The bulk of the evidence for a causal relationship between television violence and violent behavior comes from the research of Leonard Eron of the University of Illinois and Rowell Huesmann of the University of Michigan. Beginning in 1960, Eron and his associates began a large-scale appraisal of how aggression develops in children and whether or not it persists into adulthood; the issue of television violence was initially a side issue to the long-term study. Eron's data reveal that aggressive kids who turn into aggressive adults like aggressive television, but this is a correlation; it is not proof of a causal influence. Many of the studies cited to support the contention that TV violence causes violent behavior do report that children show higher levels of interpersonal aggression shortly after watching violent, energetic entertainment. A 1971 study by Feshbach and Singer had boys from seven schools watch preassigned violent and nonviolent shows for 6 weeks. The results were not constant from school to school, and the boys who watched the nonviolent shows tended to be more aggressive. Another protocol, conducted in Belgium as well as the United States, separated children into cottages at an institutional school and exposed certain groups to violent films. Higher aggression was noted in all groups after the films were viewed, but it returned to a near-baseline level after a week or so. Not only is the link between violent TV programming and subsequent behavior weak and short-lived, but the research does not provide a clear-cut mandate for the specific content of television drama. There can be no guidelines that would ensure that people will write, produce, or perform nothing that can be used by a psychotic to inflict harm. There is no way to predict how various individuals will react to various program content. There is also the danger that significant and constructively informative drama will be gutted by removing all negative portrayals of human behavior.