The Psychological Record , 1998, 48, 591-600
AGGRESSIVE RESPONDING IN THE LABORATORY MAINTAINED
BY THE INITIATION OF A PROVOCATION-FREE INTERVAL
DONALD M. DOUGHERTY, DON R. CHEREK, and SCOTT D. LANE
University of Texas-Houston Health Science Center
Four adults participated in an experiment designed to study
how the duration of an interval free from provocation , following
aggressive responding, affected the maintenance of such
responding . In testing sessions, two response options were made
available on separate buttons labeled "A" and "B." Pressing
Button A 100 times incremented a point accumulation counter,
and these pOints were later exchanged for money (one point = 10
cents). Pressing Button B 10 times subtracted a pOint from a
fictitious other subject. These B button responses were defined
as aggressive because they ostensibly presented an aversive
stimulus (i.e., point subtraction) to another person. Subjects were
provoked by subtracting pOints from their pOint counter at random
intervals (ranging from 6 s to 120 s) , and instructions attributed
these point subtractions to the responding of the other subject.
Button B responding could initiate periods free from provocation.
Subjects were exposed to two provocation-free interval durations
(5 sand 62 .5 s) in order to determine what role the provocationfree interval had in maintaining aggressive responding. We found
that brief (5 s) provocation-free intervals (a) can initially engender
high rates of aggressive responding ; but (b) cannot maintain
aggressive responding during extended periods of exposure.
These findings demonstrate the importance of the consequences
of aggressive responding in determining the probability of an
aggressive response following provocation .
I n both humans and nonhumans aggressive and/or escape
responses occur at increased probability follow ing aversive stimulus
presentation (Azrin , Hutchinson , & Hake, 1967; Cherek & Dougherty,
1997). Furthermore, responding occasioned by the presentation of
aversive stimuli typically occurs at a rate that is directly related to the
frequency of aversive stimulus presentation. Earlier studies with
nonhuman subjects have shown that the frequency of fighting among
This research was supported by Grant DA-03166 from the National Institute on Drug
Abuse . Reprint requests can be obtained from Don R. Cherek, Dept. of Psychiatry &
Behavioral Sciences, University of Texas-Houston , Health Science Center, 1300 Moursund
Street, Houston, TX 77030.
592
DOUGHERTY ET AL.
pairs of rats (Ulrich & Azrin, 1962) and aggressive biting in squirrel
monkeys (Azrin, Hutchinson, & McLaughlin, 1965) was directly related to
the frequency of electric shock presentation. Our recent study with
human subjects (Cherek & Dougherty, 1997) has shown that the
frequency of aggressive responding made toward a fictitious subject was
directly related to the frequency of provocation (i.e., point subtractions)
by a fictitious opponent. Together, these studies have demonstrated that
the frequency of provocation is an important determinant of aggressive
response rates.
Although provocation is sufficient to generate aggressive
responding, the probability that aggressive responding is maintained is
largely dependent upon the consequences. A number of previous human
laboratory studies have examined the role in which various
contingencies affect aggressive responses. For example, previous
studies have shown that aggressive responses decrease when the
response requirement for an aggressive response is increased (Cherek,
Spiga, & Egli, 1992), and higher rates of aggressive responding are
maintained by an avoidance contingency rather than an escape
contingency (Cherek, Spiga, Steinberg, & Kelly, 1990).
In the present study, using the © Point Subtraction Aggression
Paradigm (Cherek, 1981), we examined the role a contingency attached
to aggressive responding, a temporary escape (or time-out) from
provocation, plays in the maintenance of aggressive responding. In a
recent study (Cherek & Dougherty, 1997), while examining the effect of a
wide range of provocation frequencies (point subtractions) on rates of
aggressive responding, we found aggressive responding varied as a
function of provocation frequency. Interestingly, this relationship held true
even during conditions of high provocation, when aggressive responding
initiated only brief periods of escape from further provocation. Here we
examined how the duration of the provocation-free interval (PFI), initiated
by the subject's aggressive responding, affected rates of aggressive
responding during extended exposure to a particular PFI value (5 s or
62.5 s). The PFI duration was manipulated in order to determine the
importance of the PFI contingency in the maintenance of aggressive
responding.
Method
Subjects
Four subjects, three males and one female, between the ages of 20
and 34 participated after informed consent was obtained. One subject
(S-1204) had previously participated in research in our laboratory and
had been exposed to 155 sessions of the same experimental paradigm.
Apparatus
During experimental sessions, subjects sat in a 1.2-m by 1.8-m
sound-attenuated chamber which was illuminated overhead by a 75-W
light bulb. Masking noise was provided by a ventilation fan. Located
HUMAN AGGRESSIVE RESPONDING
593
within the chamber were an Apple lie monochrome monitor and a
customized response panel (10.0 cm x 43.0 cm x 25.0 cm). The
response panel contained three push buttons (MicroSwitch) labeled "A,"
"B," and "C" (Button C was not used) mounted in a straight line 10.0 cm
apart. Both the monitor and response panel were controlled by an Apple
IIGS computer equipped with an Applied Engineering I/O 32 interfacing
system located in an adjacent room.
Procedure
Recruitment and screenings. Potential subjects were recruited
through advertisements for behavioral research placed in the classified
employment section of local newspapers. Subjects were excluded if
either a psychiatric disorder, including substance abuse, or medical
illness was detected during screening exams. The psychiatric screening
consisted of a mental status exam and the Schedule for Affective
Disorders and Schizophrenia-Lifetime Version (SADS-L), a standardized
psychiatric interview (Endicott & Spitzer, 1978). A medical questionnaire,
administered by the interviewer was used to determine if any significant
medical conditions existed that would preclude subject participation.
Drug usage was monitored by collecting urine and expired air
samples daily each morning shortly after each subject's arrival (8:00
a.m.). A urine drug screen (Enzyme Multiple Immunoassay Technique Drug Abuse Urine Assay by SYLVA Corporation, Palo Alto, CA), capable
of detecting the presence of illicit and therapeutic compounds or their
metabolites, was performed on daily samples. Recent alcohol intake was
monitored by collecting an expired air sample using an Intoximeter
Model 3000-111 (Intoximeter, Inc., St. Louis, MO).
As part of the screening and recruitment process information regarding
potential earnings, urine drug testing, breath alcohol testing, and the
psychiatric and physical exams was provided. Subjects were told that they
could earn approximately $4-$5 per session through point accumulation
(paid daily) and would be paid a $10 per day bonus at the completion of
their participation for maintaining drug-free urines while in the study. The
study's purpose was described as a study of motor performance.
Instructions. Subjects were shown a diagram of the testing
apparatus and read a standard set of instructions describing the
experimental paradigm prior to the subject's first session of participation.
In these instructions the two response options were briefly described.
These instructions appear verbatim elsewhere (Cherek & Dougherty,
1997) and informed subjects that they would be paired with another
person who might take away points worth money during the course of
the experimental sessions.
Point Subtraction Aggression Paradigm (PSAP). During
experimental sessions subjects were tested using the PSAP which
makes two response options available to the subject: (1) Completing a
fixed-ratio (FR) 100 on Button A was reinforced by adding one point
(worth 10 cents) to the subject's point-accumulation counter, displayed
594
DOUGHERTY ET AL.
on a computer monitor; and (2) completing a FR 10 on Button B
ostensibly subtracted a point from a fictitious other subject. This
responding was defined as aggressive because it ostensibly resulted in
the presentation of an aversive stimulus to another person. Once either
response button was chosen it was necessary to complete its ratio
requirement before a different option could be chosen. Subjects were
provoked by the PSAP program subtracting points from their counter at
randomly selected intervals between 6 sand 120 s. These point
subtractions were attributed to the fictitious partner in the instructions
and occurred while the subject was in either response option.
Besides ostensibly resulting in point subtractions, completing a FR
10 on Button B could also initiate an interval free from point subtractions
(provocation-free intervals). A provocation-free interval (PFI) could be
initiated only after at least one point had been subtracted from the
subject's counter. Completing Button B's response requirement either
before a point had been subtracted, or after a PFI had already been
initiated, had no programmed consequence but these responses were
included in the data analysis. In the absence of PFI initiation, points
were subtracted at intervals ranging from 6 s to 120 s; all intervals were
equally represented. Point subtractions were scheduled to occur 6 s to
120 s after a PFI had timed out. Because a PFI could be initiated only by
the subject following at least one point loss, not all point subtractions
could be avoided. However, the total number of point subtractions could
be reduced by Button B responding, because B responding immediately
following each point loss would result in fewer total point subtractions in
a given session.
Provocation-free interval manipulations. Subjects participated in six
sessions per day Monday through Friday. Twenty-five minute sessions
were conducted at 9:00, 10:00 and 11 :00 a.m., and 1:00, 2:00, and 3:00
p.m., and lunch was provided at noon. In this experiment, we alternated
duration of the PFI between two values. All subjects began in a condition
where the PFI was equal to 62.5 s, and remained under this condition for
a minimum of 10 sessions. After 10 sessions, or after B button
responding stabilized, the PFI interval was decreased to 5 s. At the 5-s
PFI essentially no time-out from point loss was provided. The reason for
including 5 s instead of 0 s was to decrease the likelihood that Button B
responses would be superstitiously suppressed (punished) by a point
subtraction. Subjects remained in the 5-s PFI condition until responding
either stabilized or extinction occurred. In the next phase, subjects were
returned to the PFI 62.5-s condition. Then, in the final phase, the PFI 5-s
condition was reinstated. Data were analyzed by measuring B
responding across all sessions completed during the study.
Evaluation of instructional deception. In order to assess whether or
not the instructional deception regarding the other fictitious subjects had
been established and maintained in these procedures subjects
completed a questionnaire at the end of the day which asked them (1) to
estimate the number of subjects they had been paired with that day, (2)
HUMAN AGGRESSIVE RESPONDING
595
to describe these other subjects, and (3) to estimate whether they or the
other subjects subtracted more points, Any report on the questionnaire
indicating the subject did not believe that he or she had interacted with
another individual would result in removal from the study, However, all
30
8-1204
25
PFI=62.5 s
PFI=5 s
PFI =
62.5
20
PFI=5 s
15
10
5
0
45
40
'S 35
,~ 30
~ 25
'- 20
Q)
a.. 15
en 10
Q)
en 5
c: 0
0
8-1267
PFI=62.5
Q)
PFI=5 s
PFI=62.5 s
c..
~ 30
a:
8-1269
PFI=62.5 s
25
PFI=5 s
PFI=62.5 s
PFI=5 s
Q)
>
'eenn
Q)
'C')
C')
«
20
15
10
5
0
30
8-1287
25
f/)
l!)
20
PFI=62.5 s
II
u::
PFI=62.5 s
PFI=5 s
a..
15
10
5
0
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
Session
Figure 1. Rates of aggressive responding under two PFI duration conditions: 62 .5 sand 5
s; each condition was repeated.
596
DOUGHERTY ET AL.
subjects' reports endorsed interaction with other individuals during
experimental sessions.
Results
All subjects reported having been paired with one or more subjects
during each day of testing. None of the subjects were removed for
alcohol or drug use.
The dependent variable of primary interest was the rate of
aggressive responding emitted on Button B. Rates of responding on this
button under the two provocation conditions (PFI = 62.5 sand PFI = 5 s)
are depicted in Figure 1. Initially, Button B responses were maintained
by the 62.5-s PFI for all subjects, and when the 5-s PFI was introduced
Button B responding extinguished. For three of the four subjects, the
transition from the longer PFI (62.5 s) to the shorter PFI (5 s) resulted in
an initial increment in aggressive responding. Except for Subject S-1267,
baseline levels of Button B responding were recaptured after extinction
when the PFI of 62 .5 s was reintroduced and then again extinguished
when the PFI of 5 s was reintroduced. Subject S-1267's Button B
responding never recovered when the PFI of 62.5 s was reintroduced,
after the initial extinction. Subject S-1204, having the extensive history
with this schedule, took considerably more sessions to extinguish than
did any of the other subjects.
Rates of responding on Button A (point-maintained responding)
remained relatively stable across both PFI conditions. Rates of
responding under the 62.5 condition averaged 4.53 (SEM = 0.03) and
4.62 (SEM = 0.03) responses per second in the 62.5 and 5 s PFI
conditions, respectively.
Discussion
The purpose of this experiment was to examine the relationship
between the provocation-free interval, which provides a period of escape
from an aversive stimulus (point subtractions), and the rates of
aggressive responding. The initiation of very brief periods of time-out
from provocation (PFls) did not maintain aggressive (Button B)
responding over several sessions. This finding is meaningful because it
demonstrates that escape from provocation, or other aversive
stimulation, is an important variable in maintaining aggressive behavior.
Generally, the results of this experiment further the notion that both
aversive stimulation (provocation) and escape from subsequent
provocation (PFI) are important variables in the facilitation and
maintenance of aggressive responding in the laboratory. The delivery of
an aversive stimulus, for example, shock, loud noise, or point
subtraction, is necessary for engendering aggressive responses; and a
period free from provocation (an escape contingency) is necessary for
maintaining aggressive responses. Subjects will not typically emit
HUMAN AGGRESSIVE RESPONDING
597
aggressive responses in the absence of provocation (Azrin, Hutchinson,
& McLaughlin, 1965; Kelly, Cherek, Steinberg, & Robinson, 1988), and
escape from an aversive stimulus (e.g., provocation) appears to be a
critical consequence responsible for maintaining aggressive responding
in the laboratory (Dengerink & Covey, 1983). This is consistent with
preliminary findings described in another study where, in the absence of
periods of escape from further point subtractions, aggressive responding
extinguished (Cherek, Spiga, Steinberg, & Kelly, 1990). This escape
contingency between aggressive responding and provocation is similar
to contingencies that exist in the natural environment.
One of the most interesting findings was that for most subjects
aggressive responding initially increased (before subsequently
extinguishing) when transitions were made from the longer to the shorter
PFI values. These increases were sometimes very large. Such increases
are similar to the bursts of responding observed when conditions of
extinction are introduced following an established reinforcement history,
or when abrupt transitions are made from dense to sparse reinforcement
schedules (Ferster & Skinner, 1957).
In the vast literature on aggression, the consequences of aggressive
behavior as a factor in maintaining and determining the rate of that
behavior are seldom considered (see Baron & Richardson, 1994, cf.
Cherek & Steinberg, 1987). Relatedly, most current experimental
preparations with humans and nonhumans focus on the role of evocative
or eliciting stimuli rather than the consequences of aggressive behavior
(Miczek & Winslow, 1987; Taylor & Chermack, 1993). In such
experiments, investigators often arrange environmental conditions that
establish favorable consequences for aggressive responding. Typically,
experiments carefully control the consequences of aggressive behavior,
for example by using inanimate target objects, both with nonhuman and
human subjects (Azrin, Hutchinson, & Sallery, 1964; Bandura, Ross, &
Ross, 1963). Also, in the resident intruder paradigm (see Brain, Benton,
Howell, & Jones, 1982) the intruder is commonly defeated. Thus,
although not emphasized, the consequences for aggressive responding
probably play an implicit but important role in many laboratory studies of
aggression. Germane to the present experiment, data from several
nonhuman studies support the importance of reinforcing consequences
in maintaining aggressive responding (Cherek & Pickens, 1970;
Hutchinson, Azrin, & Hunt, 1968; Scott, 1975). Our data are
complementary to those nonhuman studies and support similar findings
with human subjects (Dengerink & Covey, 1983).
One could classify aggressive behavior in this context functionally as
an escape behavior, idiosyncratic to laboratory conditions. However, it
should be pointed out that laboratory paradigms of aggression have
been externally validated in a number of studies that relate this behavior
to aggressive behavior occurring outside the laboratory. These
validations include positive correlations between aggressive responding
and self-rated aggression and violence (Malamuth & Check, 1980;
598
DOUGHERTY ET AL.
Shemberg , Leventhal , & Allman , 1968); peer ratings of aggression
(Williams , Meyerson, Eron , & Semler, 1967); antisocial behavior
(Hartmann, 1969); and criminal records (Wolfe & Baron, 1971). We have
shown that response characteristics differ when instructions attribute the
source of provocation to a machine rather than another individual
(Cherek, Steinberg , Kelly, Robinson, & Spiga, 1990; Kelly & Cherek,
1993). And two recent studies in our lab have supported the external
validity of the PSAP by showing that parolees with violent histories make
significantly more aggressive responses than parolees with nonviolent
histories (Cherek, Moeller, Schnapp , & Dougherty, 1997; Cherek,
Schnapp, Moeller, & Dougherty, 1996).
Collectively, the results of the current study and those noted above
suggest that (a) both provocation and escape from subsequent
provocation are necessary conditions for the occurrence and
maintenance of aggressive behavior; and (b) both the rate of provocation
and the duration of the escape period can affect the frequency of
aggressive responding. Future studies of aggressive behavior should
consider the important role of these variables.
References
AZRIN , N. H. , HAKE , D. F. , & HUTCHINSON , R. R. (1965). Elicitation of
aggression by a physical blow. Journal of the Experimental Analysis of
Behavior, 8, 55-57.
AZRIN , N. H., HUTCHINSON , R. R. , & HAKE, D. F. (1967). Attack, avoidance
and escape reactions to aversive shock. Journal of the Experimental
Analysis of Behavior, 10, 131 -148.
AZRIN , N. H . , HUTCHINSON , R. R. , & MCLAUGHLIN, R. (1965). The
opportunity for aggression as an operant reinforcer during aversive
stimulation. Journal of the Experimental Analysis of Behavior, 8, 171-180.
AZRIN , N. H., HUTCHINSON, R. R, & SALLERY, R. D. (1964). Pain-aggression
toward inanimate objects . Journal of the Experimental Analysis of
Behavior, 7, 223-228.
BANDURA, A. , ROSS, D., & ROSS , S. A. (1963). Imitation of film-mediated
aggressive models. Journal of Abnormal and Social Psychology, 66, 3-11.
BARON, R. A., & RICHARDSON , D. R. (1994). Human aggression. New York:
Plenum Press.
BRAIN, P F., BENTON , D. , HOWELL, P A. , & JONES, S. E. (1982). Resident
rats' aggression toward intruders. Aggressive Behavior, 8, 371-383.
CHEREK, D. R. (1981). Effects of smoking different doses of nicotine on human
aggressive behavior. Psychopharmacology, 75, 339-349.
CHEREK , D. R. , & DOUGHERTY, D. M. (1997) . The relationship between
provocation frequency and human aggressive responding. The
Psychological Record, 47, 357-370.
CHEREK, D. R. , MOELLER, F. G., SCHNAPP, B. , & DOUGHERTY, D. M. (1997).
Studies of violent and nonviolent male parolees: I. Laboratory and
psychometric measure of aggression. Biological Psychiatry, 41 , 514-522.
HUMAN AGGRESSIVE RESPONDING
599
CHEAEK, D. A., & PICKENS, A. (1970). Schedule-induced aggression as a
function of fixed-ratio value. Journal of the Experimental Analysis of
Behavior, 14,309-311.
CHEAEK, D. A., SCHNAPP, W., MOELLEA, F. G., & DOUGHEATY, D. M.
(1996). Laboratory measures of aggressive responding in male parolees
with violent and nonviolent histories. Aggressive Behavior, 22, 27-36.
CHEAEK, D. R., SPIGA, A., & EGLI, M. (1992). Effects of response requirement
and alcohol on human aggressive responding. Journal of the
Experimental Analysis of Behavior, 58, 577-587.
CHEAEK, D. A., SPIGA, A., STEINBEAG, J. L., & KELLY, T. H. (1990) . Human
aggressive responses maintained by avoidance or escape from point loss.
Journal of the Experimental Analysis of Behavior, 53, 293-303.
CHEAEK, D. A., & STEINBEAG, J. L. (1987). Effects of drugs on human
aggressive behavior. In G. D. Burrows & J. S. Werry (Eds.), Advances in
human psychopharmacology: A research annual, Vol. 4 (pp. 239-290).
Greenwich, CT: JAI Press.
CHEAEK, D. A., STEINBEAG, J. L., KELLY, T. H., AOBINSON, D. E., & SPIGA,
A. (1990). Effects of acute administration of diazepam and d-amphetamine
on aggressive and escape responding of normal male subjects.
Psychopharmacology, 100, 173-181.
DENGEAINK, H. A., & COVEY, M. (1983). Implications of an escape-avoidance
theory of aggressive responses to attack. In A. G. Geen & E. I.
Donnerstein (Eds.), Aggression: Theoretical and methodological issues:
Vol. 1 (pp. 163-188). New York: Academic Press.
ENDICOTT, J., & SPITZEA, A. L. (1978). A diagnostic interview: The schedule
for affective disorders and schizophrenia. Archives of General Psychiatry,
35, 837-851.
FEASTEA, C. 8., & SKINNEA, B. F. (1957). Schedules of reinforcement. New
York: Appleton-Century-Crofts.
HAATMANN, D. P. (1969). Influence of symbolically modeled instrumental
aggression and pain cues on aggressive behavior. Journal of Personality
and Social Psychology, 11, 280-288.
HUTCHINSON, A. A., AZAIN, N. H., & HUNT, G. M. (1968). Attack produced by
intermittent reinforcement of a concurrent operant response. Journal of
the Experimental Analysis of Behavior, 11, 489-495.
KELLY, T. H., & CHEAEK, D. A. (1993). The effects of alcohol on free-operant
aggressive behavior. Journal of Studies on Alcohol, 11 (Supp), 40-52.
KELLY, T. H., CHEAEK, D. A., STEINBEAG, J. L., & AOBINSON, D. (1988).
Effects of provocation and alcohol on human aggressive behavior. Drug
Alcohol Dependence, 21,105-112.
MALAMUTH, N., & CHECK, J. V. (1980). The effect of mass media exposure on
acceptance of violence against women: A field experiment. Paper
presented at the meeting of the American Psychological Association,
Montreal, CN.
MICZEK, K. A., & WINSLOW, J. T. (1987). Psychopharmacological research on
aggressive behavior. In A. J. Greenshaw & C. T. Dourish (Eds .),
Experimental pharmacology: Concepts and methods (pp. 27-114). Clifton,
NJ: Humana Press.
SCOTT, J. P. (1975). Aggression. Chicago: University of Chicago Press.
SHEMBEAG, K. M., LEVENTHAL, D. B., & ALLMAN , L. (1968). Aggression
machine performance and rated aggression. Journal of Experimental
Research in Personality, 3,117-119.
600
DOUGHERTY ET AL.
TAYLOR, S. P, & CHERMACK, S. T. (1993). Alcohol, drugs, and human physical
aggression. Journal of Studies on Alcohol, 11 (Supp), 78-88.
ULRICH, R. E., & AZRIN, N. H. (1962). Reflexive fighting in pairs of rats.
Journal of the Experimental Analysis of Behavior, 5, 511-520.
WILLIAMS, J. F., MEYERSON, L. J., ERON, L. D., & SEMLER, I. J. (1967). Peerrated aggression and aggressive responses elicited in an experimental
situation. Child Development, 38, 181-190.
WOLFE, B. M., & BARON, R. A. (1971). Laboratory aggression related to
aggression in naturalistic social situations: Effects of an aggressive model
on the behavior of college student and prisoner observers. Psychonomic
Science, 24, 193-194.