Abstract
Understanding how teams perform has been the goal of organisational psychologists for the
past century, to date there is no effective model or approach that incorporates both internal
and external factors. Since the late 19th century psychologists have studied individuals in the
workplace to measure their abilities and understand how these relate to performance. During
this time a detailed understanding of the individual has emerged, but not of teams. Part of the
reason for this is the type of teams studied and the approach of the research.
The initial objective of this research was to extend the current thinking of the role of
individuals within teams with the aid of a number of case studies from industry. During the
course of this approach two key themes emerged; team performance was not influenced by
the individual personalities and a one size fits all model would not work for the range of teams
involved in the study. This study focussed solely on external factors influencing teams and did
not explicitly consider internal factors.
Based on interviews and questionnaires with staff in industry, as well as observations from
within the workplace, the importance of organisational variables was identified for predicting
team performance. The data collected from using qualitative interviews with a range of staff
in each organisation was summarised using context analysis into a number of ‘organisational
variables’. These variables were aspects of the organisation that were unique to the particular
workplace. They may occur as a result of the industry, culture, climate, leadership attributes
or recruitment strategy. For both organisations involved, 20 organisational factors were
summarised into an extended version of the McKinsey 7-S framework, informally named the
“UNSW McKinsey 8-S framework”. From this framework, a detailed organisation specific
questionnaire was developed.
Traditional statistical methods for determining relationships among variables (Structural
Equation Modelling) cannot accurately assess complex models or work with small sample
sizes. The models for the organisations contained over 200 variables, each with less than
20 participants. Through the use of Partial Least Squares Path Modelling (PLS PM) and
Redundancy Analysis (PLS-PM RA) the questionnaire methodology was validated. In
addition to the use of PLS PM for the purpose of the research, a standardised approach to
reporting PLS-PM RA is included when there is a not a base model available. This thesis
presents an entirely new methodology for analysing and predicting team performance,
combining a contingent but templatised organisational modelling approach with PLS analysis
for estimating the variables and the path loading. The approach and method was validated
with a second organisation. The value of this work is as an addition to existing work, and does
not necessarily replace the existing body of research.
The research provides two distinct applications to industry. Firstly, based on the path
loading, the weights of the variables included in the PLS RA models can be used as levers for
improving team performance. In addition to the PLS-PM RA model identifying the levers,
the relative proportions can also be determined. The application of PLS PM to larger more
complex team models is also discussed. The second application to industry is for any domain
where there is expertise and there is a constraint (time, money etc). The same methodology to
build PLS -PM RA models for teams can be applied and the results used to determine where
funding or resources should be allocated (and in what proportion).
If a truly generic team model of predicting performance exists, the model would contain
both internal and external influences. The research provides a holistic and replicable approach
for determining the external influences and their relative strength. Future research in the field
should aim to integrate the two approaches.