Using latent semantic analysis (LSA) to study expert design teams' verbal communication
Date of Issue2009
School of Mechanical and Aerospace Engineering
Centre for Human Factors and Ergonomics
Teams of people design products. They bring their own language, experiences, perspectives, and detailed knowledge to the design. Thus, design in a team can be regarded as a social process of reaching a “shared understanding”, through activities such as information exchange, compromise and negotiation. Latent Semantic Analysis (LSA) is a computational linguistic technique. It is well known for its ability of modeling human knowledge representations and capturing the latent relations between words and passages. LSA was applied in design team research. Previous research showed that by mining design teams’ formal documentation, LSA can successfully capture different designers’ semantics. Quantification metrics used to measure semantic coherence of design documentation could predict successful design outcomes. This research focuses on the verbal communication in design teams and aims to develop assessment tool(s) that can be used to enhance the communication in design teams. This study expands the previous LSA-based design team communication framework by decoupling the design process from the design outcome. Moreover, a new quantification metric is proposed to complement existing metrics. Data from two experimental studies were used to examine our postulations. The results showed that LSA is able to capture the semantics of different design team members’ verbal communication. The semantic coherence of a design team is more correlated to design process quality than design outcome quality.