摘要: Abstract In summary, then, 6-year driving records (accidents and convictions for traffic violations) were obtained a representative sample of 7841 California drivers, both sexes all ages. From analysis these record data the following conclusions appear warranted: (1) By far vast majority accidents involve previously accident-free that is. accident-repeater is responsible only small portion accident problem. Therefore, safety efforts directed toward so-called “normal” driver will have better chance pay-off. (2) Conviction experience more stable over time than experience, are males females. predicting future performance be successful accidents, (3) Driving information becomes reliable as period which it accumulated increases. prediction performance, or generalizing any other purpose, desirable to use long possible base.