An Empirical Study about the Influence of Project Manager
Personality in Software Project Effort
Daniel Tadeu Martínez C. Branco
1,2
, Edson Cesar Cunha de Oliveira
1
, Leandro Galvao
1
,
Rafael Prikladnicki
3
and Tayana Conte
1
1
USES Research Group, Instituto de Computacao, Universidade Federal do Amazonas, Manaus, Brazil
2
PRODAM, Processamento de Dados Amazonas S.A., Manaus, Brazil
3
Faculdade de Informatica (FACIN), PUCRS, Porto Alegre, Brazil
Keywords: Software Project Manager, Software Project Management, Software Development, MBTI Personality Test,
Belbin Team Role, Human Factors.
Abstract: Project effort is a main concern on software organizations. The project budget is derived from project effort
which in turn is based on the software engineers’ effort cost. Project manager is responsible for planning
and controlling this effort estimation. Some researches relate how project manager can influence the project
success, specially when considering project manager personality. This research aims to evaluate the
influence between project manager personality and teamwork behavior over project’s effort deviation. A
case study was performed with 65 real projects collected from a software company dedicated to develop
software projects for its local government. Unlike previous researches our results show no statistically
significant influence of project manager personality, assesed by MBTI test, over project’s effort deviation.
However, our results show the project manager teamwork behavior, assesed by Belbin’s BTRSPI, has a
statistically significant influence on the project’s effort deviation.
1 INTRODUCTION
Over the last few decades only under half of
software projects end up in success, i.e., they are
delivered on time, on budget and with all required
features. CHAOS Manifesto (Standish Group
International, 2013), which has a database of 80.000
real software and IT projects since 1997, shows that
in 2012, 39% of projects achieved success; 18% of
projects are either cancelled prior to completion or
delivered and never used; and 43% of projects are
delivered late, over budget and/or with less required
features. For the last case, problems with budget
overrun increased from 46% in 2010 to 59% in
2012.
A software project budget has as its main
element the amount of software engineers involved
(PMI, 2013). After project scope is set out with
client, the estimated delivery date of a software
project is dimensioned according to the effort of
every engineer in the project. Project budget is
estimated from engineers’ effort cost. Project
manager is responsible for planning and controlling
these estimations, and for selecting engineers from
company that will compose the project’s software
team.
Acuña and Juristo (2004) performed a study
about the influence of assigning people to
predetermined functional roles in software
development process. Personality was assessed by
the projective 16PF personality test (Russell and
Karol, 1994) that measures 16 primary personality
traits identified by Cattell et al. (1993). Acuña and
Juristo (2004) study’s premise was that personal
behavioral competencies and characteristics of
professional conduct has influence in the
effectiveness and efficiency of software
development. The result showed that assignin people
to functional roles according to their capabilities,
based on behavior and characteristics, and habilities
demanded by software project role can help
organization to develop systematic long-term
competences.
Cruz et al. (2011) performed a systematic
literature review about the influence of individual
personalities on individual tasks and team work in
the software engineering. Data were extracted from
102
Martínez C. Branco D., Cunha de Oliveira E., Galvão L., Prikladnicki R. and Conte T..
An Empirical Study about the Influence of Project Manager Personality in Software Project Effort.
DOI: 10.5220/0005373001020113
In Proceedings of the 17th International Conference on Enterprise Information Systems (ICEIS-2015), pages 102-113
ISBN: 978-989-758-097-0
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)
42 studies published between 1970 and 2010.
Results showed that pair programming and team
building were the most recurrent topics surveyed and
that MBTI (Myers-Briggs Type Indicator)
personality test was the most used.
Karn and Cowling (2006) carried out a study
about the effects of personality on performance of
software engineering teams. In their conclusions, the
authors reinforced in their results the notion that the
psychological type is an important factor to consider
in software development teams’ performance.
Wang and Li (2009) studied the effects of
software project manager personality in 116 real
software development projects in Shanghai’s
software companies. In this work, personality was
assessed by Neo-Five-Factor Inventory (Costa and
McCrae, 1992) for the Five Factor Personality
Model (FFM or Big Five) (Goldberg, 1992). Their
results showed that software project manager’s
personality is one of the indicators for software
project success and it also influences in success
through mediating effects of the leadership.
Previous works emphasized the importance of
studying personality in software engineering and
how it affects software project success. This paper
attempts to answer if software project manager
personality influences on software project effort.
Moreover, we also study if software project manager
teamwork behavior influences software project
effort and their software development performance,
according to the work developed by Belbin (2010a).
This study was performed based on a dataset of 65
real projects of a government software development
company. Thus, project result based on effort
deviation was assessed with project manager’s
MBTI psychological test and Belbin’s team role.
This analysis is important to verify if the difference
between the planned and actual effort can be related
by project manager behaviors.
This paper is organized as follows: Section 2
presents concepts of personality types and their
relation with software engineering. Section 3
describes research method used in this study and
data sample characterization. In Section 4, study
results and evidence obtained are presented and
analyzed. Section 5 describes threats to validity
related to this research. Finally, in Section 6
conclusions and opportunities for future work are
presented.
2 HUMAN FACTORS IN
SOFTWARE ENGINEERING
Researches relating human factors with software
engineering have been increasing over years,
according to Cruz et al. (2011). Some of these
researches aim to identify the appropriate
psychological types for specific activities in
software development (Acuña and Juristo, 2004).
França and Silva (2007), Wang and Li (2009),
Brewer (2005) and Acuña and Juristo (2004)
researches are based on different psychological
theories to study human behavior and characteristics
in software engineering. This work will focus on two
specific psychological theories: theory of personality
based on Myer Briggs Type Indicator (MBTI, 2014),
and theory of team roles based on Belbin’s work
(Belbin, 2010a). In next sections, we will present
concepts related to each one of these theories and
how they have been studied within software
engineering context.
2.1 Personality Theory
Each person has his own behavior when dealing
with same stimuli during his social life, such as
impulsive people, people with difficulties in
communication or people who make emotional
decisions. These observed behaviors may only be a
situational reaction, but Morin and Aubé (2009)
state that it can also be derived from his personality.
Bloch (2002) et al. apud Morin and Aubé (2009)
conceptualized personality as a set of relatively
stable affective, emotional and dynamic
characteristics; these characteristics are the person’s
usual way of being.
Personality systematic study has been done
since late 1930s through surveys of Henry Murray
and Gordon Allport reseaches at Harvard University
and also by contributions of the Swiss psychiatrist
Carl Jung (Schultz and Schultz, 2002 and Morin and
Aubé, 2009).
Jung in his studies describes that much of
perception and reaction to environment is
determined by opposing mental attitudes of
extraversion and introversion. These attitudes relate
primarily to how energies are directed in certain
situations. Extroversion behaviors are characterized
by an orientation to outside world and people, while
introversion is characterized by a preference to its
own ideas and feelings (Schultz and Schultz, 2002).
Realizing the differences between extrovert and
introvert people, Jung proposed a separation of these
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
103
behaviors in psychological functions (Morin and
Aubé, 2009).
Psychological functions refer to the way the
individual’s energy is organized and structured.
These functions are divided by Jung in sensation,
intuition, thinking and feeling.
Sensation and intuition were classified as non-
rational functions because their actions are based on
fact interpretation and world perception. The
sensation function feels “what is happening” and
reacts through this external stimuli and previous
experiences. Intuition function emphasizes
inspiration and world perception, and their possible
impacts into the future (Morin and Aubé, 2009).
Jung classified thinking and feeling functions as
rational because of how they are related to
evaluation of previous experiences. The thinking
function assigns to the individual the judgment
based on principles, rules, or laws, that are
considered relevant to logic and situational analysis.
In opposite, the feeling function allows the
individual to evaluate situations based on the
involved personal values and consequences, such as
embarrassment or future conflicts.
Although the functions described by Jung follow
a pattern of duality, this does not mean that these
functions are mutual exclusive. Jung uses the
concept of dominant function, which exists in the
conscious and rules the majority of human behavior,
and secondary function, which has influence on the
personal unconscious and can affect behavior in
certain occasions (Morin and Aubé, 2009).
2.2 Myers-Briggs Type Indicator
(MBTI)
Katharine Briggs and Isabel Briggs Myers continued
in the 30s the work begun by Jung. They introduced
a new psychological function in addition to those
previously described. Another important
contribution of their work was the creation of an
assessment test of those functions. The result of this
assessment indicates the individual's psychological
type. These kinds of tests are also known as
inventories in psychology (Morin and Aubé, 2009).
The Myers-Briggs Type Indicator (MBTI)
integrates Jung’s theory with Myers-Briggs’s
concept of "lifestyle" . Currently, MBTI is used by
research community as well as in industry, where it
has been already used by 89 organizations that
compose the Fortune 100 (Ash, 2013).
MBTI test result indicates one of 16 possible
psychological types. Each possible type has a
structure that combines the four different
psychological functions: one attitude function
(introversion or extroversion), one rational or
judgment function (thinking or feeling), one non-
rational or worldview perception function (sensation
or intuition) and one lifestyle function, which is the
new dimension described by Myers-Briggs. This
new dimension aims to assess in the lifestyle of
individual’s behavior the dominant dimension
between judgment and perception.
The psychological type, indicated by an MBTI
test, is represented by a four-letter combination,
where each letter describes one psychological
function. As an example, consider a person with
psychological type INTP. This acronym can say that
this person has an attitude of introversion (I), and
uses intuition (N) as non-rational function, has
thinking as rational function (T) and a lifestyle based
on perception (P). Morin and Aubé (2009) describe
all the 16 possible psychological types, also know as
profiles, and their noticeable characteristics.
2.3 Belbin’s Team Role Theory
In the late 60s, Meredith Belbin began studies of
people’s behavior at teamwork when developing
business games. During the execution of these
business games, Belbin observed that the results
could be influenced by individual’s behavior, i.e.,
by the relationship with other team members and the
way they contributed to the team (Belbin, 2010a).
Belbin defined team role as the person
characteristic behavior in the context of teamwork
(Belbin, 2010b). Belbin’s team role theory states that
in teamwork individuals tend to play different type
of roles beyond the usual functional roles associated
with his technical activities (Fernandes and Silva,
2007).
Belbin’s team role identification is accomplished
through an inventory called Belbin’s Team Role
Self-Perception Inventory (BTRSPI). The BTRSPI
design aims to assess the behavioral characteristics
of individuals when working in a team (Belbin
Associates, 2013).
BTRSPI is a form composed by a situational-
based questions and alternative answers. Individuals
when participating must distribute a total of ten
points over the alternatives answers that are closer of
his own behavior. To assess the test, the total points
for each Belbin’s team role is calculated. From the
distribution of percentile scores, the BTRSPI
proposes to each team role one of four possible
values: "Very High", "High", "Average" and "Low"
(Riding and Rayner, 2001). The values "Very High"
and "High" indicates that the individual has the
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
104
descriptive characteristics of that Belbin’s team role,
which describes their strengths and weaknesses in
teamwork.
Belbin’s team roles has the following
classification (Belbin, 2010b):
Completer Finisher (CF) is meticulous, has
self-control and self-discipline, look for errors
and omissions in his own work and try to finish
it without any delays.
Implementer (IMP) is disciplined, trustable and
know how to turns ideas into pratical,
pragmatic and efficient actions. He tends to be
inflexible and slow to respond to new
possibilities.
Teamworker (TW) is cooperative, friendly,
diplomatic, non-competitive, compliant and has
good perception.
Monitor evaluator (ME) is sober, strategic, and
is able to visualize options and judge them
accurately.
Co-ordinator (CO) is dominant, confident
explainer of goals to others and delegates well
the tasks. He is also considered the leading role
for teams.
Plant (PL) is creative, unorthodox, confident
and have radical thinking. He has difficulties in
expressing himself because he is too focused in
his thoughts.
Shaper (SH) is challenging, dynamic,
impatient, self-confident, thrives under
pressure and has courage to overcome
obstacles. He may offend people's feelings
because he is susceptible to provocations. He is
also considered a leadership role with an
aggressive style.
Resource Investigator (RI) is extroverted,
communicative, flexible and like to explore
opportunities. He tends to be optimistic and to
quickly lose interest of everything after an
initial enthusiasm.
Specialist (SP) is not interested in any other
team member. He is focused, self-motivated,
dedicated and likes to discuss only technical
issues.
In the latest BTRSPI revision some roles were
renamed and a new one was included, the Specialist
(SP) role, as was described above. This study used
an BTRSPI version, where this new role was not
present and therefore it was not evaluated in this
research.
Belbin’s team role theory, beyond describe the
team roles, also proposed grouping them according
to common characteristics (Belbin, 2010b):
leadership (Co-ordinator, and Shaper), creativity
(Plant and Resource Investigator), execution
(Teamworker and Implementer) and balance
(Monitor Evaluator and Completer Finisher).
2.4 Personality Studies in Software
Engineering
The influence of personality in software engineering
is a recurrent topic covered by several studies. We
highlight the studies of Gorla and Lam (2004),
Brewer (2005), França and Silva (2007) and the
systematic mapping study performed by Cruz et al.
(2011).
Gorla and Lam (2004) studied the effects of
project leader’s personality in relation to the project
outcome. Project result’s metrics were based on six
project’s effectiveness measures: staff workload,
quality of work, efficiency of team operations,
effectiveness of user interaction, frequency of
schedule adherence and frequency of budget
adherence. These metrics were collected through
questionnaire-based surveys answered by team
members, based on their perception, of 20 software
development teams in Hong Kong. This study
performed the application of survey and the MBTI
personality test to those software teams. Their
results evidenced that project leaders with NF
(intuition-feeling) dimensions obtained better project
outcomes than the others. Furthermore, it was also
observed that functional role of systems analyst is
influenced by judgment (thinking-feeling)
dimension, while energy (introversion-extroversion)
dimension is crucial for developers.
Brewer (2005) analyzes whether a person can be
trained for the project manager role or if he already
has the skills necessary and how these skills can be
improved. Beyond knowledge they also studied the
impact of project manager’s behavior. MBTI
psychological test is one inventory considered for
this work to assess project manager personality.
Their results suggest the ENTJ and ESTJ
psychological types as ideal for the role of project
management. Another psychological test is refered
at this paper, named FIRO-B (Waterman and
Rogers, 2004), but no suggestion of this
psychological assessment was presented.
França and Silva (2007) studied the influence of
RUP functional roles and Belbin’s team role on
software factories. This research evaluates system
analyst, software architect, project manager and
implementer roles characteristics and the expected
psychological behavior of each one. The
characteristics of project manager’s functional role
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
105
considered in this work were decision,
communication, leadership, time management and
goal orientation. As result, França and Silva (2007)
suggest that project manager must be Co-ordinator
(CO) but Resource Investigator (RI) and
Teamworker (TW) roles could also have positive
results.
Considering the importance of this topic and
diversity of papers published in the last 50 years,
Cruz et al. (2011) carried out a systematic mapping
study, which we can highlight some of its main
findings:
Most studies carried out experiments;
Type of participant were well balanced
between students and professionals;
The main topics found were pair programming,
effectiveness of development teams and
individual performance;
The top personality tests used were MBTI and
Big Five (Goldberg, 1992);
In this systematic mapping, the effectiveness of
the software project manager related to his
psychological type was a topic within the lowest
quantity of studies. This finding indicates the need
for further research that relate human factors of the
project manager and the results achieved by its
software project.
3 CASE STUDY
This present study was based on the following
research question: does the personality of the project
manager can influence the outcome of a software
project, with respect to the project’s effort
deviation? Such research question has two main
goals: to evaluate the influence of the MBTI
psychological type and also evaluate Belbin’s team
role of the software project manager in relation to
the planned project effort. In order to evaluate these
goals we propose the following null hypothesis H01
and H02, contrasted by the alternative hypothesis
HA1 and HA2:
H01 = MBTI psychological types don’t affect
the planned project effort.
HA1 = MBTI psychological types affect the
planned project effort.
H02 = Belbin’s team roles don’t affect the
planned project effort.
HA2 = Belbin’s team roles affect the planned
project effort.
These hypothesis were analyzed considering a
dependent variable (project’s effort deviation) and
independent variables (MBTI psychological type
and Belbin’s team role). The instuments used to
collect independent variables were psychological
type, assessed by an online MBTI based personality
test, and Belbin’s team role, assessed by an based
Belbin’s Team Role Self-Perception Inventory
(BTRSPI) assessment.
We performed an in vivo study in the context of
a real software development company, in order to
assess the statistical significance of the relationship
between these variables.
3.1 Research Procedures
This case study was performed with projects
collected from the project management system of a
software development company dedicated to
develop software projetcs for its local government.
In this system the company records project data such
as schedules, estimated dates and estimated hours
needed to perform each task, hours consumed by the
task and the employees involved, and the assigned
functional role of each employee. The effort
consumed for a task is the result of daily record
made by each software team member of how much
time was spent and the description of the activities
performed. All these daily records represent, at the
end of the project, the actual project effort.
All team members who participated in the
software projects that we studied participated in a
presentation about human factors in software
projects and its uses in software engineering. This
was done aiming to contextualize them into the
concepts of the this present study. In this
presentation the MBTI psychological test and the
Belbin’s team role theory were presented. Later the
participants were instructed to identify their
personality types. This identification of
psychological type was done according to a based on
MBTI test, using the form available on the site
http://www.inpiira.org, and Belbin’s team roles,
based on the Belbin self-perception inventory
(BTRSPI). Participants were instructed to send by e-
mail the results of their assessments.
After collecting all data, we calculated the effort
deviation of each project. This calculation is the
difference between the total hours actually
consumed and the total hours estimated for the entire
project. As projects differ on size, we proceed with
the normalization of the scale using the ratio
between that difference calculated and estimated
hours of the project. The formula used is presented
below:
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
106
.
. .
.
After each project’s effort deviation calculated,
we grouped all the projects according to the MBTI
dimensions of the project manager. We also grouped
according to the Belbin’s team role. For each group
obtained we calculated the mean and variance of the
project’s effort deviations.
We used Kruskal-Wallis statistical test (Siegel
and Castellan, 1988) in order to compare the
project’s effort deviation between the groups
obtained, first for groups acordding to the MBTI
psychological test and then according to Belbin’s
team roles,. This non-parametric test evaluates
whether the groups provided have homogenity in
variance, i.e., test if the variance of groups come
from the same distribution and because of this we
can consider the groups having no statistically
difference between their means and variances. This
test was chosen because our sample data does not
follow the normal distribution according to Shapiro-
Wilk test (Shapiro and Wilk, 1965; Razali and Wah,
2011). Finally, we performed all these statistical
analysis using the SPSS software (IBM, 2013).
In following subsections we will present in more
detail the sample characteristics, the metrics used for
evaluation, and in the next section we present the
results obtained.
3.2 Sample Characterization
This research was performed in a technology
company with ISO 9001/2000 certification who acts
as an agency for perform the information technology
government policies. This company is responsible
for the majority of software systems developed and
currently in use by the state executive branch. The
company’s portfolio includes software systems for
education, human resources, public safety,
administration, government’s strategic planning and
public health.
3.2.1 Projects
This empirical study evaluated data from 65
projects, allowed by the company, started after
March 2011 and completed by March 2012 by a
domain expert team in developing software solutions
for the government. This time period for analyzed
projetcs was established by the company. These
projects have been chosen because they have
common characteristics, such as the same
application domain, the same customer, they were
based on the same programming language and the
same technology. There were no further process for
selecting or rejecting the data collected, thus all
collected projects were used for this study. These
projects involve goals like the development of new
modules of a management system, corrective
maintenance, perfective maintenance and new
software projects.
To better study all projects collected, they were
classified in respect of their duration in order to
finish. The classification system was created with
the help of the supervisor of these projects. He was
reponsible for coordinate the software teams of these
projects. He proposed the following classification:
Figure 1: Histogram of project distribution according to
size classification proposed by company’s supervisor.
•
Very small projects: lasting up to 7 days;
•
Small projects: lasting more than 7 through
10 days;
•
Medium projects: lasting more than 10
through 50 days;
•
Big projects: duration greater than 50 days.
Figure 1 shows the software project distribution
studied in this research, according to the
classification described. This histogram shows a
balance in this project distribution according to size,
which allows a proper analysis of the results
obtained.
3.2.2 Project Effort’s Metrics
Software organizations need to evaluate their
projects in order to continuously improve their
internal processes to achieve better results. This
evaluation has some perspectives, such as customer
perspective and organization perspective.
Customer perspective usually considers a success
when the software project are delivered on time, on
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
107
budget and with all features requested. In this
company, if any replanning with customer occurs
during project execution, the delivery date and
features will reflect this latest agreement.
When evaluating collected data, from the
customer perspective, we noticed that the majority
of the projects had little or no delay on project’s
delivery date. To investigate this we plotted a graph
showing the distribution of the difference between
date actually delivered and date estimated from the
last replanning. These data were normalized by a
ratio between the difference above explained and
project’s estimated date, so that projects can be
compared regardless of its duration. We called this
metric as delivery date deviation where negative
values represent that the project was completed
ahead of schedule and positive values indicate
delays in the project.
Figure 2: Histogram of the projects distribution in respect
of deadline deviations.
Figure 2 shows the histogram of delivery date
deviation calculated from all collected projects. In
this figure we can confirm that almost all projectes
were delivery on date, and this was probably caused
by this project replanning, where the date estimation
was renegotiated and commited into the company’s
project management system. Because of this internal
project’s cost overrun can be hidden from the
customer perspective.
Organization perspective evaluates internal
development process and cost overrun. Cost
management allows to track the internal resources
involved in the project and the final project price for
customer. These data, in the company’s project
management system, are not affected by any project
replanning realized with the customer. These
internal resources are related to the project’s effort
planned and the amount of software engineers
planned for the project. To evaluate the
organization’s perspective, we assume project’s
effort deviation as our main variable.
Figure 3: Histogram of project’s effort deviations.
The project effort deviation is calculated from
the difference between the total hours worked and
the total hours estimated for the project. This metric
evaluates the quality of the estimate made in the
original planning phase.
Figure 3 shows the histogram of project effort
deviation. The existence of negative project effort
deviations. This negative effort deviation represents
projects that took less time than estimated and a
positive deviation effort indicates projects that has
used more hours than was originally estimated.
These values are presented with the normalization of
the scale using the ratio between the actually
realized and estimated, in this way all projects can
be compared regardless of its actual duration.
These metrics when evaluated together allow the
company to assess the final status of their projects.
This kind of evaluation gives possibility of new
analysis, such as identify projects completed on time
but generated loss by allocating more resources than
planned. This work focuses on organization
perspective aiming to improve internal proccess and
avoid financial losses for the company.
3.2.3 Participants
The software team selected cosisted of 22 software
engineers, being 16 men and 6 women. These
engineers were assigned in these project for
functional roles such as software project manager,
systems analyst, developer and tester. The software
project manager in this company, despite being
directly responsible for the coordination and
implementation of the project, has no direct
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
108
authority over resource allocation. The company
supervisor was responsible for the distribution of
engineers between projects and for assigning their
functional roles. Each of the software engineers
could be involved in more than one project
simultaneously by playing different functional roles.
For a software engineer to be selected as a
project manager some criterias were used, such as
the current work overload and degree of knowledge
about business processes involved in the product to
be developed. In cases where projects had
innovative business processes, this criteria included
project management in-house experience
Project manager main duties in this company are
based on a predictive model for project management
(PMI, 2013): scope management, time management,
quality management, risk management, allocation of
tasks to the selected team, stakeholder management
and communication management.
After collected, the software engineers MBTI
test results were analysed and the distribution of the
22 participants were:
•
Source of Energy: 60% introversion (I) and
40% extroversion (E);
•
Worldview Perception: 75% Sensation (S)
and 25% Intuition (N);
•
Evaluation: 55% Thinking (T) and 45%
Feeling (F);
•
Lifestyle: 75% Judgement (J) and 25%
Perception (P).
Figure 4 shows the distribution of MBTI
psychological types of the participants who took
project manager functional role on the 65 projects
analyzed. In this distribuition, the same participant
can appear multiple times as much as he was
assigned to the project manager functional role.
Figure 4: Histogram of project manager MBTI
psychological type for each project collected.
Figure 5 show the distribution of same project
managers according to their Belbin’s team role. This
histogram is based on the dominant role of each
project manager, i.e., where his final test score of
BTRSPI shows the level "Very High (VH)". We
assumed only the “Very High” level because it
better highlights the dominant characterics of each
role.
During selecting the dominant role, in some
cases we had more than one role with "Very High
(VH)" level for the same project manager. In these
cases we duplicated the project record for each
Belbin’s team role identified with “Very Hig” of its
project manager.
Figure 5: Histogram of project managers according to
Belbin’s team role with score at the "Very High (VH)".
In Figure 5, the "None" category describes
managers who did not have any score level with
"Very High (VH)" in any of Belbin’s team role.
4 RESULTS
To answer the research question (does the
personality of the project manager can influence the
outcome of a software project, with respect to
project planned effort?), we adopted the following
criteria for analysis:
As dependent variable of experiments, we
take the project effort deviation;
As explanatory variable (factor), we analyzed
the following characteristics:
o MBTI psychological type;
o Belbin’s team role.
As we can not assure the assumption of
distribution normality of the samples collected, then
we do not use the statistical test Analysis of
Variance (ANOVA) to compare the influence of the
factors (MBTI psychological type and Belbin’s team
role) on the project effort deviation. We adopt the
non-parametric Kruskal-Wallis statistical test (Siegel
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
109
and Castellan, 1988).
Figure 6 shows the project effort deviation
distribution related for each personality type
assessed by the MBTI of each software project
managers. With this boxplot we can notice the
difference for each psychological type.
Figure 6: Boxplot with project effort deviation related to
the MBTI pscyhological type.
To determine if there is a significant difference
between the MBTI psychological types in the
variance of project effort deviation, we applied the
non-parametric Kruskal-Wallis test (Siegel and
Castellan, 1988) using α = 0.05 as confidence level.
As a result, we obtained a p-value of 0.691, thus
rejecting the alternative hypothesis, so that nothing
can be stated about the effects of the MBTI
psychological type on the project effort deviation on
our sample.
We also analyzed the influence of Belbin’s team
roles of project managers on the project effort
deviation using our sample.
Figure 7: Boxplot with project effort deviation related to
Belbin’s team roles.
From Figure 7, we notice these group distributions,
according to Belbin’s team role scored at "Very High
(VH)", have visually large differences on variance in
the project effort deviation.
To determine if this differences were significant,
we applied the non-parametric Kruskal-Wallis test
(Siegel and Castellan, 1988) again using a
confidence level of α = 0.05 confidence. As a result,
we obtained a p-value equal to 0.000, thus rejecting
the null hypothesis and accepting that there is no
homogeneity of variances, i.e., the variance and
mean have significant difference. Therefore, we can
assume that the variance should be the effect of one
of the Belbin’s team roles on the project effort
deviation, and this variance does not occur by
chance in our sample.
Table 1 provides a ranking of variance by
dominant Belbin’s team roles. This table shows that
some roles have less variance with respect to project
effort deviation than others. Table 1 also shows that
in our sample the role of Co-ordinator (CO),
indicated by Belbin as one role of leadership, was
not determinant in order to obtain smaller variances
of project effort deviation. Additionally, we notice
the prominence of the roles Plant (PL) and
Implementer (IMP), with the smallest variances of
only 0.113 and 0.119 respectively.
Table 1: Ranking of Belbin’s team roles with respective
effort deviation variance.
Belbin’s team role
Variance
Plant (PL) 0,113
Implementer (IMP) 0,119
Teamworker (TW) 0,289
Co-ordinator (CO) 0,401
Completer Finisher (CF) 0,847
NONE 3,209
4.1 Result’s Discussion
In this section we discuss our findings and
contributions with this research. For this analysis, by
assessing the personality of the manager of the
project, we can compare our research with studies by
Brewer (2005), and Gorla and Lam (2004).
Brewer (2005) discusses whether a person can be
trained for the project manager role and if he already
has the skills necessary, then how its skills can be
improved. Brewer’s work (2005) highlights the
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
110
ENTJ and ESTJ psychological types as ideal for the
role of project management. ENTJ psychological
type refers to people with strong leadership
characteristics. ESTJ psychological type refers to
people aware with rules. Their result was concluded
by comparing the necessary skills for such a
function and characteristics of the psychological
type. Despite the proposed recommendation, no
evidence was presented to support the proposed
statement was presented. Our work differs from
Brewer (2005) because he do not observe the
influence between MBTI psychological type and
project’s effort deviation. Thus, we can not assume
that ENTJ and ESTJ influence whole project
management activities.
Gorla and Lam (2004) evaluated the influence of
the project manager personality, based on the MBTI
psychological type, on team performance. Their
results, based on assessement of 20 teams from
Hong Kong, identified the relations of MBTI
psychological type with theirs software project’s
success metrics used. Gorla and Lam (2004) indicate
the relevance of the MBTI psychological ENFJ type
in the project’s success using qualitative measures of
success. ENFJ psychological type refers to
charismatic people with guiding people skill.
Therefore differently from our study, no statistically
signifcant result of these relations were presented
and no quantitative metrics for effort deviation were
reserached. Our findings, based on used sample,
differs from them indicating that there is no
influence between MBTI psychological type and
project’s effort deviation. Thus, we can not assume
ENFJ as a main psychological type to manage
projec’s effort deviation.
Another result obtained in our research refers to
the influence of Belbin's team role of the software
project manager. This research identified evidence
of the relationship between the role of the Belbin’s
Plant (PL) and Implementer (IMP) role of project
manager and project’s effort deviation.
Belbin’s (2010b) Plant (PL) role, classified as a
creativity role, has as its main characteristic the
ability to solve problems. Although not classified by
Belbin as a role with coordination skills, its
creativity may have been important for obtaining a
minor variance in project’s effort deviation. As
hypothesis from this result we think that in scenarios
with problems during software development which
could cause an increase in project’s effort, the role
Plant (PL) may have find fast or simple solutions
which avoided the increased the project’s effort.
Belbin’s (2010b) Implementer (IMP) role,
classified as an execution role, has as important
characteristics discipline, inflexibility and the ability
to turn ideas into pratical actions. Our results also
show that this team role achieved a small variance
on effort deviation. We think that is possible that its
skills may also have contributed to better align the
project’s scope originally designed, influencing
positively on the project’s effort deviation.
França and Silva (2007) evaluated the relation
between RUP functional roles, including project
manager, and Belbin’s team role. Franca and Silva
(2004) considered characteristics as communication,
leadership, time management and goal oriented for
project managers. In order to reach these skills, the
researchers suggested Co-ordinator (CO), Resource
Investigator (RI) and Teamworker (TW) Belbin’s
team roles. Our work also show evidence between
project manager Belbin’s team role and project
effort deviation. However, our findings differs from
França and Silva (2004) on the suggested roles. This
result can be explained because project’s effort
deviation is a subset of a project managerial activity,
requiring specifics skills for this.
Thus, the evidence obtained shows that the
Belbin’s team role of the software project manager
can influence the effort deviation in a software
project. Furthermore, our results suggest that
characteristics such as creativity and the ability to
turn ideas into pratical actions, from the Belbin’s
team roles Plant (PL) and Implementer (IMP) are
important factors. This indicates that during the
project, in relation to the deviation effort, these
characteristics maybe more relevant than just the
ability to delegate tasks or challenge the team, like
the Belbin’s team roles of coordination.
5 THREATS TO VALIDITY
Running a case study brings threats that can affect
the validity of its results (Castello Branco et al,
2012). Below we discuss the threats to validity for
this case study.
One threat to validity of the results obtained in
this study is about the amount of participants,
projects and company in the sample. This threat
causes can prevent the findings obtained to not be
generalized outside the same scope our sample.
Regarding the sample of software projects used in
this study, the results reflect an analysis over a year
of activity provided by the software company.
Another threat to validity is that these software
projects maybe influenced by cyclical processes that
may affect the sample. Also about these projects,
another limitation of this study refers to data quality
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
111
according to hours informed to complete project’s
tasks and consequently it can affect the metric effort
deviation. This information is obtained directly from
the declaration of each team member, so there is a
little possibility that certain values may not reflect
the reality.
Other threat to validity is if the distribution of
Belbin’s team roles of participants of this study does
not reflect the distribution found in society. The
accuracy of personalty identification is another
possible threat of this work. This limitation occurs
because the assesment of tests used in this study was
not accompanied by a psychologist. Another
possible factor that may have influenced the
personality identification in this study was the
presentation made for explaining how the tests
should be used. Additionally, the MBTI
psychological types identification and Belbin’s team
role assesment were carried out by participants in a
non-monitored way without the accompaniment of a
researcher.
6 CONCLUSIONS
This research aims to contribute for the assessment
of human factors in software engineering. To
achieve this goal, we studied how to relate MBTI
psychological types and Belbin’s team role of the
software project manager in relation to the project
effort deviation.
The results obtained, based on our sample,
showed evidence that the effort deviation is not
statistically significant in relation to the project
manager’s personality, assesed by MBTI
psychological test. This indicates that MBTI
dimensions like source of energy, worldview
perception, evaluation and lifestyle do not impact on
in project’s planned effort.
The results obtained also give evidence that the
way the software project manager behaves in
teamwork, according to Belbin’s team roles theory,
has a statistically significant impact on the variance
of project effort deviation. This result is important
because Belbin (2010b) highlights for management
activities only the roles Co-odinator (CO) and
Shaper (SH). However, our results suggest that
characteristics such as creativity and ability to turn
ideas into pratical actions, of the roles Plant (PL)
and Implementer (IMP) respectively influenced the
variance reduction of the effort deviation in our
software projects sample.
As future work, we can suggest studies to
identify if the variance of the Belbin’s team roles has
the same characteristic when it relates to the project
size. Furthermore, this study can be extended to the
assessment of psychological personality using other
instruments such as the Big Five (Goldberg, 1992).
ACKNOWLEDGMENTS
We thank the company and the software engineers
who participated in this study. A special thanks to
the efforts of all researchers involved in this research
and for contributions done by USES Research
Group. We would like to thank the financial support
granted by CAPES process AEX 10932/14-3 and by
FAPEAM through processes number:
062.00146/2012, 062.00600/2014, 062.00578/2014;
and 01135/2011.
REFERENCES
Acuña, S., Juristo, N., 2004. “Assigning people to role in
software projects”. Software: Practice and Experience
34, no. 7 (June 2004): pp. 675-696.
Ash, L., 2013. Personality tests: Can they identify the real
you? BBC News. British Broadcasting Corporation.
Website: http://www.bbc.co.uk/news/magazine-
18723950.
Belbin, Meredith R., 2010. “Management Teams: Why
they succeed or Fail”. 3
rd
Edition. Elsevier
Butterworth-Heinemann Ltd.
Belbin, Meredith R., 2010. “Team Roles at Work”.
Elsevier Butterworth-Heinemann Ltd.
Belbin Associates, 2013. Method, Reliability & Validity,
Statistics & Research – A comprehensive review of
Belbin Team Roles.Website: http://www.belbin.com/
content/page/5835/BELBIN(uk)-2013-
A%20Comprehensive%20Review.pdf.
Brewer, J. L, 2005. Project Managers, Can We Make
Them or Just Make Them Better? Proceedings of the
6th Conference on Information Technology Education.
Castello Branco, D. T. M., Prikladnicki, R., Conte, T. U.,
2012. A preliminary study about personality types on
Scrum teams (in Portuguese). Proceedings of the XV
Ibero-American Conference on Software Engineering
(CIbSE 2012) v.1.
Cattell, R. B., Cattell, A. K., Cattell, H. E. P., 1993.
Sixteen Personality Factor Questionnaire 5
th
Edition.
Institute for Personality and Ability Testing:
Champaign, IL.
Cohen, Y., Ornoy, H., Keren, B., 2013. MBTI Personality
Types of Project Managers and Their Success: A field
survey. Project Management Journal, Volume 4,
Number 3.
Costa, P. T., McCrae, R. R., 1992. Revised NEO
Personality Inventory (NEO-FFI) Professional Manual
of Psychological Assessment Resources, Odessa. FL.
ICEIS2015-17thInternationalConferenceonEnterpriseInformationSystems
112
Cruz, S., Da Silva, F. Q., Monteiro, C., Santos, P. C. F.,
Dos Santos, I.R.M., 2011. Personality in Software
Engineering: preliminary findings from a systematic
literature review. Proceedings of XV Annual
Conference on Evaluation & Assessment in Software
Engineering (EASE).
Fernandes, F. L. M., Silva, F. Q. B. da, 2007. Relationship
between personal competences and project
management personality types (in Portuguese).
Proceedings of the 2º Congresso Brasileiro de
Gerenciamento de Projetos. Salvador, Bahia.
França, A. C., Silva, F. Q. B., 2007. A study about RUP
functional roles and personal behavior on software
factory teamwork (in Portuguese). III Workshop Um
Olhar Sociotécnico sobre a Engenharia de Software.
Goldberg, L. R., 1992. The development of markers for
the big-five factor structure. Psychological
Assessment, 4(1), 26-42.
Gorla, N., Lam, Y.W., 2004. Who should work with
whom?. Communications of the ACM 47, no. 6 (June
2004): pp. 79-82.
International Business Machines (IBM), 2013. Statistical
Package for the Social Science (SPSS) - Version
22.0.0.0.
Karn, J., Cowling, T., 2006. A follow up study of the
effect of personality on the performance of software
engineering teams. Proceedings of the ISESE’06.
Morin, E. M., Aubé, C., 2009. Psychology and
management (in Portuguese). Translated by Maria
Helena C. V. Trylinski. Editora Atlas, São Paulo.
Myers-Briggs Type Indicator (MBTI), 2014. Website:
http://www.myersbriggs.org.
Project Management Institute (PMI), 2013. Software
Extension to the PMBOK Guide Fifth Edition.
Razali, N. M., Wah, Y. B., 2011. Power comparisons of
Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and
Anderson-Darling tests. Journal of Statistical
Modeling and Analytics, 2(1), 21-33.
Riding, R. J., Rayner, S., 2001. International Perspective
on Individual Differences. Volume 2 – Self
Perception. Ablex Publishing.
Russell, M. T., Karol, D. L., 1994. 16PF Fifth Edition
Administrator’s Manual. Institute for Personality and
Ability Testing. Champaign, IL.
Waterman, J. A., Rogers, J., 2004. Introduction to the
FIRO-B Instrument. CPP Publishing.
Schultz, D., Schultz, S., 2002. Personality theories (in
Portuguese). Translated by Eliane Kanner. Pioneira
Thomsom Learning, São Paulo.
Shapiro, S. S., Wilk, M. B., 1965. An Analysis of
Variance Test for Normality (Complete Samples).
Biometrika, 52(3/4), 591-611.
Siegel, S., Castellan, J., 1988. Nonparametric Statistics for
the Behavioral Sciences, 2nd ed. McGraw-Hill
International Editions, New York.
Standish Group International, 2013. CHAOS Manifesto
2013: Think Big Act Small. The Standish Group
International Inc.
Wang, Y., Li, F., 2009. How does project manager’
personality matters? Building the linkage between
project managers’ personality and the success of
software development projects. Proceedings of the
24th ACM SIGPLAN conference companion on Object
Oriented Programming Systems Language and
Applications (OOPSLA’09). ACM Press.
Waterman, J. A., Rogers, J., 2004. Introduction to the
FIRO-B Instrument. Consulting Psychology Press,
Palo Alto, CA.
AnEmpiricalStudyabouttheInfluenceofProjectManagerPersonalityinSoftwareProjectEffort
113
