If you've taken time out to tweet in the last few days, you're hanging out in a hot, black neighborhood, or at least that's the latest news from Edison Research. African Americans make up roughly 12 percent of the population, but evidently, we occupy nearly 25 percent of the real estate on Twitter—a pretty strong showing.
There's no objective reason to doubt Edison's motives in reporting such an interesting statistic. The firm has a solid, 12-year history of producing reports like Twitter Usage in 2010, the source of the new statistics on how much time black people supposedly spend tweeting, or at least closely monitoring the tweets of others.
It also helps that the random sample of Americans selected for the Twitter study was taken from data gathered by Arbitron, an organization with a long track record in media research. So why do some statisticians question the Twitter study results?
The primary issue is the size of the population sample the people at Edison used to draw their conclusion. Edison gathered statistics from a randomly selected group of 1,753 Americans, originally surveyed by Arbitron for their Internet and Multimedia Series. Not a huge number, but large enough to make reasonable projections using reliable statistical models.
''The larger study sample is a hybrid of two sources,'' explains Bill Rose, senior vice president of marketing for Arbitron. ''Roughly one half of the sample were people who filled out Arbitron diaries, the other half were gathered from a random-digit dialing survey using landlines and cell phones,'' said Rose, adding that 12.8 percent of the group was African American, a near-perfect representation of the U.S. population. But there is some doubt about Arbitron's ability to assess African-American media trends when you look at its history. The company has struggled for years to accurately represent African-American and Hispanic radio listeners. It even settled a lawsuit related to the issue, brought by New York's attorney general, as recently as 2009. The settlement states that Arbitron had to include a larger number of cell phone users (who are more likely to be African Americans or Hispanics) in its surveys.
Assuming Arbitron made the needed adjustments to its methods, there's still a reason to question what Edison did with that sample of 1,753 people. ''The statistic [more than 24 percent of Twitter users are African-American], was based on the number of monthly Twitter users in the study,'' explained Jason Hollins, a vice president at Edison. That means the data that was used to assume that African Americans comprise one-quarter of Twitter's 17 million users in the United States was based on a survey of 105 people who reported that they visit Twitter at least once a month. No matter how you crunch the numbers, it's difficult to make a solid assumption about 17 million people from reviewing the habits of such a small group.
''There's a risk that as you get a smaller and smaller sample, there's a greater chance that you do not have a representative sample of a population,'' said David Swanson, a statistician and professor of sociology at the University of California at Riverside and expert in population estimation forecasting. ''The real question is 'what is the margin for error at that point,''' Swanson said. ''There's a huge amount of uncertainty with really small samples.''
So that's what The Root did. We discovered that the group was not really representative of Americans, let alone African Americans, in many ways. Thirty percent of the 105 have four-year degrees. That's above the national average, and far above the four-year degree rate for African Americans—roughly 20 percent. More than 30 percent of the group earns $60,000 or more a year, again, a number above the national average and far above the African-American median of about $35,000 per year. About 63 percent of the 105 use mobile phones to access the Internet and 35 percent update their social networking status several times a day. Not surprisingly, 81 percent of monthly Twitter users are under age 45, and 33 percent are 25 to 33. This is hardly a picture of average folks, but it does seem to be a reasonable portrait of African Americans who spend a great deal of time online. When comparing Edison's report to other highly respected studies, similar patterns emerge.
The Pew Internet and American Life Project, for instance, reports that 29 percent of African Americans use the Internet on a handheld device each day. Edison's users were most likely to go online via mobile phone. A large survey by eMarketer, Inc., designed to help businesses understand African-American Internet usage, reported that 51 percent of African Americans are online—19.1 million people—adding that, ''younger African-American Internet users (under age 35) are most comfortable with all forms of digital communication: visiting social networks, downloading and streaming music and videos, blogging and playing games.'' Both studies are in line with an increasing number of reports that the racial digital divide is narrowing.
In Edison's case though, they are really telling us that there's a high probability that young, well-educated, high-earning, young African Americans love Twitter, the Internet and social networking sites. That's not really news. It just means they are lot like young, well-educated, high-earning people from virtually every culture in the world. Surveying the bigger picture painted by Pew, eMarketer and other savvy demographers, tells us that while our actual participation in Twitter's daily chatter may be slightly more or less than Edison reported, we have become a force on the Internet. A trend that's sure to grow.
Sheree Crute is a writer and editor based in Brooklyn, N.Y.