Vous êtes sur la page 1sur 7

BEI-IAVIOIt TItEIIAPY (1973) 4, 110--116

BRIEF NOTE
Aversion by Fiat: The Problem of "Face Validity"
in Behavior Therapy
JoaN M. BERECZ 1
Andrews University
A brief survey of punishment theory is presented, and the inadequacy
of many aversion therapy experiments is outlined. The experimental
demonstration of the alleged aversiveness of proposed stimuli is seen as
basic to making correct judgments regarding the usefulness of aversion
therapy. This has not generally been done and most experimenters have
simply defined aversion by fiat. Typical studies from the area of modifica-
tion of smoking behavior are considered and suggestions are made for
implementation of appropriate control groups.
A numbe r of r ecent r evi ews ( Chur ch, 1963; Sol omon, 1964; Azri n &
Hol z, 1966; Campbel l & Chur ch, 1969; and Dunham, 1971) have con-
t r i but ed much to syst emat i zi ng knowl edge gai ned f r om numer ous puni sh-
ment st udi es wi t h animals. Al t hough aversi on exper i ment er s quot e
puni s hment st udi es in t hei r bi bl i ographi es, t he act ual pr ocedur es em-
pl oyed of t en seem onl y vaguel y r el at ed to puni s hment pr i nci pl es de-
r i ved f r om t he ani mal l aborat ori es. I nt ui t i on seems in most cases t o have
di ct at ed t he t echni ques empl oyed, and t he consequent l ack of success
is of t en at t r i but ed t o t he al l eged compl exi t y and unr el i abi l i t y of effects
r esul t i ng f r om t he use of puni shi ng stimuli.
Chur ch (1963) has st at ed t hat t he so-cal l ed "par adoxi cal " effect s of
puni s hment ar e most l i kel y t o occur when t he pr oc e dur e rei nst at es a
condi t i on of ori gi nal t r ai ni ng or elicits a r esponse si mi l ar t o t he act whi ch
is bei ng puni shed. I t is s omewhat surpri si ng, i n vi ew of this, t hat inves-
t i gat ors have empl oyed smoke and hot ai r as "aversive'" stimuli in at t empt -
i ng t o hel p peopl e stop smoking. Ot her s have empl oyed r api d smoki ng
as a puni shi ng pr ocedur e. Ani mal st udi es need not be r i gi dl y appl i ed t o
huma n pr obl ems, and some of t hese pr ocedur es ma y be useful , but in
~Requests for reprints should be addressed to: John M. Berecz, Department of
Psychology, Andrews University, Berrien Springs, MI 49104.
110
Copyright (~) 1973 by Academic Press, Inc.
All fights of reproduction in any form reserved.
BRIEF NOTE 111
each ease, investigators claiming to employ "aversion" techniques need
to design experimental operations which rigorously demonstrate that
subjects experience the stimuli as aversive. Typically this has not been
done. Stimuli have been defined by fiat as "aversive" and many writers
have concluded that punishment techniques have been given a fair trial
and shown to be ineffective.
Keutzer, Lichtenstein, and Mees (1968), for example write, "It ap-
pears that many experimenters have either been unclear as to what
stimuli control smoking or have been unable to manipulate the relevant
stimuli. Punishment techniques therefore, have been heavily relied upon
to suppress the smoking response. However, laboratory studies indicate
that the effects of punishment on behavior are complex . . . Techniques
other than those relying on response punishment might well be con-
sidered" (p. 530). Rachman and Teasdale state that "We have to extend
the search for suitable and less unpleasant substitutes for aversion ther-
apy" (1969, p. 320). "Life is too full of punishment to make it necessary
or advisable to administer more of it in the clinic," writes Sheehan
(1970). "Therapists who use punishment are probably incompetent to
use anything else, or have a neurotic need to assume the role of the
punisher as a reassurance against their own fear of being in the role
of the one punished" (p. 132).
Thus, a zeitgeist seems to exist in which aversion techniques are often
seen as either ineffective, or effective but ethically unacceptable. Cer-
tainly, ethical issues must always be kept in sharp focus, but such con-
siderations are more easily approached if the effectiveness of various
techniques is clearly understood. Better experimental designs are needed
to demonstrate aversiveness. The issues involved in delineating these
parameters are complex and not amenable to "one-shot" explications.
Rather, converging operations within single experiments, and series of
related experiments tend to establish the usefulness of new techniques
and to demonstrate experimentally that the aversiveness of the stimuli
are related to changes in target behaviors.
ILLUSTRATIVE STUDIES OF SMOKING BEHAVIOR
Schmahl, Lichtenstein, and Harris (1972) conducted a methodologi-
cally unsatisfactory experiment in which treatment components were con-
founded within groups. In this study warm, slightly ozoned, mentholated
air or warm smoky air was blown into smokers' faces while they smoked
at an accelerated rate. In addition, subjects received considerable social
reinforcement for abstinence and were given high expectancies of suc-
cess. Relationship factors were maximally exploited. There were no sig-
nificant differences in smoking for the two groups after treatment, but
112 SaIEF NOTE
even if differences had been found, it woul d have been impossible to
have delineated which of the many components had been effective.
From the present perspective, however, the most serious probl em of
this st udy is t hat the authors attribute "aversiveness" to various experi-
mental procedures in a manner which is unwarrant ed by the data.
Throughout the paper there are comments whi ch either directly state
or strongly imply t hat warm, smoky air was an "aversive" stimulus
[e.g., "The procedure was identical for t he aversi ve smoke and menthol
air groups." (italics mi ne) ] . Even the title of the paper, "Successful
Treat ment of Habi t ual Smokers wi t h Warm, Smoky Air and Rapi d
Smoking," carries an implication regarding t reat ment components which
is unsupport abl e by the data. The authors further the confusion by re-
ferring to Lubl i n and Joslyn's (1968) rapid-smoking procedure as an
"aversion treatment." This has not been demonstrated using appropriate
control groups.
In an unsuccessful st udy whi ch was nonetheless offered by its authors
(Marst on & McFall, 1971) as "demonstrating techniques for minimizing
or overcoming t he methodological problems previously associated with
research on smoking reduction" (p. 154), the issue of experimentally
demonstrating aversiveness was not mentioned. The authors pai d meticu-
lous attention to such issues as accurate record keeping, etc., but seemed
content with t he following statements regarding aversiveness: "The first
of the control conditions empl oyed a pill which is designed to produce
an aversive taste when dissolved in the mout h for two minutes prior to
smoking a cigarette" (p. 154). "The tablets are designed to coat t he
mouth with an herb compound (ginger, licorice, coriander, clove, and
menthol) which produces an aversive taste when followed by a cigarette.
The distributor claims, t hat the program' s success is not influenced if
Ss get used to the taste of the tablets" (p. 156).
There are a number of methodological problems with this study, such
as confounding of t reat ment procedures within groups, but t he most
basic is the failure to deal seriously with the probl em of aversiveness.
It is a noteable commentary on experimenter biases t hat t wo investiga-
tors, sophisticated in experimental design problems were willing to
accept a drug distributor' s claim ("the program' s success is not influ-
enced if subjects get used to t he taste of the tablets") as sufficient for
defining aversiveness. If this were an isolated example, one could dis-
miss the present writer' s contentions as mere picayunishness, however
it is a general finding t hat experimenters seem content to define aver-
siveness by fiat. It does not seem inordinately "hard-nosed" to suggest
that something beyond a drug distributor' s statement be offered as evi-
dence for aversiveness.
BRIEF NOTE 113
Unfortunately, these kinds of studies are marshaled as evidence to
support the premat ure conclusion t hat aversion techniques are ineffective
in modifying smoking. Commenting on the Marston and McFall (1971)
study, Mausner (1971) writes, "One can conclude from this st udy that
an emotional jolt followed by an opport uni t y for catharsis in group dis-
cussions is more effective if an elaborate procedure of scheduling or of
aversive conditioning does not stand in its way" (p. 170). In a more
general comment, Mausner (1971) states "The failure of behavior ther-
apy to modi fy smoking behavior has been ascribed to t he i nadequacy
of aversive control. This i nadequacy is probabl y due to t he fact t hat
internal contingencies of an aversive nature rarely persist outside t he
setting in which the t herapy is given . . . However, it is possible t hat
no aversive conditioning, even one ca/ling upon internal contingencies,
woul d be effective in changing smoking" (p. 168).
Mausner is correct in stressing t he need to consider "internal contin-
gencies," but this in no way diminishes the importance of aversive stimu-
lation. The present writer (Berecz, 1972a, 1972b) has shown t hat if
electric shock is pai red wi t h actual smoking of cigarettes, it is generally
ineffective in changing the behavior. However if, under the same condi-
tions, shock is pai red with internal cravings or thoughts whi ch occur
just prior to smoking, a highly significant reduction in smoking occurs.
It is possible t hat this woul d be the case for other behaviors as well, but
further research is needed.
DESIGN CONSIDERATIONS
Quantitative and Qualitative Parameters of Aversion
Researchers have generally been qui t e clear in specifying quantita-
tive aspects of shock, concentrations of distasteful liquids, smoke, etc.,
empl oyed in experiments. Less attention has been given to qualitative
differences among various stimuli. Certainly sucking an allegedly bi t t er
lozenge must produce an "aversive" experience quite subjectively differ-
ent than t hat resulting from holding one's breat h or being shocked.
These kinds of differences ought not to be ignored since t hey render
studies noncomparabl e on the aversion dimension.
The Relationship of the Aversive Stimulus to Target Behaviors
This probl em has been especially acut e in smoking research where hot,
smoky air, breat h holding, and rapid smoking have all been used as
allegedly aversive stimuli. As previously mentioned, it is best from a
punishment perspective to empl oy stimuli whi ch are unrel at ed to the
t arget behavior. However, even when this is not the case, it is possible
i i 4 BmEV NOTE
to employ control groups to disentangle the aversive components. In
t he hot, smoky air procedure (Wilde, 1964; Grimaldi & Liehtenstein,
1969; Lublin & Joslyn, 1968; and Schmahl e t al., 1972) if smoke were
t hought to be aversive, t hen nonsmoky air woul d need to be empl oyed
with a control group. If heat were thought to be aversive, t hen cool air,
etc. Typically these components have not been systematically disen-
tangled and this has rendered most studies inconclusive.
In t he breath-holding t echni que subjects are instructed to hold their
breat h until it becomes aversive each time t hey desire a cigarette
(Keutzer, 1968). It remains unclear whet her it is t he aversiveness or
t he distracting quality of t he procedure whi ch is effective. It woul d be
possible to use a yolked control group where subjects' tightly clenched
their fist for a duration corresponding to that of subjects' breath-holding
durations in t he experimental group. An analog could be used where
a yolked control group received gradually increasing levels of shock or
noxious odors or tastes corresponding to the aversiveness of t he breat h
holding. Again, this would introduce interpretive problems, but if t he
aversive control stimulus were self-administered (as is breat h holding)
and if prior to conducting the experiment a number of appropriate
analogs were devised, such research would enhance underst andi ng of
t he breath-holding technique.
The rapid-smoking procedure confounds quant i t y of smoke with rapid-
ity of smoking. If experimental subjects were requi red to smoke twice
as rapidly as usual, it woul d be possible to supply control subjects with
filters (M.D. associates, 1970) that make it possible to smoke 50g air
and 507~ smoke. Ot her combinations woul d also be possible.
Many of these complex design problems could be avoided by using
aversive stimuli unrel at ed to t he t arget behavior. One could still use
a convenient technique such as having a subject wear a thick rubber
band on his wrist and snap himself each time the desire for a cigarette
intlaaded. This woul d be portable, immediate, and avoid problems in-
herent in procedures related to smoking.
Av e r s i on Pl acebos
It is important to design studies whi ch incorporate all of t he charac-
teristics of t he aversive stimulus except t he aversiveness into a t reat ment
procedure. The "sugar pill" of drug studies is long overdue in aversion
t herapy studies. The fact that in behavior therapy, t he aversive stimulus
is distinctly presented makes it possible to design plausible placebos.
In using electric shock, for example, it is possible to ask subjects to t ur n
it down until t hey cannot feel i t --st at i ng that this will work on their
"subconscious." Using this kind of control it becomes possible to rat her
BRIEF NOTE 115
cl ear l y speci f y wha t degr ee of t r e a t me nt effect i veness is due t o t he
aver si veness of t he shock ( Ber ecz, 1972a). Si mi l ar pl acebos coul d be
us ed f or ot her aver si ve st i mul i as well.
CONCLUSIONS
Al t hough ma n y st udi es have be e n ci t ed f r om t he smoki ng l i t er at ur e,
ot her ar eas of aver si on t he r a py have be e n h a mp e r e d b y si mi l ar pr obl ems .
I t is s ugges t ed t hat exper i ment er s pr es ent nonei r eul ar exper i ment al evi -
dence f or demons t r at i ng aver si veness, pr e f e r a bl y i nde pe nde nt l y of out -
come measur es. Aver si on is a eompl ex issue, r equi r i ng t he s ame car ef ul
at t ent i on whi ch has be e n aceor ded ot her const r uct s such as anxi et y,
r ei nf or cement , etc. Mor e pa r a me t r i c st udi es are ne e de d in whi ch t he
gener al aver si veness of var i ous st i mul i ar e st udi ed on di f f er ent popul a-
t i ons be f or e usi ng t he m in appl i ed st udi es. Hope f ul l y t he er a of "f ace
val i di t y" in aver si on t he r a py wi l l soon be past .
REFERENCES
AzRIN, N. H., & HoLz, W. C. Punishment. In W. K. Honig (Ed.), Operant be-
havior: Areas of research and application. New York: Appleton-Century-Crofts,
1966. Pp. 380-447.
BEnECZ, J. Modification of smoking behavior through self-administered punishment
of imagined behavior: A new approach to aversion therapy. Journal oJ Con-
sulting and Clinical Psychology, 1972a, 38, 244-250.
BERECZ, J. Cognitive conditioning therapy: A behavioral technique. Behavioral Sci-
ence: An interdisciplinary journal, 1972b, 17, 549-552.
CAMPBELL, B. A., & CHURCH, R. M. Punishment and aversive behavior. New York:
Appleton-Century-Crofts, 1969.
CHURCH, R. M. The varied effects of punishment on behavior. Psychological Review,
1963, 70, 369-402.
DUNHAM, P. J. Punishment: Method and theory. Psychological Review, 1971, 78,
58--70.
GmMALDI, K. E., & LmHTENSTEIN, E. Hot, smoky air as an aversive stimulus in the
treatment of smoking. Behaviour Research and Therapy, 1969, 7, 275-282.
KEUTZEn, C. S. Behavior modification of smoking: The experimental investigation
of diverse techniques. Behaviour Research and Therapy, 1968, 6, 137-157.
KEUTZER, C. S., LmrrrENSTEIN, E., & MEES, H. L. Modification of smoking be-
havior: A review. Psychological Bulletin, 1968, 70, 520-533.
LUBLIN, I., & JOSLYN, L. Aversive conditioning of cigarette addiction. Paper pre-
sented at the meeting of the American Psychological Association, San Francisco,
CA, September, 1968.
MARSTON, A. R., & McFALL, R. M. Comparison of behavior modification approaches
to smoking reduction. Journal of Consulting and Clinical Psychology, 1971, 36,
153-162.
MAUSNER, B. Some comments on the failure of behavior therapy as a technique for
modifying cigarette smoking. Journal o'f Consulting and Clinical Psychology,
1971, 36, 167-170.
M. D, Associates, Stop smoking by smoking method, Westminster, CA, 1970.
116 BRIEF NOTE
PtACHMAN, S. J. , T~ASDALE, J. Aversion t herapy: An appraisal. I n C. M. Franks
( Ed. ) , Behavior therapy: Appraisal and status. New York: McGraw-Hi l l , 1969.
Pp. 279-320.
SCHM-A-aL, D. P,, LICHTENSTEm, E., & tL~a~Is, D. E. Successful treatment of habi t ual
smokers with warm, smoky air and rapi d smoking. Journal of Consulting and
Clinical Psychology, 1972, 38, 105-111.
SrmEHAN, J. G. Reflections on the behavioral modification of stuttering. I n S. Ains-
worth ( Ed, ) , Conditioning and stuttering therapy: Applications and limitations.
Memphis, TN: Speech Foundat i on of America, 1970. P. 132.
SOLOMON, R. L. Punishment. American Psychologist, 1964, 19, 239-253.
WILDE, G. J. S. Behavior t herapy for addi ct ed cigarette smokers. Behaviour Besearch
and Therapy, 1964, 2, 107-110.

Vous aimerez peut-être aussi