> It's like flipping a coin: if you flip a fair coin just a few
>times, it is possible that you could get a run of flips that
>deviate substantially from 50% heads and 50% tails.
Its possible, but statistically quite unlikely. It is remotely possible that a DNA match is co-incidence too, but statistically so unlikely that we ordinarily discount it.
> In the same way, even if there were no substantial racial bias or
>disparity in homicide sentencing, you could get 12 murderers whose
>victim pool does not perfectly represent the existing proportions of
>ethnic groups living in their area.
If they were simply executing murderers randomly, I suppose it is possible that by chance it might happen that only those who murder one particular ethnic group might be selected. Without any prejudice involved. Awfully unlikely though, the smart money wouldn't be on that horse.
However the selection of which murderers to execute is not supposed to be random. Though looks like that might be a fairer way of choosing who to execute.
>On the other hand, I recall larger scale studies that show the same
>effect: in regression models that include a variety of
>sociodemographic factors, the race of the victim is a statistically
>signficant predictor of a death penalty sentence in homicide cases.
>If I recall this correctly, Woj's criticism is valid for the example
>that Dennis C. brought up, but not for these larger scale archival
>studies.
No, it isn't logically a valid criticism. On its face. His criticism was based on possible sampling error, but it wasn't a sample, but the entire number. The only way his criticism would have any merit is if he could show that the 27 white victims of those executed was a roughly representative sample of all murder victims over that same period, in terms of the relative seriousness of the extenuating circumstances. Which seems so unlikely that it is not even worth the trouble to check it out.
The numbers are compelling proof, beyond reasonable doubt (the remote possibility that in those years the only murders involving extreme circumstances with no mitigating factors, involved only white victims is not a "reasonable" doubt). The system is guilty, denial is simply mischievous.
Bill Bartlett Bracknell Tas