University of New South Wales Law Research Series

BIG DATA FOR SECURITY: A

CULTURAL ANALYSIS

JANET CHAN AND LYRIA BENNETT MOSES

[2020] UNSWLRS 80

UNSW Law
UNSW Sydney NSW 2052 Australia

E: [email protected]
W: http://www.law.unsw.edu.au/research/faculty-publications
AustLII: http://www.austlii.edu.au/au/journals/UNSWLRS/
SSRN: http://www.ssrn.com/link/UNSW-LEG.html

Electronic copy available at:

 BIG DATA FOR SECURITY: A CULTURAL ANALYSIS

Janet Chan and Lyria Bennett Moses

UNSW Law (UNSW Australia, Sydney NSW 2052 Australia), and Data to Decisions Cooperative
Research Centre
This is the original version of a paper (March 2016) that was revised and published as Chan J;Moses
LB, 2017, 'Making sense of big data for security', British Journal of Criminology, vol. 57, pp. 299 - 319,
http://dx.doi.org/10.1093/bjc/azw059.

Electronic copy available at: https://ssrn.com/abstract=3742516

 BIG DATA FOR SECURITY: A CULTURAL ANALYSIS

Abstract
Big Data technology is said to hold great promise for improved efficiency and effectiveness for law
enforcement and security intelligence agencies. This article aims to develop a cultural analysis of the
potential impact of Big Data on the production of national and international security. Building on a
Bourdesian framework for analysing police and new technologies, the article draws on empirical data
from an Australian study to examine how security agents made sense of the capability and value of
Big Data and developed technological frames that envisaged how this new technology could enhance
or change their practices. The analysis demonstrates the importance of understanding the habitus of
security agents in negotiating technological change in the field of security production.

Keywords
Big Data, Law enforcement, Security intelligence, Sensemaking, Technological frames, Cultural
analysis

Electronic copy available at: https://ssrn.com/abstract=3742516

Manuscript

BIG DATA FOR SECURITY: A CULTURAL ANALYSIS

1. Introduction
‘Big Data’ is an amorphous concept that has been used to refer both to large, diverse, rapidly-
changing datasets, or to the analytic techniques employed to extract information from such
datasets. While some have attempted to refine the definition of ‘Big Data’ (see e.g. Kitchin 2014),
others avoid the term, preferring ‘data science’ or ‘data analytics’. Boyd and Crawford (2012: 663)
take a broader view and describe ‘Big Data’ as ‘a cultural, technological, and scholarly phenomenon’
involving technology, analysis and mythology. The concept of ‘Big Data’ thus remains flexible,
subject to the different interpretations among those who seek to analyse or employ data-related
technologies for a wide variety of scholarly, commercial and government purposes (Bennett Moses
and Chan 2014; Chan and Bennett Moses 2016).
It has been suggested that Big Data holds great promise for improving the efficiency and
effectiveness of law enforcement and security intelligence agencies. For example, the Executive
Office of the President (US) (Podesta et al. 2014: 29, 58) has claimed that ‘[b]ig data can be a
powerful tool for law enforcement’ and that it ‘holds the potential to … substantially strengthen
national security’. Similar claims about the use and potential of Big Data appear on websites and in
other publications (see Wyllie 2013; Olesker 2012; Staniforth and Akhgar 2015).
This article aims to develop a cultural analysis of the potential impact of Big Data technology on law
enforcement and security intelligence by conceptualising Big Data as a new technique of security that
is being introduced into national and international security projects (Valverde 2014). Building on
Chan’s (2003) integration of Orlikowski and Gash’s (1994) notion of technological frames into a
Bourdieusian analysis of police’s reception of new technologies, the article draws on empirical data
from an Australian study to examine how security agents in law enforcement and national security
agencies made sense of the capability and value of Big Data and developed technological frames that
envisaged how this new technology could enhance or change their practices. The analysis
demonstrates the importance of understanding the habitus of security agents in negotiating
technological change in the field of security production.
The article is divided into nine parts. Part 2 develops a conceptual framework employed in our
analysis. In particular, it integrates Valverde’s (2014) ‘security projects’, Orlikowski and Gash’s (1994)
‘technological frames’, Weick’s (1995) ‘sensemaking’, and Bourdieu’s theory of practice to explain
how different social groups respond to the possibility of technological change. Part 3 summarises the
findings of the available empirical research on the impact of technology on security practices. Part 4
briefly describes the research method and research participants. Part 5 analyses the working
assumptions of security agents about current purposes of using data for security production. Part 6
describes security agents’ framing of Big Data—its nature, capability and value—compared with other
stakeholders. Part 7 explores their expectations about Big Data technology, while Part 8 examines
their perceptions of the impact of Big Data on the distribution of capital in the field of security
production. Part 9 concludes by summarising the findings and discussing their broader implications.

2. Making Sense of Big Data for Security

A useful starting point for analysing the potential impact of Big Data on security practice is to regard
Big Data as an instance of ‘technology’. Research in science and technology studies has concluded
that ‘technology’ is not only a physical given (artefacts and technical systems), but also comprises
knowledge about such systems as well as practices of handling them (MacKenzie and Wajcman 1985).
Technology is constructed in the sense that it is ‘made’ and also in the sense that it is interpreted and
understood through social groups influenced by a range of physical, social, political and
organisational factors that may change over time (Bijker 2010). To understand how Big Data is
constructed in the

Electronic copy available at: https://ssrn.com/abstract=3742516

context of law enforcement and security intelligence, it is useful, following Valverde (2014), to
conceive of Big Data as a technique that is being introduced into one or more security projects in the
governance of society. To set up a framework for examining the logic and practices of current
Australian security projects, we build on the analytic tools used by Chan (2003) for understanding the
impact of information technology on police practice. In particular, we use the notion of technological
frames from science and technology studies and sensemaking from organisational studies and
integrate them with concepts from Bourdieu’s theory of practice. The development of this
framework is explained below.

Security Projects
Valverde (2014: 382) has suggested that a fruitful way for researchers to study the governance of
crime and security is to focus on security projects—‘the governing networks and mechanisms that
claim to be promoting security at all scales’. Instead of focusing on ‘security’ as a concept, she argues
that we should look at the ‘very wide variety of activities and practices that are being carried out
under the name of “security”’ (Valverde 2014: 383–4). In particular, it is important to examine the
logic, spatiotemporal scale, jurisdiction and techniques of security projects. These aspects of security
are more than what their labels suggest. For example, the notion of logic in this formulation goes
further than the instrumental, rational dimension of governance to include its affective and aesthetic
dimensions: ‘the aims and the assumptions of a project – that which tells us what counts as relevant
information – but also the culturally specific fears and moods that pervade the field of security’
(Valverde 2014: 384). Similarly, scale has both spatial (or geographic) and temporal (both direction
and duration) dimensions. A distinction identified by Valverde that is highly relevant to our discussion
is that between past-focused exercises such as crime detection and criminal investigation and future-
oriented activities such as crime prevention in the governance of security. Jurisdiction is not
necessarily tied to geographical space but involves specifying ‘the proper authority for space X or
problem Y’ and thus ends up determining how X or Y should be governed (Valverde 2014: 388).
Finally, techniques of security encompass more than technologies or equipment; they can denote
reporting formats, as well as law, architecture, bodily habits and other governance tools.
While Valverde’s dimensions are useful for analysing security projects in general, concepts from
science and technology studies and theories of practice can provide additional tools for examining
the logic and practices in projects that involve technological change. The next section is devoted to a
discussion of technological frames and sensemaking.

Technological Frames and Sensemaking

The notion of ‘technological frames’1 (Orlikowski and Gash 1994) is a useful tool for documenting how
different social groups conceive of technology and respond to technological change. Drawing on the
idea of frames in social cognitive research, Orlikowski and Gash (1994: 178) define the technological
frame as ‘that subset of members’ organizational frames that concern the assumptions, expectations,
and knowledge they use to understand technology in organizations’. While people hold individual
interpretations about technology, members of a professional or occupational group may also have
assumptions and beliefs that are shared within the group. Technological frames can be powerful in that
they ‘will strongly influence the choices made regarding the design and use of those technologies’
(Orlikowski and Gash 1994: 179). Hence, the ‘success’ or otherwise of a technological change can be
explained by the congruence or incongruence of technological frames between, for example, the
architects and the users of a technology (Orlikowski and Gash 1994: 180).
In their own empirical research, Orlikowski and Gash (1994: 183–4) found three (overlapping) frame
domains that characterise participants’ interpretations of technology: (i) nature of technology,
people’s perception of the technology and its capabilities; (ii) technology strategy, people’s views of
the motivation behind their organisation’s adoption of the technology and the value of this
technology to the organisation; and (iii) technology in use, people’s understanding of ‘how the
technology will be used on a day-to-day basis and the likely or actual conditions and consequences
associated with such use’. The authors found incongruence in all three domains between the users
and the technologists; these incongruences had led to unanticipated outcomes ‘such as an initial
barrier of skepticism and frustration and the perception that the office had not realized the benefits
that were anticipated at the acquisition of [the technology]’ (Orlikowski and Gash 1994: 198).

Electronic copy available at: https://ssrn.com/abstract=3742516

In assessing Orlikowski and Gash’s (1994) contributions, Davidson (2006) suggests emphasising
framing as a dynamic process involving ‘interpretive power’ and environmental triggers, and
investigating the cultural and institutional foundations of technological frames. The idea of
sensemaking (Weick 1995; Weick et al. 2005) is helpful in this regard. Technological frames are not
static or frozen in time; they are formed as part of sensemaking—an ongoing process that people
engage in to explicate the world and give it a sense of order. Technological change creates an
‘occasion’ for sensemaking. While people can draw on any information or cue to make sense of
change, in practice they tend to draw on categories that summarise past experience such as cues
found in traditions, standard procedures and assumptions that are salient in their group or
organisation.
Davidson (2006) also sees important benefits for technological frames analysis to go beyond
organisational boundaries, given that information technology increasingly requires the involvement of
multiple organisations or whole industries. This is a particularly important point for understanding Big
Data technology as it includes a conglomerate of techniques, modalities and applications that are
applicable to myriad industries. In the next section, we propose broadening the technological frames
concept to incorporate structure and the interplay between structure, sensemaking and frames.
Technological Change in the Field of Security Production
An integrated framework for understanding the potential impact of Big Data technology on law
enforcement and security intelligence can be built from Chan’s (2001, 2003) analysis of police
responses to technological change. Chan (2003) draws on Bourdieu’s notions of field, capital and
habitus to focus on both structural and cultural determinants of social practice. Bourdieu’s field is a
‘social space of conflict and competition, where participants struggle to establish control over
specific power and authority’ (Chan 2003: 663). The field is often compared to a ‘game’ with
different types of ‘capital’ (economic, cultural, social, symbolic) that are valued (Bourdieu 1987).
Associated with each field is a system of dispositions (habitus) that agents in the field have
acquired through family, education system or professional socialisation; it internalises the
external structures and provides the dominant frame through which agents make sense of and act
in the world. Habitus both ‘sets structural limits for action’ and ‘generates perceptions,
aspirations and practices that correspond to the structuring properties of earlier socialization’
(Swartz 1997:101).
The field of security production is made up of various subfields, including (public or private)
agents and agencies such as police and intelligence organisations concerned with maintaining
order, preventing crime, enforcing laws or protecting lives and properties (cf McCahill’s (2015) use
of Bourdieu’s theory to theorise the crime control field). Agents and agencies are differentiated
not only in function, but also in power and resources. The logic of security projects (Valverde
2014)—their aims, assumptions, fears and moods—is manifest in the shared habitus of agents
who operate in their (sub)field. Similarly, their technological frame (Orlikowski and Gash 1994)—
the assumptions, expectations, and knowledge they use to understand technology—is a subset of
agents’ habitus.
For the purpose of this article, it is useful to conceive of the use of Big Data for law enforcement
and security intelligence as a change in the field of security production. Chan (2003) has suggested
that technological change can bring about changes in the field since technology is a much-valued
resource. For example, in the field of policing, technology can be a ‘power-amplifier’ (Nogala
1995). Technical expertise can be a form of cultural capital: where previously police leaders were
predominantly drawn from the criminal investigation branch, the ascendency of officers with IT
expertise may threaten the traditional power structures of these organisations (Chan 2003).
Technological change is, however, a double-edged sword. It can create problems and constraints for
policing, such as leading to greater internal and external demands for information and putting limits
on police discretion (Ericson and Haggerty 1997). Hence technological change can alter the field of
security production by creating resources as well as constraints (positive and negative capital).
Technological change can also transform the habitus of security production. For example, the
introduction of Big Data could bring about changes in security agents’ assumptions about the
purpose of information, what they regard as relevant information, how information is obtained

Electronic copy available at: https://ssrn.com/abstract=3742516

and used, and how they think information should be obtained and used (cf Sackmann’s (1991)
dimensions of cultural knowledge).

3. Research on Impact of Technological Change on Security Practice

There is a dearth of empirical research on the impact of technological change on security practice
(Manning 2014:2512). What is available relates mostly to policing and law enforcement.
Chan’s (2003) review of the literature found that while technological change in some instances
can ‘radically alter the structure of police organization by levelling hierarchies, blurring traditional
division of labor, dispersing supervisory capacities and limiting individual discretion’ (Ericson and
Haggery 1997:388), it had generally resulted in continuities more than changes in police practices.
For example, the availability of more data did not lead to a more proactive style of policing: police
continued to regard ‘information as useful only if it leads to arrests’ and problem-oriented
policing as ‘soft’ and ‘marginal’ (Chan 2003: 666). Similarly, what is relevant information was often
restricted by police assumptions about the purpose of information, so that data on community
profiles, economic conditions, community attitudes, etc was not usually regarded as relevant.
Even though police were aware of the potential of using technology for ‘smarter’ policing
strategies, ‘they said there was not sufficient time or resources to realise this potential’ (Chan
2003: 666). The availability of better information technology had also led to few changes in how
information was obtained and used, and how police thought it ought to be obtained and used. For
example, there was a ‘cultural aversion’ to depersonalised and decontextualised data generated by
crime analysts (Cope 2003). Similarly, detectives had a tendency to ‘co-opt crime analysis for the
purposes of crime investigation’ (Sheptycki 2004:324).
Chan (2003: 668) concludes that ‘the prevalent attitude of police appears to still favour case-by-case
investigation rather than crime analysis, evidence gathering rather than intelligence analysis, secrecy
rather than openness in information sharing’. In fact, her study (Chan 2001) demonstrated a classic
case of the clashing of technological frames between users (who expected technology to make their
work easier) and architects (who had intended the organisation to use information in a more
sophisticated way). Chan (2003), however, does not see technological frames as immutable. Rather,
the introduction of new technology is ‘merely the beginning of a “technological drama” (Manning
1992, 1996) of normalization, adjustment, reconstitution and reintegration’ (Chan 2003: 673).
Technological change can destabilise the status quo in terms of power balance, and can even lead to
resistance or sabotage in some circumstances (Ericson and Haggerty 1997).
More recently published research similarly confirms that the impact of information technology on
police practices can be uneven or unpredictable. For example, results of a multi-site study of police
technology in the US demonstrate that ‘the effects of technology are complex and that technological
advancements do not always produce obvious or easy improvements in productivity, communication,
cooperation, management, or job satisfaction’ (Koper, Lum and Willis 2014:212). Similarly, research in
six Canadian police services found that the use of ‘crime science’ and analytic technologies to support
‘intelligence-led policing’ (ILP) is more rhetorical than real (Sanders, Weston and Schott 2015:711).
More specifically, the ‘occupational culture of information hoarding… has shaped the use and
functioning of police innovations’ (2015:718). In line with previous research on the ‘poorly
understood and appreciated’ role of crime analysts (eg Cope 2004), the lack of knowledge and
training about crime analysis on the part of police managers and officers ‘has rendered many analysts
to engage in simple crime counting and mapping instead of advanced analytics’ so that instead of
adopting a new approach to policing (ILP), new technologies are used to support ‘traditional modes of
policing’ (Sanders et al. 2015:724).
Nevertheless, with the production of security being a major global concern, public institutions such as
police and national security agencies are the prime producers and communicators of security
knowledge (cf Ericson and Haggerty 1997). There is therefore enormous pressure for these agencies
to make use of new technological tools associated with Big Data.

Electronic copy available at: https://ssrn.com/abstract=3742516

4. Research Method
This article draws on a research project Big Data Technology and National Security conducted under
the auspices of the Data to Decisions Cooperative Research Centre (D2D CRC). The empirical data
includes 31 semi-structured interviews2 conducted with 38 stakeholders including law enforcement
and intelligence officials, policymakers, computer technologists, and officers in relevant civil society
organisations. The research team worked with various government agencies and the D2DCRC’s
partners to identify potential interviewees in Australia. Twenty-four of the interviews were recorded
with the consent of the research participants and verbatim transcripts prepared. For those who did
not consent to recording, notes were taken to recreate as closely as possible the words used by
research participants.
We classified research participants according to the nature of the organisation for which they
worked. In each case, there were three potential classifications: Operational (O), Technical (T) and
Policy (P). Where a participant was being interviewed in relation to a recent former role, the coding
matched the former organisation. While participants were asked different questions depending on
their role and organisation, all participants were asked similar questions in relation to their framing of
Big Data.
Among the 38 participants, 19 were from operational, 7 from technical and 12 from policy
organisations. Since the sample was not randomly selected, the results presented here should be
regarded as indicative rather than representative of the population of stakeholders. Another
limitation of the sample is that interviewees from operational organisations were mainly in
managerial or higher positions rather than security agents operating at the coalface. This was in spite
of our request to have both managerial and operational officers represented.
Quotations from interviews have been altered in various ways. In order to protect the identity of
research participants, any details of their organisation or team that appeared in the transcript were
removed. In order to increase the fluency and relevance of selected quotations in the report, we have
also used ellipses and square brackets to indicate the omission or replacement of words respectively.
In accordance with the framework described in Part 2, the following analysis will focus on several
dimensions of security agents’ habitus: their perception of the purpose of data in general, their
conception of Big Data and its capability and value, and their expectations of how Big Data will affect
their work. It will also discuss participants’ perceptions as to how the field of security production is
likely to be affected, i.e. the winners and losers in the use of Big Data.

5. The Purpose of Data

Data may be regarded as a form of ‘security object’ that is relatively banal (i.e. commonplace and
taken for granted) yet powerful nevertheless (Goold, Loader and Thumala 2013). Research
participants in our study worked with a wide range of data, from telecommunication metadata,
official data, data from international partners, internal databases, information provided by the
community, geospatial or financial data to open source or online data and communication signals..
Not unexpectedly, the purpose of data was very much tied to investigations:

Any data that we can collate online, whether it be that online evidence that may indicate the
commission of offence or assist in making a nexus, a link, to that offence, such as
photographs, emails, whether it be data these days, obviously text messages, contacts.
Realistically it’s comprised of anything that’s online that we can, again lawfully, collate for
the purpose of our investigation… we use any data that we can get our hands on lawfully,
certainly, to assist in our investigation. (O, emphasis added)

In terms of the temporal scale (Valverde 2014) of their security projects, research participants from
operational organisations nominated a range of past-focused and future-oriented purposes of using
data. Past-focused purposes include investigation, arrest and prosecution (nominated by 9
participants), reporting (1), and event evaluation (1); while future-oriented purposes include
prevention or disruption of incidence or mitigation of risks (6), intelligence gathering (4),
identification of trends or risks (3), policy or service decisions (3) and trust building (1). Often, data
would serve multiple purposes, such as:

Electronic copy available at: https://ssrn.com/abstract=3742516

 [Data is mainly used for] security intelligence. You’re looking at two or probably three areas.
One is event prevention, so trying to foresee something or stop something from happening.
Two, you’re looking at event evaluation, so what happened and now retrospectively that we
know something has happened—can we better analyse it to see what we missed, what we
could have done better around who was involved now that more players might be exposed?
Then I guess the third one was creating the bigger picture, so you find one person of interest
and then you look at 10 people that they regularly contact and you look at the 10 people
that they regularly contact and then you look at the 10 people they regularly contact and
you create a global map or an organisational map of contacts and then you … cut out who’s
not of interest…. Then you start to look for the links across the networks … so that informs
then this strategic analysis … it’s that kind of longer term analysis rather than event based
analysis. (O, emphasis added)
Yeah, we talk about a spectrum of activity. So we’ve got … traditional law enforcement so we
always go for prosecution. If we can’t prosecute … depending on where it is in the cycle, we’ll
look for an intervention like a control order or a preventative detention order … Then we’ll
go into the middle where we’re looking at this sort of disruption thing. … So you have to sit
back, bring that data together and actually work out what’s the risk we’re going to have and
how are we going to play that? What’s the way to intervene? (O, emphasis added)
We do what we call security intelligence investigations. In the course of that, we produce lots
of intelligence … It depends on the national security outcome we are trying to achieve. We
deal very much in shades of grey. Jail may be the best national security outcome where
police assisted by [name of agency] have enough evidence to put someone in jail. But in
some situations, it may be just as good to stop them doing something. We can sometimes
convince them to stop, that is, disruption that does not involve prosecution. (O, emphasis
added)
As one research participant noted, the purpose of using data was not static but potentially evolving:
[W]e're very focused on prosecution. There's a real desire within parts of the agency to move
away — certainly with some crime types — move away from prosecution and be more
imaginative in terms of the strategies around disruption, deterrence, target hardening and
the likes. … But if we weren't so focused in on prosecution all the time, I suggest that we
would look for different data sources and we would ask different questions of the data,
because we'd have a different mission if you like. (O)

This highlights that different missions involve different data and different tools. For example,
investigation and disruption both involve identifying individuals to whom data pertains. Even where
research participants described using data for future-focused activities, the kind of analysis being
done very much revolved around investigating individuals for past conduct or identifying individuals
who may be involved in future conduct rather than understanding broader trends among groups. This
is consistent with the fact that almost all research participants were only interested in identified,
rather than de-identified data:

Rarely is it de-identified, because the only reason we'd be sharing information is for
investigative action or in support of an investigative outcome. … [T]he purpose … is to
identify who, what, where, how, etc. … [I]t has to be for the purposes of the conduct of an
investigation with an intent to prosecute, and where all other avenues have dried up. (O)

It's no value if it's de-identified. (O)

With few exceptions, where sharing of de-identified was discussed, it was generally as a future
possibility or rare practice:

It is possible that, in the future, police could use our data to predict trends....(O)

We typically share reports … reports based on identified, classified data. (O)

Electronic copy available at: https://ssrn.com/abstract=3742516

This was consistent with accounts of participants from technical organisations who stated that their
systems are not concerned with de-identifying data or that such data is not of central interest to
government clients:

The systems that we're currently looking at aren't trying to de-identify data. (T)

We don't provide the capability to de-identify data. (T)

The above findings suggest that the importance of using data for case-by-case, investigative or
disruptive purposes, rather than for identification of trends, predictions, or strategic analysis, is a
shared assumption among security agents, a key dimension of their habitus, even at the managerial
level.

6. Big Data and Its Capability and Value

To understand how agents in the field of security production conceived of Big Data, we examine what
they thought Big Data is and what capability and value it will bring to their work.
Definition of Big Data
Research participants were asked how they would define Big Data. This was done not to measure
awareness of a fixed definition, but rather to explore diverse interpretations of a flexible concept.
Figure 1 shows the main responses broken down by the type of organisation participants worked in.
The most frequently mentioned attribute of Big Data was in terms of its volume, followed by its
analytic or predictive capacity, the fact that it consists of aggregated or integrated data from different
sources, and that the volume of data makes its handling beyond the capacity of humans, the skills of
existing analysts or current technology. Some mentioned velocity and variety as characteristics of Big
Data. Five participants—all from technical organisations—saw Big Data as a marketing term that
covers a variety of techniques. Only two participants mentioned veracity as a challenge of Big Data.
[Figure 1 about here]
Volume was often mentioned together with the need for new technology, particularly in relation to
the need to employ advanced techniques in analysis:
An ever-increasing, an exponentially-increasing volume of information which is beyond the
capability of a human to analyse without computer assistance. (O)
My understanding is Big Data is enormous data sets or combinations of data sets that require
advanced and analytic techniques in order to make sense of them or analyse them. (P)
Some participants saw the analytic capacity of Big Data, the capacity to ‘unlock the secret that’s
within the data’ (O), as its defining feature:
I would describe Big Data as the consolidation of large amounts of information in a single or
multiple repositories that can be manipulated or explored to release information or findings
that could not be found if the data [were] analysed individually or separately. It’s primarily
associated with trying to find much more complicated, complex relationships – it’s this trying
to unlock the secret that’s within the data. (O)
While four of the participants from technical organisations mentioned volume as one of the
characteristics of Big Data, five of them admitted that they disliked the term Big Data which some saw
as a marketing term. They would prefer to focus on analytics or predictive techniques—the size of the
dataset may or may not be essential:
Big data is something that’s obviously more of a marketing term than anything specific … Big
data means I guess the collection and exploitation of significant volumes of data to drive
outcomes that are relevant to the business or organisation that is trying to use it ... I think
Big Data can be a bit of a distraction because it’s not so much necessarily about the volume
of data but about the signal in the data that’s actually of primary interest … ultimately it’s the
analysis and that cuts across both what you might term analytics… to derive whatever the
outcomes are that are relevant for the situation that you’re working through. (T)

Electronic copy available at: https://ssrn.com/abstract=3742516

Almost one in four research participants defined Big Data in terms of the aggregation or integration of
data from different sources. The majority of these worked in operational organisations:
So Big Data is everything — because I actually consider what we hold, agencies, as Little
Data. The Big Data that’s out there is a lot of the stuff that sits in the public space, like
Facebook, like Twitter and things like that. Then you move to the next phase which data
being held on all of us which is held in different sort of areas like our licensing material, our
passports material, our movements material. All this sort of stuff that sits in silos which when
analysed on its own really means nothing but when you aggregate it all up can actually build
a pretty good picture about somebody. (O)
…it’s the accumulation of large datasets beyond what would normally be held within one
organisation or entity containing many different aspects of information within that dataset.
Therefore the analytics of it is how can you identify those useful pieces of connections
between those disparate pieces of information. (O)
Thus Big Data can generate diverse meanings among those working on or formulating policies for
security projects. As well as the obvious question of size, the term captures ideas about analytic
capacity, the integration of data sets, the technological horizon and buzz word scepticism.
Capability and Value of Big Data
Research participants were also asked questions designed to elicit their perceptions as to the capacity
and value of Big Data, particularly for law enforcement and security intelligence (see Figure 2 for the
distribution of responses).
[Figure 2 about here]
Half of the participants referred to Big Data’s more advanced analytic capacity:
Most importantly I think predictive trending potentially. Retrospective networking or even
real-time networking analysis. Associational analysis of individuals, groups, and
organisations. (O)
What makes it different is that increasingly using better tools and techniques we can gain
insight from data that to date are not humanly possible. So it helps us join the dots where we
can’t possibly do it ourselves. (O)
A smaller proportion mentioned its ‘richness’ or ‘completeness’ as the advantage of Big Data, which
was then linked to its analytic capacity:
So what these big data analytics allow you to do is to use all of your data, rather than a
subset… and therefore get a much more complete picture of what’s there, and therefore get
much more valuable and highly qualified insights. The other is that it allows you to
implement a lot more real time analytics… So that’s the big change really. (T)
The difference is between populational data as opposed to a sampled data set. If you have
complete data, the potential for analysis is not subject to the risk of modelling but is
observed fact. In other words, richer analytical capacity. (P)
The ‘richness’ of Big Data was also linked to an investigative advantage by participants who worked in
operational organisations, by providing historical and contextual details, as well as the ability to cross-
check information and identify new targets:
… let’s say for example we were analysing past activities of what terrorist suspects have been
like. So analyse their movements against when they applied for their passports, who went
guarantor for them, where they went to school, what their license is, when they got it, all
that sort of stuff and we analysed all that for individuals. Out of that we might be able to pick
up some patterns. … Then if you … develop an algorithm from that and then you ran it right
across all the systems you may very well pick up targets that are previously unknown to us …
So using … an analysis of previous data in terms of predictive type activity. None of us have
the capability to do that at the moment even though the information is probably out there to
do so. (O)

Electronic copy available at: https://ssrn.com/abstract=3742516

 [Big Data’s] interrogation potential. You can cross check against multiple things. (O)
You can actually build profiles and identify people, patterns of life, things like that. I mean …
that actually gives you a greater level of understanding about somebody before you actually
approach them. … Whether we’re setting up a surveillance plan or we’re planning when
we’re going to execute a search warrant and all those sorts of things which are traditional
law enforcement, trade craft which we do for a whole range of reasons, a lot of it around
safety and our own people in the broader community, but also about actually understanding
… a situation. Because while it gives you the opportunity to target in on people it also gives
you the opportunity to not target people … (O)
Four out of the seven research participants who worked in technical organisations and two from the
policy group mentioned proactive/preventative policing as an opportunity that Big Data could open
up for law enforcement and security intelligence:
In the counter-terrorism space it’s going to be manifested through a better ability to prevent
– although you can never prevent but to be able to intercept, detect and otherwise intervene
in potential events before they happen. (T)
The biggest one, I think, is prediction and probably the most challenging. Trying to change
the way policing works so that it’s not reactive, but is more proactive about looking for
potential anomalies or indicators that something might be occurring. (T)
I guess the shift that everybody talks about, which I think is a reasonable way of putting it, is
to move towards preventative policing and preventative law enforcement national security.
In other words not to react to an occurrence, but to be able to anticipate occurrences,
anticipate where law enforcements resources may be required ahead of time. So to move
from a lagging indicator regime to a leading indicator regime. (T)
Surprisingly, this was not mentioned by any of the research participants who worked in operational
organisations, not even those in a technical role. This suggests that there may be an incongruence
between the technological frames of the operational and those of the technical participants.
Participants in operational organisations saw the investigative advantage that Big Data could bring,
but participants from technical organisations were suggesting that Big Data could offer a different
way of doing policing and intelligence work.
Improved efficiency or effectiveness was mentioned by six research participants as what Big Data can
provide. Four participants (three were from technical organisations) pointed to the opportunity for
governments to make better decisions or provide better services, while three mentioned confidence
or accuracy as an advantage of Big Data. One participant with an operational role cautioned against
having unrealistic expectations about the predictive capability of Big Data in the context of crime.
Finally, one participant (from the policy group) confessed they didn’t know what the advantage of Big
Data was.
More than half of the security agents who took part in the research said that they were not currently
using Big Data. This suggests that their conceptions of Big Data and its capability and value were not
necessarily based on first-hand knowledge or experience with this technology. In spite of the small
sample sizes, the apparent incongruence in technological frames between the operational and the
other (technical and policy) participants is consistent with the findings in Part 5 that an important
dimension of the habitus of security agents was their shared assumption that data was useful for
case-by-case investigative or disruptive purposes in relation to known events or individuals, rather
than for predictive modelling to anticipate new threats. This raises another question about their
habitus: what do security agents expect from Big Data technology? This is examined in the next Part.

7. Expectations regarding Big Data

Our interviews with operational participants suggest that security agents have different expectations
about what Big Data technology will bring. Since more than half of them had not been using Big Data
in their work, the advent of Big Data has created a ‘occasion’ for a new focus for sensemaking, which,
as Weick (1995: 14) argues, is to ‘construct, filter, frame, create facticity … and render the subjective

Electronic copy available at: https://ssrn.com/abstract=3742516

into something more tangible’. To use Bourdieu’s terminology, sensemaking is the process through
which the habitus constructs and interprets changes in the field.

Participants were both attracted to new technology and wary of what it may bring. To make sense of
this new scenario, ‘they simultaneously interpret their knowledge with trusted frameworks, yet
mistrust those very same frameworks by testing new frameworks and new interpretations’ (Weick et
al. 2005: 412). Trusted frameworks can be based on traditions or standard procedures, naïve
expectations or informed understanding, familiar experience or a leap of faith. As one technologist
participant pointed out, unrealistic expectations can lead to ‘pretty bad outcomes’:

I think there are risks around expectations … historically in [technical organisation] we’ve
talked about it as the ‘find terrorist’ button. …[O]ur organisation was working on counter-
terrorism problems and probably for the first three or four years of our existence the most
requested feature was some manifestation of which button do I press to actually find the
bad guy? There is, particularly among non-technical people, a yearning desire based on
movies and otherwise to actually think that there is an ability to just automatically do their
job for them. The reality is that that’s far, far from the truth and far, far from desirable … So
there’s a risk that the expectation will never be met and there’s a risk that the expectation is
just misguided to begin with. … There’s no substitute for having an intelligent human being
in their intuitions and understanding of the world and you very much… want that person to
be there. There’s a risk that that is not well understood and there’s a risk of software
companies coming to the table and saying, well, technology is the answer and due to that
mismatch of expectation too much willingness on behalf of these agencies to accept that as
true which could ultimately have pretty bad outcomes. (T)
Some participants focused less on capability and more on the ability to use the tools conveniently,
referencing features like mobility and speed:
More and more they want the data immediately or in real time. … [A]t least some agencies
are moving towards the need to have data on device when out on the streets when doing
their job. (O)
That comes down to again speed access to that data, to download that or to upload that
information, and to access it, and capacity. (O)
This perspective is very much grounded in existing frameworks and modes of practice.
Given the importance of case-by-case investigation or disruption of crime or disorder, operational
participants who considered capability improvements generally expected Big Data to bring more
diverse data (as implicit in the understanding of Big Data as data aggregation), and thus were
concerned that it also comes with better facilities for sorting or prioritising information, which may
require a higher degree of automation:

Law enforcement organisations are thirsting for more data and the right data at the right
time. They are nervous about being swamped with data so getting right data is important.
(O)
What you want is that if there are a thousand pieces of information we want the analytical
tools to do the analysis, to understand the context and prioritise to say do this one first, this
one second, this one third ... (O)
[T]he amount of Big Data that that is going to create is very large but there’s going to be a lot
of useful information in there and an extreme amount of un-useful information. So our
ability to gain access to that data and have a mechanism to find the needle in the haystack or
the valuable information from the rubbish is going to be extremely important. (O)
Related to the idea of better sorting was improved tools for human-driven search and data
exploration. These were generally framed by comparisons with existing and familiar commercial
products, such as Google and IBM Watson, sometimes by direct reference and sometimes by similarly
described functionality:

Electronic copy available at: https://ssrn.com/abstract=3742516

 There is no capability to put in a name and draw from various different sets of data – No
POOGLE [Police Google]– can’t put in a name and get an answer like you can with Google.
(O)
Big Data will come alive when it adapts to what is being searched for, once you get into the
first layer, you get further result sets getting the spider’s web of structured index data, it self-
learns machine-learns each time but we don’t really have this yet. (O)
The thing that struck me about Watson … is you get to a point where [you have set up the
appropriate rules] then you should be very comfortable that the more data you provide to an
engine like that, … it would take you to the answer of that question, … it gives you better
surety about the judgements you make about what's going on in the environment. Or what
might be going on in the environment, or what might go on in the environment sometime in
the future. Or what might have gone on in the environment that you don't have real visibility
of. (O)
Again, as with the case of sorting and prioritisation tools, there was an expectation of increased
automation. However, a number of participants thought that fully automated Big Data tools would
not be effective and emphasised the importance of human evaluation and reasoning as part of the
process, so that automation would assist rather than replace human analytic reasoning:
You need a system flexible enough to show near hits and create associations but a person
makes the decision. (O)
I would be cautious about relying on anything computer generated to gain a complete
understanding about intent and so forth. Not in my lifetime. (O)
I think one of the greatest things [Big Data] can deliver to law enforcement and to the
national security community agencies is reducing the amount of time that humans spend
curating data, managing data, getting it into the right formats and into the right places
before they analyse it. The most critical thing that new tools, techniques and procedures in
big data analytics will do for those agencies is to increase the amount of time analysts spend
analysing rather than managing data. (O)
I think there is a risk that smart systems will deskill people. It scares the hell out of me. (O)
There is thus some realisation among operational participants that, as much as automation of some
functions is useful and necessary in the context of large, diverse data sets, a ‘find terrorist’ button is
both unlikely and undesirable.
While most participants linked their projections to trusted frameworks and existing practices, one
participant from an intelligence organisation described how a ‘Big Data’ business model would change
their current operation:
We operate under a ‘join the dots’ business model. … If we have a lead, we follow the lead to
its logical conclusion, joining the dots. We don’t access data until we have cause to use it
(need to link another dot). … Under a Big Data business model, we [would] have all the data
available all the time, we traverse it constantly looking for trends, patterns, anomalies and
red flags. (O)
According to this participant, Big Data would change the approach to how intelligence work is done.
Interestingly, however, the use of ‘we’ as a subject of the verb ‘traverse’ suggests that this process
would be human, not machine, driven.
What is important to note here is that, outside the contexts of technologies for prioritising and
decrypting data, and possibly distributed data storage systems, none of the future visions described
by the security agents would involve significant technological breakthroughs. Much of it draws on
existing commercial tools, such as Google search or IBM Watson. This is unsurprising – agents made
sense of the possibilities of Big Data technologies in light of what they understood about existing
technologies. Mostly, agents were concerned with new ways of doing things that can be linked to
better access to more data, being permitted to use it in new ways, better IT resourcing (for example,
through mobile access and faster speeds), or better project management. They wanted more data,
but also the tools to manage it. Mostly, they wanted to use data to make their work easier (e.g. tools

Electronic copy available at: https://ssrn.com/abstract=3742516

that can be accessed quickly and remotely, better prioritisation of information, better search
functionality) rather than to change the nature of their work (e.g. from investigation, prosecution or
disruption of individuals to broader identification of trends). A notable exception is the suggestion
that intelligence could move from ‘joining the dots’ to Big Data approaches. There were some general
concerns about negative implications of Big Data thinking, and some more specifically linked to
excessive automation (which raises issues of job losses and thus reduced status for current employees
as well as questions about the quality of decision-making). But overall, agents were looking to
existing, commercially available technologies in imagining the future of Big Data technology in their
own work.

8. Winners and Losers in Technological Change

The field of security production is made up of a multitude of agencies and agents charged with
responsibilities to maintain order and security, preventing crime and protecting people and property.
Big Data technologies are likely to be taken up to different extents by different agencies depending on
their capacities, resources and purpose. The fact that not all intelligence and law enforcement
agencies had equal capacity to access Big Data technologies was noted by two research participants:
… there is a broad spectrum of sophistication across the law enforcement and … the national
security community or agencies … So within the national security community there’s a vast
difference in the maturity and sophistication of the use of Big Data and Big Data analytics.
(O)
…it’s very largely a question of resources. The intelligence services have many more
resources proportionately speaking than the police do for this kind of work. (P)
Resources and capabilities can be a source of prestige, both for the agencies themselves and security
agents within them.
As discussed in section 2, technological change can be a resource as well as a constraint, providing
advantages to certain groups while posing risks for others. We asked all participants to identify the
risks of using Big Data for law enforcement or security intelligence. The risks identified are set out in
Figure 3 and they vary by the type of organisation participants represented. While those from
operational organisations mentioned most of the listed risks, only two mentioned misuse of data.
Participants from policy organisations were most concerned with privacy, data security, and misuse of
data, with no more than one person mentioning any of the other identified risks. Those from
technical organisations were most concerned with misuse of data and misplaced trust in Big Data
technology. Overall, privacy, data security and integrity, misuse of data and misplaced trust in
technology or algorithms were the most significant concerns, while only one research participant was
concerned about the risk of discrimination.
Differences among the three groups are interesting, while not surprising. Those in operational
organisations seemed to be less concerned about misuse of data and more concerned about their
own potential loss of capital (through political and reputational risks, lower public perceptions and
overload) compared to other groups. More surprising is the fact that those in operational and
technical organisations were conscious of misplaced trust in technology, an issue of less note to those
in policy organisations (which include government agencies focussed on policy). Also surprising is the
relatively low level of concerns around the potential for discrimination, despite these being raised in
the literature (Barocas and Selbst 2016).
[Figure 3about here]
For research participants in operational organisations, not engaging with Big Data technologies could
itself be a risk:
The real question is what are the risks of not taking a Big Data approach. … Whether [the
public] are comfortable on a bulk data perspective, if we don’t have access to data we
cannot stop things happening. It is about national security outcomes. We know people want
to breach national security. If the data is there, we will be able to stop terrorism or
espionage (O)

Electronic copy available at: https://ssrn.com/abstract=3742516

Yet having access to Big Data also presents a risk to operational participants: agencies could face a
loss of symbolic capital (e.g. public trust, reputation) if they did not act on the data they had, they
made use of data that ‘society didn’t think they had access to’, or they acted on a ‘false positive’:
Again, that’s probably one of the biggest risks, I think, of Big Data, that from an
organisational point of view, if you know something you haven’t acted on that information,
therefore are you liable in the public sphere that you knew the risk and you did nothing with
it, regardless of where that data sits, or where that risk sits. (O, emphasis added)
I think there’s a risk for the police and law enforcement and intelligence agencies — really
about perception. If they are seen to be using data that society didn’t think they had access
to. (T)
I think a reasonable paranoia on the part of the law enforcement national security agencies
themselves is the embarrassment of a false positive. A lot of the sophistication in the system
really is about reducing the number of false positives, and for that matter of false negatives.
So that’s an area that gets a lot of attention. That’s really an area where you might even end
up catching totally innocent people, just through a set of circumstances that made a
suggestion that turns out to be unwarranted. (P)
While risks to agencies dominated among operational participants, all categories of research
participants identified more general risks . The most frequently mentioned concern was that
‘everyone, the community or citizens’ were exposed to the risks of using Big Data. The risks here were
largely around privacy, poor data integrity and security and misuse of data, including the risks of
‘vendetta policing’, harassment and over-enforcement. Other categories of people exposed to risks
include minorities and people at the margins, people of interest to law enforcement and security
agencies, academics and researchers, and people identified through data.
The analysis above suggests that technological change in the ‘game’ of security production generates
both winners and losers. Stakeholders, including security agents, were cognisant of the range of risks
Big Data technology could bring to the field of security production, e.g., the risks of invasion of
citizens’ privacy, compromises to data security and integrity, misuse of data, misplaced trust in
technology, and various political and reputational risks to governments and security agencies. Yet for
security agents, these risks must be weighed against the potential benefits of Big Data, at least as
they understood them.

9. Conclusion
In spite of the promises of Big Data for improving the efficiency and effectiveness of policing and
security agents, very little is known about how Big Data is understood or imagined by these agents.
Empirical findings from our Australian study suggest that Big Data is a security technique that is both
novel and contested (cf Goold et al. 2013). Less than half of the security agents who participated in
the research reported that Big Data was being used in their unit; their conception of Big Data was
therefore not necessarily based on experience with this technology. Different stakeholders perceived
the value of Big Data in slightly different ways: while all participants emphasised its analytic capacity,
security agents tended to focus on the richness of data and the investigative advantage it affords,
while participants in policy and technical organisations saw Big Data as opening up opportunities for a
more proactive approach to security based on inferences from trends and patterns. This is consistent
with an important dimension of the habitus of security agents: while data was considered useful for
both past-focused activities (e.g. detection and investigation) and future-oriented exercises (e.g.
prevention, disruption or risk reduction), the focus was almost always on identifying and learning
about individuals rather than understanding broader trends. Security agents expected Big Data to
provide better access to more data and a range of improvements over current methods, without any
fundamental change in approach. Their visions of what Big Data could offer were primarily based on
their current technological frames and their experience with existing commercial tools. While they
were aware of community concerns around issues such as privacy and data security, they were
especially conscious of the political and reputational risks in raising public expectations and not
delivering the outcomes through technical or human errors.

Electronic copy available at: https://ssrn.com/abstract=3742516

Valverde (2014: 389) has argued that it is ‘dangerous’ to focus on techniques only in our analysis of
security projects, as their logic, scale and jurisdiction ‘cannot be read off from the techniques’. Our
analysis confirms this — stakeholders have different expectations of what Big Data can provide as a
security technique. While security agents may see the purpose of using data as split between (past-
focused) detection and investigation and (case-based future-oriented) prevention or risk reduction,
among developers of software tools, the ‘selling point’ of Big Data analytics has primarily been future
oriented and risk-based rather than case-based (see Bennett Moses and Chan 2016). This potential
incongruence in technological frames between the users and the architects of Big Data is likely to
pose problems for future implementation. Thus, technologists could pay more attention to their own
expectations and assumptions and whether they are aligned with those of users and managers
(Orlikowski and Gash 1994).

A better understanding of how cultural assumptions (part of habitus) can influence the impact of new
technology is not only important for managing technological change within organisations, but also for
designing regulatory or governance regimes (other techniques of security) for the benefit of the
broader community. Goold et al.’s (2013:987) analysis of how surveillance cameras in the UK have
become a banal security object, taken for granted by citizens as ‘an integral part of the infrastructure
of public life’, even a new kind of ‘security blanket’, should alert us to the possibility of the
‘securitisation’ of Big Data going down the same path, either through ignorance or apathy.

There are many potential futures for greater use of Big Data in national security, both from the
perspective of potential access to larger, integrated datasets, and the increased capacity to extract
information from data. There are also a number of risks associated with different pathways, which
will affect the likelihood and extent of impact on the potential ‘losers’ here. The habitus of different
players in the field is crucial because it will affect how security practices change in response to Big
Data ideas. Because of the diversity among stakeholders, the outcome is partly a question of
jurisdiction in Valverde’s sense, and not only of the technical performance of the various possibilities.
Our study has revealed where there are gaps in participants’ understandings of risks or impact on
capital. For example, while there is an awareness within operational organisations of the limits of
fully automated decision-making, there is also a strong sense that algorithms can be used to prioritise
targets and perhaps even identify new ones. Surprisingly, only one research participant (from the
policy group) raised the issue of discrimination as a risk in this context. The study thus also suggests
that, while arguments about privacy are familiar (although subject to disagreement), other risks are
either unexplored or only considered within particular types of organisation. If Big Data is to make a
difference to security practice, there needs to be greater alignment among sectors regarding
understandings of the technologies involved, whether and how they ought to change approaches to
security, and, most crucially, where the risks lie and how negative impacts can be diminished.

Funding
This work was supported the Data to Decisions Cooperative Research Centre (D2D CRC) [DC52001].

Acknowledgment
The authors would like to thank Professor Louis de Koker and Dr Alana Maurshat for contributing to
the project design and the fieldwork. We also thank the D2D CRC for their assistance with the
recruitment of research participants. The valuable research assistance provided by Daniel Cater and
Brigid McManus is also gratefully received. We thank all research participants for their generosity in
responding to our invitations and Gavin Smith for his insightful comments on an earlier draft of this
article. The views expressed in this paper do not represent those of the D2D CRC.

References
Barocas, S. and Selbst,A.D. (2016), ‘Big Data’s Disparate Impact’, California Law Review, 104,
forthcoming.

Electronic copy available at: https://ssrn.com/abstract=3742516

Bennett Moses, L., and Chan, J. (2014), ‘Using Big Data for Legal and Law Enforcement Decisions:
Testing the New Tools’, University of New South Wales Law Journal, 37: 643–78.

Bennett Moses, L., and Chan, J. (2016), ‘Algorithmic Prediction in Policing: Assumptions, Evaluation,
and Accountability’, unpublished paper.
Bijker, W. E. (2010), ‘How is Technology Made? – That is the Question!’, Cambridge Journal
of Economics, 34: 63–76.
Bourdieu, P. (1987), ‘What Makes a Social Class? On the Theoretical and Practical Existence
of Groups’, Berkeley Journal of Sociology, 32: 1–18.
boyd, d., and Crawford, K. (2012), ‘Critical Questions for Big Data: Provocations for a
Cultural, Technological, and Scholarly Phenomenon’, Information, Communication and
Society, 15: 662–79.
Chan, J. (2001), ‘The Technological Game: How Information Technology is
Transforming Police Practice’, Criminology and Criminal Justice, 1: 139–59.
Chan, J. B. L. (2003), ‘Police and New Technologies’, in T. Newburn, ed., Handbook of
Policing, 655–79. Willan.
Chan, J., and Bennett Moses, L. (2016), 'Is Big Data Challenging Criminology?', Theoretical
Criminology, 20: 21–39.
Cope, N. (2003), ‘Crime Analysis: Principles and Practice’, in T. Newburn, ed., Handbook of
Policing, 340–62. Willan.
Davidson, E. (2006), ‘A Technological Frames Perspective on Information Technology and
Organizational Change’, Journal of Applied Behavioural Science, 42: 23–39.
Ericson, R. V., and Haggerty, K. D. (1997), Policing the Risk Society. Oxford University Press.
Goold, B., Loader, I and Thumala, A. (2013), ‘The Banality of Security: The Curious Case of
Surveillance Cameras’, British Journal of Criminology, 53: 977–996.
Kitchin, R. (2014), The Data Revolution: Big Data, Open Data, Data Infrastructures and Their
Consequences. Sage Publications.
Koper, C.S., Lum, C. and Willis, J.J. (2014), ‘Optimizing the Use of Technology in Policing: Results
and Implications from a Multi-Site Study of the Social, Organizational, and Behaviourial Aspects of
Implementing Police Technologies’, Policing, 8(2): 212–221.
Mackenzie, D., and Wajcman, J. (1985), The Social Shaping of Technology: How the
Refrigerator got its Hum. Open University Press.
Manning, P.K. (2014), ‘Information Technology and Police Work’, in G. Bruinsma and D.
Weisburd, eds., Encyclopedia of Criminology and Criminal Justice,2501–2513. Springer.
Manning, P. K. (1992), ‘Information Technologies and the Police’, in M. Tonry, and N. Morris,
eds., Modern Policing – Crime and Justice: A Review of Research, vol. 15, 349–98. University of
Chicago Press.
Manning, P. K. (1996), ‘Information Technology in the Police Context: The “Sailor” Phone’,
Information Systems Research, 7: 52–62.
McCahill, M. (2015), ‘Theorizing Surveillance in the UK Crime Control Field’, Media and
Communication, 3(2):10–20.
Nogala, D. (1995), ‘The Future Role of Technology in Policing’, in J. P. Brodeur, ed.,
Comparisons in Policing: An International Perspective, 191–210. Aldershot: Avebury.
Olesker, A. (2012), ‘White Paper: Big Data Solutions for Law Enforcement’. CTOlabs.com,
available online at http://ctolabs.com/wp-
content/uploads/2012/06/120627HadoopForLawEnforcement.pdf.

Electronic copy available at: https://ssrn.com/abstract=3742516

Orlikowski, W. J., and Gash, D. C. (1994), ‘Technological Frames: Making Sense of Information
Technology in Organisations’, ACM Transaction on Information Systems, 12: 174–207.
Podesta, J., Priztker, P., Moniz, E. J., Holdren, J., and Zients, J. (2014), Big Data: Seizing
Opportunities, Preserving Values. Executive Office of the President (US), available online at
https://www.whitehouse.gov/sites/default/files/docs/big_data_privacy_report_may_1_2014.pdf
Sackmann, S. (1991), Cultural Knowledge in Organizations. Sage Publications.
Sanders, C.B., Weston, C. and Schott, N. (2015), ‘Police Innovations, “Secret Squirrels” and
Accountability: Empirically studying intelligence-led policing in Canada’, British Journal of
Criminology, 55: 711–729.
Sheptycki, J. (2004), ‘Organizational Pathologies in Police Intelligence Systems: Some
contributions to the lexicon of intelligence-led policing’, European Journal of Criminology 1(3):
307–332.
Staniforth, A., and Akhgar, B. (2015), ‘Harnessing the Power of Big Data to Counter International
Terrorism’, in B. Akhgar, G. B. Saathoff, H. Arabnia, R. Hill, A. Staniforth, P. Bayerl, eds.,
Application of Big Data for National Security: A Practitioner’s Guide to Emerging Technologies,
23–38. Elsevier.
Swartz, D. (1997), Culture & Power: The Sociology of Pierre Bourdieu, Chicago: University of
Chicago Press.
Valverde, M. (2014), ‘Studying the Governance of Crime and Security: Space, Time and
Jurisdiction’, Criminology & Criminal Justice, 14: 379–91.
Weick, K. E. (1995), Sensemaking in Organizations. Sage Publications.

Weick, K. E., Sutcliffe, K. M., and Obstfeld, D. (2005), ‘Organizing and the Process of Sensemaking’,
Organization Science, 16: 409–21.
Wyllie, D. (2013), ‘How “Big Data” is Helping Law Enforcement’. PoliceOne.com, available
online at https://www.policeone.com/police-products/software/Data-Information-Sharing-
Software/articles/6396543-How-Big-Data-is-helping-law-enforcement/.

Electronic copy available at: https://ssrn.com/abstract=3742516

Tables and Figures
Figure 1 Conception of Big Data by Type of Organisation Employing Participants (n=38)

25

20

15

10

0
Aggregated Beyond
Analytic/
Volume / integrated human/est. Velocity Marketing Variety Veracity
prediction
data technology
Operational 10 8 6 4 2 0 2 1
Technical 4 2 0 2 2 5 1 0
Policy 8 3 3 0 1 0 1 1
TOTAL 22 13 9 6 5 5 4 2

*Note: Multiple responses can be coded for each research participant. There were 19 participants from Operational, 7 from Technical and 12 from Policy organisations.

Electronic copy available at: https://ssrn.com/abstract=3742516

Figure 2 Perceived Capability and Value of Big Data by Type of Organisation (n=38)

20

18

16

14

12

10

0
Better
Proactive/pre Improved Need caution
Analytic Richness of Investigative government Confidence in
ventative efficiency/ re Don’t know
capacity data advantage decisions/ data
policing effectiveness expectations
services
Operational 7 4 0 5 2 0 1 1 0
Technical 4 3 4 0 3 3 0 0 0
Policy 8 2 3 1 1 1 2 0 1
TOTAL 19 9 7 6 6 4 3 1 1

*Note: Multiple responses can be coded for each research participant. There were 19 participants from Operational, 7 from Technical and 12 from Policy organisations.

Electronic copy available at: https://ssrn.com/abstract=3742516

Figure 3: Risks of Using Big Data by Research Participant Organisation (n=38)

14
12
10
8
6
4 Operational

2 Technical
0 Policy
Data Misplaced trust Political and
Public TOTAL
Privacy security/integrit Misuse of data in technology / reputational Overload Discrimination
perceptions
y algorithms risks
Operational 5 5 2 6 6 5 4 0
Technical 1 0 2 3 0 0 0 0
Policy 6 6 6 1 1 1 0 1
TOTAL 12 11 10 10 7 6 4 1

*Note: Multiple responses can be coded for each research participant. There were 19 participants from Operational, 7 from Technical and 12 from Policy organisations.

Electronic copy available at: https://ssrn.com/abstract=3742516

Notes
1
Orlikowski and Gash’s (1994) notion of ‘technological frames’ differs from Bijker’s (1995) in that
the former involves socio-cognitive structures, whereas the latter involves only social structures
(see Davidson 2006: 37).
2
The project received human research ethics approval from the universities involved in December
2014.

Electronic copy available at: https://ssrn.com/abstract=3742516