Modelling the Emotional State of a Group

Predictive Analytics

Predictive Analytics
The emotional state of a group is a significant determinant of the workability and
performance of any group. The emotional state informs on the cohesiveness and probability of
success of groups. This therefore makes it important to understand and predict the emotional
state of a group. Several research works have been done in the area of modelling the emotional
state of a group with the aim of understanding and predicting emotional state.
Gerben (2009) considers the emotional state of a group in a wider sense that goes beyond
the business perspective. The research investigates the effect that the emotional expressions of an
individual can have on the behavior of observers. Gerben (2009) is relevant to this study since it
is interested in determining whether the emotional expressions of an individual can be used as a
predictor of the behavior and by extension the emotional state of those around them. The
research introduces the application of the EASI (Emotion As Social Information) model for
prediction of the emotional state of groups. This research is unique because it considers groups
from points of view such as; leadership, negotiation and conflict, which is a strength because it
broadens the applicability of the model and findings. This unique approach is however a
weakness as well, since the performance of the model considerably lower compared to models
that are area-specific and consider all the factors within that area of interest.
The approach by Wang & Huang (2009) focuses of developing a three-variable model
with the factors being Emotional Intelligence, Transformational Leadership and Group
Cohesiveness. The research evaluates the possibility of determining one of the three factors by
using the other factors. This research is relevant since it touches on two aspects of interest to this
study; group cohesiveness and emotional intelligence. It examines how transformational
leadership and emotional intelligence relate to group cohesiveness. The research is unique in its
consideration of transformational leadership and considers the team leader’s emotional
2

Predictive Analytics
intelligence and their engagement in transformational leadership as predictors of the team’s
(group’s) emotional state and cohesiveness. The research has a weakness since it greatly limits
the variables in the study to just three variables. This research specifically considers the business
context for the research hence making the resultant model more specific.
Stéphane (2014) examines emotional intelligence in organizations in terms of work
criteria, performance and leadership. Emotional intelligence refers to abilities relating to
emotions of an individual as well as the emotional information (Stéphane, 2014). The research
focuses on the role that emotional intelligence plays in organizations and the general
performance of the organization. It is relevant to this study since among the roles of emotional
intelligence focused on is how it determines the emotional state of an organization. The unique
character in this research is the focus of emotional intelligence across the board without giving
preference to the team (group) leaders, which strengthens the research by giving a wider picture.
The weakness of this research is in the measurement of emotional intelligence (EI), where two
approaches are given and used instead of considering the best approach.
The agent-based modeling approach is applied by Bosse, Duell, Memon, Treur & Van
der Wal (2015) in the prediction of the emotional state of groups. The research develops the
agent-based modelling through the simulation of the contagion processes of the emotions within
teams (groups). This research is relevant to the study since the agent-based model developed can
be used as referencing for the predictive model for emotional state of groups. The agent-based
approach in this research represents a unique approach in this research area. The simulation, as
opposed to actual collection of data, is a weakness for this research because simulated data is
often not random enough to represent real world situations. The research is however
3