-existant, however, the digital divide is a cliff that probably has to be scaled in increments. Indeed, there are at least three cliffs that need to be scaled by "digital have not" countries. The first, literacy, is an issue for the 100+ countries that contain 3/4s of the the world's population. The second, infrastructure (as measured by telephone use), is an issue for over 130 countries. It is only when we scale the third cliff and reach the world's 50 or so "digital have" countries, that we see a clearly discernable continuum of use.

Introduction

"Could someone please define the Digital Divide?"

Not exactly a first choice for a question to face after speaking at the "Defining the Digital Divide" session of the IAMCR-ICA Digital Divide Symposium (Austin, TX, 2001). The need for the question suggests a collective failure, on the part of session presenters (Stanley, 2001; Newhagen and Bucy, 2001; Foulger, 2001), to successfully engage the topic. This was understandably a source of confusion for the speakers at the session, as each of the three papers had provided fairly strong statements about the nature of the digital divide. There were differences, of course, but none that would suggest that the digital divide is about anything other than the profound differences in access to digital resources (e.g. the Internet) that separates "digital have" countries and people from "digital have not" countries and people. People on the "digital have" side of this divide, in general, have computer or terminal based access to a wide variety of new communication media and information that is unavailable to people on the "digital have not" side. Where then, might the confusion come from?

The Digital Divide Symposium Divide

It can be argued, with some small risk of oversimplification, that the presentations at the IAMCR-ICA Digital Divide Symposium can be classified into two groups (that there was a Digital Divide Symposium Divide :-). A smaller set of papers looked at the "global digital divide", either on a global (Foulger, 2001; Jacobson and Barnett, 2001; Monge and Matei, 2001; Shukla, Kandath, and Rogers, 2001; Thussu, 2001) or regional (Akbar, 2001; Brasil, 2001; Rao, 2001; Sinha, 2001, Sonaike, 2001) basis. Papers in this group discussed the differences between the smaller number of "digital have" countries and the larger set of "digital have-not" countries. People in digital have countries, in general, have fairly high levels of internet availability. While Internet access may or may not be ubiquitous in homes, it is certainly commonly found in business, school, and community access settings such that people who want Internet access can get it. Digital have not countries, by contrast, tend to have such low levels of Internet access that one can reasonably assert that, for most people, Internet access is not a choice. Most people would find it difficult or impossible to get to digital resources even if they wanted to.

A much larger set of papers focused on the remaining digital divide within digital have countries (Aguayo, 2001; Findahl, 2001; Göktepeli and Christensen, 2001; Fuentes and Straubharr, 2001; Hope and Hoar, 2001; Horrigan, 2001; Krasnoboka, Iossifov, and Rathmann, 2001; McConnaughey, 2001; Mun, 2001; Niece and Mansell, 2001; Rudd, 2001; Weerakoddy, 2001). In these papers the focus was not countries, but people (Ball-Rokeach, Kim, Loges, and Jung, 2001; Berquist and Pugh, 2001; Hawkins, Pingree, Gustafson, Julesberg, McTavish, and Stengle, 2001; Kaigo, 2001; Kim, Kim, Lin, and Cheong, 2001; Stanley, 2001), communities (Abbott, 2001; Crick, 2001; Drake and Battaglia, 2001; Mäkinen, 2001; Park, 2001; Pecora, Riffe, Real, and Krendl, 2001), organizations (Scott, 2001), schools (Seiter, 2001; Webber and Clark, 2001), cultures (Biocca, 2001; Krotz, 2001), demographics (Cintorino, 2001; Gopalakrishnan, 2001, Nicholas, 2001), content (Bucy, 2001; Newhagen and Bucy, 2001), ownership (Bates, Miller, and Raber, 2001; Kleinwaechter, 2001; Stein and Sinha, 2001), privacy (Colby, 2001; Phillips, 2001), policy (Brown and Bernt, 2001; Courtright and Robbin, 2001; Steward, Pileggi, and Gil-Egui, 2001), and the media (Gustafson, 2001). The primary question in these papers was not one of making it possible for people to choose to have Internet access, but understanding why some people choose not to take advantage of digital resources in countries where large numbers of peoople do.

There were other obvious differences between papers in the smaller set concerned with the "global digital divide" when contrasted with the larger set of "local digital divide" papers:

Implications for Defining the Digital Divide

In these differences one finds a possible source of the confusion about the definition of the digital divide. The digital divide may be the differences in access to digital resources (e.g. the Internet) that separate "digital haves" from "digital have nots". The nature of the separation is different, however, when it is a matter of choosing to use digital technology rather than having no choice at all. In digital have countries, the digital divide is, as many speakers at the conference observed, really a continuum of choice. Some people choose to make extensive use of digital resources. Others do not. Most people fall somewhere in between. In digital have not countries, there is little or no choice. A small number of people may have access to digital resources, but most people have none. Choice is not an issue.

Is should not be surprising, given these differences, that a study focused on the digital divide in a digital have country would ultimately define the digital divide somewhat differently than would a study focused on the global digital divide. The differences will be subtle, and may not be explicit. A discussion of the global digital divide might note that 90% of the world's population that lack even the choice to access digital resources. A discussion of a digital have country's digital divide might discuss the minority (large or small) that choose not to access digital resources. While the issue, in both definitions, remains the divide between digital haves and digital have nots, the differences in scale (world versus country), direction (majority versus minority), and choice (have choice versus have none) change the fundamental assumptive ground of these two very different kinds of studies. The authors of papers in each group really are using different definitions of the digital divide.

Continuum or Cliff: A methodology

It is possible to illustrate this difference using the dataset constructed by Foulger (2001). This dataset, which is drawn from publicly available UNESCO spreadsheets and other public sources, describes a number of statistical measurements for 190 countries. The important measures, for the purposes of this study, are the three variables that most closely describe a country's Internet readiness (number of Internet hosts, number of Personal Computers, and number of telephone lines), all expressed as some variant of units per thousand population. This study uses k-means cluster analysis (as offered in SPSS), to draw a continuum of Internet readiness. A series of cluster analyses were performed on these variables, with each requesting a different number of clusters ranging from 2 to 20. Even with pairwise deletion of missing values, only 188 countries clustered (e.g. two countries had missing values for all three variables). These clusters will be presented here using a combination of tables and graphs. For reasons that will become apparent during the discussion, the results of a selection of these cluster analyses will be presented as both "raw" clusters and as a reduced set that combines strongly related clusters.

Number of Clusters
Largest number of hosts
Intermediate Cluster Sizes (N of Countries)
Smallest number of hosts
2
10
178
3
4
9
175
4
3
3
9
173
5
2
3
7
16
160
6
2
3
3
4
17
159
7
2
1
2
3
4
17
159
8
2
1
2
1
3
3
17
159
9
2
1
2
3
1
3
13
18
145
10
2
1
2
2
2
3
5
8
18
145
11
2
1
2
2
2
1
3
5
7
18
145
12
2
1
2
2
2
1
3
5
2
7
24
137
13
2
1
2
2
2
1
3
5
2
7
11
38
112
14
2
1
2
2
1
1
1
3
5
2
6
12
37
113
15
1
1
2
2
1
1
1
3
5
5
2
3
11
38
112
16
2
1
2
2
1
1
1
3
5
5
2
3
9
14
33
104
17
2
1
2
1
1
1
1
3
2
4
5
2
3
9
14
33
104
18
2
1
2
2
1
1
1
3
5
1
5
5
3
7
5
7
33
104
19
2
1
2
1
1
1
1
3
2
4
1
4
2
3
2
8
7
35
108
20
2
1
2
1
1
1
1
3
2
4
1
1
4
5
3
3
7
7
35
104
Table 1: The size (in countries) of clusters resulting from the k-means cluster analysis of three variables for 188 countries. The variables include Internet Hosts per hundred thousand population, computers per thousand population, and telephone lines per thousand population. In this table, clusters have been resequenced based on the mean value for internet hosts per hundred thousand population. Countries with higher numbers of internet hosts per 100,000 appear to the left. Countries with lower numbers appear on the right. This same left to right structure will be retained for all charts and tables in this paper.

 

Table 1 shows the evolution of the clusters over the course of the cluster analysis. The countries represented by the rightmost column dominate the cluster analysis results, capturing at least 104 of the 188 countries in every cluster analysis. These 104 countries can be regarded as true "digital have not" countries which not only lack Internet hosts, but both the computers to run them on and the telephone lines through which they might connect to the Internet. The columns to the left reflect a breakdown of at least comparatively "digital have" countries into related clusters. As will be seen, some of these countries aren't much better off than the 104 "digital have not" countries that emerge as a stable cluster at 16 clusters. Because the clusters are, for the most part, fairly stable starting at 16 clusters, and to a considerable extent up to 16 clusters, we will use the 16 cluster solution for the remaining analysis. Table 3, at the bottom of this paper, shows the 16 clusters, the countries associated with each cluster, and 5 converged clusters which will be described below.

The Continuum

There is, as numerous speakers at the digital divide symposium indicated, a digital continuum. This continuum is clearly visible in Figure 1, which shows the number of Internet hosts per hundred thousand people for a series of country clusters. As can be seen in the figure, there is, depending on how you prefer to view it, either a curvilinear relationship or a series of three continua. The first of this series (or the steep part of the curve) entails the first three clusters (five countries: the United States, Finland, Iceland, Norway, and Taiwan). The second series (or the middle of the curve) entails at least 5 clusters (13 countries including Canada, Australia, New Zealand, Sweden, Denmark, United Kingdom, Netherlands, Germany, Switzerland, and Austria). A third continuum (the flat part of the curve) the remaining clusters, and such countries as Japan, Korea, France, Italy, Spain, and Ireland, as Internet hosts approaches zero.

Bar of mean(hkhosts) by ihost16
Figure 1: Internet hosts per hundred thousand, clustered into 16 groups.

 

In the interests of simplifying the remaining analysis, we will recombine these continua into five clusters as shown in Table 2. The value of doing this should be plain for the clusters that describe only one or two countries (the norm for members of the first and second recombined clusters). Three of the new clusters are built from the three continua of Figure 1 (the first three old clusters, the next six old clusters, the next five old clusters). The last two new clusters, which between them account for almost 3/4's of the countries in the analysis, correspond to the last two clusters of the old analysis. The alert reader will note that old cluster 9 has been associated with new cluster 2 rather than new cluster 3 (the cluster which Figure 1 appears to suggest it should belong to. The cluster analysis is based on three variables. Figure 1 only shows one of them. The other variables suggest that old cluster 9 may be more closely related to old clusters 4 to 8. In Table 1, moreover, cluster 9 (of 16) turns out to be something of an isolate, but one whose subsequent decomposition ties it more tightly to old clusters 4 to 8. This recombined clustering will be used for most of the remaining analysis.

16 Raw Clusters
2
1
2
2
1
1
1
3
5
5
2
3
9
14
33
104
5 Clusters Combined from 16
5
13
33
33
104
Table 2: The recombined clusters based on the clearly visible continua of Figure 1.

Figure 2 shows the data of Figure 1 displayed in five clusters. There is still a steep continuum through the first three clusters (the old clusters 1 through 14). More obvious in this rendition just how internet host poor the 137 countries in clusters 4 and 5 are when compared with the 51 countries of clusters 1 through 3. Countries in these last two clusters are not operating under the same rules as the countries in the first three clusters.

Bar of mean(hkhosts) by ihost516
Figure 2: Internet hosts per hundred thousand, clustered into 5 groups.

These last two clusters are important, however, as between them they represent almost 90% of the world's population (see Figure 3). Indeed, we see from Figures 2 and 3, an intersection of rapid declines that might be called cliffs. A cliff of digital have countries (Figure 2) declines rapidly through approximately 50 digital divide countries while a cliff of digital have not populations declines rapidly through about 135 others (Figure 3).

Bar of sum(kpop) by ihost516
Figure 3: Population, in thousands, of the countries in each cluster.

There is, then, a cliff at the end of the continuum formed by the world's 50 or so most Internet-ready countries. Countries in the continuum may vary in the extent to which digital media resources are distributed within their populations, but there is some level of general access upon which further growth can build. In the most digitally ready of these countries (those in clusters 1 and 2), bridging the digital divide really is more an issue of choice than access. People who want Internet access in these countries can, in general, achieve it, whether or not they have access from their homes. In the least ready of these countries (cluster 5), however, choice is not the problem. In countries where the number of Internet hosts is less than 50 per hundred thousand, most people simply do not have the opportunity to use digital media. As Figure 4 shows, moreover, this divide may be difficult to remediate. When the mean GNP per capita for a country is less than the cost of an Internet ready PC, it is difficult to see how people in that country can obtain access to computers and the internet in anything resembling large numbers. Countries that already have digital resources, on the other hand, have the resources to obtain more, and continuing declines in price of digital media access should bring the Internet within the reach of increasing numbers of people.

Bar of mean(capigdp) by ihost516
Figure 4: The mean Gross National Product (GNP) per capita for the various clusters. Note the sharp falloff in resources available per capita in the digital have not countries of cluster 5.

The cliff is, however, more pernicious than a simple lack of resources with which to buy digital resources. There are, in fact, much more problematic cliffs separating digital have countries from digital have not countries.

The Cliffs

The first of these cliffs is literacy. As can be seen in Figure 5, literacy rates in the countries of cluster 5 are much lower than they are for countries in the first cluster. Literacy rates in the countries of first four clusters exceed 90%. Hence literacy is not a barrier to entry for most people in these countries. In cluster 5, however, literacy rates average only 65%, and are much lower in some countries, particularly those in Africa. Literacy is a fundamental in the use of today's Internet resources, and while progress continues to be made on computer hardware and software that can convert speech to text, text to speech, and one language to another, they currently work for only a few languages. The mechanisms that do such translations work far better today than they did twenty years ago, but they are imperfect at best, and you can't really know if they are working properly unless you are literate enough to be able to cross-check the results.

Bar of mean(plitrat) by ihost516
Figure 5: Literacy rates for the various clusters. Note the sharp falloff in literacy for the digital have not countries of cluster 5.

A second cliff can be found in the existing telephone infrastructure. Telephone infrastructure matters to digital access because it is, in general, the telephone network, that provides the first level of access to Internet services. There are other choices for provision of such access, including cable television and power line infrastructures, where they are available, but the telephone infrastructure is generally the infrastructure of first resort because of the ready availability and low cost of modems. Indeed, telephone infrastructures usually depend upon, and are built on the backbone of, the power infrastructure. Cable infrastructures most generally ride the same infrastructure, and as the premium alternative to broadcast television, usually trail the availability of telephones. Figure 6 shows that the cliff for telephone infrastructure falls through cluster 4. The average number of telephone lines per thousand in cluster 5 is very low - on the order of 25 main telephone lines per thousand people. Cluster 4 does better, with an average approaching 200 main telephone lines per thousand, but at this level it cannot reasonably be expected that telephone distribution penetrates much beyond the business and government sectors, or that there is any room for dedicating existing phone lines to anything but voice traffic.

Bar of mean(kcaptel) by ihost516
Figure 6: Main telephone lines per thousand people for the various clusters. Note the sharp falloff in existing telephone lines after cluster 3. While the digital have not countries of cluster 5 are particularly lacking in such infrastructure, the relative digital have countries of cluster 4 also lack adequate telephone infrastructure to fully bridge the digital divide.

A third cliff, visible in Figure 7, reflects the number of computers available, per thousand people, in countries in the various clusters. In this case, the cliff runs through cluster 3, an indication that, even in digital have countries, access to digital media is not as high as it is in the eighteen countries of clusters 1 and 2. Cluster 3 is a particularly interesting datapoint, as its internet hosts (Figures 1 and 2) are low by comparison to its computers, even though its telephone network access (Figure 6) is relatively strong. The difference, one suspects, is in the mean GNP per capita (Figure 4). People in these countries may well have adequate resources to handle one time expenses like computers, but may have difficulty maintaining additional regular costs such as those associated with computer network access.

Bar of mean(kcappcs) by ihost516
Figure 7: Mean computers per thousand people for the various clusters. Note the sharp falloff in computers per thousand between clusters 2 and 3. While clusters 4 and 5 are particularly lacking in computer resources, countries in cluster 3 clearly lack the computer resources they need to keep up with countries in clusters 1 and 2.

The Cliffs and the Continuum

There clearly is a digital continuum that divides the populations of digital have countries, on an individual basis, into digital haves and have nots. There are variations in the levels of digital access in digital have countries, and the reasons for these variations in access almost certainly differ somewhat as one moves from the high GNP per capita countries of cluster 1 and 2 to the moderate GNP per capita countries of cluster 3. Variations in cluster 1 and 2 countries are more likely to revolve around issues of personal priorities and preferences, including various social pressures, a result that comes out strongly in several of the papers presented at the Digital Divide Symposium. The lower GNP per capita of cluster 3 countries makes personal budgetary priorities a more likely player in the decision making of individuals in these countries. The decline in computers per thousand is consistent with this kind of economic decision making. So is the mild decline in telephone lines per thousands associated with this cluster. The still steeper declines in the GNP per capita of cluster 4 and 5 countries presents even greater challenges to individuals. Indeed, cluster 4 and 5 countries will have difficulty providing general access to the Internet on any basis.

This is the cliff at the end of this continuum; the cliff that separates digital have not countries from digitally wired countries; the cliff that separates 5 billion people from the Internet resources that 500 million of us take for granted. People in cluster 5 countries do not, by and large, have the choice of using Internet media or other digital resources. If they had the money, and they don't, they aren't sufficiently literate to ensure that digital media would be broadly usable. If they were literate sufficient to using such media they don't have a telephone infrastructure (or reasonable alternative) sufficient to ensure that digital media could be widely accessed. If they had a telephone infrastructure sufficient to enabling wide access to internet resources, they still don't have the computers they would need to take advantage of the telephone infrastructure. All of this can be bought, of course, but the low GNP per capita of these countries is where the problems begin.

The Cliff and the Continuum

The digital divide is, then, both cliff and continuum. Indeed, it is not one cliff but several, and the continuum is found not only in the personal decision making that characterizes the digital divide of digital have countries, but in the mounting of the individual cliffs of resources, literacy, network infrastructure, and computer access that characterize the different clusters of countries in this analysis. Here then, I offer two definitions of the digital divide:

For digital have countries: The digital divide is the continuum of use of Internet and other digital media that separates those that choose, for whatever reason, to use such media from those who choose not to use such resources.
For digital have not countries: The digital divide is the cliff that separates the five billion people who cannot, for whatever reason, choose to use Internet and other digital media from the half billion or so people who can choose to use such resources.

The key difference here is is the issue of choice. The digital divide is one thing when you do not have the option of crossing it. It is something very different when you have the option but choose not to do so. The digital divide (or digital cliff) of digital have not countries is a divide without choice, and it can only be bridged by resolving the economic, infrastructure, computer access, and literacy gaps that separate digital have not countries from digital have countries. The digital divide (or digital continuum of digital have countries is a one that is increasinglly a matter of choice. It will be bridged by making the use of Internet resources so inexpensive, appealing, and indispensable that even those who would prefer not to use such resources will feel obligated to do so.

One can argue, as many do when facing the digital continuum of digital have countries, that the digital divide does not need to bridged. People should be free to make their own decisions about which media work for them, much as they make decisions about food, transportation, housing, and other elements of daily living. It is, after all, their choice. This same argument does not apply to the digital cliff of digital have not countries. The world would be enriched if everyone who would choose to make use of the Internet and other digital media was able to do so.

References

All papers referenced are 2001 Papers presented at the IAMCR-ICA Digital Divide Symposium in Austin, TX.

Abbott, Erik. "Digital Divide or Digital Quilt?" Rural Communities, Farmers and the Digital Transformation".

Aguayo, Lucía Castellón. The Digital Divide in Chile.

Akbar, Md Shahid Uddin. Beyond the Digital Divide: Bangladesh Aspect.

Ball-Rokeach, Sandra, Yong-Chan Kim, William E. Loges, and Joo-Young Jung. Measuring Ecological Relationship Between the Internet and Individuals: Revisiting the Internet Connected-ness Index.

Bates, Benjamin, Tamara Miller, and Douglas Raber. Copyright Policy and the Digital Divide: A Social Economic Perspective.

Berquist, Lon and Rondella Pugh. The City Role in Creating Digital Opportunities.

Biocca, Frank. HomeNetToo: Using Cultural and Cognitive Style Research to Help Close the Digital Divide.

Brasil, Antonio Claudio. Exploring Alternative Television News in the Internet: A New Approach to Restraining the Digital Divide in Brazil.

Brown, Duncan H. and Phyllis W. Bernt. Framing the Debate: The Use of the Term ‘A Digital Divide’ in Congressional Hearings.

Bucy, Erik P. The Warmer Side of the Digital Divide: Emotional and Evaluative Responses to Online Content.

Cintorino, Margaret. Lessons From a Computer Skills Program for Low Income Teenagers.

Colby, Dean. Closing the Digital Divide: The Imperative of Anonymous Networking.

Courtright, Christina and Alice Robbin. Deconstructing the Digital Divide in the United States: An Interpretive Policy Analytic Perspective.

Crick, Gene. Community Networking Efforts.

Drake, Barbara A., and Peter A. Battaglia. How Community-Based Organizations Impact the Digital Divide.

Findahl, Olle. What Does the Digital Divide Look Like? The Example of Sweden.

Foulger, Davis A. Media Aristocracies, Network Resources, and the Global Digital Divide: Seven Bridges Across the Digital Divide.

Fuentes, Martha & Joe Straubhaar. Improving Public Internet Access in Brazil: Moving Beyond Connectivity.

Göktepeli, Miya and Christian Christensen. Defining and Curing the Digital Divide: EU and US Approaches.

Gopalakrishnan, T.R.. Beyond Digital Divide: An Exploration of ICT Application for Poverty Alleviation Initiatives.

Gustafson, Karen (2001). Changing Conceptualization of the Digital Divide in Mainstream US Press Coverage.

Hawkins, Robert, Suzanne Pingree, David H. Gustafson, Karen Julesberg, Fiona McTavish, William Stengle. Is It Feasible for the Disadvantaged to Ride the Information Highway? Preliminary Report of a Pilot Project for Breast Cancer Patients.

Hope, Wayne, and Peter Hoar. Internet Development and the "Digital Divide" in New Zealand.

Horrigan, John. Internet Access and Surfing Patterns Among Different Groups.

Jacobson, Thomas and George A. Barnett. A Political Solution To the Global Digital Divide.

Kaigo, Muneo, and Teruyoshi Sasaki. Cognitive and Affective Factors of New Information and Communication Technology Usage and the Digital Divide in Japan.

Kaiser, Scott and Kenneth Rogerson. Bridging the Digital Divide Through Local Area Computer Technology Centers.

Kim, Yong-chan, Joo-Young Kim, Wan-Ying Lin, & Pauline Cheong. (2001). Internet Connectedness of Teenagers in Seoul, Singapore and Taipei.

Kleinwaechter, Wolfgang. Digital Divide in the Domain Name System.

Krasnoboka, Natalya, Iordan Iossifov, and Tim Rathmann. A Digital dimension of the European Divide: Ukraine and the Netherlands.

Krotz, Friedrich. Paths Into the Future Media Society: How the Diffusion of PC and CMC Depends on Social and Cultural Conditions and What May Be Concluded for the Digital Divides.

Mäkinen, Maarit. The Internet as a Community Media: Bridging the Digital Divide - Making Information Networks Part of Everyday Life in the Neighbourhoods of Tampere, Finland.

McConnaughey, James. The Multiple Years of Analysis Using Census Data and Why the Digital Divide Continues to Exist.

Monge, Peter and Sorin Matei. The Impact of Globalization on the Digital Divide.

Mun, Seung-Hwan. Bridging a Two-Tiered Information Society: A Study of the Bandwidth Divide in Texas.

Newhagen, John E. and Erik P. Bucy. Routes to Media Access: Apprehending Internet Content.

Nicholas, Kyle. Rural Access Issues.

Niece, David C., and Robin Mansell. Inside Tier II of the ‘Digital Divide’.

Park, Han W. Digital Divide Among Community Network Users in Korea.

Pecora, Norma, Daniel Riffe, Michael Real, and Kathy Krendl. (2001). The Reality of the Digital Divide: The Appalachian Region of Ohio.

Phillips, David J.. Constructing a Privacy Divide: Structuring Differential Protection of Privacy Interests.

Rao, Sandhya. Urban Digital "Haves" and "Have Nots" in India: What Difference Does Internet Access and Usage Make?

Rudd, Tim. Digital Divides in Britain and the Work of Bourdieu.

Scott, Craig R. Digital Divides Within and Between Organizations: Differences in Access to Various Technologies, Key Communication Partners and Relevant Organizational Information.

Seiter, Ellen. The Digital Divide at Elementary School: Ethnography of an After-School Computer Class.

Shukla, Pratibha, Krishna P. Kandath & Everett M. Rogers. (2001). The Internet and the Digital Divide in Africa, Latin America and Asia.

Sinha, R.M.K. Multilinguality and the Global Digital Divide.

Sonaike, S. Adefemi. Internet and the Dilemma of Africa’s Development.

Stanley, Laura D. Beyond Access: Defining the Digital Divide.

Stein, Laura, and Nikhil Sinha. Information Access vs. Information Control: Intellectual Property Policy in the Digital Era.

Stewart, Concetta M., Mary S. Pileggi & Gisela Gil-Egui. (2001). Examining the Digital Divide: Toward a New Theoretical Framework for Policy-Making in the Cyberage.

Thussu, Daya. The Global Digital Divide and a Privatized Intelsat.

Webber, Scott, and Lynn Schofield Clark. At Least He’s Reading: Ethnography, the WWF and Computer Use at Public Schools and a Community Center.

Weerakoddy, Niranjala D. Technology and Power: The Intranet and Marginalisation at a State-Owned Organization in Rural Australia.

Appendix: The Clusters from the Cluster Analysis

Converged Cluster
Size
Cluster
Size
Countries
1 5 1 2 Taiwan (Asia), Finland (Europe)
2 1 United States (America, North )
3 2 Iceland, Norway (Europe )
2 13 4 2 Andorra (Europe), New Zealand (Oceania )
5 1 Sweden (Europe)
6 1 Australia (Oceania)
7 1 Denmark (Europe)
8 3 Canada (America, North), Netherlands, Switzerland (Europe )
9 5 Israel, Singapore (Asia), United Kingdom, Germany, Austria (Europe)
3 33 10 5 Ireland, Belgium, Estonia, Slovenia (Europe), Japan (Asia)
11 2 San Marino, Luxembourg (Europe )
12 3 Czech Republic, Hungary (Europe), Tonga (Oceania)
13 9 Republic of Korea, Cyprus (Asia ), Greece, Portugal, Spain, Italy, France, Monaco (Europe), Antigua and Barbuda (America, North )
14 14 Malta, Croatia, Slovakia, Latvia, Lithuania, Poland, Bulgaria (Europe ), Kuwait, United Arab Emirates (Asia), Uruguay (America, South), Bahamas, Barbados, Saint Kitts and Nevis (America, North ), South Africa (Africa)
4 33 15 33 Chile, Brazil, Colombia, Venezuela, Argentina, Suriname (America, South), Brunei Darussalam, Qatar, Bahrain, Turkey, Saudi Arabia, Malaysia, Georgia, Armenia, Lebanon (Asia), Seychelles, Mauritius (Africa), Costa Rica, Panama, Trinidad and Tobago, Saint Vincent and the Grenadines, Belize, Dominica, Jamaica, Grenada, Saint Lucia (America, North), Belarus, The FYR of Macedonia, Republic of Moldova, Ukraine, Yugoslavia, Russian Federation, Romania (Europe)
5 104 16 104 Tunisia, Gabon, Libyan Arab Jamahiriya, Sierra Leone, Ethiopia, Somalia, Eritrea, Rwanda, Comoros, Congo, United Rep. of Tanzania, Madagascar, Burundi, Mali, Zambia, Mozambique, Sudan, Malawi, Nigeria, Niger, Gambia, Angola, Chad, Burkina Faso, Guinea-Bissau, Liberia, Uganda, Sao Tome and Principe, Guinea, Djibouti, Benin, Cape Verde, Dem. Rep. of the Congo, Equatorial Guinea, Kenya, Senegal, Central African Republic, Togo, Côte d'Ivoire, Ghana, Mauritania, Cameroon, Lesotho, Zimbabwe, Egypt, Morocco, Botswana, Namibia, Swaziland, Algeria (Africa), Dominican Republic, Mexico, Haiti, Cuba, Honduras, Nicaragua, El Salvador, Guatemala (America, North), Ecuador, Guyana, Bolivia, Paraguay, Peru (America, South ), Iran, Thailand, Oman, Cambodia, Yemen, Afghanistan, Tajikistan, North Korea, Bhutan, Nepal, Myanmar, Laos, Bangladesh, Turkmenistan, Azerbaijan, India, Viet Nam, Maldives, Pakistan, Kyrgyzstan, Mongolia, Iraq, Syrian Arab Republic, Sri Lanka, Uzbekistan, Indonesia, Kazakhstan, Jordan, Philippines, China (Asia ), Nauru, Fiji, Tuvalu, Kiribati, Vanuatu, Marshall Islands, Samoa, Papua New Guinea, Solomon Islands (Oceania), Albania, Bosnia and Herzegovina (Europe)
Table 3: 188 countries, grouped into 16 clusters and 5 converged clusters based on "Internet Readiness" (using the variables Internet Hosts per hundred thousand, personal computers per thousand, and main telephone lines per thousand). The sixteen clusters were identified in an SPSS k-means cluster analysis. The 5 converged clusters are based on an examination of Figure 1 (above). Note that, within each cluster, countries are grouped by region (e.g. Europe, Asia, North America, South America, Africa, Oceania), with the region identification following the grouping.