Re: Evaluating Abuse Cases

["Thomas R. Chibucos" <tchibuc@xxxxxxxxxxxxxx>]

Fellow Netters

Mark Chaffin's comment below is as accurate a statement as can be made on
the subject.

I would add and extend Chaffin's comments in the following way.  In my
view, the statement also provides a very strong case for a focus on primary
prevention analogous to the approach that has worked extremely well in the
medical field.  We don't try to figure out who is "at risk" for TB, for
example.  Everyone gets innoculated.  To be sure, there is not a single
innoculant to prevent physical child abuse, or, indeed, any of the other
kinds of child abuse.  

However, we do know some  important demographic, historical and behavioral
variables, attention to which would dramatically reduce risk.  Here are a
few: substance use; social isolation; lack of preparation for parenthood;
absence of sustained support for parenting; parent's own mistreatment when
coupled with not having had a meaningful developmental support experience
to deal with that mistreatment; mental illness; sustained economic
deprivation; sexism; racism.  

Our collective focus should be on moving the entire  "continuum of
treatment of children" from the less favorable to the more favorable
direction--moving the mean of favorable treatment and all deviations from
the mean in a positive direction.  Yes, there will always be kids at the
low end, and yes, this entire approach requires sustained attention at all
levels of society.  But, the alternative seems to me to be the continued
search for scapegoats, and for putative easy fixes.  Mistreatment is
complex and difficult to predict and to understand, and it is
multidimensional in nature--just as "good" treatment of children is.  Why
should the fix be easy?  Rather than continued feeble attempts to predict
risk (and, again, the data are absolutely clear as to how useless this
process is), why not tackle the problem from an
"individual/family/community/society assets" perspective?  To wit: what
variables predict healthy(ful) behavior in children and healthy(ful)
treatment of children (by parents, teachers, "the community," society)?
Social science research actually does know a lot about this.

The research-clinical-policy nexus is clear, and neither researchers nor
clinicians are intelligently informing policy on these matters.  But, this
is not new.  Some 30 years ago when I was a graduate student, the data were
clear that clinical prediction was worse than useless.  This is still the
case: the data are stronger than ever, yet people keep using the oxymoronic
clinical prediction!   For reasons that (what?..are
delusional?...self-serving?...make people feel useful?) the irrelevance
continues.  It is like looking for child abusers among strangers to
children when the empirical evidence is overwhelming that it is close
relatives and significant others who account for all but a tiny proportion
of direct child abuse.  Ethics, indeed!   How ethical is it for researchers
and clinicians to pretend that they can do something they cannot do.  It's
like obtaining a significant r = .20 between two variables, then writing a
discussion section as if this means something useful.  The saving grace of
the latter is that hardly anyone reads the stuff--unlike the pap that leaks
out because it addresses a significant social problem.

We have a collective responsibility to effect an immediate paradigm shift
in this area, the hallmark of which should be to transfer to the
prevention, treatment, and policy realms research information that is,
indeed, reliable and valid.  When we don't have such data-based
recommendations to make about applied matters, we should at least shut up
about it, or at best, actively discourage unsupported use of approaches,
instruments, etc.

Sincerely,  Tom Chibucos


At 10:18 AM 2/9/99 -0600, you wrote:
>	We are currently conducting an NCCAN-funded controlled treatment
>outcome study for physically abusive parents and have struggled with the
>question of assessment.  In addition to looking for measures related to
>treatment outcome (Eyberg's DPICS-II observational coding system, Conflict
>Tactics Scale P/C, Child Abuse Potential Inventory, etc.), we were
>interested in instruments measuring empirically documented risk-relevant
>constructs (e.g. depression, antisocial personality, substance abuse) and
>consequently are using the BDI and substance abuse and ASP modules from the
>Diagnostic Interview Schedule (DIS) in addition to  the usual risk-relevant
>historical and demographic factors (age, number of children, number of
>previous reports/incidents, etc.). 
>	However, I must say that this entire discussion is a bit
>disconcerting because, as far as I'm aware, psychological testing has been a
>poor performer in predicting anything of forensic relevance in child abuse
>cases.  My reading of the risk literature is that demographic, historical
>and behavioral factors are the most robust predictors of risk--not tests.
>Possibly the only thing less accurate than tests is clinical judgement.  I'm
>sure we're all aware of the literature documenting the poor track record
>clinical impressions have in predicting any kind of future violent behavior.
>The CAPI has good predictive validity for screening purposes, but is not
>usually recommended for assessing a forensic case.  Perhaps what we should
>be doing is more of an actuarially based historical-behavioral analysis and
>not muddying the waters with test information which is mostly error
>variance. This is exactly what people are doing in assessment of sexual
>abusers (i.e. Quinsey, et al's VRAG, Hanson, et al's RRASOR, etc.), which
>makes me wonder why we are wasting so much public money giving physical
>abusers MMPI's and Rorschachs to determine......well, to determine what?
>	If anyone is aware of any empirical evidence that projective tests,
>or any other personality test, offer any incremental predictive validity in
>child abuse cases, I'd love to see it.  In the absence of this, I must
>respectfully disagree with the author who suggested that he could give any
>test he pleases.  My reading of our ethical requirements is that
>psychologists should only use tests which have clear empirical support for
>the specific purposes and with the specific population where they are used.
>I think we owe our clients, our courts, and our CPS systems something better
>than "clincal judgement," regardless of whether or not its haphazardly
>butressed with MMPI's and Rorschachs.
>
>Mark Chaffin, Ph.D.
>Center on Child Abuse and Neglect
>University of Oklahoma Health Sciences Center
>mark-chaffin@xxxxxxxxx
>
******************************************************
Views expressed are mine alone, unless otherwise noted.
******************************************************
Thomas R. Chibucos, Ph.D.
Professor, Human Development and Family Studies
Chair, School of Family and Consumer Sciences
217 Johnston Hall
Bowling Green State University
Bowling Green, OH  43403-0254    Phone: (419) 372-7823    FAX: (419)
372-7854