By Douglas J. Besharov and Jacob W. Dembosky

This article originally appeared in Slate, October 3, 1996.

With a new government-funded study in hand, Secretary of Health and Human Services Donna Shalala diagnosed a rising epidemic of child abuse last month. She reported that "child abuse and neglect nearly doubled in the United States between 1986 and 1993"--and that was only the beginning of the ugly news. The number of "serious" cases had quadrupled, and the percentage of cases being investigated by the authorities had actually declined by 36 percent, trends that she called "shameful and startling."

Is Shalala right? Is an unheeded child-abuse epidemic raging in America? Or, as I think is more likely, is the methodology behind the study and the interpretation of its numbers flawed? And if Shalala overstated the child-abuse peril, is she undermining public interest in the problem by making it appear too big and difficult to fix at a reasonable cost?

A Look at the Numbers

The secretary drew her statistics from the National Incidence Study of Child Abuse and Neglect, which was conducted by Westat Inc., a consulting firm that conducted similar studies in 1980 and 1986. In the new study, about 5,600 professionals, a representative sample, were asked by Westat whether the children they had served appeared to have suffered specified harms or to be living under specified conditions. Westat then determined if the reported "harms" and "conditions" met the study's definitions of "abuse" and "neglect," and generated estimates of incidence.

The odd thing about Shalala's claim that the number of children abused and neglected doubled from 1.4 million in 1986 to 2.8 million in 1993 is that no other signs point to such a dramatic increase in child abuse and neglect. Fatalities arising from child abuse have held roughly steady, ranging from 1,014 in 1986 to 1,216 in 1993, according to the National Committee to Prevent Child Abuse.

Of the 1.4 million additional cases reported, almost 80 percent fall into three suspect categories. (Anywhere from 13 percent to 34 percent of the 2.8 million children suffered more than one type of abuse or neglect. Unfortunately, the study did not "unduplicate" these reports. Nevertheless, the proportions I describe below provide a general picture of what is happening.)

Endangered children account for 55 percent of the increase. These are cases where the child was not actually harmed by parental abuse or neglect, but was "in danger of being harmed according to the views of community professionals or child-protective service agencies" [emphasis added]. (See Figure 1.)

Emotional abuse and neglect account for another 15 percent of the increase. The great majority of emotional-abuse cases, according to the 1986 Westat study, involved "verbal assaults," and more than half of the emotional-neglect cases involved "the refusal or delay of psychological care."

Educational neglect--the chronic failure to send a child to school--added another 8 percent to the total of new cases.

"Definitional Creep"?

All these cases warrant attention, but the explosion of numbers may be caused by the growing reportorial sensitivity of professionals, that is, "definitional creep." Professionals who become more sensitive to possible abuse, or more adept at noticing it, would make more reports to Westat--even if the actual incidence had not risen. In endangerment cases, at least, the study seems to accept this explanation.

According to Shalala, the number of "serious" cases increased between 1986 and 1993 from "about 143,000 to nearly 570,000." Her comments left the impression that the cases involved life-threatening assaults, but the study defines "serious" cases as any in which the child suffered "long-term impairment of physical, mental, or emotional capacities, or required professional treatment aimed at preventing such long-term impairment." Emotional maltreatment accounted for fully half of the increase in serious cases. (See Figure 2.)

In cases labeled as serious physical abuse, the reported injury could be mental or emotional.

Even in these "serious" cases, the study seems affected by definitional creep. For example, in three categories (sexual abuse, physical neglect, and emotional neglect), the number of cases described as "moderate" declined even as the number of "serious" ones increased--strongly suggesting that cases once viewed as only moderately threatening have now been "upgraded" to the most dire category.

Uninvestigated Cases?

Shalala's assertion that investigations of child abuse and neglect cases have dropped by 36 percent deserves closer scrutiny. In producing the number of uninvestigated cases, the study compared the number of cases identified by professionals with those known to local agencies. Of the cases not investigated, 33 percent involved educational neglect. (See Figure 3.)

The main flaw here is that most educational-neglect cases are handled by the schools; reports are made to protective agencies only when all else fails.

Another 30 percent of uninvestigated cases involved emotional abuse and neglect. But child-protective agencies usually avoid these cases because they tend to involve subjective judgments, and there is little that a quasi-law-enforcement agency can do about them.

Definitional creep is clearly at play here, too. Professionals who are increasingly willing to identify situations as harmful aren't necessarily ready to equate them with the sort of abuse and neglect they are legally obliged to report. And even if they did report these instances, child-protective agencies would still be expected to screen them out.

Does Shalala Believe Her Own Hype?

Probably not. Radical action would be required if Shalala's figures were even roughly correct. But instead of proposing radical action when she released the report, she outlined modest steps that had long been planned and budgeted.

Having worked in the field for 30 years, I can testify firsthand that the problem of child abuse and neglect is real. But however well meant, exaggerating the severity of abuse endangers children. In the late '80s, for example, the nation was told that 375,000 drug-exposed babies were born each year; Washington policy-makers were immobilized by estimates that tens of billions of dollars were needed to protect these children. In fact, the true number was closer to 35,000, and a decade later, the government has yet to mount a meaningful program for the children of addicts.

Overstatement may also obscure genuinely worrisome findings. Some of the increases in sexual abuse, physical abuse, and physical neglect uncovered by Westat may well reflect a true deterioration of conditions in disorganized, poverty-stricken households. But Shalala paid scant attention to this possibility.

And to claim recklessly that too few cases are investigated is to play with fire. Child-protective agencies are already overwhelmed investigating about 2 million reports a year, two-thirds of which are dismissed as unfounded or inappropriate. For many in the field, the most pressing need is to discourage inappropriate reporting--not to blithely call for more.

Figures for specific types of maltreatment exclude endangerment cases. Percentages total more than 100 because some children counted under more than one type of abuse.

Percentages total more than 100 because some children counted under more than one type of abuse.

These exclude cases of endangerment.