Publish to Newswire
Add an EventFrom the Open-Publishing Calendar
From the Open-Publishing Newswire
Indybay Feature
Privatizing the Public Trust: A Critical Look At Connected Nation
This report is by Public Knowledge, The Media and Democracy Coalition and Common Cause.
As a result of the passage and signing of the new stimulus legislation, there is now up to $350 million available to map the deployment of broadband services across the country. The data collected as a result of this effort will be one of the important factors in the national broadband strategy plan the law directed the Federal Communications Commission (FCC) to construct.
As a result of the passage and signing of the new stimulus legislation, there is now up to $350 million available to map the deployment of broadband services across the country. The data collected as a result of this effort will be one of the important factors in the national broadband strategy plan the law directed the Federal Communications Commission (FCC) to construct.
Across the country, states have already begun their own efforts to determine where broadband service is being offered and have already allocated millions of dollars to the effort. As a general matter, trying to figure out the lay of the land is a productive exercise. However, there is a great danger that the process of data collection and, as a result, the national broadband map and plan, will be harmed by an organization known as Connected Nation.
In order to be effective, a national broadband data-collection and mapping exercise should be conducted by a government agency, on behalf of the public, with as granular a degree of information as possible and be totally transparent so that underlying information can be evaluated.
Connected Nation is none of those and represents none of those characteristics. It is an organization sponsored by the telephone and cable companies and represents their interests in deciding what data to collect and how information should be displayed. They are quite up front about their company sponsorship and, in fact, believe it is an asset, if in a way counter to solid public policy.
It would be a setback for our broadband policy if Connected Nation were to take a prominent role in broadband mapping and data collection if it continues on its present policy course because the organization does not represent wise public policy and because it distorts its results. Kentucky Gov. Steve Beshear (D) was correct in April, 2008, when he vetoed a $2.4 million appropriation for Connect Kentucky, which until then had received almost $7 million from the commonwealth. Beshear said that the program was being rejected for state financing because it had asked for funds “without specifically identifying any services to be rendered to the state or providing for any oversight, control or performance measures relative to the services being rendered.”
<strong>Connected Nation Represents Bad Public Policy</strong>
Connected Nation is not a neutral broker in broadband information. It is run by, and boasts of its connections to, telephone and cable companies. Yet, it accepts public funds in the millions of dollars to conduct a public function—mapping of broadband.
The end result is a project from Connected Nation which, instead of reflecting neutral information on which good public policy can be based, instead represents only the information that the most interested of parties wants reported.
Quite simply, Connected Nation’s strategy is to accept public funds for collecting information from its sponsors which is then kept largely private, hidden behind strict non-disclosure agreements (NDA). This privatized function is a violation of the public trust.
It was through the lobbying of Connected Nation that legislation (S. 1492) requiring broadband mapping has such restrictive terms that allow the companies involved, not the government, to determine which information should be made public and would best serve the public interest. It is at Connected Nation’s insistence that it, and its state-level operations, such as Connect Tennessee or Connect Ohio, insist on keeping much of the information they collect hidden from public review. The original Connect Kentucky was also a lobbyist on behalf of telephone companies for pro-industry, anti-consumer legislation.
Let’s take a look at the Connect Board of Directors. There are 12 outside directors, eight of which are directly in the orbit of network operators. They are not small players.
James W. Cicconi – AT&T senior executive vice president-external and legislative affairs
Steve Largent – CTIA – The Wireless Association president and CEO
Joseph W. Waz – Comcast senior vice president, external affairs and public policy counsel
Larry Cohen – Communications Workers of America president. CWA is in frequent agreement with telecom companies on policy issues.
Thomas J. Tauke – Verizon executive vice president for public affairs, policy and communication
Walter B. McCormick – United States Telecom Association president
Kyle E. McSlarrow – National Cable and Telecommunications Association president
Grant Seiffert – Telecommunications Industry Association president. (The members are the equipment makers who sell their gear to the telecom industry.)
These individuals, and others, are listed as “national advisors” on the Connected Nation Web site. They are listed as “directors” in their filing with the Kentucky Secretary of State.
As the Kentucky Public Service Commission told the FCC:
“As often recognized by public policy makers across the country, specific private business interests are not always consistent, or even compatible, with broader public interests. The Kentucky Commission believes a sound and honest public policy initiative to promote broadband deployment must anticipate and account for such discrepancies. In going forward with this national broadband mapping collection effort, which the Kentucky Commission supports, the Kentucky Commission petitions the FCC to remain diligent in keeping consumer interests at the forefront of this endeavor. ”
The PSC recommended that any data mapping information be verified independently from its source. That’s a good suggestion, and it runs counter to the philosophy of Connected Nation. Instead, CN relies on very strict non-disclosure agreements to limit what can be done with the data it collects. In North Carolina, the NDA requires that the maps and web sites used to show broadband coverage “may not differentiate between general broadband service types (such as DSL, cable, fixed wireless, BPS and others) and may not, at a pinpoint, address level, identify broadband Providers at a given location.” (The agreement also stipulates that any information collected remains the property of AT&T, and can be returned or destroyed at any time.) The NDA attached to this report was used by AT&T to collect information on behalf of the Connect North America organization it sponsors.
This clause in the agreement means that the broadband maps produced by Connected Nation or its franchise operations around the country simply show that a company has some service on some street. For consumers or policymakers, there may be no indication of what the technology is, at what speed, or at what price. If the carriers have their way, there never will.
The American Public Power Association, in a 2008 resolution, observed that CN “is ‘public’ only in the sense that it receives taxpayer money from the state [Kentucky] government; its activities are purely private.”
Indeed, Connected Nation will go to great lengths to keep its information collected from its supporters shielded from public view. It argues that only because of its close connections with telecom carriers does a state get any information at all. As one CN official told a gathering of state utility commission staff about whether states could decline to work with carriers or even do the mapping themselves:
“The question would go to provider cooperation in this process. And unless there is legislation that passes in the individual states that mandates the collection of this information, there is [sic] a lot of things that can happen that the provider community might not want to participate. They may not want to provide [the information]. It is not just as easy as the states utilizing the resources they have; there is the other consideration to consider, and that is provider participation in the process. It may be, and I am not saying that this is good or bad, and providers definitely do not definitely want to provide their proprietary information to a state agency, that [carriers] could tie this up in the courts for a while, too. And that is something that would have to play itself out.”
Indeed, in North Carolina, a Connected franchise is working to duplicate work being done by a state agency, the e-NC authority, which attempts to compile its own maps from information supplied by carriers and from its own estimates. The Connect North Carolina operation, run by AT&T, supplies information to itself that it would not supply to the state.
The maps compiled by Connect are inadequate and inaccurate. It is some times hard to discern which definition fits at any given moment. There is a distinct lack of useful information on the maps, such as what data speeds are being offered at what price at any given location.
Indeed, the basic information on the maps, that service of whatever type is available, is open to question because CN, rather than collect granular information by door-to-door canvass, assumes that every spot within a range of a cell tower or telephone company wire center is being served. That is not the case. And it can take dozens of steps and clicks through the cumbersome map interface to reach the inadequate or inaccurate information.
In sum, as a group of municipal utilities told FCC Commissioner Copps in July, 2008, “Broadband data must be collected and delivered in a transparent, verifiable manner. The CK/CN model doesn’t do that: Data is collected, interpreted and reported by a private non-profit entity and shielded from government and public input, oversight and verification.”
<strong>Connection Nation’s Achievements are Over-Stated</strong>
Before state governments hand the reins of policymaking to Connected Nation and their corporate backers, they should ask the basic question: Does the Connected Nation program actually work? That is, do the millions of taxpayer dollars spent by Connected Nation actually impact broadband deployment or adoption beyond what would have occurred otherwise?
Despite the state and Federal embrace of the Connected Nation model as the ideal approach to promoting broadband deployment and adoption, there is absolutely no proof whatsoever that their program actually produces positive benefits, whether from outside evaluators or from Connected Nation itself.
One outside study, “Closing the Rural Broadband Gap,” done by Michigan State University (available at <a href="https:/~larose/ruralbb">https://www.msu..edu/~larose/ruralbb), examined four rural counties that received broadband deployment grants from the USDA. This study found that one of the four counties, Pike County Kentucky, “registered the strongest gains among older adults and those with less than a high school education compared to the other three counties.” The study attributed the changes to “a possible impact of the Connect Kentucky initiative targeting disadvantaged populations in that state,” and then generalized the conclusion to the rest of the state.
If the study can only point to one of Kentucky’s 120 counties in which CK accomplished its goal, then the organization’s claims of success are open to challenge. In addition, there could be any number of external factors involved, which the study did not attempt to account for. For example, in two of the four counties studied, the USDA grants were awarded to non-traditional wireless ISPs, while the grant recipient in the third county went out of business. Only in the fourth county—Pike County Kentucky—did the grant go to an established provider, South East Telephone Company. Thus, the changes observed by the researchers are likely due to the impact of the USDA funding and South East Telephone Company’s own efforts, and not the result of the Connect project. Merely citing the “possible impact” of Connect Kentucky without proper controls and without considering alternative causes is not rigorous economic analysis.
The organization’s only claims of success are contained in a single “study” that is so flawed it has been characterized as committing “methodological malpractice” in an Aug. 4, 2008 filing at the Commission by Consumers Union, Consumer Federation of America, Free Press, and Public Knowledge. And this “study” comes from a group that blatantly attempted in its comments to the Commission to hijack the FCC’s data collection and mapping proposals and replace them with their own, industry-governed and supported, mechanisms.
The Connected Nation study from February 2008 claims “Kentucky’s broadband adoption rate is higher than the national trends due to Connected Nation’s first statewide broadband expansion program, Connect Kentucky.” However, the methodology used to support this claim is fatally flawed. Unfortunately, several press outlets have uncritically repeated the results of this bogus study, leading to the false impression that the Connected Nation model has been proven a success.
Connected Nation could have conducted a proper programmatic evaluation using established and accepted quasi-experimental techniques. But it did not. Instead the group chose an analytical approach that would easily be recognized as flawed by a first year statistics student. The entire claim for success is based on the fact that between 2005 and 2007 the broadband adoption growth rate was 83 percent in Kentucky versus 57 percent for the U.S. as a whole. That is, in 2005 (according to surveys conducted by Connect Kentucky) approximately 24 percent of Kentucky homes subscribed to broadband, which had increased to 44 percent by 2007 -- a “growth rate” of 83 percent. At the same time the national broadband adoption level (according to surveys by the Pew Internet and American Life Project) went from 30 percent to 47 percent -- a “growth rate” of 57 percent.
From this, Connected Nation claims that if “the national growth rate between 2005 and 2007 were applied to the 2005 Kentucky baseline (24%), then Kentucky’s expected statewide adoption in 2007 would be 37%. However, Kentucky’s broadband adoption percentage is actually 44% in 2007, which represents 297,000 more subscribers above the expected adoption rate. The intervening factor has been Connect Kentucky”.
But there is a major flaw in this approach, one of such gravity that it is hard to see how an organization of Connected Nation’s supposed professional caliber could make innocently.
The flaw lies in the comparison of “growth rates” of an initially low performing state against the national average. It should be obvious that improvement by a subject with a low performing metric almost always results in greater percentage (not percentage point) gains when compared to improvements made by a subject that starts with a higher performing metric. This is especially true when comparing the improvement of a low performing state with the average improvement of all states.
The national rate average is just that -- an average of all states, both low and high performing. It is composed of many states that already had high broadband penetration in 2005, and thus didn't have much room to improve to 2007. It would certainly be expected that the national rate of improvement to be somewhat lower than the rate of improvement of some of the lower ranking states -- simply because they had more room to improve. Indeed, using Form 477 data to examine changes in broadband penetration over the 2002 to 2006 reveals that the states with the highest percentage change in broadband penetration were the worst performing states in 2002, and their percentage improvement were far higher than the national average.
The nationwide improvement over this period was 179 percent. Kentucky, which was ranked 50th in penetration in 2002 improved 544 percent; but Montana (ranked 49th in 2002), which had no mapping program intervention had a 568 percent improvement. Alaska had the lowest improvement over this period (97 percent), but was also ranked 3rd in penetration in 2002. Simply stated, a big percentage improvement by a low performing state is unremarkable, especially compared to the nationwide average improvement.
When examining an entity (such as a state or nation) that is maturing towards a saturation point (such as technology adoption), those with lower adoption rates almost always have larger percent increases in growth over time, when compared to entities that begin with higher adoption rates. To illustrate this, consider the following example: According to the U.S. Census Bureau, the percent of urban homes subscribing to broadband went from 11 percent in 2001 to 54 percent in 2007, a 400 percent (not percentage point) increase. During this same time, the percentage of rural homes with broadband went from 6 percent to 39 percent, a 555 percent increase. Also during this time the national household level broadband adoption rate went from 10 percent to 51 percent, a 430 percent increase.
Thus, if we are to measure broadband success using the Connected Nation standard, the rural areas of the United States are greatly outperforming the “national average” and greatly outperforming the urban areas. By the same logic, America should seek to emulate the broadband policies of nations like Turkey and Greece, who experienced infinite percent increases in broadband adoption since 2001 (because in 2001, these countries had no broadband at all, and percentage growth from zero is by definition and infinite percent increase). No one of credibility would make such claims, but it is the exact methodological basis that underlies the “proof” of Connected Nation’s success.
These examples indicate is that the percent change (not percentage point change) in broadband adoption is a completely meaningless statistic.
Connected Nation has continued to make these bogus claims despite critiques of their underlying methodology. In comments to the Federal Communications Commission (urging the Commission to not collect its own broadband data) Connected Nation boasts of the success of the Connected Tennessee program, stating “[i]n just the first six months of the Connected Tennessee program, home broadband adoption has doubled the national growth rate. Rural areas have seen the most significant increases [in growth rate], as home broadband adoption increased by 37% over a six-month period.” It is frankly embarrassing that the group would make such a statement; because we certainly would expect the rural areas to have the most significant percentage increases in adoption -- precisely because rural areas start from the lowest levels of adoption.
If Connected Nation really desired to conduct a proper evaluation of their program they should compare (using accepted quasi-experimental program evaluation techniques) the change in performance of a state where they have intervened to the change in performance of other similar states (in terms of initial broadband adoption and other relevant characteristics) that had no such program. It is telling that Connected Nation has avoided conducting or commissioning such an evaluation, despite having the reason and resources to do so.
But even if we ignore for the moment the bogus nature of Connected Nation’s data, we must focus on the possibility that outside forces other than the non-profit’s are responsible for changes in Kentucky and Tennessee. In their comments, Connected Nation states “Kentucky increased from an estimated 60% of households passed in the state of Kentucky prior to the beginning of the program, to 95% at the end of 2007.”
FCC data indicates that the percent of ILEC lines that were DSL-capable in the state of Kentucky went from 60 percent in June of 2005 to 87 percent in June of 2007. But in December of 2006, as a part of its consent decree to merge with BellSouth, AT&T committed to 100 percent DSL availability in the entire BellSouth territory. Thus, it is quite possible that much of the improvements seen in Kentucky and Tennessee (former Bell South territory) are due to AT&T’s commitments, and not to Connected Nation’s mapping and demand stimulation efforts.
<strong>Additional Resources</strong>
Connect Kentucky Provides Uncertain Model for Federal Legislation – Jan. 9, 2008
http://www.publicknowledge.org/node/1334
Connect Kentucky Update: Broadband Tax Plan Ditched – Feb. 4, 2008
http://www.publicknowledge.org/node/1382
Connect Kentucky Disconnected at Home – April 14, 2008
http://www.publicknowledge.org/node/1521
Connected Nation's Private Interests Hit in FCC Comments – July 24, 2008
http://www.publicknowledge.org/node/1675
Connected Nation Takes Aim At Stimulus Broadband Mapping; Rural Areas Could Be Hurt – Feb. 17, 2009
http://www.publicknowledge.org/node/1998
Virginia Uses Self-Help Program for Rural Broadband – Feb. 20, 2009
http://www.publicknowledge.org/node/2003
Chicago Tribune: $7.2 billion plan to wire rural America holds promise, pitfalls
http://www.chicagotribune.com/news/chi-wired-for-webfeb22,0,4823241.story
Column from Charlotte Observer: A Blind Curve on the Information Highway – Feb. 20, 2009
http://www.charlotteobserver.com/346/story/548379.html
Coverage from IndyWeek, Durham, N.C. Stories by Fiona Morgan
State Lawmakers Debate Using Industry-backed Info on Broadband Access – Feb. 4, 2009
http://www.indyweek.com/gyrobase/Content?oid=oid%3A279831
Building a Better Broadband Map – Feb. 5, 2009
http://www.indyweek.com/gyrobase/Content?oid=oid%3A280381
Telecom Industry Brings Connected Nation to North Carolina – Dec. 31, 2008
http://www.indyweek.com/gyrobase/Content?oid=oid%3A272764
Article by Drew Clark of Broadband Census.com
http://arstechnica.com/tech-policy/news/2009/02/infrastructure-investment-decisions-need-transparency.ars
Connected Nation affiliates (as listed on Web site)
Connected Tennessee
ConnectKentucky
Connect Minnesota
Connect Ohio
Connect South Carolina
Connect West Virginia
In order to be effective, a national broadband data-collection and mapping exercise should be conducted by a government agency, on behalf of the public, with as granular a degree of information as possible and be totally transparent so that underlying information can be evaluated.
Connected Nation is none of those and represents none of those characteristics. It is an organization sponsored by the telephone and cable companies and represents their interests in deciding what data to collect and how information should be displayed. They are quite up front about their company sponsorship and, in fact, believe it is an asset, if in a way counter to solid public policy.
It would be a setback for our broadband policy if Connected Nation were to take a prominent role in broadband mapping and data collection if it continues on its present policy course because the organization does not represent wise public policy and because it distorts its results. Kentucky Gov. Steve Beshear (D) was correct in April, 2008, when he vetoed a $2.4 million appropriation for Connect Kentucky, which until then had received almost $7 million from the commonwealth. Beshear said that the program was being rejected for state financing because it had asked for funds “without specifically identifying any services to be rendered to the state or providing for any oversight, control or performance measures relative to the services being rendered.”
<strong>Connected Nation Represents Bad Public Policy</strong>
Connected Nation is not a neutral broker in broadband information. It is run by, and boasts of its connections to, telephone and cable companies. Yet, it accepts public funds in the millions of dollars to conduct a public function—mapping of broadband.
The end result is a project from Connected Nation which, instead of reflecting neutral information on which good public policy can be based, instead represents only the information that the most interested of parties wants reported.
Quite simply, Connected Nation’s strategy is to accept public funds for collecting information from its sponsors which is then kept largely private, hidden behind strict non-disclosure agreements (NDA). This privatized function is a violation of the public trust.
It was through the lobbying of Connected Nation that legislation (S. 1492) requiring broadband mapping has such restrictive terms that allow the companies involved, not the government, to determine which information should be made public and would best serve the public interest. It is at Connected Nation’s insistence that it, and its state-level operations, such as Connect Tennessee or Connect Ohio, insist on keeping much of the information they collect hidden from public review. The original Connect Kentucky was also a lobbyist on behalf of telephone companies for pro-industry, anti-consumer legislation.
Let’s take a look at the Connect Board of Directors. There are 12 outside directors, eight of which are directly in the orbit of network operators. They are not small players.
James W. Cicconi – AT&T senior executive vice president-external and legislative affairs
Steve Largent – CTIA – The Wireless Association president and CEO
Joseph W. Waz – Comcast senior vice president, external affairs and public policy counsel
Larry Cohen – Communications Workers of America president. CWA is in frequent agreement with telecom companies on policy issues.
Thomas J. Tauke – Verizon executive vice president for public affairs, policy and communication
Walter B. McCormick – United States Telecom Association president
Kyle E. McSlarrow – National Cable and Telecommunications Association president
Grant Seiffert – Telecommunications Industry Association president. (The members are the equipment makers who sell their gear to the telecom industry.)
These individuals, and others, are listed as “national advisors” on the Connected Nation Web site. They are listed as “directors” in their filing with the Kentucky Secretary of State.
As the Kentucky Public Service Commission told the FCC:
“As often recognized by public policy makers across the country, specific private business interests are not always consistent, or even compatible, with broader public interests. The Kentucky Commission believes a sound and honest public policy initiative to promote broadband deployment must anticipate and account for such discrepancies. In going forward with this national broadband mapping collection effort, which the Kentucky Commission supports, the Kentucky Commission petitions the FCC to remain diligent in keeping consumer interests at the forefront of this endeavor. ”
The PSC recommended that any data mapping information be verified independently from its source. That’s a good suggestion, and it runs counter to the philosophy of Connected Nation. Instead, CN relies on very strict non-disclosure agreements to limit what can be done with the data it collects. In North Carolina, the NDA requires that the maps and web sites used to show broadband coverage “may not differentiate between general broadband service types (such as DSL, cable, fixed wireless, BPS and others) and may not, at a pinpoint, address level, identify broadband Providers at a given location.” (The agreement also stipulates that any information collected remains the property of AT&T, and can be returned or destroyed at any time.) The NDA attached to this report was used by AT&T to collect information on behalf of the Connect North America organization it sponsors.
This clause in the agreement means that the broadband maps produced by Connected Nation or its franchise operations around the country simply show that a company has some service on some street. For consumers or policymakers, there may be no indication of what the technology is, at what speed, or at what price. If the carriers have their way, there never will.
The American Public Power Association, in a 2008 resolution, observed that CN “is ‘public’ only in the sense that it receives taxpayer money from the state [Kentucky] government; its activities are purely private.”
Indeed, Connected Nation will go to great lengths to keep its information collected from its supporters shielded from public view. It argues that only because of its close connections with telecom carriers does a state get any information at all. As one CN official told a gathering of state utility commission staff about whether states could decline to work with carriers or even do the mapping themselves:
“The question would go to provider cooperation in this process. And unless there is legislation that passes in the individual states that mandates the collection of this information, there is [sic] a lot of things that can happen that the provider community might not want to participate. They may not want to provide [the information]. It is not just as easy as the states utilizing the resources they have; there is the other consideration to consider, and that is provider participation in the process. It may be, and I am not saying that this is good or bad, and providers definitely do not definitely want to provide their proprietary information to a state agency, that [carriers] could tie this up in the courts for a while, too. And that is something that would have to play itself out.”
Indeed, in North Carolina, a Connected franchise is working to duplicate work being done by a state agency, the e-NC authority, which attempts to compile its own maps from information supplied by carriers and from its own estimates. The Connect North Carolina operation, run by AT&T, supplies information to itself that it would not supply to the state.
The maps compiled by Connect are inadequate and inaccurate. It is some times hard to discern which definition fits at any given moment. There is a distinct lack of useful information on the maps, such as what data speeds are being offered at what price at any given location.
Indeed, the basic information on the maps, that service of whatever type is available, is open to question because CN, rather than collect granular information by door-to-door canvass, assumes that every spot within a range of a cell tower or telephone company wire center is being served. That is not the case. And it can take dozens of steps and clicks through the cumbersome map interface to reach the inadequate or inaccurate information.
In sum, as a group of municipal utilities told FCC Commissioner Copps in July, 2008, “Broadband data must be collected and delivered in a transparent, verifiable manner. The CK/CN model doesn’t do that: Data is collected, interpreted and reported by a private non-profit entity and shielded from government and public input, oversight and verification.”
<strong>Connection Nation’s Achievements are Over-Stated</strong>
Before state governments hand the reins of policymaking to Connected Nation and their corporate backers, they should ask the basic question: Does the Connected Nation program actually work? That is, do the millions of taxpayer dollars spent by Connected Nation actually impact broadband deployment or adoption beyond what would have occurred otherwise?
Despite the state and Federal embrace of the Connected Nation model as the ideal approach to promoting broadband deployment and adoption, there is absolutely no proof whatsoever that their program actually produces positive benefits, whether from outside evaluators or from Connected Nation itself.
One outside study, “Closing the Rural Broadband Gap,” done by Michigan State University (available at <a href="https:/~larose/ruralbb">https://www.msu..edu/~larose/ruralbb), examined four rural counties that received broadband deployment grants from the USDA. This study found that one of the four counties, Pike County Kentucky, “registered the strongest gains among older adults and those with less than a high school education compared to the other three counties.” The study attributed the changes to “a possible impact of the Connect Kentucky initiative targeting disadvantaged populations in that state,” and then generalized the conclusion to the rest of the state.
If the study can only point to one of Kentucky’s 120 counties in which CK accomplished its goal, then the organization’s claims of success are open to challenge. In addition, there could be any number of external factors involved, which the study did not attempt to account for. For example, in two of the four counties studied, the USDA grants were awarded to non-traditional wireless ISPs, while the grant recipient in the third county went out of business. Only in the fourth county—Pike County Kentucky—did the grant go to an established provider, South East Telephone Company. Thus, the changes observed by the researchers are likely due to the impact of the USDA funding and South East Telephone Company’s own efforts, and not the result of the Connect project. Merely citing the “possible impact” of Connect Kentucky without proper controls and without considering alternative causes is not rigorous economic analysis.
The organization’s only claims of success are contained in a single “study” that is so flawed it has been characterized as committing “methodological malpractice” in an Aug. 4, 2008 filing at the Commission by Consumers Union, Consumer Federation of America, Free Press, and Public Knowledge. And this “study” comes from a group that blatantly attempted in its comments to the Commission to hijack the FCC’s data collection and mapping proposals and replace them with their own, industry-governed and supported, mechanisms.
The Connected Nation study from February 2008 claims “Kentucky’s broadband adoption rate is higher than the national trends due to Connected Nation’s first statewide broadband expansion program, Connect Kentucky.” However, the methodology used to support this claim is fatally flawed. Unfortunately, several press outlets have uncritically repeated the results of this bogus study, leading to the false impression that the Connected Nation model has been proven a success.
Connected Nation could have conducted a proper programmatic evaluation using established and accepted quasi-experimental techniques. But it did not. Instead the group chose an analytical approach that would easily be recognized as flawed by a first year statistics student. The entire claim for success is based on the fact that between 2005 and 2007 the broadband adoption growth rate was 83 percent in Kentucky versus 57 percent for the U.S. as a whole. That is, in 2005 (according to surveys conducted by Connect Kentucky) approximately 24 percent of Kentucky homes subscribed to broadband, which had increased to 44 percent by 2007 -- a “growth rate” of 83 percent. At the same time the national broadband adoption level (according to surveys by the Pew Internet and American Life Project) went from 30 percent to 47 percent -- a “growth rate” of 57 percent.
From this, Connected Nation claims that if “the national growth rate between 2005 and 2007 were applied to the 2005 Kentucky baseline (24%), then Kentucky’s expected statewide adoption in 2007 would be 37%. However, Kentucky’s broadband adoption percentage is actually 44% in 2007, which represents 297,000 more subscribers above the expected adoption rate. The intervening factor has been Connect Kentucky”.
But there is a major flaw in this approach, one of such gravity that it is hard to see how an organization of Connected Nation’s supposed professional caliber could make innocently.
The flaw lies in the comparison of “growth rates” of an initially low performing state against the national average. It should be obvious that improvement by a subject with a low performing metric almost always results in greater percentage (not percentage point) gains when compared to improvements made by a subject that starts with a higher performing metric. This is especially true when comparing the improvement of a low performing state with the average improvement of all states.
The national rate average is just that -- an average of all states, both low and high performing. It is composed of many states that already had high broadband penetration in 2005, and thus didn't have much room to improve to 2007. It would certainly be expected that the national rate of improvement to be somewhat lower than the rate of improvement of some of the lower ranking states -- simply because they had more room to improve. Indeed, using Form 477 data to examine changes in broadband penetration over the 2002 to 2006 reveals that the states with the highest percentage change in broadband penetration were the worst performing states in 2002, and their percentage improvement were far higher than the national average.
The nationwide improvement over this period was 179 percent. Kentucky, which was ranked 50th in penetration in 2002 improved 544 percent; but Montana (ranked 49th in 2002), which had no mapping program intervention had a 568 percent improvement. Alaska had the lowest improvement over this period (97 percent), but was also ranked 3rd in penetration in 2002. Simply stated, a big percentage improvement by a low performing state is unremarkable, especially compared to the nationwide average improvement.
When examining an entity (such as a state or nation) that is maturing towards a saturation point (such as technology adoption), those with lower adoption rates almost always have larger percent increases in growth over time, when compared to entities that begin with higher adoption rates. To illustrate this, consider the following example: According to the U.S. Census Bureau, the percent of urban homes subscribing to broadband went from 11 percent in 2001 to 54 percent in 2007, a 400 percent (not percentage point) increase. During this same time, the percentage of rural homes with broadband went from 6 percent to 39 percent, a 555 percent increase. Also during this time the national household level broadband adoption rate went from 10 percent to 51 percent, a 430 percent increase.
Thus, if we are to measure broadband success using the Connected Nation standard, the rural areas of the United States are greatly outperforming the “national average” and greatly outperforming the urban areas. By the same logic, America should seek to emulate the broadband policies of nations like Turkey and Greece, who experienced infinite percent increases in broadband adoption since 2001 (because in 2001, these countries had no broadband at all, and percentage growth from zero is by definition and infinite percent increase). No one of credibility would make such claims, but it is the exact methodological basis that underlies the “proof” of Connected Nation’s success.
These examples indicate is that the percent change (not percentage point change) in broadband adoption is a completely meaningless statistic.
Connected Nation has continued to make these bogus claims despite critiques of their underlying methodology. In comments to the Federal Communications Commission (urging the Commission to not collect its own broadband data) Connected Nation boasts of the success of the Connected Tennessee program, stating “[i]n just the first six months of the Connected Tennessee program, home broadband adoption has doubled the national growth rate. Rural areas have seen the most significant increases [in growth rate], as home broadband adoption increased by 37% over a six-month period.” It is frankly embarrassing that the group would make such a statement; because we certainly would expect the rural areas to have the most significant percentage increases in adoption -- precisely because rural areas start from the lowest levels of adoption.
If Connected Nation really desired to conduct a proper evaluation of their program they should compare (using accepted quasi-experimental program evaluation techniques) the change in performance of a state where they have intervened to the change in performance of other similar states (in terms of initial broadband adoption and other relevant characteristics) that had no such program. It is telling that Connected Nation has avoided conducting or commissioning such an evaluation, despite having the reason and resources to do so.
But even if we ignore for the moment the bogus nature of Connected Nation’s data, we must focus on the possibility that outside forces other than the non-profit’s are responsible for changes in Kentucky and Tennessee. In their comments, Connected Nation states “Kentucky increased from an estimated 60% of households passed in the state of Kentucky prior to the beginning of the program, to 95% at the end of 2007.”
FCC data indicates that the percent of ILEC lines that were DSL-capable in the state of Kentucky went from 60 percent in June of 2005 to 87 percent in June of 2007. But in December of 2006, as a part of its consent decree to merge with BellSouth, AT&T committed to 100 percent DSL availability in the entire BellSouth territory. Thus, it is quite possible that much of the improvements seen in Kentucky and Tennessee (former Bell South territory) are due to AT&T’s commitments, and not to Connected Nation’s mapping and demand stimulation efforts.
<strong>Additional Resources</strong>
Connect Kentucky Provides Uncertain Model for Federal Legislation – Jan. 9, 2008
http://www.publicknowledge.org/node/1334
Connect Kentucky Update: Broadband Tax Plan Ditched – Feb. 4, 2008
http://www.publicknowledge.org/node/1382
Connect Kentucky Disconnected at Home – April 14, 2008
http://www.publicknowledge.org/node/1521
Connected Nation's Private Interests Hit in FCC Comments – July 24, 2008
http://www.publicknowledge.org/node/1675
Connected Nation Takes Aim At Stimulus Broadband Mapping; Rural Areas Could Be Hurt – Feb. 17, 2009
http://www.publicknowledge.org/node/1998
Virginia Uses Self-Help Program for Rural Broadband – Feb. 20, 2009
http://www.publicknowledge.org/node/2003
Chicago Tribune: $7.2 billion plan to wire rural America holds promise, pitfalls
http://www.chicagotribune.com/news/chi-wired-for-webfeb22,0,4823241.story
Column from Charlotte Observer: A Blind Curve on the Information Highway – Feb. 20, 2009
http://www.charlotteobserver.com/346/story/548379.html
Coverage from IndyWeek, Durham, N.C. Stories by Fiona Morgan
State Lawmakers Debate Using Industry-backed Info on Broadband Access – Feb. 4, 2009
http://www.indyweek.com/gyrobase/Content?oid=oid%3A279831
Building a Better Broadband Map – Feb. 5, 2009
http://www.indyweek.com/gyrobase/Content?oid=oid%3A280381
Telecom Industry Brings Connected Nation to North Carolina – Dec. 31, 2008
http://www.indyweek.com/gyrobase/Content?oid=oid%3A272764
Article by Drew Clark of Broadband Census.com
http://arstechnica.com/tech-policy/news/2009/02/infrastructure-investment-decisions-need-transparency.ars
Connected Nation affiliates (as listed on Web site)
Connected Tennessee
ConnectKentucky
Connect Minnesota
Connect Ohio
Connect South Carolina
Connect West Virginia
For more information:
http://www.media-alliance.org/article.php?...
Add Your Comments
We are 100% volunteer and depend on your participation to sustain our efforts!
Donate
$75.00 donated
in the past month
Get Involved
If you'd like to help with maintaining or developing the website, contact us.
Publish
Publish your stories and upcoming events on Indybay.
Topics
More
Search Indybay's Archives
Advanced Search
►
▼
IMC Network