Behav. Res & Therapy. 1975. Vol. 13. pp.

101-l I I Pergamon Press Prmtcd m Great Bratam

ELIMINATING SELF-INJURIOUS BEHAVIOR BY

EDUCATIVE PROCEDURES

N. H. AZRIK* L. GOTTLIEB. L. HUGHART.

M. D. WESOL~WSKI and T. RAHN~
Anna State Hospital

Summary-Self-injury is a common problem among autistic and severely retarded persons. The
most effective treatment has been pain-shock punishment. To provide a possible alternative treat-
ment, modifications were made in previously developed treatments for autisttc behavior. The
revised method included positive reinforcement for non-self-mjurious behavior. a period of
required relaxation or incompatible postures upon each occurrence of a self-injurious episode. and
a hand-awareness training procedure. The treatment procedure was used with I I clients. ten of
whom were very severely retarded. No dents were excluded. The mean number of self-injurious
episodes was reduced by 90 per cent on the tirst day. by 96 per cent at the end of one week and
by 99 per cent by the end of three months. For four of the clients self-injury was eliminated almost
entirely. The new procedure appears to be an effective method of treating self-injurious behavior
and avoids the general reluctance to use pain-shock.

Self-injurious behavior by the profoundly retarded or mentally ill is one of the most severe
psychological disorders since physical injury always results and sometimes. even death, if
ignored. Yet, this problem persists in spite of a large number of recent reports of effective
treatment by learning therapy procedures. One frequently used method is pain-shock
punishment which has been effectively used by Risley (1968). Tate and Baroff (1966).
Yeakel et al. (1970) Corte, Wolf and Locke (1971). and see review by Bucher and
Lovaas (I 968). A second method of treating self-injury is timeout from positive reinforce-
ment which has been used effectively by Wolf. Risley and Mees (1964), Hamilton,
Stephens and Allen (1967) Myers and Deibert (1971). Wolf et al. (1967) but has been
fairly ineffective in studies by Corte. Wolf and Locke (1971) Risley (1968) and Tate
and Baroff (1966). In a few instances. effective treatment has resulted from a third
method, that of reinforcement for non-injurious behavior (Lovaas et al., 1965; Lane
and Domrath, 1970; Peterson and Peterson. 1968).
Unfortunately, the general applicability of the above treatments is an open question
since all of the above reported applications have been case studies in which only one
selected client was used. except for two reports where the same method was used with three
clients (Bucher and Lovaas, 1968) and four clients (Corte et al., 1971). Also, most
of the studies have eliminated the self-injurious behavior during restricted time
periods ofan hour or less per day. Exceptions are the all-day elimination obtained by Tate
and Baroff (1966). Hamilton et (11.( 1967). Wolf clt rrl. ( 1964).
Of the three methods, shock seems to have the advantage of extreme rapidity in eli-
minating self-injury, often within I hr. Perhaps the greatest restraint on the use of shock
* Author from whom reprints may be obtamod. at Behavior Rescorch Laboratory. Anna State Hospital, Anna,
Illinois 62906, U.S.A.
t Currently at St. Lawrence State Hospital. Ogdenshurg. New York 13669. U.S.A.
101
102 N. H. AZRIN. L. GOTTLIEB, L. HUGHART. M. D. WISOLOWSKI and T. RAHY

has been the reluctance of therapists to resort to this physical punishment (see discussion
by Bucher and Lovaas. 1968:and Risley. 1968). As Lovaas and Bucher have stated. regard-
ing the use of shock-punishment by the ward staff members. “all have approached the task
with extreme reluctance and anxiety” (p. 140). In addition. the spectre of a ward attendant
carrying an electric prod discourages the widespread use of this demonstrably effective and
rapid treatment.
The alternatives to shock have been less satisfactory. Physical restraint by tying the
client to a chair (Lane and Domrath, 1970) is not as rapid a treatment as shock and also
suffers from the characteristic of being excessively aversive. Timeout from positive re-
inforcement has not been as rapid or as effective as shock and seems to suffer from the
disadvantage that the client can continue to injure himself during the timeout. thereby
precluding its use with severe self-injury (Corte et ~1.. 1971). Reinforcement of incompatible
behavior has the advantage of being totally non-aversive but has not been used effectively
alone, only in combination with other methods (Lane pt al., 1970. and Lovaas cr al.. 1965).
The present study devised a new treatment program based largely on three recently de-
veloped procedures that appear to hold promise as a relatively non-aversive treatment for
self-injury. The first method is chat of Autism Reversal (Azrin. Kaplan and Foxx. 1973:
Foxx and Azrin, 1972) which has been found to be effective as a general treatment for autis-
tic behavior, of which self-injury may be considered a sub-class and is based on the Over-
correction principle (Foxx and Azrin. 1972; Foxx and Azrin. 1973). In the Autism Reversal
procedure, the client is required by instruction and manual guidance to engage in several
different fixed postures which are non-self-stimulatory. This required practice is given
upon each self-stimulatory episode on a response contingent basis. When the client is not
self-stimulating, he is given positive reinforcement for alternative. incompatible activities.
The second promising method, the Required Relaxation procedure, is also derived from
the Overcorrection principle, and has been effectively used to eliminate a variety of agita-
tive-disruptive behavior, including one client who injured himself (Webster and Azrin,
1973). This Required Relaxation procedure was found to be especially favored by hospital
ward staff as a humane and meaningful type of treatment. The third promising procedure
was Hand-Awareness Training. In a recent treatment developed for eliminating nervous
habits of normal clients (Azrin and Nunn, 1973). the lack of awareness by the client of the
location of his hands seemed to be contributing toward nervous habits involving the
hands. Consequently the normal clients were given training in being continually aware of
the position of their hands as part of the treatment. Since self-injury almost always includes
striking oneself with the hands, the Hand-Awareness Training might be expected to help
the self-injurious client to control this problem.
The present study modified these three promising procedures of Required Relaxation.
Autism Reversal, and Hand-Awareness Training for use with a larger number of self-injur-
ious clients in an attempt to eliminate self-injury on an all-day basis.

METHOD

Eleven clients were obtained in response to an offer to several institutions to provide

assistance in treating clients who repeatedly inflicted injury on themselves that resulted
in evident tissue damage. No clients were excluded. Five were from the same institution.
 Eliminating self-injurious behavior by educative procedure 103

the remaining six from four other institutions. Treatment was given in the client’s institu-
tion. Table 1 shows the age. sex. diagnosis. IQ. years of institutionalization. years of exhi-
biting self-injury. the type of self-injury and the frequency of the behavior. Ten clients were
diagnosed as severe or profoundly retarded. having an average IQ of 13. the highest IQ
being 26 and the lowest was 6. The 11th client was diagnosed as schizophrenia, childhood
type with an IQ of 89; he exhibited many of the diagnostic signs of autism. The retarded
clients had an average age of 30 yr with an average duration of institutionalization of 18 yr,
all having been institutionalized before the age of 15. Self-injury was reported to have been
a problem for an average of 12 yr except for the schizophrenic boy who developed the
problem only a year earlier. One noteworthy instance was the 1S-year-old female who was
reported in her records to have started hitting her head against the sides of her crib at
2 yr of age and had evidently self-inflicted scratches on her cheeks and ears during her
first year of life. All clients had visible swelling and most also had scratches. scabs. bruises,
or open wounds. All clients struck themselves on the face or head or on one part of their
head such as the ears, side of the face, or eyes, usually with their fist or open hand. In
addition. 2 clients banged their head on a floor or wall (listed as head-banging in Table
l), one of them as his predominant method of self-injury. Five of the clients had been given
protective clothing such as a helmet or gloves or put in physical restraints such as special
jacket or wrist restraints. Ten of the 11 clients were receiving tranquilizing or sedative
medication. For 4 of the clients. treatment has also been given previously in the form of
Electra Convulsive Shock therapy or timeout seclusion. or manually holding the client’s
hands behind the back. For 2 of the clients. the ward staff was only mildly interested in
treating the problem. For one of these clients they felt that the protective helmet was ade-
quately preventing self-injury: for the other client, the problem of eye-gouging usually was
subordinated to other pressing ward problems.

Recording
The extent of self-injurious behavior prior to treatment was directly recorded by observ-
ing the client for as long as was feasible. For 2 clients who were in restraints or protective
clothing, the restraint or clothing was removed and the client observed until the self-injury
responses appeared to be causing damage. For one of these clients. only 10 min of observa-
tion was feasible; for the other, 38 min. For the other 9 clients, the duration of observation
was adjusted to the frequency of the behavior. One client who hit himself continuously was
observed for 30 min. whereas another client who had a few episodes per day was observed
for 8 hr per day for 12 days. by a time-sample procedure. A response was considered self-
injurious if the client struck, bit or scratched himself. The nature of the behavior was so
unusual that the observers felt little difficulty in differentiating these self-injurious actions
from normal on-going activities. In every instance. the ward staff reported that the
recorded frequency was representative of the client’s usual frequency. For one client, the
self-injurious response was static. namely pressing her hand hard against her eyeball; this
response was recorded in terms of duration. When the self-injury consisted of spaced
blows. the measure was number of blows. If the self-injury consisted of a rapid flurry of
blows, the measure was the number of such episodes (see last column of Table 1).

Respo~w detectiorl
After treatment was initiated. two instructors continually observed the client for about
12 hr/day for the first 2 or 3 days. always keeping the client in full view, and within arms’
104 N. H. AZRIN, L. GOTTLIEB. L. HUGHA~T. M. D. WE~~L~WSKI and T RAH\

Table 1. Description and frequency of occurrence of the self-injurious behaviors for each of the 11clients

Years Prr-treatment
Client Years exhibiting frequent)
age and sex Diagnosis IQ inst’d self-injury Nature of injur! of behavior

2%yr-old male Mentally II 32 14 face hitting. self-chokmg. 32 da!

retarded bitmg. kicking
profound
44-yr-old female Mentally 8 31 6 face slapping 6 da>
retarded
profound
26-yr-old female Mentally 26 19 6 scratching. face slapping 9 da)
retarded
severe
25-yr-old male Mentally 14 20 70 face and head slapping. 74s day
retarded punchmg
profound
32-yr-old female Mentally 6 19 8 eye gougmg 86”,, of the da)
retarded
profound
46-yr-old female Mentally 11. 31 31 face. arm. leg slapping. finger 200 day
retarded biting
profound
24-yr-old male Mentally 12 14 20 hand biting. head hitting and Sday
retarded banging
profound
IO-yr-old male Mentally 9 9 7 ear punching 3578:daq
retarded
profound
I S-yr-old female Mentally 6 18 18 face. ear. and head slapping 3500, da)
retarded
profound
I Syr-old male Schizophrenia. X9 3 0.65 face punching 4X’da)
childhood
type
I7-yr-old male Mentally 8 8 7 head banging 4liday
retarded
profound

length since the instructors were required to give continuing positive reinforcement for
appropriate non-injurious behavior. Members of the ward staff assisted the instructors in
recording and carrying out the treatment during these first few days and were encouraged
to assume this responsibility when the instructors were absent. For 5 of the clients, the
special instructors were present for at least 2 weeks. For the other 6 clients who were at
remote institutions, the instructors returned periodically for direct confirmatory observa-
tion of reported benefits.

On each ward, the special instructor taught those employees who were interested how
to conduct the procedure and supervised their performance. The instructors role-played
the procedure with the staff members prior to application of the procedure to a client. The
staff was advised of the importance of using the procedure immediately upon detection
of a self-injurious response and upon every self-injurious response. One ward employee
was typically designated as the coordinator but all were instructed to record the self-injur-
 Eliminating self-injurious behavior by educative procedure 105

ious behavior and to initiate the treatment procedure. The ward staff was advised to con-
tinue the treatment for several weeks after the special instructor departed but his role was
advisory only. Telephone contact was made daily with the employees to encourage their
continuation of the treatment. For all clients, the treatment lasted for at least 12 days. For
4 clients, the ward employees were not motivated to continue the procedure after the spe-
cial instructors were absent. The ward employees discontinued treatment for I client after
12 days. another after 1 month. and the third and fourth clients after 2 months.

Positive rrir?fbmwwt,for outward-diwctrd actiritics

The client was given positive reinforcers for engaging in a variety of outward-directed
and incompatible responses. For the clients whose behavior was most ‘inner-directed’
these activities included eye contact with the instructor. looking at specific objects when
instructed, sitting down or arising from a chair when instructed. walking. banging drum-
sticks together, grasping the armrests of their chair. catching and throwing a ball, or even
simply sitting still without injuring themselves. For the clients who were more ‘outward-
directed, the responses included playing with a jig-saw puzzle. educational games, toilet-
ing, dressing, grooming. washing oneseff. word recognition. ward cleaning chores, making
beds. trips to the ward commissary. playing with simple music-making instruments. group
recreational activities, swimming at a local pool. and bus rides. The guiding principle was
to select those responses that involved active interaction with the physical and social en-
vironment, especially responses that were functional and could be expected to be main-
tained later because of their potential enjoyment or utility in their own right. If the ward
program included regular supervised activities or classes. every attempt was made to enroll
the client in these activities or classes. The reinforcers selected for use included verbal
praise. back-stroking, and desired snack items such as candy. pudding, coffee, and juice.
The ward staff usually knew what was reinforcing for a given client. In general. the snack
items proved to be the best reinforcers for the more inward-directed clients. Reinforcers
were delivered very frequently at first, the verbal praise was almost continuous. Once the
client began spending extended periods without self-injury. the reinforcers were made
more intermittent by reinforcing for longer response sequences.

Required Rr1a.yation
The Required Relaxation was essentially the same as described elsewhere as a treatment
for agitative-disruptive conduct (Webster and Azrin. 1973). When the client injured him-
self. he was told that he was over-excited and agitated and was required to go relax in
his bed. He was assisted in putting on a hospital gown and directed to his own bed where
he remained for 2 hr. The instructor stood behind the head of the bed and assured that
the client did not leave the bed.
A modification in the previously described procedure was made because of the severe
nature of the self-injurious conduct. The client was required not only to remain in bed
but to maintain his arms in an extended downward position with the hands alongside his
legs away from his head. a position that was incompatible with striking one’s head. The
instructor used verbal instruction and gentle manual guidance. according to the Gra-
duated Guidance Method which provides no more manual contact than is minimally
necessary to obtain the required posture. This gentleness of contact as prescribed by the
Graduated Guidance procedure was essential : otherwise great resistance resulted. His
fixed posture was required for an uninterrupted 10 min in this arms-extended position. If
106 N. H. AZRIS. L. GOTTLIEB. L. HUGHART. M. D. Wrso~owss~ and T. RAHX

he moved his hands toward his head or hit himself, 10 more min were required. All clients
required considerable manual guidance initially, but after 1 or 2 days they usually per-
formed the Required Relaxation Procedure upon verbal direction and with minimal
manual contact. After the client began spending several hours on the ward without injur-
ing himself. the Required Relaxation was given for any emotional or agitated conduct that
was found to be a usual precursor to self-injury. such as excitedly pacing or rocking. mut-
tering, screaming, or cursing in which case the Required Relaxation was given only for
10 min. again explaining to the client that he was overexcited and had to calm himself.
The instructor’s presence at bedside was usually required only during the first 1 or 2 days
until the client learned to lay fairly still with the arms extended. Thereafter. the instructor
usually remained with the client for only about 5 min until he was assured that the client
was resting in the correct posture. The instructor. or any other staff member passing by
the open door, could determine whether the client was resting as required. The pressure-
sensitive device attached to the leg of the bed. and described previously (Webster and
Azrin, 1973), sounded a signal to the staff if the client left the bed.

The Hand Control Procedure was very similar to the arm exercises described in the pre-
vious report (Azrin et al., 1973) for eliminating hand autisms. In the previous report. when
the client exhibited an autism he was immediately reprimanded and told that he must now
practice holding his hands away from his body. The instructor stood behind him and
guided him. The client was required to hold his arms extended at his sides, then out-
stretched horizontally to the front, then to the side, then extended over his head. Thirty
seconds was required in each position preceded by a verbal instruction prior to each
change in position. This cycle was repeated for 20 min in the standing posture. No conver-
sation occurred between the instructor and the client, except the instruction every 30 set
as to the change in hand posture. The client received no praise or cajoling from the instruc-
tor during the exercises.
Several modifications were made in the above procedure to make it suitable for use with
self-injurious clients. The arms-forward position was eliminated since a head-striking
movement could be made easily from that position leaving only the arms down, side, and
up position. Secondly. the arm-down position was modified to include clasping of the
hands together behind one’s back in order to make the response more incompatible with
striking oneself than was the hands-by-side position. This clasping of the hands also
seemed easier to teach and seemed to be used spontaneously after training by the clients
as a method of self control. Another modification was to require a position change every
10 set rather than every 30 set in order to have the client more active and to give more
opportunity to react to the instructions. A fourth modification was to conduct the exercises
in the sitting posture rather than standing for those clients who were physically unable
to stand easily or who became too fatigued or emotionally upset by having to stand. A
fifth modification was omission of the head-orientation exercises at the start of the practice
period since the self-injurious clients could easily hit themselves while the instructor was
manually guiding their head. A sixth modification was to terminate the arm exercise period
while the client was in the arms-down position since this posture merged more naturally
with his usual posture. This last posture was maintained for a longer period, 30sec to
1 min, until the client was standing calmly with his arms down with no need for the instruc-
tor to hold his arm there. As in the previous report. the clients usually learned after 1 or
 Eliminating self-injurious behavior by educative procedure 107

2 days to move their arms to the new position upon hearing the verbal instruction with
a need for only minimal manual guidance.
In the event that the client became very upset during the Hand Control exercises. the
instructor attempted to continue but paying special note to the need for gentleness of con-
tact during the Graduated Guidance. If the agitation still persisted. the client was seated
in a chair and the practice continued. In the event the agitation still persisted. the client
was given the Required Relaxation procedure in his bed for a few minutes. until he was
calm, at which time he completed the remainder of the 20-min practice period. The general
rule was that the client should learn that the full 20min of practice would be required
whenever he injured himself.

The Hand-Awareness Training procedure of Azrin and Nunn (1973) was modified for
use with the present type of clients. At the start of treatment the instructor continually
made comments to the client regarding the need to position his hands away from his head.
The instructor used gestures, pointing. and touching as well as verbal statements. When
the hands were away from the client’s head. the instructor praised him for keeping them
there. Conversely, when his hand moved upward for any reason, the instructor directed
him to lower them. To maintain awareness of their hands. the clients were instructed to
walk with their hands clasped behind their back and to clasp the armrests of their chairs
while seated. As in the other procedures. the reinforcers were snack treats, praise, and
stroking. As the client learned to maintain his hands away from his head, the instructor
commented on the client’s hand position progressively less often, but after having com-
mented almost continuously for the first 2 days. This awareness was also being taught in-
directly as an integral part of the Hand Control procedure when the client changed his
hand position every 10 set in response to the direct instruction regarding his hands. Simi-
larly, in the Positive Reinforcement procedure. the client was being reinforced for using
his hands in a functional manner.

At the start of treatment, the protective helmet or mittens were removed and the client
was given the Hand-Awareness training. and the Positive Reinforcement for outward-dir-
ected activities. When the client injured himself. the instructor immediately reprimanded
him in a stern tone of voice that conveyed his displeasure and gave the Hand Control Pro-
cedure for 20min or the Required Relaxation procedure for 2 hr. (See Results section as
to which clients received which procedure.) The Positive Reinforcement for outgoing beha-
viors and the hand-awareness training were then reinstated after indicating to the client
that he should not injure himself and that the Hand Control or Required Relaxation
would be needed if he did injure himself. Initially. all instruction was conducted in special
locations on the ward to obtain a more distraction-free atmosphere. After a few trials the
locations were varied so that the client would learn that the procedure would eventually
be applied in any locations in which he might self-injure.

In both the Required Relaxation and Hand Control procedures, the instructor ‘faded
out’ the need for his guidance or even his presence. In the Required Relaxation procedure,
once the client attained the fixed-posture. and his behavior was agitation-free while in bed,
108 N. H. AZRIN. L. GOTTLILB. L. HUGHART. M. D. WESOLOWSII and T. RAN

the instructor moved himself gradually to the rear of the client. and eventually completely
out of the client’s view. In the Hand Control procedure. once the client was responding
to the postural instructions and maintaining the positions unassisted. the instructor
reduced his guidance of the client’s arms to merely a touch and then to just ‘shadowing’
the client’s movements.
When the client had spent one day without self-injury. the duration of the scheduled
Required Relaxation or Hand Control was reduced to about 5 min on the next day. then
to 2 min. and then to a simple warning on successive days. providing no self-injurious re-
sponses had occurred on the preceding day. Even in the final stage. a warning or reminder
was given to the client for any attempt at self-injury.

RESULTS

Figure 1 shows the change in self-inlurious conduct averaged for the 11 clients. Each
data point is expressed as a percentage of the baseline level. On the first day of training.
the self-injurious responses decreased by 90 per cent from the pre-treatment level and de-
creased further by about 96 per cent by the end of the first week. By the fourth week of
training, the self-injurious responses had decreased by 98 per cent and by 99 per cent by
the third month. A r-test of differences showed that all of the data points were significantly
less than the baseline level (p < 0.001). (Four clients received no treatment by the ward
staff after the special instructors were absent. The data for these 4 clients are therefore in-
cluded only up to the date that treatment was terminated.)
The Relaxation Procedure was used as the treatment for the first 6 clients. For 3 of these
clients, the Relaxation Procedure was very effective. but none of the other 3 clients was
benefitted substantially and 2 of them began injuring themselves in a seemingly deliberate
fashion in order to obtain the bed-rest indicating that the bed-rest involved in the pro-
cedure was serving as a reinforcer for self-injury. In addition. these 3 clients continued to
attempt to injure themselves while in bed unless very closely supervised. Consequently.
the Relaxation Procedure was discontinued for these 3 clients and the Hand Control pro-
cedure substituted for it. The last 5 clients were given only the Hand Control procedure
and no Relaxation Training. The data treatment points in Fig. 1 are for the Hand Control
procedure for all but the first 3 clients whose data points are for the Relaxation Procedure.
Analysis of the individual benefits showed that 4 of the 11 clients were almost totally
free of the self-injurious responses. either having no further self-injurious responses or less
than one per week. One of these was discharged to a shelter-care facility where the opera-
tors reported he has not exhibited any self-injury. Each of the other 7 clients averaged less
than 4 self-injurious responses per day by the second week of treatment. The client who
had exhibited self-injurious behavior for the longest period, 31 yr. had the highest level
of self-injury after 2 weeks. an average of four self-injurious responses per day. The client
who had exhibited self-injurious responses shortly after birth, and a high pre-treatment
rate of 3500 responses per day. exhibited an average of only one self-injurious response
per week after 2 months of treatment.
Three of the clients often physically aggressed against other residents or staff members
at the same time that they exhibited self-injury. The physical aggression by all 3 clients
decreased substantially once treatment was initiated for the self-injury.
General improvements in the clients’ overall manner were evident when the self-injur-
ious behavior was decreased. Ten of the 11 clients seemed to greatly increase their social
interactions and social responsiveness. The notable exception was one client located on
 Eliminating self-injurious behavior bj educative procedure lo9

BoselIne I Instructions for

,oo, observotlon/; non-Injury

P I
i+i 0 I
1234567
I2 34 5 67 8 9101112
Days
Weeks

Fig. I. Self-injurious behavior of IO retarded and I schizophrenic persons. The frequency of self-
injury is expressed as a percentage of the frequency recorded prior to treatment. The self-injury
prior to treatment (Baseline Observation) was recorded for varying durations dictated by the safety
of the client and the frequency of the behavior. The data points are for the average number per
day for the first 7 days and weekly thereafter. Each data point is for 1 I clients for the first 2 weeks.
9 clients for the 3rd week. 8 for the 4th to 7th weeks and 7 clients thereafter. During the ‘Baseline
Observation’, the self-injury was simply recorded; during the ‘Interruptions for Non-Injury’, the
instructors used positive reinforcement for non-injurious behavior. a Required Relaxation pro-
cedure for agitated states. a Hand Control procedure for self-injury and a Hand-Awareness pro-
cedure. The vertical dashed line designates the time that the treatment procedure started.

a ward with virtually no programmed activities and whose pre-treatment routine was to
sit limply in a chair or to lie in a corner asleep wearing his protective helmet. He exhibited
little muscle tone when the instructor manually guided him during attempts at reinforce-
ment during the treatment phase.

DISCUSSION
The new procedure was effective in eliminating, or greatly reducing, the self-injurious
behavior of the mentally ill and retarded clients. The treatment was fairly rapid as seen
by the average reduction of about 90 per cent on the first day and about 96 per cent by
the seventh day. The treatment appears applicable to the general population of self-injur-
ious clients as seen by its effectiveness with all 11 clients in the present unselected sample.
The extent of the benefit was substantial in that after three months of treatment, the self-
injurious behavior was reduced by an average of 99 per cent. Self-injury was virtually eli-
minated for all 4 clients. The general acceptability of the treatment was evidenced by the
positive reaction of the clinical personnel in all five institutions in which the clients were
treated.
Speculatively, the degree of benefit for a given client seemed to be greater if he had a
pre-existing high level of outward-directed behavior, or if the ward environment strongly
encouraged outward-directed activity. All 4 of the clients who were virtually ‘cured’ had
considerable social and attention-getting behavior, including aggression toward others by
three of them, whereas the clients who benefitted less, were seemingly oblivious to the pres-
ence or actions of others. The ‘custodial’ type of wards in which little attention was paid
to residents often abandoned the treatment effort, whereas the clients in the treatment-
oriented ward situations continued to receive the instruction and continued to benefit. The
I10 N. H. AZRIN. L. GOTTLIEB. L. H~GHART. M. D. WFSOLOWS~I and T. RAH\

Required Relaxation procedure seemed most appropriate for the outward-directed client
whereas the Hand Control procedure seemed more appropriate with the inward-directed
client. The 3 clients with whom the Required Relaxation was successful were all of the
outward-directed type. whereas the 3 clients with whom it was unsuccessful were inward-
directed.
Comparison of the present procedure with alternative procedures must be somewhat
tentative since the clients treated in the previous reports are of unknown comparability.
Nevertheless. the present method appears to be more acceptable as a treatment than either
shock or timeout seclusion. in that no physical punishment is used and the emphasis is
on instruction and adding reinforcers. In all of the institutions included in this stud). shock
was viewed as a last resort and the staff were apprehensive about its abuse. They were
eager to use this procedure before considering shock punishment. With respect to the
speed of treatment, the present method appears at least as rapid as has been reported for
the alternative methods but far less rapid than the almost instantaneous benefit obtained
in most reports of shock (Bucher and Lovaas. 1968; Corte er al., 197 1; Tate and Baroff.
1966; Risley. 1968). With respect to general applicability to a variety of clients, the present
method would be considered sirperior at this time if only because the previous reports have
been case studies that have not as yet reported results for a large number of unselected
clients. In general, the present method seems to provide many advantages over the alterna-
tive methods.

Ackrlo~c/rd~~~tt~~if~~This research was supported by the Illinois Department ofMental Health. Grateful acknow-
ledgement 1s given to the many staff members of the several Institutions in the State of Illinois who participated
in this project.

REFERENCES
AZRIS N. H.. KAPLAN S. J. and FOXX R. M. (1973) Autism reversal: Eliminating stereotyped self-stimulation of
retarded individuals. AIII. J. ww. Defic. 78, 241-248.
AZRIX N. H. and NUNN R. G. (1973) Hablt-reversal: a method of eliminatmg nervous habits and tics. Bchur.
Res. & Thmpy 11, 619-628.
BUCHER B. and LOVAAS 0. 1. (1968) Use of aversive stimulation in behavior modification. In Miatni Sw~posiwt~
011f/w Prrdictim qfBrharior. lYh7: Arer.s~w Stiwrrltrtior~ (Ed, M. R. JONES). pp. 77-145. University of Miami
Press. Florida.
CORTE H. E., WOLF M. M. and Locw B. J. ( 1971) A comparison of procedures for eliminating self-injurious be-
havior of retarded adolescents. JI. q>pI. Bchtrr-. .4/w/ 4. 20 I-2 13.
Foxx R. M. and AZRIN N. H. (1972) Restitution: A method of eliminating aggressive-disruptive behawor 01
retarded and brain damaged patients. Brlltrr. Res. d Thop!~ IO, I i27.
FOXX R. M. and AZRIN N. H. (1973) The elimination of autistic self-stimulatory behavior b!, overcorrection. J.
uppl. &htrr. AJIU/. 6, I 13.
HAMILTON J.. ST~PH~.NS L. and ALLEN P. (1967) Controlling aggressive and destructive behavior m severely
retarded institutionalized residents. A,II. J. I~II. DI+IC,.71. X52-X56.
LAVE R. G. and DOMKATH R. P. ( 1970) Behavior Therapy: A cast history. Hospittrl UM/ C~~WII~III~~~ Ps~~c/rrtrrr~~
21, l50-1.53.
LOVAAS 0. 1.. FREITAG G.. GOLI~ V. J. and KAS%)KI.A I. C. (1965) Experimental studies in childhood schizo-
phrenia: Analysis of self-destructwe behawor. Bci~trr. Rc.\. d: T/~cwp~~ 2. 67-84.
MYERS J. J. and DCIRERT A. N. (1971) Reduction of self-abusive behavior in a blind child by using a feeding re-
sponse. J. Beh. Thwup~~ & cy. P.s~rhiur. 2, I41 -144.
P~TEKSON R. F. and P~:TERSOUL. R. (1968) The use ol positive reinforcement in the control of self-destructive
behavior in a retarded boy. J. r.vp. C/II/C/ P.s~rhl. 6, 35 l-360.
RISLI.~ T. (1968) The effects ol punishmg the autistic behaviors ol”n dcwant child. J. trppl. Bch. Arwl. I, 21 -34.
TATE B. G. and BAROFF G. S. (1966) Aversive control ol self-injurious behavior in a psychotic boy, Bchar. Rc.s.
& Thupy 4, 2x I 2x7.
 Eliminating self-injurious behavior by educative procedure 111

WERS~~K D. R. and AZKIY N. H. (1972) Required relaxation: A method of mhlbiting agitatwe-disruptive behawor
of retardates. Behur. Rex 6: TIwrupv 1 I, 67-78.
WOLF M.. RISLEYT.. JOHNSTONM.. HARRIS F. ~~~ALLEK E. (1967) Application of operant conditionmg procedures
to the behavior problemsof an autistic child: A follow-up and extension. Brhu~. Rrs. & Tlwtwp~~ 5, 103-I 1 I.
WOLF M.. RISLEY T. and MEES H. (1964) Application of operant conditioning procedures to the behavior pro-
blems of an autistic child. Behcrr. Rrs. & T/IL’wI)_I~
1, 305-3 12.
YEAKEL M. H.. SALISBURYL. L.. GREER S. L. and MARCUS L. F. (1970) An appliance for autoinduced adverse
control of self-injurious behavior. J. rrp. Chikl. Ps!~hol. 10, I59- 169.