"good," only 39 percent of blacks offered that evaluation. Although most Americans express confidence in the police, such expressions may have more to do with a sense of gratitude toward or dependency on the police than with an objective assessment of their performance (e.g., Reiss, 1967).
As we have seen in this paper, the state of knowledge on public perceptions and reactions to violence has improved substantially during the past two decades. Although this trend is certainly encouraging, a good deal of both descriptive and analytical work remains to be done.
Many of the most important unresolved questions concerning fear of victimization are questions of process, meaning that they pertain to the sequence, timing, or duration of events. For example, as a consequence of fear, individuals commonly take steps to reduce their risk, steps that may have the effect of reducing or eliminating the fear that initially provoked them. (Indeed, it may be that people commonly "manage" their fear by taking those steps necessary to reduce it to acceptable levels.) If the adoption of avoidance/precautionary measures is typically followed by a reduction in fear, then conventional cross-sectional measurements of fear may seriously misrepresent the nature of fear in a population. Although such measures may accurately capture instantaneous levels of fear, they overlook the history of fear among members of that population. There is a pressing need, then, for longitudinal data to establish the sequential relations between fear and precautionary/avoidance behaviors. In the meantime, cross-sectional measurements of the prevalence of fear might prove to be much more informative if they were to express the proportion of the population "ever afraid" (or, perhaps, afraid during the past six months) as well as the proportion currently afraid.
Longitudinal data would also permit answers to questions concerning the duration of fear or anxiety. For example, in the wake of a frightening event (e.g., a violent attack by a spouse), what is the natural rate (i.e., absent any intervention by the victim or others) at which fear diminishes or decays? How does this rate differ for events of different types (e.g., a rape versus a robbery) and for victims of different types (e.g., males versus females)?