Another difficulty encountered at this point is the fact that such direct measures of vocational success as have been utilized in these comparisons are in themselves subject to very large error. Only in recent years, and as a result of the emphasis of the human factor in industry, has it come to be the common practice to secure adequate records of the work of the individual as contrasted with the work of the gang. Even today such records are available in accurate form for only the simpler operations, in which standardized conditions of work can be maintained. The relative success of salesmen, for example, is not fairly measured in terms of the amounts of their sales, the number of prospects interviewed, or the frequency with which the assigned tasks are accomplished, unless the local trade conditions of the respective territories are fully taken into account. Inasmuch as such errors of measurement tend to reduce the apparent correlation between the traits measured, it is extremely probable that the psychological tests are even more significant than their present results indicate. Refinement of the tests must be accompanied by more accurate and precise measurement of the actual working efficiency of individuals in the industrial field, if the results of the one are ever to represent the amount of the other. In this as in many other respects, the development of vocational tests depends as much upon the active and intelligent cooperation of industrial concerns as it does upon the enthusiasm and diligence of the psychological investigators. From the point of view of the employer, the incompleteness of the correlation between tests and direct measures is of little concern. Even a very small positive correlation affords him a degree of guidance in the selection of his workers that was far from forthcoming under the haphazard methods of employment that have been traditional. But from the point of view of the individual who is seeking guidance, or who is accepted or rejected on the basis of his performance in psychological tests, any correlation which is imperfect may lead to occasional injustice and misdirection.