Published Tuesday, April 22, 1997, in the San Jose Mercury News   How to fight crime   Remember midnight basketball? During the debate over the 1994 federal crime bill, it was touted as a remedy for juvenile crime. Well, a new study suggests midnight basketball is no ``impact player'' against crime, just a bench warmer. But wait. The nation's huge prison construction binge doesn't appear to be reducing crime either. The same goes for D.A.R.E., the drug education campaign now used in 70 percent of the nation's schools, including many in the Bay Area. The most comprehensive study ever of crime prevention came out last week. Its findings discredit some of the most popular anti-crime strategies around, both liberal and conservative. Asked by Congress to evaluate federal crime spending, a team of University of Maryland criminologists set out to determine what works, what doesn't work, and why. What doesn't work are juvenile boot camps, neighborhood watches, drug education for schoolchildren, and prison construction. These tend to be short-term programs that don't change the fundamental thinking or behavior of youths or improve the conditions in which they live, the researchers said. Here's what works and why: Intensified police patrols in high-crime neighborhoods; drug treatment in prisons; and home visits by nurses, social workers and others for infants in troubled families. These programs focused directly on problems and prevented greater problems. Concentrated policing stops petty crimes as a way of heading off felonies. Infant visitation has lasting effects for children and their parents. If there's a weakness to this study, it's that the researchers didn't have an ideal amount of information available to them. The sad fact is that Congress has never asked for the kind of scientific evaluation of crime programs that it demands, for example, in the testing of new drugs. Still, this study rings true. It's consistent with what some top law-enforcement experts and researchers have been saying fora long time. Here are results of key surveys done in the 1990s: Three out of every four big city police chiefs, and three out of five in the overall sample of police chiefs, said the best way to reduce crime and violence is to increase investment in programs that help all children get a good start. A Northeastern University study showed that intensive early intervention programs can reduce the later delinquency and criminal behavior of at-risk youths by as much as 80 percent. The nonpartisan RAND Corp. think tank concluded that every dollar for treatment of cocaine addiction was worth seven dollars in law enforcement. President Clinton and the Republican Congress ignored that good advice when they passed the $30 billion crime bill that, among other things, is supposed to put 100,000 new cops on the streets. Instead of sending those cops where they're needed most, let's say to the top 20 crime-ridden cities, the officers will be spread throughout every congressional district in the land. But this study is important even if federal anti-crime dollars disappear. Crime policy is more of a local and state affair than a national one. Many county jails and prisons in California offer at least some drug treatment. Some cities, including San Jose, have adopted concentrated ``community policing.'' Many of the programs that scored highly or badly in the Maryland study are being tried somewhere in the Bay Area. One of the University of Maryland study's authors states: ``We need to put the money where the crime is, not just where the votes are.''