A new research is reveling a major theory of 'randomised' arrest of experiment at 23 years ago discover so as to domestic violence wounded whose associates were under arrest on wrong charges – mostly without cause injury and the cases were more than 64% who died early, as compared with the victims whose partners be warned though not detached by police force.
many famous study's authors are saying causes which are currently not known though such health impacts can create consistent health impacts on chronic stress that can be amplified by arrested partner. They may can call for a "robust review" of UK and US mandatory capture policies in the cases of domestic violence.
Professor Lawrence Sherman from Cambridge University's Institute of Criminology, who authored the study with his colleague Heather M. Harris from University of Maryland said "It remains to be seen whether democracies can accept these facts as they are, rather than as we might wish them to be,".
I'm actually a little troubled by the fact that the study's author is so confident in claiming that there is an actual causal link. Maybe it's a "medical mystery" because there is no causal mechanism.
ReplyDelete(http://www.cam.ac.uk/research/news/mandatory-arrest-in-domestic-violence-call-outs-causes-early-death-in-victims)
For one, random assignment is not the same thing as random sampling. (They teach this stuff in undergraduate courses.) Random assignment can help average out individual differences between the two groups (arrested vs. warned), but it does not help eliminate the effects of confounding variables. I hope Dr. Sherman doesn't teach a research and methods course.
The study in question did not have random sampling; it was a convenience sample. You did not randomly assign these people into a "you get a physically violent partner" or "you do not get a physically violent partner" group.
And it seems absurd even bringing this up but you cannot randomly assign someone to be "black" or "white". Related to this, this was also not a representative sample. Your sample was 70% "black" whereas the population is somewhere around 40.0%. (Correct me if I'm wrong on that one; I used Wikipedia to check)
The huge difference between "black" and "white" victims is a strong indicator that something other than the arrest itself is causing the earlier deaths of these people. I say "earlier" because there was only a 6.2% chance ([70/1125]*100) of dying if your partner was arrested compared to 1.9% chance ([21/1125]*100). Are you really confident that some other mixture of confounding variables (individual or historical) can't account for 4.3%?
Had the coin flip gone the other way and you had 2/3 of the participants in the "warned group" instead of the "arrested group", there probably wouldn't have been a story here. Again, this was not an experimental study. There was no random sampling and there was no proper control.
A proper control would be comparing these victims to randomly selected, demographically-similar people who did not have their partners arrested. That is, what is the death rates per 1000 people in the general public over the same period of time? What does it say about your study if it's more than 4%?
It's not stated in the article but were other confounding variables even
controlled for: age and overall health; socioeconomic status; access to health care; number of children; etc.? Were such variables even recorded? Was there even a multivariate analysis done on the data? Or did it all come down to comparing a single group of people (domestic violence victims) on two independent variable (arrested or warned) for a single outcome (death or not). If so, that's some weak sauce.
The central claim being made is that it is the simple act of having your partner arrested makes it more likely that you'll die. The study's methodology does not allow for such erroneous conclusions and I hope this story doesn't get too much attention.
If you want to make casual claims, you'll have to at least get an actual random sample (of ALL POSSIBLE PEOPLE IN A POPULATION) and start randomly arresting peoples' partners for minor infractions (preferably different kinds of crimes) and see if the simple act of arresting one's partner makes it more likely you'll die. I imagine it doesn't.