Reporters and editors have a love-hate relationship with statistics.
On the one hand, they sound so specific. So factual. They lend substance to a story.
On the other hand, statistics can be misused or raise more questions than they answer. To be useful to reporter and reader alike, statistics need explanation. That takes time.
Without such support, a story can lose credibility or confuse. Most of us have no way to determine whether a statistic has become distorted in repetition or was meaningful in the first place.
The subject of sexual abuse and assault is ripe for such confusion. Experts caution that the actual numbers of victims and offenders cannot be known with certainty. Where does that leave the rest of us?
Last Sunday a story in the Living Today section discussed child molesters, their methods and ways to prevent children from being victims. The package was a useful approach to an important topic.
In the course of the story, Cory Jewell Jensen — who with her husband, Steve Jensen, treats adult sex offenders at their Center for Behavioral Intervention in Beaverton — cited some statistics.
She said that according to some studies, one in 10 men has molested children. The most conservative estimate, she told Cheryl Martinis, a correspondent for The Oregonian, puts the figure at one in 20.
Some colleagues in the newsroom found the one in 10 not believable, and maybe the one in 20 as well. They wanted to know where those numbers originated.
Martinis properly noted in the story that data on sex abuse varies widely. She explained that the Jensens base their estimates on various studies.
Still, that one in 10 figure cried out for further checking.
Cory Jensen told me last week that the figures are “even shocking to people who work with sex offenders,” as she and her husband do. “It’s not a very well published or talked about part of the field.”
Her primary source, she said, is a 1988 book by John Crewdson, now a senior writer with the Chicago Tribune, called “By Silence Betrayed: Sexual Abuse of Children in America.”
Crewdson, in turn, cited a study by the Los Angeles Times. The 1985 study surveyed 1,260 adult men. Needless to say, men who molest children are not likely to respond “yes” if asked about it. The survey used a complicated method that allowed respondents to answer without revealing themselves. (It is described in chapter four of the book, published by Little, Brown and Co.)
“The number (the study) came up with was astonishing,” Crewdson wrote, referring to the one in 10 figure. With a standard margin of error figured in, the author said, the figure might be one in 15.
The poll takers then added in still more factors that might have caused error. The result was that an “absolute minimum” of one in 25 men has molested a child.
(Definitions of “molest” might vary from study to study. Jensen said her references are to “hands-on” actions.)
One in 25 is less startling than one in 10 (if hardly reassuring), but how do those 1985 findings compare with others, and are they credible?
Jim Hopper, who has a doctorate in clinical psychology, is a research associate at the Boston University School of Medicine. He also is a researcher with the Trauma Center in Brookline, Mass.
His work is focused on the effects of child abuse and other trauma, rather than on offenders. However, he said, in an e-mailed response to my questions, he also has done two smaller studies that involved men graduate and undergraduate students with an average age of 26.
In one study, 5.9 percent said they had committed sexual abuse against a child, and in the other 5 percent admitted as much. In either case, that would be about one in 20.
(Hopper’s definition of abuse included any abuse — including noncontact abuse — that used force or threats; sexual contact by a person 5 years older than a child of 13 or younger; or contact by a person 10 years older than a child of 14 or 15.)
Given those studies (one published, the other not), Hopper said he is “not surprised by a 1 in 20 number, but 1 in 10 definitely sounds high.”
Hopper is candid about the misuse of research and the ease with which it can be misleading.
“Even the most objective scientific research is imperfect,” he wrote in a paper, “Child Abuse: Statistics, Research and Resources” (available online at www.jimhopper.com/abstats).
Those experts “who claim to be without bias are fooling themselves or trying to fool you,” he wrote.
Furthermore, “All statistics on the incidence and prevalence of child abuse and neglect are disputed by some experts.”
The media don’t help in the matter of using statistics thoughtfully, he said.
People often do not believe the extent to which a problem is significant, he wrote, unless someone backs it up with “impressive-sounding statistics.”
As a result, the media often insist on such numbers for their stories “even if no good ones exist.”
Furthermore, when newspapers, magazines and broadcast outlets do cite studies, they often do not explain how the problem was defined, what methods and questions were used, who was studied or who asked the questions.
That is Hopper’s view. However, the fact is that editors and reporters know it, too. We can be our own worst enemies when we spot an attractive set of numbers.
Perhaps a news story cannot include lengthy detail on a study. However, reporters can educate themselves on the research and evaluate its usefulness to the story. They also can give readers enough information to make their own judgments.
The Oregonian and other newspapers use such practices in reporting political polls. We can do the same when citing studies on other significant matters.
In any event, journalists and readers alike should keep in mind Hopper’s cautionary note: Even the most objective research is imperfect.
