The year was 1967, and Detroit was on fire. Police had raided an illegal bar, sparking a five-day riot that left forty-three people dead, hundreds injured, and thousands arrested. Over two thousand buildings were burned to the ground.

Click to read photo caption. Photo by Mark Derewicz; ©2007 Endeavors.

What caused the raid to spiral into a riot was up for immediate speculation. Experts thought that it was the frustration of southern migrants struggling to assimilate. Editorial writers thought that it was the desperation of uneducated poor folks. But Philip Meyer, a reporter assigned to cover the riots for Knight Newspapers, wondered if these assumptions told the whole story.

He suggested to editor Derick Daniels, a Carolina alum like Meyer, that the Detroit Free Press could find deeper reasons for the riot if the paper conducted surveys in the African American community. Meyer thought reporters could use the same social-science techniques that UCLA researchers had used after the Watts Riots in 1965. Daniels liked the idea, and under Meyer’s leadership the paper debunked the speculation: people born or raised in the North were three times as likely to riot as the immigrants from the South. Education and income were not good predictors of whether a person would riot, although unemployment was.

The survey also helped Meyer ferret out specific grievances in the African American community, and these also contradicted assumptions. Police brutality was a main grievance, as were poor housing and overcrowded living conditions.

The Free Press staff won a Pulitzer Prize for “both the brilliance of its detailed spot news staff work and its swift and accurate investigation into the underlying causes of the tragedy,” according to the Pulitzer committee. Meyer’s social-science reportage gained instant credibility, and was even given a cool name—precision journalism—which Meyer liked and used as the title of his 1973 book, now in its fourth edition.

At the time, publishing in-depth surveys and analyses was completely novel and in direct contrast to the sort of storytelling articles—the New Journalism—that Tom Wolfe and others were making popular. There’s a definite art to writing narratives and even traditional news articles that quote sources, but Meyer says newspapers don’t need more art; they need a dose of science.

To be precise

Meyer, a professor of journalism at Carolina since 1981, says that the most common applications of precision journalism are survey research, analysis of public records, and field experiments. Each method can uncover precise, accurate information that cannot be gleaned from confidential sources or experts, or by haphazardly surveying citizens.

Meyer’s career offers a few examples.

Before his days in Detroit, Knight Newspapers assigned Meyer to cover the governor’s race in Ohio for the Akron Beacon Journal. But he had never been to Ohio so his editor sent him and a political reporter on a two-week junket across the state to learn about Ohio and her voters. They filed stories, conducted polls, and reported in the Beacon Journal that the race was too close to call. Turned out they were wrong; Republican challenger James Rhodes won in a landslide. Two years later, they conducted more polls and predicted that Robert Taft would unseat Senator Stephen Young. Wrong again—really wrong.

Meyer figured he was doing something, well, wrong, so he documented his troubles in an application to Harvard’s Nieman Fellowship Program for mid-career journalists. He was accepted, and the social-science methods he learned—including statistical analysis and how to properly poll a diverse population—gave him the confidence a year later to pursue the Detroit Riot stories from a different angle.

In 1968, he was back at it in Miami. After Martin Luther King, Jr. was assassinated, Meyer was skeptical of experts who assumed that violence would replace MLK’s passive nonviolent resistance. Meyer had already conducted a survey on attitudes in the African American community, so he contacted his respondents again to ask questions about activism. “I had to raise the money for this project on my own,” Meyer says, “but I did get back in the field, and produced a series of stories showing that King’s death had strengthened his following, and that the violent minority was steady at 10 percent.”

On election night in 1976, when wire services were reporting a high voter turnout—some even predicted a record turnout—Meyer decided to analyze sample precinct data from ABC News, and wound up accurately describing the lowest presidential voter turnout in a generation.

The precious truth

“One of the great benefits of precision journalism is that it can give you the confidence to break from the master narrative,” Meyer says. “Being different can lead to eventual success, but you need to be sure you’re right and have the patience for acceptance of the truth to emerge.”

Precision journalism has little to do with covering city council meetings or reporting on fires and accidents, but if a city councilman uses data to support a claim, a reporter can and should look at those data instead of merely reporting on what the councilman says. This, Meyer says, was the problem with most of the White House reporting in the run-up to the Iraq War.

Precision journalism, though, takes money, Meyer says. In his 2004 book, The Vanishing Newspaper, Meyer rails against publishers who slash fixed costs, such as editorial staff, so that their newspapers can stay afloat in a world awash in new media. This approach, he says, won’t lead to success or long-term survival.

Again, Meyer derives his opinion from science. In the 1970s, political scientist Herbert Simon found that the growing abundance of information available in the media was leading to a scarcity of attention—that is, the more attention people give to something, the less attention they can give to something else. In today’s world, if a reader can find instant news on the internet, he may never find time to read the paper.

“The response from the media of all kinds has been to do wilder and crazier things to get our attention, even if it means sacrificing accuracy,” Meyer says, pointing to entertainment media obsession with celebrity, violence, and sex.

“The outcome is that now truth is becoming a scarce good. The cure for this is better trained journalists who can apply scientific method to their work instead of being mere hunter-gatherers of information that’s processed by someone else.”

In The Vanishing Newspaper, Meyer writes that instead of cutting editorial staff, newspapers should cut variable costs, such as paper, ink, and transportation by migrating distribution to the web.

“If truth is scarce, there should be money to be made by finding it, branding it, and selling it,” Meyer says. “To do so would require journalists of greater skill than is the norm now.”

The art of polling

These days, most common social-science reportage in newspapers involves polling and editorializing about poll results, neither of which guarantees precise analysis, depth of coverage, or even accuracy.

Meyer, who has studied public-opinion research for decades, says that polling has been evolving ever since political scientist George Gallup accurately predicted the winner of the 1936 presidential election by using a special method with a much smaller sample size than other pollsters used.

But in 1948, Gallup predicted that Thomas Dewey would defeat Harry Truman by five to fifteen percentage points. His mistake was captured in that iconic photo of Truman holding up the next day’s edition of the Chicago Daily Tribune with the headline “DEWEY DEFEATS TRUMAN.”

Meyer says that Gallup didn’t have a bad sample of voters; Gallup messed up because he stopped polling three weeks before the election, and so did not take into account Truman’s effective whistle-stop tour or the use of mass media. Also, some Democrats didn’t like Truman, Meyer says, and told pollsters they would vote for Dewey instead. But as Election Day approached, they changed their minds, and voted Democratic. For newspapers, this problem was easily solved—conduct polls right up until Election Day.

But other problems popped up.

Polls were a lot easier to conduct back when homeowners expected uninvited guests and daily deliveries of mail, dry cleaning, ice, and milk; pollsters were merely part of the parade of visitors. As those halcyon days faded away, Meyer says, pollsters began using the telephone, which worked well at the time because 95 percent of the population could be contacted by phone and one phone per household was the norm. And random-digit telephone dialing worked like a charm because the theory of polling depends on each member of the population having an equal chance of being polled.

“But now, so much technology has gotten between pollsters and recipients,” Meyer says. Answering machines, call waiting, multiple phone lines, fax machines, and especially people who use only cell phones all screw up the polling sample. “The only survey I trust nowadays is one done by mail, and—for certain populations of people—the internet,” Meyer says.

But news organizations could still use the mail and internet (and in-person interviews, if they are so bold) to conduct in-depth surveys and polls, and then tease out the surveys with analysis and interpretation to come up with fascinating story angles that would inform readers.

Meyer points to Thomas Frank’s book What’s the Matter with Kansas as a good example.

“Frank’s research showed that many working class people in Kansas voted for President Bush in 2000 against their own economic interests, and they did this based on their social values,” Meyer says. But Frank’s book was published in 2004. Meyer says that newspapers could have done this sort of reporting before the election because the Kansas phenomenon had developed over time. Polls would have shown that Republicans wisely built a coalition between poor people with strong feelings on social values and relatively affluent, educated people with opposite opinions on social issues. Together, they sealed victory for Bush.

“If polls showed reasons why people vote the way they do, voters might become more aware of conflicts of interest,” Meyer says. “Voters could ask themselves, ‘Which group do I feel more comfortable with? Where do my interests really lie?’ It could change people’s minds about how they vote.”

Such analysis would better serve readers, but reporting only on who’s winning a presidential race isn’t necessarily shallow or harmful, Meyer says. Polls attract attention. “You wouldn’t expect people to stay interested in a basketball game if the scoreboard was covered up,” he says. “And polls can help voters interpret what candidates are saying. A candidate who’s way behind might say more desperate things. Voters should know that.”

And, Meyer says, polls are good for democracy. If a voter knows that his first choice for presidential nominee is way behind in the polls, but his second choice has a real chance of winning, then the person could vote for his second choice, making his vote more meaningful and building a coalition of supporters. “But you could probably find some people who disagree with me,” he says.

And the award goes to…

Newspapers don’t gain readership, earn trust, or get accolades for conducting “horse-race” polls without analysis or interpretation. Newspapers get attention and gain credibility when reporters break accurate and amazing stories, and a lot of these stories are right under our noses—in public records or confusing reports or even scientific data that may never see the light of day. It’s tough work, Meyer admits, but the results are worth it.

In 2005, reporters at the St. Petersburg Times found that federal government records of wetland depletion were full of questionable science and accounting. As the reporters were trying to decipher those data, they decided to use before-and-after satellite imagery and geographical information systems software to measure the loss of 84,000 acres of wetlands partly due to residential development, even though federal law was supposed to have stopped such development in 1990.

Last year, Wall Street Journal reporters dug through public records and analyzed raw data to help expose illegal backdating of stock options in publicly held companies. The stories won the Philip Meyer Journalism Award, which cites the best use of social-science methods in news reporting. A few months later the Journal won the Pulitzer Prize for Public Service for the same story.

Meyer says these stories show how advanced precision journalism has become, but using such methodology is still the exception to the rule. The proportion of adults who read the newspaper every day has declined an average of nearly one percentage point a year for more than forty years, and many publishers, citing the bottom line, have circled the wagons.

“The industry might be reverting to the hunter-gatherer model,” Meyer says. “I think a reporter who has precision journalism skills can still get a good job, provided he’s also good at something else.”

Philip Meyer plans to retire in 2008 as Knight Chair in Journalism in the School of Journalism and Mass Communication, a post he has held since 1993. He was a correspondent in the Washington Bureau of Knight Ridder, Inc., from 1962 to 1978. He is also past president of both the American Association for Public Opinion Research and the World Association for Public Opinion Research. The Philip Meyer Award is administered by the National Institute for Computer-Assisted Reporting.