Safety Science 105 (2018) 192–211

Contents lists available at ScienceDirect

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

Review

Maturity models and safety culture: A critical review T

a,⁎ b
Anastacio Pinto Goncalves Filho , Patrick Waterson
a
Ministry of Labour, Federal University of Bahia, Salvador/Bahia, Brazil
b
Human Factors and Complex Systems Group, Loughborough University Design School, Loughborough University, Loughborough LE11 3TU, United Kingdom

A R T I C L E I N F O A B S T R A C T

Keywords: The available evidence suggests that maturity models are a popular means of assessing safety culture in orga-
Safety culture nisations. The aim of the present study was to review their conceptual underpinnings and roots, as well as
Maturity model provide details of how they have been used to assess safety culture (e.g., types of methods used, coverage of
Safety culture assessment safety domains). A total of 41 publications were reviewed based on a set of selection criteria (e.g., studies which
explicitly reported data or a case study which used a maturity model). The findings indicate steady growth in the
use of maturity models to assess safety culture particularly within domains such as construction, the oil and gas
industries and healthcare. We also found that most studies focus on providing a descriptive account of safety
culture using maturity models and make limited attempts to assess the reliability/validity of outcomes from their
use. We discuss the strengths and weaknesses of maturity models in the light of our findings, alongside iden-
tifying a number of new directions for future work of relevance to safety researchers and practitioners (e.g., the
need for more detailed case studies of the use of maturity models to assess safety, as well as more attention to the
underlying theory guiding use of maturity models).

1. Introduction (e.g., Reiman and Rollenhagen, 2014; Dekker, 2018) suggest that a
preoccupation with safety culture has shifted the focus away from more
Some of the most compelling arguments that culture and safety systemic accounts of the causes of accidents and encouraged a rather
might contribute to accidents and disasters were made in the late 1970s ore superficial account of how safety is related to system levels and
by Barry Turner in his pioneering work ‘Man-Made Disasters’ (Turner, other organisational dynamics (e.g., how safety culture changes over
1978; Pidgeon, 1988). Following the 1986 Chernobyl nuclear disaster time). Finally, Antonsen (2009b) compared qualitative and quantitative
the term ‘safety culture’ started to be regularly used amongst a broad descriptions of the safety culture in the same organisation (a Norwegian
community of safety scientists, psychologists and other groups (Silbey, oil and gas platform) and found them to be dramatically different,
2009). There are a number of different explanations for the rise in in- leading him to cast doubt on the predictive validity of safety culture
terest in the construct of safety culture including increasing recognition assessments. In the present paper we focus on a review of one popular
of the importance of cultural aspects of health and safety management tool or approach which is used to assess safety culture, namely maturity
(Cooper, 2000; Cox and Cheyne, 2000; Flin et al., 2000; Reason, 1998) models. A later section of the paper discusses the findings from our
and the shift in the last few decades towards a focus on organisational review in the light of contemporary criticisms of the safety culture
factors governing risk and safety (Borys et al., 2009; Waterson et al., construct, alongside a consideration of how maturity models fit within
2015; Robertson et al., 2016). As a result, many contemporary orga- debates centred on research-practice gaps within safety science and
nisations strive to understand and improve their safety culture in order human factors (Chung and Shorrock, 2011; Waterson, 2016).
to deliver effective health and safety management and enhance their
safety performance (Antonsen, 2009a; Reason, 1998, 2016). 2. Safety culture: some current challenges
At the same time, amongst researchers and academics, there have
been a number of criticisms levelled at the construct of safety culture. 2.1. Defining ‘safety culture’
Henriqson et al. (2014) for example, argue that the study safety culture
encourages the view that safety is a widely shared norm, value or set of Despite the considerable literature covering theoretical and em-
beliefs within organisations which masks important conflicts and dis- pirical aspects of safety culture (Antonsen, 2009a, b; Cox and Flin,
agreements which may exist amongst employees and managers. Others 1998; Díaz-Cabrera et al., 2007; Flin et al., 2000; Guldenmund, 2000;

⁎
Corresponding author at: Ministério do Trabalho, Rua Ewerton Visco, 190, Caminho das Árvores, Salvador, Bahia CEP 41.820.022, Brazil.
E-mail address: anastaciofilho@ufba.br (A.P. Goncalves Filho).

https://doi.org/10.1016/j.ssci.2018.02.017
Received 23 March 2017; Received in revised form 21 December 2017; Accepted 16 February 2018
Available online 23 February 2018
0925-7535/ © 2018 Elsevier Ltd. All rights reserved.
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

Hopkins, 2006; Lee and Harrison, 2000; Mearns et al., 2009) there is and interactions within an organisation and the dynamic interplay be-
still a lack of consensus and agreement about how to define the con- tween these which shapes and influences culture. The pragmatic ap-
struct as well as assessment methods and on the overall structure of proach also places emphasis on the types of processes that an organi-
safety culture assessment (Mkrtchyan and Turcanu, 2012). The differ- sation should have in order to achieve a mature or advanced status with
ence between safety climate and safety culture, for example, has been regard to safety culture. These processes are reflected in Geller’s ap-
debated over decades by a number of safety researchers (Flin et al., proach towards Total Safety Culture (Geller, 1994) and safety culture
2000; Griffin and Curcuruto, 2016). Safety culture typically refers to maturity models such as the Shell Hearts and Minds programme
the underlying assumptions and values that guide behaviour in orga- (Hudson and Willekes, 2000; Hudson, 2007).
nisations rather than the direct perceptions of individuals (Griffin and
Curcuruto, 2016). Safety climate, by contrast, is sometimes regarded as 2.3. Aims, objectives and organisation of the current study
the surface features of the safety culture discerned from the workforce’s
attitudes and perceptions at a given point in time (Flin et al., 2000). The focus of the current study is to outline the results of carrying a
Andrew Hale (2000) refers to these and a range of other discussions literature review on one particular approach towards safety culture,
centred on safety culture (e.g., the relation of culture to other aspects of namely the use of maturity models for safety culture assessment. There
safety management and behaviour) as examples of ‘culture’s confu- is some evidence to suggest that maturity models are increasing in
sions’. More recently, Hale stated:“… safety culture is problematic in popularity (e.g., Fleming, 2001, 2017; Goncalves Filho et al., 2010;
many of the same ways that ‘accident proneness’ was in the last century; in Parker et al., 2006; Health and Safety Technology and Management,
terms of its attributional consequences, the difficulties of defining it and the 2017; Office of Rail and Road, 2017). Previous reviews have been
difficulties of deciding what you should measure as the outcome of its pre- carried out on the subject of maturity models and their use within
sence or absence; either accidents or other intermediate measures of safety” domains such as software, management, business process management,
(Waterson, 2017). information management and information technology management
(Becker et al., 2009; Maier et al., 2012; Wendler, 2012). Little work
2.2. The theoretical status of safety culture however, has assessed the extent to which maturity models have been
used to assess safety culture, as well as their scope and coverage. With
A number of authors have attempted to characterise the various this in mind, the specific aims of the review are:
theoretical approaches and methods which have been used to assess
safety culture. Silbey (2009) for example, describes three dominant 1. To provide a better understanding of how maturity models to assess
‘lenses’ which characterise what she terms as ‘talk about safety and safety culture have been developed; their conceptual underpinnings
culture’. The first ‘lens’, ‘culture as causal attitude’, view safety culture as and roots; the range of safety domains in which they have been
something that is measureable and comprises the values, competencies, applied; and, characteristics of their use;
attitudes and behaviours about safety which exist within organisations. 2. To examine the methodological properties of maturity models to
From this point of view culture “determine[s] the commitment to, and assess safety culture and the extent to which the outputs from using
the style and proficiency of, an organisations’ health and safety pro- maturity models are evaluated (e.g., assessment of validity and re-
grams” (Silbey, 2009, p. 350 quoting Reason, 1997, p. 194). By con- liability);
trast, the second ‘lens’, ‘culture as engineered organisation’ whilst simi- 3. To use the outcomes from the review to offer some reflections on the
larly focusing on the importance of cultural factors on safety outcomes, theoretical status of the use of maturity models to assess safety
places more emphasis on how an organisation configures its processes culture and suggest new directions for future research and practice.
and practices in order to improve safety, reliability and resilience.
Proponents of the High Reliability Organisations (HROs) approach to- 3. Maturity models and safety culture
wards safety are viewed by Silbey (2009) as examples of the ‘culture as
engineered organisation’ approach to safety culture (e.g., Eisenhardt, 3.1. Definition and scope
1993; La Porte and Rochlin, 1994; Weick, 1987). A third ‘lens’ refers to
‘culture as emergent and indeterminate’. From this point of view, safety Maturity models involve defining maturity stages or levels which
culture is understood to be socially constructed and mediated by arte- assess the completeness of the analysed objects, usually organisations or
facts and material, both mental and representational (Gherardi and processes, via different sets of multi-dimensional criteria (Wendler,
Nicolini, 2000). 2012; Becker et al., 2009). Hudson (2007) defines the use of maturity
An alternative characterisation of safety culture and ways in which models in safety culture in terms of a continuum ranging from orga-
it is conceptualised and assessed which draws partly on Burrell and nisations that have unsafe cultures (‘pathological’ organisations)
Morgan’s (1979) analysis of sociological paradigms is provided by through to those who manage safety proactively (‘generative’ organi-
Guldenmund (2010, 2016). Guldenmund describes three approaches: sations) and those who are an intermediate stage of development
(1) interpretative or anthropological approaches – these often treat culture (‘bureaucratic’ organisations). Organisations are seen as progress se-
as a system of meanings and symbols shared between groups of in- quentially through the stages, by building on the strengths and re-
dividuals who participate in this social process. Culture cannot be moving the weaknesses of the previous levels (Fleming, 2001). A ma-
changed easily and cannot be assessed easily using scientific methods turity model is a descriptive model in the sense that it describes
(Geertz, 1973; Alvesson, 2007; Martins, 1992). Qualitative methods, essential, or key, attributes that would be expected to characterise an
such as a narrative study, phenomenology, grounded theory, ethno- organisation at a particular level.
graphy or case studies (Antonsen, 2009a, b), or various combinations of The application of this concept is not limited to any particular do-
these approaches, are methods used by an interpretative approach. Data main (Wendler, 2012) and maturity models can be used both as an
collection such interviews, observational studies, document analysis are assessment tool and as an improvement tool (Maier et al., 2012). Focus
typically used to provide clues to underlying cultural assumptions (e.g., groups, interviews, audits and checklists support maturity models in
Scott et al., 2003); (2) analytical or psychological approaches – this is safety culture as well as questionnaires. Assessment can be also struc-
similar to Silbey’s notion of ‘culture as causal attitude’, however, speci- tured around a matrix or grid, where levels of maturity are allocated
fically relates to the use of questionnaires and to assess safety culture against key aspects of performance or key activities, thereby creating a
and the analysis of dimensions, factors and other statistical and psy- series of cells. An important feature of this maturity matrix approach is
chometric properties of the survey instrument being used; (3) pragmatic that the cells contain descriptive text for the characteristic traits of
or experience-based approaches – this approach focuses on the structure performance at each level. One stated advantage of the use of a

193
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

maturity matrix is that it is less complex than diagnostic and im- with a heavy focus on written rules and procedures. An “independent
provement tools (Maier et al., 2012). Based on the results of the ana- culture” where the focus is on a personal commitment to and respon-
lysis, recommendations for improvement measures can be derived and sibility for safety. The final stage is “interdependent” where there is a
prioritized in order to reach higher maturity levels (Becker et al., 2009). team commitment to safety with everyone having a sense of responsi-
bility for safety beyond their own work and by caring for the safety of
3.2. Origins others.

The use of maturity models as a tool to assess safety culture can be 3.2.2. Westrum’s typology of organisational cultures
traced back to two main roots, namely, previous work on the ‘Quality Typology of organisational was proposed by Ron Westrum (1993,
Management Maturity Grid’ and Westrum’s ‘Typology of 2004). Westrum’s model distinguished three types of organisation –
Organisations’. pathological, bureaucratic and generative. Table 1 explains these ca-
tegories and provides an example how their characteristics apply within
3.2.1. Quality Management Maturity Grid organisations. This model was extended from three to five stages in
The Quality Management Maturity Grid (QMMG) was first proposed sequence, replacing the label ‘bureaucratic’ with ‘calculative’ and in-
by Philip Crosby in 1979 (Wendler, 2012). In the QMMG, Crosby ad- troducing two extra stages, the reactive and the proactive stages
vocated that organisations go through five successive stages of quality (Hudson, 2001, 2007 – Fig. 2). This was done in order to allow for more
maturity as they approach the maximum level of quality in all phases of subtle and accurate classification, and at the same time increasing the
organisational activity, these are: uncertainty, awakening, enlight- accessibility of the framework to industry employees by including terms
enment, wisdom and certainty. In the first stage, management has no they would be familiar with (Parker et al., 2006).
comprehension of quality as a management tool. The intermediate
stages are characterised by a transformation in management under- 4. Method
standing and attitude towards quality, how quality appears within an
organisation, how organisational problems are handled, the cost of 4.1. Search strategy and sources
quality as a percentage of sales, quality improvement actions taken by
management, and how management summarizes the organization’s A literature search, including publications in academic journals,
quality problems. In the final stage of the QMMG, Total Quality Man- conference papers and grey literature covering the period January
agement (TQM) is viewed as an essential part of the company system 2000–January 2017 was carried out using the following electronic
(Calingo, 1996; Fraser et al., 2002). QMMG was adapted for process of databases: Google Scholar, Science Direct, UK Health and Safety
building by the Carnegie-Mellon Software Engineering Institute (SEI) in Executive (HSE), Web of Science, British Library, Open Grey (in-
1986, where the concept of the Capability Maturity Model (CMM) was formation on grey literature in Europe), ProQuest (dissertations and
added and developed the foundation for its current use throughout theses within the UK and Ireland), Canadian Centre for Occupational
industry (Paulk et al., 1993). The CMM was designed to guide software Health and Safety (CCOHS, include OSHLINE and NIOSHTIC), NIOS-
organisations in selecting process improvement strategies by de- HTIC-2 and Occupational Safety and Health Resource Database.
termining current process maturity and identifying the few issues most
critical to software quality and process improvement. It has five levels: 4.2. Search terms and exclusion/inclusion criteria

Initial: The software process is characterised as ad hoc, and occa- The following search terms were used: “safety culture maturity”,
sionally even chaotic. Few processes are defined, and success de- “safety culture maturity model”, “safety culture AND maturity”, “safety
pends on individual effort; AND culture AND maturity”, “framework and safety culture maturity”,
Repeatable: Basic project management processes are established to “safety culture maturity AND model”. The reference sections and bib-
track cost, schedule, and functionality. The necessary process dis- liographies of articles were also searched. When publications were not
cipline is in place to repeat earlier successes on projects with similar found directly from databases, the authors were contacted and copies of
applications; their articles were requested. The publications were filtered using a set
Defined: The software process for both management and engineering of inclusion and exclusion criteria. Publications where the development
activities is documented, standardized, and integrated into a stan- of a maturity model was described as a tool to assess safety culture or
dard software process for the organisation. All projects use an ap- presented an application of a maturity model were included within the
proved, tailored version of the organization's standard software review. Similarly, publications where the dimensions used to describe
process for developing and maintaining software; and assess safety culture were presented were included. Publications
Managed: Detailed measures of the software process and product where actual development and/or application of a maturity model were
quality are collected. Both the software process and products are not described were rejected, as were those written in languages other
quantitatively understood and controlled; than English.
Optimising: Continuous process improvement is enabled by quanti-
tative feedback from the process and from piloting innovative ideas 4.3. Categorisation framework
and technologies. These five levels provide a framework for orga-
nizing these evolutionary steps into five maturity levels which are In order to provide a structure for the review, we categorised pub-
put forward as laying successive foundations for continuous process lications using a framework made up of the following nine elements:
improvement (Paulk et al., 1993, Fig. 1).
(1) Publication title, author(s) and year of publication;
One of the first maturity model developed to assess safety was (2) Source (e.g., journal, thesis);
Dupont Bradley Curve which was examined by Fleming (2001) to de- (3) Country of origin;
velop a safety culture maturity model. The four stages in this model are: (4) Application domain;
(i) reactive; (ii) dependent; (iii) independent; and, (iv) interdependent (5) Methods used to develop, evaluate and apply the model – this involved
(Foster and Hoult, 2013). In the first “reactive” stage people do not take identification different methods used to develop, evaluate and
responsibility. They believe that safety is more a matter of luck than apply the maturity model (e.g., questionnaire, interview, literature
management, and that “accidents will happen.” In a “dependent cul- review and focus group);
ture” there is an emphasis on management and supervisory control, (6) Type of publication – in this case we made use of the distinction

194
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

Fig. 1. Quality Management Maturity Grid (Crosby, 1979) and Capability Maturity Model (Paulk et al., 1993).

Table 1
Typology of organisation cultures (adapted from Westrum, 2004).

Typology of Characteristic Example

organisation

Pathological Preoccupation with personal power, needs, and Information is hidden, messengers are “shot”, responsibilities are shirked, bridging is
glory discouraged, failure is covered up, new ideas are actively crushed
Bureaucratic Preoccupation with rules, positions, and Information may be ignored, messengers are tolerated, responsibility is compartmentalised,
‘departmental turf’ bridging is allowed but neglected, organisation is just and merciful, new ideas create problems
Generative Concentration on the mission itself, as opposed to a Information is actively sought, messengers are trained, responsibilities are shared, bridging is
concentration on people or positions rewarded, failure causes inquiry, new ideas are welcomed

Fig. 2. Typology of organisations (Westrum, 1993).

195
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

drawn by Wendler (2012) between: ‘maturity model development’ 5.1. Year of publication, country of origin and application domain
(defined as ‘publications where the main objective was developing
or constructing a new maturity model or adaptation of an existing In order to provide an overview of trends (e.g., growth, country of
maturity model’); ‘maturity model application’ (defined as ‘pub- origin) in maturity models, the publications were analysed by their year
lications where the main aim is the application of maturity models of publication. The distribution over the last two decades (January
in several contexts or specific domains’); ‘maturity model valida- 2000–January 2017) is shown in Fig. 3. Between 2000 and 2005, the
tion’ (defined as ‘publications where the main purpose is vali- number remained stable with at least one publication a year. Beginning
dating existing maturity models and includes empirical as well as in 2006, a steady rise is noticeable, reaching its peak with 7 publica-
conceptual validation, comparisons of maturity models and other tions in 2011. The country which published the most on maturity
approach and simulations’); and, ‘maturity model reliability’ (de- models was the United Kingdom with 8 publications (1, 8, 9, 12, 16, 24,
fined as ‘publications where the main purpose is reliability testing 27 and 31) followed by Canada with 7 publications (3, 5, 7, 15, 18, 34
of existing maturity models’); and 39). One maturity model was applied in two countries, Canada and
(7) Structure – this involved checking the structure of maturity United Kingdom (Kirk et al., 2007, 8).
models used in this study (e.g., number of level, description or There is fairly uniform application across a range of domains, but
summary of the characteristics of each level). the trends indicate areas of recent growth in the healthcare (3, 4, 7, 8,
(8) Level descriptor – this involved examining the different types of 9, 18, 20, 38, 39 and 41), oil and gas (1, 12, 13, 24, 26, 27, 28, 34 and
descriptors for the maturity levels used in the study sample (e.g., 35) and the construction industry (10, 11, 25, 30, 31 and 37).
‘pathological’, ‘uncommitted’ and ‘amoral calculators’).
(9) Aims and objectives of using the maturity model – this category was
included in order to capture the main drivers or intentions for 5.2. Methods used to develop, evaluate and apply the maturity model
using the maturity model (e.g., assessing overall safety manage-
ment or risk perceptions; assessing safety leadership). In order to A wide variety of methods were used in order to develop, evaluate
generate a set of codes which could be used to categorise the and apply the maturity models in the sample (Table 4). Accordingly, the
reasons for using the model, each author independently compiled most used method (30 in total) was where the developer sought in-
a list of safety culture related keywords based on our individual formation from previous literature in order to develop the maturity
reading of the study sample. The authors then met and agreed models. Some studies (1, 4, 5, 8, 12, 14, 15, 19, 31, 32 and 34) used
upon a final set of 16 keywords; literature mixed with other development methods, e.g., Parker et al.
(10) Maturity model reliability and validity – this involved checking (2006, 1) used literature and interview and Kirk et al. (2007, 8) used
which of the study sample had carried out an assessment of mature literature review, interview and consultation with experts (where safety
model component reliability (e.g., by calculating values of internal and risk management experts are consulted for comments and opinion
reliability using Cronbach’s). The different types of validity on maturity model by developer).
checked as well as validity testing are described in Table 2. Questionnaires were the most common method used to evaluate and
apply the maturity model. Four publications used comparison to eval-
uate the maturity model, e.g., Gordon et al. (2007, 6) presented 2
5. Findings maturity models, one maturity model to assess a safety culture by
questionnaire and one to assess the safety culture by interview and both
A total of 2026 publications were retrieved from the databases de- maturity models were evaluated their results by comparison. A com-
scribed in Section 4.1. Duplicate publications were removed and a total bination of two or more methods (e.g., interviews and questionnaires),
of 135 (67 publications in academic journals, 68 reports and other was used to apply 7 maturity models. Tappin et al. (2015, 2) for ex-
publications in grey literature) were selected by reading the title, ab- ample, used semi-structured interviews combined with documentary
stract and key words. The remaining 135 publications were filtered out analysis of occupational safety and health (OSH) systems in order to
using the set of criteria, described in Section 4.2. A total of 41 pub- assess OSH maturity levels.
lications were selected using the criteria for detailed review. Twenty-
two publications were in academic journals and the remaining 19 were
grey literature, including 5 reports. Ten publications were conference 5.3. Type of publication
papers. In addition, the search yielded 4 theses (one PhD and three
Masters Theses). Appendix A lists the complete set of review publica- Table 5 shows the distribution of the four types of publications
tions. Table 3 is summary of the results of applying the framework to among the sample. Maturity model development is the most common
the 41 studies. In subsequent Sections of the paper, we refer to the type of publication (33 out of 41), whereas the number of publication of
studies with reference to their number in Table 3 (e.g., article 6 is maturity model application (4 out of 41), validation (2 out of 41) and
Gordon et al., 2007). reliability (2 out of 41) was smaller.

Table 2
Types of validity and testing (Trochim, 2006).

Type of validity Definition Testing

Face Refers to the degree to which a measurement tool subjectively appears to To assess the face validity of a safety culture measure, it would be sent the
measure the construct that it is supposed to measure (e.g., safety culture). In measure to a carefully selected sample of experts on safety culture measure and
other words, face validity is when a measurement appears to do what it they all reported back with the judgment that the measure appears to be a good
claims to do measure of safety culture
Content Refers to how accurately a measurement tool taps into the various aspects of If a particular measurement tool is designed to measure safety culture, a group
the specific construct in question (e.g., safety culture). In other words, do the of experts on safety culture would evaluate each dimension chosen to measure
questions really measure the construct in question safety culture and provide an opinion on how well each dimension taps into
measuring of safety culture
Convergent Examines the degree to which a measurement is similar to (converges on) Measurement tool should be able to show a correspondence or convergence
another measurement that it theoretically should be similar to between similar safety culture
Discriminant Examine the degree to which measurement is not similar to (diverges from) Measurement tool should be able to discriminate between dissimilar safety
another measurement that it theoretically should be not be similar to culture

196
 Table 3
Applying the categorisation framework.

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

1. A framework for understanding the Safety Science United Oil and gas Literature review and Focus group Questionnaire or X Level 1: Pathological
development of organisational Kingdom interview matrix
A.P. Goncalves Filho, P. Waterson

safety culture
Parker et al. (2006)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

2. An implementation evaluation of a Applied Ergonomics New Zealand Non-specific domain Literature review Interview and Interview and X Level 1: Pathological
qualitative culture assessment tool documental analysis documental
Tappin et al. (2015) analysis
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

3. Assessment of safety culture Healthcare Quarterly Canada Healthcare Questionnaire Questionnaire Questionnaire X Level 1: Pathological
Law et al. (2010)
Level 2: Reactive
Level 3: Bureaucratic
Level 4: Proactive
Level 5: Generative

197
4. Assessing patient safety culture in Journal of Primary Health Care New Zealand Healthcare Literature review and Not reported Matrix and work X Level A
New Zealand primary care: a pilot meeting group group
study using a modified Manchester
Patient Safety Framework in
Dunedin general practices
Wallis and Dovey (2011)
Level B
Level C
Level D
Level E

5. Measurement of food safety culture Food Control Canada Food Literature review and Questionnaire Questionnaire X Level 1: Doubt
using survey and maturity profiling consultation with
tools expert
Jespersen et al. (2016)
Level 2: React
Level 3: Know
Level 4: Predict
Level 5: Internalize

6A. Measuring safety culture in a Safety Science France Air traffic control Literature review Questionnaire Questionnaire X Level 1: Emerging
research and development centre: A
comparison of two methods in the
Air Traffic Management domain
Gordon et al. (2007)
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
Level 5: Continually
improve
Safety Science 105 (2018) 192–211

(continued on next page)

 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

6B. Measuring safety culture in a Safety Science France Air traffic control Literature review Interview Interview X Level 1: Uncertainty
research and development centre: A
comparison of two methods in the
A.P. Goncalves Filho, P. Waterson

Air Traffic Management domain

Gordon et al. (2007)
Level 2: Core
Level 3: Extended
Level 4: Leading Edge
Level 5: World Class

7. Patient Safety Culture Improvement Healthcare Quarterly Canada Healthcare Literature review Interview Matrix X Level 0: Pathological
Tool: Development and Guidelines
for Use
Fleming and Wentzell (2008)
Level 1: Reactive
Level 2: Calculative
Level 3: Proactive
Level 4: Generative

8. Patient safety culture in primary care: Quality and Safety in Health United Healthcare Literature review, Interview and focus Matrix X Level 1: Pathological
developing a theoretical framework Care Kingdom consultation with group
for practical use expertise and interview
Kirk et al. (2007)
Level 2: Reactive

198
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

9. Safety culture assessment in Quality and Safety in Health United Healthcare Focus group Focus group Matrix X Level 1: Pathological
community pharmacy: Care Kingdom
development, face validity, and
feasibility of the Manchester Patient
Safety Assessment Framework
Ashcroft et al. (2005)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

10. System dynamics modelling of Engineering, Construction and Thailand Construction Literature review No reported Questionnaire X Level 1:Uncommitted
construction safety culture Architectural Management
Mohamed and Chinda (2010)
Level 2: Drifters
Level 3: Improvers
Level 4: Award
winners
Level 5: World Class

11. Research on the Maturity of Real Journal of Applied Sciences China Construction Literature review Not reported Questionnaire X Level 1: Emerging
Estate Enterprises Safety Culture
Zhang et al. (2013)
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
(continued on next page)
Safety Science 105 (2018) 192–211
 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

Level 5: Continually
improve

12. Safety culture maturity model Offshore technology report n. United Oil and gas Literature review and Interview No reported X Level 1: Emerging
A.P. Goncalves Filho, P. Waterson

Fleming (2001) 2000/049 for Health and Safety Kingdom focus group
Executive
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
Level 5: Continually
improve

13. A safety culture maturity model for Safety Science Brazil Oil and gas Literature review Interview and Questionnaire X Level 1: Pathological
petrochemical companies in Brazil questionnaire
Goncalves Filho et al. (2010)
Level 2: Reactive
Level 3: Bureaucratic
Level 4: Proactive
Level 5: Generative

14. Incorporating organisational safety Ergonomics New Zealand Utilities company Literature review, Not reported Interview and X Level 1: Pathological
culture within ergonomics practice interview and documental
Bentleya and Tappin (2010) documental analysis analysis
Level 2: Reactive
Level 3: Calculative

199
Level 4: Proactive
Level 5: Generative

15. Cultural Maturity Model: Health and The Seventh SPE International Canada Non-specific domain Literature and Workshop Matrix X Level 1: Documenting
Safety Improvement through Conference Calgary, Alberta, consultation with
Involvement Canada experts
Fleming and Meakin (2004)
Level 2: Controlling
Level 3: Engaging
Level 4: Participating
Level
5:Institutionalising

16. A Tool to Diagnose Culture in Food Report from Greenstreet United Food Literature review Not reported Documental X Level 1: Amoral
Business Operators Berman Ltd for the Food Kingdom analysis and calculators
Wright et al. (2012) Standards Agency, Reference observation
GSB CL2567 R1 V6 FCA
Level 2: Dependent
Level 3: Doubters
Level 4: Proactive
compliers
Level 5: Leaders

17. Enterprise development through the International Journal of Thailand Non-specific domain Literature review, Questionnaire and Questionnaire X Level 1: Pathological
safety culture Management and Enterprise consultation with consultation with
Phusavat et al. (2015) Development experts, documental expert
analyses and
questionnaire
Level 2: Reactive
Level 3: Calculative
(continued on next page)
Safety Science 105 (2018) 192–211
 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

Level 4: Proactive
Level 5: Generative

18. Understanding Safety Culture in Journal Patient Safety Canada Healthcare Application of the See Publication Interview, Focus X Level 1: Pathological
A.P. Goncalves Filho, P. Waterson

Long-Term Care: A Case Study maturity model number 7 group and

Halligan et al. (2014) developed in questionnaire
Publication 7
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

19. Qualitative evaluation of a The Journal of Occupational New Zealand Utilities company Literature review, Not reported Interview; focus X Level 1: Pathological
framework for understanding the Health and Safety Australia and interview and group and
development of organisational New Zealand documental analysis documental
safety culture analysis
Bentleya and Tappin (2008)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

20. Patient safety dialogue: Evaluation Journal of Patient Safety Sweden Healthcare Literature review Not reported Interview and X Level 1: Pathological
of an intervention aimed at documental
achieving an improved patient analysis

200
safety culture
Ohrn et al. (2011)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

21. A preliminary study on safety Proceedings of 2011 China Coal mining Literature review Not reported Not reported X Level 1: Initial
culture maturity model of Chinese International Conference on
coal mine Management Science and
Xu and Pan (2011) Intelligent Control
Level 2: Managed
Level 3: Preventive
Level 4: Matured
Level 5: Optimized

22. Development of safety culture Proceedings of 2011 South Korea Nuclear power Literature review Not reported Not reported X Level 1: Pathological
assessment model using safety International Topical Meeting
culture maturity model and 4P-4C on Probabilistic Safety
matrix Assessment and Analysis
Sheen and Chung (2011)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

23. Application of Safety Maturity Transactions of the Korean South Korea Nuclear power Literature Review Not reported Not reported X Level 1: Pathological
Model and 4P-4C Model in Safety Nuclear Society Autumn
Culture Assessment Meeting
Choi et al. (2010)
(continued on next page)
Safety Science 105 (2018) 192–211
 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
A.P. Goncalves Filho, P. Waterson

Level 5: Generative

24. Towards a mature safety culture Symposium series n. 148 of the United Oil and gas Application of the See Publication Workshop X Level 1: Emerging
Lardner et al. (2001) Institution of Chemical Kingdom maturity model number 12
Engineers developed in
Publication 12
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
Level 5: Continually
improve

25. Health and Safety Culture Report 2014 of the RMIT Australia Construction Literature review Not reported Not reported X Level 1: Pathological
Lingard et al. (2014) University – Centre for
Construction Work Health and
Safety Research
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

26. Safety Culture Maturity and Risk XVIII International Conference Brazil Oil and gas, footwear Application of the See Publication Questionnaire X Level 1: Pathological

201
Management Maturity in Industrial on Industrial Engineering and Manufacturing and maturity model number 13
Organizations Operations Management infrastruture developed in
Goncalves et al. (2012) Publication 13
Level 2: Reactive
Level 3: Bureaucratic
Level 4: Proactive
Level 5: Generative

27. Investigating employee perceptions Safety Science United Oil and gas Application of the Questionnaire See Publication X Level 1: Pathological
of a framework of safety culture Kingdom maturity model number 1
maturity developed in
Lawrie et al. (2006) publication 1
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

28. Lessons Learned from Applying International Journal of Thailand Oil and gas Literature review Questionnaire Questionnaire X Level 1: Pathological
Safety Culture Maturity Model in Synergy and Research
Thailand
Vongvitayapirom et al. (2013)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

29. A Framework for Assessing Safety Master’s thesis in Chemical Sweden Non-specific domain Literature review Accident analysis Not reported X Level 1: Pathological
Culture Engineering
Gotivall (2014) University of Technology
Gothenburg
(continued on next page)
Safety Science 105 (2018) 192–211
 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
A.P. Goncalves Filho, P. Waterson

Level 5: Generative

29. A Framework for Assessing Safety Master’s thesis in Chemical Sweden Non-specific domain Literature review Accident analysis Not reported X Level 1: Pathological
Culture Engineering
Gotivall (2014) University of Technology
Gothenburg
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

30. Criteria for the Development of a Proceedings of the International Australia Construction Literature review Not reported Questionnaire X Level 1: Emerging
Safety Culture Maturity Model for Symposium on Information
the Construction Industry Technologies in Safety
McGeorge et al. (2011) Management of Large Scale
Infrastructure Projects
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
Level 5: Continually
improve

202
31. Development of the people first Research Report RR895 for the United Construction Literature review, Interview Matrix X Level 1: Pathological
toolkit for construction small and Health and Safety Executive Kingdom meeting discussion,
medium sized enterprises consultation with
Lunt et al. (2011) expert, focus group and
interview
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

32. Tool to be used to survey and Thesis in Safety, Health and Norway Rail Literature review, Workshop Questionnaire X Level 1: Pathological
improve safety culture in the Environment University of interview and focus
European railway industry Science and Technology group
Bergersen (2003)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

33. Evaluation of safety culture maturity Proceedings of the 2012 JKUAT Kenya Education Literature review Not reported Questionnaire X Level 1: Emerging
levels of the universities in Kenya Scientific, Technological and (Universities)
Njeru et al. (2012) Industrialization Conference
Level 2: Managing
Level 3: Involving
Level 4: Cooperating
Level 5: Continually
improve

(continued on next page)

Safety Science 105 (2018) 192–211
 Table 3 (continued)

Title/Author Source Country of Domain Method Type of publication Level descriptor

origin (Wendler, 2012)

D E A D A V R

34. Developing safety culture Final project report Saint Mary's Canada Oil and gas Literature review and Questionnaire, scored Interview X Level 1: Pathological
measurement tools and techniques University Halifax Nova Scotia consultation with card, interview and
based on site audits rather than experts interrater reliability
A.P. Goncalves Filho, P. Waterson

questionnaires
Fleming (2007)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

35. Safety culture maturity in upstream: Proceedings of 2013 Thailand Oil and gas Literature review Questionnaire Questionnaire X Level 1: Pathological
oil and gas industry in Thailand International Conference on
Vongvitayapirom and Phusavat Technology Innovation and
(2013) Industrial Management
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

36. Safety culture in a service company: Journal of Occupational Health The Non-destructive Application of the Questionnaire See Publication X Level 1: Pathological
a case study and Safety – Australia and New Netherlands research maturity developed in number 1
Guldenmund (2008) Zealand publication 1
Level 2: Reactive
Level 3: Calculative

203
Level 4: Proactive
Level 5: Generative

37. Meaningful and effective Construction Management and Australia Construction Literature review Not reported Documental X Level 1: Pathological
consultation and the construction Economics analyses
industry of Victoria, Australia
Ayers et al. (2013)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

38. A multidimensional approach to Proceedings of the International Canada and Healthcare Application of the Questionnaire See Publication X Level 1: Pathological
studying safety culture in health Conference of Healthcare United maturity model number 8
care Systems and Ergonomics and Kingdom developed in
Law et al. (2008) Patient Safety Publication 8
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

39. Understanding and Changing the Thesis submitted for the degree Canada Healthcare Application of the Questionnaire See Publication X Level 1: Pathological
Patient Safety Culture in Canadian of Doctor of Philosophy maturity developed in number 3
Hospitals University of Toronto publication 3
Law 2011)
Level 2: Reactive
Level 3: Calculative
Level 4: Proactive
Level 5: Generative

(continued on next page)

Safety Science 105 (2018) 192–211
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

5.4. Structure

Bureaucratic

Bureaucratic
Level 1: Pathological

Level 1: Pathological
Generative

Generative
Proactive

Proactive
Reactive

Reactive
Level descriptor

The maturity models have a conceptual overlap, such as aspect at

discrete stages or maturity levels, with a description of characteristic

2:
3:
4:
5:

2:
3:
4:
5:
performance at various levels of maturity. The following structure is

Level
Level
Level
Level

Level
Level
Level
Level
common in maturity models to assess safety culture: (a) a number of
levels (typically 5); (b) a descriptor for each level which are describe
following Section 5.5; (c) a generic description or summary of the
R
Type of publication

characteristics of each level as a whole; (d) multi-dimensional; and, (f) a

(Wendler, 2012)

description of each activity as it might be performed at each maturity

V

level.
A

5.5. Level descriptor

D

More than two-thirds of maturity models (68%, n = 28) were de-

veloped based on Westrum’s model and adopted pathological, reactive,
Questionnaire

Questionnaire

calculative or bureaucratic, proactive, and generative or sustainable

levels (Table 6). Filho et al. (2010, 13) relabelled ‘generative’ as ‘sus-
tainable’ in order to make it more understandable in their country. The
A

‘bureaucratic’ level was also retained in maturity models developed by

Law et al. (2010, 3) and Filho et al. (2010, 13). Six maturity models
were developed based on the CMM and adopted ‘emerging’, ‘managing’,
consultation with
Focus group and
See Publication

‘involving’, ‘cooperating’ and ‘continually improving’ levels. Eight ma-

number 13

turity models adopted different levels from Westrum’s model and CMM.
Wright et al. (2012, 16) adopted the levels ‘amoral calculators’, ‘de-
expert

pendent’, ‘doubters’, ‘proactive compliers’ and ‘leader’ because these

E

labels were seen as more suitable for the domain of food safety. Simi-
larly, Jespersen (2016, 5) adopted the levels ‘doubt’, ‘react’, ‘know’,
‘predict’ and ‘internalize’ in their models. The maturity model devel-
Application of the
maturity model

oped by Lunt et al. (2011, 31) combined elements of the Westrum

Publication 13
developed in

Not reported

model and CMM in order to make it more suitable for use in the con-
Method

struction industry. Consequently, the names for each level were

changed to ‘starting blocks’, ‘getting going’, ‘walking’, ‘running’ and
D

‘sprinting’, respectively.
Non-specific domain

Gordon et al. (2007) developed two maturity models in same pub-

lication (6), one based on CMM (6A) and other one based on the DNV
(Det Norske Veritas) auditing system in which the levels adopted were
Healthcare

‘uncertainty’, ‘core’, ‘extended’, ‘leading edge’ and ‘world class’ (6B).

Domain

Fleming and Meakin (2004, 15) adapted the maturity model from the
model presented in the ‘Changing Minds’ toolkit (Step Change in Safety,
2017). Their model includes the following levels: ‘documenting’, ‘con-
trolling’, ‘engaging’, ‘participating’ and ‘institutionalising’. Mohamed
Philippines
Country of

and Chinda (2010, 10) adopted the EFQM (European Foundation for
Note: D = Development E = Evaluation A = application V = Validity R = Reliability.
origin

China

Quality Management) Excellence model, consequently, their maturity

The

model includes the following levels: ‘uncommitted’, ‘drifters’, ‘im-

provers’, ‘award winners’, and ‘world-class’.
Journal of Sciences, Technology
Thesis for the degree of Master
in Politecnico di Milano

5.6. Aims and objectives of using the maturity model

and Arts Research

The most common aim was general safety management assessment,

which appeared in 33 (80%) publications followed by assessment of
communication about safety (68%), management commitment to safety
Source

(68%) and safety training (66%). Assessing organisational learning

accounted for more than half of the publications in our sample (56%),
whilst accident/incident reporting (37%), safety prioritisation (37%),
healthcare providers in hospitals in
40. Towards the definition of an OHS
management maturity model and

Jabonete and Concepcion (2016)

employee involvement (34%) and accident/incident analysis (34%)

made up more than a third of the publications. Organisational trust
41. Perceived safety culture of

(22%), safety audit (22%), periodic review (20%) represented ap-

proximately a fifth of the sample, followed by a smaller proportion of
publications which focused on risk perceptions (17%), risk analysis
assessment tool

the Philippines

(15%) safety leadership (15%) and work pressure (12%).

Table 3 (continued)

Chen (2016)
Title/Author

5.7. Maturity model reliability and validity

The 41 publications were specifically checked regarding evidence in

terms of the assessment of reliability and validity were carried out

204
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

8
7
F 7
r
6
e 5
5
q 4 4
u 4
3 3 3 3
e 3
n 2 2
2
c 1 1 1 1 1
1
y
0
2000 2001 2003 2004 2005 2006 2007 2008 2010 2011 2012 2013 2014 2015 2016

Year of publication

Fig. 3. Year of publication.

(Table 7). Eleven (27%) studies carried out some form of reliability Table 5
assessment (e.g., where a maturity model had involved a questionnaire Type of publication (Wendler, 2012).
and assessed reliability using Cronbach’s Alpha). Fourteen (34%) pub-
Type of publication Publication Frequency
lications carried out both content validity (e.g., by assessing how close
the maturity model aligned with other measures or indices of safety Maturity model 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 33
such as an employee safety culture perception survey) and face validity development 15, 16, 17, 19, 20, 21, 22, 23, 25, 28, 29,
(e.g., by asking employees to assess the extent to which the components 30, 31, 32, 33, 34, 35, 37, 41
Maturity model 18, 24, 26, 40 4
of the model mapped on to aspects of safety in their organisation). Four application
publications carried out convergent validity (i.e., testing whether ma- Maturity model 36, 39 2
turity model components supposed to be either related or unrelated, in validation
fact, related or unrelated – Campbell and Fiske, 1959) and two carried Maturity model 27, 38 2
reliability
out discriminant validity. Only one study (34) provided evidence about
the procedure used to assess reliability and all four types of validity.
Eighteen (44%) of the publications report no evidence that assessments been a steady growth in the use of maturity models to assess safety
of either reliability or validity were carried out. culture across a wide range of industries. The period 2008–2011 re-
presents a peak of activity in using maturity models with steady use
6. Discussion characterising the succeeding years (Fig. 3). Similarly, articles drawn
from the Anglophone countries are the most frequent ones using ma-
6.1. Growth and expansion of maturity models turity models, but with some evidence that their use is consistently
wide in terms of other countries and global regions. In terms of appli-
Our findings demonstrate that over the last two decades there has cation domain, construction, oil and gas and healthcare are the most

Table 4
Methods used to develop, evaluate and apply the maturity model.

Method Develop Evaluate Apply

Literature 1, 2, 4, 5, 6, 7, 8, 10, 11, 12, 13, 14, 15, 16, 17, 19, 20, 21, 22, 23,
25, 28, 29, 30, 31, 32, 33, 34, 35, 37
Interview 1, 8, 14, 19, 31, 32 2, 6, 7, 8, 12, 13, 31, 34 2, 6, 14, 18, 19, 20, 34
Matrix 1, 4, 7, 8, 9, 15, 31
Consultation with expert 5, 8, 15, 17, 31, 34 41
Focus group 9, 12, 31, 32 1, 8, 9, 36, 41 4, 12, 18, 19
Questionnaire 3, 17, 34 3, 5, 6, 13, 17, 27, 28, 34, 35, 36, 38, 3, 5, 6, 10, 11, 13, 17, 18, 26, 28, 30,
39 32, 33, 35, 40, 41
Documental analysis 17, 19 2 2, 14, 16, 19, 20, 37
Meeting discussion 4, 31
Observation 16
Scored Card 34
Inter-rater reliability 34
Workshop 15, 32 24
Triangulation 2, 4, 14, 16, 18, 19, 20
Accident analysis 29
Comparison 6, 36, 34, 39
No reported 41 4, 10, 11, 14, 16, 19, 20, 21, 22, 23, 12, 21, 22, 23, 25, 29
25, 30, 33, 37

205
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

Table 6 growth of specific types of tools associated with maturity models since
Level descriptor. the beginning of the millennium, as well as application domains in
which they are used and links back to earlier work in the 1970s, 1980s
Publication Level descriptor Frequency
and 1990s (Section 3.2).
1, 2, 3, 7, 8, 9, 13, 14, 17, 18, 19, 20, Level 1: Pathological 28
22, 23, 25, 27, 26, 28, 29, 32, 34, Level 2: Reactive
36, 35, 37, 38, 39, 40, 41 Level 3: Calculative or
6.2. The theoretical status of maturity models
bureaucratic
Level 4: Proactive
Level 5: Generative 6.2.1. Maturity models as tools for organisational development
6A, 11, 12, 24, 30, 33 Level 1: Emerging 6
The range of methods used to develop, evaluate and apply maturity
Level 2: Managing models, alongside the finding that most of the publications in the re-
Level 3: Involving view sample fall into ‘development’ (i.e., where the main objective was
Level 4: Cooperating to build a maturity model, rather than necessarily describe its appli-
Level 5: Continually
cation or evaluation – Wendler, 2012), would seem to provide some
improve
evidence maturity models illustrate what Guldenmund’s (2010, 2016)
4 Level A 1
characterised as a ‘pragmatic or experience-based’ approach with regard
Level B
Level C to safety culture. This ‘pragmatism’ extends to numerous attempts
Level D which were made to change the various elements of the maturity
Level E models and the variety of level descriptors used within the study sample
5 Level 1: Doubt 1 (Table 6). Similarly, the fact that the main driver for using the model in
Level 2: React most of our sample was assessment of general levels of safety and core
Level 3: Know elements of safety culture (e.g., communication, management com-
Level 4: Predict
mitment) as opposed to detailed aspects (e.g., levels of organisational
Level 5: Internalize
trust, risk perceptions) might be seen to imply that maturity models are
6B Level 1: Uncertainty 1
seen as ways of assessing the ‘wider, bigger picture’ of how safety might
Level 2: Core
Level 3: Extended
work in a specific organisational culture (c.f., ‘total safety culture’ –
Level 4: Leading Edge Geller, 1994). Finally, the low numbers of studies which report details
Level 5: World Class of reliability and validity, might be seen to imply that the ‘process’ of
10 Level 1: Uncommitted 1 using a maturity model is more important that the actual ‘outcome’.
Level 2: Drifters From this point of view, maturity models might be construed as tools
Level 3: Improvers for ‘organisational development’, as compared to a means of ‘mea-
Level 4: Award winners
suring’ safety culture, again reinforcing their ‘pragmatic ‘advantages as
Level 5: World Class
compared to survey instruments (‘analytical/psychological’ approaches
15 Level 1: Documenting 1
– Guldenmund, 2010, 2016, ‘culture as causal attitude’ – Silbey, 2009)
Level 2: Controlling
Level 3: Engaging
and qualitative methods such as interviews or ethnographies (‘inter-
Level 4: Participating pretive/anthropological’ approaches Guldenmund, 2010, 2016; ‘culture
Level 5: Institutionalising as emergent/indeterminate’ – Silbey, 2009).
16 Level 1: Amoral 1
calculators
Level 2: Dependent 6.2.2. Static vs. dynamic accounts of safety culture
Level 3: Doubters
Aside from the resource-intensive drawbacks (i.e., the time taken to
Level 4: Proactive
compliers organise discussion groups or workshops and analyse/evaluate results)
Level 5: Leaders of using maturity models, it may also be the case that maturity models
21 Level 1: Initial
may encourage a misleading picture of how safety culture operates in
Level 2: Managed organisations. Maturity models imply that organisations make steady
Level 3: Preventive progress on a journey that takes them from low levels of safety towards
Level 4: Matured an optimal state. This might in some ways be interpreted as a ‘Whig-
Level 5: Optimized 1
like’ interpretation of cultural improvement1, when in reality values
31 Level 1: Starting blocks 1 and beliefs held by employees about safety may spontaneously wax,
Level 2: Getting going
wane and in some cases relapse over short periods of time (Turner,
Level 3: Walking
Level 4: Running 1978; Busch, 2016). Vincent and Amalberti (2016) for example, refer to
Level 5: Sprinting safety as a ‘moving target’ as opposed to a fixed entity. Maturity models
by their very nature may tend to overemphasise aspects of a static view
of safety and may underplay the subtlety with which safety may dy-
common areas in which maturity models are applied. To some extent namically vary according to a host of influencing factors (resources,
this might be explained by the fact that industries such as oil and gas workload, policy changes, economic and political influences –
have a well-established track record in using specific type of models and Rasmussen, 1997). Internal/external benchmarking of the outcomes of
developing tools which have influenced other attempts to use maturity safety culture assessment may also play to hidden agendas within or-
models (e.g., the Hearts and Minds toolkit – Hudson and Willekes, ganisations and potential lead to manipulation and ‘gaming’ (e.g., the
2000). Likewise, the growth of maturity models in the healthcare sector misuse of safety culture in ‘league tables’ – Shorrock, personal com-
might be attributable to the popularity of the Manchester Patient Safety munication).
Framework (MaPSaF) and its promotion up until 2012 within the UK by
bodies such as the National Patient Safety Agency (NPSA), alongside
other trends such as increased efforts to assess patient safety culture and 1
The term ‘Whig’ refers to a 17th Century political faction. A Whig view of history
drives to accelerate the integration of human factors within healthcare presents the past as an inevitable progression towards ever greater liberty and enlight-
(Waterson, 2014; Waterson and Catchpole, 2016). Fig. 4 illustrates the enment.

206
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

Table 7
Validity and reliability.

Publications with evidence that Frequency Publications without evidence that was carried out Frequency
was carried out

Reliability 2, 3, 13, 17, 27, 28, 31, 34, 35, 11 1, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 18, 19, 20, 21, 22, 23, 24, 30
38, 41 25, 26, 29, 30, 32, 33, 36, 37, 39, 40
Content validity 1, 3, 5, 7, 8, 9, 12, 15, 17, 29, 14 2, 4, 6, 10, 11, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 27
31, 32, 34, 41 30, 33, 35, 36, 37, 38, 39, 40
Face validity 1, 3, 5, 7, 8, 9, 12, 15, 17, 29, 14 2, 4, 6, 10, 11, 13, 14, 16, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 27
31, 32, 34, 41 30, 33, 35, 36, 37, 38, 39, 40
Convergent validity 6, 34, 36, 39 4 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 37
23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 35, 37, 38, 40, 41
Discriminant validity 2, 34 2 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 39
23, 24, 25, 26, 27, 28, 29, 30, 31, 33, 35, 36, 37, 38, 39, 40, 41
Reliability + validity (face, content, convergent 34 1 4, 10, 11, 14, 16, 18, 19, 20, 21, 22, 23, 24, 25, 26, 30, 33, 37, 40 18
and discriminant)

Fig. 4. Maturity model development over time.

6.2.3. The absence of theory in maturity models ‘pathological’ or ‘bureaucratic’ states, stages of phases) which either
In common with other well-known frequently used constructs in predicts the nature and course of improvement or set out a steps on the
safety science and human factors (e.g., the systems approach – road to improvement. In the absence of a theoretical basis it is difficult
Waterson, 2009), we found few attempts within our review sample to to see how this can be justified. Theories by their very nature are used
articulate what type of theory of safety culture was being used. Authors to predict behaviour and levels of safety, as well as supporting com-
tended to offer a definition of safety culture, but made few attempts to parison with other, sometimes unrelated, phenomena (so-called ‘dis-
link their work to theoretical frameworks and approaches which are ciplined imagination – Weick, 1989). Sutton and Staw (1995) point to a
common within safety science and human factors (e.g., the theory of number of misunderstandings about the use of theory in the social
High Reliability Organization (HRO) – Weick and Sutcliffe, 2007; sciences including the view that data on its own does not constitute a
Normal Accident Theory – Perrow, 1984). In many ways this criticism contribution to theory. In many respects the findings from our review
might be levelled at much of the literature on safety culture (Flin, 2007; suggest that using a maturity model and reporting the outcomes from
Reiman and Rollenhagen, 2014), however the use of maturity models this use, does not in itself constitute a theory of safety. Studies which
appears to imply that there is some mechanism underlying improve- report the use of maturity models need to be more explicit in stating the
ment and that particular components of cultural development (e.g., type of theoretical stance (e.g., epistemology, ontology – Haavik, 2014)

207
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

they are adopting and move away from a non-committal approach to- be characterised as ‘experimenting’ with maturity models, particularly
wards theory. Part of this involves providing more details of the process as a result of disappointing efforts using other methods for measuring
of change and the mechanisms which underpin safety improvement. safety culture (e.g., low response rates from questionnaire studies).
Many of these ‘experiments’ appear as internal reports or other forms of
6.3. Maturity models – academic and practitioner perspectives informal feedback within these organisations and hence do not appear
in published form. Maturity models, we would argue, focus on the
6.3.1. Maturity models as a ‘cool medium’ process aspects of safety culture and in line with other areas of safety
In many respects the findings from our review of the use of maturity science there is a need for more case study or ‘experienced-based’ ac-
models to assess safety culture also touch on a number of debates which counts of how they are used and what sorts of organisational outcomes
have surrounded safety culture, as well as the assessment of work-based arise from their usage and how this might change and develop over
attitudes and the wider world of safety over the past few decades. Many time (Hale, 2006):
of these debates are framed in terms of binary distinctions and opposing
‘Maturity scales … represent the steps leading from the pathological
positions. Wilson (1987) provides a typical example in his account of
state or face to the smiling, generative one, but we know little or
the various methods (e.g., questionnaire instruments, semi-structured
nothing about whether it is indeed possible to mount that scale, and
discussions with employees) which are used to assess job attitudes and
if so how. We have not defined whether we should be trying to shift
inform job redesign decisions. This time drawing on the work of Mar-
companies all the way from one to another end of the scale, or that
shall McLuhan (1964), Wilson argues that ‘formal job attitude instruments
we would be happy if the bulk of companies made it to the halfway
used alone are 'hot media', in the sense of being well-filled with data but low
point and became ‘calculative’. Longitudinal research studies of
in participation. Informal discussion methods could be seen then as a 'cool
companies to plot such shifts and how to facilitate them are des-
medium', high in participation or completion by the audience’ (Wilson,
perately needed. We might also need to expand our discipline base
1987, p. 385). Wilson further argues that informal methods can en-
to do it.’ (Hale, 2006)
courage employees to develop ‘ownership’ of the outcomes from dis-
cussions and workshops (Shipley, 1987) as well as providing detailed In terms of the practice of safety culture assessment, would also
insights into the reasons why particular attitudes may be shared or argue that the decision to use a maturity model to assess safety culture
dominant at one particular time. It might be argued that maturity instead of other type of methods (e.g., focus groups, questionnaire
models, as compared to other methods of measuring safety culture (e.g., surveys), should not be seen in terms of a straightforward question of
questionnaires) fall under Wilson’s category of ‘cool media’ and a si- either/or. Rather, we would encourage potential users of maturity
milar set of trade-offs may be at play (e.g., promoting participation as models to view them pragmatically and to acknowledge the value of
compared to providing more scientific or ‘objective’ measures of safety using them in combination with other methods (see for example Kirwan
culture). In some cases, it may be that from the safety practitioner’s and Shorrock, 2014, for a case study description of how this applies).
point of view it may be that in using a maturity the process may be more Finally, the strengths and weaknesses of maturity models and their
important that the actual outcome. How practitioners use maturity application within the field of safety culture needs to be more widely
models, their goals and what they do with the results is something, acknowledged (e.g., compromises and trade-offs – Waterson et al.,
alongside other aspects of the day-to-day work of the safety profes- 2014; Amalberti, 2013).
sionals, which is worthy of further attention (Hale, 2006). From an academic or scientific point-of-view we would argue for
greater attention be given to the theoretical aspects and justification for
6.3.2. Flexibility and ‘bricolage’ using maturity models. As noted in Section 6.2.3 we would also suggest,
An additional strength of maturity models might be said to be the in common with many other authors, that more attention to theory is
fact that they are flexible and can be tailored to fit the specific re- also a need for the field of safety culture more widely. The findings from
quirements and objectives of organisations and their employees. Our our review showed that assessments of the reliability and validity of the
findings as they relate to level descriptors for example (Table 6), show use of maturity models to assess safety culture tend to be the exception,
that the types of maturity models used to assess safety culture varied a rather than the rule (44% of the sample did not report any evidence of
great deal. This flexibility and adaptability is in many ways similar to reliability or validity assessments). A key weakness of the maturity
other methods within safety science (e.g., Rasmussen’s work on the model approach may be that results obtained during one point in time
Accimap method of accident analysis and subsequent attempts to may not prove to be repeatable during another. The studies in our
‘remix’ his work – Waterson et al., 2017). This process might be seen as sample do not allow us to draw firm conclusions about the reliability,
akin to what the anthropologist Claude Levi-Strauss (1966) called validity and overall robustness of using maturity models. Further in-
‘bricolage’, that is ‘the construction [e.g., an artefact, narrative, tool] or vestigation probing deeper into these factors is likely to be one area for
creation from a diverse range of available things’ (Concise Oxford English future research. It seems likely however, that discussion of focus groups
Dictionary, 12th Edition, 2011). The process of bricolage is also very with different groups and/or members are unlikely to produce out-
much in keeping with the pragmatic approach towards safety culture comes which are difficult to compare and hence very difficult to assess
assessment. in terms of reliability.
We should also acknowledge, as mentioned in Section 6.2 of the
7. Conclusions, limitations and future work paper, that the flexibility afforded by maturity models may be a
strength rather than a weakness as compared to other methods of safety
In this paper we have attempted to review published literature culture assessment. We note in passing, however, that the reliability
which has made use of maturity models of one sort or another to assess and validity of other types of maturity model has been the subject of
safety culture over the last two decades or so. Our findings show on the some debate in the last few years (e.g., Bach’s criticism of the Capability
one hand growth in terms of the use of maturity models to assess safety Maturity Model – Bach, 1994; Finkelstein, 1992). The question of va-
culture, but also significant variation in the ways in which they are used lidity (e.g., how well do maturity assessments correlate with safety
and reported within the literature. A key limitation of our work, which outcome data) is however, possible to assess and, in common with other
was difficult to avoid, is that we focused on published research. types of safety culture assessment (Flin et al., 2000), is something that
Anecdotal evidence suggests that use of maturity models may be under- should be pursued in the future, at least for those organisations (e.g.,
reported. Our work in a variety of sectors (e.g., oil and gas, healthcare) those who did carry out validity/reliability assessment) who wish to
suggests that maturity models are widely used by organisations eager to pursue this goal. In conclusion, we would say that our understanding of
gain some insight into their safety culture. Much of this activity might the scientific and practice-oriented aspects of maturity models remains

208
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

relatively ‘immature’. Much more effort needs to be given over in the and safety improvement through involvement. In: The Seventh SPE
future to some of the gaps in our current understanding which have International on Health, Safety and Environment in Oil and Gas
been revealed by our literature review. Exploration, Calgary, Alberta, Canada.
16. Wright, M., Leach, P. and Palmer, G., 2012. A tool to diagnose
Acknowledgements culture in food business operators. Report from Greenstreet Berman
Ltd for the UK Food Safety Agency. Reference GSB CL2567 R1 V6
The work described in this paper was carried out whilst the first FCA.
author was an academic visitor in the ‘Human Factors and Complex 17. Phusavat, K., Vongvitayapirom, B. and Hidayanto, A.N., 2015.
Systems Group’ at Loughborough University during the period Enterprise development through the safety culture maturity model.
2016–2017. International Journal of Management and Enterprise Development,
We would like to acknowledge the advice and guidance given to us 14, 2, 89–102.
by the handling editor for Safety Science (Dr. Guldenmund), as well as 18. Halligan, M. H., Zecevic, A., Kothari, A. R., Salmoni, A. W. and
two anonymous reviewers. We believe the paper is greatly improved as Orchard, T., 2014. Understanding safety culture in long-term care:
a result of this advice. Finally, thanks to Dr. Steve Shorrock for some a case study. Journal of Patient Safety, 10, 192–201.
discussion on safety culture which helped shaped some of the conclu- 19. Bentley, T. and Tappin, D., 2008. Qualitative evaluation of a fra-
sions of our paper. mework for understanding the development of organisational
safety culture. The Journal of Occupational Health and Safety
Appendix A. Review sample Australia and New Zealand, 24, 3, 231–220.
20. Ohrn, A., Rutberg, H. and Nilsen, P., 2011. Patient safety dialogue:
1. Parker, D., Lawrie, M. and Hudson, P., 2006. A framework for Evaluation of an intervention aimed at achieving an improved pa-
understand the development of organisational safety culture. Safety tient safety culture. Journal of Patient Safety, 7, 4, 185–192.
Science, 44, 551–562. 21. Xu, W. and Pan, J., 2011. A preliminary study on safety culture
2. Tappin, D.C., Bentley, T.A. and Ashby, L.E., 2015. An im- maturity model of chinese coal mine. In: Proceedings of 2011
plementation evaluation of a qualitative culture assessment tool. International Conference on Management Science and Intelligent
Applied Ergonomics, 47, 84–92. Control (ICMSIC 2011), Beijing, China, pp. 245–247.
3. Law, M. P., Zimmerman, R., Baker and G. R., Smith, T., 2010. 22. Sheen, C. and Chung, D.W., 2011. Development of safety culture
Assessment of safety culture maturity in a hospital setting. assessment model using safety culture maturity model and 4P-4C
Healthcare Quarterly, 13, 110–115. matrix. In: Proceedings of 2011 International Topical Meeting on
4. Wallis, K. and Dovey, S., 2011. Assessing patient safety culture in Probabilistic Safety Assessment and Analysis, Volume 2, pp.
New Zealand primary care: a pilot study using a modified 2143–2152.
Manchester Patient Safety Framework in Dunedin general prac- 23. Choi, K.S., Lee, Y.E., Ha, J. T., Chang, H. S. and Kam, S. C., 2010.
tices. Journal of Primary Health Care, 3, 1, 35–40. Application of Safety Maturity Model and 4P-4C Model in Safety
5. Jespersen, L., Griffiths, M., Maclaurin, T., Chapman, B. and Culture Assessment. In: Transactions of the Korean Nuclear Society
Wallace, C. A., 2016. Measurement of food safety culture using Autumn Meeting, Jeju, Korea, pp. 1115–1116.
survey and maturity profiling tools. Food Control, 66, 174–182. 24. Lardner, R., Fleming, M. and Joyner, P., 2001. Towards a mature
6. Gordon, R., Kirwan, B. and Perrin, E., 2007. Measuring safety cul- safety culture. In: Symposium series n. 148 of the Institution of
ture in a research and development centre: A comparison of two Chemical Engineers (IChemE), pp. 635–642.
methods in the Air Traffic Management domain. Safety Science, 45, 25. Lingard, H., Zhang, R., Harley, J., Blismas, N. and Wakefield, R.,
669–695. 2014. Health and Safety. Culture Report 2014 of the RMIT
7. Fleming, M. and Wentzell, N., 2008. Patient safety culture im- University – Centre for Construction Work Health and Safety
provement tool: development and guidelines for use. Healthcare Research.
Quarterly, 11, 10–15. 26. Goncalves, A. P., Kanegae, G. and Leite, G., 2012. Safety culture
8. Kirk, S., Parker, D., Claridge, T., Esmail, A. and Marshall, M., 2007. maturity and risk management maturity in industrial organisations.
Patient safety culture in primary care: developing a theoretical In: XVIII International Conference on Industrial Engineering and
framework for practical use. Quality and Safety in Health Care, 16, Operations Management, pp. ID131.1-ID131.10.
313–320. 27. Lawrie, M., Parker, D. and Hudson, P., 2006. Investigating em-
9. Ashcroft, D.M., Morecroft, C., Parker, D. and Noyce, P. R., 2005. ployee perceptions of a framework of safety culture maturity.
Safety culture assessment in community pharmacy: development, Safety Science, 44, 259–276.
face validity, and feasibility of the Manchester Patient Safety 28. Vongvitayapirom, B., Sachakamol, P., Kropsu-Vehkapera, H. and
Assessment Framework. Quality and Safety in Health Care, 14, Kess, P., 2013. Lessons learned from applying safety culture ma-
417–21. turity model in Thailand. International Journal of Synergy and
10. Mohamed, S. and Chinda, T., 2010. System dynamics modelling of Research, 2, 1, 5–21.
construction safety culture. Engineering, Construction and 29. Gotivall, J., 2014. A Framework for Assessing Safety Culture.
Architectural Management, 18, 3, 266–281. Master’s thesis in Chemical Engineering. Department of Product
11. Zhang, F., Zhang, Y., Liu, J. and Luo, H., 2013. Reseach on the and Production Development of Chalmers University of
maturity of real estate enterprises safety culture. Journal of Applied Technology, Gothenburg, Sweden.
Sciences, 13, 11, 2019–2044. 30. McGeorge, D., Sunindijo, R.Y. and Zou, P.X.W., 2011. Criteria for
12. Fleming, M., 2001. Safety culture maturity model. Offshore tech- the Development of a Safety Culture Maturity Model for the
nology Report 2000/049, Health and Safety Executive, UK. Construction Industry. In Proceedings of the International
13. Goncalves Filho, A.P., Andrade, J.C.S. and Marinho, M.M.O., 2010. Symposium on Information Technologies in Safety Management of
A safety culture maturity model for petrochemical companies in Large Scale Infrastructure Projects, Hubei, China.
Brazil, Safety Science, 48, 615–624. 31. Lunt, J., Bennett, V., Hopkinson, J., Holroyd, J., Wilde, E., Bates, S.,
14. Bentley, T. and Tappin, D. 2010. Incorporating organisational Bell, N. and Webster, J., 2011. Development of the people first
safety culture within ergonomics practice. Ergonomics, 53, 10, toolkit for construction small and medium sized enterprises.
1167–1164. Research Report RR895 for the UK Health and Safety Executive.
15. Fleming, M. and Meakin, S., 2004. Cultural maturity model: health 32. Bergersen, C. E. B., 2003. Tool to be used to survey and improve

209
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

safety culture in the european railway industry. Thesis in Safety, Dekker, S.W.A., 2018. The Safety Anarchist. Routledge, London.
Health and Environment. Faculty of Social Sciences and Díaz-Cabrera, D., Hernández-Fernaud, E., Isla-Díaz, R., 2007. An evaluation of a new
instrument to measure organisational safety culture values and practices. Accid. Anal.
Technology Management, Department of Industrial Economics and Prevent. 39 (6), 1202–1211.
Technology Management at Norwegian University of Science and Eisenhardt, K., 1993. High reliability organizations meet high velocity environments:
Technology. Trondheim, Norway. common dilemma in nuclear power plants, aircraft carriers, and microcomputer
firms. In: Roberts, K.H. (Ed.), New Challenges Understanding Organizations.
33. Njeru, P. M., Mailutha, J. T., Gatebe, E. and Mburu, C., 2012. Macmillan, New York.
Evaluation of safety culture maturity levels of the Universities in Foster, P., Hoult, S., 2013. The safety journey: using a safety maturity model for safety
Kenya. In: Proceedings of the 2012 JKUAT Scientific, Technological planning and assurance in the UK coal mining industry. Minerals 3, 59–72.
Finkelstein, A., 1992. Software process immaturity model. SIGSOFT Software Engineering
and Industrialization Conference, Nairobi, Kenya, pp. 349–361. Notes.
34. Fleming, M., 2007. Developing safety culture measurement tools and Fleming, M., 2001. Safety culture maturity model. Offshore Technology Report 2000/
techniques based on site audits rather than questionnaires. Final project 049, Health and Safety Executive, UK.
Fleming, M., 2017. A practitioner’s guide to creating a high performance safety culture.
report Saint Mary's University in Halifax, Nova Scotia, Canada. http://
Keynote Presentation at the 12th International Symposium on Human Factors in
www.pr-ac.ca/files/files/133_FinalReport_12May07.pdf (August, 04, Organizational Design and Management (ODAM 2017), Banff, Canada, July
2016). 31st–August 3rd.
35. Vongvitayapirom, B. and Phusavat, K., 2013. Safety culture ma- Flin, R., 2007. Measuring safety culture in healthcare: a case for accurate diagnosis. Saf.
Sci. 45 (6), 653–667.
turity in upstream: oil and gas industry in Thailand. In: Proceedings Flin, R., Mearns, K., O’Connor, P., Bryden, R., 2000. Measuring climate: identifying the
of 2013 International Conference on Technology Innovation and common features. Saf. Sci. 34, 177–192.
Industrial Management, Phuket, Thailand, pp. S4-264–290. Fraser, P., Moultrie, J., Gregory, M., 2002. The use of maturity models/grids as a tool in
assessing product development capability. In: IEEE International Engineering
36. Guldenmund, F. W., 2008. Safety culture in a service company: a Management Conference, pp. 244–249.
case study. Journal of Occupational Health and Safety – Australia Geller, E.S., 1994. Ten principles for achieving a Total Safety Culture. Prof. Saf. 18–24.
and New Zealand, 24, 3, 221–235. Gherardi, S., Nicolini, D., 2000. To transfer is to transform: the circulation of safety
knowledge. Organization 7, 329–348.
37. Ayers, G. F., Culvenor, J. F., Sillitoe, J. and Else, D., 2013. Geertz, C., 1973. The Interpretation of Cultures. Basic Books, New York.
Meaningful and effective consultation and the construction in- Goncalves Filho, A.P., Andrade, J.C.S., Marinho, M.M.O., 2010. A safety culture maturity
dustry of Victoria, Australia. Construction Management and model for petrochemical companies in Brazil. Saf. Sci. 48, 615–624.
Griffin, M.A., Curcuruto, M., 2016. Safety climate in organizations. Annu. Rev. Org.
Economics, 31, 6, 542–567. Psychol. Org. Behav. 3, 191–2012.
38. Law, M. P., Parker, D., Lawrie, M. and Baker, G. R., 2008. A mul- Guldenmund, F.W., 2000. The nature of safety culture: a review of theory and research.
tidimensional approach to studying safety culture in health care. In: Saf. Sci. 34, 193–214.
Guldenmund, F.W., 2010. Understanding and Exploring Safety Culture. Thesis (PhD).
Proceedings of 2008 the International Conference of Healthcare
Delft University, The Netherlands.
Systems and Ergonomics and Patient Safety (HEPS), Strasbourg, Guldenmund, F.W., 2016. Organizational safety culture. In: Clarke, S., Probst, T.M.,
France, pp. 1–4. Guldenmund, F., Passmore, J. (Eds.), The Wiley Blackwell Handbook of Occupational
39. Law, M. P., 2011. Understanding and Changing the Patient Safety Safety and Workplace Health. John Wiley, Chichester.
Haavik, T.K., 2014. On the ontology of safety. Saf. Sci. 67, 37–43.
Culture in Canadian Hospitals. Thesis submitted for the degree of Hale, A.R., 2000. Culture's confusions. Saf. Sci. 34, 1–14.
Doctor of Philosophy Department of Health Policy, Management Hale, A.R., 2006. Method in Your Madness: System in Your Safety. Rede, Delft University
and Evaluation University of Toronto, Canada. of Technology, 15th September. Retrieved from < https://www.researchgate.net/
publication/27351879_Method_in_your_madness_system_in_your_safety > (last ac-
40. Chen, S., 2016. Towards the definition of an OHS management cessed 15th March, 2017).
maturity model and assessment tool. Thesis submitted for the de- Health and Safety Technology and Management (Hastam). 2017. Safety Culture
gree of Master in Mechanical Engineering of School of Industrial Benchmarking Conference, Loughborough University, September 6th 2017.
Henriqson, E., Schuler, B., van Winsen, R., Dekker, S.W.A., 2014. The constitution and
and Information Engineering, Politecnico di Milano, Italy. effects of safety culture as an object in the discourse of accident prevention: a
41. Jabonete, F. G. and Concepcion, L. R., 2016. Perceived safety cul- Foucauldian approach. Saf. Sci. 70, 465–476.
ture of healthcare providers in hospitals in the Philippines. Journal Hopkins, A., 2006. Study organizational cultures and their effects on safety. Saf. Sci. 44,
875–889.
of Sciences, Technology and Arts Research, 2, 1–14.
Hudson, P., Willekes, F.C., 2000. The Hearts and Minds Project in an Operating Company:
developing tools to measure cultural factors. Paper Presented at SPE International
References Conference on Health, Safety and Environment in Oil and Gas Exploration and
Production, 26–28 June, Stavanger, Norway.
Hudson, P., 2001. Aviation safety culture. Safeskies 1–23.
Alvesson, M., 2007. Understanding Organizational Culture. Sage, London. Hudson, P., 2007. Implementing a safety culture in a major multi-national. Saf. Sci. 45,
Amalberti, R., 2013. Navigating Safety Necessary Compromises and Trade-Offs – Theory 697–722.
and Practice. Springer, Dordrecht. Kirwan, B., Shorrock, S., 2014. A view from elsewhere – safety culture in European air
Antonsen, S., 2009a. Safety Culture: Theory, Method and Improvement. Ashgate, traffic management. In: Waterson, P.E. (Ed.), Patient Safety Culture. Ashgate,
England. Farnham.
Antonsen, S., 2009b. Safety culture assessment: a mission impossible? J. Contingencies La Porte, T.R., Rochlin, G., 1994. A rejoinder to Perrow. J. Contingencies Crisis Manage.
Crisis Manage. 17 (4), 242–254. 2, 221–227.
Bach, J., 1994. The immaturity of CMM. American Programmer, September, 7, 9, 13–18. Lee, T., Harrison, K., 2000. Assessing safety culture in nuclear power station. Saf. Sci. 34,
Becker, J., Knackstedt, R., Pöppelbuß, J., 2009. Developing maturity models for IT 61–97.
management – a procedure model and its application. Bus. Inform. Syst. Eng. 3, Levi-Strauss, C., 1966. The Savage Mind. University of Chicago Press, Chicago, IL.
213–222. Mearns, K., Kirwan, B., Kennedy, R.J., 2009. Developing a Safety Culture Measurement
Borys, D., Else, D., Leggett, S., 2009. The fifth age of safety: the adaptive age? J. Health Toolkit (SCMT) for European ANSPs. Proceedings of the Eighth USA/Europe Air
Saf., Res. Pract. 1, 19–27. Traffic Management Research and Development Seminar.
Burrell, G., Morgan, G., 1979. Sociological Paradigms and Organizational Analysis. Maier, A.M., Moultrie, J., Clarkson, P.J., 2012. Assessing organizational capabilities: re-
Ashgate, Aldershot. viewing and guiding the development of maturity grids. IEEE Trans. Eng. Manage. 59
Busch, C., 2016. Safety Myth 101. Mind the Risk, Mysen. < www.mindtherisk.com > . (1), 138–159.
Calingo, L.M.R., 1996. The evolution of strategic quality management. Int. J. Qual. Martins, J., 1992. Cultures in Organisations: Three Perspectives. Oxford University Press,
Reliab. Manage. 13 (9), 19–37. New York.
Campbell, D.T., Fiske, D.W., 1959. Convergent and discriminant validation by the mul- McLuhan, M., 1964. Understanding Media. Routledge, London.
titrait-multimethod matrix. Psych. Bull. 56, 81–105. Mkrtchyan, L., Turcanu, C., 2012. Safety culture assessment tools in nuclear and non-
Chung, A.Z.Q., Shorrock, S.T., 2011. The research-practice relationship in ergonomics nuclear domains. Review of safety culture tools. Report of the Belgian Nuclear
and human factors – surveying and bridging the gap. Ergonomics 54, 413–429. Research Centre. Nuclear Science and Technology Studies (NST), Belgium.
Cooper, D., 2000. Towards a model of safety culture. Saf. Sci. 36, 111–136. Office of Rail and Road (ORR). 2017. RM3 – The Risk Management Maturity Model.
Cox, S.J., Flin, R., 1998. Safety culture – philosophers' stone or man of straw? Work Stress Report available from: < http://orr.gov.uk/rail/health-and-safety/health-and-safety-
12 (3), 189–201. strategy/risk-management-maturity-model-rm3 > (last accessed 22nd September
Cox, S.J., Cheyne, A.J.T., 2000. Assessing safety in offshore environments. Saf. Sci. 34, 2017).
215–257. Parker, D., Lawrie, M., Hudson, P., 2006. A framework for understand the development of
Crosby, P.B., 1979. Quality is Free. McGraw-Hill, New York. organizational safety culture. Saf. Sci. 44, 551–562.

210
A.P. Goncalves Filho, P. Waterson Safety Science 105 (2018) 192–211

Paulk, M.C., Chrissis, C., Weber, M.B., 1993. Capability Maturity Model for Software, research. Ergonomics 52 (10), 1185–1195.
Version 1.1. Software Engineering Institute, Carnegie-Mellon, Pennsylvania. Waterson, P.E., 2014. Patient Safety Culture. Theory, Methods and Application. Ashgate,
Perrow, C., 1984. Normal Accidents. Princeton University Press, Princeton. Farnham.
Pidgeon, N., 1988. Shaking the kaleidoscope of disasters research – a reply. J. Waterson, P.E., 2016. Bridging the gap between the gap between research, policy and
Contingencies Crisis Manage. 6, 97–101. practice in health and safety. Pol. Pract. Health Saf. 14 (2), 97–98.
Rasmussen, J., 1997. Risk management in a dynamic society: a modelling problem. Saf. Waterson, P.E., 2017. ‘I came into safety by accident’ (interview with Professor Andrew
Sci. 27, 183–213. Hale). The Psychologist 64–66.
Reason, J., 1997. Managing the Risks of Organizational Accidents. Ashgate, Aldershot. Waterson, P.E., Catchpole, K., 2016. Human factors in healthcare: welcome progress, but
Reason, J., 1998. Achieving a safe culture: theory and practice. Work Stress 12 (3), still scratching the surface. BMJ: Qual. Saf. 25, 480–484.
293–306. Waterson, P.E., Clegg, C.W., Robinson, M., 2014. Trade-offs between reliability, validity
Reason, J., 2016. Organizational Accidents Revisited. Ashgate, Farnham. and utility in the development of human factors methods. In: Broberg, O., Fallentin,
Reiman, T., Rollenhagen, C., 2014. Does the concept of safety culture help or hinder N., Hasle, P., Jensen, P.L., Kabel, A., Larsen, M.E., Weller, T. (Eds.), Human Factors in
systems thinking in safety? Accid. Anal. Prevent. 68, 5–15. Organizational Design and Management XI. IEA Press, Santa Monica.
Robertson, K., Black, J., Grand-Clement, S., Hall, A., 2016. Human and Organisational Waterson, P.E., Robertson, M.M., Cooke, N.J., Militello, L., Roth, E., Stanton, N.A., 2015.
Factors in Major Accidents. RAND Europe Report. Available at: < http://www.rand. Defining the methodological challenges and opportunities for an effective science of
org/randeurope/research/projects/major-accident-prevention-factors.html > (last sociotechnical systems and safety. Ergonomics 58, 650–658.
accessed February 27th 2017). Waterson, P.E., Jenkins, D.P., Salmon, P.M., Underwood, P., 2017. ‘Remixing Rasmussen’:
Scott, T., Mannion, R., Davies, H., Martin, M., 2003. The quantitative measurement of the evolution of Accimaps within systemic accident analysis. Appl. Ergon. 59 (Part B),
organizational culture in health care: a review of the available instruments. Health 483–503.
Serv. Res. 38 (3), 923–945. Weick, K., 1987. Organizational culture as a source of high reliability. Calif. Manage. Rev.
Shipley, P., 1987. The methodology of applied ergonomics: validity and value. In: Wilson, 29, 112–127.
J.R., Corlett, E.N., Maneca, I. (Eds.), New Methods in Applied Ergonomics. Taylor and Weick, K.E., 1989. Theory construction as disciplined imagination. Acad. Manage. Rev.
Francis, London. 14, 516–531.
Silbey, S.S., 2009. Taming Prometheus: talk about safety and culture. Annu. Rev. Sociol. Weick, K.E., Sutcliffe, K.M., 2007. Managing the Unexpected: Resilient Performance in
35, 341–369. and Age of Uncertainty, second ed. Jossey-Bass, San Francisco, CA.
Step Change in Safety, 2017. Changing Minds Guidance. Available from: < https://www. Wendler, R., 2012. The maturity of maturity model research: a systematic mapping study.
stepchangeinsafety.net/safety-resources/publications/changing-minds- Inform. Softw. Technol. 54, 1317–1339.
guidance > (last accessed February 27th, 2017). Westrum, R., 1993. Cultures with requisite imagination. In: Wise, J.A., Hopkin, V.D.,
Sutton, R.I., Staw, B.M., 1995. What theory is not. Admin. Sci. Quart. 40, 371–384. Stager, P. (Eds.), Verification and Validation of Complex Systems: Human Factors
Trochim, W.M.K., 2006. Research Methods Knowledge Base, second ed. Internet WWW Issues. Springer-Verlag, New York.
page, at URL: < http://www.socialresearchmethods.net/kb/ > (version current as of Westrum, R., 2004. A typology of organizational cultures. Qual. Saf. Health Care 13,
October 20, 2006). 22–27.
Turner, B.A., 1978. Man-Made Disasters. Wykeham Science Press, London. Wilson, J.R., 1987. Whose attitudes to which aspects of work? or Cool media for high
Vincent, C., Amalberti, R., 2016. Safer Healthcare. Springer, New York. participation. Work Stress 1, 385–395.
Waterson, P.E., 2009. A critical review of the systems approach within patient safety

211