See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/333123516

The main challenges of safety science

Article in Safety Science · May 2019

DOI: 10.1016/j.ssci.2019.05.006

CITATIONS READS

28 262

6 authors, including:

Ji Ge
Northeastern University (Shenyang, China)
8 PUBLICATIONS 49 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Study on Safety Management Evaluation of Productive Enterprises View project

All content following this page was uploaded by Ji Ge on 09 September 2020.

The user has requested enhancement of the downloaded file.

 Safety Science 118 (2019) 119–125

Contents lists available at ScienceDirect

Safety Science
journal homepage: www.elsevier.com/locate/safety

The main challenges of safety science T

a,b,c,⁎ a,c,⁎ a,c a,c a,c a
Ji Ge , Kaili Xu , Xin Zheng , Xiwen Yao , Qingwei Xu , Bohan Zhang
a
School of Resources and Civil Engineering, Northeastern University, NO. 3-11, Wenhua Road, Heping District, Shenyang 110819, PR China
b
School of Resources and Environmental Engineering, Jilin Institute of Chemical Technology, No. 45 Chengde Street, Longtan District, Jilin 132022, PR China
c
Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines, NO. 3-11, Wenhua Road, Heping District, Shenyang, Liaoning 110819, PR China

A R T I C LE I N FO A B S T R A C T

Keywords: Safety science can be viewed as a messy field, although it has developed many different types of accident models
Safety science (or theories and metaphors) and methodologies of analysis in recent decades. As a young field of research, safety
Object science currently faces many challenges. First, among safety researchers, there is no consensus on how to answer
Scope the question “what is the object of safety science?”. In this article, we address the wide-ranging debate over this
Paradigm
issue. Second, many safety researchers still debate the scope of safety science. The main viewpoints on this issue
Models
Professional identity crisis
are briefly reviewed and discussed. Third, it is controversial whether there is a uniform paradigm of safety
science. Fourth, if safety science is viewed as a pure science, it seems that there is no falsifiable model or theory,
although many models and theories do exist. Last but not least, there is a professional identity crisis in the safety
science community because the field has probably not helped people see this scientific community as a group of
safety researchers and professionals. This article is neither a review nor a research study but rather an overview
of the main challenges of safety science that we should prioritize in the future.

1. Introduction Neither a review nor a research work (Zio, 2018), this article instead
offers an overview of the main challenges of safety science; it aims to
Safety science can be described as a relatively young multidiscipline present open questions for consideration and discussion rather than to
(Le Coze et al., 2014; Hale, 2014; Aven, 2014). Hudson (2010) suggests provide complete analyses of the issues addressed. There are five main
that safety science is much harder than rocket science. Safety science is issues in safety science that safety researchers and practitioners should
also viewed as a messy field, although it has developed many different prioritize in the future. We propose that these issues are as follows: (i)
types of accident theories (or models) and analysis methodologies in controversies among safety researchers regarding the object of safety
recent decades (Le Coze et al., 2014). At present, safety science faces science; (ii) discussions about the scope of safety science; (iii) argu-
many challenges, and many scientists and researchers in the field ments about the paradigm of safety science; (iv) debates about the
(Hollnagel, 2014; Aven, 2014; Hale, 2014) are still working to answer validity of safety models or theories; and (v) the professional identity
the question “what is the object of safety science?” Despite these dis- crisis in the safety science community.
cussions, they have not yet reached an agreement on this issue. The The remainder of this article is organized as follows. First, in Section
scope of safety science seems to be broad in many ways, although the 2, we address a wide-ranging debate among safety scientists about the
study of safety has been established as a domain of applied knowledge. object of safety science. Then, in Section 3, the main viewpoints re-
Metaphors, models and theories of safety science have relatively weak garding the scope of safety science are reviewed and discussed. In
evidence to support themselves (Hale, 2014). Safety science has not yet Section 4, we discuss the paradigm and paradigm shifts in safety sci-
developed a uniform paradigm as a mature science, nor has it formed a ence. In Section 5, we discuss the validity of models or theories in safety
falsifiable theory or model to explain many phenomena in safety-re- science. In Section 6, we present the occupational identity crisis of
lated fields (Almklov et al., 2014; Hollnagel, 2014; Dekker & Nyce, safety researchers and safety practitioners in the scientific community
2014; Grant et al., 2018). As a young field of research, safety science of safety. In Section 7, we provide some conclusions.
has probably not helped people to understand that safety scientists
constitute a scientific community focused on safety (Le Coze et al.,
2014).

⁎
Corresponding authors at: School of Resources and Civil Engineering, Northeastern University, NO. 3-11, Wenhua Road, Heping District, Shenyang 110819, PR
China.
E-mail addresses: [email protected], [email protected] (J. Ge), [email protected], [email protected] (K. Xu).

https://doi.org/10.1016/j.ssci.2019.05.006
Received 31 October 2018; Received in revised form 20 March 2019; Accepted 6 May 2019
0925-7535/ © 2019 Published by Elsevier Ltd.
J. Ge, et al. Safety Science 118 (2019) 119–125

Table 1
Chronological listing of the main viewpoints on the object of safety science.
Viewpoints and proponents Brief description

Man-machine-environment systems (Kuhlmann, 1986) The man-machine-environment system is composed of three components, “human, machine and
environment” and is influenced by the interactions of these three
Socio-technical system (Rasmussen, 1997; Leveson, 2004, 2012, 2017) The solutions to safety problems require that we combine social systems and technical systems and
understand their interactions and relations
Accidents (Fu et al., 2004; Fu, 2017) Three main types of losses result from accidents: human safety and health, economic losses and
environmental damage
Zaishi (Lü, 2006) The term “zaishi,” which is a composed word in Chinese, is a phenomenon or event that refers to the
human body, human mind, human society and human living conditions suffering from external factors
that can damage them
Safety matters (Zhang et al., 2007; Yu et al., 2011) People's safety in the fields of life, production and survival
Safety II (Hollnagel, 2014) The object of safety science is accordingly “how people are able to provide the required performance
under expected and unexpected conditions alike”
Safety II means working safely or with safe operations
Safety I (Hale, 2014; Aven, 2014) Safety I is the study of why things go wrong.
Safety I is defined as the opposite of risk or as ‘lack of accidents’
The trinity research objects “hazard + yinhuan + accident” (Xu et al., The trinity of research objects of “hazard + yinhuan (a Chinese phrase meaning hidden trouble in
2015) English) + accident (including occupational disease)”

2. Controversies regarding the object of safety science scientific investigation and as “a multifaceted topic that can be ap-
proached from many different perspectives”. In our opinion, the sci-
A science should have a well-defined object that can be studied entific investigation of safety entails two sides of the same coin: Safety I
(Hollnagel, 2014). An explicit object of research is the symbol of a and Safety II, with Safety II being “heads” and Safety I being “tails” (or
mature science. Without explicitly defining an object of research and the reverse side). Safety science should focus not only on Safety I but
education, any attempt to reach consensus on safety-related issues will also on Safety II. Therefore, it is necessary for safety scientists to create
be hampered by confusion and by different understandings and inter- a new term to describe the object of safety science, which can include
pretations of the value and effectiveness of safety investigations (Stoop both Safety I and Safety II.
et al., 2017). However, safety scientists still debate the question “what Of course, other safety researchers may have many other perspec-
is the subject of safety science?” It is clear from viewing Table 1 that tives on this issue. Reviewing all of them is not the main purpose of this
there are eight main perspectives on the object of safety science. In the article and is beyond the scope of our research. To date, there is no
early days, Kuhlmann (1986) proposed that safety science should focus consensus on the question of “what is the object of safety science?”
on the man-machine-environment system, which is composed of three among safety researchers. Thus, answering this question is the first
components “human, machine and environment” and is influenced by major challenge we must address in safety science.
their interactions. However, Rasmussen (1997) and Leveson (2004,
2012, 2017) think that safety science should focus on socio-technical 3. Discussions on the scope of safety science
systems. Fu et al. (2004) and Fu (2017) believe that the object of safety
science should be accidents, a category that also includes incidents and Safety researchers and practitioners still often face intractable and
occupational health (or occupational hygiene) because safety science is fundamental questions such as “what is the scope or boundary of safety
equal to “accidentology”, which is defined as the scientific study of science?” and “does X belong to the scope of safety science?” For ex-
accidents. Lü (2006) views the object of safety science as the composed ample, the reviewers and editors of safety journals can often be faced
word ‘zaishi’ in Chinese. This word refers to a phenomenon or event in with the question “Is this article within the scope of this journal?”
which the human body, human mind, human society and living con- Different journals focus on different areas. Sometimes, safety scholars at
ditions suffer from external factors that can cause them damage. Zhang universities also need to make a judgement on the issue of “Does this
et al. (Zhang et al., 2007; Yu et al., 2011) argue that safety science graduate thesis belong to the safety discipline?” Students whose major
should focus on matters of safety that affect people's lives, production is safety science and engineering may be confused by the broadness of
and survival. Xu et al. (2015) propose that safety science should focus the safety discipline and the lack of clear boundaries; it may seem that if
on a trinity of research objects: “hazard + yinhuan (a Chinese phrase they want to become qualified safety professionals, there are no limits
meaning “hidden trouble” in English) + accident”. to what they must learn. In enterprises, safety practitioners are usually
Hollnagel (2014) suggests that safety science should study Safety II, puzzled about the scope of their duties, as they might be punished to
which means working safely or ensuring safe operations, rather than varying degrees once an accident occurs in their company, no matter
Safety I, according to which safety is viewed as a condition where the how hard they work and how responsible they are (Zhang et al.,
number of adverse outcomes is acceptably small. Hollnagel proposes 2018b). In total, all of these issues can be ultimately be ascribed to the
that the object of safety science is accordingly “how people are able to unanswered questions of the scope of safety science.
provide the required performance under expected and unexpected The idea that there should be a single definite “science-based”
conditions alike”. While Hale (2014) seems to argue that Safety I, which prescription in safety science may be misleading. Some safety re-
is the study of why things go wrong, is sufficient to form a clear focus searchers debate the scope of safety science (see Table 2). Hopkins
for safety science. Hale (2014) also suggests that the subject of safety (2014) proposes that the scope of safety science is inherently negotiable
science is defined by the communities of safety science that have been and that it depends on the composition of the safety science commu-
formed by safety researchers, safety practitioners and safety teachers at nity. Hale (2014) also suggests that the communities of safety science
any one time. If safety science communities have changed with the have a major influence on the scope of safety science. Therefore, the
development of society, the subject of safety science has also changed scope of safety science is dynamic because safety science is open to
(Hale, 2014). From a risk perspective, Aven (2014) considers that change, negotiation and conflict rather than being a fixed entity (Hale,
safety, defined as the opposite of risk or as ‘lack of accidents’, can offer 2014). Yu et al. (2011) propose that the boundary of safety science is
a sufficiently stable definition to support the work of safety science. Le still in a state of dynamic change and is even blurred in some respects.
Coze et al. (2014) argue that safety can be viewed as an object of That is, the scope of safety science can be transformed from one field to

120
J. Ge, et al. Safety Science 118 (2019) 119–125

Table 2
Chronological listing of views on the scope of safety science.
Primary proponents Brief description

Yu et al. (2011) The boundary of safety science is still in a state of dynamic change and is even blurred in some respects
Hopkins (2014) The scope of safety science is inherently negotiable and depends on the composition of the safety science community
Aven (2014) The scope of safety science is dynamic because safety science is open to change, negotiation and conflict rather than being a fixed entity

Table 3
Chronological listing of the definitions of safety science.
Viewpoint proponents Brief description

Liu (1994) Safety science is a science that studies the essentials of safety as well as its transformation and the conditions that guarantee it
Zhao and Luo (1995) Safety science is a multidisciplinary subcategory that studies the contradictory law of movement between safety and danger caused by the
application of technology, adopts countermeasures to control technological hazards within the allowable limits, promotes the sustainable
and stable development of production, and achieves the goal of protecting the physical and mental health and safety of personnel,
avoiding material property losses and protecting the environment
Zhang et al. (2007); Yu et al. (2011) Safety science is a kind of comprehensive interdisciplinary field that studies the law of human safety activities and its application
Aven (2014) Safety science can be understood as the totality of safety education programmes, researchers, research groups, journals, papers, etc
Hollnagel (2014) Safety science can be viewed as “a conceptual umbrella term that is common to other scientific disciplines” instead of replacing any of
them
Fu (2015) Safety science is a comprehensive system of knowledge about “the occurrence and development rule of accident and means of preventing
accidents”.
Stoop et al. (2017) Safety science is an integral part of science, defined as a scientific discipline with a dedicated object of research and methodology

another; it has shifted from technology to management, from human research groups, journals, papers, etc. Hollnagel (2014) argues that
errors to organization errors, from safety to health, etc. Furthermore, safety science can be viewed as “a conceptual umbrella term that is
safety science has expanded to include life safety, leisure safety, and common to other scientific disciplines” instead of replacing any of
network safety. In the future, the research scope of safety science may them. Fu (2015) proposes that safety science is a comprehensive system
expand to some new fields, such as robot safety and autonomous ve- of knowledge about “the occurrence and development rule of accident
hicle safety. However, what we want to determine is the boundary and means of preventing accidents”. Stoop et al. (2017) think that
between safety science and other scientific disciplines rather than safety science is an integral part of science, defined as a scientific dis-
which fields safety science can include. cipline with a dedicated object of research and a dedicated metho-
If safety researchers and practitioners want to explicitly understand dology. It seems that the definitions of safety science are different for
the scope of safety science, they must first understand the definition of different safety researchers. There may be no definitive answer to the
safety science. Table 3 shows that many safety scientists and re- question of the scope of safety science. Therefore, the second main
searchers have conducted extensive and in-depth discussions on the challenge of this discipline is that its scope is uncertain and con-
definitions of safety science. Liu (1994) proposes that safety science is a troversial.
science that studies the essentials of safety, including its transformation
and the conditions that guarantee it. Zhao and Luo (1995) suggest that 4. Arguments on the paradigm of safety science
safety science is a multidisciplinary subcategory that studies the con-
tradictory law of movement between safety and danger caused by the Science can be used to explain, predict, and ultimately help prevent
application of technology, adopts countermeasures to control techno- adverse events to make the world a better place and establish a “more
logical hazards within the allowable limits, promotes the sustainable just and equal social order” (Dekker and Nyce, 2014). Fuchs (2005)
and stable development of production, and achieves the goal of pro- argues that science as a systematic term is considered to be “a system of
tecting the physical and mental health and safety of personnel, avoiding organized and systematic knowledge production.” A paradigm is
material property losses and protecting the environment. From Fuchs' usually applied to many problems in a field irrespective of their specific
perspective on science and research, Zhang et al. (Zhang et al., 2007; Yu contents, while a methodology is often considered to be a problem-
et al., 2011) suggest that safety science is a type of comprehensive in- oriented procedure that contains a particular paradigm (Gigch, 1979).
terdisciplinary field that studies the law of human safety activities and A paradigm can also be viewed as “what the members of scientific
its application. Aven (2014) suggests that safety science can be un- communities have in common and what they share in values, notions,
derstood as the totality of safety education programmes, researchers, methods and techniques” (Stoop et al., 2017).

Table 4
Listing of the main viewpoints on the paradigm of science.
Viewpoint proponents Brief description

Gigch (1979) A paradigm is usually applied to many problems in a field irrespective of their specific contents
Stoop et al. (2017) A paradigm can also be viewed as “what the members of scientific communities have in common and what they share in values, notions,
methods and techniques”
Kuhn (1962) A science should have a paradigm defined as a distinct way of thinking about problems
Perrow (2009); Le Coze (2013a) Some models are incommensurable and they require separate types of empirical support rather than a unified and theoretical paradigm to hold
them together
Gigch (1979) Some complex organized systems cannot be dealt with by scientific paradigms because they assume that the relationship the relationship
between two parts of the whole system can be easily untangled due to low levels of interaction
Le Coze et al. (2014) Safety science needs to imitate natural science with its causal explanations rather than the social sciences with their intentional explanations
Stoop et al. (2017) It is necessary for safety science to develop its own paradigms
Swuste et al. (2019) Without an agreed-upon paradigm, the status of safety science has been labelled in the literature as pseudo-scientific

121
J. Ge, et al. Safety Science 118 (2019) 119–125

Table 4 shows that a debate on paradigms and paradigm shifts the development of scientific paradigms (Stoop et al., 2017). The de-
proves to be inevitable (Stoop et al., 2017). Kuhn (1962) argues that a velopment of safety science might follow the evolutionary mechanism
science should have a paradigm, defined as a distinct way of thinking of a new discipline that can be created and eventually developed when
about problems. A new scientific research focus on challenging issues new issues that cannot be countered by existing unitary disciplines need
can lead to a discourse between different disciplines, which may lead to to be addressed by the cooperation of different scientific disciplines.
the emergence of a new paradigm and the questioning of the scientific Parasuraman et al. (2008) propose that “Popper’s (1959) well-known
validity of original disciplines. For example, the struggle between the criteria for falsification may be less applicable to theory development in
social and natural sciences challenged the validity and value of most understanding human interaction with complex systems”. Those au-
mainstream science in the 19th century. If we are to cope with future thors suggest that theories should be held to “the standards of being
challenges that lie beyond existing disciplines and applications, we able to specify the constraints in which they account for more or less
must explore and develop paradigm shifts for safety science (Stoop variance in human-system performance” rather than to the standard of
et al., 2017). According to this idea, safety science can be considered as being falsified, which is the standard used in cognitive engineering and
neither a natural science nor a social science. We can further infer that Human Factors or Ergonomics (Parasuraman et al., 2008). According to
safety science may be considered a socio-natural science because it these and other criteria, safety science is still of great value in under-
involves both natural science and social science. If that is the case, the standing and predicting the performance of the human-related system.
paradigms, scientific validity and values of social-natural science may Le Coze (2013b) also suggests that safety science as a distinct and
be different from those of the conventional sciences, and the same is multidisciplinary activity is composed of different disciplines that need
true of safety science. to recognize each other's value and effectiveness in the process of de-
However, Le Coze et al. (2014) suggest that there cannot be a uni- velopment. Leiber (1998) proposes that mathematized natural science
fied paradigm of safety science that can encompass “the multifaceted is currently reminding people of its limits, and we can clearly see the
nature of the topic” because safety issues are considered complex sci- epistemological limits of our understanding of the dynamics of the
entific problems. It is difficult to find common underlying subjects that physical world. Recognizing these limits is likely to provide insight into
are shared across disciplines because of the fragmentation of perspec- the indispensable value of using qualitative arguments in the quanti-
tives in safety science (Le Coze et al. 2014). Therefore, the concept of tative sciences rather than relying on the developmental endpoint of the
various “safety sciences” can be used to replace the idea of a “safety mathematical natural sciences (Leiber, 1998). Stirling and Mayer
science” (Le Coze et al., 2014). Gigch (1979) proposes that some (2000) state that there is no one set of values or frameworks that can be
complex organized systems cannot be dealt with by scientific paradigms emphatically considered more rational than any other in the field of
because they assume that the relationship between two parts of the multi-social discourse. Stoop et al. (2017) propose that a full informa-
whole system can be easily untangled due to low levels of interaction. tion paradigm, engineering design methodologies and a multiple in-
Perrow (2009) suggests that some models are incommensurable and tervention perspective are three cornerstones of a paradigm shift in
that they require separate types of empirical support rather than a safety science. Stoop et al. (2017) seem to suggest that it is necessary for
unified and theoretical paradigm to hold them together (Le Coze, safety science to develop its own paradigms and notions through
2013a). Thus, do we consider that safety science can be divided into creative destruction of a consensus on existing paradigms. Swuste et al.
two parts, natural science and social science, and then classify as nat- (2019) propose that “the status of safety science has been labelled in the
ural science what belongs to natural science and to social science what literature as pseudo-scientific” without an agreed-upon paradigm. In
belongs to social science? Does safety science try to interpret phe- our opinion, it is still an open question whether there is a unified
nomena in the safety domain using the paradigm of natural science or paradigm of safety science. Therefore, the third main challenge of
social science? Le Coze et al. (2014) seem to suggest that safety science safety science is whether there is indeed a unified paradigm, whether
should try to imitate natural science with its causal explanations rather the field can be viewed as a pure science, and how the paradigms of
than social science with its intentional explanations. safety science shift if it is regarded as an applied science.
Safety science seems to lack the ability to centrally control and
coordinate different safety-related disciplines because of its scientific 5. Debates on the scientific validity of models and theories in
diversity (Le Coze et al., 2014). As the product of many interacting safety science
perspectives, this fragmentation is one of the wicked issues facing safety
science (Le Coze et al., 2014). There is also a lack of methodological Models and theories are the products of deduction and induction,
and theoretical unity in safety science. The ambition of safety science as which lead to objective knowledge (Le Coze, 2013b). There are many
a unitary science has been heavily criticized for its scientific disunity. accident-causation models and theories in safety science (Khanzode
Therefore, Le Coze et al. (2014) propose that a unified science in the et al., 2012; Swuste et al., 2010, 2014). For simplicity, safety models
safety domain should be kept at a distance and is not necessarily de- and theories are divided into two specific categories: older models
sirable. Haavik et al. (2016) suggest that there is no way to set up one based on causal-sequence thinking and newer models based on systems
theory (or model) that can include all the causes and methods used to thinking. The question addressed is whether newer theories based on
explain the reasons for failure and success in safety science, and safety systems thinking can play a more effective role in occupational accident
science should seek to generate and maintain different perspectives that prevention than older theories. In recent years, the most prominent
may contribute to the collective task of continuing to increase knowl- models are the newer theories and models based on systems thinking,
edge in the safety domain. Because the world changes and socio-tech- such as STAMP, risk management frameworks, Normal Accident Theory
nical systems change with it, knowledge of safety science will never be (NAT), and the Functional Resonance Analysis Model (FRAM) (Grant
complete (Haavik et al., 2016). However, Grant et al. (2018) suggest et al., 2018). These newer models have more advantages than older
that the tenets of systems thinking represent “a step toward a unified ones in terms of their explanatory power, usefulness, and extent of
model of accident causation” and could conceivably “provide a more acceptability. First, the newer model has better explanatory power than
comprehensive approach to accident investigation than the adoption of the older model. Older models can merely identify a proximate cause as
one model in isolation”. the root cause (Marais et al., 2004; Leveson, 2004), while newer models
It is important for safety scientists to identify the underlying as- can identify systemic causes that can supplement older models. Second,
sumptions and the philosophical underpinnings that are embedded in the newer models based on systems thinking are more useful than the
the theories and models of safety science (Le Coze et al., 2014). Al- older ones. For example, STAMP, as one of newer models, contributes to
though philosophers provide many theoretical frameworks, few safety better understandings and practices in preventing accidents in the do-
researchers focus on the long-term dynamics and control mechanisms of main of high-risk complex socio-technical systems when it is compared

122
J. Ge, et al. Safety Science 118 (2019) 119–125

to the domino models (i.e., the older models) (Hovden et al., 2010). safety science has been noted and doubted by some safety researchers
Third, the newer models based on systems thinking are currently more because of the tension between the scientific and social relevance of the
acceptable than the older ones. These new models have increasingly notion of safety (Stoop et al., 2017). Hovden (Hovden, 1984; Hovden
become the dominant ones within scientific research on safety et al., 2010) has questioned the usefulness of accident models because
(Underwood & Waterson, 2014; Grant et al., 2018). the models were neither scientific enough nor holistic enough to serve
What are the relationships between the newer models and the older the progress of safety science and improve accident prevention in in-
ones? There seems to be no straightforward answer to this question. On dustry. Hale (2006, 2014) suggests that the theories and models of
the one hand, accident models and theories can provide many different safety science lack high-quality proofs and the tradition of testing in-
types of “glasses” that will influence our understanding and help us to terventions, although some of them have enjoyed great popularity.
analyse recommendations for the prevention of adverse events (Hovden Hopkins (2014) proposes that without critical scrutiny, a theory might
et al., 2010). Many safety researchers and practitioners still believe in be just a matter of fashion. High Reliability Organization (HRO) theory,
the domino model and the iceberg metaphor in the occupational acci- as one of three major and popular theories in safety science, is an in-
dent prevention domain (Hovden et al., 2010). On the other hand, demonstrable theory (Hopkins, 2014). Le Coze et al. (2014) suggest that
safety scientists have raised the question “Are these most prevalent few models and theories in safety science have established normative
models still suitable for the current situation and trends in safety re- standards for evaluating their scientific validity across disciplines, al-
search?” or “Are current models up to date?” One could argue that there though many such theories currently exist. Therefore, there is a need to
is no need to develop new models because the current ones can still debate and evaluate the utility of different models because not all
affect more industry practices, while others disagree (Le Coze, 2013b). models are equally useful (Paul and Brian, 2006). Some safety scientists
Many safety scientists and researchers have puzzled over the question and researchers are trying their best to find credible evidence to prove
of “Are the theories and models of safety science complements or or falsify current theories and models so that safety science as a para-
substitutes of each other?” (Hale, 2006). digmatic science can affect things in the real world in a measurable way
However, the relationships among newer models based on systems (Dekker & Nyce, 2014). Research in safety science should increasingly
thinking seem to be complementary (Filho et al., 2017; Salmon et al., demand controlled studies to determine whether the new models are
2012; Grant et al., 2018). Filho et al. (2017) suggest that there are better than the cast-off old ones (Dekker & Nyce, 2014). The urgent
complementary elements between them, although each model has its topic of these controlled studies is the value of system auditing that is
own advantages. For example, the AcciMap does not explicitly address driven by ISO and related management standards (Almklov et al., 2014;
“all of the key systems theory concepts, but graphically presented their Hale, 2014). Grant et al. (2018) suggest that there is no universally
findings in a more succinct manner” (Underwood & Waterson, 2014). accepted theory or model in safety science, although many accident
Conversely, STAMP more clearly embodied “the concepts of systems causation models exist at present. In summary, it seems that there is no
theory but did not provide a simple graphical representation of the falsifiable model or theory in safety science if safety science is viewed as
accident” (Underwood & Waterson, 2014). The systems thinking tenets a pure science. This is the fourth main challenge that we must address.
could provide “a more comprehensive approach to accident investiga-
tion than the adoption of one model in isolation” (Grant et al., 2018).
Although the newer models and theories based on systems thinking 6. The professional identity crisis in the safety science community
have their own benefits, it is worth noting that these theories are not
conclusively proven themselves. Swuste et al. (2019) suggest that the Ibarra (1999) suggests that professional identity can be viewed as
gap between academic developments and the application of models and “an individual's self-concept about their professional role based on their
theories by safety professionals is still rather large. New methods and experiences, attributes, motives, beliefs, and values”. Professional
instruments in safety practice are often introduced. However, there is identify is distinct from an individual's organizational identity, which is
still “little evidence of so-called evidence-based proven approaches” “an indicator of an individual's personal association with where they
(Swuste et al., 2019). presently work” (Provan et al., 2018; Pratt et al., 2006). Professional
Popper (1959) suggests that our methodological decision – that is, identity is described as “the sense of being a professional” and “inter-
to leave nothing unexplained – should be deduced from the require- nalised expectations and behaviours that arise from a role or distinct
ment for the highest degree of universality and accuracy and be sim- network of relationships” (Paterson et al., 2002; Stryker & Burke 2000).
plified to give priority to the rule of the most rigorously tested theories. Nadelson et al. (2015) argue that “the extent to which an individual has
Parasuraman et al. (2008) propose that normal science seeks to con- internalized the elements of a profession reflects the level to which s/he
tinually improve itself by falsifying current theories and models based has developed her/his professional identity”. Professional identity helps
on credible and acceptable evidence. The fundamental question is what safety professionals understand who they are by looking at the com-
collection of data can support these theories and models in safety sci- bination of the above with other aspects of themselves as individuals
ence (Swuste et al., 2014). It is clear from viewing Table 5 that some (Snook et al., 2011). The identity of safety professionals is shaped by
safety researchers take a sceptical attitude towards the falsifiability and “an individual’s experiences, attributes, beliefs, motives, and values”
the usefulness of current theories and models in safety science. From (Provan et al., 2018; Ibarra 1999; Schein, 1978) and is situated “within
the perspective of already existing sciences, the scientific validity of the context of their organization and their profession more broadly”
(Provan et al., 2018). The experiences of working safety professionals

Table 5
Chronological listing of views on the scientific validity of models and theories in safety science.
View proponents Brief description

Paul and Brian (2006) There is a need to debate and evaluate the utility of different models
Hale (2006, 2014) The theories and models of safety science lack high-quality proofs and the tradition of testing interventions
Hovden et al. (2010) The models were neither scientific enough nor holistic enough to serve the progress of safety science and improve accident prevention in industry
Hopkins (2014) A theory might be just a matter of fashion without the critical scrutiny
Le Coze et al. (2014) Few models and theories in safety science have established normative standards for evaluating their scientific validity across disciplines
Dekker & Nyce (2014) Some safety scientists and researchers are trying their best to find credible evidence to prove or falsify current theories and models
Stoop et al. (2017) The scientific validity of safety science has been noted and doubted by some safety researchers because of the tension between the scientific and social
relevance of the notion of safety

123
J. Ge, et al. Safety Science 118 (2019) 119–125

vary considerably across organizations, industries and nationalities, return occupation (Shen, 2013). Hale (1995) proposes that the assign-
although safety professionals have been defined, standardized, trained ment of responsibility between safety professionals and line managers is
and accredited by the INSPHO (International Network of Safety and complex and dynamic. Safety professionals bear great responsibilities
Health Practitioner Organizations) (Pryor et al., 2015). when an accident occurs even though their rights are limited (Zhang
As safety researchers and professionals, we have to acknowledge et al., 2018b). A safety professional normally reports directly to top
that some of us are currently experiencing a sort of professional identity management but has no hierarchical power in the company beyond
crisis (Le Coze et al. 2014). Le Coze et al. (2014) suggest that safety overseeing his own team (Wybo and Van Wassenhove, 2016). Research
science has probably not helped people to see members of the safety has shown that most safety professionals have “a low level of involve-
science community as safety researchers. Researchers in the safety ment with senior management, and low attendance in management
domain often introduce themselves to the public as scientists of other forums or participation in critical decision-making and planning pro-
disciplines (Le Coze et al., 2014); thus, safety researchers and teachers cesses” (Brun and Loiselle, 2002; Pryor, 2014; Provan et al., 2017). A
are suffering from an occupational identification crisis. First, safety greater gap between the actual prevention work accomplished and the
researchers at universities are facing an existential crisis under the official safety mission has resulted from the increase in tasks and
pressure to promote their positions. To achieve higher positions, safety functions assigned to safety practitioners (Jézéquel, 1999).
researchers are required to provide high-grade research results, and In total, safety science can be strongly influenced by research
they are usually evaluated by how many papers they have published in funded by governments, industry and research councils (Hale, 2014).
high impact-factor journals (Hale, 2014). As we all know, most safety- Safety-related journals are one of the most important influences on the
related journals do not meet this criterion (Hale, 2014). Consequently, development of safety science (Aven, 2014; Hopkins, 2014; Hale,
safety researchers face more disadvantages than other researchers in 2014). Compared to researchers in other disciplines, it is more difficult
the domain of publication. As a result, safety researchers in universities for safety researchers in universities and research institutions to con-
have to return to their original disciplines or transfer into other dis- duct their professional research when they are required to publish in
ciplines (Hale, 2014). Second, safety researchers must also apply for high impact-factor journals. At the same time, safety professionals in
funding for their projects, yet it seems doubtful that there are methods enterprises also have their own problems with professional identity.
in safety science that can be understood and accepted unanimously by Therefore, the fifth main challenge of safety science is that there is a
both society and the scientific community (Le Coze et al., 2014). These professional identity crisis in the safety science community.
problems lead to the competitive disadvantage of safety researchers
when applying for scientific research funds and consultation. For ex- 7. Conclusion
ample, with enhanced safety investments in the coal-mining industry in
China (Zhang et al., 2018a), most research projects are about the ap- Stoop et al. (2017) suggest that safety researchers and practitioners
plication of safety engineering and technology in certain fields, espe- may remain confused about the object of their research, as well as about
cially coal mining. This is the reason why some Chinese safety re- the definitions and legitimacy of safety science, if safety scientists at-
searchers prefer to focus on applied safety technology instead of the tempt to restrict safety science to a ‘pure’ science. Safety science faces
fundamental research of safety science. Third, in some countries, the many challenges, and this article reviews and discusses the five most
departments of safety science research at universities and institutes are important ones. The first main challenge of safety science is that among
born with the appointment of safety professors and die with the re- safety researchers, there is no consensus on the object of safety science.
tirement of those professors (Hale, 2014). The communities of safety The second main challenge is that safety researchers and practitioners
science as multidisciplinary groups seem to be “undervalued in uni- often confuse the question “what is the scope of safety science?” The
versities and are subject to being broken up and sent back to their answers to this question are vague and controversial, although the
mono-disciplinary original disciplines” (Hale, 2014). scope of safety science is considered to be dynamic by some safety
Understanding the professional identity of safety professionals is scientists. The third main challenge of safety science is whether there is
pivotal for “understanding how professionals embed themselves in or- a uniform mature science paradigm in safety science if safety science is
ganizations” (Webb, 2015) and “exploring safety professional practice” viewed as a pure science, or how the paradigms shift in safety science if
(Provan et al., 2018). Safety professionals often play an important role it is considered an applied science. Many safety scientists are sceptical
in the application of OHS initiatives (Brun and Loiselle, 2002). Some- that existing theories and models of safety science are falsifiable be-
times, they can also be called “OHS professionals”, “HSE professionals” cause there have been few proofs that can strongly support such the-
or “safety practitioners” (Wybo and Van Wassenhove, 2016; Pryor, ories. The fourth main challenge of safety science is it does not have a
2019; Brun and Loiselle, 2002). The safety profession is approximately testable model, although many accident causation models currently
30 years old, and by and large, it remains immature and fragmented exist. The fifth main challenge of safety science is that some safety re-
(Provan et al., 2018). First, the role of safety professionals within the searchers and safety professionals are experiencing a professional
organization of a company is unclear. In industrial enterprises, the role identity crisis in the safety science community.
of safety professionals is not clear, which gives these individuals a
confused professional identity (Provan et al., 2018). The professional Author contributions
identity of safety professionals is “rife with unresolved contradictions
and tensions” because they are “unsure of their place in the organiza- Ji Ge: conceived of the study, designed the study and collected the
tion” (Provan et al., 2018). Provan et al. (2017) suggest that the role of data, performed the research, analyzed data, and wrote the paper.
a safety professional is socially and organizationally complex and that it Kaili Xu: designed the study and collected the data, performed the
varies significantly across industries. Therefore, defining and aligning research, analyzed data, and wrote the paper.
around a common role of safety professionals would “assist all stake- Xin Zheng: designed the study, performed the research, analyzed
holders to support and interface with the safety profession” (Provan data, and wrote the paper.
et al., 2018). Second, the role of safety practitioners is changing. Safety Xiwen Yao: analyzed data, and wrote the paper.
professionals are no longer regarded as “technical experts but rather as Qingwei Xu: analyzed data, and wrote the paper.
generalists with strong human relation and management skills” Bohan Zhang: involved in writing the manuscript.
(Frederick et al., 1999; Brun and Loiselle, 2002). They are generally
perceived as advisors and specialists who can supply support and in- Acknowledgements
formation at all hierarchical levels (Wybo and Van Wassenhove, 2016).
Third, some safety practitioners view safety as a high-risk and low- This research is funded by The National Key Research and

124
J. Ge, et al. Safety Science 118 (2019) 119–125

Development Program of China (Grant No. 2017YFC0805100), China Liu, Q., 1994. Safety science. Labor Saf. Health 6, 14–17.
and The Science and Technology Research Project of Jilin Institute of Lü, B.H., 2006. Study on some basic problems of safety science. Eng. Sci. 8 (11), 57–69.
Marais, K., Dulac, N., Leveson, N., 2004. Beyond normal accidents and high reliability
Chemical Technology (Grant No. 2016040; Grant No. 2017052), China. organizations: the need for an alternative approach to safety in complex systems. In:
The authors are grateful to Professor Leveson in MIT for her feedback. Paper presented at the Engineering Systems Division Symposium. MIT,
Thanks to two anonymous reviewers for their useful comments and Cambridge, MA.
Nadelson, L., McGuire, S., Davis, K., Farid, A., Hardy, K., Hsu, Y., Wang, S., 2015. Am I a
recommendations to earlier versions of this paper. This article is dedi- STEM professional? Documenting STEM student professional identity development.
cated to my father Xuan Ge. Finally, thanks to everyone who offers us Stud. Higher Educ. 42.
support to finish this article. Parasuraman, R., Sheridan, T.B., Wickens, C.D., 2008. Situation awareness, mental
workload and trust in automation: viable, empirically supported cognitive en-
gineering constructs. J. Cogn. Eng. Decis. Mak. 2 (2), 140–160.
Declaration of Competing Interest Paterson, M., Higgs, J., Wilcox, S., Villenuve, M., 2002. Clinical reasoning and self-di-
rected learning: key dimensions in professional education and professional sociali-
sation. Focus Health Prof. Edu. 4 (2), 5–21.
There are no conflicts of interest.
Paul, J., Brian, S.M., 2006. Description, diagnosis, prescription: a critique of the appli-
cation of co-evolutionary models to natural resource management. Environ. Conserv.
References 33 (4), 281–293. https://doi.org/10.1017/S0376892906003444.
Perrow, C., 2009. What’s needed is application, not reconciliation: a response to
Shrivastava, Sonpar and Pazzaglia. Human Relations 62 (9), 1391–1393. https://doi.
Almklov, P.G., Rosness, R., Størkersen, K., 2014. When safety science meets the practi- org/10.1177/0018726709339120.
tioners: does safety science contribute to the marginalization of practical knowledge? Popper, K.R., 1959. The logic of scientific discovery. London: Hutchinson & Co.
Saf. Sci. 67, 25–36. https://doi.org/10.1016/j.ssci.2013.08.025. Pratt, M.G., Rockmann, K.W., Kaufmann, J.B., 2006. Constructing professional identity:
Aven, T., 2014. What is safety science? Saf. Sci. 67, 15–20. the role of work and identity learning cycles in the customization of identity among
Brun, J.P., Loiselle, C.D., 2002. The roles, functions and activities of safety practitioners: medical students. The Academy of Management Journal 49 (2), 235–262.
the current situation in Quebec. Saf. Sci. 40, 519–536. Provan, D.J., Dekker, S.W.A., Rae, A.J., 2018. Benefactor or burden: exploring the pro-
Dekker, S., Nyce, J., 2014. There is safety in power, or power in safety. Saf. Sci. 67, 44–49. fessional identity of safety professionals. J. Saf. Res. 66, 21–32. https://doi.org/10.
Filho, A., Jun, G., Waterson, P., 2017. Four studies, two methods, one accident—another 1016/j.jsr.2018.05.005.
look at the reliability and validity of Accimap and STAMP for systemic accident Provan, D.J., Dekker, S.W.A., Rae, A.J., 2017. Bureaucracy, influence and beliefs: A lit-
analysis. In: The 2nd International Conference on Engineering Sciences and erature review of the factors shaping the role of a safety professional. Saf. Sci. 98,
Technologies. 98–112.
Fu, G., Zhang, J.S., Xu, S.R., 2004. Critical Examination of safety science. Engineering Pryor, P., 2019. Developing the core body of knowledge for the generalist OHS profes-
Science 8, 12–16. https://doi.org/10.3969/j.issn.1009-1742.2004.08.003. sional. Saf. Sci. 115, 19–27.
Fu, G., 2015. Studies on the Structure of Safety Science. Safety Science Publishing Pty. Pryor, P., 2014. Towards an Understanding of the Strategic Influence of the OHS
Fu, G., 2017. Safety science and accident causation. Safety 38 (2). https://doi.org/10. Professional. Federation University.
3969/j.issn.1002-3631.2017.02.001. Pryor, P, Hale, A., Hudson, D., 2015. The OHS professional: a framework for practice –
Fuchs, C., 2005. Science as a self-organizing meta-information system. In: Dobronravova, role, knowledge and skills. International Network of Safety and Health Practitioner
I., Hofkirchner, W. (Eds.), Science of Self-Organization and Self-Organization of Organisations, Park Ridge, IL, USA: INSHPO.
Science. Abris, Kiew, pp. 126–199. Rasmussen, J., 1997. Risk management in a dynamic society: a modelling problem. Saf.
Frederick, L.J., Winn, G.L., Hungate, A.C., 1999. Characteristics: employers are seeking in Sci. 27 (2–3), 183–213.
today’s safety professionals. Prof. Saf. 27–31. Salmon, P.M., Cornelissen, M., Trotter, M.J., 2012. Systems-based accident analysis
Gigch, J.P.V., 1979. A methodological comparison of the science, systems and meta- methods: a comparison of Accimap, HFACS, and STAMP. Saf. Sci. 50 (4), 1158–1170.
system paradigms. Int. J. Man Mach. Stud. 11 (5), 651–663. Schein, E., 1978. Career Dynamics: Matching Individual and Organizational Needs.
Grant, E., Salmon, P.M., Stevens, N.J., Goode, N., Read, G.J., 2018. Back to the future: Addison-Wesley, Reading, MA.
what do accident causation models tell us about accident prediction? Saf. Sci. 104, Shen, Y.X., 2013. Brief discussion on career development of safety practitioners.
99–109. https://doi.org/10.1016/j.ssci.2017.12.018. Theoretical Research on Urban Construction: Electronic Version (15).
Haavik, T.K., Antonsen, S., Rosness, R., Hale, A., 2016. HRO and RE: a pragmatic per- Snook, S., Nohria, N., Khurana, R. (Eds.), 2011. The Handbook for Teaching Leadership:
spective. Saf. Sci. https://doi.org/10.1016/j.ssci.2016.08.010. Knowing, Doing, and Being. Sage, Thousand Oaks, CA.
Hale, A., 2014. Foundations of safety science: a postscript. Saf. Sci. 67 (8), 64–69. Stirling, A., Mayer, S., 2000. Precautionary approaches to the appraisal of risk: a case
Hale, A.R., 2006. Method in Your Madness: System in Your Safety. Valedictory Lecture. study of a genetically modified crop. Int. J. Occup. Environ. Health 6 (4), 296–311.
Delft University of Technology, Delft, Netherlands, ISBN 90-5638-160-1. Stoop, J., Kroes, J.D., Hale, A., 2017. Safety science, a founding fathers’ retrospection. Saf.
Hale, A.R., 1995. Occupational health and safety professionals and management: identity, Sci. 94, 103–115. https://doi.org/10.1016/j.ssci.2017.01.006.
marriage, servitude or supervision? Saf. Sci. 20, 233–245. Stryker, S., Burke, P., 2000. The past, present and future of an identity theory. Soc.
Hollnagel, E., 2014. Is safety a subject for science? Saf. Sci. 67, 21–24. Psychol. Q. 63 (4), 284–297.
Hopkins, A., 2014. Issues in safety science. Saf. Sci. 67, 6–14. Swuste, P., Gulijk, C.V., Zwaard, W., 2010. Safety metaphors and theories, a review of the
Hovden, J., 1984. Behøver vi ulykkesmodeller Fordeler og ulemper ved ulike modeller. In occupational safety literature of the US, UK and the Netherlands, till the first part of
Norwegian [Do we need accident models?] In: Fjärde Nordiska the 20th century. Saf. Sci. 48 (8), 1000–1018.
Olycksfallsforskningsseminariet, Technical Research Centre, Symposium, 55, Espoo. Swuste, P., Gulijk, C.V., Zwaard, W., Oostendorp, Y., 2014. Occupational safety theories,
Hovden, J., Albrechtsen, E., Herrera, I.A., 2010. Is there a need for new theories, models models and metaphors in the three decades since world war II, in the United States,
and approaches to occupational accident prevention? Saf. Sci. 48 (8), 950–956. Britain and the Netherlands: a literature review. Saf. Sci. 62, 16–27.
https://doi.org/10.1016/j.ssci.2009.06.002. Swuste, P., Zwaard, W., Groeneweg, J., Guldenmund, F., 2019. Safety professionals in the
Hudson, P., 2010. Safety science: it's not rocket science, it's much harder. Delft University Netherlands. Saf. Sci. 114, 79–88. https://doi.org/10.1016/j.ssci.2018.12.015.
of Technology. Underwood, P., Waterson, P., 2014. Systems thinking, the Swiss Cheese Model and ac-
Ibarra, Herminia, 1999. Provisional selves: experimenting with image and identity in cident analysis: a comparative systemic analysis of the Grayrigg train derailment
professional adaptation. Adm. Sci. Q. 44, 764–791. using the ATSB, AcciMap and STAMP models. Accid. Anal. Prevent. 68 (Sp. Iss. SI),
Jézéquel, Bernadette, 1999. É tude sur les besoins de formation des fonctionnels sé curité 75–94.
et hygie‘ne industrielle. INRS, Nancy. Webb, S.A., 2015. Professional identity and social work. 5th International Conference on
Khanzode, V.V., Maiti, J., Ray, P.K., 2012. Occupational injury and accident research: a Sociology and Social Work.
comprehensive review. Saf. Sci. 50 (5), 1355–1367. Wybo, J.L., Van Wassenhove, W., 2016. Preparing graduate students to be HSE profes-
Kuhlmann, A., 1986. Introduction to Safety Science. Springer, New York. sionals. Saf. Sci. 81, 25–34.
Kuhn, T.S., 1962. The Structure of Scientific Revolutions. University of Chicago Press, Xu, M., Wu, Z.Z., Luo, Y., Kang, R.X., Guan, L., 2015. Discussion on research objects of
Chicago. safety science. China Saf. Sci. J. 25 (12), 3–8.
Le Coze, J.C., Pettersen, K., Reiman, T., 2014. The foundations of safety science. Saf. Sci. Yu, X.W., Zhang, G., Nie, W., 2011. Study on interdiscipline and architecture of safety
67 (10), 1–5. science. J. Saf. Sci. Technol. 07 (3), 48–53.
Le Coze, J.C., 2013a. a. Outlines of a sensitising model for industrial safety assessment. Zhang, J.J., Cliff, D., Xu, K.L., You, G., 2018a. Focusing on the patterns and characteristics
Saf. Sci. 51 (1), 187–201. of extraordinarily severe gas explosion accidents in Chinese coal mines. Process Saf.
Le Coze, J.C., 2013b. b. New models for new times. An anti-dualist move. Saf. Sci. 59, Environ. Prot. 117, 390–398. https://doi.org/10.1016/j.psep.2018.05.002.
200–218. Zhang, J.L., Wang, J.Y., Huang, H., 2007. The object of study of safety science and its
Leiber, T., 1998. On the impact of deterministic chaos on modern science and philosophy knowledge system. China Saf. Sci. J. 17 (2), 16–21.
of science: implications for the philosophy of technology? Phil. Tech. 4, 48. Zhang, J.S., Fu, J., Hao, H., Chen, N., Zhang, W.Y., Kim, Y.C., 2018b. Development of
Leveson, N.G., 2004. A new accident model for engineering safer systems. Saf. Sci. 42 (4), safety science in Chinese higher education. Saf. Sci. 106, 92–103.
237–270. Zhao, Y.S., Luo, Z.J., 1995. On several basic issues of safety science (Continuation). Geol.
Leveson, N.G., 2012. Engineering a Safer World: Applying Systems Thinking to Safety. Explor. Saf. 1, 38–41.
MIT Press, Massachusetts. Zio, E., 2018. The future of risk assessment. Reliab. Eng. Syst. Saf. 177 (C), 176–190.
Leveson, N.G., 2017. Rasmussen's legacy: a paradigm change in engineering for safety. https://doi.org/10.1016/j.ress.2018.04.020.
Appl. Ergon. 59 (Pt B), 581–591. https://doi.org/10.1016/j.apergo.2016.01.015.

125

View publication stats