are thought to be especially risky for other reasons as well. Theories, issues, and sta­ tistical methods in vogue at the time they were planned may not be in vogue when they are finally executed. Perhaps worse, according to some perspectives, the structure of causal factors may shift during the execu­ tion of a longitudinal project such that in the end its findings apply to a reality that no longer exists. These fears and expectations assume an ever-changing world and a corresponding conception of research as a more or less disciplined search for news. Such ideas belittle the contributions of past research and leave us vulnerable to theories, programs, policies, and research agendas that may have only tenuous connections to research of any kind.