5. Students design a third experiment based on knowledge gained previously and within the resources of their budget. The report on this design goes into Vol. 3. 6. Students review the "literature" generated in their class and write a review paper. The paper not only discusses their own research, but that conducted by other classmates. 7. An optional exercise is to have students write a report to their sponsor justifying their own research and the expenditure for that information. This may be reserved for students who have overspent their "budgets" or who have conducted a series of disjointed or trivial experiments which the instructor believes needs to be rationalized. *** Building New Models Both MESS and LESS have been written so that new models can be added to the library by writing FORTRAN subroutines to operate with the supervisor or management program. In other words, a professor or advanced student can write his own model as a FORTRAN subroutine for the supervisor program. He need not concern himself with the programming required to interface the student with the data-generating model. For instructors or students who do not program or do not wish to commit themselves to a programming task, it is possible to develop and build new models for the library making use of the Simulation Writer Interactive Program (SWIP) written by Robert Stout and developed at the University of Michigan. For smaller computers, LESS has features that would enable the building of certain kinds of models. Both are possible in the interactive mode. The development of SWIP and the LESS programs makes it possible to add models to the library that are concerned with physics, chemistry, biology, medical sciences, and engineering as well as psychology, sociology, economics, education, and political science. Such models can be used in methodology and statistics courses as well as specific content courses where it is desirable to teach students by way of a method of inquiry. **** Costs The cost of using MESS and LESS once they are implemented is very small compared to the cost of maintaining a traditional laboratory course. The programs are designed for batch processing (requiring the use of a keypunch or equivalent) as well as interactive processing (requiring the use of a teletype or other user terminal). At the University of Michigan both methods are used simultaneously in the psychology course. (A sample batch run follows this article.) The parameters of an experiment can be conveyed to the computer on about a dozen cards, many of which can be used over and over again. In the interactive mode the student is prompted for the input required, and error messages are immediately available. The cost of batch processing student experiments is usually many times cheaper than the interactive mode. Its disadvantages are a longer turn-around time and delayed access to error messages. Although the interactive mode is more expensive, the student obtains "results" from his experiment more quickly. Even with this more expensive mode, a budget of $7.00 to $10.00 per student has been sufficient for a given term. Cost figures on SWIP will be available soon. The programs can be obtained for a nominal fee. For information on MESS and SWIP write: Dr. Dana B. Main, 3435 Mason Hall, Department of Psychology, University of Michigan, Ann Arbor, Michigan 48104. For information on LESS write: Dr. Arthur Cromer, Computing Center, University of Louisville, Louisville, Kentucky 40208. **** Acknowledgement EXPER SIM was developed in conjunction with the center for Research on Learning and Teaching at the University of Michigan with support from the Exxon Education Foundation. [Ed. note: the IMPACT Program of the Exxon Education Foundation has selected EXPER SIM as a tried and effective innovation for which adoptions are encouraged. IMPACT provides modest grants to potential adopters of EXPER SIM (and other items on their adoption "menu") on the basis of proposals received. Contact Ms Caryn Korshin at the Exxon Education Foundation for further details.]