Search results

The importance of the soc sci's has been much exaggerated. The soc sci'ts are particularly bad at prediction, & so at helping us to forestall the SP crises that occur more & more often. The rate of soc change has gone up very sharply for a number of reasons, so surprises are more frequent. This rate should ae reduced, at least in rich countries, which face far more unknown situations. Poor countries, on the other hand, mainly face situations known from those slightly more advanced. What poor countries mostly need, therefore, is better information, of old kind, not new kinds of information. Crisis-anticipating information will mostly be valueless, since no one knows how to winnow it. It also requires many low-level informers, & the identity of these people is a major problem of pol'al power. So is the identity of the processor & publisher of their data. Substantial improvements in information & its use might abolish democracy. HA.

A symposium contribution on, "The Supreme Court Forecasting Project," highlights the statistical model's erroneous prediction that the Court would affirm the Texas Court of Appeals' judgment that Texas law criminalizing gay sex is constitutional in the gay rights case Lawrence v. Texas. Although the prediction resulted from a programming error that was later fixed to produce a revised prediction of a 5-4 reversal, the corrected model still missed Justice Kennedy's vote to reverse the lower court even though Kennedy's empathy for gay rights was evidenced in earlier cases. A look at why Lawrence v. Texas was the most predictable of all major cases of the 2002 term is followed by an account of personal predictions for 27 cases covered during the same term as a member of the Supreme Court press corps. Specific cases are detailed & the personal predictions are compared with those of the statistical & expert models. J. Lindroth