Remember the great reefer madness scares in the late 20th Century? Maybe marijuana itself wasn’t so bad, it was said, but it’s a gateway drug!
Now that people have finally caught on that much of what family policing agencies (a more accurate term than “child protective services” agencies) do is confuse poverty with neglect, the child welfare establishment has come up with a similar excuse to justify all that surveillance of impoverished families and removal of their children: Neglect, they suggest, is a gateway allegation. And that’s even worse. Because while a gateway drug, it was said, might lead people to worse in the future, a gateway allegation is a claim that something far worse has already happened to innocent children.
Making that argument requires taking data out of context — either by presenting numbers without ratios or ratios without numbers.
The out-of-context data make for good sound bites. Understanding why they’re so misleading requires a trip into the weeds.
Numbers without ratios
One of the arguments used by those claiming neglect is somehow much more serious than “just” poverty is that, they say, more children die of neglect than abuse. For reasons I’ll get to below, we don’t know if this is true. But let’s assume for now that it is. By taking that figure out of context, it might leave the false impression that every time workers go out on a case in which the allegation is neglect, the child is in more danger than when the allegation is abuse.
In fact, the opposite is true. The data show that any given neglect allegation actually is far less likely to involve a fatality than an abuse allegation. To explain what seems like a paradox, let’s look at the data.
According to the latest Child Maltreatment report, 630 fatalities involved physical abuse, while 1,217 are said to have involved either physical or medical neglect. But those raw numbers alone tell us nothing about the odds of any given case leading to a fatality. For that, we have to look at what caseworkers actually are investigating.
In 2020, 7 million children were reported to family police agencies. We know that among those cases in which workers investigated and then checked the “substantiated” box on a form, 16.5% involved abuse and 76.1% involved neglect.
That suggests that about 1,155,000 calls alleged abuse and 5,327,000 alleged neglect.
Now, divide the fatality numbers into the hotline call numbers. Here’s what you get: One fatality for every 1,833 calls alleging abuse — but one fatality for every 4,377 calls alleging neglect. Thus, when a hotline operator is deciding whether to screen in a given call and when a family police investigator knocks on a door to begin an investigation, the odds that the case will involve a family where a parent will kill a child are extremely low if the allegation involves abuse — and far lower if the allegation is neglect.
How can both be true? How can there both be more fatalities attributed to neglect but proportionately fewer fatalities in cases alleging neglect? Simple. Most of the time the caseworkers are not investigating neglect at all — they’ve investigating poverty.
You might be able to make a case that when there is actual honest-to-God neglect — children locked in closets, or starved, or where there’s a meth lab in the basement — such cases are more dangerous than typical cases of physical abuse, which can often involve a spanking that leaves a bruise. But the overwhelming majority of the time, horrific cases of starvation and the like are not what caseworkers are investigating.
Further complicating matters: Even in a fatality case, deciding if the cause is neglect can be subjective. Consider a hypothetical: A three-year-old wakes up before his parents one Sunday morning. The parents don’t know he’s figured out how to unlatch the back door. He wanders out of the house, falls into a body of water and drowns.
Accident or neglect? I suspect it’s far more likely to be labeled neglect if the body of water is a pond behind a trailer park than if it’s a pool behind a McMansion.
Ratios without numbers
The misuse of ratios without numbers occurs in Dee Wilson’s expenditure of 2,500 words trying to scare us into thinking that neglect is merely the tip of a horrifying iceberg. So he cites a study that included this:
“Of particular note was the strong relationship between physical neglect and sexual abuse, with any physical neglect associated with 9.07 times the risk of sexual abuse,” compared to children without reported neglect.
Oh my God! Sounds like every parent investigated for neglect is also probably a rapist, doesn’t it?
But again, take a closer look. This time what’s left out are the raw numbers. The study used data from the National Surveys of Children’s Exposure to Violence. The total sample size was 7,872 children from ages 2 to 17. Of that number, 530 experienced what the study defined as physical neglect and 826 experienced what the study defined as “supervisory neglect.”
The number who experienced sexual abuse that would come under the jurisdiction of a family policing agency, that is, abuse by a parent, sibling or other relative: 41. (In contrast, the study found a much higher rate of sexual abuse by peers — which may help explain the high rates of abuse in foster care, group homes, and institutions.)
So I emailed the lead researcher, Prof. Heather Turner of the University of New Hampshire, and asked this question:
“[O]f the 530 victims of physical neglect how many also were victims of sexual abuse? Of the 826 victims of supervisory neglect, how many also were victims of sexual abuse?”
Prof. Turner replied that this question “would require additional analyses that I am not able to provide at this time, as my coauthor/analyst for this paper is currently unavailable.”
But this much seems clear: Overwhelmingly, the 1,356 young people in this study deemed victims of physical or supervisory neglect were not, in fact, also victims of sexual abuse by a parent or guardian.
Wilson cites another study for which he does cite raw numbers — selectively — while avoiding percentages. Perhaps he did not want to write that in this study, of all those reported for the one form of neglect he singled out, 82.4% were not subsequently reported for sexual abuse. (When all forms of neglect discussed in the study are counted, 85.6% were not reported for sexual abuse.)
Fearmongering by using giant ratios involving very small numbers is nothing new. Take, for example, a study claiming that “Children residing in households with adults unrelated to them were 8 times more likely to die of maltreatment than children in households with 2 biological parents.”
What the study doesn’t say is that, even if you double official estimates of child abuse and neglect fatalities, in any given year 99.995% of American children won’t die of child abuse or neglect. So this study means only that if there’s a boyfriend in the house the odds of a child being killed by an adult in the home go from infinitesimal to slightly less infinitesimal. But the headline on a story citing this study is: “Why Are ‘Mothers’ Boyfriends So Likely to Kill?”
Similarly, Wilson reinforces his out-of-context data with a horror story about a case in which the family police had gotten 52 reports before a child died.
Fearmongering with figures endangers all children. It endangers those subjected to the trauma of needless investigation and the even greater trauma of needless foster care. It deluges caseworkers with needless investigations, leaving them even less time to find those few cases in which children are in real danger. But it also keeps an entire industry going — an industry of scholars and administrators and consultants all invested in convincing themselves and then convincing the rest of us that things that are relatively simple are incredibly complex.
The great filmmaker Costa Gavras, known for landmark political films such as “Z” and “The Confession,” once said:
“The issues in politics are not complex, even though the politicians tell us so in order to convince us of the politicians’ importance … and to keep us from criticizing them.’”
Turns out, in child welfare, sometimes, when you scratch a scholar or an administrator, you find a politician.