Re: Severity scores in Child abuse data

[Brett Drake <brettd@xxxxxxxxxxxxxxx>]

> I would like your advise on how one should handle data on children with
> multiple child abuse severity scores.  My first thought would be to merely
> add up the scores from each type of abuse, but I am worried that there may
> be some negative implications caused by going this route.  

In my view, you should not mix subtypes of abuse at all.  If possible, 
you should always break everything down by Sex/Phy/Neg and analyze 
separately.  You may even want to include a "multiple types" category, 
which in practice will be mainly Phy+Neg.  Failing this, I would just 
look at the single most serious rating.

Many studies have shown different dynamics in abuse subtypes, and 
etiological theory definitely suggests that they should be viewed 
differently.

I would *not* add ratings cumulatively, as this confuses severity of abuse
with "breadth" of abuse.  In my view, you don't want a kid with moderate
physical abuse and moderate neglect being graded as equal to a kid with
with very serious physical abuse or very serious neglect. 

This is especially bad if emotional abuse is included, since a strong 
case can be made that all types of child maltreatment imply emotional abuse.

-Brett Drake
Washington University