Modelling environmental behaviour 69

4 Modelling environmental
behaviour
Socio-psychological simulation
Hans-Joachim Mosler

Introduction
In economically highly developed countries, many of the conditions that would
allow people to behave in environmentally responsible ways are already in place.
We have a lot of knowledge; for years now, surveys have shown that people give
top priority to the need to act on environmental issues. The necessary technical
and economical resources are also available. But there is little sign of a real about-
face except in limited areas. We believe that the much-cited discrepancy between
cognition and behaviour, between lip service and a person’s own contribution to
conserving the environment, can be better understood if we also take people’s
perceptions of the social surround into consideration. To view human beings as
Homo economicus (Diekmann 1996; Harsanyi 1977) falls short, because the economic
man approach does not explicitly take into account inner psychological factors,
such as motives, values, norms and attitudes, nor does it consider social influences
on a person’s behaviour. The present research conceives a model of behaviour
that in addition to economic factors includes personal and social factors. Environ-
mental consciousness is determined to a significant degree by social systems, or
that is to say, by people’s corresponding social representations. An individual’s
personal contribution seems insignificant in the face of massive destruction of the
environment caused by many. This perception – that there is nothing we can do
personally, that each one of us is powerless – as well as a reluctance to be the
‘sucker’, are important causal factors in behaviours that overuse environmental
resources. It does not seem rational to exercise personal restraint (for example, by
not driving), because not only will we suffer from the harm caused by the general
public’s overuse (the consequences of air pollution), but from a reduction in our
own direct return as well (time saved, comfort). However, as this state of affairs
applies equally to all individuals in a society overusing environmental resources,
people mutually trap each other in patterns of actions that harm the environment.
It is for this reason that we are particularly interested in examining the psychological
conditions that would form the basis of a collective reorientation towards environ-
mentally sustainable behaviour. Starting out from new, environmentally friendly
behaviours of some ‘pioneer’ individuals, we wish to discover the social psycholog-
ical conditions that would ensure that the number of persons joining ranks with
70 Hans-Joachim Mosler
such pioneers would continue to automatically increase and result in a true, large-
scale ‘turn-around’ of previous, environmentally harmful patterns of behaviour.
The focal question of our research can thus be framed as follows: what are the
conditions that foster widespread, effective inner dynamics that change collective
environmental overuse (in thinking and action) to collective, environmentally
responsible thinking and acting? Findings generated by this research will lend
scientific support to the planning of environmental protection campaigns.

Method

Procedure
If we start from the assumption that environmental problems originate in the
overusing behaviours of very many individuals, we need to consider how new
solutions might be tested in a problem area of this magnitude. The instrument of
the questionnaire, based upon imaginary situations or conditions, seems ill-suited
(‘How expensive would gasoline have to be for you to change to public
transportation?’). Massive field experiments that translate the issue directly in real
social systems can also be eliminated; given our present state of scientific knowledge,
such experiments would be both financially and ethically irresponsible. Laboratory
experiments, which would not require intervention in existing social systems, cannot
be carried out with large groups of persons (1000 and more). Computer simulation
provides a possible solution. Simulation aims to reconstruct the relevant cause-
and-effect relationships in a problem area in the form of a model. With the aid of
empirical data, the relationships can then be validated. In this way, we can test
‘experimentally’ the most various and unconventional ideas of ways to spread
environmentally responsible thinking without incurring the risk of intervention in
real social systems.
It is important to remember, however, that simulation results are not empirical
results found in reality. Simulation results can generate fruitful hypotheses that
must be tested in reality. The great advantage is that the hypotheses are derived
consistently using a clearly defined procedure. Another important point to consider
is that findings gained through the simulation method allow only comparative
conclusions, for example that under the conditions studied, Measure A produced
better results than Measure B. The simulation method, furthermore, does not
allow statements to be made about the means to be used to implement certain
measures. For this, real experiments with real people are required. Despite all of
this, computer simulation is the only method available for investigating, by means
of many systematically varied experiments, how environmentally responsible
changes can be instituted in large populations.
Our simulation procedure consisted of the following steps.
Following a preliminary selection of the most important, empirically well-
founded theoretical approaches within the field of social influence, the theories’
most relevant and significant core statements were formulated, according to content
criteria.
 Modelling environmental behaviour 71
In the ensuing modelling, the core statements of the theories were described
with the existing variables and set in relation to one another according to certain
systems-theoretical rules (see ‘Modelling of the theoretical concepts’). For lack of
space, we here dispense with discussion of programming/technical implement-
ations of the theories (but see Mosler 2000 for details).
The design of the simulation model was validated through the aid of experts’
evaluations and replication of findings from the fields of environmental and social
psychology. Again, due to space considerations, the validation process is not
reported here (but see Mosler 2000 for details).
Experiments with various strategies for the spread of behaviours were then
conducted. The following will report on the most important and meaningful
experiments.
From the findings of the simulations, conclusions were drawn pertinent to both
environmental practice and basic research. The most well known forms of inter-
vention stemming from environmental research receive a new interpretation; well-
founded recommendations for the field of intervention are derived.
The reader may wish to note that the following deals intensively with psycho-
logical concepts (to psychologists, however, the descriptions will appear somewhat
superficial). The author sees the present chapter as a contribution that takes a
perspective that differs greatly from the economic viewpoint, although some
relevant connections are pointed out. This becomes clear in the next section. The
reader interested in the exact derivation of the findings should read all sections,
while others may wish to read only the sections on the simulation experiments and
results. The most important findings are reported and discussed in the concluding
section.

The framework model

A great many simulations work from the very start at the level of collective variables
(macroanalytic simulation, such as of, for example, the influence of prices on total
consumption by the public). However, if we aim to tap into the dynamics of the
spread, or dissemination (or non-spread, respectively), of environmentally
responsible thinking and behaving, we must first of all begin at the level of
individuals. We then allow them to interact with each other within the framework
of the simulation (microanalytic aggregative simulation). The difficult task is to
model the relevant processes of the individual in this area of conflict on the
computer. We assume that humans are in principle beings that are free to make
their own decisions, and we hold that a computer program cannot represent the
richness of human individuality, as it is primitive in comparison. On the other
hand, in many areas we do find empirically well-founded uniformities in human
behaviour. Our way of proceeding springs from the will and hope to develop,
from such empirically proven knowledge, a useful working model of the processes
taking place within the individual. Usefulness is measured according to whether
or not the model can be validated by the behaviour of real collectives and in terms
of whether the simulation based upon the model widens and furthers our
understanding of the dynamics of these processes.
72 Hans-Joachim Mosler
For this reason, we first design a basic model of an individual (see Mosler 2000
for details), which – as 10,000 identically structured copies (equipped, however,
with individually different characteristics) – serves as the basis for the simulated
influencing and resource-use processes (see Figure 4.1). These persons differ
individually only in their values of the variables. They all function according to
the same social psychological principles. These principles are based on a few central
and well-founded essences of theories, which are presented in detail in the
presentation of results. The framework model first specifies the input and output
variables of the theory-based processing by individuals. The model yields inform-
ation about the inner psychological processes that take place when people use
environmental resources (for example, use the resource of air when they drive or
heat their homes) or influence each other mutually, whether deliberately or
unintentionally. Inner psychological processes are triggered as people communicate
with one another in daily life and observe themselves and others. These processes
change, in dependency on further internal and external conditions, the ways in
which we feel, think, and argue about the environment and the way we act toward
the environment. These simulated individuals have at their disposal differently
structured social contact nets (number of friends, acquaintances, neighbours, and
strangers they observe).
In its basic form, the model is an expanded ‘rational choice’ model (Diekmann
1996; Becker 1976). This means that the variables and processes specified in the
model more or less consciously enter into the decisions relevant to using behaviour.
The variables and processes thus having an effect are not, however, purely cognitive
components. They include, in addition, motives, emotions, norms, and so on. This
is what distinguishes our model from the usual ‘rational choice’ models: with a
psychological conception, it is not a problem that non-rational components enter
into rational decision-making. The following explains the external and internal
input variables as well as the inner processes.
External input variables are the influences that exert upon a person from the
outside and are perceived by the person in some form. Possible distorted perceptions
of individuals on the basis of biases are not taken into consideration in this model.
For each individual some output variables are calculated that function as external
input variables for other persons with whom the individual has contact (‘contacts’)
(see Figure 4.1). Also, some non-social input variables, such as specific situational
parameters and the momentary state of the resource being used, enter into the
calculation of an individual’s output value.

• Use, contact: A summary of all environmentally related behaviours shown by a

contact. These behaviours towards the environment are conceived as resource-
using behaviours. For resource utilization, a particular resource can be entered
into the model with its characteristic parameters (resources: for example, water,
air, wild game populations; parameters: for example, rate of regeneration).
• Attitude, contact: A summary of a contact’s opinions and evaluations of
environmental issues, objects, and behaviour towards the environment, as
expressed in various ways.
 ss
Indi vidu als wi th ` ´ýP ` ´ýP ne us
wh om In divi dua l 1 Submodels at Individual 1
Va
sive St e
ha s a ny fo rm o f lue ud
s Processing of ua tit
con ta ct => contacts rs At
group Pe
influences
Interna l stru ctu re Use, contact Attitude
id entica l to Kn
ow
Indi vidu al 1 le
dg Processing of
e
discrepancies

Attitude, contact Subjective

norm
Processing of
observation of
other's
behaviour
Persuasiveness, Behaviour Using
contact control be havi our

Volition
re
sp S
on elf Processing of
si - resistance
Behavioural Intention
bi
lit
y Cost-bene-
Status, contact fit analysis
Processing of
s M collective use
rs or ot
to ct ive
ac fa s s
l f es nal tive
na tiv io n Sustaina-
tio en
a c at ce Processing of bility motive
tu in tu in
Si nd Si nd social influen-
a a ces with know-
ledge and
concern

State of the resource

Figure 4.1 Framework model of behaviour toward the environment

Modelling environmental behaviour
73
74 Hans-Joachim Mosler
• Status, contact: This variable represents the sum of a contact’s relevant personal
resources (for example, social competence, trust, prestige, knowledge, power,
possessions).
• Persuasiveness, contact: The intensity and quality with which a contact makes a
case for various attitudes related to the environment. While direction is given
for an attitude (in other words, ‘for’ or ‘against’ environmental protection),
persuasiveness indicates the intensity and quality with which these attitudes
are presented.
• Situational factors and incentives: These include all influences from the societal-
institutional surround (rules and prohibitions, positive and negative incentives,
and the like). From the entire spectrum of the possible scope of a behaviour,
situational factors ‘filter’ a partial spectrum.
• State of resource: State of the environment, or state of a particular environmental
resource, as ‘noticed’ and established by the person. Internal input variables
have an effect on the simulated psychological processes ‘from the inside’. The
emergence of individual degrees of markedness of these variables due to the
individual’s learning history is not examined in this simulation.
• Values: Stable orientation with regard to environmental facts, objects, and
behaviours.
• Knowledge: Extent and quality of information about the environment, such as
knowledge of the regeneration parameters of a specific resource.
• Self-responsibility: Describes the extent to which people attribute responsibility
to themselves for the state of the environment (as opposed to holding other
persons, organizations, or institutions responsible).
• Motives: A summary of various motives, such as curiosity and laziness, that
enter into a person’s readiness to act in certain ways. Dependent upon these
motives, an act – independently of its consequences – will be viewed as ‘easy’
or ‘desirable’.

External and internal input variables are processed in different sub-models (see
Figure 4.1) according to the theory being applied. The output values of the sub-
models have either a direct outward effect on other persons with whom the
individual has contact (attitude, persuasiveness), or they affect variables leading
up to behaviour and decisions on the use of resources and consequent actions.
The following sub-models were designed and simulated (simulations of sub-models
II and IV will not be presented):

• Processing of group influences upon attitudes towards the environment:

application of the theory of Social Comparison Processes.
• Processing of discrepancy between environmental behaviour and
environmental attitude: application of Dissonance Theory.
• Processing of observation of others’ behaviour towards the environment:
application of Bandura’s Social Learning Theory.
• Processing of resistance to environmental protection measures: application
of Reactance Theory.
 Modelling environmental behaviour 75
• Processing of information on the collective utilization of an environmental
resource: application of Commons Dilemma research.
• Processing of communicative influence in view of the individual’s feeling of
concerned consternation about the state of the environment, knowledge about
the environment, and biases: application of the Elaboration Likelihood Model.

Variables leading to behaviour are direct preliminary steps towards behaviour.

On the basis of the Theory of Resource Mobilization (Klandermans 1984) and
of the Theory of Planned Behaviour (Frey et al. 1993b; Ajzen 1991; Ajzen and
Madden 1986), we start from the assumption that the following five components
play an important role in behavioural intentions relating to the environment:

• Attitude towards the environment as the attitude towards environmental

protection and individual behaviour regarding the environment
• The subjective norm as ‘perception of pressure from the social surround’.
This expresses the expectations that confront the individual and the degree to
which the individual is willing to fulfil these expectations
• Behavioural control as the subjective conviction that one can in fact carry out
the behaviour
• Cost-benefit analysis as a motive for behaviours, which results from weighing
the direct costs and direct benefits of actions
• Sustainability as a motive, leading to willingness to restrict personal use of a
resource. This readiness depends upon both the absolute value that a person
places on an environmental resource and on the current discrepancy between
collective, sustainable use and the actual pattern of use shown by others. The
higher the subjective value of the environment, or the smaller the discrepancy
between actual and sustainable patterns of utilization, the greater the effect
of the sustainability motive in the sense of a person’s own restraint in use.

All of these variables are conceptualized at a general level, thus according with
Ajzen and Fishbein’s (1977) correspondence hypothesis, which states that there is
a high correlation between behaviour and the determining variables if they have
been operationalized with the same degree of specification.
As a fundamental extension of the theory of planned behaviour (Ajzen and
Madden 1986), we have added an additional factor, cost-benefit analysis. Although
the theory of planned behaviour does contain a kind of cost-benefit analysis, in
that its components are broken down in ‘expectation ´ evaluation products,’ we
conceptualise this factor as an independent component. With this we expect to
achieve better explanation of the variance for intention, but more importantly, to
gain added starting points for intervention measures. Cost-benefit analysis models
people’s ‘economic’ considerations, as they determine whether it will pay for them
to act in certain ways. The model is thus the first to integrate economic and social
psychological components (see also Mosler and Tobias 2000).
These components all influence behavioural intention, which leads to behaviour
towards the environment. Additionally, we introduce volition into the model, which
76 Hans-Joachim Mosler
is closer to actual behaviour than behavioural intention. Psychological research
(Frey et al. 1993b) has shown that the act of making a decision commits a person to
the when and where of attempting to realise an intention. As the entire model is
deterministic in construction, random variables play absolutely no role in deter-
mining individual behaviour towards the environment.

Simulation experiments and results

In the following, after a brief introduction of the various sub-models, the conditions
for the simulation experiments are described, and the results are presented. In
describing the sub-models, we focus on experiments with populations of 10,000
persons and omit the experiments that we conducted with 1 to 10 persons in order
to reach a basic understanding of the processes and to validate the model with the
aid of existing empirical findings. All variables range within a dimension from 0
to 100. For variables related to environmental orientation, 100 signifies maximum
environmental friendliness, and 0 indicates maximum lack of environmental
friendliness; 50 represents a point of ‘neither/nor’. In the graphics presenting the
findings, average behaviour towards the environment, or the average attitude
towards the environment by the population, is always put on the ordinate. If a
curve rises, this means that the population is changing in the direction of
environmental responsibility. On the abscissa, we find the simulation steps. A step
means that in a complete calculation cycle, all individuals exert influence on their
contacts, or are themselves influenced, and – together with the others – utilize a
resource. The linear progression of the steps should be interpreted as the
progression of time; more exact temporal pinpointing is not possible.
For the experiments, we usually started out from populations that were rather
eco-unfriendly (mean of 40), whereby the values of this variable in the individuals
strew around this mean. In order to avoid the reproach that we may have built the
results of the simulation right into the program, all experiments were conducted
according to a baseline/control group design. The results of a number of steps
with and without manipulations (control group) form the basis for comparison;
populations were always identical (which is only possible in computer simulation).
In the following, we will examine important issues in social intervention and
campaign planning with the aid of the simulation model. The following sections
on the sub-models are each organized according to a research issue and a reference
to the theories being applied.

Group influences upon attitudes

How must a minority of people behaving in an environmentally friendly way be
distributed and networked within a population in order that the environmentally
unfriendly majority comes to change its attitudes and behaviour? Frey (1993a:
114ff.) developed an integrative concept of social comparison processes that
encompassed Festinger’s theory (1954), Tajfel’s theory of social identity (1978,
1979; cited in Frey et al. 1993a), and Thibaut and Kelley’s theoretical concept of
 Modelling environmental behaviour 77
the comparison level for alternatives (1959; cited in Frey et al. 1993a). According
to the integrative conception of social comparison processes (Frey et al. 1993a:
111ff.), persons in groups change their own attitudes in dependency upon the
existing pressure to conform, the attractiveness of the group, and the perception
of threat to their self-concepts should they change their own positions. A change
in attitude results from the pressure to conform, weighted by the attractiveness of
the group. If, however, the person’s self-concept is threatened, there will be no
attitude change (see Figure 4.2).
Threats to a person’s self-concept arise when their values differ to a certain
degree (a value of 30) from the average attitudes of other members of the group.
Individuals experience too great a discrepancy between their own values and the
group’s attitude as threatening: in order to adapt to the group, the individual would
have to deviate too far from personal values. Under social pressure, they would
become untrue to themselves. People are more likely to change an attitude the
more attractive the group is and the greater the group pressure to conform. Attitudes
will change in the direction of the average attitude of the group. Attitude change
occurs, however, only if the change does not pose a threat to self-concept. The
attractiveness of the group results from two components: (1) the difference between
a person’s own attitude and the average attitude of the group – a group is all the
more attractive the more that person and group are on the same ‘wavelength’;
and (2) the average status of the group – its prestige, power, social resources. The
conformity pressure of a group can be calculated according to Tanford and Penrod’s
(1984) well-documented formula. Pressure to conform, in reference to the total

INDIVIDUAL

PROCESSING OF GROUP INFLUENCES

Values Threat to self-concept

Attitudes of contact Pressure to

persons conform Attitude

Status of contact
persons Attractiveness of group

Figure 4.2 Representation of the ‘processing of group influences’

78 Hans-Joachim Mosler
number of group members, is calculated in an exponential function separately for
a minority and a majority member.
The population is structured in groups of 10 persons, whereby each group is in
contact with two other groups. For the intervention, we assume that an
environmental campaign can bring 10 per cent of the population to clearly embrace
more friendly attitudes towards the environment than previously (an increase in
25 points), for a limited period of time (for example, for one week = one step in
the simulation). These persons we call (attitude) ‘pioneers’. Pioneers are, moreover,
more strongly sensitized to deviations from their own positions. Through this,
they acquire a particularly low susceptibility to social influencing attempts: even
minor deviations of other individuals in the group from their own, pioneer positions
imply a potential endangering of their sense of self-esteem. This leads to the less
susceptible attitudes of the ‘pioneers’ in the group.
As a further form of intervention, pioneers are variously organized. Following
the short environmental campaign: they remain in their groups (isolated) with no
wider contacts to the outside; they remain in their groups, but are in contact with
other pioneers (networks); or they form their own groups (core groups), which
may have few or many contacts with other groups. Figure 4.3 documents the
experiments conducted as well as their results. The intervention applies only to
the fifth step; the dynamics described below result from this short-term intervention.
It is extraordinarily important for pioneers’ susceptibility to be low, in order
that they do not too soon once again adapt to the surrounding majority. If
susceptibility is not low, there will be no positive effects from being networked with
other pioneers and from no longer experiencing themselves as a minority. Core
groups (also with low susceptibility) only then have an effect on the surrounding
majorities in the group if they have numerous outer contacts (curve F in Figure
4.3 continues to climb in further steps, but not as high as curves B and D). In sum,
pioneers have to hold to a stronger environmentally responsible attitude, and their
susceptibility to influence must be low. In effect, they feel quickly threatened in
their self-concept. Pioneers should be activated or active in as many groups as
possible and should not be concentrated in just a few groups. Further measures
are not required.

Social learning of behaviour

What factors lead the population to follow the examples of behavioural role models?
According to Bandura (1979) people learn not only through direct personal
experience, but also through observing others. In this way they gain an idea of
how an action is performed. The probability that a learned behaviour will be
carried out depends on motivation, and motivation is dependent upon expectations
of efficacy (self-efficacy, Bandura 1986: 390ff.) and determined by anticipated
self-regulation (Stalder 1985; Bandura 1986: 390ff.). Bandura does not, however,
give any indication of the effect on a person’s behavioural repertory when there
are several, contradictory behavioural models. For example, someone might observe
environmentally responsible behaviour in some people and behaviour that is
 Modelling environmental behaviour 79
Environmentally
responsible
55

50
Attitude of the population

45

40

35
Not 0 5 10 15 20 25
environmentally
Runs
responsible
Intervention

Control population D: Networked ‘pioneers’ with

low susceptibility
A: Isolated ‘pioneers’, E: ‘Pioneer’ core groups, few
normal susceptibility contacts, low susceptibility
B: Isolated ‘pioneers’, F: ‘Pioneer’ core groups, many
low susceptibility contacts, low susceptibility
C: Networked ‘pioneers’ with
normal susceptibility

A. In each group there is an isolated pioneer whose susceptibility to influence is the same in degree
as that of the members of the group majority.
B. In each group there is an isolated pioneer whose susceptibility is lower than that of the members
of the group majority.
C. In each group there is a pioneer whose susceptibiity is the same in degree to that oo group
members; in addition, the pioneer is in contact with 10 other ‘pioneers’ in other groups. They
thus no longer experience themselves to be a minority.
D. As in C above, but here pioneers have a lower degree of susceptibility.
E. All pioneers are concentrated within their own core groups and show low susceptibility
toinfluence. Each pioneer is inn contact with another person outside of the core group.
F. As in E above, whereby here pioneer group members are in contact with 10 other persons
outside of the core group.

Figure 4.3 Change in average attitute due to various group-oriented interventions

damaging to the environment in others. In order to model contradictory influences

on self-efficacy, we turned to Latané’s theory of social impact (Latané 1981; Latané
and Wolf 1981; further specified in Nowak and Latané 1994; Nowak et al. 1990).
In this theory, social influence bases upon the following factors, which stand in a
multiplicative relationship:

• Strength: power, importance, intensity, unusual quality or features of the source

person for the target person
80 Hans-Joachim Mosler
• Immediacy: directness, immediacy in space and time, absence of barriers or
filters
• Number of sources: number of group members, number of persons present.

The multiplicative relation of the three variables expresses the fact that the effect
of one of the variables is greater, the greater the value of the other variables.
There is no effect at all if one of the variables equals zero. According to the
theory, moreover, the effect of the variable N (number) is not linear, but rather is
an exponential function: I = sN t, where I is impact, s is a constant, and the exponent
t is a value less than 1. The parameters s and t are different for each situation and
have to be determined empirically. The factor ‘number’ thus has the effect that
the first person has the greatest impact and each person thereafter has ever less of
an impact. With an increasing number of influencing source persons, the social
impact on a person rapidly decreases.
Expectations of efficacy can be represented approximately by behaviour control,
because we define behaviour control as a person’s conviction that they can indeed
execute the behaviour. In accordance with the framework model, behaviour control
affects the behavioural intention and, thus, behaviour. According to Latané’s theory
of social impact (1981), behaviour arises from observation of the using behaviour
of contact persons: social influence is a multiplicative function of the number,
strength, and immediacy of contact persons. In our model (see Figure 4.4), the
variable ‘number’ is determined by the number of the individual’s contact persons
who behave in an environmentally responsible, or irresponsible, way. The ‘strength’
of contact persons’ influence is defined by their status. ‘Immediacy’ is conceived

INDIVIDUAL

PROCESSING OF OBSERVED BEHAVIOUR

Using behaviour

Using
behaviour of Behaviour control
contact persons Intention

Status of contact persons

Values Self-confirmation
Volition

Self-responsibility
Using
behaviour

Figure 4.4 Representation of ‘processing of observed behaviour twards the environment’

 Modelling environmental behaviour 81
of as psychological immediacy. This means that friends have an influence of 1/1,
acquaintances of 3/4, neighbours of 1/2, and strangers of 1/4. The social
influence of contact persons who behave in an environmentally responsible way
and that of those who behave irresponsibly are calculated separately according to
the formula and then compared by dividing (compare Nowak and Latané 1994;
Nowak et al. 1990). This ratio determines the behaviour control of the person: a
person’s belief that they are able to perform the concrete environmentally friendly
behaviour (that is, their behaviour control is the more environmentally responsible)
is stronger the greater the number and the higher the status of individuals behaving
environmentally responsibly in comparison to persons not behaving
environmentally responsibly that they observe.
Self-confirmation arises when people compare their own values (personal choice
of behavioural standards) with their behaviour. Self-confirmation is also weighted
by self-responsibility in order to incorporate the element of feeling responsible. A
particular behaviour is all the more important to a person the better it accords
with the person’s values and the more strongly the person feels responsible. Self-
confirmation affects volition and, thus, behaviour.
Within the framework of an intervention, incentives can move persons having
above-average status, for example, to change their behaviours during the inter-
vention phase in the direction of environmental acceptability. On the other hand,
it is possible to raise the status of persons displaying above-average environmentally
friendly patterns of behaviour through various measures (for example, by means
of public commendation awards, coverage in the media, and so on). The experi-
ments were conducted with 300 or 500 persons and included varying degrees of
‘visibility’, in that selected persons displayed their environmental behaviours to
many (15) or a few (5) contacts (Figure 4.5). Visibility of resource use has already
been demonstrated to be an effective factor (Jorgenson and Papciak 1981; Mosler
1993).

A. In this experiment, 500 persons of high status are selected and caused to
adopt a behaviour that is more eco-friendly for the duration of the intervention.
They are to demonstrate this behaviour to many others (15 contacts).
B. As in A above, but here only 300 persons are selected.
C. As in A, but here the behaviour is demonstrated to only 5 contacts.
D. Here we selected 500 persons showing very environmentally sustaining
behaviour patterns, which they demonstrate to 15 contacts. In this case, their
status is raised for the duration of the intervention.

It appears to be more effective to select persons having high status as role models
and to induce them to behave in a more environmentally friendly way for the
duration of the intervention than to temporarily raise the status of persons already
showing such behaviours. An increase in the number of contacts, that is, an increase
in the visibility of the environmentally sound behaviour, achieves relevant effects.
An increased number of role models also has a strong effect. The dissemination
process continues to progress automatically for some time after the intervention,
as the entire social system must again adapt to the changes.
 82

Environmentally
responsible 56
A. 15 CP/500 P/Act+
54
B. 15 CP/300 P/Act +
52
Hans-Joachim Mosler

C. 5 CP/500 P/Act +
50
D. 15 CP/500 P/Status +
48

46

44
Control situation
42

Not 40
environmentally 0 5 10 15 20 25 30 35 40 45 50 Runs
responsible
Intervention

CP: contacts per person; P: number of persons addressed (role models); Act+, Status+: more responsible acts towards environment.

Figure 4.5 Average population behav iour with differing interventions based on observational learning
 Modelling environmental behaviour 83
In developing concepts for environmental campaigns, simulation can contribute
support with regards to the implementation of efficient means. It can answer
questions as to the number of role models required, how these should be selected,
and what changes they should show, the degree of visibility necessary, and so on.
In particular, simulation allows us to estimate the effects that can be achieved by
means of compensation in other values, which makes a direct contribution to an
increase in efficiency (for example, an increase in visibility at the expense of the
number of role models). However, the simulation method does not allow conclusions
about implementing the results, such as statements on how realistic it is to induce
high status people to behave in a more environmentally friendly way. As discussed
above, the simulation method permits only comparative conclusions.

The commons dilemma

How must the social surround be perceived for people to become willing to use an
environmental resource sustainably? When people act upon the environment, they
usually utilize an environmental resource. Utilization of a resource available to all
(common property) involves an interpersonal conflict (Liebrand et al. 1992; Mosler
1995). It is in each individual’s interest to keep the personal benefit as large as
possible, while the harm incurred by or the depletion of the resource must be
borne by all (Spada and Ernst 1990; Mosler and Gutscher 1996).
The sustainability motive expresses the degree to which a person is willing to
protect common property, the environment. This motive depends upon both the
absolute value a person places on environmental goods and the current discrepancy
between collective, sustainable use and actual patterns of use shown by others
(expectation of sustainable utilization). The more highly the individual values the
common property of the environment, or the smaller the discrepancy between
actual and sustainable patterns of use seems, the greater the effect a sustainability
motive will have. Under these conditions, the individual is motivated to make a
personal contribution to sustainable patterns of use.
The sub-model of the commons dilemma (compare Figure 4.6) yields a sustain-
ability motive as an expression of the desire to use the common property in a
resource-sustaining way through cautious use. The sustainability motive results
from a multiplicative connection between the expectation of sustainable utilization
and valuation of the resource ‘environment’.
Valuation of the resource ‘environment’ is determined by the perceived state
of the resource (state of the environment) and a person’s environmental values.
Environmentally-oriented people use exhausted resources less. Various studies have
shown the legitimacy of equating social orientation with environmentally-oriented
values (Liebrand 1986; Liebrand et al. 1986; Mosler 1990). The degree of exhaust-
ion of a resource is calculated on the basis of its state, as a percentage of deviation
from its maximal state, and weighted by values. This means that a person’s eval-
uation of the resource is higher the more depleted a resource is and the more the
person is environmentally oriented. Our model presently conceives of the percep-
tion of the state of the resource as direct perception, that is, without any distortion
84 Hans-Joachim Mosler
INDIVIDUAL

PROCESSING OF INFORMATION ON COLLECTIVE USE

Estimated proportion
Using behaviour of resource- Expectation of
of contact sustainable Sustainability
sustaining users
persons utilization motive

Required proportion Intention

of resource-
Knowledge about sustaining users
the resource

Perception of the Volition

state of the Valuation of
resource the resource

Values Using
behaviour

Figure 4.6 Representation of ‘processing of information on the collective use of a resource’

stemming from inner psychological biases or from external influences in the form
of media reports, disputes among the experts, and so on.
The expectation that one’s own way of using the environmental resource will
make a contribution towards preservation of the resource is formed from the required
proportion of resource-sustaining users in the population and the estimated
proportion of resource-sustaining users. In other words, the greater the proportion
of non-sustaining users of a resource as estimated by the person in comparison with
the required proportion of resource-sustaining users in the population, the lower
the person’s expectation that they themself can make an effective contribution to
sustainable use of the resource. This expectation rises in the degree to which the
proportion of estimated resource-sustaining contact persons exceeds the required
proportion of resource-sustaining users. Knowledge of the number of resource-
sustaining users required in order to preserve the resource comes from resource
knowledge and the perception of the state of the resource. The model presumes
people to have – at present not a very realistic assumption! – knowledge about the
regeneration rate, regeneration characteristics, and the maximal state of the resource.
From the discrepancy between the maximal state of the resource and the current
state of the resource, the person derives the required proportion of resource-sustaining
users. The greater the discrepancy between the current and the maximal state of
the resource, the greater must be the proportion of the population that uses the
resource in a sustainable way if the resource is to recover.
The estimated proportion of resource-sustaining users in the population is based
upon the behaviour of contact persons perceived by the person. According to the
ratio of contact persons using the resource sustainably and non-sustainably, the
person makes an estimate for the entire population. This factor is at present not
 Modelling environmental behaviour 85
yet weighted by socially, or environmentally, oriented values. For example, people
who are not environmentally oriented overestimate the number of like-minded
people.
If no particular interventions take place, there is a danger that dynamics such
as those shown in Figure 4.7a will develop. With an optimistic starting value with
regard to average use within a population (over 50 = environmentally friendly
use), the state of the resource briefly improves. Due to its improved state, its value
declines (only goods in short supply are valuable), and the individuals in our
simulated population resume stronger exploitation of the resource. From the fifth
step onwards, this tendency results in clear over-utilization (average use under 50).
This increases the discrepancy between actual and sustainable use patterns, and
expectations of sustainable use correspondingly decrease. The individual is less
motivated to make a personal contribution to sustainable patterns of utilization
(‘… personal restraint on my own part would not make any difference; no one else
is showing restraint, so it is better for me to help myself to the resource so long as
it is still up for grabs …’). And so the state of the resource deteriorates. Its value
rises, which does not, however, lead to a marked reduction in utilization. There is
no stopping the course of these ‘downhill’ dynamics, and the resource is destroyed.
To counteract these negative dynamics, we ran a campaign in the fifth step that
aimed to (a) lead persons, in their own use behaviour, to orient themselves less to
other persons’ patterns of use and (b) lend heavier weighting to the importance of
the resource. Through this, the common property becomes more highly esteemed.
We found that the utilization behaviour of the population becomes more
environmentally sustaining (see Figure 4.7b). If individuals use a resource in an
environmentally sustainable fashion on the average, the state of the resource
improves. This line of development continues until the resource has regenerated.
At this time, the resource is available in over-abundance so to speak, whereby its
value again declines, and intensity of use increases. Downhill dynamics develop
until that crucial point where they are again brought under control. A dynamic-
stable balance has emerged, in which utilization continually adapts to the state of
the resource by means of inner personal factors. The resource will never be
completely destroyed.
Through the interaction of the social system and the resource, the resulting
system behaviour shows large fluctuations. Following Forrester (1972: 48ff.),
fluctuating system behaviour results when in a system of interlocking feedback
loops two or more temporal delays occur. In the present simulation model, there
are delays both within the social system and between the social system and the
resource system. The average expectation that one can make an effective
contribution to sustainable use of a resource reacts with a delay to the average
resource use behaviour of the social surround. The rise and fall of the average
value of the common good reacts with a delay to the development of the resource.
Our simulation approach is not directly comparable to Forrester’s approach.
Forrester observes only one macrosystem, while the simulation builds upon
numerous microsystems (individual persons) which join to form the macrosystem
(the social system).
86 Hans-Joachim Mosler

Environmentally
responsible/
resource maximal
100

80

60

40

20

0
Not 0 5 10 15 20 25 30
environmentally Runs
Resource in %
responsible/ Intervention
resource minimal Average use
Average value of common resource
Average expectation of sustainable use

Environmentally
`´ ýP

responsible/
resource maximal
100

80

60

40

20

0
Not 0 5 10 15 20 25 30
environmentally
responsible/ resource in % Runs
resource minimal average use
average value of common resource
average expectation of sustainable use

Note: Also illustrates the course of significant inner variables. 4.7a (top): Control situation, 4.7b
(bottom) with intervention.

Figure 4.7 Use of an environmental resource

 Modelling environmental behaviour 87
Processing communicative influence
What is the effect of convincing attempts (persuasion) in populations, according
to individuals’ concern about the environment, knowledge of the environment,
and biases? Here we apply the Elaboration Likelihood Model (Petty and Cacioppo
1986a, b). People process information in varying degrees of thoroughness: the
‘intensity’ of processing is a function of processing motivation and processing
ability; areas of great personal relevance (concerned consternation about the
environment) and self-responsibility (here responsibility for the state of the
environment) increase motivation to process (Stahlberg and Frey 1993). Where
deep, complete processing of information occurs, the results of campaigns to
convince are primarily dependent upon the quality of the arguments presented;
in the case of superficial processing, superficial cues, such as status, credibility and
attractiveness of the communicator, gain more weight. Furthermore, individuals
holding extreme attitudes or values tend to process information in a biased way:
when a person’s own position is too removed from the attitude advocated by the
influencing campaign, there can be a ‘boomerang’ effect, whereby the person
changes their attitudes in a direction opposing the persuasive arguments. If the
advocated position is close to one’s own, there will be convergence between them.
The principle mechanisms involved may be illustrated by a simplified, information-
campaign experiment.
The four external input variables in Figure 4.8 represent external influences
on the processing individual. The three variables in the upper section of the figure
represent influences stemming from the source of persuasion, a communicator.
The communicator communicates a particular attitude position with a certain
degree of persuasiveness. This persuasiveness variable corresponds to ‘argument
quality’ in Petty and Cacioppo (1986b). The fourth external input variable that is
processed is the (perceived) current state of the environment. From the left, two
free, internal input variables – self-responsibility and values – flow into the sub-
model. These are not affected by the processes within the model, just as the status
of the individual also remains constant.
Concern arises from the processing of information on the state of the
environment and from a person’s own values as related to the environment. Concern
corresponds to the discrepancy between these personal values (the desired value
for the state of the environment) and the actual value for the state of the
environment (the effective, perceived state of the environment). This means that
the greater the deviation between the individual’s values in relation to the environ-
ment and the perceived state of the environment, the greater the feeling of
concerned consternation. This is true only as long as those values are at a higher
level than the state of the environment, or in other words, only as long as the
environment is in worse shape than the individual would like to see.
The knowledge that the individual has changes depending upon the degrees of
persuasiveness the individual has. Environmental knowledge, in its function as an
ability variable, determines the degree of processing depth, or intensity. Processing
motivation results from the sum (addition) of the feeling of self-responsibility –
88 Hans-Joachim Mosler
INDIVIDUAL
PROCESSING OF SOCIAL INFLUENCING UNDER CONDITIONS OF
KNOWLEDGE AND CONCERN

Communicator's persuasiveness

Communicator's attitude

Communicator's status

Self-responsibility
Processing
Status
motivation
Perception of state Concern
of environment
Processing intensity Attitude

Values

Persuasiveness
Processing ability
Knowledge

Figure 4.8 Representation of ‘processing of communicative influencing under conditions

of knowledge and environmentl concern’

how strongly the person experiences a sense of responsibility for the state of the
environment – and environmental concern. Processing intensity (elaboration
likelihood) arises in multiplicative fashion from processing motivation and
processing ability. In addition, it is influenced by the status of the communicator
(but only where neither processing motivation nor processing ability equals zero).
A communicator’s positive status raises processing intensity, while negative status
lowers processing intensity.
The difference between the attitude position of the individual and the position
advocated by the communicator determines attitude change – in dependency,
however, upon the influences from central and peripheral processing. If processing
intensity is low, peripheral processing will predominate, and the status of the
communicator is the main determinant of attitude change. High status of the
communicator leads to a change of attitude to the position advocated. Low status
of the communicator leads to change of attitude opposing the advocated position.
If processing intensity is high, however, central processing takes over, and the
persuasiveness of the communicator is the main determinant of attitude change.
High persuasiveness is expressed by high quality arguments, and the result is a
change of attitude in the desired direction. Low persuasiveness, accompanied by
weak arguments, leads to a change of attitude in the opposing direction.
Biased processing, moreover, modifies central processing. The individual’s
environmental values are compared with the attitude position advocated by the
communicator. If they agree (when both are environmentally friendly or
environmentally unfriendly), the communicator will successfully influence the
 Modelling environmental behaviour 89
person in the desired direction. If there is a discrepancy, however, the person’s
attitude will change in the opposing direction.
To the individual’s initial persuasiveness, a central and peripheral proportion is
added in proportion to processing intensity. Where processing intensity is high,
the individual’s persuasiveness increases if the persuasiveness of the communicator
is higher than the individual’s persuasiveness. Here the individual integrates the
communicator’s good arguments into his own argumentation schema. Where
processing intensity is low, the individual’s persuasiveness increases if the
persuasiveness of the communicator is higher than the individual’s persuasiveness.
In contrast, if the persuasiveness of the communicator is low, the individual’s
persuasiveness decreases. This assumption is our own and is not based upon the
Elaboration Likelihood Model. The idea is that unreflected acceptance of weak
arguments tends to weaken a person’s persuasiveness.
We assume to this purpose (as an exception) that the individuals do not influence
each other mutually, but that in five steps they are presented with information
campaigns (the individuals do not discuss the subject of the environment among
themselves, but they all stand under the influence of the campaigns). The campaigns
promote strong environmentally friendly attitudes by means of good arguments
(leading to high persuasiveness) and/or by means of strong peripheral cues
(communicators having high status, as for example esteemed public figures). In
addition, we assume that in the campaigns, an appeal can cause persons’ self-
responsibility to increase. Average attitudes and values are low on environmental
friendliness, the extent of the sense of self-responsibility is low, and the state of the
environment is poor. Persons having environmentally friendly values experience
high concern under the condition of poor state of the environment. The following
interventions were ‘tested’ (see Mosler et al. 1998):

• An information campaign presenting upon good arguments (Figure 4.9,

persuasiveness+)
• An information campaign which works with well-known and esteemed
personalities (Figure 4.9, status+)
• An information campaign based on good arguments and well-known, esteemed
personalities (Figure 4.9, persuasiveness+/status+)
• An information campaign with well-known, esteemed personalities that
moreover appeals to people’s sense of environmentally-oriented self-
responsibility (Figure 4.9, self-responsibility+/status+).

Let us first examine results with individual persons selected from the population.
Person A – who is biased due to an environmentally non-friendly attitude – places
a negative value even on good arguments (curve Persuasiveness+). In contrast,
good arguments effect even more environmental consciousness in Person B, who
has environmentally responsible attitudes from the start, as information processing
is hardly biased here.
A campaign using esteemed public figures is effective with both Persons A and
B, if they tend to process information relatively superficially (curve Status+). But
90 Hans-Joachim Mosler
Environmentally
responsible
70 70
Person A Person B C C
60 60 B B
D D
A A
50 50
40 BC
40
DA
30 B B
C 30
C
20 20
BC
10 A A 10
DA
0 D D 0
Not 0 5 Intervention 10 15 0 5 Intervention 10 15
environmentally Runs
responsible Runs

Environmentally
responsible Campaigns with:
90
Population A Campaign: persuasiveness+
80 B Campaign status+
70 C Campaign: persuasiveness+/status+
C C
B B D
60 Campaign: self-responsibility+/status+
A A
50 D D Control situation

40

30
Not
environmentally 5 Intervention 10 15
responsible Runs

Note: Above individuals in the population. Below population values. The three graphs present
courses of attitude change in four different experiments and the control situation.

Figure 4.9 Information campaign experiments without environmentally relevant

communication

if Person A’s depth of processing is increased by means of self-responsibility (curve

Self-responsibility+/Status+), they then show biased processing, and their attitude
changes in the direction away from that advocated by the campaign. In the total
population, the average attitude becomes most changed by means of good
arguments and presentation by esteemed personalities. But as shown in Figure
4.9, we must always take into consideration that these average values within the
population hide very contrasting effects upon differing individuals.
For various experiments with multiplicators (persons who advocate their environ-
mental attitudes to others), a communication net was constructed. Each simulated
person ‘talks’ about environmental issues with five friends and one stranger during
each step, whereby the friendships are assumed to be lasting and reciprocal. The
relationships to strangers are different in each step. Initial values in the population
are the same as those described above for the information campaign experiment.
Within the framework of an environmental action campaign, 500 persons
(multiplicators) having very environmentally friendly attitudes receive training in
argumentation (increasing persuasiveness, see Gonzales et al. 1988). In a variant
 Modelling environmental behaviour 91
of this experiment, the status of these persons is raised (as for example through
public commendation awards). The following interventions were carried out:

• Multiplicators with good arguments (Figure 4.10, Persuasiveness+)

• Multiplicators with raised status (Figure 4.10, Status+)
• Multiplicators with good arguments and raised status (Figure 4.10,
Persuasiveness+/Status+)
• Control situation.

In Figure 4.10 we again see two selected individuals A and B, who both stand
under the influence of the action campaign. A glance shows that these graphs are
not as smooth as the information campaign graphs. The curves would be smoother
if, after a short levelling-off phase, there were only the influence of a constant
group of friends. But the ever-changing influence of strangers, upon Persons A
and B and also upon their friends, causes fluctuations in attitudes. We can still
recognize certain tendencies in the development of attitudes, however. Person A’s
attitude becomes less environmentally friendly relatively quickly. Later changes to
a more responsible attitude are only brief and are repeatedly destroyed by the
environmentally unfriendly social surround. Person B shows initial swings in
attitude, but then develops a tendency towards a less environmentally friendly
attitude. Both Persons A and B react to the introduction of multiplicators having
high status (curve Status+) with an increasingly environmentally friendly attitude.
This effect is even stronger if multiplicators in addition show high persuasiveness
(curve Persuasiveness+/Status+). But if multiplicators demonstrate only
persuasiveness (curve Persuasiveness+), they have no effect upon Person A. There
is an initial effect in this case upon Person B, but the influence of the rather
environmentally unfriendly social surround cannot be cancelled out. The same
effects find expression at the level of the population: multiplicators with raised
status are just as convincing as those who also show increased persuasiveness.
Multiplicators who have only high persuasiveness at their disposal have a counter-
productive effect.

Summary and conclusions

The concept of the model, which is based on an elaborated rational choice theory,
has proved to be – in connection with the simulation method – a valuable tool for
the drafting, designing, and testing of environmental psychological interventions.
The most important results of the experiments with the various sub-models can
be summarized as follows:
People who as a minority wish to effect more environmentally responsible
attitudes in their group absolutely must have a low susceptibility to influence. If
this condition is fulfilled, no measures for networking the minority persons need
be undertaken.
Role models can be implemented more successfully if these models have high
status and their behaviour is highly visible.
92 Hans-Joachim Mosler
Environmentally
responsible
Person A Person B
100 100 C C C
C C
B B B
80 B
80
C CB B C
B A B
60 C B C 60 D
C B DA
B D
40 40 D D A
A A A D
20 20
C
D
A A A D
B D A
0 D D A A D 0
Not
environmentally 10 20 30 10 20 30
responsible Runs Runs
Intervention Intervention

Environmentally
responsible
80
Population
70
A Multiplicators: persuasiveness+
60
B Multiplicators: status+
50 C C C C C
B B C Multiplicators: persuasiveness+/status+
B B B
A D D D
40 D D D
A Control situation
A A A
30

20
Not
environmentally 10 20 30
responsible Runs
Intervention

Note: Individual graphs show the course of attitude change in three different experiments and the
control situation.

Figure 4.10 Experiments with multiplicators

To achieve appropriate use of a resource, it is important to lead persons to

orient themselves less to other people’s patterns of use and to pay more attention
to the state of the resource.
Both in general campaigns and in specific action campaigns using multiplicators,
it is advantageous to work with people who have high status. If attitudes within a
population tend to be environmentally irresponsible, there is little to be gained
through the use of persuasiveness or appealing to people’s sense of environmentally-
oriented self-responsibility.
A common denominator of the results can be formulated as follows, in analogy
to Latané’s (1981) theory of social impact, according to which social influence is a
function of the strength, immediacy, and number of sources.

• Strength: the concepts of susceptibility to influence, suggestibility, status, and

persuasiveness belong here.
• Immediacy: this involves the way that persons are networked, the visibility of
behaviour, and the degree to which a person may be influenced by others.
 Modelling environmental behaviour 93
• Number: while it is always advantageous to work with a large number of
pioneers, role models, and multiplicators, the financial means to do so may be
limited. In this case, the same effect can be achieved with a smaller number
of persons if measures are implemented to improve strength or immediacy.

In summary, find a certain number of people equipped with a certain degree of

strength and immediacy who will stimulate others to behave in more
environmentally responsible ways. Success in getting 3, 5, or even 10 percent of a
community to become active does, however, seem an impossible task. According
to the Swiss Environmental Survey (Diekmann and Franzen 1995), 16 percent of
people in Switzerland are members of environmental groups. This existing
potential, dispersely distributed, needs to be seen as a resource that can be activated
and brought together within the framework of concerted environmental efforts.
If efforts are insufficiently co-ordinated, they fall flat.
The conception of our experiments, with various forms of intervention as well
as the application of terms suited to translation into action, might create the impres-
sion that we underestimate the problems of implementing in reality the forms of
intervention studied. We are conscious of the fact that there are great gaps
(theoretical as well as empirical) and incalculabilities between ‘population
simulations’ based upon simulated individual behaviour and possible processes in
real populations. Many sub-components are missing in our model or are as yet not
adequately designed or validated. In spite of all this, from our perspective there is
no other alternative: we should indeed take on the challenge of examining the
complexity of such dynamic processes by means of appropriate instruments.
Simulation is an instrument well suited to enlarging our understanding of basic,
underlying social processes and developing further those explanatory approaches
that are almost exclusively static and based on one-person models. Against the
possible view that our approach is far from reality, it must be objected that the
forms of intervention we propose are founded upon empirical knowledge gained
in small-scale field experiments (see Dwyer et al. 1993; Mosler and Gutscher 1998).
In the next phase of research, we plan stage-by-stage testing of the simulation
model as a logical step in its proof in practice. Here, in the sense of a practice-
oriented ‘acid test’, we will strive towards process-oriented validation. To gain
well-founded understanding of the dynamics of existing social systems, data must
be collected in real social systems that are relevant to indications resulting from
the simulation and that can then be entered into the simulation itself. The focus of
the investigation would be the topic of dissemination of new behaviour patterns
throughout existing social networks. The optimal method would consist in large-
scale field experiments.
It is not our intention to propagate as ‘tested’ or ‘problem-free’ the application
of those forms of intervention judged effective on the basis of our simulations.
Rather, our purpose is to show environmentally conscious people, responsible
persons, and politicians how worthwhile it can be to expand the old triad of
‘traditional information campaigns’, ‘legal measures’, and ‘economic measures’
94 Hans-Joachim Mosler
by means of additional, novel strategies and to test these potentially successful
forms of intervention in practice.

References
Ajzen, I. (1991) ‘The theory of planned behaviour. Some unresolved issues’, Organizational
Behaviour and Human Decision Processes, 50: 179–211.
Ajzen, I. and Fishbein, M. (1977) ‘Attitude-behaviour relations: a theoretical analysis and
review of empirical research’, Psychological Bulletin, 84: 888–918.
Ajzen, I. and Madden, T.J. (1986) ‘Prediction of goal directed behaviour: attitudes,
intentions, and perceived behavioural control’, Journal of Experimental Social Psychology,
22: 453–74.
Bandura, A. (1979) Sozial-kognitive Lerntheorie, Stuttgart: Klett-Cotta.
Bandura, A. (1986) Social Foundations of Thought and Action, New York: Prentice-Hall.
Becker, G.S. (1976) The Economic Approach to Human Behavior, Chicago: University of Chicago
Press.
Diekmann, A. (1996) ‘Homo ökonomicus. Anwendungen und Probleme der Theorie
rationalen Handelns im Umweltbereich’, in A. Diekmann and C. Jaeger (eds)
Umweltsoziologie, Sonderband 36 der Kölner Zeitschrift für Soziologie und Sozialpsychologie: 89–
118.
Diekmann, A. and Franzen, A. (1995) Der Schweizer Umweltsurvey 1994, Codebuch, Bern:
Institut für Soziologie.
Dwyer, W.O., Leeming, F.C., Cobern, M.K., Porter, B.E. and Jackson, J.M. (1993) ‘Critical
review of behavioural interventions to preserve the environment. Research since 1980’,
Environment and Behaviour, 25(3): 275–321.
Festinger, L. (1954) ‘A theory of social comparison processes’, Human Relations, 7: 117–40.
Forrester, J.W. (1972) Grundzüge einer Systemtheorie, Wiesbaden: Gabler.
Frey, D., Dauenheimer, D., Parge, O. and Haisch, J. (1993a) ‘Die Theorie sozialer Ver-
gleichsprozesse’, in D. Frey and M. Irle (eds) Kognitive Theorien 2, Bern: Hans Huber:
81–122.
Frey, D., Stahlberg, D. and Gollwitzer, P.M. (1993b) ‘Einstellung und Verhalten: Die Theorie
des überlegten Handelns und die Theorie des geplanten Handelns’, in D. Frey and M.
Irle (eds) Kognitive Theorien 2, Bern: Hans Huber: 361–98.
Gonzales, M.H., Aronson, E. and Costanzo, M.A. (1988) ‘Using social cognition and
persuasion to promote energy conservation: a quasi-experiment’, Journal of Applied Social
Psychology, 18: 1049–66.
Harsanyi, J.C. (1977) Rational Behaviour and Bargaining Equilibrium in Games and Social
Situations, Cambridge: Cambridge University Press.
Jorgenson, D.O. and Papciak, A.S. (1981) ‘The effects of communication, resource feedback,
and identifiability on behaviour in a simulated commons’, Journal of Experimental Social
Psychology, 17: 373–85.
Klandermans, B. (1984) ‘Mobilization and participation: social-psychological expansions
of resource mobilization theory’, American Sociological Review, 49: 583–600.
Latané, B. (1981) ‘The psychology of social impact’, American Psychologist, 86: 343–56.
Latané, B. and Wolf, S. (1981) ‘The social impact of majorities and minorities’, Psychological
Review, 88: 438–53.
Liebrand, W.B.G. (1986) ‘The ubiquity of social values in social dilemmas’, in H.A.M.
Wilke, D.M. Messick and C.G. Rutte (eds) Experimental Social Dilemmas, Frankfurt am
Main: Verlag Peter Lang: 113–33.
 Modelling environmental behaviour 95
Liebrand, W.B.G., Wilke, H.A.M. and Wolters, F.J.M. (1986) ‘Value orientation and
conformity. A study using three types of social dilemma games’, Journal of Conflict
Resolution, 30(1): 77–97.
Mosler, H.-J. (1990) Selbstorganization von umweltgerechtem Handeln: der Einfluss von Vertrauensbildung
auf die Ressourcennutzung in einem Umweltspiel, Zürich: Zentralstelle der Studentenschaft.
Mosler, H.-J. (1993) ‘Self-dissemination of environmentally-responsible behaviour: the
influence of trust in a commons dilemma game’, Journal of Environmental Psychology, 13:
111–23.
Mosler, H.-J. (1995) ‘Umweltprobleme: eine sozialwissenschaftliche Perspektive mit
naturwissenschaftlichem Bezug’, in U. Fuhrer (ed.) Oekologisches Handeln als sozialer Prozess,
Basel: Birkhauser: 77–86.
Mosler, H.-J. (2000) Computersimulation sozialpsychologischer Theorien. Studien zur Veränderung von
Umwelteinstellung und Umweltverhalten, Weinheim: Psychologie Verlags Union.
Mosler, H.-J. and Gutscher, H. (1996) ‘Kooperation durch Selbstverpflichtung im Allmende-
Dilemma’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft Umweltsoziologie,
36: 308–23.
Mosler, H.-J. and Gutscher, H. (1998) ‘Umweltpsychologische Interventionsformen für
die Praxis’, Umweltpsychologie, 2: 64–79.
Mosler, H.-J. and Tobias, R. (2000) ‘Die Organization kollektiver Aktionen durch
Beeinflussung der individuellen Teilnahmeentscheidung. Eine Simulationsstudie’, Kölner
Zeitschrift für Soziologie und Sozialpsychologie, 52(2): 264–90.
Mosler, H.-J., Ammann, F. and Gutscher, H. (1998) ‘Simulation des Elaboration Likelihood
Model (ELM) als Mittel zur Entwicklung und Analyse von Umweltinterventionen’,
Zeitschrift für Sozialpsychologie, 29: 20–37.
Nowak, A. and Latané, B. (1994) ‘Simulating the emergence of social order from individual
behaviour’, in N. Gilbert and J. Doran (eds) Simulating Societies, London: UCL Press:
63–84.
Nowak, A., Szamrej, J. and Latané, B. (1990) ‘From private attitude to public opinion: a
dynamic theory of social impact’, Psychological Review, 97(3): 362–76.
Petty, R.E. and Cacioppo, J.T. (1986a) ‘The elaboration likelihood model of persuasion’,
in L. Berkowitz (ed.) Advances in Experimental Social Psychology, New York: Academic Press:
123–205.
Petty, R.E. and Cacioppo, J.T. (1986b) Communication and Persuasion: Central and
Peripheral Routes to Attitude Change, New York: Academic Press.
Spada, H. and Ernst, A.M. (1990) ‘Wissen, Ziele und Verhalten in einem ökologisch-
sozialen Dilemma’, Forschungsbericht, Freiburg: Psychologisches Institut der Universität
Freiburg.
Stahlberg, D. and Frey, D. (1993) ‘Das Elaboration-Likelihood-Modell von Petty und
Cacioppo’, in D. Frey and M. Irle (eds) Kognitive Theorien 2, Bern: Verlag Hans Huber:
327–60.
Stalder, J. (1985) ‘Die soziale Lerntheorie von Bandura’, in D. Frey and M. Irle (eds) Gruppen-
und Lerntheorien, Bern: Verlag Hans Huber: 241–72.
Tajfel, H. (1978) Differentiation Between Social Groups, London: Academic Press.
Tajfel, H. (1979) ‘Individual and groups in social psychology’, British Journal of Social and
Clinical Psychology, 18: 183–90.
Tanford, S. and Penrod, S. (1984) ‘Social influence model: a formal integration of research
on majority and minority influence process’, Psychological Bulletin, 95(2): 189–225.
Thibaut, J.W. and Kelley, H.H. (1959) The Social Psychology of Groups, New York: John Wiley.