available is always incomplete and that the predictions are always subject to error.

Standardized tests should be considered in thiscontext. They provide a quick, objective method of getting some kids of information about what a person learned, the skills he has developed, or the kinds of person he is. The information so obtained has, qualitatively, the same advantages and shortcomings as other kinds of information. (73) Whether to use tests, other kinds of information, or both in a particular situation depends, therefore, upon the evidence from experience concerning comparative validity and upon such factors as cost and availability.

(74)In general, the tests work most effectively when the qualities to be measured can be most precisely defined and least effectively when what is to be measured or predicted can not be well defined.Properly used, they provide a rapid means of getting comparable information about many people. Sometimes they identify students whose high potential has not been previously recognized, but there are many things they do not do. (75) For example, they do not compensate for gross social inequality, and thus do not tell how able an underprivileged youngster might have been had he grown up under more favorable circumstances.

1996年英译汉试题

 

The differences in relative growth of various areas of scientific research have several causes. (71)Some of these causes are completely reasonable results of social needs. Others are reasonable consequences of particular advances in science being to some extent self-accelerating. Some, however, are less reasonable processes of different growth in which preconceptions of the form scientific theory ought to take, by persons in authority, act to alter the growth pattern of different areas. This is a new problem probably not yet unavoidable; but it is a frightening trend. (72)This trend began during the Second World War, when several governments came to the conclusion that the specific demands that a government wants to make of its scientific establishment cannot generally be foreseen in detail. It can be predicted, however, that from time to time questions will arise which will require specific scientific answers. It is therefore generally valuable to treat the scientific establishment as a resource or machine to be kept in functional order. (73)This seems mostly effectively done by supporting a certain amount of research not related to immediate goals but of possible consequence in the future.

This kind of support, like all government support, requires decisions about the appropriate recipients of funds. Decisions based on utility as opposed to lack of utility are straightforward. But a decision among projects none of which has immediate utility is more difficult. The goal of the supporting agencies is the praisable one of supporting “good ” as opposed to “bad” science, but a valid determination is difficult to make. Generally, the idea of good science tends to become