In carrying out such complex investiga tions some mistakes are made, such as those due to the tempta tion to rely on opinions obtained by questioning workers and their employers. This "questionnaire method" has received severe con demnation from scientists and should be used sparingly by inves tigators in the field of job analysis. In the first place, it yields chiefly opinions showing what some one thinks the work is like. Secondly, it usually states the components of the job in general, abstract words such as "accuracy," "quickness" and the like, which do not characterize any occupation in particular, being required in hundreds of different jobs.
Too great dependence on psychological tests is another error. There is a procedure whereby a number of psychological tests are chosen which seem to resemble the work being analyzed. Those tests in which the good workers excel and the poor workers fall short are considered to call forth the same mental activities as those required in the job and to constitute therefore the analyzed components of the work. As a matter of fact, such an assumption is not valid. The activities called forth by the tests are merely test-activities and nothing more.
In formulating a procedure which prom ises scientifically valid results, and will be applicable regardless of the ends for which an analysis may be made, the underlying re quirement is that the procedure must conform to the rigid tech nique of scientific method. It must be unbiased and dominated only by the desire for the exact determination of conditions. Thorough analysis can be carried on only by trained scientists, though they must have the co-operation and assistance of experts in the occupational fields as well.
In conducting an analytic investigation the first task is to sur vey the entire vocational field in which the activity lies, to deter mine its relation to the social organization, then to divide it into its occupational sections. These are to be described in terms of their relations with each other and their relation to the whole field.
Then each occupational section is described and divided into smaller units. This division may be made by mere observation, and the operations may at first be described in qualitative terms, but the descriptions should also be couched in quantitative terms.
Measurement is an indispensable part of the process. The early quantitative analyses of F. W. Taylor were made with a stop-watch reading in fifths of a second. As the principle of measurement has become more widely extended, how ever, there has come a demand for measurements of greater pre cision, made with instruments used in the psychological laboratory for the measurement of intervals of time as small as one-thou sandth of a second. For example, pictures have been made of the eye-movements made by proof-readers. The chronocinemato graphic method of F. B. and L. M. Gilbreth offers much promise for the exact measurement of motions and time intervals involved in work operations. Motion pictures of the worker at his work are taken, in which is placed a clock that measures time in thousandths of a second. The background in each picture is cross-sectioned into squares of predetermined size. By examining the exact position of the worker in each succeeding picture in relation to these squares and to the time-recorder, it is possible to determine how much he has moved and at what rate.
For developments in Great Britain see the article on FATIGUE IN INDUSTRY.
Some indication of the fundamental part which job analysis can play will be furnished by this list of ends which have motivated the various analyses that have been made : elimination of waste motions; determining standard day's tasks; wage setting; figuring accurate costs ; arranging grades and routes of promotion ; invent ing trade tests; establishing a curriculum for vocational education ; determining methods of teaching vocational activities; securing a basis for choice of a vocation; securing job specifications for the selection of workers. See also INDUSTRIAL PSYCHOLOGY ; INTELLI GENCE TESTS.