.
InThe efficacy of computer-assisted instruction (CAl) for increasing test performance in an experimental design course was tested using two commercially available programs to supplement course material. CAl was used to supplement course content for one of three examinations. Comparisons were also made with performance on the exams from 2 previous years. The results indicate a significant enhancement of performance by CAL A number of logical, insightful articles have been written on how computers could, theoretically, be effectively used to teach advanced courses in psychology (e.g., Butler, 1988;Castellan, 1986Castellan, , 1988Eamon, 1986;Peterson, 1989). However, only limited research on the efficacy of computer-assisted instruction (CAl) has been published, and the research which has been conducted has yielded mixed results. Some (Belmore, 1983; Goolkasian, 1989;Lowman, 1990;Ware & Chastain, 1989) have found no effect of CAl on performance criteria when compared with that obtained with other methods of teaching. Other investigations have shown an enhanced effect of CAl on performance (Petty & Rosen, 1990Rosen, , 1991, on attitudes toward the course supported by CAl (Goolkasian, 1989;Ware & Chastain, 1989), and on attitudes toward computers (Spivey, 1983). Welsh and Null (1991) found that traditional methods of teaching a laboratory component in a research design course were superior to computerbased methods of teaching.Although a meta analysis on the effectiveness of CAl at the college level found moderate effects (Kulik & Kulik, 1987), the statement by Eamon (1986) that "there is precious little research showing that computer assisted instruction in psychology laboratory courses makes any difference at all," constitutes a valid summary of the present situation.Given the expense and time required to implement CAl, it is imperative that more applied research be conducted to determine its efficacy. The research presented here describes the use and effectiveness of CAl in a laboratory course in research design.
METHOD SubjectsThe subjects were 139 junior and senior psychology majors enrolled in an advanced course in research design.This research was partially supported by Grant P120A80053 from the Minority Science Improvement Program of the Department of Education. The author wishes to thank Perry Duncan for his editorial comments and Benton Gardner for assistance in data entry. Requests for reprints should be sent to Nancy C. Duncan, Department of Psychology, Hampton University, Hampton, VA 23668.
Software and ApparatusTwo software packages written by Fazio andBackler (1986, 1987) and available from CONDUIT, Power and Main Effects and Interactions, were used to teach concepts of research design. Power, a simulation based on an actual experiment, provides for the manipulation of six design features (increasing sample size, matching, use of aggregate measures, homogeneity of the sample, including an important source of variation as an independent variable and use of a pretest) to demonstrate how the power of an...