Bully Factories?   

A behind-the-scenes look at how researchers' disagreements, simplistic and sensational coverage and a host of other problems clouded public understanding of a major study of child care

JUST AS SARAH FRIEDMAN reached for the phone to dial into a conference call with 22 journalists, a buzzer released staccato blasts through the halls of the National Institute of Child Health and Human Development. Evacuate! Evacuate! The developmental psychologist scooped up her papers and rushed down six flights of stairs. She hurried to a neighboring building in Rockville, Maryland, found an available office on the fourth floor and belatedly joined the call.
The ill-timed fire drill that Wednesday, April 18, was just a warm-up for trouble.
The real clamor began after the press conference introducing the latest conclusions from the NICHD Early Child Care Study, the nation's most comprehensive long-term assessment of its kind. Scientific coordinator Friedman and University of London professor Jay Belsky, already in Minneapolis for the meeting where they would formally present their scientific findings the next day, outlined the highlights.
Reporters seized on a provocative finding: The more time preschoolers spent in day care, the more problem behavior reported as they entered kindergarten. "Researchers Find a Link Between Behavioral Problems and Time in Child Care," the New York Times headlined the next day. "Day care linked to aggression: More hours add to kids' problems," USA Today fretted over a short but restrained front-page story.
As others picked up the news, the tone grew more ominous. "From Babies to Bullies," the Washington Times summarized in an editorial. A newscast on which Friedman appeared introduced the segment with brief footage of the Columbine school shooting and a statement "that our study verified long-held suspicions that more hours in child care predict aggressive behavior," she recalls with dismay. "This was the frame of mind."
Left out of some early coverage were a healthy dose of perspective and some other key findings. Among them: The more hours in care, and the better its quality, the higher children scored on language and cognitive tests. And those "aggressive" behaviors that had surfaced in 17 percent of the kids racking up more than 30 hours a week in care--versus the 6 percent in care less than 10 hours--included not only "defiant," "fights" and "disobedient" but also the far less menacing "can't stand waiting" and "wants attention."
Such nuances came through more clearly in a second round of stories and talkfests as reporters tracked down elusive reports, NICHD shifted into damage-control mode and a whole spectrum of interest groups began spinning. "I knew the data, but I just was not prepared for the emphasis that [reporters] gave it and the context in which they were embedding it," Friedman says.
But plenty of other factors--including the press conference's focus on preliminary findings, the researchers' disagreements over interpretations and NICHD's reluctance to disseminate unevaluated reports or their details--played a role in the confusion.
How the child care findings were released and subsequently reported provide "a textbook case of the conflicting forces and errors and intensity that can attach to science journalism," says David Murray, director of the Statistical Assessment Service, a Washington, D.C., nonprofit that aims to improve public understanding of science and statistics. "All this research became more interesting because it could be plugged into the culture wars and political battles."
The case underscored longstanding tensions between researchers and journalists about who shapes the popular notions involving science. It resurrected complaints about the media's oversimplifying or misrepresenting findings. "Does the press want to inappropriately worry the public and lose credibility as a consequence? Because that's what happened," says Ellen Ruppel Shell, codirector of Boston University's science journalism program and a correspondent for The Atlantic Monthly. Author of a 1992 book on day care, she contends some journalists blithely parroted the results without sufficient analysis or context.
The episode fueled grumbling by journalists frustrated in their efforts to get data on deadline. It bred finger-pointing among the study's 25 investigators, some of whom publicly faulted Belsky--an outspoken critic of long hours in care--for the media emphasis on aggression. He, in turn, accused them of trying to soft-pedal the bad news. And the backlash prompted NICHD's director, Duane Alexander, to call for restrictions on how and when these investigators release information to the media in the future.
"There were distortions in the presentations and in the press," Alexander says. "In the future, the group is going to be a lot more careful about how they report their results."

WITH MOST MOTHERS of young children in the workforce, child care has become the norm for American kids. Seventy-five percent of those younger than 5 regularly are watched by someone other than their parents, according to 1995 figures from the U.S. Census Bureau.
Concerns about the relationship between child care and child well-being spawned the government study. In 1991, researchers at 10 U.S. universities began tracking 1,364 children, following their development and their care arrangements from infancy onward. Data have been analyzed only through the kindergarten experience, though 80 percent of the remaining 1,100 subjects will enter the fifth grade this fall. The $80 million project--which considers the impact of care by anyone other than moms--has yielded dozens of papers and presentations on subjects such as mother-child attachment and links between child care and self-control. Because of the study's scope and complexity, and because of emotionally charged attitudes about child care itself, the NICHD reports provoke keen interest.
The most recent findings proved no exception.
The Society for Research in Child Development accepted four preliminary reports from the NICHD Early Child Care Research Network (as its investigators are collectively known) for presentation at the society's April meeting in Minneapolis. Team members would discuss the early conclusions cited in these working papers, an accepted practice that serves the dual purpose of staking a research claim while soliciting valuable feedback before a report undergoes the rigorous peer review needed for publication in a scholarly journal.
Journalists who routinely cover SRCD's biennial meetings had indicated, in a mail survey, their strong interest in early childhood issues. So, the society invited the network to showcase its findings in the April 18 press conference--a step NICHD's Alexander discouraged because the results hadn't been independently ascertained. The institute's policy is to publicize only research that has been accepted by, or published in, a journal. But because the child care investigators operate under a cooperative agreement with the government, not a more restrictive grant or contract, they decide how to handle everything from research methodology to the dissemination of results. Their steering committee, which includes a representative from each of the 10 research sites, three middle childhood investigators, a data specialist and Friedman of NICHD, determines network policy. "We have one of 15 votes," Alexander points out. "We don't have veto power."
With a green light on the press conference, Lauren Fasig, SRCD's director of policy and communication, asked whether NICHD would furnish a news release. The press office, again citing lack of a paper "in press," offered a semantic compromise: four brief "backgrounders" intended for distribution only to press-conference participants "to focus this from our perspective," explains John McGrath, NICHD's top press officer. Fasig summarized these in a single press release--leading with the problem-behavior findings--and posted it on the SRCD Web site. Limiting access "was not communicated to me," she says.
Fasig lined up press-conference speakers just days in advance. There would be two from the child care study and two from a separate analysis of the Early Head Start program. Fasig placed a handful of calls--discovering that researchers would be in transit to the Minneapolis meeting and thus unavailable--before securing Jay Belsky to explain the child care findings with Sarah Friedman.
Belsky gained notoriety in 1986 when, as a professor of human development at Pennsylvania State University, he raised concerns about near full-time child care in the first year of life. But Friedman, who described her study colleague as "a very good scientist," welcomed his participation--and supported his solo presentation of the study's findings at a small seminar April 17 at the National Institutes of Health. I was in the NIH audience as Belsky described links between time in child care and more behavior problems--all kinds of care, he stressed--as well as better language and intellectual skills. Friedman, who also attended, later agreed Belsky gave an appropriately balanced talk then and to SRCD members two days later.
For the press conference, Friedman asked him "simply to talk about the aggression findings--well, the effects of quantity of care," Belsky says. She would make the points linking quality child care to higher achievement.
After the Early Head Start researchers presented their findings, Belsky and Friedman each had three minutes to lay out the child care study results. Belsky emphasized that there was a correlation between more time in care and elevated levels of problem behavior, from "talking too much" to "argues a lot" to "gets in lots of fights."
"There was some back and forth" that helped focus attention on the aggression findings, says Diane Debrovner, a Parents magazine senior editor who participated in the call. She says Friedman tried to downplay the severity of the behaviors, noting most fell within the range of normal for young children. "But he was saying on the edge of normal is not [necessarily] pathological, but problematic."
Belsky dominated the conversation, say several reporters. They suggest the Early Head Start researchers were virtually ignored and Friedman was at a disadvantage. Because of the disruptive fire drill, "she was harried, she was late," says the Lincoln Star Journal's Erin Andersen. "She was a little more vague. I think she was not as eager to make a conclusion," Debrovner adds.
Debrovner describes Belsky as "a lot more media savvy," more prone to "sweeping statements." One that was widely quoted helped propel the research onto national news budgets: "As time in care goes up, so do problem behaviors."
At the press conference--and, Belsky says, in subsequent interviews with reporters--he was questioned about the study's evidence and his own inferences. "The issues were juxtaposed. What was I supposed to do, say, ŒNo comment'?... I said there are two core findings that come out of our study: One is improve the quality of care, and the second is to allow families to rely on care less."
"We hoped the researchers would present this in as factual a way as possible without excessive interpretation," says NICHD's Alexander, adding that "the investigators themselves were not in agreement as to what the interpretation was. Basically, it was not yet ready for prime time."
Belsky's take on the brouhaha? "This blew up because the bad news got the ink, and the good news was the secondary consideration."

"SMART AND NASTY" was the unattributed phrase that boiled down the findings and reverberated in sound bites, over dinner tables and inside day care centers. It surfaced the day after the press conference in an April 19 Washington Post story by Shankar Vedantam that led with the problem- behavior finding.
H.J. Cummins of Minneapolis' Star Tribune says she "tried to stress the limitations" of the study, folding the child care results into a broader story on key research emerging at the SRCD conference. Her article acknowledged both the drawbacks and benefits of care, and attributed to Friedman the notion that "the effects of good parenting were almost four times as strong" as the effects of outside child care. Cummins says she tried to be circumspect, unlike "the TV news teasers, 'Your child care could be turning your child violent.' "
For the Thursday evening news, ABC News correspondent Michele Norris produced a two-minute "Closer Look" segment that used the aggression findings to segue into a comparison of U.S. and European child care policies. "This sparked a very yeasty discussion here at ABC," she says of the study. "As a few of us noted, we weren't sure we would be sitting in these chairs if we were not aggressive."
David Crary, a national writer in the Associated Press' New York office, was assigned a reaction story that Thursday and initially worried there wouldn't "be anything fresh to say," he recalls. "It turned out there was a lot." His April 19 article had a calming tone, stating in the second paragraph that some experts both "urged parents...not to overreact" and recommended "a nationwide push for improved child care options."
The NICHD study generated tense debate and lively copy for days, for weeks. Belsky--and, to a lesser extent, Friedman and other principal investigators--made the rounds of the talk shows, from "Today" to "Talk of the Nation" to "CNN Talkback Live."
Columnist Mike Barnicle of New York's Daily News denounced the study as "the latest example of how federal bureaucrats can come together with pinhead intellectuals to waste money on foolish exercises that provide bogus results." An editorial in Pennsylvania's Lancaster New Era called for less day care, on the grounds that "many kids in child care are exhibiting anti-social behavior because they are not receiving enough love."
"A group of trigger-happy scientists leaped to the podium to publicize results that have yet to be published or subjected to peer review," Ann Hulbert scolded in the May 14 New Republic. "And the media eagerly spread findings that were tailor-made for sensationalizing."

IF SOME EARLY STORIES fell short on context or statistics, their authors weren't solely to blame.
"It was very difficult to get the actual study," points out ABC's Norris. Based in Washington, D.C., she obtained it for backgrounding that first Thursday only because "we had hired a producer on the ground who was able to collar people in Minneapolis."
USA Today's Marilyn Elias, who filed a short and thorough front-page story from the press conference, wanted to do a longer takeout for that same day's paper but was stymied until she got the researchers' reports. "Had we had more detailed information, I would have done an [additional] inside piece," Elias says. Writing from Los Angeles, she came back at the topic with a well-rounded Q&A--on May 3, two weeks later.
Communications staffs of the National Institute of Child Health and Human Development and the Society for Research in Child Development scrambled to accommodate interview requests with various researchers. But the institute referred inquiries about the research papers to SRCD, which bounced callers back to the NICHD Web site and its roster of investigators. "Everyone was on this circuitous runaround," says SRCD's Fasig, who more than a week later got researchers' permission to let her distribute the reports via e-mail.
As media calls surged, the NICHD news office began to compile a Q&A for the institute's Web site. But by late May, neither it nor any other document on the most recent findings had been posted. "When [news] breaks, when it's out there, you've got to get the information to reporters so they can accurately report," says a perplexed Barbara Rice, deputy director of news at the Washington-based National Academies, which advise the nation on science, health policy and technology. "You're dealing with complicated issues."
Robert Bock, an NICHD spokesman, confides that most of the news accounts he saw, even the early ones, "accurately reflected the information that was presented." The problem, several observers suggest, is that some coverage stopped short of important context. For instance, the long-term study had shown a slight uptick in problem behavior at age 24 months--think terrible twos--that dissipated after age 3. "You've got to be really careful about looking at short-term findings in child care studies," cautions Arthur Reynolds, a University of Wisconsin-Madison professor of social work. He directs the Chicago Longitudinal Study, which has tracked the social and academic progress of roughly 1,300 disadvantaged youngsters who took part in public early-childhood education programs. "A number of early intervention studies show the kids who get a lot of center-based care do have elevated rates of difficulty in getting used to the school situation," Reynolds points out, "but a lot of that completely disappears in first and second grade.... The most important thing is what happens down the road."
One reporter who took part in the press conference says she focused solely on what she heard. "I mostly just reported on what the findings were," says Joanne Boeckman, family life reporter for the Des Moines Register. "I handled it as a news reporter."
Charles A. Nelson, a University of Minnesota professor of child psychology and pediatrics who served as program chair of the SRCD conference, says one April 19 story he read was all too typical of the way journalists approach science. The piece, he says, "was almost a recitation of the findings. It was accurate, but it didn't go further than the data and the scientists' interpretation."
In contrast, Nelson cites "a model" story he found several days later in the New York Times' Week in Review section. There, reporter Sheryl Gay Stolberg examined the day care findings and the shortcomings of such observational studies to explain the challenges of social-science research. "Studies of human behavior make good headlines, but they can be fraught with complications," she wrote.
As the debate swirled, Belsky typically was cast in the black hat.
The University of London professor sorely objects to reporters and others discounting the aggression findings on the grounds that "this is an unpublished study. That was completely true two years ago as well," Belsky says. At a 1999 press conference, researchers shared generally favorable early findings. "Higher Quality Care Related to Less Problem Behavior," read part of the lengthy headline on a press release that NICHD issued three days later. Then, Belsky says, he didn't publicly criticize the presentation as too positive, instead taking his complaints in-house.
NICHD's Alexander, a pediatrician by training, complains that, because of Belsky's emphasis, "what the press picked up on was the most aggressive of behaviors--fighting, biting, kicking--when these were only among many, many items on the list of behaviors."
"By not qualifying this, we scared families," says Margaret Burchinal, the child care network's lead statistician and a scientist at the University of North Carolina. She says that while 17 percent of preschoolers in full-time care demonstrated problem behavior, that's consistent with the general population. "Even our kids with full-time care are really looking like typical 4- and 5-year-olds."
But Murray of the Statistical Assessment Service contends that even minor elevations of risk merit attention: "In observational studies [such as this], you're not going to get strong signals, so a small effect is noteworthy." An anthropologist, Murray says the study reflects "really good research. These are the best people in the field, and [they] try to control for all the [confounding] factors. But we're making the best out of a terrible situation, which is to try to make sense of human behavior."

AT THE HEART OF THIS episode lie the conflicting demands of journalism and social science.
"In their search for newsworthy stories, journalists seek what is new, interesting and unexpected. This approach contrasts with scientists' efforts to contribute to cumulative, enduring, generalizable knowledge that will stand the test of time and critical review," Nelson wrote with Ross A. Thompson in the January 2001 issue of American Psychologist. "Journalists report information to the public that is practical and relevant, whereas scientists report new knowledge to their peers that may be incomplete, abstract, or esoteric."
Scientists tend to qualify and hedge, journalists to simplify and amplify--with both groups citing public service as a motivation.
Alexander suggests the public was ill-served by unduly negative press about child care, which could erode credibility in the government study now funded through 2004. "This is a serious concern," he says. The study makes a "major impact on [the understanding of] child development and on what public policies ought to be."
He doesn't want it jeopardized. So, at a regularly scheduled meeting in May with the child care study steering committee, Alexander charged its members with developing a policy that "at minimum" ensures a paper has been peer reviewed "before we do any oral presentations in the future."
"We cannot take a risk like that again," agrees Friedman, the study's science coordinator. "Reporters will not know about the study until papers are in press. This is probably the policy that we will adopt."
Is that restriction warranted? And is it wise?
"It's not altogether a bad thing. We don't want gadfly social scientists to just put out things," concedes Shell, the Boston University science journalist. "But it has a chilling effect on getting information out to the public. It's putting us in the position of being policed by other professions, and that really diminishes our freedom and our professionalism. It's the information they want us to know. We [can] become conduits for ideologues."
In general, "an organization that attempts to control the message, when it's not an evidentiary issue, is simply going to look bad," says Sharon L. Dunwoody, director of the school of journalism and mass communication at the University of Wisconsin-Madison. "I don't think that works. It's a politically unsavory thing to do." She adds that it could be even more detrimental to research because it would curtail the "really crucial sharing that needs to go on among scientists."
Nelson, the Minnesota professor, agrees that early findings should be emphasized as such--or delayed until they've been evaluated by other scientists. He supports Alexander's call for a revision on how the child care researchers present results. "To have a press conference elevates early research to a level [to which it] should not be elevated," Nelson says. "The stakes are too high, the policy implications are too great. Science should be cool and objective."
Shell says journalists invite limitations "if we just take what we're given and put it out without concern for the outcome." Responsible journalists show enterprise, review the evidence, discuss it with other experts and report it in context.
"It's always our responsibility," Shell says. "What the press has to do is weigh these kinds of reports very carefully."