It's just been reported that girls, but not boys, who walk or ride to school tend to do better in school. The reasons for these results are not entirely clear, and the study has quite a few limitations. Regardless, such is one more reason to support biking and walking, and the safe routes to schools program. Making it safe enough for students to bike and walk to school is good not only for health, but for academic performance as well.
On the topic of risk, it's notoriously difficult to estimate in an objective manner. The most extreme example is probably people being afraid to have their kids walk/ride to school due to a fear of stranger abduction. Such abductions happen, but are extremely rare. But because those few abductions receive much press, the fear of such gets exaggerated. On the other hand, benefits from increased socialization and improved health are less visible in the short run so parents have a tendency to opt out due to the former when they don't have all the facts.