An opposite example is offered by the man who made an excellent score in a battery of tests for salesmen. Moreover, experience had taught her that she did not last at jobs very long, except for her present one, and so she was content to stay on even though the psychological test predicted poor ability and her production records verified the prediction. An interview with her revealed that she was under extreme financial pressure and simply could not afford to be out of work one day. Nevertheless, she was among the workers who had the longest tenure of employment. She was also rated as a poor worker by her supervisor. This girl actually had a poor production record, and since salary was determined on a piece-rate basis, her earnings were very low.

In this instance better test scores were correlated with better job performance. For example, a girl who worked in a certain watch factory had an abnormally poor score on a finger and tweezers dexterity test. Anyone who expects perfection is unaware of the real situation. It is no wonder that psychological tests in industry cannot do a perfect job at the present time. Furthermore, the claim cannot be made that even ability is perfectly measured. Most industrial testing programs limit the functions to be measured to certain specific abilities and do not consider the other contributing factors. Success or failure depends upon other things besides ability. On the other hand, for many reasons a person may be a failure on a job even though he has the ability to be successful. The relationship is nevertheless useful, for even such slight help in predicting success is better than if no tests were given, as was made clear in the earlier discussion of the selection ratio.Д¬ompelling reasons for continuing on a job can compensate for an employeeвЂ™s limited ability. Note that only three of the values in Table 4.1 exceed 0.40 in size, and that two variables are not tests but are personal history items: While the process of averaging used by Ghiselli definitely masks some very substantial validities, it should be sufficiently clear that testing is not a complete solution to the selection problem.ДЇrom the point of view of the statistician, such correlations are, of course, too low for individual prediction purposes. All validities within a particular category were then averaged to give the figures shown in Table 4.1. Ghiselli examined large numbers of validation studies and classified them in terms of the type of test being used and the type of criterion involved.

Table 4.1 shows some data collected by Ghiselli (1955) on the efficiency of testing for various types of industrial tests. Although better results are sometimes obtained, this is the exception rather than the rule. However, an improvement in job performances of 10 to 20 percent is often all that can be expected. Testing is no panacea, although testing in industry can have value if used correctly. However, the mere introduction of a valid testing program cannot act like a magic wand to wave away all difficult problems. The expected result of improving the selection of applicants is that training costs and labour turnover are decreased, production is increased, accidents and the probВability of accidents are reduced, and morale may be increased. Even when tests are used, it is advisable to conduct an interview, check the application form information, and use the various other techniques recommended by good personnel procedure. In many instances the desire to introduce a, testing program is fostered in the hope that a solution will be obtained to a вЂњno solutionвЂќ problem. Testing is often regarded as a fascinating subject by the novice.

Psychological testing, in industry should be approached with caution.

Do psychopaths pass phychological tests