Enhanced training with special system software at the L. L. Thurstone Psychometric Laboratory

Abstract

There are many facets to training students in the use of on-line computers at the L. L. Thurstone Psychometric Laboratory. It is felt that competency in the use of computers is, for psychologists, a vital research skill, but not a curriculum in itself. However, course structure, faculty research, professional staff, and the facilities serve as a conducive framework for students to have many goal-directed encounters with computers. In order to promote many varied contacts with computers, not only must the hardware be competent, but also the system programs, e.g., operating systems and languages, must promote effective and efficient use of both computer and student time. The premise of this paper is that special system software is one of the most important assets to such a training program. An example of a typical student encounter with on-line computers can provide rationale for this premise. A student decides to embark on an investigation of sequential decision making in human Ss. He feels that he should use his department's minicomputer to administer the task since there is fairly complicated computation and record keeping between stages. Moreover, he already knows FORTRAN and there is a FORTRAN on the computer he plans to use. After several days of programming in FORTRAN, this student finds that his program can handle only a small subset of the conditions he had hoped to study. Integer arithmetic in his FORTRAN does not handle the range of numbers that he had planned to use, and floating-point arithmetic suffers from rounding errors. Some of the S's input was to have been characters rather than numbers, and FORTRAN was never designed to manipulate characters efficiently. Moreover, the FORTRAN input statements were restrictive in that they required that the 5 enter his responses in specific columns (at best, too much to expect from Ss), and that should as, even in error, enter an alphabetic character where a numeric character was expected, the run time system stopped the program and returned to the monitor. The student decided to accept these limitations and continued to add to the program to store data about each decision the 5 made. With this addition, the program became too large to fit into the small amount of core that the minicomputer had. The student had