Impact of Group Size and Project Duration on Capstone Design, The

Journal of Engineering Education, Jul 2004 by Griffin, Paul M, Griffin, Susan O, Llewellyn, Donna C

ABSTRACT

This paper discusses how group size and project duration impact capstone design in terms of learning objectives for the student, value to industry sponsors, and faculty resources. The analysis is based on survey results and an external faculty evaluation comparing a one-semester offering with a two-semester offering of capstone design in the School of Industrial and Systems Engineering at the Georgia Institute of Technology. In addition, we examine whether the stated learning objectives (technical writing skills, presentation skills, and technical analysis) for the course provide value to sponsors. Our findings suggest that the one-semester offering was preferred by both students and industry sponsors and required fewer resources. We also found that although students prefer smaller group sizes, sponsors do not have a definitive preference. Finally, we found that although technical analysis is significantly correlated with dollar value to the sponsor, technical writing skills and presentation skills were not.

Keywords: capstone design, group size, course duration

I. INTRODUCTION

Criterion 4 of the ABET engineering criteria requires a student to participate in a major design experience [I]. Capstone design course offerings are a common way engineering programs meet this criterion. For many programs, including those at the University of Michigan, Purdue University, Pennylvania State University, Northeastern University, and the University of Virginia, the capstone design course consists of student groups working on industry-based projects under the supervision of a faculty member. Christensen and Rundus [2] recently provided a survey of some of the industry-based capstone design courses. The goals of such a course were to enhance student's technical writing skills, ability to work in teams, and ability to solve "real world" engineering problems, and to benefit the industry sponsor.

There is a significant body of literature about capstone design courses for engineering education. The vast majority of this literature focuses on content, delivery mechanisms, and learning objectives. Todd et al. [3] surveyed capstone courses given in North America and Dutson et al. [4] provided an extensive review of the literature concerning the teaching of engineering design through capstone courses. Techniques for addressing common problems in engineering design projects are discussed in [S].

One key issue in a capstone design course is the determination of the most appropriate size of a project group. There have been several studies on the impact of class size on successful learning, particularly in primary and secondary education (e.g., see [6] for discussion). In addition, there has been a great deal of work on group size and dynamics from a sociological and behavioral perspective. There is little consensus, however, on the optimal group size. Several researchers [7-9] have argued that groups need to have sufficient numbers but not be too large. There are several justifications for this argument. First, as groups become large, there are problems of free-riding, social loafing, and conformity [10-12]; factors which Steiner [13] refers to as process losses. Steiner suggested that the actual performance of a group is determined by the potential performance minus these process losses. Some researchers have argued that there is an increasing (nonlinear) relationship between the size of a group and its ability to innovate. Very small groups, then, lack the diversity of views and skills needed for innovation [14], though this relationship is concave and hence as group size increases the returns to innovation diminish. Other studies have shown that smaller groups allow for better team dynamics since there is a deeper knowledge of the members and a better sense of responsibilities and objectives [15,16].

The empirical evidence about the impact of group size on project work is also unclear. Ziller [17] showed that as group size increases the quality of the output (measured in terms of errors) decreases with group size. There was, however, no real significant difference once groups exceeded five members. Yetton and Bottger [18] also found no significant performance increases in groups beyond five members. Thomas and Fink [19] and Schellenberg [20], however, showed that there was a slight tendency towards higher satisfaction in smaller groups. Stoneman and Dickinson [21] studied groups of size two to eight and found no significant difference in performance but did find evidence of free-riding and social loafing. Curral et al. [22] found that team size was negatively correlated with team processes such as participation, clarity of objectives, and emphasis on quality, though these results where less significant for the case of projects requiring a high amount of innovation. Cosse, Ashworth and Weisenberger [23], however, found that for an undergraduate marketing class, team size was strongly related to performance, though they did not examine groups with more than four members.