Content uploaded by Sonia Livingstone
Author content
All content in this area was uploaded by Sonia Livingstone on Jan 28, 2017
Content may be subject to copyright.
Sonia Livingstone
Online risk, harm and vulnerability:
reflections on the evidence base for child
Internet safety policy
Article (Published version)
(Refereed)
Original citation:
Livingstone, Sonia (2013) Online risk, harm and vulnerability: reflections on the evidence base
for child Internet safety policy. ZER: Journal of Communication Studies, 18 (35). pp. 13-28.
ISSN 1137-1102
© 2013 University of the Basque Country/ Euskal Herriko Unibertsitatea
This version available at: http://eprints.lse.ac.uk/62278/
Available in LSE Research Online: June 2015
LSE has developed LSE Research Online so that users may access research output of the
School. Copyright © and Moral Rights for the papers on this site are retained by the individual
authors and/or other copyright owners. Users may download and/or print one copy of any
article(s) in LSE Research Online to facilitate their private study or for non-commercial research.
You may not engage in further distribution of the material or use it for any profit-making activities
or any commercial gain. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE
Research Online website.
Online risk, harm and vulnerability:
Reections on the evidence base for child Internet
safety policy
Internet darabilten adingabekoen segurtasun
politikarako informazioaren inguruko gogoetak
Riesgos, daños y vulnerabilidad online:
Reexiones sobre la información para la política de
seguridad de los menores en Internet
Sonia Livingstone1
Recibido el 17 de septiembre de 2013, aceptado el 8 de noviembre de 2013.
Abstract
After a decade or more in which research has examined the opportunities and risks encoun-
tered by children on the internet, this article assesses the contribution and challenges of pro-
ducing an evidence base to inform policy in a hotly contested eld. It offers critical analysis
and new ndings, drawing on the EU Kids Online project, a major study of children’s inter-
net use in 25 countries. Building on the distinction between risk (a calculation based on the
probability and severity of harm), and harm itself, research and policy on children’s online
risk faces particular problems in measuring harm and, therefore, risk. Further complications
arise from the interdependencies among opportunity, risk-taking, resilience and vulnerability.
Such complexities must be recognised if we are to advance beyond the entrenched positions
that so often polarise debate.
Kewwords: Internet, minors, childhood, risk, security.
Laburpena
Hamarraldi bat edo gehiago eman du ikerkuntzak adingabekoek Interneten aurkitzen dituz-
ten aukera eta arriskuak aztertzen. Hori kontuan izanik, eremu oso eztabaidagarri batean
politika sustatzeko informazio-base baten sorrerak dakartzan ekarpen eta erronkak aztertzen
ditu artikulu honek. Hala, EU Kids Online egitasmoa –adingabekoen Interneten erabilera 25
herrialdetan aztertzen duen ikerlan garrantzitsu bat– oinarri harturik, analisi kritiko bat eta
1 The London School of Economics and Political Science (LSE), s.livingstone@lse.ac.uk
zer
Vol. 18 - Núm. 35
ISSN: 1137-1102
pp. 13-28
2013
14
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
emaitza berriak eskaintzen ditu. Arriskua (probabilitatearen eta kalte-mailaren arabera neurt-
zen dena) eta kaltea bera argi bereiziz, online dabiltzan adingabekoen arriskuen inguruko
ikerketak zein politikak kaltea –eta, ondorioz, arriskuak ere– neurtzeko arazo espezikoak
jasaten dituzte. Halaber, arazo gehiago sortzen dira aukera, arrisku onartze, erresistentzia
eta zaurgarritasunaren arteko interdependentziatik. Zailtasun horiek aintzat hartu behar dira
sarri-askotan eztabaida polarizatzen duten jarrera hertsiak gainditu nahi baldin baditugu.
Gako-hitzak: Internet, adingabekoak, haurtzaroa, arriskua, segurtasuna.
Resumen
Después de una década o más en la que la investigación ha examinado las oportunidades y
los riesgos que los menores encuentran en Internet, este artículo estudia la contribución y re-
tos que implican la creación de una base de información para motivar la política en un campo
muy controvertido. Presenta un análisis crítico y nuevos resultados, basados en el proyecto
EU Kids Online, un importante estudio sobre el uso de Internet de los menores en 25 países.
Haciendo una clara distinción entre el riesgo (que se calcula sobre la base de la probabilidad
y el grado de daño), y el daño mismo, la investigación y la política sobre los riesgos de los
menores online afrontan problemas especícos para medir el daño y, por lo tanto, los riesgos.
Más complicaciones surgen de las interdependencias entre oportunidad, asunción del riesgo,
resistencia y vulnerabilidad. Tales complejidades deben ser reconocidas si queremos superar
las posturas atrincheradas que tantas veces polarizan el debate.
Palabras clave: Internet, menores, infancia, riesgo, seguridad.
15
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
1. Developing the agenda for research and policy2
Most children and young people in the world’s wealthy countries use the internet
at home, school and elsewhere. As ever more families, schools and communities
gain broadband and mobile access, online activities are becoming thoroughly em-
bedded into the timetables and spaces of children’s daily lives. Researchers and po-
licy makers, along with the wider public, are working hard to grasp the signicance
of the resulting socio-technical changes in the conditions of communication, socia-
lisation, learning and participation. On the one hand, society is beginning to recog-
nise the considerable opportunities the internet affords its users, including the intri-
guing and complex digital literacies that children are gaining and, in consequence,
the many and diverse benets of going online. On the other hand, there is growing
concern that these online opportunities are accompanied by an equally diverse array
of risks. Professionals in law enforcement, clinical practice and child protection, as
well as the general public, are increasingly calling for action and, to guide this, for
better knowledge of the actual harms associated with internet use.
Implicitly, if not always explicitly, policy initiatives assume particular motives,
knowledge and practices on the part of children. These assumptions may be well
founded or, instead, unnecessarily anxious or already dated. If we think young peo-
ple are living their leisure lives alone in their bedrooms, we will take a different view
of social support they may need compared with if we see them as richly embedded
in their peer group. If we see them as strong and able to cope with what life throws
at them, the policy agenda will take a different direction compared with if we see
them as at risk. Here lies the value of direct research with children and their lifeworld
online as well as ofine.
Ten years ago, when I reviewed the research literature on children and the inter-
net, I found so little I could barely write a review (Livingstone, 2003). Much more
has been conducted since, but there have still been difculties. Each new survey,
not always with a robust methodology, risks engendering a media panic. Research-
ers have asked all sorts of questions of children, not knowing really what to look
for or how to ask children ethically difcult questions. Too much research has been
couched in universalistic terms, though in reality most of it is from the US, raising
uncertainties about relevance elsewhere and underplaying contexts and comparisons.
Research seemed to go out of date the minute it was published – so we now know a
lot about getting a personal computer at home, learning html or visiting chatrooms,
but does any of this still matter?
Policy makers have had their parallel struggles, particularly regarding the moral
visions of childhood innocence, parental competence or trust in government that
frame decision making. These also make life difcult for researchers trying to build
a useful evidence base. For instance, discussions of risk and safety initially occurred
in domains wholly disconnected from discussions of educational or civic benet,
each seemingly ignorant of their signicance for the other. Too few adults (in the
2 This article draws on the work of the EU Kids Online network (www.eukidsonline.net), funded
by the EC Safer Internet Programme. I thank those in and beyond the network with whom I have
discussed these ideas, especially David Finkelhor, Anke Goerzig, Leslie Haddon, Ellen Helsper and
Janis Wolak.
16
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
early days, less so today) knew as much about the internet as children, so the rheto-
ric of the ‘digital native’ – now roundly critiqued (see Helsper and Eynon, 2010)
– found fertile ground, undermining the ability of parents and teachers to manage
this medium with the condence that they had managed previous media in the home.
In regulatory debates, opposing opinions from libertarians or the moral majority
pre-dominated, the rst fearing any government intervention, the second calling for
national control over the global internet. Neither, it seemed, welcomed a nuanced,
context-dependent account of how the internet use both shapes and is shaped by
children’s lives (Livingstone, 2011). The dominant metaphors were too extreme: the
virtual or cyber – an unreal realm oating in the ether where nothing really matters;
the Wild West – a vision of the internet where the natural wilderness impedes soci-
ety’s efforts to regulate, and rightly so if innovation is to ourish; or a paradise for
anarchists or inventors, pornographers or paedophiles.
But any serious examination of how the internet could or should be regulated,
why and by whom requires recourse to the evidence base. Fortunately, in recent
years, there has been an explosion in the volume of research, and some key insights
have resulted. Much of this focuses on online opportunities; while important, this is
not my present focus. Rather, this article examines three issues: the nature of online
risk, of harm and, linking the two, the nature of vulnerability. In each case, the focus
is on the particular conditions afforded to children by the internet.
Consider an illustrative survey nding. The Pew Research Center’s Internet
& American Life project reported that 15% of 12- to 17-year-olds with a mobile
phone had “received sexually suggestive nude or nearly nude images of someone
they know” (boyd and Hargittai, 2010: 2). ‘Sexting’ quickly became the latest
risk, with policy makers, law enforcement and educators springing into action. But
some pressing questions arise. How can we measure the prevalence of ‘sexting’
(how is it dened, can we ask young people ethically, will they report it truthful-
ly?)? Does it matter, and is it harmful? If it is, which teenagers fall into the 15%?
In other words, who is vulnerable? Last, what can be done? Is this harm new, or
worse than before, and is the internet or mobile phone therefore culpable? If we
don’t ask these questions, public perceptions may conclude that all children are
‘at risk’, thereby fuelling the media-amplied moral panics that result in anxious
calls to restrict children’s internet access, increase surveillance or legislate against
online freedoms.
So ask them we must. But rather than building in an assumption that the internet
is to blame (by asking what the internet is doing to childhood) or even, grounding
our inquiry in research on the internet, since this is relatively new, I suggest that
we begin by learning from the long-established tradition of research and policy on
the nature of risk, harm and vulnerability ‘ofine’, including the psychological and
sociological analysis of risk in children’s everyday lives (Bradbrook et al., 2008;
Coleman and Hagell, 2007; Feinstein and Sabates, 2006; Finkelhor, 2008; Munro,
2008; Schoon, 2006). In what follows, I show how this literature is useful, both
for its body of empirical ndings regarding the array of risk and protective fac-
tors that may apply online as they do ofine, and also conceptually, for clarifying
the confusion that still reigns in relation to the internet (Millwood Hargrave and
Livingstone, 2009).
17
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
2. On the nature of risk
What is risk? Beck (1986/2005: 21), social theorist of the ‘risk society’, argues that,
“risk may be dened as a systematic way of dealing with the hazards and insecuri-
ties induced and introduced by modernization itself.” Until the modern era, he ar-
gues, societies were preoccupied with natural hazards (such as ooding, volcanoes
or plagues). Since these are uncontrollable in themselves, people can only seek to
manoeuvre around them. By contrast, societies today are increasingly preoccupied
with risks of our own making, being “concerned no longer exclusively with making
nature useful, or with releasing mankind from traditional constraints, but also and es-
sentially with problems resulting from techno-economic development itself” (p. 19).
A risk, in short, stems from the conditions of modern life rather than from out-
side them (in a similar vein, Smillie and Blissett, 2010, distinguish a ‘natural haz-Smillie and Blissett, 2010, distinguish a ‘natural haz-
ard’ from a ‘technological hazard’). Unlike the problem of dealing with a volcano,
where one can only respond to an inevitable if unpredictable hazard, dealing with
the problem of online grooming or bullying invites us to anticipate risk when de-
signing the online environment as well as to consider how to respond to harm after
the event. Following Giddens (1991), in today’s reexive modernity one cannot be
innocent of even the unintended consequences of institutional actions, especially
given foreknowledge of the harms research has already revealed. To take a pertinent
example – to design a social networking site for small children without anticipating
possible abuses by ill-intentioned adults would be naïve. But how should such risks
be anticipated?
This claries the at-times confused discourse of child online safety. Being ex-
posed to pornography online is a risk in the sense that it is associated with a certain
likelihood and magnitude of harm. Hence it is important that the evidence base mea-
sures these. However, the identication of online risk does not imply that harm will
follow, and nor that all users will be equally affected. Rather, risk may be judged
(according to a simpler or more complex calculation; see Hansson, 2010) by taking
into account the particular and contingent interaction between user and environment.
Then, the risk may be dealt with by conducting a risk evaluation (asking to whom
this risk matters and why), which, in turn, establishes the legitimacy of risk manage-
ment (such as the development of regulatory institutions or user tools and tactics; see
Klinke and Renn, 2001). Following Beck, online risk does not arise inevitably but as
a result of human design. This risk management can be proactive as well as reactive;
and it can also focus on the actions of the individual (the user) or the design of the
socio-technical environment (such as the online site or service), or both. Or indeed,
nothing may be done – the risk may be judged acceptable (i.e., it may be tolerated,
up to a point), or there may be political or economic impediments unrelated to the
risk assessment.
Thus far, matters seem straightforward. But recall what is typically measured
in online safety surveys – that 15% of teenagers receive sexual messages or that a
certain fraction have seen online pornography or been cyberbullied. These ‘online
risks’ in ‘cyberspace’ are often compared with the ‘ofine risk’ of a child crossing
the road. For the latter, the risk is rst calculated and then managed (by inuen-
cing children by teaching them to cross safely and by regulating the environment
18
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
– cars, roads, town planning). The parallel for cyberspace is productive, and many
have talked of road safety rules online as well as ofine (Criddle, 2006). But the
analogy faces a problem in relation to children’s online safety. In the case of road
accidents, the risk to the child is dened as the probability of an accident (calcu-
lated by dividing the number of children hurt in a particular way on the roads by
the number of children in the population) multiplied by its severity (in terms of
consequences, which can range from minor bruising to death). Risk, harm and the
relation between them are as clear (or unclear) as the measurements of probability
and severity are accurate.
But on the internet, we do not know how many children are hurt, or how severe
are the consequences; there are no accident gures. If the ofine were like the online,
it would be like knowing, only, how many children report crossing a road and per-
haps, how many report that something bad happened in consequence. On the other
hand, if the online were like the ofine, we would also know the online equivalent of
how many cars were on the road and how fast they were driving (e.g. exactly what
pornography they saw or how they were cyberbullied or groomed); most important,
we would know whether an accident resulted (i.e., whether the child suffered harm-
ful consequences, for how long and with what severity). But for the most part, over
a decade of surveys have asked children whether they saw something inappropriate
but have generally not asked exactly what they saw, and few have asked (if it can
be asked) whether this exposure harmed them. Some researchers have made greater
efforts to gain a more exact picture of what happened – what they saw or what was
said to them, in relation to pornography (Peter and �alkenburg, 2009), sexual harass-(Peter and �alkenburg, 2009), sexual harass-, sexual harass-
ment (Mitchell, Finkelhor and Wolak, 2007) and cyberbullying (Smith, Mahdavi and
Carvalho, 2008).
But this is still to get a closer picture of what was happening on the road rather
than what happened to the child. How, then, can risk be calculated? This is where the
road analogy breaks down, for in relation to online risk, survey researchers dene
risk not as the probability of harm, but as the probability of an encounter that might
(or might not) result in harm. In other words, they calculate the number of children
who encounter pornography or a cyberbullying message or a grooming attempt, and
divide that number by the number of children online. What is reported, therefore, is
not the actual risk (i.e., the probability of harm to the child population) but the risk
of the risk (the probability of something happening – commonly called ‘online risk’)
that might result in harm; but whether it does, and for how many it does, remains
unknown. It is like reporting the risk of road accidents in terms of the likelihood of
children crossing the road (i.e., the proportion of children who cross the road divided
by the number of children altogether) rather than in terms of the risk of their being
hurt. No wonder sceptics ask, so what? And no wonder policy makers are hesitant
about the legitimacy of risk management.
3. Confusion about the nature of harm
Without good evidence of harm resulting from online encounters, without even
a clear picture of the nature of those encounters, we cannot really speak of risk,
at least not if we were to rely on established approaches to the calculation and
19
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
management of risk in society. Why is harm proving so difcult for research into
children and their use of the internet, notwithstanding the huge public and policy
interest in this question?
Most obviously, what is meant by ‘harm’ in relation to online ‘risks’ is often
unclear. The nature of a road accident is far less contested, and its consequences
are generally straightforwardly assessed, although there can be arguments in
court about the time required for recovery or the psychological distress that may
accompany physical damage. But society has established mechanisms, instituted
by reputable authorities (clinical, legal, actuarial), to address these. For exposure
to pornography, or receiving hostile or racist messages, or visiting a self-harm
chatroom, or having one’s social networking prole trashed, what is the harm?
And is intervention justied? Opinion varies, one reason being that many of the
harms society is concerned about in relation to the internet are relatively minor;
indeed, bullying or exposure to pornography are generally ‘under the radar’ of
teacher or parent or welfare intervention when they also occur ofine. For the
severest risks (such as grooming, or prolonged exposure to extreme pornogra-
phy, or such sustained bullying that a child is driven to self-harm), it is often
assumed that, while the risk is encountered online, the harm will occur ofine,
so that child welfare organisations can respond to the harm according to familiar
practices. Another reason is that the claimed harms raise unresolved moral is-
sues over whether early exposure to the adult world is ‘normal,’ even desirable,
or problematic. The debates over internet-related harm, in other words, do not so
much concern the internet as societal conceptions of childhood – particularly in
relation to the place of sexuality and violence in childhood. The recognition that
we as a society have only recently created – and are further re-designing – the
internet, is stimulating some soul-searching about the childhood we have – and
still could – create for our children.
There are limits to the debates ongoing in society regarding internet-related harm
to children. It is widely accepted that receiving hostile or nasty messages from peers,
especially if sustained over time, constitutes a harm that merits institutional inter-
vention (by the schools, regulators or industry expected to reduce cyberbullying).
It is also generally held that parents are responsible for preventing young children
from accessing pornography, even if it is less agreed whether such access is inher-
ently damaging. In many countries it is also held that an online approach from an
adult to a child for sexual purposes is not merely unacceptable but illegal. Moreover,
while the boundaries – regarding a child’s age and development, the expectations on
parental responsibility, where to draw the line on the extremity of certain types of
content – remain difcult, these are after all familiar debates from the ofine world;
they just seem more intense because the rapidity with which the internet has entered
children’s lives means that society must engage in these debates anew, and often
from a position of ignorance regarding both the technology and the evolution of
children’s practices of use.
It is striking how little these debates over harm refer to evidence or theory from
experts on child welfare, especially by comparison with the blizzard of statistics
from researchers regarding online risk. Every day brings a new survey on children’s
exposure to pornography or strangers or bullying. But there is remarkably little
20
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
discussion of what exactly is thought to be the problem – is pornography a problem
because it upsets or shocks children, or because it distorts their conception of sexual-
ity, or because it puts pressure on girls to perform certain sexual acts? My intention
is not necessarily to question any of these claims, but to invite their explication in
discussions of online risk so that it is clear what, online, is the equivalent of accidents
on the road, ofine. Only then can we as researchers strengthen the evidence base
regarding robust indicators of harm. At present, the literature is sparse regarding the
harms that may result from internet use, especially if one applies the standards of
high quality research – representative samples, careful questioning of children about
the nature of harm, longitudinal research designs that permit assessment of harm
over time, procedures to identify risk and protective factors that pinpoint which chil-
dren are harmed and why.
So where is the equivalent of road accident statistics in relation to the internet?
Such ‘objective’ evidence of harm might be expected from law enforcement, clini-
cians or child welfare services, for example, in cases where the internet is involved
in incidences of sexual abuse or criminal abduction, youth suicide or self-harm at-
tempts. But surprisingly little such evidence has been forthcoming, partly because
the particular involvement of the internet in cases addressed by such authorities is
not reliably recorded in police or clinical records (although see Wolak and Finkelhor,
2013). Thus while surveys can tell us how many children go to an ofine meeting
with an online contact, it is difcult to link this to the cases of actual abuse that come
to the attention of law enforcement or welfare services. For instance, in the UK, 4%
of 9- to 16-year-olds report having gone to an ofine meeting with someone rst
met online (Livingstone et al., 2010), while the UK’s Child Exploitation and Online
Protection Centre (CEOP, 2013) receives reports from around 1,000 children each
year concerning online victimisation by adults. Assuming around seven million in
this age group, a rough-and-ready calculation might put this as a risk of 1:300, as a
ratio of those who go to such a meeting and do or do not come to harm (and ofcially
report it). But this is to make many assumptions!
Not only is it difcult to discover whether a risk results in harm, but it is also
difcult to discover when the internet plays a role in known harm. The UK ma-
jor children’s charity, the NSPCC, estimates that 5% of UK children suffer contact
sexual abuse at some point during childhood, with some 10,000 new victims each
year (Harker et al., 2013); if we try to link this to the above gure of 1,000 children
who report online victimisation by adults, we might assume abuse by perpetrators
known to the child ofine remains much more common than grooming by strangers
online. But again, this makes many assumptions since, on the one hand, many cases
of sexual abuse go unreported and, on the other, when they are reported the possible
role of the internet is rarely considered. In short, even 10 or more years after the ad-
vent of mass internet, researchers and policy makers still rely on incidents learned of
ad hoc, often from the mass media, and it is very difcult to gauge what proportion
of the population they represent. Clearly, the situation regarding online risk is quite
different from the situation regarding road accidents, where policy makers rely less
on surveys about whether children cross roads than on objectively veried statistics
of road accidents involving children3.
3 Of course, I acknowledge the lively and difcult debates over the veracity of crime and health statistics.
21
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
4. Pragmatic solutions
How, then, can we proceed, in the interests of evidence-based policy making?4.
Let us make a start. Three dimensions of harm are surely pertinent. First, the type
of harm: this may include any or all of physical harm (e.g. bodily attack), emo-
tional (e.g. feeling upset, threatened or distressed), psychological (e.g. low self-
esteem, distorted sense of sexuality, aggressiveness) and/or social harms (e.g. loss
of friends, being ostracised). The second and third dimensions are simple in con-
ception, if difcult to measure – harm associated with online risk varies in terms
of severity and longevity (from immediate or short-term consequences to longer
or even lifelong effects). If every study henceforth claried its assumptions about
harm in these terms, our assessment of the evidence base would be far more in-
formative. In terms of theory, this should be possible. But the measurement chal-
lenge is a substantial one. There is an emerging consensus on how to ask children
about online risk (the range of risks, agreement over phrasing, response options,
etc.) but not yet about how to measure harm, or even whether this is possible at all.
There is much one cannot ask children about, for ethical reasons, and so there are
difculties in establishing just what children mean by ‘pornography’ or being ‘up-
set’. Children may not be in a position to judge the harm done to them, especially
as it may take years to be revealed.
Nonetheless, asking children about harm directly is one way forward and, while
it has limitations, it is equally problematic not to ask them; children’s voices, and
their experiences, must also be heard in the public policy debates regarding their
well-being and their best interests. In the EU Kids Online survey, conducted in 25
countries in 2010 (Livingstone et al., 2011a5), we sought to gain the most that could
be learned from a self-report survey, asking children not only whether they have
encountered pornographic or hostile messages, for instance, but whether this had
upset or bothered or distressed them. The ndings showed that 14% of online 9-
to 16-year-olds have seen images online that are ‘obviously sexual – for example,
showing people naked or people having sex’ in the past 12 months, 6% have been
sent nasty or hurtful messages online, and 9% have met an online contact ofine in
the past year. Further, 15% of 11- to 16-year-olds have received peer-to-peer ‘sexual
messages or images … [meaning] talk about having sex or images of people naked
or having sex,’ and 21% have been exposed to one or more types of potentially
harmful user-generated content: hate (12%), pro-anorexia (10%), self-harm (7%),
drug-taking (7%) or suicide (5%). But our gures on self-reported harm are equally
important: of the 9- to 16-year-olds who had been exposed to online sexual images,
one in three said that they were bothered by the experience (and of those, half were
fairly or very upset by what they saw); of those who had received nasty or hurtful
messages online, between half and two thirds had been fairly or very upset; and
of those who had met an online contact ofine, one in six were bothered by what
happened (and about half of those said that they were very or fairly upset by what
happened). Similarly, of the 11- to 16-year-olds who had seen or received a sexual
4 The trials and tribulations of evidence-based policy making merit a separate examination (see Li-
vingstone, in press).
5 For the Spanish report, see Garmendia et al. (2011).
22
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
message online, nearly a quarter had been bothered by this (and nearly half of those
were fairly or very upset)6.
Our effort to construct a condential, detailed survey with children allowed us to
build on the insights of research on risk in childhood. The resulting evidence base
has substantially informed policy makers’ decisions and public understanding of on-
line risks to children, clarifying both the prevalence of online risk (broken down by
demographic and country variables) and the prevalence of harm (Livingstone, in
press). It was especially important to establish that only a subset of those children
exposed to online risks reported experiencing any harm, for this helped to diffuse
the moral panics (‘all children are at risk’, ‘the internet is bad for kids’) and open up
more productive questions about vulnerability and resilience. In other words, it is
timely to ask, which children are more at risk of harm, and why? In our survey, ex-
amination of the diverse contexts of children’s lives helped to identify which risk and
protective factors help account for patterns of risk, harm and vulnerability online. In
terms of risk factors, the research revealed the importance of social psychological
factors on the part of the child (such as facing psychological difculties or having a
tendency to sensation-seeking); as for protective factors, children’s self-esteem and
their parents’ strategies for mediating the internet were shown to matter, though not
in any simple fashion.
Overall, the growing evidence base suggests that those children who are vulner-
able or at risk ofine are more likely also to be at risk online, thereby compounding
cycles of disadvantage (Bradbrook et al., 2008), although some specic factors must
be considered in relation to the nature or affordances of the online environment (see
the evidence and analysis collected in Livingstone et al., 2012). This strengthens
the claim that online risk, harm and vulnerability (or, its opposite, resilience) can
be researched by building on the literature for ofine risk in children’s lives. Many
questions remain. Since it seems that children in disadvantaged or ‘at risk’ life cir-
cumstances are more likely than those in ‘normal’ circumstances to be vulnerable of-
ine, more research is needed to understand how children’s life circumstances shape
their online experience and whether the same risks ‘migrate’ online (e.g. sexual risks
encountered ofine are similar to or even linked to sexual risks encountered online;
see Görzig, 2011; Livingstone and Görzig, 2012), creating a vicious cycle. Then,
since the variance explained by ofine risk factors is fairly low, it remains to be
understood whether some children are also newly at risk now that they have internet
access – do they behave differently, or reveal different sides of their identity online,
putting them more at risk online than ofine? Last, it may be that for children who
are vulnerable ofine, the internet provides a safe haven, a space where risk does not
follow them, so that new ways of behaving, even new sources of resilience, can be
developed, possibly beneting their circumstances ofine as well as online.
For the present, it has been useful to establish that not all exposed to online risk
report harm as a result. The evidence thus counters the assumption of some policy
makers that risk and harm are one and the same – that to see pornography is to be
harmed by it, to be approached by a stranger online is to be damaged inevitably, and
so forth. Rather, the conditions under which risks (whether seeing online pornography
6 Note that due to time constraints, the survey did not ask children about possible harm associated with
exposure to negative user-generated content.
23
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
or crossing the road) result in harm are complex. What is needed is an analysis of
the complex set of contingencies that mediate the relation between risk and harm
in accounting for whether, when and why some children are vulnerable to online
risk. This analysis must include an account both of individuals (themselves diverse,
depending on life contexts) and the socio-technical environment with which they
engage (and behind which lie the institutions that shape them).
While there is no escape from the limits of self-report data (in which both under-
reporting and over-reporting by children is likely)7, we refuse the strong claim that
harm to children associated with online risk simply cannot be measured. This would
result in evidence-free rather than evidence-based policy, permitting sceptics to dis-
miss further consideration of the internet as affording children harm or permitting
the moral majority to call for action on the assumption that all risk results in harm.
Rather, the EU Kids Online network has taken the view that the onus is on research-
ers to do the best job they can in terms of devising appropriate methods to ask sensi-
tive questions of children (Lobe, Livingstone and Haddon, 2007), being transparent
as to methodological decisions and their limitations, in the knowledge that policy
makers can invoke the precautionary principle on occasion to legitimate policy ac-
tion in the absence of evidence (Klinke and Renn, 2001).
5. Complications
Three important problems remain unresolved by the above analysis. The rst is that
the tripartite analysis of risk assessment, evaluation and management does not take
into account the benets of internet use, failing to balance opportunities against risks.
It can seem that if a child is to avoid online harm, they must avoid almost all online
activities: to post content online, you must provide personal details; to make new
friends, you must contact ‘strangers’; to explore diverse information may expose
you to inappropriate content; to seek guidance on dieting will result in receipt of pro-
anorexic advice, and so forth. As Livingstone and Helsper (2010) showed, children’s
take-up of online opportunities is positively correlated with their exposure to online
risk, with digital skills acting to increase the likelihood of both. One reason is that
the same act (e.g. making a new contact) can result in either an opportunity (more
friends) or the risk of harm (meeting an abusive stranger). Making a new contact
cannot, therefore, be straightforwardly described as either harmful or harmless – the
context is crucial.
Here the road accident analogy is again useful. Society neither prevents children
from crossing the road nor permits them to run freely across the motorway. Rather,
it takes the concerted efforts of parents, teachers, car designers, road authorities and
town planners to strike an acceptable balance between children’s freedom to navi-
gate their neighbourhood and the attendant risks (Criddle, 2006; Livingstone, 2009).
Online, unless children are to live in heavily ltered environments with Facebook
and YouTube banned and adults always peering over their shoulders, a better reso-
lution than a simple trade-off of opportunities and risks must be found. In policy
terms, this points to the need for further development of parental and school mediation,
7 See Livingstone et al. (2011b) for our best efforts rst to minimise, and then to assess the conse-
quences of the methodological limitations of the research.
24
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
media literacy education, provision of technical tools for users, and improved in-
built or default safety and privacy in the design of online sites and services. Such
improvements are underway, but the importance of keeping the main purpose in
mind – namely, to facilitate children’s online opportunities – is crucial. In relation to
children’s outdoor play, for instance, it appears that the efforts towards risk manage-
ment (soft surfaces, safety rails, playground attendants, etc.) have not, in practice,
freed children to play as they would wish; instead, an overly risk-averse culture has
resulted that prevents children climbing trees or even swinging on swings without an
onerous risk assessment undertaken by supervising adults (Gill, 2007).
This leads to a second complication – the particular sensitivities over risk evalu-
ation in relation to children, for whom society nds it difcult to accept any degree
of risk above zero. Shaped by the media’s tendency to amplify risks, framing them
as threatening the innocence of children (Kitzinger, 2004) and undermining the hope
of an idealised, risk-free childhood (Kehily, 2010), for many parents risk anxiety
has become “a constant and pervasive feature of everyday consciousness” (Jackson
and Scott, 1999: 88). There is, it seems, an unmanageable gulf between the rational
balancing of probabilities matched to available policy tools and the unacceptability
of harm that will still occur to a particular child. Although researchers and policy
makers have learned to take great care in disseminating ndings and recommenda-
tions to the media and public, this remains a difcult issue to be factored into any risk
management strategy (Smillie and Blissett, 2010).
Third, as foreshadowed in the above, a world without risk is undesirable. Chil-
dren must learn to take calculated risks and, insofar as is possible, cope with the
consequences. Developmental psychologists are clear that facing and coping with
risk is important, for “resilience can only develop through exposure to risk or to
stress” (Coleman and Hagell, 2007: 15). As Luthar, Cicchetti and Becker (2000:
543) dene it, resilience is ‘a dynamic process encompassing positive adaptation
within the context of signicant adversity.’ The latter part of this denition is im-
portant – without experience of adversity, a child may be protected but has nothing
to adapt to positively and so will not become resilient. A risk-averse society will,
paradoxically, exacerbate rather than reduce the very vulnerabilities it seeks to pro-
tect by undermining the development of resilience. And for teenagers, risk-taking is
also important developmentally and culturally (Green, Mitchell and Bunton, 2000:
123-4). As Lupton (1999: 156) adds, the dominant discourse’s excessive emphasis
on safety generates its own counter-discourse: “risk-taking may be regarded as the
ipside of modernity, a response to the ever-intensifying focus on control and pre-
dictability of modernity.”
6. Conclusions
In reecting on the lessons to be learned for evidence-based policy, following the
conduct of the EU Kids Online survey, I have argued for a fundamental distinc-
tion between risk (a calculation based on probability and the likely consequences of
harm), and harm (a distinct outcome, whether measured objectively or subjectively).
I have further noted that the eld of children’s online risk faces particular problems
in measuring harm, and therefore also struggles to measure risk, instead tending to
25
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
measure the risk of a risk, and often leaving unknown the relation between risk and
harm. It is also worth making explicit that while risk can occur online as ofine, the
focus of harm is the child rather than the internet – put simply, harm is always suffe-
red ‘in the real world’. In this sense, while the internet has added new sources of risk
to children’s lives, the history of harm is as old as childhood. The harmful effects of
any online or online risk are to be understood, as they always have been, in terms
of physical harm, emotional distress, adverse psychological consequences or nega-
tive social outcomes. But today, as online and ofine increasingly intersect or blur
in fast-changing cycles of mutual inuence and connection, the risk and protective
factors that mediate the relation between risk and harm must be rethought.
In many ways, we can rely on the established literature on childhood risk to
propose the likely factors that increase or protect against risk of harm. In other
ways, still little explored, the changing socio-technological environment may add
new factors, or new interactions among factors, that researchers should explore
and that policy makers need to know about. It is encouraging to learn that, in terms
of the long-term statistics on child welfare (their mental or physical health or rates
of crime victimisation), there is little evidence that the conditions of childhood are
worsening (Finkelhor, 2008; Madge and Barker, 2007). Thus the internet is not,
in any simple terms, making matters worse. But the public’s fear of the internet
does seem to be restricting children’s online opportunities and, therefore, their life
chances in the long term. And it is likely that widespread use of the internet is also
altering the conditions under which risk of harm adversely affects children’s lives.
Robust, independent evidence is important to guide policy makers in their task of
risk assessment, a crucial precursor to then evaluating acceptable levels of risk and
developing policies to manage risk. I have further argued that they are impeded
in this work by a public reluctance to accept any risk to children (notwithstanding
growing public disquiet over a risk-averse culture of childhood) and, even more
important, by the inability of traditionally-framed risk calculations to take into ac-
count the benets of internet use in general and of learning to cope with tolerable
levels of online risk in particular.
Last, evidence of benets as well as harms is needed to enable a proportionate
balance between the opportunities and risks that the internet affords to children,
recognising that the opportunities and risks often go hand in hand when using the in-
ternet, and that striking this balance should be achieved differently for children who
are more vulnerable or more resilient. In short, there can be no simple translation of
online risks – or opportunities – into predictable outcomes, and each can result in
positive or negative outcomes for children. Here the concept of internet affordances
is valuable in reminding us that the internet is not intrinsically risky – everything
depends on the interaction between users and their socio-technological environment,
and the ways in which this interaction has been shaped. In some cases, online risks
may afford harm (whether measured subjectively or objectively), but in others, they
may facilitate resilience. Moreover, while online opportunities generally afford po-
sitive benets for children, the existence of those same opportunities can, if children
are restricted in accessing them, result in the negative outcome of digital exclusion.
Policy makers should, therefore, seek to address the challenges of online risk wi-
thout increasing children’s digital exclusion or leaving them vulnerable to harm. And
26
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
that means taking action to both improve the design of the online environment and
to enhance children’s resilience.
References
BECK, U. (1986/2005). Risk society: Towards a new modernity. London: Sage
Publications.
BOYD, D. and HARGITTAI, E. (2010). ‘Facebook privacy settings: Who cares?’
In: First Monday, 15(8).
BRADBROOK, G., ALVI, I., FISHER, J., LLOYD, H., MOORE, R., THOMPSON,
�., et al. (2008). Meeting their potential: The role of education and technology
in overcoming disadvantage and disaffection in young people. Coventry: Becta.
CEOP (Child Exploitation and Online Protection) Centre (2013). Threat assessment
of child sexual exploitation and abuse. London: CEOP Centre.
COLEMAN, J. and HAGELL, A. (eds) (2007). Adolescence, risk and resilience:
Against the odds. Chichester: Wiley.
CRIDDLE, L. (2006). Look both ways: Help protect your family on the internet.
Redmond, WA: Microsoft Press.
FEINSTEIN, L. and SABATES, R. (2006). Predicting adult life outcomes from ear-
lier signals: Identifying those at risk. London: Prime Minister’s Strategy Unit.
FINKELHOR, D. (2008). Childhood victimization: Violence, crime, and abuse in
the lives of young people. Oxford: Oxford University Press.
GARMENDIA, M., GARITAONANDIA, C., MARTÍNEZ, G. and CASADO, M.
(2011). Riesgos y seguridad en internet: Los menores españoles en el contexto
europeo. Available at www.sociologia.ehu.es/s0018-eukidshm/es/
GIDDENS, A. (1991). Modernity and self-identity: Self and society in the late mo-
dern age. Cambridge: Polity Press.
GILL, T. (2007). No fear: Growing up in a risk averse society. London: Calouste
Gulbenkian Foundation.
GÖRZIG, A. (2011). Who bullies and who is bullied online? A study of 9-16 year
old internet users in 25 European countries. London: EU Kids Online, Lon-
don School of Economics and Political Science. Available at http://eprints.lse.
ac.uk/39601/
GREEN, E., MITCHELL, W. and BUNTON, R. (2000). ‘Contextualizing risk and
danger: An analysis of young people’s perceptions of risk.’ In: Journal of Youth
Studies, 3(2), 109-126.
HANSSON, S.O. (2010). ‘Risk: objective or subjective, facts or values.’ In: Journal
of Risk Research, 13(2), 231-238.
HARKER, L., JÜTTE, S., MURPHY, T., BENTLEY, H., MILLER, P. and FITCH,
K. (2013). How safe are our children? London: NSPCC.
HELSPER, E.J. and EYNON, R. (2010). ‘Digital natives: where is the evidence?’
In: British Educational Research Journal, 36(3), 502-520.
27
Online risk, harm and vulnerability: Reections on the evidence base for child Internet safety policy
Zer 18-35 (2013), pp. 13-28
JACKSON, S. and SCOTT, S. (1999). ‘Risk anxiety and the social construction of
childhood.’ In: D. Lupton (ed.). Risk (pp. 86-107). Cambridge: Cambridge Uni-
versity Press.
KEHILY, M.J. (2010). ‘Childhood in crisis? Tracing the contours of “crisis” and its
impact upon contemporary parenting practices.’ In: Media, Culture & Society,
32(2), 171-185.
KITZINGER, J. (2004). Framing abuse: Media inuence and public understanding
of sexual violence against children. London: Pluto Press.
KLINKE, A. and RENN, O. (2001). ‘Precautionary principle and discursive strate-
gies: classifying and managing risks.’ In: Journal of Risk Research, 4(2), 159-174.
LIVINGSTONE, S. (2003). ‘Children’s use of the internet: Reections on the emerg-
ing research agenda.’ In: New Media & Society, 5(2), 147-166.
LIVINGSTONE, S. (2009) Children and the internet: Great expectations, challeng-
ing realities. Cambridge: Polity Press.
LIVINGSTONE, S. (2011). ‘Regulating the internet in the interests of children:
Emerging European and international approaches.’ In: R. Mansell and M. Raboy
(eds). The handbook on global media and communication policy (pp. 505-524).
Oxford: Blackwell.
LIVINGSTONE, S. (in press) ‘“Knowledge enhancement”: On the risks and op-
portunities of generating evidence-based policy.’ In: B. O’Neill, E. Staksrud and
S. McLaughlin (eds). Children and internet safety in Europe: Policy debates and
challenges. Gothenburg: Nordicom.
LIVINGSTONE, S. and GÖRZIG, A. (2012) ‘Sexting.’ In: S. Livingstone, L. Had-
don and A. Görzig (eds). Children, risk and safety on the internet: Research and
policy challenges in comparative perspective (pp. 151-164). Bristol: Policy Press.
LIVINGSTONE, S. and HELSPER, E.J. (2010). ‘Balancing opportunities and risks
in teenagers’ use of the internet: The role of online skills and internet self-efca-
cy.’ In: New Media & Society, 12(2), 309-329.
LIVINGSTONE, S., HADDON, L. and GÖRZIG, A. (eds) (2012). Children, risk
and safety online: Research and policy challenges in comparative perspective.
Bristol: Policy Press.
Livingstone, S., Haddon, L., Görzig, A. and ÓLAFSSON, K. (2010). Risks and safe-
ty on the internet: The UK report. London: EU Kids Online, London School of
Economics and Political Science.
LIVINGSTONE, S., HADDON, L., GÖRZIG, A. and ÓLAFSSON, K. (2011a).
Risks and safety on the internet: The perspective of European children. Full
ndings. London: EU Kids Online, London School of Economics and Politi-
cal Science.
LIVINGSTONE, S., HADDON, L., GÖRZIG, A. and ÓLAFSSON, K. (2011b) Te-
chnical report and user guide: The 2010 EU Kids Online Survey. London: EU
Kids Online, London School of Economics and Political Science.
LOBE, B., LIVINGSTONE, S., and HADDON, L., with others (2007) Researching
Children’s Experiences Online across Countries: Issues and Problems in Me-
thodology. Deliverable D4.1 for the EC Safer Internet plus programme. LSE,
London: EU Kids Online.
LUPTON, D. (1999). Risk. London: Routledge.
28
Sonia LIVINGSTONE
Zer 18-35 (2013), pp. 13-28
LUTHAR, S. S., CICCHETTI, D. and BECKER, B. (2000). The construct of resi-
lience: A critical evaluation and guidelines for future work. Child Development,
71(3), 543-562.
MADGE, N. and BARKER, J. (2007). Risk and childhood. London: The Royal So-
ciety for the Encouragement of Arts, Manufactures & Commerce.
MILLWOOD HARGRAVE, A. and LIVINGSTONE, S. (2009). Harm and offence in
media content: A review of the empirical literature (2nd edn). Bristol: Intellect Press.
MITCHELL, K.J., FINKELHOR, D. and WOLAK, J. (2007). ‘Online requests for
sexual pictures from youth: Risk factors and incident characteristics.’ In: Journal
of Adolescent Health, 41(2), 196-203.
MUNRO, E. (2008). Effective child protection (2nd edn). London: Sage Publications.
PETER, J. and �ALKENBURG, P.M. (2009). ‘Adolescents’ exposure to sexually
explicit internet material and notions of women as sex objects: Assessing causa-
lity and underlying processes.’ In: Journal of Communication, 59(3), 407-433.
SCHOON, I. (2006). Risk and resilience: Adaptations in changing times. New York:
Cambridge University Press.
SMILLIE, L. and BLISSETT, A. (2010). ‘A model for developing risk communica-
tion strategy.’ In: Journal of Risk Research, 13(1), 115-134.
SMITH, P.K., MAHDAVI, J. and CARVALHO, M. (2008). ‘Cyberbullying: its natu-
re and impact in secondary school pupils.’ In: Journal of Child Psychology and
Psychiatry, 49(4), 376-385.
WOLAK, J. and FINKELHOR, D. (2013). ‘Are crimes by online predators different
from crimes by sex offenders who know youth in-person?’ In: Journal of Adoles-
cent Health. [Available online 26 July 2013, 1-6.]
