JOURNAL OF APPLIED BEHAVIOR ANALYSIS

BEHA VIORAL ENGINEERING: TWO APPARA TUSESFOR TOILET TRAINING RETARDED CHILDREN'

N. H. AZRIN, C. BUGLE, AND F. O'BRIEN

ANNA STATE HOSPITAL

Daytime incontinence is a major problem for retarded children. A training procedurefor eliminating this problem should be facilitated by an apparatus that provided thetrainer with an immediate signal when the child voided so that the trainer could reactimmediately. Two apparatuses were developed for this purpose: a toilet-chair apparatusto signal proper toileting and a portable pants-alarm apparatus to signal wetting of thepants. A reprimand was given when pants wetting occurred whereas positive reinforcementwas given for proper toileting. Results with four profoundly retarded children indicatedthe reliability of the apparatuses in practice and the effectiveness of a toilet trainingprogram that used the two apparatuses.

Nocturnal incontinence of normal childrenhas received extensive study as a result ofwhich effective treatments based on condi-tioning procedures have been developed(Mowrer and Mowrer, 1938), refined, andsystematically evaluated (see review by Lovi-bond, 1964). Daytime incontinence appar-ently has not been a major problem for nor-mal children. For the retarded child, however,daytime incontinence causes major economicconcern and personnel time to custodial insti-tutions. Of even greater importance is the hu-mane consideration that the mentally re-tarded should not be obliged to live in theirown waste. Only recently have reinforcementprocedures been described for toilet trainingprofound retardates. These procedures haveincluded positive reinforcement for toiletingin the appropriate location, Hundziak,Maurer, and Watson (1965), Giles and Wolf(1966), Bensberg (1965), Van Wagenen,Meyerson, Kerr, and Mahoney (1969) and anegative consequence, a reprimand, for void-ing in one's clothing (Van Wagenen et al.,1969). These two procedures should be moreeffective if an apparatus were available toprovide an immediate signal for the elimina-tory responses so that the parent or trainercould provide consequences immediately. A

'This research was supported by the State of IllinoisDepartment of Mental Health and by Grant MH-17981from the National Institute of Mental Health. Re-prints may be obtained from N. H. Azrin, BehaviorResearch Laboratory, Anna State Hospital, Anna, Il-linois 62906.

somewhat complex apparatus for detectingpants-wetting has been developed by VanWagenen and Murdock (1966) and used ef-fectively with normals (Madsen et al., 1969)and retardates (Van Wagenen et al., 1969).A complex photoelectric apparatus for signal-ling correct toileting by retardates has beendescribed by Watson (1968) but no datagiven regarding its reliability or effectiveness.The present report describes two other ap-paratuses that seemed simpler in design, morefoolproof, and more economical than the pre-viously described apparatuses.

METHODPants-Alarm Apparatus

Figure 1 (top) is a representation of thepants-alarm apparatus. No special pants wereneeded. Rather, two ordinary metal clothingsnaps (size 16, male) were fastened to ordi-nary cotton briefs in the area of the crotchand spaced about 1.5 in. (4 cm) apart usinga commonly available device for attachingsnap fasteners to clothing. The two malesnaps were connected by female snaps (Nu-Way snaps available from Burstein-Applebee,Kansas City, Missouri) to a flexible #20-gauge insulated wire leading to the circuitbox worn by the child on a belt beneath theouter blouse or dress. Neither the wires northe small (8 by 5 by 2 cm) and lightweight(4 oz) circuit box was visible. Figure 2 (top)is a schematic of the electrical circuit in thecircuit box. When urine moistened the pants,the decreased electrical resistance between

249

1971, 4, 249-253 NUMBER 3 (FALL 1971)

N. H. Azrin et al.

WET-ALARM PANTS

MOISTURE DETECTING SNAPS

TOILET-SIGNAL CHAIR

CIRCUITBOX

-wl

MOISTURE DETECTING STUDSFig. 1. Wet-alarm pants for toilet training-upper fig-

ure. The pants are ordinary cotton training pants.Two ordinary clothing snap fasteners are attached tothe crotch area. Snap leads connect the snaps to thecircuit box, which is worn on a belt on the waist. Thecircuit box contains an alarm that sounds when urinemoistens the cloth between the two snaps. Dry pantsare substituted for wet ones by simply pulling thesnap leads off the wet pants and snapping them onthe snaps of the dry replacement pants. Toilet-alarmchair for toilet training-bottom figure. Side view ofthe apparatus for immediately signalling urination ordefecation. The chair is a commercially available"potty" chair. Two studs are attached to the bottomof the "potty". Manually connected snap leads connectthe studs to the circuit box which is fixed to the backof the chair. A speaker in the circuit box gives animmediate signal when urine or feces falls into the"potty".

the two snaps completed the low voltage cir-cuit and the pants alarm sounded. The elec-trical current flow was too slight to be de-tected or felt by the child. The alarm was ableep-tone of about 2500 Hz and could beheard easily above the usual noise on a ward.The total cost of all components including thesounding device was about $10.

Toilet Signal ApparatusFigure 1 (bottom) is a representation of the

toilet-signal apparatus. No special chair wasneeded. Rather, a standard children's toiletchair, costing about $5 was used that had astandard plastic "potty" container beneath theseat rest. Two male snap studs were fastenedto the bottom center of the container about0.5 in. (1 cm) apart. The small (8 by 5 by 2cm) signal package was fastened to the backsupport of the chair and was connected to thetwo metal studs by female snaps attached toinsulated wire leads. When the child urinatedinto the "potty", the moisture completed a lowvoltage circuit, see Fig. 2 (bottom), betweenthe metal studs and produced a clickingsound from a speaker in the circuit box. Thesound was qualitatively different from thatgenerated by the pants-alarm and could alsobe heard easily at a distance. The total costof all components, excluding the $5 chair, wasabout $102. A modification of this apparatusfor use in a regular toilet bowl is described byAzrin and Foxx (1971).

SubjectsFour profoundly retarded females aged 3,

5, 6, and 6 yr were the subjects. The 3-yr oldchild had Hurler's syndrome. All were non-verbal, retarded from birth, had been resistantto previous intensive toilet training efforts, didnot learn to dress or feed themselves withoutintensive training, did not play with otherchildren, and had locomotor deficits. The two6-yr olds had visual deficits and were treatedin a day-care center. The 3 and 5-yr old chil-dren were institutionalized in a special facil-ity for retarded children. The IQs of the two6-yr olds were too low to produce a score onthe Stanford-Binet test administered by an in-

"We have been informed by Lehigh Valley Elec-tronics, Inc. that they can supply both of the ap-paratuses described here at an individual cost of about$40: address Box 125, Fogelsville, Pennsylvania 18051.

25}0

TOILET TRAINING APPARATUS

WET-ALARM PANTS CIRCUIT

S - 1

Si

TOILET-SIGNAL CHAIR CIRCUIT

3

S ljaSNAPS (8) B-1a

R-2a

Fig. 2. Schematic of wet-alarm pants circuit (top) and toilet-signal chair circuit (bottom); component identifica-tions are as follows: R-l and R-la, 100 ohm. i/8 watt resesistor; R-2 and R-2a, 15,000 ohm, i/8 watt resistor; R-3,22,000 ohm. 1/8 watt resistor; C-1, 100 mfd capacitor, 15 volts; C-2, 2.2 mfd capacitor, 15 volt; T-l, T-la and T-3.transistor #GE-2; T-2, transistor #GE-7; S-1, "Bleep-tone" signal tone device available from C. A. Briggs Company,Glenside, Pa. S-la, speaker, 1.5 in., 0.1 watt, 8 ohm; B-i and B-la Battery, Eveready #216, 9 volt, or equivalent;Snaps: Nu-Way available from Burstein-Applebee Co., Kansas City, Missouri. The A-Snaps attach to matchingsnaps on the training pants; the B-Snaps attach to matching studs on the toilet chair.

dependent testing agency. IQs were not avail-able for the other two children.

ProcedureThe apparatuses were used as part of a toi-

let training program for two to four monthswith the two children in the day-care pro-gram. The baseline frequency of wetting forthese two children was 10 and seven pants-wettings per day. During training, manychanges were made in the apparatus, the con-sequences for urinating, the toileting sched-ule, etc. in an effort to develop an effectiveprogram. By the end of the training, both

children ceased wetting their pants almostentirely, averaging less than once per week.They continued to be dry after the trainingwas discontinued.The final procedure developed with the

first two children was then applied to the twoinstitutionalized children. A baseline recordof incontinence was obtained for at least twoweeks, during which the child wore the pants-alarm apparatus (inoperative) and wasplaced on the toilet chair (also inoperative)for 5 min at 2-hr intervals such as might benormal toileting routine in an institution.After the 2-wk baseline, the training proce-

251

N. H. Azrin et al.

dure was initiated and the two apparatuseswere made operative: the pants alarmsounded and the toilet-signal clicked whenwet. The child was required to drink a glassof water every half-hour to increase the natu-ral frequency of urination; similarly, Gilesand Wolf (1966) used laxatives to increasebowel movements. The scheduled toilet pe-riods were changed from the baseline fre-quency of one every 2 hr to one every half-hour to provide more opportunity for thechild to urinate in the toilet-chair. The teacherimmediately praised and hugged the childand gave the child a bit of candy upon hear-ing the signal caused by the child urinatingin the toilet chair. When the child urinated inher pants, the alarm sounded, whereupon theteacher gave her a single spank on the but-tocks and then arranged a timeout from posi-tive reinforcement by not interacting with herfor 10 min.Both children required training to eliminate

unassisted in the toilet chair. For one of thechildren, instructions were sufficient for thechild to lower her pants, sit on the chair anddress herself afterward. Manual guidance andverbal approval was used to train the secondchild to dress and undress herself. In bothinstances, the trainer faded out the assistanceuntil the child toileted herself with no needfor manual guidance or verbal or gesturalinstructions.

RecordingThe children's pants were inspected by hav-

ing a specifically assigned attendant feel thepants at each of the regular toileting periodswhich occurred every 2 hr before and aftertraining and every half-hour during training.A second adult was available to serve as analternate or assistant, thereby providing ad-ditional assurance of the recording. Addi-tional confirmation and assurance was pro-vided by the experimenter who directlyobserved the procedure for about 2 hr per dayon more than half of the days on an irregularbasis for the first child and every day for thesecond child. These supervisory checks re-vealed that the procedure and recording werebeing conducted appropriately and consist-ently. The adults also recorded any instancesof wetting between the scheduled check peri-ods. All instances of wetting were recordedin writing at that moment on special forms.

RESULTSFigure 3 (upper) shows that the toilet

training procedure reduced the number ofwet pants per day for the child from aboutfour per day during the baseline to near zeroafter three days of training. The effect was,however, reversible; wetting returned slightlywhen training was prematurely discontinued.When toilet training was reinstated, the childceased wetting and remained dry thereafter.A similar training program using the same

two apparatuses was tested with an additionalchild. Figure 3 (lower) shows that the second

NO ITRMININGX~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~6-

5 -

3 -

2-

1 -

0 --

TRAINING

TOILET TOILETTRAINING TRAMIdINGAPPRATUS IAPPARATUS

~

1 5 10 15 20

DAYS

_, 4

i3

2

NOTRAINING

12 DAYS

I0 - 0 0

25 30 35

0 5 10 15 20 25 30 35

DAYS

Fig. 3. The number of times that two profoundlyretarded children wet their pants during a 6-hr periodeach day. Toilet training utilized a wet-alarm pantsapparatus and a toilet-signal chair apparatus. For S-3(upper), at regular intervals the child was given 5 minto eliminate in the toilet chair. For S-4 (lower), the5-min interval was in effect during the first seven daysof training but was then extended to 20 min.

Il

252

TOILET TRAINING APPARATUS 253

child was wet about two times daily beforetraining. This level was not reduced duringinitial training. The child almost completelyceased wetting, however, when the procedurewas changed such that the child was given upto 20 min to eliminate in the chair rather thanthe previous 5 min. Following training, thechild remained dry for the eight remainingdays of the study.Some problems were presented by the ap-

paratuses initially, but were solved. False posi-tives (the alarm sounding when no wettinghad occurred) were never present with thetoilet-chair apparatus. For the pants alarm,false positives did result when the exposedparts of the snaps on the pants touched eachother. This was solved by (1) spacing thesnaps 1.5 in. (3.7 cm) apart and arrangingone in front and the other behind rather thanside by side; (2) by covering the top of thesnap with insulating paint or plastic, and, (3)using the resistance values shown in Fig. 2 sothat ordinary perspiration would not activatethe alarm. False negatives (alarm not sound-ing when wetting had occurred) were not aproblem for the toilet chair except that onechild occasionally excreted only a few drops,which was insufficient to activate the circuit,until a plastic funnel was substituted for thepot and the electrodes were located at thebottom narrow portion of the funnel. Falsenegatives for the pants alarm resulted frombreaking of the connecting wires and low bat-tery voltage. The wire breakage problem wassolved by using standard 20-gauge wire; thebattery voltage problem was solved simply bychanging batteries every two months since thecircuits were designed for minimal currentdrain. A search for false negatives was system-atically conducted by examining the chil-drens' pants at each scheduled toilet period aswell as other times. No instance of false posi-tive or false negative occurred for either ap-paratus during the last month of use.

DISCUSSIONThe pants alarm apparatus differs from the

apparatus described by Van Wagenen andMurdock (1966) in that no special sensingelectrode grid need be sewn into the pantsand no special laundering is needed to protectthe electrode sensors. No reliability or costcomparison can be made because that infor-mation was not given for the other apparatus.

The report of the photoelectric apparatus(Watson, 1968) also gave no informationabout reliability or cost. In addition, fairlyelaborate precautions were needed to preventthe patients who used that apparatus fromartifactually interrupting the light beams.The usefulness of the present apparatuses

seemed to be in (1) the freedom they per-mitted the trainer to attend to other tasks untilthe alarm sounded, rather than continuouslywatching the children for signs of wetting,(2) the immediacy with which they permit-ted consequences to be delivered, (3) theirreliability and maintenance-free operation,(4) their low cost of about $10 each, and,(5) their potential application to early toilettraining of normal children (one recent suchattempt was effective after 10 hr of trainingover a 3-day period).

REFERENCESAzrin, N. H. and Foxx, R. M. A rapid method of

toilet training the institutionalized retarded. Journalof Applied Behavior Analysis, 1971, 4, 89-99.

Bensberg, G. J. Teaching the mentally retarded, ahandbook for ward personnel. Atlanta: SouthernRegional Education Board, 1965.

Giles, D. K. and Wolf, M. M. Toilet training institu-tionalized, severe retardates: an application of be-havior modification techniques. American Journalof Mental Deficiency, 1966, 70, 766-780.

Hundziak, M., Maurer, R. A., and Watson, L. S. Op-erant conditioning in toilet training severely men-tally retarded boys. American Journal of MentalDeficiency, 1965, 70, 120-124.

Lovibond, S. H. Conditioning and enuresis. NewYork: The Macmillan Co., 1964.

Madsen, C. H., Hoffman, M., Thomas, D. R., Koropsak,E., and Madsen, C. K. Comparisons of toilettraining techniques. In D. M. Gilfond (Ed.), Sociallearning in childhood. Belmont: Brooks/Cole, 1969.Pp. 124-132.

Mowrer, 0. H. and Mowrer, W. M. Enuresis: a methodfor its study and treatment. American Journal ofOrthopsychiatry, 1938, 8, 436-459.

Van Wagenen, R. K., Meyerson, L., Kerr, N. J., andMahoney, K. Field trials of a new procedure fortoilet training. Journal of Experimental Child Psy-chology, 1969, 8, 147-159.

Van Wagenen, R. K. and Murdock, E. E. A transis-torized signal-package for toilet training of infants.Journal of Experimental Child Psychology, 1966, 3,312-314.

Watson, L. S., Jr. Application of behavior-shapingdevices to training severely and profoundly mentallyretarded children in an institutional setting. MentalRetardation, 1968, 6, 21-23.

Received 7 October 1970.(Revised 30 June 1971.)