Customize Risk Communication Since the General Public Is Heterogeneous. Lars-Erik Warg, Kjell Andersson, Gunnar Ahlborg Jr., and Inger Fagerlund, Department of Occupational and Environmental Medicine Örebro Medical Centre Hospital, S-701 85 ÖREBRO, SWEDEN
INTRODUCTION
As the recognition of risks in our society increases, so do the writing and talking about those risks. A world wide example of this is all the brochures, information videos and seminars aiming to increase peoples knowledge about AIDS and proper behaviour actions in order to protect one self. If we add these information to the continuing flow of risk communication in every day life (about drugs, nutrition, environmental threats etc.), we can easily get an idea of the importance of risk communication in modem society. In this paper, focus is on the process of risk communication and some means by which it can be improved.
In today's society, there are often disagreements between governments/experts and groups within community concerning risks due to certain events or activities. Maybe most often this disagreements are about "new technology" or more general environmental risks. The pattern in this disagreements seems to be similar in most cases; experts think that the community worry about "wrong" risks i.e. risks considered tiny by the experts and/or the community seem frustrated by the governments seeming disinterest in their concerns. Consequences due to these disagreements range from oral/non-violent (normally) to physical/violent (seldom).
Bearing in mind the potential power of disagreements of risks, politically, economically (just think of the mad-cow disease) and socially and the potential value of improvements, it is easy to understand a growing literature about how to communicate environmental health risks to the public. In fact, it has been claimed that how we communicate about risks is of importance for the democratic foundations of our society. If this is true and we remind ourselves of the discrepancy between expert/technical and public assessments of risk, it is hard to see more important areas to investigate.
Although there are many questionnaires developed to measure risk perception, we are not aware of any focusing on what Peter Sandman labels "outrage factors". These type of factors represent a sound knowledge in cognitive psychology which have proved to be of significant importance in how people perceive risks.
The aim of the study presented in this paper was twofold: 1) to develop a questionnaire focusing on outrage factors, which governments may use to improve their dialogues with the community concerning environmental risks, and 2) testing the impact on risk perception of a questionnaire focusing on general environmental factors.
MATERIAL AND METHODS
Sample and procedure
A total sample of 2.500, consisted of a simple randomised
sample of the population in Örebro county between 25 - 65 years
available for identification in registers (156.505 persons), was
produced for a series of distributed questionnaire survey study
in risk communication and risk perception. This sample was then
divided in seven subgroups, all with the important
characteristics of the total sample and thus the population in
Örebro County. In the present study, a total of 1.600
persons/four subgroups were included, comprising four
experimental conditions, two consisted of 300 persons each
(conditions 1 and 2) and two of 500 persons each (conditions 3
and 4).
Data were collected by two questionnaires, "Public's Outrage Questionnaire" (henceforth POQ) and "My Environment Questionnaire" (henceforth MEQ). Each person in the conditions 2 - 5 received POQ by mail and were asked to complete the questionnaire and return it directly to the researchers in a postage-paid envelope. With the same requirements, persons in conditions 6 and 7 received MEQ in addition to POQ. After three weeks, a reminder in the form of a post card was sent to those not yet been answering the questionnaire. The response rate in this study was 44.8% (1.116).
Measures and design
POQ was designed to assess the impact of outrage factors on
risk perception. It requested the subjects to read a fictive
news-paper article and then answer 14 questions with fixed
response alternatives, about perceived risk related to the news
article. There were also four manipulation check items and three
questions requesting information about the subjects age, gender
and present occupation. In a covering letter, subjects were
instructed to start by reading an article comprising the first
side(s) of POQ. This article with a news paper layout, told the
fictive story about RODMAN AP, a plastic industry and their plans
to build a plant in Karlslund, a beautiful recreation area in
Örebro. As background information, to the subjects in condition
1 and 3 it was suggested that people would not get much control
over and insight in the production process, that RODMAN AP would
benefit much more of the plant than the local citizens, that the
management of the plant induced very little credibility and that
the local citizens were very critical to the company's responses
to worries expelled by the community (in conditions 2 and 4, the
opposite states were at hand). This was ill since they were using
styrene (C8H8) in the turn out process.
Styrene has some well known negative health effects on humans and
can have severe effects on the environment if not handled
correctly. The risk questions were answered on 7-point scales
ranging from 1 (e.g. Would absolutely not affect plans to
buy house) to 7 (e.g. Would absolutely affect plans to buy
house).
The second questionnaire, MEQ, required subjects to answer 30 questions about risks and 10 questions concerning demographic data, on 7-point scales ranging from 1 (e.g. Risk of being injured by chemical additives in food-stuffs is non-existent) to 7 (e.g. Risk of being injured by chemical additives in food-stuffs is very great). Four were about risk aversion, twelve concerned knowledge and interest among local authorities in radon, chemical additives in foodstuffs and industrial contamination in addition to personal influence. 14 questions were about risk with a variety of environmental factors (many of them found in studies by Sjöberg and Drottz-Sjöberg). Subjects in conditions 3 and 4 received MEQ in addition to POQ. Thus, the design in the present study was a 2 (extent of outrage) by 2 (administration of MEQ) design.
RESULTS
The "architecture" in the two questionnaires was tested by factor analyses. As expected, three factors were extracted in the analysis of POQ: Risk perception, impact on nature and health and outrage factors. All items loaded on their proposed dimensions and the three factors accounted for 67% of the variance. Two factor analyses were conducted on MEQ. Part one focusing on risk aversion and perception on local authorities extracted four factors: Local authorities, public's influence, health effects and risk aversion. These four factors accounted for 61.3% of the variance. Part two, environmental risk factors, extracted three factors: external environmental factors, life style/domestic threats and violence/accidents. Together these three factors accounted for 56.8% of the variance.
In addition to these analyses, a discriminant analysis conducted on POQ showed that the questionnaire discriminated correctly between high and low outrage groups in 78% and 80% of the cases for conditions 1 and 2 respectively 3 and 4.
Table 1 shows the results of nine t-tests for independent samples of groups. No values for t or p are given. However, the criteria for significance was at least p<.05. In the table, these are the shaded areas. The results may be summarised as follows: Women perceived more risks than men; married and singles differed in certain issues but perceived risks much the same; people without children perceived more risks with lifestyle/domestic threats and of being assaulted; those living in apartments scored higher on some environmental factors as well as on life style/domestic threats compared to those living in houses; townspeople perceived more risks in industrial contamination and air pollution compared to those living in the country; people with elementary schools were more risk aversive and scored higher on some environmental factors while those with collage/university showed a little more trust in local authorities knowledge and believed themselves to exercise more influence over environmental decisions; people with a history of illness/accidents perceived more risks with greenhouse effect, ozone layer, radon indoors and of being assaulted; older people showed more trust in local authorities, were more risk aversive, and perceived more risks on a number of environmental factors compared to younger people.
Table 2 shows significant effects of high outrage (at least
p<.05, shaded areas) for almost every question in the
conditions receiving MEQ in addition to POQ. However, giving POQ
only produced no significances on risk perception, just on the
manipulation check items.
DISCUSSION
The results in this study is congruent with positions that
the general public is heterogeneous and that this must have an
impact on risk communication. MEQ clearly shows that demographic
data affects risk perception. Further analyses of multivariate
type, will give answers to questions whether the population can
be classified in clusters with typical variable patterns. The
study indicate the possibility to develop a questionnaire (POQ)
that considers outrage factors and thus will improve the
understanding of peoples risk perception in environmental issues.
Interesting is the finding that just adding MEQ to POQ, produced
a different risk perception. Maybe it is so that just
"structural thinking" about environmental factors is
enough to change risk perception. If so, it will focus on the
need to develop instruments to describe the population and
customize the risk communication between experts and communities.
Even if there have been attempts in this direction, it is our
belief that this can be done more often and in a more explicit
way.