Abstract and Figures

In the analysis of stimulus competition in causal judgment, 4 variables have been frequently confounded with respect to the conditions necessary for stimuli to compete: causal status of the competing stimuli (causes vs. effects), temporal order of the competing stimuli (antecedent vs. subsequent) relative to the noncompeting stimulus, directionality of training (predictive vs. diagnostic), and directionality of testing (predictive vs. diagnostic). In a factorial study using an overshadowing preparation, the authors isolated the role of each of these variables and their interactions. The results indicate that competition may be obtained in all conditions. Although some of the results are compatible with various theories of learning, the observation of stimulus competition in all conditions calls for a less restrictive reformulation of current learning theories that allows similar processing of antecedent and subsequent events, as well as of causes and effects.
Content may be subject to copyright.
Competition Between Antecedent and Between Subsequent Stimuli in
Causal Judgments
Francisco Arcediano
Auburn University
Helena Matute
Universidad de Deusto
Martha Escobar
Auburn University
Ralph R. Miller
State University of New York at Binghamton
In the analysis of stimulus competition in causal judgment, 4 variables have been frequently confounded
with respect to the conditions necessary for stimuli to compete: causal status of the competing stimuli
(causes vs. effects), temporal order of the competing stimuli (antecedent vs. subsequent) relative to the
noncompeting stimulus, directionality of training (predictive vs. diagnostic), and directionality of testing
(predictive vs. diagnostic). In a factorial study using an overshadowing preparation, the authors isolated
the role of each of these variables and their interactions. The results indicate that competition may be
obtained in all conditions. Although some of the results are compatible with various theories of learning,
the observation of stimulus competition in all conditions calls for a less restrictive reformulation of
current learning theories that allows similar processing of antecedent and subsequent events, as well as
of causes and effects.
Stimulus competition is defined as the phenomenon in which
responding to a target stimulus (X), on the basis of its signaling
some event, is weakened as a consequence of X’s being trained in
the presence of another stimulus (A) that better signals the same
event (e.g., Kamin, 1968; Pavlov, 1927; Wagner, Logan, Haber-
landt, & Price, 1968; Wasserman, 1974). Overshadowing (Pavlov,
1927) is one example of stimulus competition. In overshadowing,
presentations of two stimuli in compound, A and X (e.g., a tone and
a light), signaling an impending event (e.g., food) results in weaker
conditioned responding to X compared with a condition in which
X is presented alone as a signal for food. That is, although X is
presented signaling the food an equal number of times in both
conditions, participants seem to attribute greater signal value to X
for the impending food when X is presented alone than when it is
presented in compound with A.
The phenomenon of stimulus competition, which was originally
observed in animal conditioning, has also been demonstrated in
human causal judgment experiments (e.g., Baker, Mercier, Valle´e-
Tourangeau, Frank, & Pan, 1993; Dickinson, Shanks, & Evenden,
1984; Matute, Arcediano, & Miller, 1996; Shanks, 1985; Van
Hamme, Kao, & Wasserman, 1993; Waldmann & Holyoak, 1992;
Wasserman, 1990). These experiments with human participants
were similar to the examples mentioned above, but here, instead of
lights, tones, and food, participants were exposed to a series of
different causes (e.g., fictitious medicines) signaling the occur-
rence of an effect (e.g., an allergic reaction), and then, at test, they
were asked to indicate the degree to which they thought each one
of the potential causes was a cause of the effect. Stimulus com-
petition in causal judgment is of theoretical importance because it
speaks to the conditions necessary for causal judgment (and learn-
ing in general) to occur. Moreover, stimulus competition is widely
used nowadays as a tool with which to discriminate between the
two most prevalent frameworks that seek to explain causal judg-
ments in humans.
On the one hand, causal judgment has been frequently explained
in the framework of the modern associative learning theories (e.g.,
Cobos, Lo´pez, Can˜o, Almaraz, & Shanks, 2002; Shanks & Lo´pez,
1996). In this framework, the same principles that are applied to
animal conditioning are applied to causal judgment. In the majority
of associative theories (e.g., Mackintosh, 1975; Pearce & Hall,
1980; Rescorla & Wagner, 1972), it is assumed that the establish-
ment of an association (i.e., learning) between two events (e.g., a
cause and an effect) requires that the two events occur temporally
proximate to each other, and that the event presented first in the
event sequence (the antecedent event or cue) predicts nonredun-
dantly the occurrence of the event presented second in the event
sequence (the subsequent event or outcome). Thus, in this frame-
work, stimulus competition occurs unidirectionally between ante-
cedent events predicting the same subsequent event, regardless of
whether these events are causes or effects (either of which could be
Francisco Arcediano, Department of Computer Science and Software
Engineering, Auburn University; Helena Matute, Department of Psychol-
ogy, Universidad de Deusto, Bilbao, Spain; Martha Escobar, Department
of Psychology, Auburn University; Ralph R. Miller, Department of Psy-
chology, State University of New York at Binghamton.
Support for this research was provided by Departamento de Educacio´n,
Universidades e Investigacio´n of the Basque Government Grant PI-2000-
12, and by an Auburn University Competitive Research Grant award. We
thank Robert Cole, James Denniston, Mirko Gerolin, Lisa Gunther, Phil-
ippe Oberling, Nuria Ortega, Oskar Pinen˜o, Gonzalo Urcelay, and Sonia
Vegas for comments on a draft of the article, and Abraham Arias, Kenya
Castellanos, Jamie Francis, Leanne Scalli, and Tyson Platt for their help in
conducting the experiments.
Correspondence concerning this article should be addressed to Ralph R.
Miller, Department of Psychology, SUNY-Binghamton, Binghamton, NY
13902-6000. E-mail: rmiller@binghamton.edu
Journal of Experimental Psychology: Copyright 2005 by the American Psychological Association
Learning, Memory, and Cognition
2005, Vol. 31, No. 2, 228 –237
0278-7393/05/$12.00 DOI: 10.1037/0278-7393.31.2.228
228
the antecedent event in the information sequence provided to the
participant). In this account, causal judgments proceed indepen-
dent of any general and abstract causal knowledge that people may
have. That is, the causal relationship between the events is not
relevant to stimulus competition, but their temporal order of pre-
sentation is critical.
On the other hand, human causal judgment has been considered
by some researchers as a special type of learning with principles
different from those of associative learning. In the causal model
theory (Waldmann & Holyoak, 1992), it is assumed that people are
sensitive to general and abstract causal knowledge (i.e., people are
sensitive to how causal relationships between events are defined),
and that people use this knowledge to attribute causal relationships
between events. According to this account, the acquisition of
causal knowledge cannot proceed independent of the knowledge of
the underlying causal structure of the events. Thus, according to
Waldmann and Holyoak, people are aware that multiple potential
causes of an effect can compete as causes of that effect, indepen-
dent of their temporal order of presentation (i.e., their being
antecedent or subsequent events is irrelevant) and that multiple
potential effects of a cause cannot compete as consequences of the
cause regardless of whether they are presented first or second (e.g.,
Waldmann, 2000; Waldmann & Holyoak, 1992).
Experiments analyzing stimulus competition within causal judg-
ment situations have become the benchmark with which to support
or reject these two different families of accounts of causal judg-
ments. Most contemporary theories of associative learning (e.g.,
Rescorla & Wagner, 1972) assume that associations are unidirec-
tional, from antecedent events (cues) to subsequent events (out-
comes). This assumption has led to their postulating learning
mechanisms that predict stimulus competition only between ante-
cedent events, regardless of whether they are causes or effects.
Thus, stimulus competition is expected in animal conditioning
between multiple conditioned stimuli that signal an impending
unconditioned stimulus but not between multiple unconditioned
stimuli that follow a conditioned stimulus (e.g., Mackintosh, 1975;
Pearce & Hall, 1980; Rescorla & Wagner, 1972); in associative
neural networks, competition is expected to occur between inputs
but not between outputs (e.g., Shanks & Lo´pez, 1996); and in
causal judgment situations, competition is expected to occur be-
tween antecedent stimuli (causes or effects) but not between sub-
sequent stimuli (e.g., Cobos et al., 2002). In contrast, as mentioned
above, in the causal model theory (Waldmann & Holyoak, 1992),
it is postulated that competition occurs between causes (antecedent
or subsequent events) but not between effects.
As can be seen in Figure 1, empirical data often seem to provide
support for (but sometimes refute) the claims from each family of
accounts of causal judgment. The associative accounts are sup-
ported by findings of stimulus competition between antecedent
stimuli (Panels A and D) regardless of whether they are causes
(with predictive training, i.e., causes followed by effects; e.g.,
Cobos et al., 2002; Shanks & Lo´pez, 1996; Waldmann, 2000,
2001; Waldmann & Holyoak, 1992) or effects (with diagnostic
training, i.e., effects followed by causes; e.g., Cobos et al., 2002;
Price & Yates, 1993; Shanks, 1991; Shanks & Lo´pez, 1996).
However, this evidence has been challenged by failed demonstra-
tions of competition between antecedent effects (e.g., Van Hamme
et al., 1993; Waldmann, 2000, 2001; Waldmann & Holyoak, 1992)
and by evidence of stimulus competition between subsequent
effects (e.g., Esmoris-Arranz, Miller, & Matute, 1997; Matute et
al., 1996; Miller & Matute, 1998; but see Cobos et al., 2002).
In contrast with the associative account, the causal model theory
has been supported by the findings of stimulus competition be-
tween causes (e.g., Cobos et al., 2002; Shanks & Lo´pez, 1996;
Waldmann, 2000, 2001; Waldmann & Holyoak, 1992) and the
failed demonstrations of competition between antecedent effects
(e.g., Van Hamme et al., 1993; Waldmann, 2000, 2001; Waldmann
& Holyoak, 1992). However, this last observation has been chal-
lenged by repeated demonstrations of competition between effects
(e.g., Cobos et al., 2002; Matute et al., 1996; Price & Yates, 1993;
Shanks, 1991; Shanks & Lo´pez, 1996) and by the failed demon-
strations of competition between causes with diagnostic training
(i.e., when the effect is presented first; e.g., Cobos et al., 2002;
Price & Yates, 1995). Note that the causal model theory predicts
competition between causes even when they are presented as
subsequent events because, for this account, the causal structure is
the relevant factor, and the temporal order of the stimuli is irrel-
evant. However, competition between subsequent causes is not
predicted by the associative accounts because these models assume
that subsequent stimuli cannot compete.
Interpreting the results from studies of stimulus competition in
causal judgments has been complicated because every reported
experimental finding seems to be susceptible to more than one
interpretation. For example, the most straightforward and best
supported evidence of stimulus competition, competition between
causes with predictive training (e.g., lower ratings to Cause X
when X was presented together with Cause A predicting a subse-
quent effect) can be interpreted as evidence for (a) competition
between causes, (b) competition between antecedent stimuli, and
(c) competition with predictive training. However, to conclude that
causes always compete (as is predicted in the causal model theory),
one must demonstrate competition between causes with diagnostic
training, not just with predictive training. To conclude that ante-
cedent stimuli always compete (as is predicted in associative
models), one must demonstrate competition between causes pre-
sented as antecedents during predictive training and competition
between effects presented as antecedents during diagnostic train-
ing. Finally, to conclude that competition occurs exclusively when
training is predictive, one must demonstrate competition when
causes are followed by effects during training and not when effects
precede causes.
The situation is further complicated by test variables. The ques-
tions used to assess stimulus competition can also be predictive
(from causes to effects) or diagnostic (from effects to causes). In
the associative account, it is assumed that competition can occur
only unidirectionally between antecedent stimuli signaling a sub-
sequent stimuli, and, probably because of this assumption, its
supporters have tested only competition between causes with pre-
dictive training and predictive test questions and competition be-
tween effects with diagnostic training and diagnostic test questions
(e.g., Cobos et al., 2002; Shanks, 1991; Shanks & Lo´pez, 1996).
That is, it has been assumed that training and testing must have the
same forward directionality to observe competition (i.e., from
antecedent events to subsequent events, regardless of their causal
status). In contrast, in the causal model theory, it is assumed that
competition between causes can occur and competition between
effects will never be observed (e.g., Waldmann, 2000; Waldmann
& Holyoak, 1992, 1997).
229
STIMULUS COMPETITION
In summary, four variables can be identified that may have
potential impact on stimulus competition in human causal judg-
ment (see Figure 1):
1. Causal status. Whether the competing stimuli are causes
(Panels A1, A2, B1, and B2 of Figure 1) or effects
(Panels C1, C2, D1, and D2), regardless of their temporal
order of presentation in training or testing.
2. Temporal order of the stimuli. Whether competition oc-
curs between antecedent stimuli (Panels A1, A2, D1, and
D2) or between subsequent stimuli (Panels B1, B2, C1,
and C2), independent of their being causes or effects.
3. Directionality of training. Whether training is predictive
(i.e., causes are presented first; Panels A1, A2, C1, and
C2) or diagnostic (i.e., effects are presented first; Panels
B1, B2, D1, and D2), independent of the competing
stimuli being causes or effects.
4. Directionality of testing. Whether the test question used
to assess competition is predictive (i.e., asking for effects
of causes; A1, B1, C1, and D1) or diagnostic (i.e., asking
for causes of effects; A2, B2, C2, and D2).
For instance, competition between causes does not necessarily
require predictive training and testing (Panel A1) but can also be
studied with any other combination of predictive or diagnostic
training and testing (Panels A2, B1, and B2). Likewise, studying
competition between effects does not necessarily require diagnos-
tic training and testing (Panel D2) but can also be accomplished
Figure 1. The eight panels show the critical variables examined in this research. The variables that may be
responsible for competition in causality judgments are (a) causal status of the competing stimuli: causes (C)
versus effects (E); (b) directionality of training: predictive (cause-to-effect) versus diagnostic (effect-to-cause);
(c) directionality of the test question: predictive (cause-to-effect) versus diagnostic (effect-to-cause); and (d)
temporal order of presentation of the competing stimuli: antecedent versus subsequent. For each form of
competition, the positive and negative evidence is listed in the corresponding panel. 1 Baker et al., 1993; 2
Chapman, 1991; 3 Chapman & Robbins, 1990; 4 Cobos et al., 2000; 5 Cobos et al., 2002; 6 Dickinson
et al., 1984; 7 Esmoris-Arranz et al., 1997; 8 Matute et al., 1996; 9 Miller & Matute, 1998; 10 Price
& Yates, 1993; 11 Price & Yates, 1995; 12 Shanks, 1991; 13 Shanks & Lo´pez, 1996; 14 Van Hamme
et al., 1993; 15 Waldmann & Holyoak, 1992; 16 Waldmann, 2000; 17 Waldmann, 2001; 18
Wasserman, 1990.
230
ARCEDIANO, MATUTE, ESCOBAR, AND MILLER
with any other combination of predictive and diagnostic training
and testing (Panels D1, C1, and C2). More generally, any condi-
tion in which there are multiple stimuli linked to one stimulus is
potentially susceptible to competition among the multiple stimuli
regardless of whether they are causes or effects, antecedent or
subsequent stimuli, whether they are trained in the predictive or
diagnostic direction, and whether they are tested in the predictive
or diagnostic direction. Each of these four potentially critical
variables in stimulus competition should be studied, without their
being confounded with any of the other three variables, and the
potential interactions between them also requires examination in
order to determine the roles they play in producing competition
and their relative weights in doing so.
In the following experiment, we manipulated the role of the
aforementioned four variables within an overshadowing proce-
dure. Training information was presented to the participants in a
trial-by-trial format. While so doing, we systematically varied the
causal status of the competing stimuli (causes vs. effects), the
directionality of training (predictive vs. diagnostic), the direction-
ality of testing (predictive vs. diagnostic test questions), and the
temporal order of presentation of the competing stimuli (anteced-
ent vs. subsequent) during training, all in a single integrative study.
Overview of the Experiment
We used an overshadowing procedure (e.g., Pavlov, 1927) in
order to analyze some of the different variables that play a role in
stimulus competition in causality judgment (see Figure 1). As
previously mentioned, in an overshadowing procedure, the target
stimulus is trained in compound with another stimulus as a signal
for an impending event, which results in lower ratings of the target
stimulus than would be obtained in a control condition in which it
was paired alone with the impending event.
We manipulated whether causes or effects were the potentially
competing stimuli (i.e., multiple causes of one effect or multiple
effects of one cause) and whether causes or effects were presented
first (antecedent) or second (subsequent) in the event sequence on
each training trial (see Figure 1 for an explanation of the variables
and Table 1 for the experimental design). Thus, we controlled for
the influence on stimulus competition of the causal status of the
competing stimuli (causes vs. effects), the directionality of training
(predictive vs. diagnostic), and the temporal order of presentation
of the competing stimuli (antecedent vs. subsequent). In addition,
we controlled for the influence of the directionality of the test
question, asking participants to answer predictive and diagnostic
test questions at the end of training.
The experiment was completely within-subjects (i.e., every par-
ticipant was exposed to all the conditions and was tested with all
the test questions). There was a single cover story for all the
conditions. Participants had to learn the relationships among eight
causes and eight effects. The causes were the ingestion of various
foods by fictitious patients, and the effects were adverse reactions
to those foods. The role of every food as a cause and the role of
every adverse reaction as an effect were counterbalanced between
participants to control for any potential impact of their familiarity
or prior experience. During training, participants had to guess,
after the presentation of the first stimulus or stimuli on each trial,
what stimulus or stimuli would follow. Thus, we were able to
record the progress of learning of these relationships during train-
ing and assess whether participants were able to learn these rela-
tionships despite the complexity of the task. This allowed us to
determine whether a low rating to a given stimulus during testing
was the result of a deficit in acquiring the information.
Method
Participants and Apparatus
Twenty-eight undergraduate students from the State University of New
York at Binghamton and 22 undergraduate students from Auburn Univer-
Table 1
Design Summary
Condition
Training variables Test variables
Training Causal status Temporal order Directionality Directionality
Competing cause 3 effect
(control)
C1 3 E1 Antecedent Predictive C1 3 E1? (pred) E1 3 C1? (diag)
Competing causes 3 effect
(experimental)
C2 C3 3 E2 Multiple causes Antecedent Predictive C2 3 E2? (pred)
C3 3 E2? (pred)
E2 3 C2? (diag)
E2 3 C3? (diag)
Effect 3 competing cause
(control)
E3 3 C4 Subsequent Diagnostic C4 3 E3? (pred) E3 3 C4? (diag)
Effect 3 competing causes
(experimental)
E4 3 C5 C6 Multiple causes Subsequent Diagnostic C5 3 E4? (pred)
C6 3 E4? (pred)
E4 3 C5 (diag)
E4 3 C6 (diag)
Cause 3 competing effect
(control)
C1 3 E1 Subsequent Predictive C1 3 E1? (pred) E1 3 C1? (diag)
Cause 3 competing effects
(experimental)
C7 3 E5 E6 Multiple effects Subsequent Predictive C7 3 E5? (pred)
C7 3 E6? (pred)
E5 3 C7? (diag)
E6 3 C7? (diag)
Competing effect 3 cause
(control)
E3 3 C4 Antecedent Diagnostic C4 3 E3? (pred) E3 3 C4? (diag)
Competing effects 3 cause
(experimental)
E7 E8 3 C8 Multiple effects Antecedent Diagnostic C8 3 E7? (pred)
C8 3 E8? (pred)
E7 3 C8? (diag)
E8 3 C8? (diag)
Note. C1–C8 are eight different causes; E1–E8 are eight different effects. Temporal order refers to whether the potentially competing stimuli were
presented in the antecedent (cue) or subsequent (outcome) position of the trial pairing. Predictive or pred predictive directionality (from cause to effect),
and diagnostic or diag diagnostic directionality (from effect to cause). Conditions listed twice (for clarity) are in italics the second time they are listed.
231
STIMULUS COMPETITION
sity participated in the study. All Binghamton students participated in the
study as partial fulfillment of a course requirement. All Auburn University
students received extra credit within a course in return for their participa-
tion. The experiment was conducted in small rooms containing one to four
personal computers.
Design
The five variables manipulated in the experiment were the following.
Experimental condition (experimental vs. control). There were four
experimental conditions and two control conditions. For the experimental
(overshadowing) conditions, a compound of two causes or two effects was
paired with its corresponding effect or cause. For the control conditions,
one cause or one effect was paired with its corresponding effect or cause.
Causal status (multiple causes vs. multiple effects). The potentially
competing stimuli were causes for half of the conditions and effects for the
remaining half.
Temporal order of presentation of the competing stimuli (antecedent vs.
subsequent). The potentially competing stimuli were antecedent (i.e.,
presented first in the event sequence; either causes or effects) for one half
of the conditions and subsequent (i.e., presented second in the event
sequence; either causes or effects) for the remaining half.
Directionality of training (predictive vs. diagnostic). Half of the con-
ditions received predictive training; that is, causes were presented before
effects during training. The other half received diagnostic training; that is,
effects were presented before causes during training. It is important to note
that directionality of training was not independent of the temporal order of
presentation of the competing stimuli and causal status but, rather, repre-
sented the interaction of these two variables. For example, if causes are
antecedent stimuli, the directionality of training is, by definition, predictive
(although, as shown in Figure 1, either competition between causes or
between effects can be studied in such a predictive training condition); if
causes are subsequent stimuli, the directionality of training is diagnostic
(although competition between causes and between effects can also be
studied in this condition). In statistical analysis, any two of these variables
could serve as independent variables, with the third variable emerging as
the interaction of the first two variables.
Directionality of test (predictive vs. diagnostic). All participants re-
ceived predictive (from cause to effect) and diagnostic (from effect to
cause) test questions for each cause and each effect. The words predictive
and diagnostic were not used on the test questions to facilitate the paral-
lelism between the two types of test questions.
In Table 1, we summarize the design. We manipulated five variables in
six within-subjects conditions (four experimental and two control condi-
tions). Note that, in Table 1, eight (not six) conditions were depicted for the
sake of clarity with respect to the pairwise comparisons. The C1 3 E1 (i.e.,
Cause 1 predicting Effect 1) condition served as the control condition for
competition between causes in the C2 C3 3 E2 condition (i.e., Cause 2 and
Cause 3 predicting Effect 2). In addition, the C1 3 E1 condition served as
the control condition for competition between effects in the C7 3 E5 E6
condition (i.e., Cause 7 predicting Effect 5 and Effect 6). Likewise, the
E3 3 C4 condition served as the control condition for competition be-
tween effects in the E7 E8 3 C8 condition (i.e., Effect 7 and Effect 8
diagnosing Cause 8) and competition between causes in the E4 3 C5 C6
condition (i.e., Effect 4 diagnosing Cause 5 and Cause 6).
Procedure
In order to manipulate whether the competing stimuli were causes or
effects and whether training was predictive (cause-to-effect direction) or
diagnostic (effect-to-cause direction), we introduced a cover story in which
the causes were foods that fictitious patients ate and in which the effects
were adverse reactions to these foods. Participants were informed that they
would first see for each patient either a record of what food(s) the patient
ate or what adverse reaction(s) the patient developed. When food(s) were
presented first, participants were asked to guess which adverse reaction(s)
that patient had developed and were then told what adverse reaction(s)
there were; when adverse reaction(s) were presented first, participants were
asked to guess which food(s) that patient had eaten and were then told what
food(s) had been eaten.
Training and testing was implemented with personal computers. Partic-
ipants entered their responses by pointing with the mouse cursor and
pressing either of the mouse buttons. There were 48 training trials, each of
which represented a different fictitious patient. These training trials were
organized in eight blocks (undifferentiated for the participant) with one
trial of each of the six conditions per block. The order of the conditions
within each block for each participant was pseudorandomized and coun-
terbalanced among participants. On each trial, participants were shown two
charts, one of which contained causes (i.e., the foods that the fictitious
patient ate) and the other of which contained effects (i.e., the adverse
reactions that the patient had developed to those foods). Each chart de-
picted one or two causes or effects, depending on the condition. The foods
that served as causes were chicken, cheese, mushrooms, bananas, fish,
onions, rice, and carrots; the adverse reactions that served as effects were
nausea, fever, rash, itching, headache, constipation, insomnia, and cramps.
The role of the foods as causes and the adverse reactions as effects were
counterbalanced between participants to control for any effect of familiar-
ity. That is, any food could be any of the eight causes, and any adverse
reaction could be any of the eight effects.
In the predictive training conditions, the first screen depicted a chart of
causes showing the names of one or two foods that the patient had eaten.
Below the chart with causes, participants could see all eight potential
adverse reactions that could follow that cause(s), and they were asked to
select the adverse reaction(s) that they thought that patient had developed.
There were no restrictions on the number of selections they could make.
That is, they could select none of the choices or all of them. Once
participants had made their selections, and after clicking with the mouse
cursor on the “accept” button on the screen, they received feedback about
the actual adverse reactions that the current patient had developed. The
diagnostic training conditions were similar to the predictive training con-
ditions, but the order of the causes and effects was reversed. In the
diagnostic conditions, participants had to select from the choices what
food(s) they guessed were the causes of the adverse reaction(s) the current
patient had developed.
After training, participants had to answer one predictive and one diag-
nostic test question for each cause and each effect. One half of the
participants received first the eight predictive test questions and then the
eight diagnostic test questions. For the other half, this order was reversed.
The predictive test question read To what degree do you think that having
eaten name-of-food is indicative that each of the following adverse reac-
tions will occur? Below the question, participants were presented with the
eight potential adverse reactions, and they had to select using the mouse
their rating for each of the effects. Participants rated each effect by
selecting a value on a response bar presented above the name of each
effect. The bar was anchored at 0 and 100; ratings could be increased or
decreased in intervals of one. The diagnostic test question read To what
degree do you think that developing name-of-adverse-reaction is indicative
of having eaten each of the following foods? As in the predictive test
questions, participants had to select with the mouse their rating for each
food in a 0 –100 scale. The order of questioning concerning each cause and
each effect was counterbalanced among participants. The order of the
specific effects and causes to be rated in each predictive and test question
was not counterbalanced, but their role in the experimental design was.
That is, for example, participants always rated nausea first when asked a
predictive question, but nausea could be any of the eight potential effects
counterbalanced between participants.
232
ARCEDIANO, MATUTE, ESCOBAR, AND MILLER
Results
Stimulus competition was observed regardless of whether the
potential competing stimuli were causes or effects, whether the
directionality of training was predictive or diagnostic, whether the
directionality of testing was predictive or diagnostic, and whether
the competing stimuli were presented in the antecedent or the
subsequent temporal order. Thus, the only variable that influenced
the participants’ judgments in an important way seemed to be
whether the test stimuli had been in the experimental (overshad-
owing) or control condition. Figure 2 depicts participants’ ratings
for each target causal relationship.
Figure 3A depicts participants’ responding during training for
each one of the conditions. As can be observed, participants were
sensitive to the causal relationships between the stimuli during
training. At the end of training, participants discriminated between
the correct and incorrect choices. Figure 3B depicts the predictive
and diagnostic ratings during testing for each cause and each
effect. As during training, participants discriminated which causes
were followed by which effects and vice versa. Thus, it seems that
participants successfully discriminated the stimuli during training,
and that stimulus competition at test did not reflect a deficiency of
acquisition but a deficit due to overshadowing per se.
For statistical comparisons, when two causes or two effects were
presented in compound, we used the average of their predictive–
diagnostic ratings as the dependent variable for their respective
comparisons. Thus, for example, when comparing competition
between causes in predictive training (i.e., C13 E1 vs. C2
C33 E2), we averaged the ratings of Cause 2 and Cause 3 pre-
dicting Effect 2, and we compared the mean with the ratings of
Cause 1 predicting Effect 1.
A 2 (experimental condition: experimental vs. control) 2
(causal status: competing causes vs. competing effects) 2 (di-
rectionality of training: predictive vs. diagnostic) 2 (direction-
ality of test: predictive vs. diagnostic) analysis of variance
(ANOVA) performed on the ratings for the target stimulus in each
condition yielded a main effect of experimental condition (exper-
imental vs. control), F(1, 49) 43.05, MSE 3,104, p .0001,
and no other main effect or interaction (all Fs 1.14, all ps
.29). Although there were no interactions, we broke down all the
relevant comparisons to see the impact of the variables on each one
of the cells of Figure 1. Planned comparisons (see Table 2 for
specific statistical values) showed that stimulus competition (lower
ratings for the experimental conditions than the control conditions)
had occurred regardless of the causal status of the competing
stimuli, direction of training, and direction of testing. Thus, the
results were indicative of stimulus competition in all possible
conditions.
In order to analyze the influence of temporal order (antecedent
vs. subsequent, independent of their directionality of training and
causal status), we performed an additional 2 (experimental condi-
tion: experimental vs. control) 2 (causal status: competing
causes vs. competing effects) 2 (temporal order: antecedent vs.
subsequent) 2 (directionality of test: predictive vs. diagnostic)
ANOVA. Although the main effects and interactions were the
same as in the former ANOVA, we needed this additional com-
parison to analyze the impact of temporal order (antecedent vs.
subsequent stimuli) on stimulus competition, because temporal
order and directionality of training were nonorthogonal variables.
As can be seen in Table 2, stimulus competition was observed with
both antecedent and subsequent stimuli regardless of whether they
were causes or effects.
Discussion
Our results show that the phenomenon of stimulus competition
in human causal judgment tasks does not categorically depend on
the causal status of the stimuli (competing causes or competing
effects), the temporal order of the competing stimuli (antecedent or
subsequent), the directionality of the training procedure (predictive
or diagnostic), or the directionality of testing (predictive or diag-
nostic). That is, stimulus competition in causal judgment can be
obtained in every possible manipulation depicted in Figure 1.
In the present research, we systematically manipulated these
variables factorially in an integrative study, and its results do not
fully support any associative account of causal judgment (e.g.,
Cobos et al., 2002; Shanks & Lo´pez, 1996) or any of the models
that treat causal learning as a special adaptation (e.g., Waldmann
& Holyoak’s causal model, 1992, 1997). Our results partially
support the assumptions of the associative account because we
observed competition between antecedent stimuli, independent of
whether they were causes (with predictive testing) or effects (with
diagnostic testing). However, our results also showed competition
between antecedent causes (i.e., predictive training) with diagnos-
tic testing (from effects to causes) and competition between ante-
cedent effects (i.e., diagnostic training) with predictive testing
(from causes to effects). These results seem to refute the assump-
tion that a common directionality during training and testing is
necessary to observe stimulus competition; that is, they are con-
trary to the view that associations are formed only from anteced-
ents to subsequent stimuli (for other data relevant to this conclu-
sion, see Arcediano, Escobar, & Miller, in press). Moreover, our
results demonstrated competition between subsequent events in-
Figure 2. Mean ratings at test. Filled bars represent the experimental
conditions, and open bars represent the control conditions. Brackets rep-
resent the standard error of the mean. Labels for the conditions represent,
from left to right, causal status (causes vs. effects), directionality of
training: predictive (Pred) versus diagnostic (Diag), and directionality of
testing: predictive (Pred) versus diagnostic (Diag).
233
STIMULUS COMPETITION
Figure 3. Mean responses during training and testing for all causes (C) and effects (E). A: Acquisition of the
relationships between causes and effects. B: The predictive (from causes to effects) and the diagnostic (from
effects to causes) ratings at test. Brackets represent the standard error of the mean.
234
ARCEDIANO, MATUTE, ESCOBAR, AND MILLER
dependent of whether they were causes or effects. Competition
between subsequent stimuli is not consistent with associative ac-
counts of causal judgment based on learning models in which
forward unidirectionality of the associations is assumed (e.g.,
Mackintosh, 1975; Pearce & Hall, 1980; Rescorla & Wagner,
1972). Alternatively, our results partially support the assumptions
of Waldmann and Holyoak’s (1992, 1997) causal model theory,
because we observed competition between causes, independent of
whether they were antecedent or subsequent events (i.e., whether
they were presented with predictive or diagnostic training, respec-
tively). However, contrary to the postulates of the causal model
theory, in which it is assumed that people are sensitive to the
causal structure of the events and known that common causes of an
effect can compete but common effects of a cause cannot, we also
observed competition between effects.
Our results are problematic for the two predominant families of
accounts of causal judgment. Two findings would allow us to
discriminate between these families of accounts: evidence of com-
petition between antecedent effects, which would support the
associative accounts but not the causal model account; and evi-
dence of competition between subsequent causes, which would
support the causal model account but not the associative account.
With respect to competition between antecedent effects, this evi-
dence has not been conclusive (see Figure 1). For example, there
have been reports of competition between antecedent effects (e.g.,
Chapman, 1991; Cobos et al., 2002; Price & Yates, 1993, 1995;
Shanks, 1991; Shanks & Lo´pez, 1996; Waldmann, 2001, Experi-
ment 1; Waldmann & Holyoak, 1992, Experiment 2). However,
there have also been reports of failure to obtain competition
between antecedent effects (e.g., Van Hamme et al., 1993; Wald-
mann & Holyoak, 1992, Experiments 1 and 3; Waldmann, 2000,
2001, Experiment 2). Concerning competition between causes
presented as subsequent events, there is little prior evidence of this
effect. Indeed, Cobos et al. (2002) and Price and Yates (1995) have
reported absence of stimulus competition in this situation. How-
ever, the results of the present experiment show evidence of
competition between subsequent causes. Finally, our most surpris-
ing finding, competition between subsequent effects (i.e., effects
presented with predictive training), is in strong contradiction to
both families of accounts of causal judgment. However, this ob-
servation is consistent with prior evidence of competition in this
situation (e.g., Esmoris-Arranz et al., 1997; Matute et al., 1996;
Miller & Matute, 1998).
In summary, the role of each of the variables investigated in the
present research was not adequately isolated in previous research
(i.e., they were not analyzed independent of the values of the other
potential variables). The present research demonstrated that none
of these variables or their interactions were critical in obtaining
stimulus competition in causal judgment. Failure to report compe-
tition in some conditions of previous studies does not mean that
competition never occurs in these conditions. Instead, it demon-
strates that these effects can be elusive. But stimulus competition
can be elusive even in the most well documented condition, that is,
in the case in which causes are presented as antecedent events
during training and testing. However, these null results, although
well known to researchers, usually are not submitted for publica-
tion. That other studies have failed to obtain competition between,
for example, effects presented as antecedents should not be taken
as strong evidence in favor of any particular theory, because there
are many other studies demonstrating that effects presented as
antecedent events can compete. There are now enough data to
conclude that stimulus competition is not limited to antecedent
stimuli or to situations with common causes competing to predict
an effect, as has been suggested by the associative account and the
causal model account of causal judgment, respectively. Our results
show that competition between stimuli occurred when the target
stimulus was trained in the presence of another stimulus, regard-
less of whether they were causes or effects, of whether they were
trained in the predictive or diagnostic direction, of whether they
were presented in the antecedent or subsequent temporal order, and
of whether they were tested with predictive or diagnostic ques-
tions. Although our results could be limited to an overshadowing
procedure, we expect that they would generalize to other stimulus
competition effects, because prior research has not found over-
shadowing to have properties that are fundamentally different from
blocking or stimulus relative validity.
Table 2
Statistical Comparisons for the Target Stimuli
Conditions F MSE p
Experimental condition 43.05 3104.30 .000001
Causes 38.64 1513.16 .000001
Effects 32.98 2296.62 .000001
Predictive training 31.14 2640.88 .000001
Diagnostic training 25.76 2058.35 .000006
Predictive testing 29.60 2104.91 .000002
Diagnostic testing 40.89 1765.95 .000001
Causes with predictive training 26.18 1578.90 .000005
Causes with predictive training and
predictive testing
17.15 1114.99 .000136
Causes with predictive training and
diagnostic testing
22.73 979.25 .000017
Causes with diagnostic training 17.02 1129.88 .000143
Causes with diagnostic training and
predictive testing
8.70 937.74 .004877
Causes with diagnostic training and
diagnostic testing
14.89 751.58 .000333
Effects with predictive training 22.61 1808.79 .000018
Effects with predictive training and
predictive testing
14.38 1199.14 .000411
Effects with predictive training and
diagnostic testing
20.73 1154.70 .000035
Effects with diagnostic training 21.60 1618.55 .000026
Effects with diagnostic training and
predictive testing
16.88 1149.71 .000151
Effects with diagnostic training and
diagnostic testing
18.08 865.58 .000095
Antecedent stimuli 33.79 2253.58 .000001
Antecedent causes 26.18 1578.90 .000005
Antecedent effects 21.60 1618.55 .000026
Subsequent stimuli 36.73 1582.09 .000001
Subsequent causes 17.02 1129.88 .000143
Subsequent effects 22.61 1808.79 .000018
Note. Planned comparisons conducted from the 2 (experimental condi-
tion: experimental vs. control) 2 (causal status: competing causes vs.
competing effects) 2 (directionality of training: predictive vs. diagnos-
tic) 2 (directionality of test: predictive vs. diagnostic) and 2 (experi-
mental condition: experimental vs. control) 2 (causal status: competing
causes vs. competing effects) 2 (temporal order: antecedent vs. subse-
quent) 2 (directionality of test: predictive vs. diagnostic) analyses of
variance performed on the ratings of the causes and effects in the experi-
ment, Fs(1, 49) 43.05, MSE 3,104, ps .000001.
235
STIMULUS COMPETITION
In our view, during training, events (causes or effects) do not
compete to become associated with the paired event (an effect or
a cause). Instead, we suggest that all antecedent and subsequent
events become associated if they are presented contiguously (e.g.,
the comparator hypothesis; Denniston, Savastano, & Miller, 2001;
Miller & Matzel, 1988). In this view, there is no competition
during training; it is at the moment of testing (i.e., responding)
when, if there is more than one association to the same event, these
associations may compete with one another with respect to being
activated. The observation of competition during responding may
also depend on the test demands. That is, if the test question fosters
competition between stimuli, it will be easier to observe compe-
tition between the associated events. However, if the test question
does not foster competition (e.g., asking about contiguity, cooc-
currences, or frequencies), then stimulus competition may not be
observed (e.g., Matute et al., 1996; Price & Yates, 1995; but see
also Cobos, Can˜o, Lo´pez, Luque, & Almaraz, 2000). In addition,
we suggest that the associations between stimuli (and any potential
competition at testing) may occur in either temporal direction of
the occurring events, either forward (from antecedent to subse-
quent events) or backward (from subsequent to antecedent events;
see, e.g., Arcediano, Escobar, & Miller, 2003, for a demonstration
of the use of both forward and backward associations, and
Esmoris-Arranz et al., 1997, for stimulus competition between
subsequent stimuli with respect to an antecedent one).
The issue that remains is why stimulus competition may be ob-
served under certain conditions and not others (see Figure 1). It is
difficult to determine which variables account for such divergent
results, because the different studies analyzing stimulus competition
have used a broad range of preparations, procedures, cover stories,
and test questions. Most of these studies manipulate more than one
variable, and frequently these variables are confounded. We believe
that the relative difficulty in obtaining certain effects (e.g., competi-
tion between subsequent events) should not be viewed as evidence of
the nonexistence of that effect. Rather, it should be viewed as indic-
ative of the great response flexibility in causal learning, which may
sometimes allow for competition between multiple effects of a cause
and sometimes prevent such competition. In our view, the study of
causal learning is still only beginning, and it would be quite difficult
to develop a formal comprehensive model to account for all the
different aspects of causal learning at this time. It is unfortunate that
very few studies have tried to isolate the impact of specific variables
affecting competition between causes or between effects and that,
even when that isolation has been attempted, the results have some-
times been contradictory (see, e.g., the disparate results of Cobos et
al., 2000; Matute, Arcediano, & Miller, 1996, regarding the impact of
test questions on stimulus competition). Likewise, there are multiple
possible interpretations for the same effect. Here we have favored the
view that causes and effects are subject to the same rules of associa-
tive learning. However, it is possible that causes and effects are
processed differently (cf. Waldmann & Holyoak, 1992), and that the
instances reporting competition between effects reflect, for example,
participants’ responding to whether there is a unique association
between the target effect and the cause rather than regarding the target
as an effect of the cause. In either case, the low ratings of the
cause–target effect relationship in our experimental conditions follow
the pattern of stimulus competition and suggest that causal inference
is sensitive to stimulus competition in situations with multiple
outcomes.
To conclude, we suggest there are both training variables and
testing variables that collectively could explain the disparity of the
reported results in this literature. Thus, some training variables
(e.g., cover stories, nature of the stimuli, training procedures, and
parameters) and some testing variables (e.g., wording that fosters
competition and directionality of the events) seem to promote, on
the one hand, different degrees of stimulus competition or its
absence, depending on their manipulations, and on the other hand,
different degrees in the perceived causality, which could affect the
directional use of the information. Further research is needed to
analyze the impact on causal learning of the variables at play.
References
Arcediano, F., Escobar, M., & Miller, R. R. (2003). Temporal integration
and temporal backward associations in human and nonhuman subjects.
Learning and Behavior, 31, 242–256.
Arcediano, F., Escobar, M., & Miller, R. R. (in press). Bidirectional
associations in humans and rats. Journal of Experimental Psychology:
Animal Behavior Processes.
Baker, A. G., Mercier, P., Valle´e-Tourangeau, F., Frank, R., & Pan, M. (1993).
Selective associations and causality judgments: Presence of a strong causal
factor may reduce judgments of a weaker one. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 19, 414 432.
Chapman, G. B. (1991). Trial order affects cue interaction in contingency
judgment. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 17, 837– 854.
Chapman, G. B., & Robbins, S. J. (1990). Cue interaction in human
contingency judgment. Memory & Cognition, 18, 537–545.
Cobos, P. L., Can˜o, A., Lo´pez, F. J., Luque, J. L., & Almaraz, J. (2000).
Does the type of judgment required modulate cue competition? The
Quarterly Journal of Experimental Psychology: Comparative & Physi-
ological Psychology, 53B, 193–207.
Cobos, P. L., Lo´pez, F. J., Can˜o, A., Almaraz, J., & Shanks, D. R. (2002).
Mechanisms of predictive and diagnostic causal induction. Journal of
Experimental Psychology: Animal Behavior Processes, 28, 331–346.
Denniston, J. C., Savastano, H. I., & Miller, R. R. (2001). The extended
comparator hypothesis: Learning by contiguity, responding by relative
strength. In R. R. Mowrer & S. B. Klein (Eds.), Handbook of contem-
porary learning theories (pp. 65–117). Mahwah, NJ: Erlbaum.
Dickinson, A., Shanks, D., & Evenden, J. (1984). Judgment of act-outcome
contingency: The role of selective attribution. Quarterly Journal of Exper-
imental Psychology: Human Experimental Psychology, 36A, 29 –50.
Esmoris-Arranz, F. J., Miller, R. R., & Matute, H. (1997). Blocking of
subsequent and antecedent events. Journal of Experimental Psychology:
Animal Behavior Processes, 23, 145–156.
Kamin, L. J. (1968). “Attention-like” processes in classical conditioning. In
M. R. Jones (Ed.), Miami symposium on the prediction of behavior:
Aversive stimulation (pp. 9 –31). Miami, FL: University of Miami Press.
Mackintosh, N. J. (1975). A theory of attention: Variations in the associability
of stimuli with reinforcement. Psychological Review, 82, 276 –298.
Matute, H., Arcediano, F., & Miller, R. R. (1996). Test question modulates
cue competition between causes and between effects. Journal of Exper-
imental Psychology: Learning, Memory, and Cognition, 22, 182–196.
Miller, R. R., & Matute, H. (1998). Competition between outcomes.
Psychological Science, 9, 146 –149.
Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis: A
response rule for the expression of associations. In G. H. Bower (Ed.),
The psychology of learning and motivation, Vol. 22 (pp. 51–92). San
Diego, CA: Academic Press.
Pavlov, I. P. (1927). Conditioned reflexes. Oxford, England: Oxford Uni-
versity Press.
Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: Varia-
236
ARCEDIANO, MATUTE, ESCOBAR, AND MILLER
tions in the effectiveness of conditioned but not of unconditioned stim-
uli. Psychological Review, 87, 532–552.
Price, P. C., & Yates, J. F. (1993). Judgmental overshadowing: Further
evidence of cue interaction in contingency judgment. Memory & Cog-
nition, 21, 561–572.
Price, P. C., & Yates, J. F. (1995). Associative and rule-based accounts of
cue interaction in contingency judgment. Journal of Experimental Psy-
chology: Learning, Memory, and Cognition, 21, 1639–1655.
Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning:
Variations in the effectiveness of reinforcement and nonreinforcement. In
A. H. Black & W. F. Prokasy (Eds.), Classical conditioning II: Current
research and theory (pp. 64 –99). New York: Appleton-Century-Crofts.
Shanks, D. R. (1985). Forward and backward blocking in human contingency
judgment. Quarterly Journal of Experimental Psychology, 37B, 1–21.
Shanks, D. R. (1991). Categorization by a connectionist network. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 17, 433– 443.
Shanks, D. R., & Lo´pez, F. J. (1996). Causal order does not affect cue selection
in human associative learning. Memory & Cognition, 24, 511–522.
Van Hamme, L. J., Kao, S.-F., & Wasserman, E. A. (1993). Judging
interevent relations: From cause to effect and from effect to cause.
Memory & Cognition, 21, 802– 808.
Wagner, A. R., Logan, F. A., Haberlandt, K., & Price, T. (1968). Stimulus
selection and a “modified continuity theory.” Journal of Experimental
Psychology, 76, 171–180.
Waldmann, M. R. (2000). Competition among causes but not effects in
predictive and diagnostic learning. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 26, 53–76.
Waldmann, M. R. (2001). Predictive versus diagnostic learning: Evidence from an
overshadowing paradigm. Psychonomic Bulletin & Review, 8, 600 608.
Waldmann, M. R., & Holyoak, K. J. (1992). Predictive and diagnostic
learning within causal models: Asymmetries in cue competition. Journal
of Experimental Psychology: General, 121, 222–236.
Waldmann, M. R., & Holyoak, K. J. (1997). Determining whether causal
order affects cue selection in human contingency learning: Comments on
Shanks and Lopez (1996). Memory & Cognition, 25, 125–134.
Wasserman, E. A. (1974). Stimulus-reinforcer predictiveness and selective
discrimination learning in pigeons. Journal of Experimental Psychology,
103, 284–297.
Wasserman, E. A. (1990). Attribution of causality to common and distinc-
tive elements of compound stimuli. Psychological Science, 1, 298 –302.
Received July 31, 2003
Revision received August 13, 2004
Accepted August 20, 2004
237
STIMULUS COMPETITION
... Second, and more generally, one common way of explaining nongraduality in learning from the associative perspective consists of adding response rules that work on the output of the associative mechanism to obtain nonlinear patterns (e.g., step-like functions, that is, the response might not occur until the associative strength reaches a certain threshold). The most notable example of this type of rules is the comparator hypothesis (Arcediano, Escobar, & Miller, 2005;Miller & Matzel, 1988;Stout & Miller, 2007). According to this theory, the association between the target cue and the outcome is compared with the association between the outcome and other cues that are present (e.g., contextual cues) at the time of responding. ...
... This result challenges the assumptions of associative theories (that is, sensitivity to this manipulation seems to be another error of omission), though it is also true that several variables were confounded in Waldmann and Holyoak's (1992) experiments, and therefore it was impossible to know which one was actually responsible for the observed effects (see, e.g., Matute et al., 1996;Shanks & López, 1996). For example, in a subsequent experiment, Arcediano, Matute, Escobar, and Miller (2005) separated the effect of several variables that had often been confounded in this research area: directionality of training (predictive vs. diagnostic), directionality of testing (predictive vs. diagnostic), causal status of competing effects (causes or effects), and temporal order of competing stimuli (antecedent events vs. subsequent events) relative to the noncompeting stimuli. Their results suggested that cue competition could be observed in all cases, which gave further support to the associative proposal. ...
Article
Many experiments have shown that humans and other animals can detect contingency between events accurately. This learning is used to make predictions and to infer causal relationships, both of which are critical for survival. Under certain conditions, however, people tend to overestimate a null contingency. We argue that a successful theory of contingency learning should explain both results. The main purpose of the present review is to assess whether cue-outcome associations might provide the common underlying mechanism that would allow us to explain both accurate and biased contingency learning. In addition, we discuss whether associations can also account for causal learning. After providing a brief description on both accurate and biased contingency judgments, we elaborate on the main predictions of associative models and describe some supporting evidence. Then, we discuss a number of findings in the literature that, although conducted with a different purpose and in different areas of research, can also be regarded as supportive of the associative framework. Finally, we discuss some problems with the associative view and discuss some alternative proposals as well as some of the areas of current debate. (PsycINFO Database Record (c) 2019 APA, all rights reserved).
... In the second case, when the previous belief is that A prevents the outcome from occurring, the detection of a causal relationship between the second cause B and the outcome will be facilitated (i.e., this particular case of cue interaction is generally known as superconditioning; Rescorla, 1971). Many cue interaction experiments, both with animals and humans, show that learning about the relationship between a potential cause and an outcome can result altered when the potential cause is presented in compound with another potential cause that has been previously associated either with the outcome or its absence (Aitken, Larkin, & Dickinson, 2000;Arcediano, Matute, Escobar, & Miller, 2005;Dickinson, Shanks, & Evenden, 1984;Kamin, 1968;Luque, Flores, & Vadillo, 2013;Luque & Vadillo, 2011;Mor ıs, Cobos, Luque, & L opez, 2014;Rescorla, 1971;Shanks, 1985). ...
... Nevertheless, what is clear given the present results is that the group who developed a high illusion about the ineffective Medicine A tended to assess the effective Medicine B as less effective than the group who developed a weaker illusion. Similar cue interaction effects have been clearly established in other causal learning research (Aitken et al., 2000;Arcediano et al., 2005;Dickinson et al., 1984;Luque & Vadillo, 2011;Luque et al., 2013;Mor ıs et al., 2014;Shanks, 1985). Moreover, cue interaction effects are beautifully predicted by current theories of learning (Rescorla & Wagner, 1972). ...
Article
Full-text available
Cognitive illusions are often associated with mental health and well-being. However, they are not without risk. This research shows they can interfere with the acquisition of evidence-based knowledge. During the first phase of the experiment, one group of participants was induced to develop a strong illusion that a placebo medicine was effective to treat a fictitious disease, whereas another group was induced to develop a weak illusion. Then, in Phase 2, both groups observed fictitious patients who always took the bogus treatment simultaneously with a second treatment which was effective. Our results showed that the group who developed the strong illusion about the effectiveness of the bogus treatment during Phase 1 had more difficulties in learning during Phase 2 that the added treatment was effective.
... However, other studies have failed to find this asymmetry (e.g. Arcediano et al., 2005;Shanks & Lopez, 1996), and several researchers have suggested causal model effects are dependent on particular task conditions. For example, one study found that causal model effects were only evident when the relevance of the causal structure was specifically emphasised to participants (Lopez, Cobos and Caño, 2005). ...
Conference Paper
Full-text available
Cue competition effects in human contingency learning appear to be sensitive to the causal nature of cue-outcome relationships. While blocking effects are reliably demonstrated in scenarios where cues are presented as causes of outcomes, several studies have failed to find blocking in scenarios where cues are presented as effects of outcomes, a finding that is typically taken as evidence for the involvement of controlled reasoning processes in cue competition. These studies typically measure blocking with continuous causal ratings about individual cues. Previous studies have found that sensitivity to causal model may depend on how the test question is phrased. In contrast, the current study tested the sensitivity of blocking to causal scenarios across different formats of the same test question. Participants completed a causal learning task with instructions suggesting either a predictive (i.e. cue causes outcome) or diagnostic (cue is caused by outcome) cue-outcome relationship. Participants were then asked about the likelihood of outcomes occurring by either giving a continuous rating of each outcome or a discrete choice about the most likely outcome. When measured by continuous ratings of individual cues, blocking was evident in predictive, but not diagnostic scenarios. However, when measured by discrete choice or using a compound negation test, blocking was robust and insensitive to causal scenario. The results suggest that contributions of predictive memory and causal reasoning to cue competition effects may depend substantially on the type of measure used.
... It has previously been demonstrated that the causal scenario employed can affect the results found in cue competition experiments (e.g. Waldmann & Holyoak, 1992), although some phenomena seem relatively insensitive to the choice of causal scenario (Arcediano, Matute, Escobar & Miller, 2005). It is an open question whether the current results would generalize to a procedure employing abstract stimuli in a causally neutral scenario. ...
Article
Full-text available
Demonstrations of retrospective revaluation suggest that remembered stimuli undergo a reduction in association with the unconditioned stimulus (US) present during learning. Conversely, demonstrations of mediated conditioning in flavor-conditioning experiments with rats suggest that remembered stimuli undergo an increase in association with the US present during learning. In a food allergy prediction task with 23 undergraduates, we demonstrated simultaneous backward conditioned inhibition and mediated conditioning effects. These results are compatible with the hypothesis that the direction of change (decrease or increase) in associative strength depends on whether the remembered stimulus was of a different category (conditioned stimulus/antecedent) or the same category (US/outcome) as the presented US.
... In a diagnostic task, the presentation of the causal information is reversed, so that participants are shown an effect (e.g., symptom) and have to predict the occurrence of the cause (e.g., disease). Although many researchers have found no differences between predictive and diagnostic scenarios (e.g., Arcediano, Matute, Escobar, & Miller, 2005; Baker, Murphy, Mehta, & Baetu, 2005; Cobos, López, Caño, Almaraz, & Shanks, 2002; Matute, Arcediano, & Miller, 1996; Price & Yates, 1995; Shanks & López, 1996 ), others have found negative mediation with predictive scenarios but not with diagnostic scenarios (e.g., Tangen & Allan, 2004; Waldmann, 2000; Waldmann & Holyoak, 1992). Thus, regardless of the symptoms– disease task being understood as a structural or as a diagnostic task, the current evidence suggests that it is more likely to obtain negative mediation with the food – allergy task; the symptoms – disease task might be more likely to yield either no effect (if it is understood as a diagnostic task) or positive mediation (if it is understood as a structural task). ...
Article
Full-text available
After presenting two cues, A and B, together, later pairings of one of the cues alone with an outcome can generate changes in the associative value of the absent cue. These changes can be in the same direction as the present cue (i.e., positive mediation) or in the opposite direction to the present cue (i.e., negative mediation). We found both mediational effects in a human contingency task. In addition, we found that the direction of the change was determined by the existence of a prior association between the absent cue and the outcome. When a prior association exists, the absent cue tends to change its value in the opposite direction to the present cue, whereas when there is no prior association, the absent cue tends to change its value in the same direction as the present cue. Recent associative models (Stout & Miller, 2007) can explain our results.
... For example, if beliefs are based on post hoc reasoning about events stored in memory, changing the temporal order in which these events are stored should not affect judgments; and yet several studies have demonstrated trial-order effects in learning (Collins & Shanks 2002;Dickinson & Burke 1996;Ló pez et al. 1998b). Similarly, the evidence relating to differences in diagnostic and predictive learning is overstated: many studies have found no impact of this manipulation (e.g., Arcediano et al. 2005;Cobos et al. 2002;Ló pez et al. 2005;Tangen et al. 2005). Nor is mention made of cancer patients who form food aversions despite being fully aware that chemotherapy treatment, rather than any food consumed, produced their illness (e.g., Bernstein & Webster 1980). ...
Article
Full-text available
Contrary to the target article's claims, social cognition research shows considerable learning (about other people) that is relatively automatic. Some of this learning is propositional (spontaneous trait inferences) and some is associative (spontaneous trait transference). Other dichotomies - for example, between learning explicit and implicit attitudes - are also important. However conceived, human conditioning is not synonymous with human learning.
Article
Two experiments were conducted using a picture–picture procedure to investigate the functional characteristics of evaluative conditioning. Images of foods were used as cue stimuli while images of body shapes were used as reinforcing stimuli. Foods paired with obese body shapes were rated more negatively than foods that had been paired with normal body shapes and the size of this evaluative conditioning effect did not differ between men and women. In Experiment 1 repeated presentation of the foods alone after training and before test did not reduce the size of the evaluative conditioning effect. In Experiment 2 there was no difference in the size of the evaluative conditioning effect between cues that were trained alone or in compound with another cue. The absence of extinction and overshadowing accords with the idea that in some cases evaluative conditioning is based on a learning mechanism sensitive only to the co-occurrence of stimuli and not to the contingency between them.
Article
The extent to which human learning should be thought of in terms of elementary, automatic versus controlled, cognitive processes is unresolved after nearly a century of often fierce debate. Mitchell et al. provide a persuasive review of evidence against automatic, unconscious links. Indeed, unconscious processes seem to play a negligible role in any form of learning, not just in Pavlovian conditioning. But a modern connectionist framework, in which "cognitive" phenomena are emergent properties, is likely to offer a fuller account of human learning than the propositional framework Mitchell et al. propose.
Article
We are big fans of propositions. But we are not big fans of the "propositional approach" proposed by Mitchell et al. The authors ignore the critical role played by implicit, non-inferential processes in biological cognition, overestimate the work that propositions alone can do, and gloss over substantial differences in how different kinds of animals and different kinds of cognitive processes approximate propositional representations.
Article
Ciiven the task of di the source of a patient's aUer^'ic reav-tion. college students jiuigcii the causal efficacy of common (A') and distinctive (A and Bj elements of compound stimuli: AX and BX. As the differential correlation of AX and BX with the occurrence and nonoccurrence ofthe allergic reaction rose from .00 to 1.00. ratings of ihe distinctive A and B elements diverged; most importantly, ratings ofthe common X element fell. These causal judgments of humans closely parallel the conditioned responses of animals in associa-tive learning studies, and clearly disclose that stimuli compete with one another for control over behavior.
Article
Three experiments tested a simple connectionist network approach to human categorization. The specific network considered consists of a layer of input nodes, each representing a feature of the exemplar to be categorized, connected in parallel to a layer of output nodes representing the categories. Learning to categorize exemplars consists of adjusting the weights in the network so as to increase the probability of making correct categorizations; weight changes are determined by the Rescorla-Wagner (1972) learning rule. The experiments used a simulated medical diagnosis procedure in which subjects have to decide which disease (the category) each of a series of patients is suffering from on the basis of the patients' symptoms (the features). After a series of trials, the subjects rated the extent to which particular symptoms were associated with particular diseases. In each of the experiments, it is shown that a process of selective learning occurs in this categorization task and that selection in turn depends on the relative predictiveness of the symptom for the disease. Such effects parallel results found in animal conditioning experiments and are readily reproduced by the connectionist network model. The results are also discussed in terms of a variety of traditional theories of categorization.
Chapter
This chapter describes the potential explanatory power of a specific response rule and its implications for models of acquisition. This response rule is called the “comparator hypothesis.” It was originally inspired by Rescorla's contingency theory. Rescorla noted that if the number and frequency of conditioned stimulus–unconditioned stimulus (CS–US) pairings are held constant, unsignaled presentations of the US during training attenuate conditioned responding. This observation complemented the long recognized fact that the delivery of nonreinforced presentations of the CS during training also attenuates conditioned responding. The symmetry of the two findings prompted Rescorla to propose that during training, subjects inferred both the probability of the US in the presence of the CS and the probability of the US in the absence of the CS and they then established a CS–US association based upon a comparison of these quantities. The comparator hypothesis is a qualitative response rule, which, in principle, can complement any model of acquisition.
Article
In the first experiment subjects were presented with a number of sets of trials on each of which they could perform a particular action and observe the occurrence of an outcome in the context of a video game. The contingency between the action and outcome was varied across the different sets of trials. When required to judge the effectiveness of the action in controlling the outcome during a set of trials, subjects assigned positive ratings for a positive contingency and negative ratings for a negative contingency. Furthermore, the magnitude of the ratings was related systematically to the strength of the actual contingency. With a fixed probability of an outcome given the action, judgements of positive contingencies decreased as the likelihood that the outcome would occur without the action was raised. Correspondingly, the absolute value of ratings of negative contingencies was increased both by an increment in the probability of the outcome in the absence of the action and by a decrement in the probability of the outcome following the action. A systematic bias was observed, however, in that positive judgements were given under a non-contingent relationship when the outcome frequency was relatively high. However, this bias could be reduced by giving extended exposure to the non-contingent schedule (Experiment 2). This pattern of contingency judgements can be explained if it is assumed that a process of selective attribution operates, whereby people are less likely to attribute an outcome to some potential target cause if another effective cause is present. Experiments 2 and 3 demonstrated the operation of this process by showing that initially establishing another agent as an effective cause of the outcome subsequently reduced or blocked the extent to which the subjects attributed the outcome to the action. Finally, we argue that the pattern and bias in contingency judgements based upon interactions with a causal process can be explained in terms of contemporary conditioning models of associative learning.
Article
In both Pavlovian conditioning and human causal judg- ment, competition between cues is well known to occur when multiple cues are presented in compound and followed by an outcome. More questionable is the occurrence of competition between outcomes when a single cue is followed by multiple outcomes presented in compound. In the experiment reported here, we demonstrated blocking (a type of stimulus competition) between outcomes. When the cue predicted one outcome, its ability to predict a second outcome that was presented in compound with the first outcome was reduced. The procedure mini- mized the likelihood that the observed competition between outcomes arose from selective attention. The competition between outcomes that we observed is problematic for contemporary theories of learning.
Article
In 5 experiments, humans played video games in which 2 events or causes covaried with an outcome. In Experiments 1 and 2, a highly correlated cause (a plane) of an outcome (success at traversing a minefield) reduced judgments of the strength of a weaker cause (camouflaging or painting a tank). In Experiment 3, similar results were found when both causes were negatively correlated with the outcome. In Experiment 4, strong positive or negative contingencies caused the subjects to reduce judgments of contingencies of the opposite polarity. These results can be accounted for by associative or connectionist models from animal learning such as the Rescorla-Wagner model. In Experiment 5, this type of model was contrasted with a representational model in which subjects are claimed to monitor accurately the various contingencies but use a rule in which the presence of a strong contingency causes them to discount weaker contingencies.
Article
Three experiments investigated whether a process akin to L. J. Kamin's (1969) blocking effect would occur with human contingency judgments in the context of a video game. 102 students were presented with sets of trials on each of which they could perform a particular action and observe whether the action produced a particular outcome in a situation in which there was an alternative potential cause of the outcome. Exp I showed that prior observation of the relationship between the alternative cause and the outcome did indeed block or reduce learning about the subsequent action-outcome relationship. However, exposure to the relationship between the alternative cause and the outcome after observing the association between the action and the outcome also reduced judgments of the action-outcome contingency (backward blocking), a finding at variance with conditioning theory. In Exp II, it was found that the degree of backward blocking depended on the predictive value of the alternative cause. Finally, Exp III showed that the backward blocking effect was not the result of greater forgetting about the action-outcome relationship in the experimental than in the control condition. Results cast doubt upon the applicability of contemporary theories of conditioning to human contingency judgment.