Schneier on Security

Jeff Dege • May 2, 2008 6:59 AM

What’s misleading here is the suggestion that the risks of attack by a family member are shared equally by everyone, and they’re not.

Most families are at zero risk of attack by a family member, some are at high risk. Throwing both groups into the same statistical pool does nothing but confuse things.

There have been studies examining prior police contacts with households in which domestic homicides have occured – more than 90% of cases had prior police calls to the address with the average having had police called to the household six times.

Domestic violence doesn’t just spring up out of nowhere, it’s a long-established pattern. If you aren’t experiencing it now, your risk of experiencing it in the near-future is effectively zero.

Jeff Dege • May 2, 2008 9:49 AM

Try this:

“Intrafamily murderers are especially likely to have engaged in far more previous violent crimes than show up in their arrest records. But because these attacks were on spouses or other family members, they will rarely have resulted in an arrest. So domestic murderers’ official records tend not to show their full prior violence, but only their adult arrests for attacking people outside their families. Most family murders are preceded by a long history of assaults.”
[http://www.ncjrs.gov/App/Publications/abstract.aspx?ID=127648]

More:
“Intrafamily murderers are especially likely to have engaged in far more
previous violent crime than show up in their arrest records. But because
these were attacks on spouses and/or other family members, they will
rarely have resulted in an arrest.{169} So domestic murderers’ official
records tend not to show their full prior violence, but only their adult
arrests for attacking people outside their families. Therefore only about
“seventy to seventy- five percent of domestic homicide offenders have been
previously arrested and about half previously convicted.”{170} As to how
many crimes they perpetrate within the family, even in a relatively short
space of time, “review of police records in Detroit and Kansas City” shows
that in

 90% of the cases of domestic homicide, police had responded at least

once to a disturbance call at the home during the two year period prior to
the fatal incident, and in over half (54%) of the cases, they had been
called five or more times.{171}

A leading authority on domestic homicide notes: “The day-to- day reality
is that most family murders are preceded by a long history of
assaults….” Studies (including those just cited) “indicate that
intrafamily homicide is typically just one episode in a long standing
syndrome of violence.”{172}”
[http://homepage.usask.ca/~sta575/cdn-firearms/Digests/600-699/v01n667]

Why the difference? I can easily see the possibility that cultural differences result in significant differences as to when and how often police are called, or how often police actually create a written record when then are called (which would be indistinguishable to after-the-fact data mining studies.)

I also get quite cautious when I see qualifications of the sort included in the study you cite:

“It found that nearly three-quarters of spousal homicide offenders had no prior contact with police for spousal violence.”

How is the study defining “spousal violence”? How many police contacts did the researchers not count, because they didn’t consider it to be “spousal violence”?

Another Canadian study:

“The study indicated that factors associated with firearm-related domestic homicide include: a history of violent disputes (however, in the majority of these disputes the police do not have previous knowledge); a history of substance abuse; drunkenness; an existing criminal record; separation or pending separation; and, employment and financial problems.”
[http://www.cfc-cafc.gc.ca/pol-leg/res-eval/publications/reports/1990-95/summaries/dom_sum_e.asp]

That the police are sometimes unaware of the pattern of violence in these cases does not mean that the victims are not.

I’ll say it again – if your domestic partner has not assaulted you in the past, the odds that he or she will do so in the future is nil.

Jeff Dege • May 2, 2008 11:21 AM

“Irelevant. I was responding to Drege’s ridiculous claim:

“There have been studies examining prior police contacts with households in which domestic homicides have occured – more than 90% of cases had prior police calls to the address with the average having had police called to the household six times.”
”

Why is it ridiculous? Because it contradicts your preconceptions? Had you done a quick Google search you’d have found much discussion over whether the results of these studies were typical, or whether what these studies had found offered police any predictive value, but you’d have clearly found that these studies existed.”

http://www.ncjrs.gov/App/Publications/abstract.aspx?ID=42215

DOMESTIC VIOLENCE AND THE POLICE – STUDIES IN DETROIT AND KANSAS CITY

Abstract: THE STUDY INDICATED THE NEED FOR FURTHER EFFORTS TO EXPLORE, IN AN EXPERIMENTAL AND CAREFULLY EVALUATED MANNER, WAYS OF REDUCING THE CHANCE THAT A DISPUTE WILL LEAD TO AN ASSAULT AND AN ASSAULT TO A HOMICIDE. IN KANSAS CITY IT WAS FOUND THAT THE POLICE CAN OBTAIN SOME EARLY WARNING OF ASSAULTS AND HOMICIDES. THE POLICE IN THAT CITY WERE CALLED TO THE LOCATION OF AN ASSAULT OR HOMICIDE AT LEAST ONCE BEFORE IN ABOUT 85 PERCENT OF THE CASES AND FIVE OR MORE TIMES IN 50 PERCENT OF THE CASES. IN SHORT, ANY GIVEN ASSAULT OR HOMICIDE ARREST IS LIKELY TO BE THE CULMINATION OF A SERIES OF POLICE INTERVENTIONS. WHILE THIS STUDY DOES NOT TEST METHODS OF HANDLING THOSE INTERVENTIONS SO AS TO MINIMIZE THE CHANCES THAT SUBSEQUENT ASSAULTS AND HOMICIDES WILL OCCUR, THE AUTHORS POINT OUT THAT THAT IS THE OBVIOUS NEXT STEP.