Thanks for your interest in this investigation. The three PIs are colleagues and I've forwarded your questions to them. Neil should be around as he is the one who just sent the piece to me. As for the others, well, we're on our January session break so I may not be able to find them until Wednesday when our Spring semester officially begins.
You do pose some interesting questions and I'm thinking that the differences in the graphical data you've notice are due in part to the difficulties that exist in defining and then measuring (or finding a proxy to capture) such variables or indexes as __social neglience__ or even __child abuse__. The __social neglience__ index includes homelessness and lack of health insurance coverage which are both rising, so I would except the index to rise as well. Are you thinking that since the child abuse variable included medical neglect, and the __social neglience__ index included lack of health insurance, that they should show similar trends?
At any rate, I'll pass along whatever info I get. Thanks again.
All best, Diane
Postscript: Hari, I hear you on the [Please, could I be spared the onslaught of simply pious lectures about reductionism etc] part :).
Hari Kumar wrote:
>Dear Diane Monaco
>Thanks for that. Could I ask you or others, some methodological
questions to help me understand the data? I wondered whether:
>i) There was a prior main hypothesis that the authors of this project
have stated a-priori? I could not see it on the site address you sent.
>See complete details at:
>http://www.manchester.edu/links/violenceindex/
>
>I suppose I should ask the PI's of the project & will if it does not
fall out of the conversation;
>But I am very uncomfortable unless large data projects specify up-front
a series of testable hypotheses that can be statistically refuted or
otherwise. [Please, could I be spared the onslaught of simply pious
lectures about reductionism etc].
>2) In the graphical data re the institutional variables for violence,
the one that is in stark apparent difference to the others is the law
enforcement & the institutional variables is social negligence. But the
social negligence is in apparent contradiction to the child abuse data.
I assume that statistical power in such vast data sets are simply of on
consequence. Is that right?
>Cheers, Hari Kumar