If you want to get people to surrender fundamental civil
liberties you’d better scare the hell out of them.
Donald Trump understands this; it’s been his m.o. since he
began his campaign. So when his administration started tearing apart families
at the Mexican border he
justified it with hype and horror stories.
The child welfare establishment gets it, too. For decades they’ve used horror stories – and
hyped, misleading statistics - to justify tearing apart families all over
America. No, it’s not identical, but there are more
similarities than differences.
As early as 1993, under the headline “Damned Lies and
Statistics,” Time magazine called
out what was then one of the worst offenders – Prevent Child Abuse America
(as well as other groups crusading for other causes) for using “flagrantly
flimsy figures.”
Ten years later, PCAA itself admitted
to this:
While the establishment of a certain degree of public horror relative to the issue of child abuse and neglect was probably necessary in the early years to create public awareness of the issue, the resulting conceptual model adopted by the public has almost certainly become one of the largest barriers to advancing the issue further in terms of individual behavior change, societal solutions and policy priorities.
It’s not that the numbers themselves are necessarily wrong –
but crucial context often is missing. So
we are often told that there are more than three million reports alleging child
abuse every year. But we’re not told that more than 80 percent of them turn out to be false and most of the rest
involve “neglect” – which often means only that the
family is poor.
The most misleading
statistic in child welfare
But the single most misleading figure in all of child
welfare is the one I discussed recently in connection with the Orange County Register’s descent
into “crack baby” journalism – the one about the proportion of child abuse deaths
“known to the system.”
It is a figure built around the worst possible tragedy –
child abuse fatalities. The only
acceptable goal for such tragedies is zero.
But if one really believes that even one such death is one too many,
then there is no need to distort data to leave a misimpression about how often such
tragedies occur right under the noses of caseworkers. And that very distortion makes it harder to
reach the goal we all share.
The more one favors a take-the-child-and-run approach, the
more one harps on this figure – because the impression it leaves for the casual
reader is vastly different from what the number really means.
So we read that “Of all the children who died in [name of
state or locality] [fill in percentage] were previously ‘known to the system.’”
Typically, the percentage will be anywhere from one-third to
two-thirds; the Register gave an even
higher percentage, but did not explain how it came to that conclusion.
So one might read in a news story: “Half of all child abuse
deaths involved children previously known to child protective services.” Wow. That sounds like half the time, children
known to CPS die. Of course the figure
means nothing of the sort. The real
meaning becomes clear when you add one crucial bit of context: The actual number
of children who become “known to the system” over the course of a year.
So now, let’s look at the numbers nationwide.
According to the latest Child
Maltreatment report from the Department of Health and Human Services, there
were an estimated 1,720 child abuse deaths in 2017. Of that total, 27.3 percent were previously
known, in some way, to authorities.
Those who promote hype and hysteria are always saying that
every figure they want to hype is an underestimate. But this time, let’s assume they’re
right. Just for the sake of argument,
let’s double the number of child abuse deaths and double the proportion “known
to the system.” That would give us a
total of 3,440 deaths, of which 1,878 involved children “known to the system.”
Adding the missing number
Sounds like children often, even typically, die right under
the noses of child protective services workers, doesn’t it? But now, let’s add the missing number – the
total number of children who become “known to the system” every year. In 2017 that number was over seven million children - 7,508,000 to be
precise. So now we’re talking not about
1,878 out of 3,440 – but 1,878 out of 7,508,000.
So every year, CPS agencies are expected to figure out which
of the 7,508,000 children will die if they don’t intervene in time. It looks a bit like the graphic at the start
of this post. This one:
But even that representation understates the difficulty
involved for frontline caseworkers. In the graphic, all of the deaths are in
one place. In real life, they are much
smaller needles scattered all over the haystack.
Even showing the needles in a different color can be misleading.
Because while some of the case files turn out to have more “red flags” than a
Soviet May Day parade, in many others there was no way to predict the tragedy
that would follow. As one former child
welfare official put it: “For every one of these [death ‘known to the system’]
cases, I can show you ten with identical circumstances where nothing went
wrong.”
And then there is the whole definition of “known to the
system.” Again, some of those deaths
were in cases where, in fact, the family was extremely well known to
authorities, the problems were obvious and anyone should have seen the
danger. But most were not. Sometimes “known to the system” means only
that a report several years earlier had been screened out at a child abuse hotline. So CPS investigators didn’t really know the
family at all.
Where do we put 7.4 million children?
So what do we do? Investigate the families of all 7.4
million children? Even if that could be
done logistically – and it can’t – that would traumatize millions more children
for no reason. And, of course, the
overload would be such that all the investigations would be sloppier and even
fewer needles in that giant haystack would be found.
Or should we skip the investigating and just take away all
7.4 million children? That way we know
we won’t miss the needles because we’ve vacuumed up the entire haystack. But where, exactly, would we put all those
children? And what would we do about
placing so many of them at high risk of abuse
in foster care itself?
So what’s the point of the most misleading statistic in
child welfare? It’s not to get us to take away all 7.4 million – but it is an
effort to get us to take away more of them, and traumatize more families with
needless investigations, and beat back any attempts to impose openness,
accountability and due process on the system.
But openness, accountability and due process are exactly
what we need. Because only a system that
relies less on coercive intervention will have the time and resources to find
those needles in the haystack, before it’s too late.
We know that because of something I’ve written
about before, a study done by a
liberal Texas think tank, the Center for Public Policy Priorities. The study looked at what does, and does not,
contribute to reducing child abuse deaths.
Let’s start with what doesn’t
contribute to more child abuse deaths. The study found:
● The rate at which people report child abuse does not
contribute to more or fewer child abuse deaths.
● The rate at which a state takes children from their
parents does not contribute to more or fewer child abuse deaths.
● The rate at which a state screens in reports for
investigation does not contribute to more or fewer child abuse deaths.
Here’s what does
contribute to more child abuse deaths:
● High rates of poverty
● High rates of teen pregnancy
● Low rates of services to prevent child maltreatment.
Which means, of course, the real solutions are pretty
obvious.
In contrast, the misuse of the most misleading statistic in
child welfare only makes things worse.
It encourages overloading caseworkers with even more false allegations,
more cases in which family poverty is confused with “neglect” and more children
needlessly torn from everyone they know and love. And that leaves the workers even less time to
find those very small needles in that very large haystack.
