Supplementary files are available for this work. For more information about accessing
these files, follow the link from the Table of Contents to "Reading the Supplementary Files”.
Table of Contents
Go Back
Modeling Generic Structures and Patterns
in Social Psychology
Ralph L. Levine, Ph.D.
Department of Resource Development
and Department of Psychology
314 Natural Resources Building
Michigan State University
East Lansing, MI 48824
Ph, 517-332-2317
Fax. 517-353-8994
leviner@pilot.msu.edu
James K. Doyle, Ph.D.
Department of Social Science & Policy Studies
Worcester Polytechnic Institute
100 Institute Road
Worcester, MA 01609
Ph. 508-831-5583
Fax. 508-831-5896
doyle@wpi.edu
ABSTRACT
System dynamics has been enriched from many disciplines. This paper describes our
effort to discover generic structures in the field of social psychology. Social psychologists
have accumulated a body of empirical studies and theories that are reproducible and
apply to a variety of social situations. Our task was to start modeling a few pivotal
dynamic effects found in the social psychology literature. We present models of the
dynamics of an important social process, namely, the “self-fulfilling prophecy.” The
structure underlying this process is associated with the drifting goals archetype. Next we
model the dynamic effects of contact between groups. Finally, we develop a set of models
that represent a key process in social psychology, namely the “fundamental attribution
error.” We hope that the approach to modeling generic structures in social psychology
will enrich future system dynamics models by including relevant biases and distortions in
perception discovered by social psychologists.
Key words: Social-psychology, social generic-structures, self-fulfilling prophecies,
attribution-errors, intergroup contact.
INTRODUCTION
System dynamicists emphasize looking for generic structures underlying problem
behaviors. Over the years, system dynamic modelers have discovered generic structures,
such as archetypes, which are associated with the areas of management (Senge, 1990),
organization theory (Sastry, 1998), and environmental problems (Meadows, 1989),
among others. The field of psychology may contribute to the thinking of system
dynamicists who might be confronted with modeling a problem that is dominated by a
significant social process. In the field of system dynamics most psychologists who have
made contributions to the discipline have contributed to the application of cognitive
psychology to the empirical and theoretical study of mental models (e.g., see Doyle,
1997; Doyle and Ford, 1998, 1999). Although there have been some discussions about
the relationship between psychology and the system dynamics discipline (e.g., Levine
1983, 2000), there has been little modeling of psychological processes to date, and
therefore there has not been much of an opportunity to discover new generic structures
that may not correspond to those already known from other fields.
This paper is a first attempt to discover, model, and describe those generic structures
underlying key processes in the discipline of social psychology. This work is also an
extension of previous efforts by Kao et al. (1998) to develop and use simple SD models
to teach social psychology. Indeed, there have been a few system dynamics models of
social psychological experiments, such as the Zimbardo prisoner experiment (Low, 1978)
and the Milgram shock experiments (Richmond, 1980). In the Milgram obedience
experiment (Milgram, 1963, 1965), the subjects, who thought they were part of a study to
see the relationship between punishment and learning, were instructed to shock the
“learner,” when learner made a mistake. Although the victim looked like they were in
extreme pain when shocked, actually the victim or learner was part of the experimental
team and received no shocks at all. Moreover, the experimental subject had to
progressively increase the shock voltage (ranging from 15 volts to 450 volts) if the
learner continued to make mistakes. If the subject hesitated, the experimenter would
pressure the subject by insisting that he or she continue in the experiment. The majority
of the subjects (65%) continued to increase the voltage throughout the entire experiment.
The subjects wanted to please the experimenter, who represented authority, and at the
same time, they were disturbed because they thought they were hurting an innocent
person. This conflict was represented in the model as two loops that dealt (1) with giving
in to the pressure of the experimenter and (2) feeling the guilt of harming an innocent
stranger.
The goal of the present paper is to describe a few common structures we have discovered
that might give social psychologists a deeper theoretical understanding of well-
established empirical findings. In addition, by starting to model social processes in
depth, we hope that some of the work in social psychology might help to augment the set
of tools used by system dynamicists.
Let us begin by defining social psychology. Social psychology is the scientific field that
attempts to understand the nature and causes of individual behavior and thought in social
situations (Baron and Byrne, 2000). From a system dynamics perspective, many of the
problems in organizations and in other domains involve emotional and cognitive
components. System dynamicists have resisted many of the assumptions of pure rational
choice theory that assumes having total informational resources in setting policies and
making decisions. Many of the empirical findings of social psychology would concur
with the system dynamicist’s rejection of those assumptions. There have been many
studies which show that people in general distort, discount, and filter out information,
especially if it does not conform to preconceived beliefs (see Meyers, 2000). Also, social
psychologists have found numerous problems in human selective memory, judgment, and
decision-making ability that could lead to problem behaviors. For example, Holmberg
and Holmes (1994) found in a repeated measurement study of newlywed couples that
most initially reported being happy. A follow-up study conducted on the same couples
two years later asked them how they felt currently about their marriage and asked them to
recall how they felt about their marriage as newlyweds. The results showed that those
whose marriages had disintegrated claimed that things had not gone well from the
beginning. Indeed, Holmberg and Holmes say, “Such biases can lead to a dangerous
downward spiral. The worse your current view of your partner is, the worse your
memories are, which only further confirms your negative attitudes.”
This type of “revisional memory” has been found in a number of studies (e.g.,
McFarland and Ross, 1985). Such a process might be very important to include in a
model. For example, currently there is interest in understanding the dynamics of ethnic
conflict. In modeling the escalation of anger and violent actions, one potential negative
loop that could decrease the escalation process might be related to the fact that in the
past, the two groups had gotten along very well, perhaps cooperated with each other in
dealing with crucial problems, or perhaps fought together on the same side to defeat a
common enemy. The “old comrade/friendship loop” might not mitigate the escalation
structures if people on both sides revise the past to match present conditions. Including a
revisional memory component process in a model of ethnic conflict would be realistic
and insightful.
Before going on, it is interesting to comment that the revisional memory process is
somewhat the opposite of a generic structure that is known and used by system
dynamicists. The revisional memory structure is associated with a weighting of recent
affairs as being more important than past affairs. The sticky/slippery perception process
(see Richmond, 2001) is a process where the past is weighted more than present
conditions. In the Zimbardo, Stanford Prisoner Experiment, students participated in a
role playing simulation, in which they were either assigned to play the role of a prisoner
or a prison guard. This was an experiment that was scheduled to be run over a continuous
series of days, but had to be terminated prematurely for ethical reasons. Initially, the
“guards” were very aggressive and hostile towards the “prisoners.” The prisoners reacted
to the guard hostility by generating a high level of resistance to the guards. At a later
point in time, the guards become much less hostile to the prisoners, yet the prisoners
remained highly resistant, even when conditions changed. The past dominated the
current situation. In the revisional memory case, there is an active process of substituting
current beliefs and attitudes for the past.
SOCIAL PSYCHOLOGY REVISITED
The social psychological perspective, as we have indicated, is focused on actual
human abilities and processes in the social context. For forty or fifty years, social
psychology as a field has accumulated a wealth of empirical research. Much of the
studies have been true experiments. A number of theories have emerged in the field,
such as attribution theory (e.g. see Kelley, 1972, Kelly and Michela, 1980), which
aids in understanding the reasons behind others' behavior, and social learning theory,
which helps us to understand how we acquire new information, behaviors, or attitudes
from other people. Yet, these theories are usually stated only in verbal form. We feel that
the time has come for going to the next step, modeling the dynamics of those processes
that are well established in the literature.
Limitations
Thus far, we have been very positive about the potential for modeling social
psychological processes, to enhance both the field of social psychology and the discipline
of system dynamics. However, we have encountered several problems and constraints in
our efforts to find and model generic structures in this area. First, many of the concepts
presented in texts and in the literature are couched in static terms. For example, social
psychologists study what is known as the “availability heuristic”, which is a rule of
thumb that assesses the likelihood of a phenomenon in terms of the ease with which
examples can be recalled in memory. News and information about shark attacks prime us
to think that shark attacks are more likely than actual base rate statistics. Normally,
textbooks on social psychology will present the availability heuristic as a bias or
constraint on thinking rather than putting it in a dynamic framework. There is little or no
thought given to the possibility that one can change the tendency to over- or
underestimate the risk involved in a particular action or situation over time.
From a system dynamics view, many of these static cognitive biases can be useful in
modeling dynamic problems. They make the model more realistic by helping to specify
human biases in perceptions, judgments, and reactions. As a minimum, these static
biases can be included in SD models as either initial values of the levels or as parameters
that describe individual differences among individuals or groups. On the other hand,
these cognitive biases can be integrated into dynamic models as constraints or as
parameters. For example, Riley (2001), working with wildlife managers, modeled the
interface between the dynamics of the bear population in New York State and the
response of landowners who may potentially come in contact with those bears. Incidents
between bears and people are in fact quite rare, but every time there was an incident, as
reported by the media, landowners in particular would put pressure on governmental land
managers to either move the bears elsewhere or to get rid of them by increasing the
number of hunting licenses. The farmers and other landowners perceived the likelihood
of an encounter with a bear as being much higher than the actual risk involved, i.e., they
unconsciously used the availability heuristic. However, in this case, Riley used this to
capture the reason for the pressure on the managers. It becomes part of the dynamics of
the management problem. Finally, we might add that, in designing policies to solve the
management problem, we might want to change the perceptions of the likelihood of
human/bear encounters. Design policies that might lessen the effect of the availability
heuristic move us to a dynamic perspective.
Experimental Rigor versus External Validity
Social psychologists have performed numerous studies using rigorous experimental
methods. From a causal point of view, there are advantages to manipulating specific
variables and keeping many other processes constant in an experimental setup. On the
other hand, there is the problem of generalization, which in this context is called
“external validity.” In the real world, any particular problem is embedded in many
processes. The advantage of the system dynamics simulation methodology is that a
model can represent a number of loop processes simultaneously, which is closer to what
happens in the real world. On the other hand, social psychological experiments go in the
direction of isolating processes, and perhaps that is the reason that the field of social
psychology has stayed primarily in the academic setting, rarely venturing out into the real
world.
ADVANTAGES OF MODELING BASIC SOCIAL PROCESSES
Social Generic Structures: Augmenting the System Dynamicist’s Toolbox
We believe that potentially most of the findings about cognitive biases and social
processes studied in isolation may be externally valid. However, in any real situation,
those social structures may not dominate the dynamics. We are suggesting that, if we
were to model a social process in isolation, we might consider the model and its loop
structure as a generic template, or "molecule," to be embellished and integrated into a
larger, more realistic model of a dynamic problem. What, then, does the system
dynamics community gain from focusing on structures found in social psychology? In
general, it would supplement our knowledge of “classical” archetypes (Senge, 1990, Kim
and Anderson, 1998) and other common patterns of behavior and structure in
organizational settings (Sastry, 1998). Again, as we have indicated, social psychology
can contribute to our thinking about the influence of emotional, nonrational reactions,
decisions, and policies, as well as capture some of the cognitive biases that may play a
leading role in dynamics of the problem.
EXAMPLES OF SOCIAL PSYCHOLOGICAL MOLECULES
When we first started this project, we initially believed that we were looking for social
archetypes. However, we have come to the conclusion that it may not be appropriate to
use the term “ archetype” for the social processes under study. First, not all of these
processes focus on a problem. Second, we would subscribe to Lane and Smart’s (1996)
suggestion that one consider archetypes as a special case of a generic structure in which
the solution to the problem frequently is associated with counterintuitive actions. We feel
comfortable in calling the patterns associated with social phenomena “social molecules.”
They capture the dynamics of small bits of social processes, gleamed from replications of
numerous social psychological studies. In this section we will present three of the models
we have created up to date. From a cybernetic perspective, some of the models are more
interesting than others. All of them seemed to qualitatively match what patterns of
behavior existed in the literature. Currently, we are submitting them to more rigorous
tests, as suggested by Forrester and Senge (1980).
We deliberately tried to formulate extremely simple models. One thing we should say
about modeling social phenomena is we have found even small models to be informative,
insightful, and interesting in themselves. We would concur with Richardson (2000) who
feels very strongly that much can be learned from initially working with small,
manageable, and insightful models. It is also noteworthy that each of these models can be
extended and embellished quite easily, almost too easily. At this stage, we want to
remain simple.
Self-fulfilling Prophecy
The first model deals with a well-established social process, namely, the self-fulfilling
prophecy. To provide a background on the self-fulfilling prophecy, we will begin with
the notion of a schema. Just as system dynamicists use the idea of mental model,
cognitive social psychologists frequently work with a somewhat similar construct, the
schema. Briefly, a schema is a mental framework that centers around a specific theme
that helps people to organize social information. Schemas, once formed, may exert
powerful influences on what the person notices, remembers, and how he or she may
relate to others. In a social context, a person’s schemas can affect the behavior of others.
One possible effect of a person’s schema is the self-fulfilling prophecy, in which social
perceptions, even when wrong, can strongly influence the behavior of people who are
targets of such perceptions. To give an example, Rosenthal and Jacobson (1968)
examined the possibility that teachers' schemas about minority students in the San
Francisco area were causing them to treat minority students differently with respect to
time and effort (and in other ways) than majority students. As a result of this differential
treatment, minority students might fall behind academically.
As system dynamicists, we recognize that in this school context, the effects of teachers’
schemas provide the motivation for a familiar pattern seen in other contexts, namely, the
success to the successful archetype. However, we are going to focus on the perceptions
and goals of a teacher only with respect to one population, namely, in the context of
minority students. What we plan to do is to describe two models that capture the
essential idea of the self-fulfilling prophecy. Although the models are in the context of a
school setting, the concept of self-fulfilling prophecy goes back into the history of social
science. For example, in economics, during the depression in the United States in the
1930s, people circulated rumors that the banks would fail. This generated major runs on
the banks, which in turn caused many sound banks to fail. Also, there is an active interest
in studying the role of self-fulfilling prophecies in macroeconomics (see, e.g., Farmer,
1999).
To continue putting this social phenomenon in a school context, let’s look at the
Rosenthal and Jacobson field experiment in more detail. The team of researchers went to
a school in the San Francisco area and administered IQ tests to the entire student body.
Then they told the teachers that some of the students (picked randomly) in his or her class
were brilliant, ie., had scored high on the IQ test, and should be blossoming in the near
future. Nothing was said about such high scores in the control group. Eight months later,
the IQ tests were administered again for all students. The “bloomer group” showed
greater gains than those who were not designated as bloomers. Further analysis indicated
that the teachers had spent more time on their bloomer students than the rest of the class,
which is the mechanism accounting for the gains in performance. In other contexts,
where studies had been done on lower income schools, lower teacher expectations for
minority students may decrease students’ confidence in themselves, which in turn
generates lower grades for those students.
Modeling the Self-fulfilling Prophecy
From the description of the social phenomenon, basically teachers may come in with
some previous expectations about their students’ ability to perform. In the first model,
we are assuming that the teacher’s expectations are fixed, and indeed some social
psychologists talk about a “perseverance effect” that makes it difficult to change one’s
schema even in the face of contradictions and information contrary to one’s beliefs.
Student
J Performance
Teacher Expectation
Time
Fig. 1. Potential reference mode describing the problem
of the self-fulfilling prophesy notion applied to teaching
Model 1: Behavioral reference mode for self-fulfilling prophecy. It was difficult to
find explicit time series data to work with in this case. The Rosenthal and Jacobson study
just measured performance at two time points. One contribution system dynamics may
make is to provide a theoretical rationale to measure key variables over a number of time
points, if possible. This would be expensive and difficult, but it certainly would be worth
the costs in time and money. In any event, if initially, the teacher’s expected
performance were low, and students were initially above the teacher’s expected
performance, the students would merge towards the low value over time, perhaps
displaying a logarithmic pattern. This reference mode is drawn in Figure 1.
Structural characteristics of the self-fulfilling prophecy, Model _1. From the
description of the phenomenon and from our best guess of the temporal behavior, our
first dynamic hypothesis is as simple as one can get, namely, a single negative loop
process accounts for the self-fulfilling prophecy. The dynamic hypothesis and the stock
and flow diagram of the model can be seen in Figure 2. For system dynamicists this is
not very insightful or particularly interesting. However, it does represent a simple model
for some very important social problem behavior.
Teacher Expectation
\C
Performance
a Student
ee
Change in
Student
Performance
Teacher expectation
performance gap
change ja student perform
8 P Student Performance
time to change perform
Fig. 2. The loop structure and the flow diagram of a simple
model of the dynamics of the self-fulfilling prophesy.
Output from Model 1.
performance was set initially to be above the level of the teacher expectations.
gradually lowered their performance to meet the teacher’s expectation level.
The output of the model is found in Figure 3. In this run student
Students
B 1: Teacher expectation 2: Student Performance
y 100
ZZ 2:Student Performance
a 1:Teacher Expectation
T T T 1
0.00 3.00 6.00 9.00 12.00
Page 1 Time 4:56 PM Thu, Oct 25, 2007
dev ?
‘Sample Run: Setf-fulfiling Prophesy
Fig. 3. Output of the simple model of the dynamics of self-fulfilling
Prophesy.
Model II: Introduction. As we have indicated there are many static processes that
might be useful to know when building traditional system dynamics models. In the
situation where the self-fulfilling prophecy is relevant, such as in the context of an
educational setting, it might be reasonable to assume that the teacher’s expectation is not
immutable, even though the perseverance effect might be operating in varying degrees,
depending on the teacher. We assumed that the teacher watches student performance and
is influenced by it over time. This leads to what we call the “reciprocal self-fulfilling
prophecy" model, where the student is influenced by the manifestation of the teacher’s
expectations and the teacher, in turn, is influenced by the student's level of performance.
Model 2: Behavioral reference mode for reciprocal self-fulfilling prophecy. The
double influence model of the self-fulfilling prophecy phenomenon may generate a
similar logarithmic response by the teacher to the performance of the students. We have
not found time series data that gives a detailed description of the time trajectory in the
literature. Clauset and Gaynor (1992) discuss the fact that some teachers may pay much
more attention to their students’s performance than external standards.
Student
/ Performance
Teacher Expectations
Time
Fig. 4. Possible behavior-over-time-graph when
teacher responds to the performance of his or her students.
Figure 4 shows our hypothesized behavior-over-time graph for this situation. We would
guess that students begin to lower their performance in response to the “negative” signals
given by the teacher. On the other hand, the teacher notes that the students are doing
better than she or he expected and begins to raise performance expectations a bit over
time.
Hypothesized Loop structure, Model 2. We have assumed that the response of the
teacher is generally the same as the response of the students. Thus, in Figure 5, we have
added a second negative loop, loop #2, to the structural map. In addition, note that when
the teacher begins to follow performance, a new, positive loop, loop #3, comes into
existence.
Gap in
Ko Expectations
Change In #2
Teacher
oe
NU, ‘eacher Expectations 6
Performance
Gap
on
+ Student
Performance
Change in oe
Student
Performance
Fig. 5. Loop structure for the reciprocal self-fulfilling prophesy
Model.
Have we seen this generic structure before in the system dynamics literature? We can
find this structure featured in two separate sources. First, substantively, this structure
played a dominant role in modeling classroom dynamics (Clauset and Gaynor, 1980,
1982, 1992; Levine, Van Sell, and Rubin, 1992; Roberts, 1978). In some actual
situations, in low income schools, teachers appear to use student performance as the basis
of working with the students to decrease the performance gap instead of using some
external set of standards for all their students.
A second place in the SD literature where one can find this loop structure is in the
description of existing archetypes, namely, drifting goals. One might want to consider
this as an example of the drifting goals, although there is no explicit, extra, delay between
the performance gap and the change in performance variable. Also, in this present
version, when we assume that the teacher is paying attention directly to student
performance, a third, positive loop, which has never been mentioned or associated with
the drifting goal archetype, is introduced.
gap in expectation!
x Teacher
change it/teacher expect
gap in performance
perseverance effect
change in Student perform
time to change perform
Fig. 6. The flow diagram for the reciprocal
self-fulfilling Prophesy.
Stock and Flow View for Model 2. It might be instructive to see the stock and flow
chart found in Figure 6. We note that the perseverance effect was explicitly represented
as an individual difference parameter, the perseverance effect. Time is absorbed into this
parameter, which has units of fraction per month. Thus, we recognize the potential for
this static social psychological effect by dealing with it as an exogenous parameter
r)
|: Teacher Expectations 2: Student Performance
100q—~
a
Ww | Student Performance
ia Teacher Expectations
4
A
eS Base Run: Reciprocal Self-futfilling Prophesy
oa/ 2
Fig.7. Results of the baseline simulation run for
the reciprocal self-fulfilling prophesy model.
Figure 7 shows the output of the simulation. In this run, the students’ initial performance
level is above the teacher’s expectations. This leads to a drop in performance, but
because the teacher is influenced by performance, his or her expectations increase to
match the performance. Eventually they come into equilibrium at a much lower student
performance level than would be the case if the teacher had started with expectations set
at a high level. Technically we note that, because we failed to put an extra delay in loop
#1, the model is incapable of displaying oscillatory behavior, unlike Dowling and
MacDonald’s (1995) version of drifting goals. Thus, there are a few differences in the
structural characteristics of both situations. However, in general, the drifting goals
archetype does fit the situation where both the students and the teacher are influenced by
each other’s behavior.
Student
Performance
j
SS
Teacher Expectation
Time
(a)
Student
Le
\
Teacher Expectation
Time
(b)
Fig. 8 Alternative patterns of behavior for drifting goals.
Other Patterns.
In concluding this discussion of the two models that deal with the social psychological
process of the self-fulfilling prophecy, we should note that the literature on eroding goals,
when discussing behavioral patterns that are associated with this archetype, sometimes
shows a different pattern from that found in Figure 7. In particular, Lane and Smart
(1996) and Kim and Anderson (1998) show other patterns. Two alternative behavior
patterns, as seen in Figure 8, shows the expectation and student performance going either
up or down in tandem. This is a very different pattern than was generated by the present
reciprocal model (see Figure 7).
This certainly is an intriguing idea. It might look like the pattern depicted in Figure 8b,
where the two curves move with each other in the positive direction. However, we feel
that the structure that underlies drifting goals, namely, two negative loops, cannot directly
generate the patterns found in Figure 8. Those patterns also do not represent the notion
of the self-fulfilling prophecy as well. Nevertheless, it is interesting and plausible
behavior. The model might have to be modified to take into consideration the direction
of the first derivative. For example, we might hypothesize that the students are also
paying attention a positive change in the teacher’s expectations. When they see that the
teacher expects more out of them, they may respond by raising their performance,
perhaps by raising their own internal goals. A model that incorporates (1) derivative
control and (2) a perceptual delay in monitoring the speed and direction of the teacher’s
expectations might lead to fruitful insights in the future. However, again the loop
structure will deviate from the simple structures associated with the drifting goals
archetype.
Models Describing the Extended Contact Hypothesis
A second example of modeling basic and fundamental social psychology processes deals
with the area of intergroup prejudice and conflict that might possibly be ameliorated by
providing settings where the two groups can come in contact for periods of time, perhaps
working together on a common task. This body of literature has policy implications in
urban planning, for example, where subdivisions can be designed to include homes of
varying price. Social psychologists asked the question whether contact itself could lower
intergroup prejudice. There is some evidence that prejudice will decrease over time as a
function of contact (Schwartzwald, Amir, & Crain, 1992). One would imagine that there
are many other processes, however, that may come in to nullify the positive effects of
contact. Recently, social psychologists have become somewhat pessimistic about the
reliability of these results. However, they have not completely dropped the idea that
contact plays a role in decreasing prejudice. Currently, they are accumulating a large
amount of evidence to show that if contact leads to friendship between members of two
groups, then members who have made friends with others have a much more positive
feeling about the other group. Moreover, even knowing that someone in your group has
made friends with members of the other group can lower your degree of prejudice for the
other group, even without contact. This is called the “extended contact hypothesis”
(Pettigrew, 1997; Wright et. al., 1997).
Given that the current literature emphasizes the development of friendships between
groups, we have designated the Degree Of Friendship as a principal stock in a model of
the contact hypothesis. The first question to ask is whether changes in the Degree Of
Friendship are caused by the total, cumulative time spent with the other group or is it the
actual process of being in contact with the other group that increases the Degree Of
Friendship. The literature was not explicit on this distinction, which is very important
for system dynamicists. In the model we decided that the process of contact itself
increased degree of friendship, not the cumulative time.
Model 1: Simple Extended Contact Model
sin ta lose Nears friendship Amnits per hour
Degree Of Friendship
&
decline in friendship,
baseline spending time
with target group
ef friendship on
spending time together
" spending time with target g
Cum Time Spent with Target Group SPeRemnE Hime With target group
Fig. 9. The stock and flow diagram describing a simple
representation of the extended contact hypothesis.
We shall start our description of the simple model of contact by viewing the stock and
flow diagram found in Figure 9. First note that we have represented the mechanism of
contact as a co-flow process. In addition, we hypothesize that as the Degree Of
Friendship increases over time, people will want to spend more time with each other,
hence the development and operation of a positive loop structure.
B vege o1rensinp
Degree of Friendship
8
ase ~ _ eats {038 AM Sat tn 1.2002
Nae 2 ves
Fig. 10. The behavior of the simple extended contact hypothesis
model over time.
The base output of this simple contact model is shown in Figure 10. Initially, the Degree
Of Friendship shows a logarithmic pattern over time, until the positive loop kicks in.
Then one observes an s-shaped curve for the rest of the time horizon, which is two years.
Note also that, using this structure, the system can go into equilibrium.
Contact Burnout
ftiendship units per hour
time to lose king
Degree OF Friendship
of target persons
making Wionds
decine in tepaip #1
basaine spending
Se with target group
enending time wih age! grou
ef friendship on ‘Cum Time Spent with Target Group a 7 ania
spending time together
ef of Bumaut on spending
time with other group
on
ntact Burnout
recuperating from burnout
"I
time to dissapate burnout
burnout hour
Fig. 11. The stock and flow structure of the effects
of burnout on contact behavior: The fatigue variation.
This simple extended contact model displays rather complex behavior over time.
However, one may find that “familiarization breeds contempt.” Too much contact may
inhibit the Degree of Friendship. We call this the “burnout effect”, and have introduced
another psychological stock, namely, a variable called “Contact Burnout,” to capture the
debilitating effects of contact over time. The main question we asked at this point was
whether burnout directly affects contact behavior or does it directly affect the Degree Of
Friendship. If contact burnout affects contact behavior, then we would consider burnout
as inducing response fatigue. On the other hand, if burnout affects Degree of Friendship,
then the decline in friendship is probably due to the aversive aspects of constantly being
with the other group. Figure 11 shows the stock and flow diagram of the fatigue model,
where burnout builds up and then inhibits the rate variable, making friends. Thus we
have augmented the original simple model of the extended contact hypothesis to include
a negative loop process that provides a mechanism for limiting the Degree of Friendship.
We also note that, when creating the table function for the effect of burnout on spending
time with the other group, we assumed that at first the degree of Contact Burnout had
little or no effect on spending time with the other group. However, as the burnout
variable became fairly high, the fatigue effect sharply cut down on spending time with the
other group.
99 +: Degree OF Friendship 2; Contact Burnout
] 10%
cee
q sal
e| of
0.00 25.00 52.00 7a.00 104,00
Page 1 Weeks 9:54AM Fr, Dect
. Untitled
\ger ?
Fig. 12. The results of a base run of the fatigue
model, whereContact Burnout directly affects contact behavior.
The results of a base simulation run indicate that the negative, fatigue loop had a very
strong effect on inhibiting the growth of the Degree of Friendship. Note also that the
Degree of Friendship does show a little bit of overshoot and then comes to a new
equilibrium point, at least with the parameter values used in this base run. In general
there is the old pattern of initial logarithmic behavior, followed by a period of
exponential growth that seems to be inhibited by the buildup of Contact Burnout.
As we discussed above, one of the major problems of attempting to find and model
generic structures found in social psychology is that time series data may not exist.
Change is usually measured at only two points in time. Thus, Contact Burnout might not
directly affect the Degree of Friendship through the input side, through a fatigue effect,
but because burnout itself is aversive. There are data to show that moods affect attitudes
and being in a bad mood may increase the Joss in friendship over time as Contact
Burnout builds up
friendship units per hour
time to lose flgndship
getting to lik§ target group
Degree Of Friendship
faseline spending
time with target group
spgnding time with target group
Cum Time Spent withTarget Group
fof friendship on
spending ime together
‘ef of burnout on losing
{fiendship: Bad Mood
Contact Burnout
recuperating from burnout
a
time to dissapate bumout
Bumout units per hour
Fig. 13. The stock and flow structure of the effects of
burnout on decreasing the Degree of Friendship: The
mood variation
The stock and flow diagram of the extended contact hypothesis mood model is shown in
Figure 13. Loop #3 inhibits the Degree of Friendship variable. This is a different
mechanism from behavioral fatigue, associated with the previous model. The table
function describing the effect of burnout on losing friendship was written in this base
case to have no effect until mid level values of the stock. Then the effect of mood kicked
in sharply. The maximum value of the multiplier was set around 3.0.
B® + degree OF Friendship 2: Contact Bumout
4y 100
1
4 50. . ,
| ; :
7
1
q °
52.00 78.00 404.00
Page 1 Weeks 7:01PM Sat, Jan 19, 2002
XN aeaFr 7? Untitled
Fig. 14. Results of the base run for the contact mood
Model.
Figure 14 represents the output of the model for the initial run. The rapid exponential
increase is somewhat startling, compared to the previous fatigue model’s trajectories.
Note also the behavioral pattern of overshooting, collapsing, and partial recovering to a
new asymptote. We have noted this behavior mode in several coflow models in the past
(Levine & Nguyen, 2000). We would suggest future work in analytically deriving these
oscillations.
Alternative parameter values. In subsequent sensitivity runs, we found that if the
maximum value of the ef of burnout on losing friendship were set to be relatively low,
say 1.6, then the trajectories look very much like the original base model in which we did
not assume a burnout process operating.
Combining the two burnout effects. We have uncovered a large degree of uncertainty
concerning whether fatigue or mood provides the fundamental mechanism for Contact
Burnout. Of course it may be a combination of the two processes that operates here. It
certain is plausible that both mechanisms are working simultaneously. We therefore
extended the model to allow both processes to affect behavior. This variation of the
model will be called the “combination model.” It should also be pointed out that in the
first case, Burnout had a direct impact on contact, per se. On the other hand, the second
process had an effect on essentially the quality of friendship.
ftiendship units per hour
time to lose friendship
xO
x
aid
eel omremee
earn
norte e
_—_n EES
ot indip on =
spending time together \=
{ef bumout on loss of
friendship: Bad Mood ef of burnout on spending
- time with target people
° we, _
time to dissapate bumout
Burnout units per hour
Fig. 15. The stock and flow structure of the effects of burnout
on contact behavior: The combination variation.
Fig. 16. Base run for model combining effects 0
fatigue and mood
The stock and flow diagram for the combination model of the extended contact
hypothesis is presented in Figure 15, This time we have included the simultaneous
operation of loops 1, 3, and 3. The output of the combined fatigue and mood model is
shown in Figure 16. Given this parameterization, one can see that the growth, overshoot,
and recovery to a new asymptote were sharper and quicker to happen over time.
Modeling the Fundamental Attribution Error
Introduction
Another pivotal process in social psychology deals with how people interpret the actions
of others. The fundamental attribution error is quite interesting. The tendency to explain
other’s behavior patterns as a stable character trait even when objectively, one should be
able to see that the behavior was caused by circumstances, is called the fundamental
attribution error (Ross, 1977). In this situation, we have the observer and the person
being observed, the target person. It is very true that sometimes people have internal,
psychological characteristics that may affect their behavior. In the real world, the
fundamental attribution error often has social consequences, like thinking that all
alcoholics lack will power or that lower income people on welfare are all lazy. Those are
internal, dispositional explanations of perceived behavior patterns. The community
psychologist, William Ryan (1971, 1981, 1994) coined the term “blaming the victim” to
describe the manifestation of the fundamental attribution error. In general, the
fundamental attribution error is a bias towards explaining behavior patterns as caused by
the biological, personal and moral attributes of the individual, rather than explaining
behavior from more external sources, such as being exposed to violence, being
discriminated against, and going to substandard schools. Ryan described the process
whereby people who are victimized by the political and social system are blamed for their
problems. Their personal characteristics are said to be the cause of their getting into
trouble and not being able to cope adequately with their problems.
Repenning and Sterman (2000) recently utilized a form of the fundamental attribution
error, which they called the “self-confirming attribution error” in a model that represents
the dynamics of process improvement. Managers may attribute low throughput to
inadequate worker effort, rather than understanding that low throughput may be due to
deeper and less observable process problems. This leads to higher levels of production
pressure and less emphasis and resources devoted to process improvement.
Research also finds that, when the observer makes errors or does not perform well, then
usually the person will attribute poor performance to outside causes and conditions
(McGuire and McGuire, 1986; White and Younger, 1988; Fiedler et al., 1991). In many
respects, system dynamicists deal with structures that determine behavior. We call this an
“internal approach” to causation, because we attribute the behavior observed to the
internal loop structure underlying the system problem. We have to be a little careful here
to point out that the term external and internal, as used by psychologists, may be used
differently. Actually, in terms of explanation, SD’s internal explanation corresponds
most of the time to social psychologists external explanation. Frequently this means that
almost anyone in the system may act in the same manner, given the constraints imposed
by the cybernetic loop structure. We also would like to point out the relationship
between the self-fulfilling prophecy example and this process. The teacher may be
attributing poor performance to laziness or stupidity of his or her students.
Introducing a dynamic perspective
In most cases, what we have found is that social psychologists treat this very powerful
bias in a static manner. However, the degree of internal explanation can change over
time. So, in developing a model of the dynamics of the fundamental attribution error, we
are going to look at the role of consistency in changing the internal explanation of the
observed behavior patterns.
Model variables
Let us start with the individual being observed. It may very well be that he or she does
have some dispositional characteristics that explain behavior patterns in one situation or
another. An example of such an internal, psychological disposition might be shyness.
We can envision a variable that represents the degree that the persons’ disposition affects
their behavior patterns. This individual difference variable has theoretically been set to
range from 0.0 to 100. It is called the 'target’s actual strength of internal attribution.”
Low values imply that external causes can explain observed behavior patterns. High
values imply that the target person’s poor performance is low because of internal causes,
such as low IQ or laziness, etc. A value of 50 means that about half the time the causes
were internal and the other half external.
One of the main stocks of this model is the “Perceived Internal Strength of Target by
Observer”. This is the perceptual counterpart of the target's actual internal strength
value. A second key process deals with the consistency of the target’s behavior over time
and situations. We assume that targets vary in their degree of consistency. Presumably,
if internal strength is high, then the person will consistently show the characteristic
behavior pattern in one setting after another, i.e., they will be very consistent. Low
consistency indicates that outside circumstances and the system characteristics explain
why performance, for example, can be good one time and fall down at other times.
Figure 17 shows the stock and flow representation of the model. A basic assumption is
that most observers assume that the target’s internal strength is very high, the
fundamental attribution error, and that, if this is true, consistency should be high.
Actually, the target person may have a low internal strength value and will eventual show
that he or she varies considerably over time with respect to behavior.
Let us look at the structural characteristics of this model in a bit more detail. Actually,
the model, from a system dynamics perspective, is not too interesting. There are only
two negative loops that are coupled together by a single connection through the variable
time to change perceived internal strength. This table functions assumes that, when the
target person is very consistent over time, there is very little evidence that the target’s
behavior is caused by circumstances, rather it is evidence for some permanent, highly
immutable character trait that explains his or her behavior.
On the other hand, when
Perceived Consistency somehow gets low, then the adjustment time should be short.
Thus, under these conditions, change in perceived internal strength occurs faster.
time to/perceive consistency Perceived Consistency
en Tas
change in perceived
gap inonsistency
targets consistency level
i to change perceived
internal strength
Perceived Internal Strength
of Target by Observer
changing flerceived
internal strength
target actual strength
of internal attribution
gap in perceiving
internal attribution
Fig. 17. The stock and flow diagram associated
with modeling the dynamics of the fundamental
attribution error.
KN.
~—— == =
Fig. 18. The output of a model of the fundamental
attribution Error.
4 —t
2
3
4
00 2,00 50.00 73100
Page 1 Weeks: ‘845 PM Sat Jan 19,2002
ae 2 Unie
The results of a simulation run are displayed in Figure 18. Initially we set the target’s
actual strength of internal attribution and the target’s consistency level at a low value of
15 each. Thus the target actually was not consistent and mainly made mistakes and
performed poorly primarily due to external causes. Figure 18 shows that at first, since
Perceived Consistency is high there was not much change in perceived internal strength,
until the observer changed his or her estimate of the target’s consistency level. As the
negative, stock adjustment process associated with Perceived Consistency move to
realistic values, the Observer began to rapidly change his or her perceptions of the
target’s actual internal strength, which in this run, was very low, 15.
CONCLUSIONS AND COMMENTS ON SOCIAL PSYCHOLOGY
Advantages to Describing Common General Structures for the Field of Social
Psychology
Thus far, we have been focusing on how these social generic structures might be
integrated into the work of system dynamicists. On the other hand, it is always nice to
see where system dynamics might prove extremely useful in bringing understanding and
insight into a traditional social science discipline, like social psychology. Indeed, we
think that a system dynamics approach to modeling empirically well-established social
processes has some added value for the field of social psychology. First, it can lay things
out so that the assumptions are clear and can be examined and assessed in more detail.
Indeed, as we began to model these isolated social phenomena, we noted that we had to
ask some hard questions about assumptions, which were not made explicit in the
literature. Models help to raise questions about assumptions that would not have be
asked without the modeling process. Along with making assumptions clearer, we suggest
that SD modeling will make verbal theories more coherent and consistent. In addition, as
we have indicated, traditionally, social psychologists focus on reliable static processes
and sometimes fail to look into changes over time. Since the emphasis is on experimental
manipulation, very little progress has been made to introduce cybernetic thinking about
those social phenomena. System dynamics adds a cybernetic dimension to the study of
those psychological processes.
Linkages. One of our observations about this field, as represented in textbooks on the
topic, is that researchers specialize so much in their research that in general there is not
much effort devoted to noting the linkages among research areas. Indeed, unlike physics
and perhaps microeconomics, one could easily open a social psychology textbook to any
chapter and randomly select further chapters without any difficulty in mastering the
material. A system dynamic approach to finding common social molecules may tie up
loosely fitting social phenomena through discovering those linkages. Accumulating these
bits of social processes should be helpful in dealing with complexities of the real world
by expanding and combining social molecules. Understanding attitude formation and
change may be very useful in understanding conflict, aggression, and impression
formation.
Further empirical research. Although we are looking for well established and
thoroughly researched social phenomena, we feel that using system dynamics to model
social processes can lead to identifying the need for further empirical research and indeed
might provide a framework for future studies. In several cases we have found that, after
developing a model of the social process, and running the model under conditions that
match the contexts described in the literature, the literature did not contained extended
time series data to look at, because researchers were only interested in demonstrating a
change from the behavior before the intervention was given to the subject and after the
intervention was given to the subject. The subtleties of the behavior-over- time graph
were reduced to a straight line between two points for the experimental and the control
group.
Many of these social processes are set in specific contexts. As part of our model analysis,
we frequently pushed the model to see what would happen under other conditions and
contexts. Although the output of the model seemed to be reasonable and consistent, we
have found that there are typically no data in the literature to directly validate the model’s
trajectories. There appears to be massive gaps in the literature. The model points out rich
areas of further research that can be done to study those social processes.
The social psychological literature provides a rich store of empirical data and interesting
theories about social behavior. This paper is a first attempt to find dynamic aspects of
this discipline and to model some of these processes. Hopefully, as we continue
modeling, some of these generic structures should become useful for system dynamicists
to incorporate some of these social processes into their models. Also, we would hope
that social psychologists would move toward collecting time series data and paying more
attention to behavior patterns over time. We can learn from each other.
REFERENCES
Baron, R. A., Byrne, D. 2000. Social psychology (9" Ed.). Allyn and Bacon: Boston.
Doyle, J. K. 1997. The cognitive psychology of systems thinking. System Dynamics
Review 13: 253-265.
Clauset, K. H., Gaynor, A. K. 1980. The dynamics of effective and ineffective schooling:
A preliminary report of a system dynamics policy study. Paper presented at the Annual
Meeting of the American Educational Research Association, Los Angeles.
Clauset, K. H., Gaynor, A. K. 1982. Improving schools for low achieving children: A
system dynamics policy study. Paper presented at the Annual Meeting of the American
Educational Research Association, New York.
Clauset, K. H., Gaynor, A. K. 1992. Educational policy analysis: An analysis of
effective and ineffective schools using the system dynamics method. In Analysis of
dynamic psychological systems, Levine, R. L., Fitzgerald, H, (eds.), Vol 2, Plenum Press:
New York.
Dowling, A.M., MacDonald, R. H. 1995. Simulation of systems archetypes. Paper given
at the Proceedings of the 1995 International System Dynamics Society Conference, 454-
463.
Doyle, J. K, Ford, D. N. 1998. Mental models concepts for system dynamics research.
System Dynamics Review 14: 3-29.
Doyle, J. K., Ford, D. N. 1999. Mental models concepts revisited: Some clarifications
and a reply to Lane. System Dynamics Review 15: 411-415.
Farmer, R. E. A. 1999 The macroeconomics of self-fulfilling prophecies 2" Ed). The
MIT Press: Cambridge, MA.
Fiedler, K, Semin, G. R., Koppetsch, C. 1991. Language use and attributional biases in
close personal relationships. Personality and Social Psychology Bulletin 17: 147-155.
Forrester, J. W., Senge, P. M. 1980. Tests for building confidence in system dynamics
models. In System dynamics, Legasto, A., Forrester, J. W., and Lyneis, J. (eds), TIMS
Studies in the Management Sciences, 14, North Holland: New York, 209-228.
Holmberg, D., Holmes, J. G. 1994. Reconstruction of relationship memories: A mental
models approach. In Autobiographical memory and the validity of retrospective reports,
Schwartz, N,. Sudman S. (Eds.),. Springer-Verlag: New York.
Kao, A. H., Gabrowski, M W., Doyle, J. K. 1998. Teaching social psychology through
system dynamics. Proceedings of the 16" International Conference of the System
Dynamics Society, Quebec City, Canada.
Kelley, H. H. 1972. Attribution in social interaction. In Attribution: Perceiving the
causes of behavior, Jones, E. E., Kanouse, D.E., Kelley, H. H., Nisbett, R. E., Valins, S.,
Weiner, B. (eds.). General Learning Press: Morristown, NJ.
Kelley, H. H., Michela, J. L. 1980. Attribution theory and research. Annual Review of
Psychology 31:57-501.
Kim, D. H., Anderson, V. 1998. Systems archetype basics: From story to structure.
Pegasus Communications: Waltham, MA.
Lane, D. C., Smart C. 1996. Reinterpreting ‘generic structure:’ evolution, application
and limitations of a concept. System Dynamics Review 12: 87-120.
Levine, R. L. 1983. The paradigms of psychology and system dynamics. Proceedings of
the 1983 International System Dynamics Society Conference, Chestnut Hill, MA, USA.
Levine, R. L. 2000. System dynamics applied to psychological and social problems.
Proceedings of the 18" International Conference of the System Dynamics Society,
Bergen, Norway.
Levine, R. L., Van Sell, M., Rubin, B. 1992. System dynamics and the analysis of
feedback processes in social and behavioral systems). In Analysis of dynamic
psychological systems, Levine, R. L., Fitzgerald, H, (eds.), Vol 1, Plenum Press: New
York.
Levine, R. L., Nguyen. H. 2000. Coflow structures: Some problems and solutions in
representing psychological characteristics and processes. Proceedings of the 2000
International System Dynamics Conference, Bergen, Norway.
Low, G. W. 1978. Simulating the Zimbardo prison experiment. Technical Report D-
3043. Sloan School of Management, M.L.T.
McFarland, S. G. & Ross, , M. 1984. The relation between current impressions and
memories of self and dating partners. Unpublished manuscript, University of Waterloo.
McGuire, W. J. & McGuire, C. V. 1986. Differences in conceptualizing self versus
conceptualizing other people as manifested in contrasting verb types used in natural
speech. Journal of Personality and Social Psychology 51: 1135-1143.
Meadows, D. H. 1982. Whole earth models and systems. The Coevolution Quarterly,
98-108.
Milgram, S. 1963. Behavior study of obedience. Journal of Abnormal and Social
Psychology 67: 371-378.
Milgram, S. 1965. Liberating effects of group pressure. Journal of Personality and
Social Psychology 1: 127-134.
Myers, D. G. 2000. Exploring social psychology. Boston: McGraw-Hill.
Pettigrew, T. F. 1997. Generalized intergroup contact effects on prejudice. Personality
and Social Psychology Bulletin 23: 173-185.
Repenning, N. P, Sterman, J.D. 2000. Getting quality the old-fashioned way: Self-
confirming attributions in the dynamics of process improvement. Pp. 201-235. In:
Improving Research in Total Quality Management . Scott, R., Cole, R. (eds). Newbury
Park, CA: Sage
Riley, S. 2001. A dynamic approach to integrating wildlife management with basic
human dimensions. Technical Report, Department of Fisheries and Wildlife, Michigan
State University.
Richardson, G. 2000. Insightful little models. Proceedings of the System ST/DM
Conference, Skamanaia Lodge.
Richmond, B. 1980. Generalization with individual uniqueness: Modeling the Milgram
experiments. Technical Report D-2508-2. Sloan School of Management, M.LT.
Richmond, B. 2002. An introduction to systems thinking. High Performance Systems:
Hanover, NH.
Roberts, N. 1978. System simulation of student performance in elementary classroom. In
Managerial Applications of System Dynamics, Roberts, E. (ed.). Productivity Press:
Cambridge MA
Rosenthal, R., & Jacobson, I. 1968. Pygmalion in the classroom: Teacher expectation
and student intellectual development. Holt, Rinehart, & Winston : New York.
Ross, I. 1977. The intuitive psychologist and his shortcomings: Distortions in the
attribution process. In Advances in experimental social psychology. (Vol 10),. Berkowitz,
L (ed.). Academic Press: New York.
Sastry, M. A. 1998. Analyzing the research on self-reinforcing processes in
organizations: Another approach to archetypes. Proceedings of the 16" International
Conference of the System Dynamics Society, Quebec City, Canada.
Schwartzwald, J., Amir, Y., & Crain, R. 1. 1992. Long-term effects of school
desegregation experiences on interpersonal relations in the Israeli defense forces.
Personality and Social Psychology Bulletin, 18: 357-368.
Senge, P. M. 1990. The fifth discipline: The art and practice of the Learning
Organization. Doubleday/Currency: New York.
White, P. A., & Younger, D. P. 1988. Differences in the ascription of transient internal
states to self and other. Journal Of Experimental Social Psychology, 24: 292-309.
Wright, , S. C., Aron, A., McLaughlin-Volpe, T., & Ropp, S. A. 1997. The extended
contact effect: Knowledge of cross-group friendships and prejudice. Journal of Personal
and Social Psychology, 73: 73-90.
Back to the Top
Model 1, Self-fulfilling Prophesy
Student_Performance(t) = Student_Performance(t - dt) + (change_in_student_perform) *
dt
INIT Student_Performance = 55
INFLOWS:
change_in_student_perform = performance_gap/time_to_change_perform
performance_gap = Teacher_expectation - Student_Performance
Teacher_expectation = 40
time_to_change_perform = 4
Model 2, Self-fulfilling Prophesy: Reciprocal Relationships
Student_Performance(t) = Student_Performance(t - dt) + (change_in_student_perform)
*dt
INIT Student_Performance = 65
INFLOWS:
change_in_student_perform = gap_in_performance/time_to_change_perform
Teacher_Expectations(t) = Teacher_Expectations(t - dt) + (change_in_teacher_expect) *
dt
INIT Teacher_Expectations = 40
INFLOWS:
change_in_teacher_expect = gap_in_expectation*perseverance_effect
gap_in_expectation = Student_Performance -Teacher_Expectations
gap_in_performance = Teacher_Expectations - Student_Performance
perseverance_effect = .25 {fraction/month}
time_to_change_perform =4
Simple Extended Contact Model
Cum_Time_Spent_with_Target_Group(t) = Cum_Time_Spent_with_Target_Group(t -
dt) + (spending_time_with_target_group) * dt
INIT Cum_Time_Spent_with_Target_Group = 0
INFLOWS:
spending time_with_target_group =
baseline_spending_time_with_target_group*ef_friendship_on_spending_time_together
Degree_Of_Friendship(t) = Degree_Of_Friendship(t - dt) + (making_friends -
decline_in_friendship) * dt
INIT Degree_Of_Friendship = 20
INFLOWS:
making_friends = friendship_units_per_hour*spending_time_with_target_group
OUTFLOWS:
decline_in_friendship = Degree_Of_Friendship/time_to_lose_friendship
baseline_spending_time_with_target_group = 4
friendship_units_per_hour = 1
time_to_lose_friendship = 8.3
ef_friendship_on_spending_time_together = GRAPH(Degree_Of_Friendship)
(0.00, 1.00), (10.0, 1.03), (20.0, 1.12), (30.0, 1.19), (40.0, 1.31), (50.0, 1.58), (60.0, 2.08),
(70.0, 2.66), (80.0, 2.89), (90.0, 2.96), (100, 3.00)
Contact Burnout: Fatigue Model
Contact_Burnout(t) = Contact_Burnout(t - dt) + (burning_out -
recuperating from_burnout) * dt
INIT Contact_Burnout = 0
INFLOWS:
burning_out = burnout_hour*spending_time_with_target_group
OUTFLOWS:recuperating_from_burnout = Contact_Burnout/time_to_dissapate_burnout
Cum_Time_Spent_with_Target_Group(t) = Cum_Time_Spent_with_Target_Group(t -
dt) + (spending_time_with_target_group) * dt
INIT Cum_Time_Spent_with_Target_Group = 0
INFLOWS:
spending_time_with_target_group =
baseline_spending_time_with_target_group*ef_friendship_on_spending_time_together*e
f_of_burnout_on_spending_time_with_other_group
Degree_Of_Friendship(t) = Degree_Of_Friendship(t - dt) + (making_friends -
decine_in_friendship) * dt
INIT Degree_Of_Friendship = 20
INFLOWS:
making_friends = friendship_units_per_hour*spending_time_with_target_group
OUTFLOWS:
decine_in_friendship = Degree_Of_Friendship/time_to_lose_liking_of_target_persons
baseline_spending_time_with_target_group = 4
burnout_hour = 1.2
friendship_units_per_hour = 1.2
time_to_dissapate_burnout = 9
time_to_lose_liking_of_target_persons = 8.3
ef_friendship_on_spending_time_together = GRAPH(Degree_Of_Friendship)
(0.00, 1.00), (10.0, 1.03), (20.0, 1.12), (30.0, 1.19), (40.0, 1.31), (50.0, 1.58), (60.0, 2.08),
(70.0, 2.66), (80.0, 2.89), (90.0, 2.96), (100, 3.00)
ef_of_burnout_on_spending_time_with_other_group = GRAPH(Contact_Burnout)
(0.00, 1.00), (10.0, 1.00), (20.0, 1.00), (30.0, 1.00), (40.0, 1.00), (50.0, 0.975), (60.0,
0.915), (70.0, 0.375), (80.0, 0.165), (90.0, 0.075), (100, 0.06)
Contact Burnout: Mood Variation
Contact_Burnout(t) = Contact_Burnout(t - dt) + (burning_out -
recuperating from_burnout) * dt
INIT Contact_Burnout = 0
INFLOWS:
burning out =Burnout_units_per_hour*spending_time_with_target_group
OUTFLOWS:
recuperating from_burnout = Contact_Burnout/time_to_dissapate_burnout
Cum_Time_Spent_withTarget_Group(t) = Cum_Time_Spent_withTarget_Group(t - dt) +
(spending_time_with_target_group) * dt
INIT Cum_Time_Spent_withTarget_Group = 0
INFLOWS:
spending _time_with_target_group =
baseline_spending_time_with_target_group*ef_of_friendship_on_spending_time_togeth
er
Degree_Of_Friendship(t) = Degree_Of_Friendship(t - dt) +
(getting_to_like_target_group - losing_friendship) * dt
INIT Degree_Of_Friendship = 20
INFLOWS:
getting_to_like_target_group =
friendship_units_per_hour*spending_time_with_target_group
OUTFLOWS:
losing_friendship =
(Degree_Of_Friendship/time_to_lose_friendship)*ef_of_burnout_on_losing_friendship:_
Bad_Mood
baseline_spending time_with_target_group = 4
Burnout_units_per_hour = 1
friendship_units_per_hour = 2
time_to_dissapate_burnout = 9
time_to_lose_friendship = 8.3
ef_of_burnout_on_losing_friendship:_Bad_Mood = GRAPH(Contact_Burnout)
(0.00, 1.00), (10.0, 1.00), (20.0, 1.00), (30.0, 1.00), (40.0, 1.20), (50.0, 2.48), (60.0, 2.65),
(70.0, 2.79), (80.0, 2.92), (90.0, 2.96), (100, 3.00)
ef_of_friendship_on_spending_time_together = GRAPH(Degree_Of_Friendship)
(0.00, 1.00), (10.0, 1.03), (20.0, 1.12), (30.0, 1.19), (40.0, 1.31), (50.0, 1.58), (60.0, 2.08),
(70.0, 2.66), (80.0, 2.89), (90.0, 2.96), (100, 3.00)
Additional run: Alternative setting of ef_of_burnout_on_losing_friendship
ef_of_burnout_on_losing_friendship:_Bad_Mood = GRAPH(Contact_Burnout)
(0.00, 1.00), (10.0, 1.00), (20.0, 1.00), (30.0, 1.00), (40.0, 1.10), (50.0, 1.17), (60.0, 1.56),
(70.0, 1.56), (80.0, 1.59), (90.0, 1.59), (100, 1.59)
Extended Contact Model: Combination of fatigue and mood
Contact_Burnout(t) = Contact_Burnout(t - dt) + (burning_out -
recuperating _from_burnout) * dt
INIT Contact_Burnout = 0
INFLOWS:
burning_out = Burnout_units_per_hour*spending_time_with_target_group
OUTFLOWS:
recuperating from_burnout = Contact_Burnout/time_to_dissapate_burnout
Cum_Time_Spent_with_Target_Group(t) = Cum_Time_Spent_with_Target_Group(t -
dt) + (spending_time_with_target_group) * dt
INIT Cum_Time_Spent_with_Target_Group = 0
INFLOWS:
spending_time_with_target_group =
baseline_spending_time_with_target_group*ef_friendship_on_spending_time_together*e
f_of_burnout_on_spending_time_with_target_people
Degree_Of_Friendship(t) = Degree_Of_Friendship(t - dt) + (making_friends -
loss_of_friendship) * dt
INIT Degree_Of_Friendship = 20
INFLOWS:
making_friends = friendship_units_per_hour*spending_time_with_target_group
OUTFLOWS:
loss_of_friendship =
(Degree_Of_Friendship/time_to_lose_friendship)*ef_burnout_on_loss_of_friendship:_B
ad_Mood
baseline_spending_time_with_target_group = 4
Burnout_units_per_hour = 1
friendship_units_per_hour = 1.8
time_to_dissapate_burnout = 9
time_to_lose_friendship = 9
ef_burnout_on_loss_of_friendship:_Bad_Mood = GRAPH(Contact_Burnout)
(0.00, 1.00), (10.0, 1.00), (20.0, 1.00), (30.0, 1.02), (40.0, 1.20), (50.0, 2.48), (60.0, 2.65),
(70.0, 2.79), (80.0, 2.92), (90.0, 2.96), (100, 3.00)
ef_friendship_on_spending_time_together = GRAPH(Degree_Of_Friendship)
(0.00, 1.00), (10.0, 1.03), (20.0, 1.12), (30.0, 1.19), (40.0, 1.31), (50.0, 1.58), (60.0, 2.08),
(70.0, 2.66), (80.0, 2.89), (90.0, 2.96), (100, 3.00)
ef_of_burnout_on_spending_time_with_target_people = GRAPH(Contact_Burnout)
(0.00, 1.00), (10.0, 1.00), (20.0, 1.00), (30.0, 1.00), (40.0, 1.00), (50.0, 0.975), (60.0,
0.915), (70.0, 0.885), (80.0, 0.87), (90.0, 0.075), (100, 0.06)
Model of Dynamics of the Fundamental Attribution Error
Perceived_Consistency(t) = Perceived_Consistency(t - dt) +
(change_in_perceived_consistency) * dt
INIT Perceived_Consistency = 100
INFLOWS:
change_in_perceived_consistency = gap_in_consistency/time_to_perceive_consistency
Perceived_Internal_Strength_of_Target_by_Observer(t) =
Perceived_Internal_Strength_of_Target_by_Observer(t - dt) +
(changing _perceived_internal_strength) * dt
INIT Perceived_Internal_Strength_of_Target_by_Observer = 90
INFLOWS:
changing_perceived_internal_strength =
gap_in_perceiving internal_attribution/time_to_change_perceived_internal_strength
gap_in_consistency = targets_consistency_level - Perceived_Consistency
gap_in_perceiving_internal_attribution = target_actual_strength_of_internal_attribution -
Perceived_Internal_Strength_of_Target_by_Observer
targets_consistency_level = 15
target_actual_strength_of_internal_attribution = 15
time_to_perceive_consistency = 7
time_to_change_perceived_internal_strength = GRAPH(Perceived_Consistency)
(0.00, 2.50), (10.0, 4.00), (20.0, 7.00), (30.0, 26.0), (40.0, 120), (50.0, 156), (60.0, 178),
(70.0, 195), (80.0, 200), (90.0, 200), (100, 200)
CC BY-NC-SA 4.0