Southern states like Mississippi, Alabama, Arkansas and Tennessee have abysmal subscription rates, according to the analysis.

While the No. 1 most-wired state is Hawaii, states in the relatively wealthy Northeast have the highest subscription rates — among them, Connecticut, New Jersey, Massachusetts, New Hampshire and Rhode Island.

Access to broadband has become critical for anyone to keep up in American society. Finding and applying for jobs often takes place entirely online. Students receive assignments via email. Basic government services are routinely offered online.

The lack of a broadband connection puts people at a profound disadvantage.

People without access, who are likely to be lower on the economic ladder, fall further and further behind, widening the “digital divide” between rich and poor.

The Workshop used Federal Communications Commission data collected from broadband carriers through the end of 2010, the most recent data available. We added demographic information — like income, race and age — from the U.S. Census Bureau’s American Community Survey, and created a map.

The map shows subscribership rates and demographic information at the Census tract level. That includes practically every neighborhood in the nation, or about 66,000 tracts. Each tract represents between 1,500 and 8,000 people.

“Broadband” is defined by how fast data streams to your computer.

An old-style dial-up connection streams at 56 kilobits per second, or Kbps. Today, speeds for home connections are more than 20 times that. The FCC defines broadband as 200 Kbps download (to your computer) or upload (sending data elsewhere). The national economic stimulus plan, which dedicated $7.2 billion in grants to assist broadband proliferation, uses a more strict definition of 768 Kbps download and 200 Kbps upload.

The Workshop used the more stringent definition for this analysis. Meanwhile, the FCC, in its most recent, highly upbeat, broadband report, was still using the 200 Kbps benchmark.

The best and worst

Nationwide, the Workshop survey found that 40 percent of households did not have broadband connection in the home through December 2010. Among the states and the District of Columbia, Mississippi had the lowest rating, a 35 percent subscribership rate, according to the Workshop’s statistical analysis of FCC data. Mississippi also is the poorest state in the nation, according to the U.S. Census Department’s American Community Survey, with a median household income of $36,850.

The other lowest-subscribing states and their median incomes were:

• Arkansas ranked 49th with a 43 percent subscribership rate. The state ranks 50th in median annual income, at $38,600.
• Alabama ranked 48th in broadband with a 47 percent subscriber rate. It ranks 43nd in income, at $42,218.
• Tennessee was 47th in broadband at 48 percent. It ranks 49th in income, at $40,026.
• West Virginia was 46th in broadband at 49 percent. It ranks 48th in income, at $40,824.
• Oklahoma was 45th in broadband at 50 percent. It ranks 38th in income at $45,577.

Wealthier states ranked near the top:

• Hawaii is first in broadband at 74 percent, and 10th in income at $59,125.
• Connecticut is second in broadband at 72 percent, and second in income at $65,958.
• New Jersey is third in broadband at 72 percent, and third in income at $65,173.
• Massachusetts is fourth in broadband at 70 percent, and seventh in income at $60,923.
• New Hampshire is fifth in broadband at 70 percent, and first in income at $66,303.

While Mississippi’s home subscribership rate is dismal, residents still find a way to get online, according to Roberto Gallardo with the Southern Rural Development Center at Mississippi State University. In a survey of more than 2,000 households, nearly 79 percent said someone in the home had used the Internet.

The usage rate “was a lot higher than we expected,” he said, but it included people who accessed the net at any outside location, including school, work and the local library.

Reasons people give for not subscribing are similar to those given in national surveys he said — cost, lack of equipment, and a lack of understanding of its value, which is key.

“We need to increase exposure to the needs of the Internet,” he said.

For example, people on Social Security, who are less likely to subscribe, may learn the value of being online, and how to get online, from their children and grandchildren.

“The research shows out there if you increase broadband awareness, then that’s the first step for non-adopters to make the jump,” he said. “They need to have a motive to do this.”

Most and least improved

From December 2008 to December 2010, broadband subscribership rates increased by an average of 6.1 percentage points per state, according to the FCC data. Some states improved a lot, some a little and two actually slipped.

The largest improvement was seen in Vermont, which jumped 17 percentage points to 60 percent.

Rural areas have historically suffered from lower broadband subscription rates. There’s less money to be made with lower population densities, and it costs more to run wires to those areas. But according to the Workshop analysis, some of the most rural states seem to be improving quickly.

Alaska, the least densely populated state in the nation, according to Census Bureau estimates for July 2011, jumped 15 percentage points to a subscribership rate of 53 percent. Montana, ranked 48th in population density, jumped 14 percentage points to 55 percent. South Dakota, 46th in density, jumped 11.1 percentage points to 61 percent.

Two states did worse. Maine dropped 2 percentage points to 57 percent. South Carolina dropped a half a point to 53 percent.

Meanwhile, Louisiana, North Carolina and Nevada edged up a point, and Ohio edged up only two points.

The South has shown only modest improvement.

Comparing broadband subscription rates in one city to another is a tricky business.  An area may show an overall high rating but still have pockets of low usage. The differences within the same metro area can be dramatic.

For example, the sprawling Bridgeport-Stamford-Norwalk, Conn., metropolitan statistical area is the most connected of the top 100 MSAs in our survey, coming in at a 79 percent subscription rate. This isn’t surprising, given it ranks third in the nation in median household income.

But that’s only part of the story.

The poverty divide

The Bridgeport MSA also ranks No. 1 when it comes to the unequal distribution of wealth, according to a Stanford University study that looked at income segregation in American cities.

That gap is reflected in the broadband map. The urban core of the city suffers from biting poverty and low rates of broadband subscribership, while the outer suburbs show sky-high incomes and correspondingly high rates of broadband subscribership.

Wealthier households subscribe at a rate of 80 percent to 100 percent, while low-income areas of the city, some exceeding a 50 percent poverty rate, subscribe at a rate of 40 percent to 60 percent.

The metro area with the lowest score was McAllen, Texas, five miles from the Mexican border in the Rio Grande Valley, with a score of 37. 

In McAllen, the library is often where people go to connect.

“Our computer lab and free Internet services are probably the largest draw into the building, said Jose A. Gamez, director of McAllen’s public libraries. “We’re adding about 50 more computers because of the demand."

The low home-subscription rate in the city is no mystery.

"Hidalgo County is one of poorest counties in the country so a lot of people here just can’t afford their own computers or the broadband connection,” he said.

In fact, McAllen possesses the highest poverty rate in the nation, according to the Census.

Every major survey has shown that the lower the income, the less likely it is that households will subscribe to the Internet.

In February 2011, the National Telecommunications and Information Administration released a survey that said 68 percent of households had a broadband connection, far more than what the Workshop analysis shows.

The survey showed those whose family income is $15,000 or less adopted broadband at a 32 percent rate while those whose families earn $150,000 or more adopt at a nearly 90 percent rate.

Why people don't connect

When the NTIA asked those who did not subscribe to explain, 46 percent said “don’t need/not interested;” 25 percent said it was too expensive; and 14 percent said it was because they didn’t have a computer.

The question of non-adoption, then, seems a little more complex than simple income.

The FCC commissioned the Social Science Research Council to study why people do not adopt. They interviewed 170 non-subscribers and learned that the “not interested” answer did not come up.

“We found no evidence of that,” said Joe Karaganis, one of the study’s authors. The reason was in part because of survey methods, he said. “We think that’s an almost entirely fictional category at this point,” he said.

In the end, the study “validated the obvious” — cost is the main reason people don’t adopt. But it’s a more complicated issue than just monthly connection charges, he said. There are unexpected costs.  

“We just found lots of ways in which the $40 dollar a month charge is just the beginning of the costs of broadband access,” he said – like taxes, special features and other fees. 

But it’s more than just cost. There are cultural issues. The more educated you are, the more likely you are to subscribe. Whites subscribe at higher rates than blacks and Hispanics. And senior citizens subscribe at lower rates than young people.

Wireless to the rescue?

While broadband subscribership rate in homes seems to have flattened, the popularity of smart phones has skyrocketed. In the Workshop survey, non-home-based wireless connections were not included.

Thirty-five percent of American adults owned “smart phones,” as of July of last year, basically a cell phone that can be used to access the Internet, according to a Pew Research Center survey. Two-thirds of them accessed the Internet or email on their handheld on a typical day.

There is no question that today’s “fourth generation” wireless networks are capable of delivering data at much greater speeds than in the past. But a smartphone is no substitute for a home computer with a wire-line connection, at least not today.

Speed aside, wireless companies often impose data limits — the more you use them to go online, the more you pay. And data streams are subject to “throttling” or cutting back on usage when networks are particularly busy, even under unlimited data plans.

In addition, there are far more tasks that can be performed on a laptop or a PC than on a smartphone. In short, surfing the Internet on a handheld is not like doing it at home — or on a laptop at a McDonald’s, for that matter.

That fact is reflected in the survey in which only 25 percent of smartphone owners say they go online “mostly” using their handheld devices.

Smartphone owners under age 30, non-whites and users with relatively low income and education levels are “particularly likely” to say that they “mostly go online” using their phones.

What's the solution?

President Barack Obama made Internet access a part of his platform when he was elected in 2008.

He directed $7.2 billion of his economic stimulus plan funding to broadband projects. Most of it went to infrastructure – deploying broadband networks in areas that have no access. That’s great for rural areas, not so great for low-income areas that have access but can’t afford it.

Some of the money — $251 million — did go to grants to “promote broadband adoption” among vulnerable populations. A total of $201 million went to grants for funding public computer centers.

But not as much attention has been paid to closing the gap between rich and poor. The highest-profile effort to date has been touted by the FCC, but is, in fact, a voluntary effort by the industry.

The “Connect to Compete” project provides $9.95 per month cable broadband access for families who have a child enrolled in the free school lunch program, have not have subscribed in the previous 90 days and do not owe the provider for unpaid bills or un-returned equipment. Cable companies say that if all eligible families take advantage, that’s $2.5 billion worth of broadband service.

The program also involves other private companies as well as nonprofits to coordinate the program. It includes the offer of a $150 refurbished computer for those without one.

Organizers hope to launch the project in late August or early September. While the effort is ambitious, it is unlikely to represent a long-term solution. The $9.95 per month offer is good for two years.

Feel-good fix?

Consumer advocates are supportive, but less than hopeful.

“We do think it’s a good idea for companies to offer low-cost broadband,” said Craig Aaron, president of Free Press, which is usually at odds with the big broadband providers. “But having a company offer low-cost broadband is not the same thing as having a broadband policy.”

The FCC released a lengthy national broadband plan, ordered by Congress, in 2010. It followed an exhaustive effort and was filled with policy recommendations. FCC spokesman Mark Wigfield released an agency statement saying that the FCC has taken “numerous steps” to increase broadband adoption, including the “Connect to Compete” program. 

The agency has also reformed its “Lifeline” program for low-income phone subscribers to include a pilot project to test ways that phone subsidies can be used to increase broadband adoption.

In addition, the controversial Comcast-NBC Universal merger required the cable company to make high-speed Internet access available to 2.5 million low-income households at $10 a month, along with low-cost personal computers and digital literacy training.

Aaron said the real root of the problem is too few players in the marketplace. He said the FCC needs to be dealing with “the incredible lack of competition when it comes to broadband — something they frankly haven’t been doing.”

The Investigative Reporting Workshop reporter Michael Lawson contributed to this report.

John Dunbar is the former director of Connected, The Workshop's Media and Broadband Project. He is now at the Center for Public Integrity, which is co-publishing this project, as managing editor for politics and finance.

Jacob Fenton, former Workshop director of computer-assisted reporting, collected and analyzed the data for this report. He is now at the Sunlight Foundation.

This project was made possible in part by funding from the Ford Foundation.