Multiple Baseline Testing: Experimental Method for Drawing Causal Inferences in Construction Engineering and Management Research

Identifying and evaluating solutions for critical industry challenges is a major theme of construction engineering and management (CEM) research. Researchers and practitioners in the construction sector often seek to invent, test, implement, and disseminate practical interventions that improve safety, productivity, quality, and other project success factors. Robust scientific research design is imperative for drawing valid causal inferences when testing the impact of new interventions. Unfortunately, the transient and dynamic nature of construction obfuscates the isolation of variables, thus making true experiments impractical or unethical. With roots in pharmaceutical research, multiple baseline testing (MBT), commonly known as multiple baseline design, is a promising and viable experimental technique that provides reinforced evidence for drawing causal inferences and analyzing change. The requisite process of replication through concurrent longitudinal studies, phased implementation, and inter-and intra-experimental unit comparisons allow researchers to isolate the impact of an intervention within constantly changing environments. This paper describes a protocol for the rigorous and valid implementation of MBT on construction projects based on the authors' collective experience with the method. This paper describes how the proper use of MBT minimizes threats to internal validity and ensures the ethical treatment of research participants in subject-based research. The paper concludes by reviewing the common challenges associated with the adoption of MBT in construction research. It is expected that the discussions will equip researchers with the tools to conduct rigorous, robust, and reliable intervention studies with improved evidence for causality. (C) 2015 American Society of Civil Engineers.