JOURNAL OF APPLIED BEHAVIOR ANALYSIS              2003, 36, 349–353                 NUMBER   3 (FALL 2003)

                          A TOKEN ECONOMY TO INCREASE
                                 FOOD ACCEPTANCE
                                          SUNGWOO KAHNG
                                       KENNEDY KRIEGER INSTITUTE AND


                              JAMES H. BOSCOE       AND   SHANNON BYRNE
                                         KENNEDY KRIEGER INSTITUTE

          Escape (termination of a meal) and token-based differential reinforcement of alternative
          behavior were used as reinforcement to increase acceptance of food. Using a changing
          criterion design, the number of bites accepted and consumed was gradually increased to
          15 bites per meal. These data suggest that, in some cases, escape may be a potent rein-
          forcer for food acceptance.
             DESCRIPTORS: food acceptance, food refusal, escape, negative reinforcement

   Reinforcement-based interventions such              negative-reinforcement-based intervention
as differential reinforcement of alternative           with 1 individual who exhibited food refus-
behavior (DRA) are common approaches to                al. They provided 2-min escape contingent
the treatment of pediatric feeding disorders           on acceptance of one or two bites of food.
(Patel, Piazza, Martinez, Volkert, & Santana,          Although the intervention was effective in
2002). In most cases, the reinforcer used is           increasing food intake, it was unclear what
irrelevant to behavioral function (e.g., Coe           effect DRA had on the participant’s food ac-
et al., 1997). Although these interventions            ceptance because the only dependent mea-
have been demonstrated to be effective, re-            sure was weight gain, a permanent-product
search on the treatment of severe behavior             measure. In addition, experimental control
disorders suggests that interventions based            was not demonstrated because an AB single-
on the function of the problem behavior                case experimental design was used. There-
may be more successful (Iwata et al., 1994).           fore, the purpose of this study was to ex-
Therefore, it is conceivable that a treatment          amine the use of a token economy in con-
based on negative reinforcement (i.e., es-             junction with a DRA intervention in which
cape) may prove to be a viable alternative             the reinforcer was escape from food presen-
treatment to increase food acceptance given            tation (i.e., negative reinforcement).
that food refusal may (in many cases) be
maintained by escape from the presentation
of food (Munk & Repp, 1994).                                           METHOD
   Kitfield and Masalsky (2000) evaluated a                Clara was a 4-year-old girl who had been
                                                       admitted to an inpatient unit for the treat-
  We thank Jonnie Brown for her assistance with data   ment of food refusal. Clara’s impairments in-
collection.                                            cluded speech delay and possible pervasive
  Reprints may be obtained from SungWoo Kahng,         developmental disorder. Prior to treatment,
Department of Behavioral Psychology, Kennedy Krie-
ger Institute, 707 N. Broadway, Baltimore, Maryland    Clara received 100% of her nutritional in-
21205 (e-mail:              take from a bottle. Meals were conducted in

350                              SUNGWOO KAHNG et al.

a treatment room equipped with an obser-          food presentation contingent on food refus-
vation window, three chairs, a table, and a       al. A new bite was presented every 30 s or
Rifton chair (a child-sized wooden chair          immediately following the escape interval.
with a seat belt, adjustable angle back, ad-      During the other baseline (praise plus escape
justable foot rests, and a clip-on tray). Ad-     extinction), the spoon remained in front of
ditional supplies included a spoon, Nuk ,         her lips until a new bite of food was pre-
timer, and token board.                           sented (or she accepted the bite). The meal
   Frequency data were collected on laptop        was terminated once 10 bites of food were
computers, and the dependent variables con-       presented or after 20 min had elapsed. A
sisted of the following behaviors: (a) accep-     variety of pureed foods such as applesauce,
tance (the entire bite was deposited in the       ravioli, pudding, and carrots were presented.
mouth within 5 s of the initial presentation
and the food was swallowed, as demonstrat-        DPRA plus Physical Guidance
ed by opening the mouth, within 30 s) and            Two DPRA plus physical guidance (PG)
(b) food refusal consisting of head turns         conditions, which were similar to the pre-
(turning the head or body 45 past midline         vious DPRA conditions, were compared to
during bite presentation), disruptions (any       examine differential effects that may have
part of the body comes into contact with the      been a function of procedural differences. In
spoon, plate, cup, food, or the experiment-       one condition (PG for refusal), Clara was
er’s hand or arm during bite presentation),       physically guided to accept the food contin-
and mouth covers (placing the bib or one or       gent on food refusal. In the other condition
both hands or arms on or within 2 in. of          (PG for nonacceptance), she was physically
the mouth during bite presentation). A sec-       guided to eat if she did not accept the bite
ond observer collected data during 43% of         within 5 s. Physical guidance consisted of
the meals. Interobserver (exact) agreement        the application of gentle pressure to the
was calculated by dividing the number of          mandibular junction of the jaw and depos-
agreements by the number of agreements            iting the bite of food in the mouth. The
plus disagreements and multiplying by             meal ended once 10 bites of food were pre-
100%. Interobserver agreement averaged            sented or after 20 min had elapsed. The
99% and 96% for acceptance and food re-           foods presented were identical to those in
fusal, respectively.                              baseline.
Differential Positive Reinforcement of            Differential Negative Reinforcement of
Alternative Behavior                              Alternative Behavior
   Two baselines consisting of differential          The food bites continued to be presented
positive reinforcement of alternative behav-      as in the DPRA and DPRA plus PG con-
iors (DPRA) were compared to examine the          ditions. However, if Clara accepted the bite
individual effects of procedural variations.      of food, she received a Blues Clues token.
During both baselines, Clara was prompted         She was not permitted to play with the to-
to take a bite of food and a level spoonful       kens or token board. Meals were terminated
of food was presented to her mouth (mid-          once she traded in the prespecified number
line), approximately 2 to 3 in. from her lips.    of tokens (i.e., differential negative reinforce-
If she accepted the bite within 5 s, the ther-    ment of alternative behavior or DNRA) or
apist delivered praise (e.g., ‘‘good job taking   30 min had elapsed. (Although this did not
your bite’’). During one baseline (praise plus    occur during this study, bites of food would
escape), she received 15 s of escape from the     have continued to be presented if she did
                                         DRA AND ESCAPE                                                351

   Figure 1. The top panel presents the number of bites accepted (accepted within 5 s of the initial presen-
tation and swallowed within 30 s). The horizontal dashed lines represent the number of acceptances required
to terminate the meal. The bottom panel presents responses per minute of food refusal.

not trade in the tokens.) If she did not ac-           A changing criterion design was used to
cept the bite within 5 s, she was physically           demonstrate experimental control.
guided to accept the bite. Initially, bite pre-
sentation was limited to applesauce. A sim-
ilar evaluation with a variety of other foods                 RESULTS AND DISCUSSION
(excluding applesauce) was later conducted.               Clara did not accept any bites of food
Clara was initially required to earn one (ap-          during the DPRA or DPRA plus PG con-
plesauce) or two (other foods) tokens before           ditions (Figure 1, top panel). A Nuk was
meals were terminated. The number of to-               eventually introduced into the DPRA plus
kens required to terminate the meal was                PG condition (Meal 21) in place of the
gradually increased after two meals in a row           spoon and was used throughout the remain-
in which she met criterion to terminate the            der of this study. The Nuk permitted the
session, and the terminal goal was 15 tokens.          therapist to deposit the bite of food more
352                              SUNGWOO KAHNG et al.

easily into the participant’s mouth when im-      the previous phases, and she did not accept
plementing physical guidance.                     it. Alternatively, it may have been the case
   Once Clara met the 15-bite criterion with      that the repeated presentation of the same
applesauce (in Meal 42), meals with other         food resulted in quicker acceptance than the
foods were interspersed with the applesauce       presentation of other foods.
meals until she accepted 15 bites during             It is unclear which component of our
both the applesauce meals and meals with          treatment package (i.e., escape or tokens)
other foods.                                      was responsible for the increase in food ac-
   During DPRA, food refusal was higher           ceptance given that a component analysis
when escape was contingent on food refusal        was not conducted. Although it is conceiv-
during the praise plus escape condition           able that the tokens functioned as rein-
(Figure 1, bottom panel). Both DPRA plus          forcement for acceptance, this is somewhat
PG conditions resulted in an initial increase     doubtful given that she always immediately
in food refusal, which eventually decreased       traded in these tokens for the back-up re-
to near-zero levels. Food refusal continued       inforcer (i.e., escape). Therefore, it is more
to remain low throughout the DNRA                 probable that the escape contingency was
phase.                                            responsible for the success of this treat-
   Finally, the average numbers of bite pre-      ment.
sentation per meal were 10 bites and 2 bites         One limitation to our intervention is that
during the DPRA and DPRA plus PG con-             it may be more time consuming than other
ditions, respectively. Average meal lengths       interventions such as DPRA plus PG be-
during the DPRA, DPRA plus PG, and                cause of the number of steps necessary to
DNRA conditions were 5.4 min, 20 min,             increase the termination criterion. There-
and 15.3 min, respectively. The average           fore, it may be that this negative-reinforce-
length of the meals at the 15-bite criterion      ment-based intervention should be used
(applesauce and other foods) was approxi-         only if other interventions fail.
mately 16 min. A maintenance meal was                Finally, future studies should examine the
conducted 6 months after discharge, during        role of instructional control on food accep-
which time she accepted 46 of 50 bite pre-        tance. That is, Clara was given instructions
sentations within 10 min.                         prior to each meal. Therefore, it may have
   These data suggest that a treatment            been the case that these instructions played
package consisting of an escape contingen-        a vital part in increasing food acceptance
cy and a token economy may be a viable            (and decreasing food refusal).
treatment for food refusal. This treatment
was implemented after attempts to use
physical guidance, a common treatment                              REFERENCES
for food refusal, failed. It is likely that in
                                                  Coe, D. A., Babbitt, R. L., Williams, K. E., Haji-
this case escape from food presentations              mihalis, C., Snyder, A. M., Ballard, C., et al.
may have been a more potent reinforcer for            (1997). Use of extinction and reinforcement to
acceptance than praise or avoidance of                increase food consumption and reduce expul-
                                                      sion. Journal of Applied Behavior Analysis, 30,
physical guidance.                                    581–583.
   It is interesting to note that Clara accept-   Iwata, B. A., Pace, G. M., Dorsey, M. F., Zarcone, J.
ed applesauce much more quickly than the              R., Vollmer, T. R., Smith, R. G., et al. (1994).
other foods, which may indicate a preference          The functions of self-injurious behavior: An ex-
                                                      perimental-epidemiological analysis. Journal of Ap-
for applesauce. However, this is improbable           plied Behavior Analysis, 27, 215–240.
because she was presented applesauce during       Kitfield, E. B., & Masalsky, C. J. (2000). Negative
                                          DRA AND ESCAPE                                                353

    reinforcement-based treatment to increase food           tion of two differential reinforcement proce-
    intake. Behavior Modification, 24, 600–608.               dures with escape extinction to treat food re-
Munk, D. D., & Repp, A. C. (1994). Behavioral                fusal. Journal of Applied Behavior Analysis, 35,
    assessment of feeding problems of individuals with       363–374.
    severe disabilities. Journal of Applied Behavior
    Analysis, 27, 241–250.                               Received June 21, 2002
Patel, M. R., Piazza, C. C., Martinez, C. J., Volkert,   Final acceptance May 6, 2003
    V. M., & Santana, C. M. (2002). An evalua-           Action Editor, Linda Cooper-Brown

To top