Like the media abroad, Japan's press and television are criticized for sensationalized crime reporting - with one important difference. Critics say they are too slow and too timid in reporting criminal behavior by the nation's police forces. At a time when random crimes of violence are occurring with alarming frequency, it may be better to err on the side of excess rather than caution. A well-informed public is better able to protect itself, especially when doubts arise about the level of police assistance.

Government bureaucrats seem to take it for granted that many crimes are not being reported, at least not to law-enforcement authorities. The Justice Ministry has just announced a nationwide survey to determine the extent of such "hidden" criminal activity, crimes not reported for whatever reason and incidents not officially recognized as crimes when they occur. Japanese society has long been hailed as safe and comparatively crime-free, and that image has been cultivated by those likely to benefit from it. As the ministry surely knows, however, recent events are dictating a change in attitude. These range from spiraling rates of violent crime by juveniles, including killing without remorse for petty "entertainment" money, cruel abductions and murders of defenseless children and, yes, the rash of illegal acts by officers of the nation's police forces.

The ministry's survey is already under way. Some question how accurate it can be, since it is based on replies to a questionnaire submitted to only 3,000 individuals. The eagerly awaited results are to be announced late next month. The timing is curious, since the survey comes just as the nation's criminal justice and law-enforcement systems are under fire from many quarters for inefficiency and lack of fairness. This may be mere coincidence, however, since it is part of a worldwide research project of the United Nations International Institute for Criminal Law Reform and Criminal Justice Policy. Thirty-three nations, including the Group of Eight, are participating.