Journal of Environmental Radioactivity 133 (2014) 86e91

Contents lists available at SciVerse ScienceDirect

Journal of Environmental Radioactivity

journal homepage: www.elsevier.com/locate/jenvrad

Radiation risk perception: a discrepancy between the experts and the

general population
Tanja Perko*
Belgian Nuclear Research Centre SCK CEN, Boeretang 200, B-2400 Mol, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: Determining the differences in the perception of risks between experts who are regularly exposed to
Received 23 November 2012 radiation, and lay people provides important insights into how potential hazards may be effectively
Received in revised form communicated to the public. In the present study we examined lay people’s (N ¼ 1020) and experts’
7 April 2013
(N ¼ 332) perception of five different radiological risks: nuclear waste, medical x-rays, natural radiation,
Accepted 13 April 2013
an accident at a nuclear installation in general, and the Fukushima accident in particular. In order to link
Available online 15 May 2013
risk perception with risk communication, media reporting about radiation risks is analysed using
quantitative and qualitative content analyses. The results showed that experts perceive radiological risks
Keywords:
Radiological risk perception
differently from the general public. Experts’ perception of medical X-rays and natural radiation is
Nuclear waste significantly higher than in general population, while for nuclear waste and an accident at a nuclear
X-rays installation, experts have lower risk perception than the general population. In-depth research is con-
Nuclear accident ducted for a group of workers that received an effective dose higher than 0.5 mSv in the year before the
Natural radiation study; for this group we identify predictors of risk perception. The results clearly show that mass media
Media don’t use the same language as technical experts in addressing radiological risks. The study demonstrates
that the discrepancy in risk perception and the communication gap between the experts and the general
population presents a big challenge in understanding each other.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Nevertheless, the expert population in these studies was iden-

Human behaviour is primarily driven by perception and not by
facts (Renn, 2008). The main communication challenge is that the
experts and the public frequently disagree when it comes to risk
assessment. Several studies related to these differences demon-
strated that experts have in general a lower perception of risks than
the general public (Slovic et al., 1980; Slovic, 1996). This has been
highlighted in studies related for instance to nanotechnology haz-
ards (Savadori et al., 2004) or biotechnology (Siegrist et al., 2007).
Moreover, a few existing studies from the radiological field show
large differences in the perception of radiation risks by the experts
and the general population. These studies examined the perception
of nuclear power (Hamalainen, 1991; Sjöberg and Drottz-Sjöberg,
1991; Kanda et al., 2012), nuclear testing (Purvis-Roberts et al.,
2007) or nuclear waste (Sjöberg, 2002) or nuclear waste disposal
by using mental models approach (Skarlatidou et al., 2012).
* Tel.: þ32 14 33 28 51.
E-mail address: tperko@sckcen.be.
URL: http://www.sckcen.be/
tified according to rather weak methodological standards (admin-
istrative support and persons who were not highly knowledgeable
about radiation risk were included in the group of experts) and
rather small population samples were compared, while the differ-
ences between the groups related to perception of non-industry
related radiation risks, for instance medical use of radiation or
natural radiation, was not investigated. In addition, the authors
didn’t make an empirical link between risk perception and risk
communication.
Our empirical study adds on to previous research and highlights
the risk communication challenges arising from the differences in
risk perception identified between the experts and the lay public.
In our study, the group of experts consists of employees of a
nuclear research center; this expert population was identified ac-
cording to the real radiological exposure (participation at regular
radiation protection trainings, educational programs, entering in
controlled areas, obligatory use of dosimeters and regular medical
check-up for possible internal contaminations). Taking into account
the characteristics of the employees (presented in Section 2.2), for
instance 10e20 years of working experiences or high level of ed-
ucation, they can be recognized as topical experts in the field of

0265-931X/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvrad.2013.04.005
 T. Perko / Journal of Environmental Radioactivity 133 (2014) 86e91
ionizing radiation. The number of respondents in the expert pop-
ulation is in our research much higher than in any other radiolog-
ical risk perception studies, comparing experts and lay population.
In addition, radiological risk perception is studied for four different
contexts, instead of a single one, as it is the case for similar studies.
An empirical analysis of media use of radiation units in reporting
about the Fukushima accident gives additional insight into risk
communication performed by the experts and transmitted by mass
media to the general population.
The study is divided in two parts. In the first part of the study we
investigated the perception of radiation risks among employees at a
Belgian nuclear research installation (n ¼ 332) who are profes-
sionally exposed, among which, there were employees that
received a dose higher than 0.5 mSv (effective dose) in the year
before the empirical study (n ¼ 49). The results obtained for these
two categories were compared with the risk perception of the
general population in Belgium (n ¼ 1020). We also compared the
perception of the following risks: an accident in a nuclear instal-
lation (including the Fukushima accident), natural radiation,
medical X-rays and nuclear waste.
In this first part of the study we tested the following hypotheses:
H1. The general population has a significantly higher perception of
all radiological risks than employees of a nuclear research center.
H2. Familiarity and personal experience with low radiation doses
decrease the perception of a radiation risks.
H3. Experts and general population use different mental models (an
explanation of person’s thought process) for the assessment of radio-
logical risks and develop different latent constructs (variables that are
not directly observed but are rather inferred from other variables and
are measured as factor models).
H4. Among the employees at a nuclear research centre, a lower
perception of radiological risks is influenced by the following hypo-
thetical predictors: i) higher personal experiences with ionising radi-
ation, ii) familiarity with ionising radiation, iii) strong feeling of being
protected from risk, iv) higher perceived control by authorities on the
safety of nuclear installations.
In the second part of the study we investigated how are radia-
tion risks communicated through mass media to the general pop-
ulation. The importance of this study lies in that information to the
general public is a key factor in the governance of ionising radiation
risks. Sound communication about ionising radiation with the
general public is rather complex, especially due to the low public
understanding and the perception of radiological risks which dif-
fers from that of the experts. Abstract and unfeeling language, for
example reporting quantitative radiation units, often offends and
confuses people (Covello, 2011), therefore the use of comparisons
of risks is advised in order to develop sound communication (IAEA,
2012). This was highlighted also by the 2011 accident in Japan
(Ropeik, 2011; Cantone et al., 2012; Kanda et al., 2012).
In this second part of the research reported here, we studied
media reporting about radiological risks during the first
commemoration of the Fukushima nuclear accident - one year after
the accident. This time period was selected since journalists not
only represent but also interpret and construct a reality. In doing so
they often make use of a collective memory, for instance the col-
lective memory related to the nuclear accident. For this case, we
investigated how often were the radiological risks presented
quantitatively e by using measurement units in media articles- and
how often were qualitative (descriptive) comparisons used instead.
For this purpose the content of eight Belgian newspapers was
analyzed for the period of first commemoration of the Fukushima
accident (n ¼ 51 articles). The hypotheses tested are the following:
87
H5. Although experts use technical language to communicate about
radiological risks , using quantitative units to present risk, mass media
don’t use these units in their reporting.
H6. Mass media present radiological risks by qualitative compari-
sons with familiar radiological exposures.
In the next section we describe the methods and the data, fol-
lowed by the results and discussions.
2. Method and data
Three data sets and two data collection methods were used to
obtain the results of the study. The first part of the research is based
on a public opinion survey conducted for the general population in
Belgium and a large sample of employees in a nuclear research
centre; the second part of the research is based on the media
content analysis of Belgian press. We describe the three datasets
used in what follows.
2.1. General population
The results for the general population presented in this paper
are based on a large scale public opinion survey in the Belgian
population. The data collection method employed was “Computer
Assisted Personal Interviewing”, consisting of personal interviews of
about 45 min carried out at the home of the respondents in the
period between 25 of May 2011 and 24 June 2011. The field work
was performed by a market research company with professional
interviewers.
The survey (Turcanu et al., 2011) included, among others,
questions related to risk perception and the relevance of the acci-
dent in Fukushima for Belgium. The population sample consists of
1020 respondents and is representative for the Belgian adult pop-
ulation (18þ) with respect to sex, age, region, province, habitat and
social class.
In order to measure risk perception, respondents were asked to
“evaluate the risks for an ordinary citizen of Belgium” for the
following radiation risks: nuclear waste, an accident at a nuclear
installation, natural radiation (e.g. cosmic radiation or radon) a
medical x-rays. Answering categories ranged from “very low” (1) to
“very high” (5). In a later section of the survey, the respondents
were asked to state their level of agreement with the following
statements: “What happened in Japan (the Fukushima accident)
makes me more worried about the dangers from Belgian nuclear
installations”, “There is sufficient control by authorities on the
safety in nuclear installations in Belgium” and “I feel well protected
against risks from nuclear installations”. The answering categories
for these items ranged from “strongly disagree” (1) to “strongly
agree” (5).
2.2. Experts e employees professionally exposed to radiological
risks at a nuclear research installation
The data collected for the expert population are based on an
opinion survey conducted in the Belgian Nuclear Research Centre
SCKCEN. The selected respondents were all employees who enter
controlled areas (research reactors, plutonium laboratory, irradia-
tion facility .) and are registered as such for monitoring. They are
all wear dose-metres, measuring possible radiation doses received,
are regularly checked for possible internal contamination and have
all received a special radiation-protection training. These people
are all regularly professionally exposed to radiological risks. The
other employees of the research center were not invited to
participate in the survey.
88 T. Perko / Journal of Environmental Radioactivity 133 (2014) 86e91

The data collection method employed was “Computer Assisted
Interviewing”, which consisted of internet interviews of about
20 min carried out at the working place of the respondent, con-
ducted in the period between 23 May 2012 and 11 June 2012. The
respondents were encouraged to participate by a personal letter
ensuring anonymity. The data base consists of 332 experts, among
which, employees that received a dose higher than 0.5 mSv in the
year before the empirical study (n ¼ 49). The employees varied in
age from 19 to 63 years, the age category 36e45 had the greatest
number of employees (33%). The education level of the respondents
reflects the nature of the research centre e 70% of them had higher
non-university or university degree. The employees in the expert
group had on average 15 years of working experience in nuclear
applications, including education (most of them having between 10
and 20 years of experience). For the most part they were occupa-
tionally exposed to radiation several times per month (31%) or
almost every day (17%). They performed a diverse tasks in the
controlled area: from laboratory work (16%), manipulation of
sources (13%), maintenance of equipment (13%), to inspection and
supervision (14%). Thus, taking into account these characteristics of
the employees, they can be recognized as topical experts in the field
of ionizing radiation.
The questions measuring risk perception and the perception of
the Fukushima accident, in the survey carried out in the nuclear
research centre were the same as the questions asked to the general
public. As regards, background variables, additional items were
introduced for the expert population; these measured personal
experiences, exposure to and familiarity with radiological risks.
2.3. Media content analysis data
The data presented in this research are part of a broader media
analysis project related to media reporting about the Fukushima
nuclear accident (Perko et al., 2011; Prezelj et al., 2012). A strict
scientific methodology, content analysis, was used to analyse media
reporting (Neuendorff, 2002). Here we report results concerning
radiation measurement units and descriptive comparisons of
radiological risks. Several known books and articles exist to instruct
in the methods of content analysis (Krippendorf, 1980; Neuendorff,
2002; Schiffrin et al., 2003).
Our sample contained 51 media articles that were either directly
or indirectly related to the first commemoration of the Fukushima
nuclear accident. The articles were collected by the following
keywords: “Fukushima” and “nuclear*” from the media data base
“MEDIARGUS”. All daily newspapers in Dutch language, published
in Belgium were included in the analysis: De Tijd, De Morgen, De
Standaard, Gazet van Antwerpen, Belang van Limburg, Het Laatste
Nieuws, Het Nieuwsblad and the free daily newspaper Metro. The
news articles were published in the time period between the 3rd of
March, 2012 and the 18th of March, 2012, when details of the
Fukushima accident were visible in the mass media.
In order to increase the validity of our results, the databases
independently produced by two coders (containing coded articles
for each newspaper) were checked and compared by a master
coder. In case of a difference of in the results by the two coders, a
consensual assessment was made among the coders and a final
master database was made. Also the inter-coder reliability coeffi-
cient was calculated based on the Krippendorff’s alpha (2004). This
is a reliability coefficient developed to measure the agreement
between observers, coders, judges, rates, or measuring in-
struments. For the two variables reported here (radiation mea-
surement units reported in the article and descriptive comparisons
used to compare radiological risks) the inter-coder reliability co-
efficient was 1 (perfect agreement).
3. Results
3.1. Radiological risk perception: general population vs. experts
First we compared the means of the risk perception items for
the two population groups by the independent group t-test. A
statistically significant difference between the radiological risk
perceptions of the two groups was confirmed (see Table 1). T-test
revealed statistically reliable differences between the risk percep-
tions for the following items: nuclear waste, an accident in a nu-
clear installation, natural radiation, medical X-rays and the
Fukushima nuclear accident. A significance was for all risk per-
ceptions bellow a level of 0.05. In addition, we identified, With
Levene’s test for homogeneity, that variances in the groups of the
general population and the experts are different for all investigated
radiological risk perceptions (sig ¼ 0.001), with the exception of
natural radiation (sig ¼ 0.53).
The general population had a higher risk perception for nuclear
waste and natural radiation than the experts. Moreover, the general
population perceived a nuclear accident in a nuclear installation as
a higher than did the experts from the nuclear research center; they
were also more concerned about Belgian nuclear installations after
the Fukushima accident. However, experts had higher perception of
medical X-rays than the general population. The hypothesis that
general population has a significantly higher perception of all
radiological risks than the employees of a nuclear research center
(H1), can only be partially accepted depending on the context of
exposure.
In addition, we analyzed the hypothetical differences in radio-
logical risk perceptions between the experts in the nuclear research
centre that received a dose higher than 0.5 mSv due to professional
exposure in the year before the empirical study and those that
received doses lower than 0.5 mSv.

Table 1
Differences in radiological risk perception between in the general population and the experts.

Risk perceptions Population group Na Mean Std. Std. error t (df ¼ 1238) Sig. (2-tailed)
deviation mean a ¼ 0.05
Nuclear waste General population 921 3.11 1.13 0.038
Experts 319 1.74 0.93 0.05 19.32 0.00
Accident in a nuclear General population 921 2.95 1.18 0.039
installation Experts 319 2.00 1.12 0.06 12.45 0.00
Natural radiation General population 921 2.54 1.02 0.03 4.06
Experts 319 2.27 1.03 0.06 0.00
Medical X-rays General population 921 2.60 1.04 0.03
Experts 319 2.83 0.97 0.05 3.65 0.00
The Fukushima General population 921 3.30 1.21 0.04 13.50 0.00
accident Experts 319 2.29 0.99 0.06
a
All respondents except those with “don’t know” answer to one or more items.
 T. Perko / Journal of Environmental Radioactivity 133 (2014) 86e91 89

A statistically significant difference in the radiological risk per- Table 3

ceptions between the two groups was confirmed for the following Latent constructs related to radiation risk perceptions: General population vs.
experts.
items: nuclear waste, accident in a nuclear installation and the
perception of the Fukushima nuclear accident. For these risk per- Risk perception Factor loadings and reliability of the scale
ceptions, the significance value of the t-test was p < 0.05. General population Experts
Experts of the nuclear research centre that received a dose of one factor
Factor 1 Factor 2
more than 0.5 mSv in the year before this study due to their pro-
Nuclear waste 0.81 0.73
fessional exposure, perceived the risk of nuclear waste significantly
Accident in a nuclear 0.71 0.79
lower than their work colleagues. Similarly to this, the perception installation
of an accident in a nuclear installation was also perceived much Natural radiation 0.52 0.71
lower among the employees that received a dose of more than Medical X-rays 0.44 0.67
0.5 mSv. Moreover, these employees were less concerned about the Cronbach’s Alpha 0.73 0.80 0.56

dangers from Belgian nuclear installations due to the events in

Japan. Attention-grabbing is the level of risk perception of natural
radiation and medical use of radiation: the experts receiving a dose
higher than 0.5 mSv had significantly higher perception of these
risks (see Table 2).
With Levene’s test for homogeneity we tested if the variances in
the two groups of employees were the same. The test revealed that
group variances are different for the risk perception of radioactive
waste (0.04) risk perception of an accident in a nuclear installation
(sig ¼ 0.03), and the perception of the Fukushima nuclear accident
(sig ¼ 0.04). Yet, the variances were similar for the risk perception
of natural radiation (sig ¼ 0.41) and the medical X-rays (sig ¼ 0.12).
According to these results we can confirm that familiarity and
personal experience with low radiation doses decrease the
perception of some radiation risks (H2).
To identify the differences in the latent constructs related to
radiological risk perceptions between the general population and
the experts (H3) we performed factor analysis with Principal Axis
Factoring, and Oblimin rotation with Kaiser normalisation method
(Table 3).
Some interesting differences in the mental models were thus
recognized. The respondents from the general population have the
same latent construct in minds when they asses different radio-
logical risks: nuclear waste, nuclear accident, medical use of radi-
ation or natural radiation (Cronbach’s a ¼ 0.73). Opposite to this,
the experts of the nuclear research centre have at least two latent
constructs in mind: the first construct links perception of nuclear
waste and perception of nuclear accidents (Cronbach’s a ¼ 0.80);
the second construct links the medical use of ionising radiation and
natural radiation. However, the reliability of the second construct is
rather week (Cronbach’s a ¼ 0.56), thus we can consider two
mental models for an expert assessment of the risks from X-rays
and natural radiation. The perception of the Fukushima accident
was excluded from the factor analysis due to low communalities
(<0.3) in both populations. Hence, we confirmed that the experts
and the general population use different mental models when
assessing radiological risks and develop different latent constructs
(H3).
Next, we analyzed the first factor highlighted above (Factor 1:
nuclear waste and an accident in a nuclear installation) in order to
identify potential explanatory variables for the perception of the
related radiation risks. A linear regression model was used to es-
timate the relationship between variables. We tested if the
perception of nuclear waste and the perception of an accident are
influenced by the following hypothetical predictors: i) number of
years of experience, ii) the average frequency of the professional
exposure to radiation, iii) the feeling of protection against risks
from nuclear installations, iv) the level of perceived control by
authorities on the safety in nuclear installations assessed by the
employee.
With a linear regression model we can confirm that the larger is
the professional experience of the employee, the lower is his/her
radiological risk perception (b ¼ 0.13, sig. ¼ 0.03). The feeling of
being protected against risks from nuclear installations was
revealed as the most important influencing factor for the radio-
logical risk perception among people professionally exposed to
radiological risks. Professionals that do not feel well protected had
higher radiological risk perceptions (b ¼ 0.24, sig. ¼ 0.00).
Furthermore, a higher perceived control by the authorities on the
safety in nuclear installations leads to lower perception of radio-
logical risks (nuclear waste and accident in a nuclear installation)
(b ¼ 0.14, sig. ¼ 0.01). Interestingly, the predictor “how often is an
employee in average professionally exposed to radiation” was
revealed as not influencing the radiological risk perception. The
results confirmed the hypothesis four: a lower perception of
radiological risks among employees at a nuclear installation is
influenced by the following hypothetical predictors: i) higher per-
sonal experiences in a nuclear application, ii) strong feeling of being

Table 2
Differences between radiological risk perceptions of the experts that received a dose higher than 0.5 mSv and experts that received lower doses.

Risk perceptions Expert group N* Mean Std. Std. error t (df ¼ 317) Sig. (2-tailed)
(received doses deviation mean a ¼ 0.05
due to professional
exposure)

Nuclear waste <0.5 mSv 270 1.79 0.96 0.06

>0.5 mSv 49 1.47 0.71 0.10 2.73 0.01
Accident in a nuclear installation <0.5 mSv 270 2.06 1.17 0.07
>0.5 mSv 49 1.76 0.78 0.11 2.28 0.03
Natural radiation <0.5 mSv 270 2.24 1.02 0.06 1.17 Not significant
>0.5 mSv 49 2.43 1.10 0.16
Medical X-rays <0.5 mSv 270 2.81 0.99 0.06 Not significant
>0.5 mSv 49 2.94 0.92 0.13 0.82
The Fukushima accident <0.5 mSv 270 2.33 1.01 0.06 1.94 0.05
>0.5 mSv 49 2.06 0.85 0.12

*All respondents except those with “don’t know” answer to one or more items.
90 T. Perko / Journal of Environmental Radioactivity 133 (2014) 86e91
protected from risk, and iii) higher perceived control by authorities
on the safety in nuclear installations. Contradictory to our expec-
tations, higher familiarity with the radiation, expressed with a
regular professional exposure to radiological risks did not revealed
as statistically influential for the perception of radiological risks
(nuclear waste and accident at a nuclear installation).
3.2. Media reporting about radiological risks: first commemoration
of the Fukushima accident
In a second part of the study we investigated how were the
radiation risks communicated from the experts to the general
public in articles reporting about Fukushima nuclear accident. With
the analysis of measurement units reported in the media articles
related to first commemoration of the Fukushima we assessed how
the media report radiation risks using quantitative information.
Radiation related quantities can be expressed using different
measurements units, that might be used to explain the risk of
radioactivity. For the analysis reported here we have used the
following list of 18 different possible measurement units corre-
sponding to measurement of activities, activity concentration,
ground depositions, dose rates or estimates of the dose received:
mSv (milli sievert), mSv/h (millisievert per hour), mSv/h (micro-
sievert per hour), (nanosievert per hour), Sv (Sievert), Sv/h (Sievert
per hour), Bq/kg (Bequerel per kilogram), Bq/g (Bequerel per gram),
(Bequerel per litre), kBq/kg (kilo Bequerel per kilogram), MBq/kg
(mega Bequerel per kilogram), Bq/m2 (Bequerel per square metre),
Bq/cm2 (Bequerel per square centimetre), kBq/cm2 (kilo Bequerel
per square centimetre), MBq/m2 (mega Bequerel per square metre),
MBq/km2 (mega Bequerel per square kilometre), TBq/km2 (terra
Bequerel per square kilometre) and another measurement unit
related to radiation (e.g. air concentration in Bq/m3).
These units were only coded if they were explicitly mentioned
in the article. Since for the non-expert population, the technical
language and jargon are major barriers to understand risks, we
investigated the risk comparisons used in the articles.
It is interesting that although the media reported about the
nuclear accident, radiological risks and danger, only about 12% of all
articles contained radiation units, as shown in Fig. 1. The most often
used measurement unit was the Becquerel (38% among articles
reporting measurement units). It is clear that, although experts are
used to communicate about radiological risks in technical language,
often using quantitative units to present risks, mass media don’t
use these units in their reporting (H5).
Since risk communication research shows that effective formats
are needed to communicate quantitative information about radia-
tion risks to the general population (Schwartz et al., 1997; IAEA,
2012), we analysed which descriptive risk comparisons had the
Fig. 1. Presence of measurement units related to radiation in media articles reporting
about the first commemoration of the Fukushima accident.
journalists used in the articles related to the first commemoration
of the Fukushima accident (one year after the accident). We found
the following radiological risk comparisons in the articles analysed:
 medical use of radiation, X-rays,
 natural radiation background,
 radiation released by the nuclear bomb at Hiroshima,
 legal norms and legal limits.
Percent of the articles it is confirmed that the mass media also
present radiological risks also by qualitative comparisons with
familiar radiological exposures. However, since these comparisons
were included only in eight percent of the articles, the hypothesis
H6 was partly accepted.
4. Discussion and conclusions
Previous research showed that experts and the general popu-
lation differ in their perception of risks, mainly associated with
complex technologies; for instance nuclear power plants (Sjöberg
and Drottz-Sjöberg, 1991), nanotechnology (Siegrist et al., 2007),
biotechnology (Savadori et al., 2004), genetically modified food
(Cook et al., 2004). However, the differences in expert’s and lay
person’s perception of radiological risks were not extensively
explored and empirically confirmed (for critical review of previous
research see Sjöberg, 2002).
In our research we explored perception of radiation risks in
various contexts and compared the risk perceptions of different
population groups: the general population, the employees in a
nuclear research centre and a group of experts that received an
effective dose higher than 0.5 mSv in the year before the study. For
the latter group we also identified predictors of risk perception.
Contrary to our expectations, results show that the general
population has a higher risk perception than the experts only for
some radiological risks. The risks from the medical use of ionising
radiation are perceived much higher by the employees of the nu-
clear research centre than by the general population. Opposite to
this, the general population had a higher risk perception for nuclear
waste, an accident in a nuclear installation and natural radiation,
and they were more concerned about Belgian nuclear installations
after the Fukushima accident. These results can be explained by the
psychometric approach (Fischhoff et al., 1978) suggesting that fa-
miliarity with hazard, knowledge, personal control and voluntari-
ness decrease risk perception (Slovic, 2000; Renn, 2008).
Experts in the research centre are more familiar with the
concept of radioactivity related to nuclear waste than general
public: they produce radioactive waste, study it and investigate
long term solutions for disposal. We assume that the risk assess-
ments of these experts’ are based more strongly on actual knowl-
edge than the assessments made by the general population. In
addition, they have also control over the risk activities and they
participate in decision-making related to these activities. However,
the experts’ higher perception of risks arising from the medical use
of radiation as compared to the perception of the general public
may be also explained by psychometric characteristic e institu-
tional control over the medical use of radioactivity; the experts
from nuclear research centre may be more aware of “what can go
wrong”, but they cannot control or influence it.
We found that experts and lay people use different mental
shortcuts when assessing radiological risks. A number of Inter-
esting differences in the respective mental models were recog-
nized. The respondents from the general population have the same
latent construct in minds when they assess different radiological
risks: nuclear waste, nuclear accident, medical use of radiation or
natural radiation. Opposite to this, the experts of the nuclear
 T. Perko / Journal of Environmental Radioactivity 133 (2014) 86e91
research centre have at least two latent constructs in mind: the first
construct links perception of nuclear waste and perception of nu-
clear accidents, the second construct feebly links a medical use of
ionising radiation and natural radiation. These results provide a
quantitative empirical evidence to support the mental models
developed by Morgan et al. (2002) and applied to radiological risks

by Zeleznik (2010) and Skarlatidou et al. (2012). In these studies the
experts’ mental model of radioactivity was compared with mental
model of lay people by qualitative mythology e individual open
interviews.
Nevertheless, we confirmed that within the group of experts,
employees have different radiological risk perceptions and,
although they are experts, they have different feelings related to
risks. We identified at least two groups in the research center: the
group of workers that received an effective dose higher than
0.5 mSv in the year before the study, having in general lower
perception of waste and accidents and a higher risk perception of X-
rays and natural radiation, and the group of their colleagues that
received less than 0.5 mSv effective dose due to occupational
exposure.
The in-depth analysis of the group of experts allowed to identify
predictors of radiological risk perception and confirmed again the
influence of psychometric risk characteristics:
 the larger is the professional experience of the employee
(voluntariness and knowledge), the lower is his/her radiolog-
ical risk perception,
 the feeling of being protected against risks from nuclear in-
stallations (personal control) leads to lower perception of
radiological risks,
 a higher perceived control by the authorities on the safety in
nuclear installations (institutional control) leads to lower
perception of radiological risks.
Remarkable, within the experts, higher familiarity with the ra-
diation, measured by the number of entries in the controlled area,
did not revealed as influential for the perception of radiological
risks.
The difference identified in the risk perception of the different
population groups is not the only challenge related to radiological
risk communication. Another challenge, recognized in our study, is
also the use of a different language to describe radiological risks. It
is clear that, although experts are used to communicate radiological
risks in a technical language, often using quantitative units to
present risks, the majority of media articles analysed don’t use
these units in their reporting. It is confirmed that mass media also
tends to present radiological risks by qualitative comparisons with
familiar exposures to ionizing radiation.
The study demonstrates that the discrepancy in risk perception
and the communication gap between the experts and the general
population presents a big challenge in understanding each other.
Acknowledgements
The author would like to thank the respondents from the
Belgian Nuclear Research Centre SCKCEN for theirs insights into
risk perception and to Master student Bart Adams for data collec-
91
tion in expert population. The field work for data collection in
general population was funded through the R&D collaboration
agreement between SCKCEN and GDF-SUEZ.
References
Cantone, M., Perko, T., Turcanu, C., Prezelj, I., Tavola, F., Sturloni, G., Geenen, D.,
Mamane, N., Van Rooy, L., Marin, L., Speh, K., Bizjak, T., 2012. Content analysis of
the media reporting on the Fukushima nuclear accident in three European
countries. In: 13th International Congress of the International Radiation Pro-
tection Association.
Cook, G., Pieri, E., Robbins, T.P., 2004. The scientists think and the public feels:
expert perceptions of the discourse of GM food. Discourse & Society 15,
433e449.
Covello, V.T., 2011. Risk communication, radiation, and radiological Emergencies.
Health Physics 101, 511e530.
Fischhoff, B., Slovic, P., Lichtenstein, S., Read, S., Combs, B., 1978. How safe is safe
enough e psychometric study of attitudes towards technological risks and
benefits. Policy Sciences 9, 127e152.
Hamalainen, R.P., 1991. Factors or values e how do parliamentarians and experts see
nuclear power. Energy Policy 19, 464e472.
IAEA, 2012. Communication with the Public in a Nuclear or Radiological Emergency.
EPR-Public Communications, IAEA, Vienna.
Kanda, R., Tsuji, S., Yonehara, H., 2012. Perceived risk of nuclear power and other
risks during the last 25 Years in Japan. Health Physics 102, 384e390.
Krippendorf, K., 1980. Content Analysis: an Introduction. Sage, Thousand Oaks, CA.
Krippendorf, K., 2004. Reliability in content analysis. Human Communication
Research 30, 411e433.
Morgan, M.G., Fischhoff, B., Bostrom, A., Atman, C.J., 2002. Risk Communication: a
Mental Models Approach. Cambridge University Press, New York.
Neuendorff, K., 2002. The Content Analysis Guidebook. Sage Publication, Thousand
Oaks, California.
Perko, T., Turcanu, C., Geenen, D., Mamane, N., Van Rooy, L., 2011. Media Content
Analysis of the Fukushima Accident in Two Belgian Newspapers. Belgian Nu-
clear Research Centre SCKCEN, Mol; Belgium, BLG-1084.
Prezelj, I., Perko, T., Turcanu, C., Cantone, M.C., Geenen, D., 2012. Nuclear emergency
communication: one emergency. Many Voices and the Problem of Coordination
21st International Conference Nuclear Energy for New Europe 1, 1e14.
Purvis-Roberts, K.L., Werner, C.A., Frank, I., 2007. Perceived risks from radiation and
nuclear testing near Semipalatinsk, Kazakhstan: a comparison between physi-
cians, scientists, and the public. Risk Analysis 27, 291e302.
Renn, O., 2008. Risk Governance; Coping with Uncertainty in a Complex World.
Earthscan, London.
Ropeik, D., 2011. Poor Risk Communication in Japan Makes the Fear Much Worse.
Psychology Today.
Savadori, L., Savio, S., Nicotra, E., Rumiati, R., Finucane, M.P.S., 2004. Expert and
public perception of risk from biotechnology. Risk Analysis 24.
Schiffrin, D., Tannen, D., Hamilton, E.H., 2003. The Handbook of Discourse Analysis.
Wiley-Blackwell.
Schwartz, L.M., Woloshin, S., Black, W.C., Welch, H.G., 1997. The role of numeracy in
understanding the benefit of screening mammography. Annals of Internal
Medicine 127, 966e972.
Siegrist, M., Keller, C., Kastenholz, H., Frey, S., Wiek, A., 2007. Laypeople’s and ex-
perts’ perception of nanotechnology hazards. Risk Analysis 27, 59e71.
Sjöberg, L., 2002. Communication du risque entre les experts et le public: intentions
et perceptions. Questions de Communication 2, 19e35.
Sjöberg, L., Drottz-Sjöberg, B.M., 1991. Knowledge and risk perception among nu-
clear power plant employees. Risk Analysis 11, 607e618.
Skarlatidou, A., Cheng, T., Haklay, M., 2012. What do lay people want to know about
the disposal of nuclear waste? A mental model approach to the design and
development of an online risk communication. Risk Analysis 32, 1496e1511.
Slovic, P., 1996. Perception of risk from radiation. Radiation Protection Dosimetry
68, 165e180.
Slovic, P., 2000. The Perception of Risk. Earthscan Publications, London.
Slovic, P., Fischoff, B., Lichtenstein, S., 1980. Facts and fears e understanding
perceived risk. Health Physics 39, 1005e1006.
Turcanu, C., Perko, T., Schröder, J., 2011. The SCKCEN Barometer 2011 e Perception
and Attitudes towards Nuclear Technologies in the Belgian Population. Belgian
Nuclear Research Centre SCKCEN, Mol; Belgium.

Zeleznik, N., 2010. Mental models of radioactivity and attitudes towards radioactive
waste. In: 8th International Conference on Nuclear Option in Countries with
Small and Medium Electricity Grids. http://ocs.nuclear-option.org/index.php/
CNS/2010/paper/view/16.