david friedman wrote: > But my proposal is not a study done by people on opposite sides of > the debate, but a study done by people who both sides of the debate > trust--which is a rather different thing. > > More specifically, my idea is for a study designed to estimate the > rate of false positives in the criminal justice system--the fraction > of the people convicted of serious crimes who are innocent. That is a > crucial statistic for opinions about the system, and one that nobody > has any accurate estimate of. Technological progress in DNA testing > makes it technically possible to reevaluate a large sample of old > cases, and get at least a lower bound for the fraction that convicted > an innocent person. > > Any such number will be even more controversial than Lott's results. > So you have the study done by two statisticians, one of whom > conservatives trust (for reasons having nothing to do with this > particular question), one of whom liberals trust. Lott would be an > ideal candidate for the first statistician. He has, so far as I know, > no particular bias on the question of how many innocent people are > convicted--but the right cannot attack him because of his work on gun > control. Sounds like a fascinating, useful study. I agree that it will be quite controversial. I don't think we disagree on how the researchers should be chosen--what you write above sounds quite sensible to me. Rather, I think that we disagree on the likelihood of an HCI/NRA collaboration actually choosing researchers in the fashion you suggest. How much do you estimate your study will cost? Does anyone know how much the Lott and Mustard study cost? How long it took to carry out? > > > >Of course, but given two equally high quality studies, one > >funded/managed by NRA > >alone, the other funded/managed jointly by NRA/HCI which do you think would be > >more widely trusted? Which do you think HCI would find harder to dismiss? > > I don't think a study that either NRA or HCI strongly disapproved of > would ever come out of such an institute. More generally, I think > creating "neutral institutes" by getting people on opposite sides of > an issue to fund them is not a very hopeful strategy for producing > truth. You may be correct--it may be too difficult to get people at organizations like HCI and the NRA to cooperate enough to fund (or release) a study that either NRA or HCI disapproved of. I think that problem could be addressed by funding the study in such a way that the researchers got paid no matter how the results turned out, such as by putting the money in an escrow account, to be released once the study results appeared in one of a number of a peer reviewed journals. > You are better off recognizing that most researchers have > views of their own, and using the usual academic machinery of > competing peer reviewed journals, rebuttals and rerebuttals to sort > things out. It doesn't work very well, but I think it works better > than "officially neutral" studies. I agree that researchers will have views of their own. And it may be difficult to get two researchers with different ideologies to cooperate. Do you think that there are intellectually honest researchers who support gun control? Do you believe there are any well-designed and carried out studies have been done that produced results that supported increased gun regulations? If you believe intellectually honest gun control supporters exist, do you think it would be possible to agree on a study design with one of them, prior to carrying it out, that you both could agree was scientifically sound? Conversely, if you're in favor of increased gun control regulations, do you believe there are any well-designed and carried out studies have been done that produced results that supported reduced gun regulations? > Consider the role of the Center for Disease Control in the firearms > conflict. They try to get authority by claiming to be neutral--after > all, the CDC has nothing to do with firearms regulation--but it's > clearly bogus. I would rather have them admit what side they are on > and then fight it out in the journals with people on the other side. I agree --claiming to be neutral is not at all the same thing as actually being neutral. And while the CDC may have nothing directly to do with firearms regulation, it is staffed by government employees who presumably are sympathetic to coercive solutions to social problems (else they wouldn't be working for the CDC). How do you think ideologically driven organizations, such as the NRA and HCI, should fund research? How can they reduce charges of bias? For example, suppose Charlton Heston came to you and asked your advice: "We at the NRA are confident that objective research would support increased gun ownership. However, given our values and mission, if we fund such studies ourselves, we're afraid that while the results of such studies may give succor to our existing supporters, they will be unpersuasive to those who have yet to make a decision on the issue. How can the NRA fund research that won't be easily dismissed as thinly veiled NRA propaganda?" How would you answer him? (Substitute the HCI, if you support gun control regulations.) Chris Rasch
