1Over the past few years the Internet has grown in use and importance as an information, entertainment, and business medium. While only a decade ago some people had not even heard of the Internet, today being able to not only go online, but to be able to go online at broadband speeds, has become important if not essential for many Internet users. Having broadband presents important opportunities to residential, business, institutional, and other users of the Internet. For people in rural locations in particular, the availability of broadband can be especially important, as it can create opportunities in areas such as education and healthcare.
2The majority of American adults now have access to and are utilizing the Internet on a frequent basis (Madden and Ranie 2003 ; U.S. Department of Commerce 2004). Email remains the primary application for most residential users of the Internet, and it is clearly an important means for interacting as well as for exchanging information. Other residential uses of the Internet are growing as well, with people increasingly going online to search for product and service information, to make purchases, and for health information, amongst other things (U.S. Department of Commerce 2002).
3Broadband, as opposed to a dial-up Internet connection, is necessary in terms of many of the audio, video, and data applications that the Internet has to offer (Hausman, Sidak and Singer 2001 ; Malecki 2003). Some ‘standard’ Internet applications are practical only through having a broadband, rather than a dial-up, Internet connection, and like the use of the Internet more generally, the use of broadband services is growing rapidly (FCC 2004). Thus, in areas where broadband is not available, or for members of the population for whom the costs of such services are prohibitive, lack of broadband access can mean being removed from a technology that has fast become a part of many peoples’ daily lives.
4In relation to social networking opportunities in particular, broadband offers tremendous possibilities for citizens and professionals living in rural areas. For example, broadband can enhance communication within and between school districts ; it allows teachers and administrators to participate in distance learning career development programs ; and it can improve opportunities for workforce training more generally (Center for Innovative Technology 2002 ; FCC 2002 ; Mackay 2001). Similarly, broadband offers possibilities for rural health care practitioners to participate in distance training and assistance, including opportunities that augment face-to-face mentorship, enhance information sharing, and allow for consultations during surgery. More generally, broadband can help diminish the professional isolation of healthcare practitioners in rural areas (Skorga 2003). For local governments, broadband offers prospects for increased and more effective communication with the public ; enhanced interaction with state and federal agencies ; and heightened opportunities to interact with businesses and industry (Unette 2003). Finally, and as previously suggested, for residential users broadband enhances opportunities for sharing information through email, and for a range of opportunities related to social networking more generally. In short, broadband enhances possibilities for decreasing social and professional isolation in rural areas.
- 1 According to the 2000 U.S. Census. Information available at www.census.gov.
- 2 For example, close to 80 percent of the U.S. population lives in urban areas, a percentage that is (...)
5As broadband continues to increase in importance, it becomes imperative to assess how its use and availability may vary, especially in relation to issues such as income, race, and geographic location. The primary task of this research, then, is to assess the availability of broadband service in rural areas. In particular, this research examines the extent to which Digital Subscriber Line (DSL) and cable modem services have been deployed throughout Pennsylvania, a U.S. state with large urban and large rural populations. Close to a quarter of Pennsylvania’s population – almost three million residents – live in rural areas1. The goal of this analysis is to compare the availability of broadband services between rural and urban areas in what might be considered a representative U.S. state2. The two types of broadband services that are examined in this analysis account for more than 95 percent of all residential broadband services and a large percentage of business broadband services that are currently being utilized in the U.S. (FCC 2004). In short, DSL and cable modem services are the most widely employed and often the only economically viable broadband options for many residential and small business users, including and perhaps especially in rural areas.
6Research has long identified a lack of advanced telecommunications services in rural areas of the U.S. (e.g., Gabe and Abel 2002 ; Malecki and Boush 2003 ; Parker et al. 1995 ; Strover 2001), including prior to broadband becoming a relatively widespread and important phenomenon. For example, utilizing data collected in 1998, Strover (2001) found less Internet availability in rural rather than in urban areas in four U.S. states. Also using data from 1998, Malecki and Boush (2003) found that in the Southeast U.S., telephone company central offices in rural areas were less likely to have digital switches than central offices in urban areas. Similarly, Gabe and Abel (2002) found that ISDN was more widely available in urban rather than in rural areas. In short, it is clear that telecommunications providers have been, over time, more likely to make investments in advanced technologies in urban rather than in rural areas, with population density being a key factor in relation to such investments (Glass, Chang and Petukhova 2003 ; Malecki 2003 ; Malecki and Boush 2003 ; NTIA and RUS 2000 ; Strover 2001).
7In conjunction with the growth in broadband use over the past few years, various analyses have assessed the differences in broadband availability between urban and rural areas in the U.S. Such analyses have come from different sources, including the federal government (FCC 2002 ; FCC 2004 ; NTIA and RUS 2000) ; state government agencies (California Public Utilities Commission 2005 ; Governor’s Office of Appalachia 2002 ; Iowa Utilities Board 2006 ; Maryland Technology Development Corporation 2002 ; Texas House of Representatives 2002) ; and academic scholars (Gabel and Kwan 2001 ; Prieger 2003 ; Strover 2003).
- 3 For example, the data collection procedures have the potential to both over- and under-represent br (...)
8The most noteworthy federal government research in terms of broadband availability is the Federal Communication Commission’s (FCC) ongoing data collection efforts and related analyses of broadband availability at the zip code level (FCC 1999; FCC 2002; FCC 2004). These data, known as form 477 data, are collected by the FCC at six-month intervals. FCC analyses utilizing these data demonstrate ubiquitous availability of broadband services in urban areas, and growing availability of such services in rural areas (FCC 2004). As indicated in Figure 1, form 477 data indicate a clear growth of broadband availability in rural areas over the first few years of this decade. There are, however, some important caveats to keep in mind in relation to these data3.
Figure 1: High speed internet availability in low-density zip codes
Note : Data from FCC 2004. According to the report, low density zip codes have fewer than six persons per square mile. For definition of high speed Internet services, see FCC 2004 or footnote 5 in this paper.
9Perhaps not surprisingly given the ready availability of form 477 data on the FCC website, a number of scholars have utilized these data for their own analyses. For example, Strover (2003) utilized form 477 data from June 2000 and found that zip codes in non- metropolitan counties as well as counties that were non-adjacent to metropolitan areas had fewer high speed providers than did zip codes in metropolitan or metropolitan adjacent counties. Amidst the broader goal of attempting to determine whether or not redlining was taking place in relation to broadband supply, Prieger (2003), who also used form 477 data from June 2000, found amongst a range of findings that lower density zip codes were less likely to have broadband service.
10Not all academic analyses have utilized FCC data as a basis for research. Grubesic (2003) utilized data primarily collected from the World Wide Web to examine the difference in cable modem and DSL service availability between urban and rural areas of Ohio. His research indicated a disparity in broadband availability between urban and rural areas. Gabel and Kwan (2001) also primarily utilized web-based information, collected in February 2000, to assess broadband availability in relation to various geographic and socioeconomic characteristics. Again, amidst various findings and similar to the previous academic analyses discussed thus far, their work indicated an urban-rural broadband divide.
11As previously suggested, various state government and regulatory agencies have sponsored research related to the availability of broadband in rural areas. Perhaps the most notable research in this regard is a recent report from the Iowa Utilities Board (2006), which has conducted a series of broadband analyses over the years through their own, community- by-community data collection efforts. The Board’s most recent data suggest that broadband is fairly widely available throughout the state. In fact, the data indicate that rural communities in Iowa actually have a higher rate of broadband availability than do non-rural communities. More than 90 percent of Iowa communities have DSL available, which apparently is at least in part due to relatively aggressive state regulatory measures that encourage telephone companies to deploy broadband (Iowa Utilities Board 2006).
12Like previous government and academic analyses, this paper explores the extent to which a broadband divide exists – and persists – between urban and rural areas of Pennsylvania. While previous academic research indicates a notable divide in such regards, and some government research in the past has indicated an almost complete lack of broadband availability in rural areas (NTIA and RUS 2000), more recent and longitudinal analyses have indicated that the broadband digital divide between urban and urban areas in the U.S is narrowing (FCC 2004). Utilizing primary data collected from telecommunications providers, this research examines the extent to which this divide exists in Pennsylvania, a state with a relatively large number of rural residents.
13A number of important details are worth noting in relation to the methods utilized to conduct this analysis. One matter that clearly needs to be addressed in “broadband” related research is the definition of broadband itself. As suggested in a report by the California Public Utilities Commission, “there are perhaps as many definitions of broadband as there are organizations and countries that have attempted to define it" (California Public Utilities Commission 2005, p. 3). The National Telecommunications and Information Administration (NTIA) suggests that the term “has many meanings depending upon application,” while the International Telecommunication Union (ITU) defines broadband as “the capacity to transmit large quantities of electronic signals (including data, video, text, and voice) rapidly.”4 Some people assign specific speeds of service to the meaning of broadband (e.g. Dodd 2002 ; Goleniewski 2002), though perhaps it is most useful to suggest that the definition of ‘broadband’ is in a constant state of flux, especially considering that telecommunications technologies and related speeds of information transmission are constantly improving. In order to have a baseline standard, however, this paper utilizes the definition for ‘broadband’ services as defined by the FCC (FCC 2002 ; FCC 2004)5. Moreover, and regardless of the speeds being offered, the cable modem and DSL technologies discussed and assessed in this research have become synonymous with the term broadband in the U.S.
- 6 At the time this research was conducted, there were virtually no Competitive Local Exchange Carrier (...)
- 7 The figure regarding total number of cable companies was based on a combination of information avai (...)
- 8 Many providers – especially the ILECs – considered information about the location of their broadban (...)
- 9 Information about this provider’s cable modem service was available from researchers at the Univers (...)
14A key element for conducting this research was information collected by the author through interviews with members of Incumbent Local Exchange Carriers (ILECs) and cable companies operating in Pennsylvania6. These interviews were conducted over the telephone, primarily during the first three months of 2003. At the time of this research, 26 ILECs and approximately 100 cable television companies were operating in the state, with providers in both industries ranging from small, independent ‘mom and pop’ companies to much larger companies with a national presence7. All of the ILECs were contacted to participate in this analysis. A total of 21 were willing to participate8. All of the medium and large cable providers in the state were contacted to participate in this research. Of these companies, only one large provider was unwilling to cooperate9. For the smaller cable providers, a systematic random sampling technique was used to determine which providers would be contacted. Two small cable companies did not return phone calls, and after leaving a number of messages they were no longer contacted. In the end, 41 cable companies provided data for this analysis.
- 10 See, for example, Dodd 2002 ; Goleniewski 2002 ; Grubesic and Murray 2002.
15Data for measuring DSL availability were collected from ILECs in relation to individual wirecenter (a.k.a. central office) locations and corresponding service areas, of which there are a total of 841 in the state. The full technical details related to providing DSL services are not discussed here, and readers are referred to other resources if they are interested in further information about such matters10. In short, interview participants from ILECs were asked whether or not their company had a functioning Digital Subscriber Line Access Multiplexer (DSLAM) in relation to any one of their given wirecenters in Pennsylvania. Respondents were also asked if they were utilizing remote DSLAMs to provide DSL services.
- 11 The 18,000 distance limitation was confirmed by providers. Technological advancements at the time o (...)
16To spatially assess DSL availability, the primary data were incorporated into a GIS. Wirecenter and wirecenter service area GIS “shapefiles” were purchased from a private company. In the GIS, a given wirecenter was coded as either being able to provide DSL, or unable to provide DSL. Buffers were then created around wirecenters that had DSL technologies. These buffers accounted for the approximate 18,000 foot distance sensitivity of DSL11. Where necessary, these buffers were then clipped to match wirecenter service area boundaries. Furthermore, where providers indicated that they were utilizing remote DSLAM technologies to provide service throughout an entire wirecenter service area, that given service area was treated entirely as a ‘DSL service area,’ and no buffering was needed. ILEC wirecenter service areas where no DSL was being provided were treated as ‘non-DSL areas.’ In short, through the utilization of GIS software, along with the primary data collected through interviews, ‘DSL coverage areas’ were created for Pennsylvania. The GIS techniques utilized in this aspect of the analysis are somewhat similar to techniques utilized elsewhere to analyze DSL availability (e.g. Grubesic and Murray 2002 ; Pinkham Group 2002).
- 12 See, for example, Abe (1997) ; Patterson and Rolland (2002).
17A somewhat similar process was utilized to determine cable modem coverage areas. Again, the technical details related to providing cable modem services are not discussed here ; readers are referred to other resources in such regards12. The data collection process for cable companies involved determining the presence or absence of cable modem service for any given cable company franchise area, which almost invariably corresponds with sub-county political units such as cities, townships, or boroughs. Individuals at cable companies were asked if they were providing cable modem service in any of the company’s given franchise areas. This data collection process typically involved reading through a list of franchise areas where an individual company was providing cable television service, and asking the respondent whether or not the company was providing cable modem service in that franchise area.
- 13 Generally, if the company was making broadband available to 50-80 percent of its cable subscribers (...)
18In short, cable companies provided information about service availability at the franchise level, with franchise areas typically being minor civil divisions such as townships, boroughs, and villages, which are often only a few square miles in geographic extent. Because cable companies do not always provide cable television service, let alone broadband service, throughout any given franchise area, in instances where providers suggested that they were providing broadband to only part of a franchise area, assessments were made between the researcher and the interviewee regarding the extent to which such services were being offered13.
19Minor civil division GIS data available through the U.S. Census website was used to map cable franchise areas and related cable modem service availability (U.S. Census 2003). Through the use of this GIS data it was possible to develop a direct match between the majority of cable franchise areas and geographic units that could be included in a GIS. Specifically, more than 93 percent of the franchise areas could be directly mapped through using this U.S. Census shapefile. As was the case with the DSL coverage areas, ‘cable modem service areas’ and ‘non-cable modem service areas’ were created in the GIS.
- 14 Of these block groups, 36 had no population and were therefore eliminated from the analysis. For fu (...)
20Having developed coverages of DSL and cable modem service availability for much of Pennsylvania, the next task in this research was to assess broadband service availability in relation to urban-rural population. Urban-rural population data were available from the U.S. Census (U.S. Census 2000), and the scale of data used in this analysis was at the census block group level. The 2000 U.S. Census included 10,387 block groups within the state of Pennsylvania14. In many instances, and given that the number of block groups in the state outnumbered the civil division and wirecenter coverage area information used to determine broadband location, it was typically the case that any single block group was entirely within or entirely outside of a broadband coverage area. In instances where this was not the case, a GIS centroid approximation technique was used to determine block groups that had, and did not have, broadband.
- 15 Definitions for this information are available at www.census.gov.
- 16 Statement based upon the author’s own analysis of data for such areas in Pennsylvania.
21This research utilized census definitions of “metropolitan area,” “urban,” and “rural” to make urban-rural determinations, and residents of any given block group were classified as living in either a “metropolitan,” “non-metropolitan urban,” or a “non-metropolitan rural” area15. Non-metropolitan block groups could contain both “urban” and “rural” residents, and such instances were accounted for in this research. Without getting into the specific details of U.S. Census designations regarding these matters, “metropolitan” areas in Pennsylvania include the state’s largest cities as well as their suburbs. They also include cities with a minimum population of 50,000. Generally speaking, the non-metropolitan urban areas in this research can be understood as urban areas that typically have populations ranging from 5,000 to 15,000, often serving as places of employment for people in outlying rural areas. The non-metropolitan rural areas are either especially small towns, frequently with populations of 1,000 or less, or are low-density population areas outside of such towns or areas without a central location in particular16.
22In summary, the metropolitan, non-metropolitan urban, and non-metropolitan rural designations as described above, along with the broadband availability information developed through the primary data collection process, were used to determine levels of broadband supply to citizens of Pennsylvania based upon where they resided. Using overlay techniques in a GIS, residents in block groups that had broadband were included as having broadband service available, while those outside of such service areas were deemed to not have broadband service available.
- 17 See Warren (2001) for information about number of homes passed. Cable companies not building in the (...)
23Finally, it is of note that in any area of the state where no data were collected during the interview process, such as in ILEC service areas belonging to companies who were unwilling to participate in this research, no assumptions were made regarding broadband availability. Only areas where broadband or cable modem service was known to be provided or not provided were included in the GIS. As is perhaps always the case, there are various caveats to keep in mind regarding the results of this analysis. Perhaps most importantly, some technical issues related to conducting a GIS analysis of this type have likely resulted in the overrepresentation of the availability of broadband services in the state. For example, a cable company may not have infrastructure passing all homes in any particular franchise area. ‘Total homes passed’ in any given franchise area is typically between 85-100 percent, and cable companies often do not build out their infrastructures to the more remote, ‘rural’ areas, should such areas exist, in a given franchise location17. For this analysis franchise areas such as townships were determined to either have or not have service, and in some instances homes in these franchise areas may not have had access to the given cable company’s infrastructure.
24With the ILECs there were issues that likely have led to the overrepresentation of the availability of DSL as well, including issues of telephone line quality that may be problematic in accessing DSL, even within the 18,000 foot buffer zone of ILEC wirecenters. Nonetheless, such problems could occur in rural, urban, or metropolitan areas. Finally, some of the providers that were unwilling to participate in this research were very likely to not be offering broadband services, and some remote rural areas of the state have no cable providers at all. Nonetheless, while the results of this analysis probably overestimate broadband availability, particularly in non-metropolitan rural areas, the techniques utilized in this research as well as the primary data collection process result in a highly accurate depiction of broadband availability in metropolitan, non-metropolitan urban, and non-metropolitan rural areas of the Pennsylvania.
25The results of this analysis suggest that there is a difference in broadband availability between metropolitan, non-metropolitan urban, and non-metropolitan rural areas of the state of Pennsylvania. This research indicates that broadband is available, from either a cable provider, ILEC provider, or both providers, to 98 percent of the state’s metropolitan population that has been included in this analysis (See Figure 1 below). Broadband is less widely available in non-metropolitan urban and non-metropolitan rural areas. However, broadband is fairly widely available in such areas. A total of 89 percent of the state’s non- metropolitan urban population and 73 percent of the state’s non-metropolitan rural population included in this analysis have broadband service available in their locales.
26In short, in terms of having broadband from at least one provider, this research demonstrates that broadband supply is virtually ubiquitous in Pennsylvania’s metropolitan areas. It is somewhat less available to non-metropolitan urban and non-metropolitan rural populations in the state. While there is a broadband divide in Pennsylvania, this divide is not especially dramatic. Although broadband is available to virtually all of the state’s metropolitan population, it is also available to much of the population living in non-metropolitan urban areas, and it is available to many residents living in non-metropolitan rural areas.
27From simply an availability standpoint, the urban-rural digital divide in broadband supply in Pennsylvania is not especially large. Generally speaking and as further confirmed through interviews with the providers, non-metropolitan urban areas, many of which serve as economic and service centers for outlying rural areas, have broadband available from either a cable or ILEC provider. Even many of the small, rural towns, many of which are classified as non-metropolitan rural areas, have broadband available. The areas of the state that do not have broadband service from either an ILEC, cable company, or both, are generally the least densely populated, ‘most rural’ areas of the state. In short, this analysis confirms trends seen elsewhere (cf. FCC 1999 ; FCC 2004 ; Iowa Utilities Board 2006 ; NTIA and RUS 2000). The urban-rural broadband divide has apparently, somewhat dramatically, narrowed.
Figure 2: Broadband availability by location
Figure 3: Broadband by location and provider
28This research also indicates a difference in broadband availability based upon type of provider. Across all areas analyzed in this research – metropolitan areas, non-metropolitan urban areas, and non-metropolitan rural areas – cable companies were more likely to be providing broadband services than were ILECs (See Figure 3). Both types of providers were more likely to be offering broadband in metropolitan rather than in non-metropolitan locations. In particular, this research found that ILECs were making DSL available to 83 percent of the residents in their metropolitan service areas, but to only 62 percent and 21 percent of the residents living in non-metropolitan urban and non-metropolitan rural service areas, respectively. The disparity in availability in cable modem service was not quite as dramatic between metropolitan and non-metropolitan areas. Cable providers were offering cable modem service to 95 percent, 71 percent, and 62 percent of their metropolitan, non- metropolitan urban, and non-metropolitan rural customers, respectively.
29As was clear through interviews with cable as well as telephone providers, the most important factor associated with broadband availability relates to population density, though factors such as income, especially for cable companies, can also play a role in whether or not broadband service is provided. Investing in the infrastructure necessary for broadband in densely as opposed to less densely populated areas offers cable as well as telephone providers a much greater return on their investments. Through the interviews it was clear that for many providers in the state, whether cable or telephone providers, ‘return on investment’ was the primary driver in broadband deployment. Clearly, in terms of the costs of deploying the infrastructure necessary to supply broadband, the returns on such investments are greater in urban rather than in rural areas.
30Nonetheless, such commentary should not detract from the major finding in relation to this analysis : many rural areas in Pennsylvania have broadband available from an ILEC, cable company, or both providers. The areas of the state that do not have such services available tend to be remote rural locations, marked by no central location or very low population density. Broadband is clearly available in Pennsylvania’s metropolitan areas, but it has arrived in many of the state’s rural areas as well. Obviously, broadband supply does not necessarily mean access or use. While this study found that the costs for broadband were similar across both urban and rural areas – typically in the range of $ 30-$ 40 per month – such costs remain prohibitive for many members of the population. In rural areas in particular, where average incomes are typically lower than in urban areas, broadband remains unaffordable for many residents. In short, broadband supply does not necessarily address all digital divide concerns. For members of the population for whom the costs of such services are prohibitive, supply does not necessarily equal access. Moreover, and as will be discussed in further detail in this paper’s conclusion, aside from simply connectivity and costs, other constraints can inhibit benefiting from the opportunities that broadband might otherwise provide as well.
31This research has provided a detailed analysis of broadband availability in Pennsylvania, including the variations in broadband availability between metropolitan and non- metropolitan areas of the state. The results indicate that broadband is offered in many of the state’s rural communities. In short, trends as can be seen through this and other analyses indicate that in much of the U.S., broadband is available to many rural users, especially in rural towns. It is the most remote rural areas that are unlikely, and that may be unlikely in the future, to have broadband services. Users in such locations will, for the time being, need to rely upon dial-up, satellite, or wireless Internet services, all or some of which might be available in any given rural location, and technologies such as WiMax hold great promise for broadband in rural areas in the near future.
32While there are clear advantages to harnessing some of the opportunities that broadband has to offer, such as telemedicine, distance learning, and increased business competitiveness, detailed and empirically robust examinations of the extent to which businesses and institutions are utilizing this technology in rural areas are limited. This is likely in part due to the relatively recent phenomenon of rural areas having access to such technologies. Considering the range of business, institutional, and residential uses of broadband, as well as the possible barriers to its use, there is a clear need to critically probe the issue of how rural communities are utilizing or might otherwise utilize broadband in ways that effectively promote economic development, social and professional networks, and quality of life in rural areas more generally. Broadband is only a tool that provides opportunity to the extent that individuals – whether in a residential, business, or institutional setting – are effectively utilizing it. Issues such as knowledge of broadband applications, or unwillingness to fully adopt new technologies, can remain a barrier to achieving some of the rewards that broadband might otherwise offer. Research and policy should focus on such matters.
- 18 It is of note here that some policymakers in Pennsylvania, and elsewhere in the U.S., arguably have (...)
33Thus, future research and policy efforts related to broadband could focus on issues in relation to broadband use in rural areas18. While continued, vigilant assessments of broadband availability in rural areas will remain important in the years ahead, researchers and policymakers could work to develop comprehensive overviews of the broadband applications available to small business, medical, educational, government, and residential users in rural areas. Through such efforts, information could be developed that would assist community leaders, businesses and institutions, and policymakers to ascertain the extent to which broadband is being utilized to promote development in such areas. The possibilities as previously discussed in this research – in relation to issues such as distance learning and telemedicine – are clearly now available to many rural users through broadband, especially in business and in institutional contexts. Members of rural communities, policymakers, and academics must now figure out a way to effectively harness this technology, especially where broadband connections are now available.