The analytical modelling of collective capability of human networks

Abstract

This thesis is an attempt to propose an analytical model for estimating and predicting capability in human networks (i.e. work teams). Capability in this context is the ability to utilise the collective inherent and acquired resources of individuals to complete a given task. The motivation of proposing a method for measuring collective capability of teams is to assist project managers and team builders to allocate and assign “The most capable teams” to a project to maximise the likelihood of success. The review of literature in engineering, human sciences and economics has led to a definition of capability. One of the key findings of this research work is that collective capability can be predicted by: 1. Demographic homophily of members of the team, 2. The diversity of skills that each member brings to the team, 3. The past experience or attainments of the members, and 4. The strength of relationship amongst the members of the team. The influence of the four predictors of capability is investigated through the design of empirical surveys conducted among postgraduate students over a period of 2 years. The data collected from the surveys are used to assess the correlation between the predictors and the dependent variable using standard statistical methods. The conclusions of the study confirm that there are positive and significant relationships between the independent predictors and collective capability of project teams. The demographic homophily of the individuals in team and their instrumental (task related) relationships’ strength become the two most effective predictors which have the highest effect on the collective capability of a team as a whole. The skills diversity of the members in a group and their previous level of attainments/experiences in similar projects were also proved to be effective factors (with lower level of effect) in increasing the capability of the whole team in fulfilling the requirements of a pre-defined project.