A hypothesis is a testable explanation for an observable phenomenon or scientific theory.
Published in Chapter:
The Use of Simulation as an Experimental Methodology for DMSS Research
Giusseppi Forgionne (University of Maryland, Baltimore County, USA) and Stephen Russell (University of Maryland, Baltimore County, USA)
Copyright: © 2008
|Pages: 11
DOI: 10.4018/978-1-59904-843-7.ch106
Abstract
Decision-making support-system (DMSS) research has been based largely on case and field studies of real organizations or entities and laboratory experiments involving human participants. Each methodology has provided useful results. Case and field studies can identify and describe relevant variables and relationships, suggest theories, and offer conceptual and practical models for evaluation (Benbasat, Goldstein, & Mead, 1987; Fjermestad & Hiltz, 2000; Orlikowski & Baroudi, 1991). Participant-based laboratory experiments can provide initial theory and model hypothesis tests, identify key evaluation assumptions, and offer suggestions for future research (Amaratunga, Baldry, Sarshar, & Newton, 2002; Das, 1983; Patton, 2002). Case and field studies and participant-based laboratory experiments, however, have limitations for decision- making support evaluation. In each setting, it is difficult to acquire and motivate the participants. Real organizations and entities are reluctant to disrupt operations or provide proprietary information. Laboratory participants often consist of students in universitybased studies, and as such frequently have little actual knowledge of the decision situation or incentive to mimic real behavior. Even decision-knowledgeable participants may, for political, psychological, or other reasons, mask real behavior. In addition, case studies, field trials, and laboratory settings may not be representative of the population. By their nature, case and field studies usually involve few organizations or entities. Given the diversity experienced in practice, the few organizations or entities, even if selected judiciously, are unlikely to capture the disparity. While great care may be exercised in selecting sample participants for laboratory experiments, there will be Type I and Type II errors that result from even well-designed experiments (Alavi & Joachimsthaler, 1992).