Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Weight Reduction As A Function Of The Timing Of Reinforcement In A Covertaversive Conditioning Paradigm
(USC Thesis Other)
Weight Reduction As A Function Of The Timing Of Reinforcement In A Covertaversive Conditioning Paradigm
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
WEIGHT REDUCTION AS A FUNCTION OF THE TIMING
REINFORCEMENT IN A COVERT AVERSIVE
CONDITIONING PARADIGM
by
Beatrice Scheinbaum Manno
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
June 1971
MANNO, Beatrice Ila Scheinbaum, 1942-
WEIGHT REDUCTION AS A FUNCTION OF THE TIMING OF
REINFORCEMENT IN A COVERT AVERSIVE CONDITIONING
PARADIGM.
University of Southern California, Ph.D., 1971
Psychology, clinical
University Microfilms, A X ER O X C om pany, Ann Arbor, Michigan
THIS DISSERTATION HAS BEEN MICROFLIMED EXACTLY AS RECEIVED
UNIVERSITY O F SO UTH ERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 9 0 0 0 7
This dissertation, written by
. . . . JBe^.tr.ice...S.chein.ba.nnQ..Marmo........
under the direction of h&X. Dissertation C o m
mittee, and approved by all its members, has
been presented to and accepted by The G radu
ate School, in partial fulfillment of require
ments of the degree of
D O C T O R O F P H I L O S O P H Y
D ean
D ate Jluii.a..JL22Jv.......
DISSERTATION COMMITTEE
...
a / / ) Chairman
PLEASE NOTE:
Some Pages have indistinct
print. Filmed as received.
UNIVERSITY MICROFILMS
ACKNOWLEDGMENTS
I would like to express my appreciation to
Steven Vincent and Stuart Greenberg for all their help
in running subjects, to Kay De Loach for her assistance
in data collection and to my husband, John, for his
never ending encouragement and support.
I would further like to thank the members of
my Guidance and Dissertation Committees for their valu
able help during the conduct of this study. They are
Dr. Milton Wolpin, Dr. Edward Connolley, Dr. Steven
Frankel and Dr. Norman Tiber.
Above all, I wish to thank Dr. Albert Marston,
the Chairman of my Dissertation Committee and Senior
Advisor. His direction, support, and advice have been
invaluable to me, not only in the development of this
Dissertation, but through the two years I have spent as
a graduate student in his department. For his continu
ous guidance, I extend my warmest gratitude.
ii
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS i i
LIST OF TABLES................................................. V
LIST OF FIGURES.......................................... vii
Chapter
I. INTRODUCTION..................................... 1
A. Problems in Traditional Aversive
Conditioning ............................ 1
B. Problems in Traditional Weight
Reduction Programs ..................... 3
C. The Development of Covert Treatment
Programs ................................ 9
D. The Covert Sensitization Procedure . . . 12
II. THE RESEARCH PROBLEM...............................23
III. METHOD..............................................29
Subjects.......................................29
Materials.......................................29
Procedure. . . .................................31
IV. RESULTS............................................39
Weight Loss.....................................39
Eating Behaviors ............................ 42
Body Images.....................................48
Motivation.....................................55
Rotter's Internal-External Scale ........... 55
Program Value.............. 56
Home Practice of Treatment Imagery ......... 57
Weight History .............................. 57
Chapter Page
V. DISCUSSION..........................................59
REFERENCES................................................ 71
TABLES.....................................................80
FIGURES.....................................................99
iv
LIST OF TABLES
Table Page
1. Mean Weights (lbs.) Across Groups at
Pre-Treatment, Post-Treatment, and
Follow-up.........................................80
2. Analysis of Variance for Pre-treatment
Weights (lbs.) ............................... 81
3. Mean Weight Loss Across Groups at
Post-treatment and Follow-up ................ 82
4. Analysis of Variance for Weight Loss.............. 83
5. Newman-Kuels for Differences in Weight
Loss Over Trials................................84
6. Analysis of Variance for Total Food Approach . . 85
7. Analysis of Variance for Habit Scores.......... 86
8. Analysis of Variance for Body Image (Real) . . . 87
9. Analysis of Variance for Body Image (Ideal). . . 88
10. Analysis of Variance for Body Image
Discrepancy...................................... 89
11. Analysis of Variance for Program Evaluation. . . 90
12. Analysis of Variance for Home Practice of
Treatment Imagery............. 91
13. Pearson Correlations: Food Approach................ 92
14. Pearson Correlations: Food Approach................ 93
15. Pearson Correlations: Body Image ................ 94
16. Pearson Correlations: Body Image ................ 95
v
Table Page
17. Pearson Correlations: Motivation and I-E. . . . 96
18. Pearson Correlations: Weight History and
Program Evaluation.......................... 97
19. Pearson Correlations: Miscellaneous............. 98
LIST OF FIGURES
Figure Page
1. Mean Weight Loss Across Trials for All
Treatment Conditions ......................... 99
2. Changes in Total Food Approach Scores
from Pre to Post-Treatment Across
Groups.........................................100
3. Changes in Mean Habit Scores from Pre to
Post-treatment Across Groups ............... 101
4. Changes in Real Body Image from Pre to
Post-treatment Across Groups ............... 102
5. Changes in Ideal Body Image from Pre to
Post-treatment Across.Groups ............... 103
6. Changes in Body Image Discrepancy from
Pre to Post-treatment Across Groups........... 104
CHAPTER I
INTRODUCTION
This study is concerned with the empirical assess
ment of weight reduction under various covert punishment
conditions. In order to fully elucidate the research prob
lems, it is first necessary to discuss some of the issues
which, in part, led to the development of covert procedures.
These issues essentially stem from problems inherent in
traditional aversive conditioning and, on the other hand,
the growing recognition that traditional weight reduction
programs (i.e., those relying on interviews, diets, and
discussions) were relatively ineffective in the control of
obesity.
A. Problems in Traditional Aversive Conditioning
Attempts to deal with inappropriate or otherwise
undesirable approach behavior (i.e., addictions or sexual
deviations) within a behavior modification framework have
centered primarily around methods utilizing aversive stimu
lation. The focus upon the negative stimulus has predomi
nated the literature with debates anchored to the varying
utilities of chemical or electrical stimulation (Rachman,
1
2
1965), This review pointed up some of the difficulties,
in terms of traditional conditioning paradigms, inherent
in chemical aversion techniques, difficulties in controll
ing the temporal and sequential aspects of CS and UCS, the
number of CS-UCS pairings, and the precision of stimulus
duration and intensity. In addition to the central depres
sant character of many of the drugs and their general
interference with CR acquisition, Rachman suggested that
the sheer unpleasantness of chemical techniques might be
sufficient reason to call them into question. There is
evidence, on the other hand, that chemical aversion might
lead to increased subject aggression (Morgenstern, Pearce,
and Davies, 1963) and anxiety arousal (Eysenck and Rachman,
1965) .
At the same time, electrical stimulation studies are
ridden with difficulties of a more elemental nature con
cerned with the topography of the stimulus itself. Garcia
and Koelling (1966) demonstrated the necessity of the
appropriateness of the cue. When ionizing radiation was
used as the UCS producing gastrointestinal disturbances,
avoidance learning did not transfer to tactile or audio
visual stimuli, while it did to gustatory stimuli. On the
other hand, avoidance learning did transfer to tactile and
audiovisual stimuli but not to gustatory stimuli when the
UCS was faradic shock.
3
To account for the many failures in the treatment
of alcoholism, Wilson and Davison (196 9) indicated that the
use of shock may condition fear responses only to nongus-
tatory attributes of the alcohol. This interpretation seems
tenable in light of the evidence found by Garcia, McGowan,
Ervin, and Koelling (1968) that food intake varied as a
function of CS and UCS topographies; food flavor paired
with illness induced by X-ray resulted in a significant
decrease in food consumption whereas size of food pellet
paired with illness did not. Combining size of pellet with
shock, on the other hand, resulted in inhibition of eating,
whereas pairing of flavor and pain did not. Lazarus (1968)
pointed out, in addition, that faradic shock may be appro
priate where tactile or visual stimuli are involved, such
as in cases of handwashing compulsions, while inappropriate
in cases of alcoholism or obesity.
It might then be that the notion of the inappro
priateness of cue function can be invoked to account for
the inability to extinguish various approach responses,
particularly if the stimulus elicits responses dissimilar
to the act being punished.
B. Problems in Traditional Weight Reduction Programs
The problems encountered in the control of addic
tions seem to be particularly striking in the case of
4
obesity where postulated causal agents range from power
needs (Suczek, 1957), to the failure to discriminate
between emotional arousal and physiological needs which
characterize hunger (Bruch, 1961), to inadequate exercise
(Mayer, 1955). The more traditional approaches to the con
trol of eating, such as diets and informational services,
fall short of all reasonable criteria of success, while
behavior therapies have offered not much more reason for
optimism. Reviewing the results of obesity studies,
Stunkard and McLaren-Hume (1959a) reported that of 100
patients seen at the Nutrition Clinic of the New York Hos
pital, only 12 lost 20 lbs. (criterion loss) at any time
during the two-year program, while 39 paints failed to
return after the first visit. Additionally, only six
people were still successful after one year of treatment
and two people successful after two years of treatment.
Interpretations of the results of other obesity studies are
complicated by the failure to report the methods of patient
selection and their degree of overweight. Patients who did
not remain in treatment were generally excluded from data
analysis. However, using a 20-lb. loss as the criterion,
few authors report success. In fact, of the studies
reviewed by Stunkard, only that of Feinstein, Dole, and
Schwartz reported success, and then in only 29% of their
patients.
5
It appears that where treatment centers around
calorie diets, interviews, and individual instruction in
food and nutrition, most of the patients who do lose weight
regain a significant amount. In view of this data,
Stunkard investigated four factors as possible indeces of
prognosis: sex of the patient, presence of the "night-
eating" syndrome, outcome of previous attempts at weight
reduction and response to the Taylor Manifest Anxiety Scale.
While the obese patients did show higher anxiety scores than
normals, only the sex of the patient seemed to have pre
dictive value; males tended to have a significantly higher
weight loss than females. However, it certainly might be
argued that the 20-lb. criterion was not an effective index
of differential success since men generally weigh more than
women at the outset. Initial weights were not given. A
better measure would have been percentage weight loss or
the arc sines of those losses.
In another attempt to find correlates of obesity,
the author was somewhat more successful. Schachter (1967)
reported several interesting studies which developed out of
the hypothesis that obese subjects do not rely upon the
same cues as non-obese subjects in the determination of
hunger. Bruch (1961) shared this opinion when she suggested
that obesity represents the failure to discriminate between
arousing emotional states and primary physiological needs
6
which characterize hunger. If it is assumed that a mother
can misinterpret emotional demands made by an infant, and
respond to each demand as if it were hunger, then it might
follow that arousal, rather than specific physiological
cues such as gastric motility, will subsequently be
labelled as hunger. In a test of this hypothesis, Stunkard
(1959, 1964) found that gastric motility was present 71% of
the time that normal subjects gave self-reports of hunger
and only 47% of the time that obese subjects reported
hunger. Schachter, Goldman and Gordon (1967) manipulated
food deprivation and fear, which has been shown to inhibit
gastric motility (Cannon, 1915) and found that, in the
fear-provoking situation, normals ate significantly less
than obese subjects who apparently were not affected by
fear in terms of food intake. Secondly, the obese subjects
ate as much when their stomachs were full as when empty,
unlike the normals. On the other hand, the Nutrition
Clinic of St. Luke's Hospital in New York reported contra
dictory data when the only food available to obese subjects
was a liquid preparation (Hashim and Van Itallie, 1965).
In this study, subjects were allowed to eat as much of the
liquid diet as they wished, as often as they wished.
Given this situation, the obese subjects ate significantly
less than a group of normals. Schachter (1967) interpreted
these results in terms of the external cues which accompany
7
eating. That is, appeal of food acts as a discriminative
stimulus for the self-report of hunger. In this case, the
food had relatively little appeal; it therefore did not
trigger the report of hunger.
If this interpretation is correct, then it might
be assumed that other external cues would be capable of
eliciting eating. He found, in fact, that "time" was a
significant cue in the determination of "hunger" for obese
subjects, whereas normals referred to time only to the
degree that it correlated with internal physiological
cycles. When a clock was changed to read 6:05 (dinner
time), obese subjects ate more significantly than
when it read 5:20, although, in actuality, the time was the
same in both cases (Schachter and Gross, 1967). Addition
ally, Nisbett (1966) found that external, non-visceral
cues such as the sight and smell of food had a significant
effect on food intake for obese subjects, while not for
normals. Apparently, external cues play a strong role in
the eating behavior of overweight individuals, an observa
tion which is not reflected in traditional approaches to
weight reduction. This might account, in fact, for the
poor results reported in Stunkard's (1959) review of treat
ment programs.
Aversion therapies, although reporting somewhat
more encouraging results, tend to rely on single case data
8
(Meyer and Crisp, 1964) making any conclusive statements
tenuous. In an empirical test of faradic aversion, Stollak
(1966) contrasted weight loss among the following groups:
(1) a no-contact group; (2) a no-contact group that was
asked to keep a diary of eating behaviors; (3) a contact-
diary group in which the review of diary records consti
tuted the focal treatment; (4) a contact-diary group given
food-associated shock; (5) a contact-diary non-specific
shock group. Only the contact-diary group showed a signi
ficant weight-loss, although at follow-up, this group had
gained enough weight to obviate any differences between the
groups.
Harris (1969) reviewed the extent of attempts to
deal with the obesity problem and suggested that, in part,
the difficulty stems from the peculiar nature of this type
of addiction. That the behavior seems to provide immediate
positive reinforcement, while reinforcement for resisting
food is usually delayed, presents a problem for therapies
oriented around the pairing of negative reinforcement with
the act of eating. As Eysenck (1960) pointed out, when the
appropriate behavior is performed in the absence of punish
ment, the fear of aversive consequences will become extin
guished. Moreover, as Ferster, Nurnberger and Levitt (1962)
have suggested, eating behavior is under the control of
various stimuli other than those physiological cues related
9
to hunger, since eating behavior is not restricted to any
one situation. The problem is further compounded by the
fact that one must eat; therefore we cannot deal exclu
sively with avoidance response acquisition such as in the
case of smoking or alcoholism (Harris, 1969), but rather,
must attend to the establishment of limiting cues which in
themselves have refraining properties.
C. The Development of Covert Treatment Programs
Part of the trend towards the understanding of cue
functions is a renewed interest in cognitive factors in
behavior modification, coupled with a growing disenchant
ment with traditional aversion techniques. This renewed
interest is to some extent related to the assertion that
covert operants are indeed responses. Skinner (1953) went
so far as to suggest that internal events should not be
thought of as having special properties simply because they
occur within the organism's skin.
That these covert responses (labeled coverants by
Homme, 1965) can, in fact, serve to reinforce behavior has
been amply demonstrated (Anant, 1966, 1967, 1968, 1968a;
Cautela, 1966, 1967, 1970; Davison, 1968; Marston, 1964,
1964a, 1965, 1965a, 1967; Rehm and Marston, 1968). Marston
(1965a) suggested "It has been an underlying assumption that
10
covert self-reactions act in self-rewarding ways parallel
to the operation of overt self-reinforcement and that
variables which affect sr also affect covert sr similarly."
(p. 2) He demonstrated that self-reinforcers are effective
in the learning of visual-motor tasks (1967), anxiety-
reduction (Rehm and Marston, 1968) and imitation learning
(1965). Not only is it now acceptable to talk about cogni
tions, it is necessitated by the literature of self-control
and contingency management.
Literature gained much impetus from Homme's (1965)
discussion of coverants and the methodology for self
monitored control systems. Essentially, Homme postulated
that when target behaviors (TB's) are followed first by
anti-TB coverants, second by pro-non-TB coverants and
finally by some high probability behavior (HPB), the pro-
non-TB coverant is reinforced. By thus invoking Premack's
(1965) differential probability hypothesis, that of a pair
of responses, the more probable response will reinforce the
less probable one, many of the difficulties inherent in
aversion procedures are circumvented, at least theoreti
cally. It will be recalled that a major shortcoming of
most aversive therapies is the problem encountered from the
properties of the stimulus itself. Homme (1965) is not
limited in his range of reinforcing stimuli, for the
strengthening of a coverant depends only upon its
11
occurrence immediately prior to the response of higher
probability.
Mahoney (197 0) distinguished between the coverant
and several other covert variables in that the coverant
has "response status" while other covert variables such as
blood sugar level and osmolarity are essentially nonresponse
variables because of their static nature. Secondly, cover
ants are observable by at least one person, the individual
who is experiencing them. As Homme suggested, "Each sub
ject is a highly sophisticated computer when it comes to
discriminating the occurrence or nonoccurrence of behavior
in himself. Whether he is thinking about a chair or think
ing about a table is a simple discrimination which he can
make with great reliability." (p. 503) That the subject
can "observe" the coverant distinguishes it from the many
inferred variables encountered in other areas of research.
Thirdly, the coverant is inferentially proximal and direct,
"...which further separates the coverant from many of the
more •dynamic1 covert variables. In coverant control, an
observing individual...need only make a one-step, direct
inference from a person's overt behavior to his covert
response. However, it is not unusual for dynamically ori
ented researchers to make several... inferences from behavi
oral manifestations to 'psychic event'." (Mahoney, 1970,
p. 512) Finally, the coverant involves voluntary
12
replicability, as opposed to respondent behavior which must
be elicited. In spite of some of the problems raised by
Mahoney in his discussion, such as the contingency-fulfill-
ment problem and the apparent failure of subjects to dis
criminate consistently the appropriate contingency imple
mentation cues, coverants appear to have great value as the
vehicles for self-contingency management. In addition,
covert methods give the subject a procedure which he can
take with him and apply under his own control whenever
temptation occurs; therefore, more frequent conditioning
trials can occur. In addition, the aversive stimulus can
be made more relevant and appropriate to the response in
question. Certainly, an important advantage is the sheer
accessibility of an appropriate reinforcer, given the range
of high probability behaviors which exist in a response
repertoire. When the issue of response probability is
obviated, we are left with covert procedures which essen
tially follow traditional operant paradigms and differ only
in the locus of the response system (i.e., interval vs.
external responses). The remainder of this paper will be
devoted to a discussion of these procedures.
D. The Covert Sensitization Procedure
Since it might be assumed that any technique aimed
at decreasing the probability of a response also affects
13
cognitions such as "desire, temptation, etc.," covert
procedures can be explicitly directed at these cognitive
responses. Gold and Neufeld (1965) developed a technique
for treating a male homosexual in which the subject was
asked to imagine potentially attractive homosexual situa
tions which were then paired with such fear provoking
images as the presence of a policeman. Cautela (19 66)
described a similar procedure that he labeled "covert
sensitization" in which the subject visualized a pleasur
able but undesirable stimulus while imagining noxious sensa
tions such as nausea and vomiting. The term "sensitization"
was "...used because the purpose of the procedure is to
build up an avoidance response to the undesirable stimulus."
(Cautela, 1967, p. 459) The following is an example of the
procedure used by Cautela (1967) in the treatment of
homosexuality:
I want you to imagine that you are in a
room with X. He is completely naked. As you
approach him you notice he has sores and scabs
all over his body, with some kind of fluid
oozing from them. A terrible foul stench comes
from his body. The odor is so strange it makes
you sick. You can feel food particles coming
up your throat. You can't help yourself and
you vomit all over the place, all over the
floor, on your hands and clothes. And now that
even makes you sicker and you vomit again and
again all over everything. You turn away and
then you start to feel better. You try to get
out of the room, but the door seems locked.
The smell is still strong, but you try desper
ately to get out. You kick at the door
14
frantically until it finally opens and you
run out into the nice clean air. It smells
wonderful. You go home and shower and you
feel so clean. (p. 464}
This would appear to be essentially a two-process
avoidance procedure in which, after repeated trials, the
subject reports the absence of temptation for the stimulus.
"The cues which have been previously associated with the
noxious stimulation of nausea and vomiting now have become
discriminative stimuli for avoidance behavior." (p. 461)
Covert sensitization has been applied in a variety
of cases: sadistic fantasy (Davison, 1968), attitude change
(Cautela and Wisocki, 1969), obesity (Cautela, 1966, 1967;
Harris, 1969; Manno and Marston, 1970), smoking (Mullen,
1968), alcoholism (Anant, 1967, 1968, 1968a; Ashem and
Donner, 1968; Cautela, 1967), homosexuality (Barlow,
Lietenberg and Agras, 1968; Cautela, 1967), and a clothes-
folding compulsion (Wisocki, 1969). However, the most
important questions concerning the efficacy of covert sen
sitization remain to be answered:
(1) Is Covert Sensitization an Effective Means
of Controlling Mai-Adaptive Approach Behavior? In an
attempt to answer this question, Ashem and Donner (1968)
employed three groups in a controlled study of covert sen
sitization with alcoholic subjects. In a forward condi
tioning group, subjects imagined raising a glass of alcohol
15
to their lips followed by the image of nausea and vomiting.
In a backward conditioning group, nausea preceeded the
image of alcohol. The third group was a no-contact control.
It became evident early in treatment that the backward con
ditioning group was, in fact, receiving treatment. The for
ward and backward groups were then collapsed and combined
and compared, as one group, with the control. Therefore,
as the study was reported, treatment (or contact, if you
will) was compared with no treatment (or no contact) and
questions of the role of expectancy or other demand char
acteristics remain unanswered. Given these limitations,
the study reported a significant decrease in alcohol con
sumption for the aversive group as compared with the
control.
In a more adequately controlled design, Manno and
Marston (1970) attempted to answer question one and
question two.
(2) Is Covert Sensitization More Effective Than
Covert Positive Reinforcement for Appropriate Behavior?
Manno and Marston compared covert sensitization with both
positive covert reinforcement and non-specific therapy in
group treatments with obese subjects. The covert sensiti
zation group followed essentially Cautela's (1967) proce
dure in which subjects were asked to imagine food which
was paired with covert nausea and vomiting. In the covert ;
16
positive reinforcement group, subjects were asked to
imagine resisting the temptation to eat the imagined food
and were covertly reinforced with positive imagery {i.e.,
feeling wonderful, losing weight, etc.). All groups met
twice a week for four weeks. It might be pointed out,
parenthetically, that the question of the appropriateness
of the reinforcing imagery arises regardless of whether you
are using a positive or negative procedure. The problem is
akin to that existing in traditional aversion therapy or
positive reinforcement training. Cautela (1967a) developed
a Reinforcement Survey Schedule to evaluate the effective-
ness of particular reinforcers, while Manno and Marston
(1970) used a reinforcement rating scale, both of which
rely upon self-report.
Returning to the results of the study at hand, both
aversive and positive groups had significantly greater mean
weight losses post-treatment than the control, with no sig
nificant differences between the two experimental groups.
However, between post-treatment and follow-up (four weeks),
only the control group showed a significant decrease in
weight, though still remaining significantly different from
the other two groups which continued to lose weight. It
was hypothesized that expectation was not equated for the
three groups initially which may account, to some extent,
for the control group's inability to perform as well during
17
the treatment period. If this is correct, then question
one yet remains unanswered. The answer to question two,
however, is clear from at least this one study: positive
covert reinforcement is at least as effective in the control
of weight as negative covert reinforcement.
Several points should be made in this regard.
Firstly, although the above studies were not able to quan
titatively assess the role of contributing variables such
as group pressure and expectation, it is clear from the
literature that the extent of their influence may be
abundant, Wellersheim (197 0) compared weight reduction in
a non-specific therapy group, a positive expectation —
social pressure groups, a no-treatment — wait control, and
a focal therapy group based on positive reinforcement,
shaping and the development of self-control. Although the
focal treatment group performed the best, all other groups
showed significant weight losses relative to the control
and clearly demonstrate the strength of environmental and
social-interactive influences.
Finally, the statement that positive reinforcement
was "at least as effective" as negative reinforcement was
used because covert sensitization employs positive rein
forcement in the form of aversion relief and the feelings
of satisfaction and well-being. This possible confounding
leads us to the next question.
18
(3) What Role Does Positive Reinforcement
(Aversion-Relief) Play in the Effectiveness of Covert
Sensitization? Several studies offer insight into this
issue. In the Gold and Neufeld (1965) study, the authors
employed, in part, an aversive deconditioning procedure
which perhaps might more appropriately have been called
aversion-relief training. A homosexual patient was
instructed to imagine a most unattractive, potentially
homosexual situation. When the patient then imagined that
he would not get involved in this situation, he was ver
bally rewarded by the therapist. Since the initial scene
was somewhat aversive, relief was provided by the subject's
abstinence. As a result of this treatment, the authors
reported that the patient had greatly improved, with no
relapse occurring during the 12 month follow-up.
On the other hand, Thorpe, Schmidt, Brown and
Castell (1964) reported encouraging results of an aversion-
relief therapy procedure which they used in treatment of
homosexuality, phobias and obsessive-compulsive behavior.
Essentially, the technique utilizes verbal representations
of stimuli and responses. Patients are presented with a
series of typed words (i.e., "homosexual," "fear of clouds,"
etc.) which are paired with shock. The last word in the
series signals the termination of shock and is a word repre
senting appropriate behavior (i.e., "heterosexual").
19
Given the success reported by these authors, it
does not seem unreasonable to expect that, in the absence
of positive reinforcement or aversion-relief, covert sensi
tization would be far less effective. The assumption here
is, of course, that the two reinforcements (positive and
negative) work additively and interact to produce avoidance
behavior, an assumption that remains to be tested.
(4) Is Covert Sensitization More Effective Than
Traditional Aversion Therapies? Unfortunately, the litera
ture offers little insight into this question. Weiner
(1965) found that both covert aversion and the overt occur
rence of the same negative events produced behavior change,
as compared with a no-feedback condition, although the
imagined outcomes had somewhat less of an effect. However,
even if it were found that the two forms of presentation
are equally effective, we would still be at a loss to know
the origin of our results since demand characteristics and
social-interactive influences have not been adequately con
trolled in either technique.
(5) Is Relaxation a Necessary Condition for the
Effectiveness of Covert Sensitization?
(6) To What Extent Does Response Suppression
Generalize to Other Similar Mai-Adaptive Behaviors? Ques
tions five and six have not been empirically tested,
20
although, in terms of question five, Cautela typically
employs Wolpe's (1958) relaxation procedure. Additionally,
there is no data in the literature on the question of gene
ralization other than that reported from single-case
studies. Cautela (1966) stated that:
An important characteristic of the covert
sensitization procedure is that its effects are
very specific. If one treats for aversion to
beer, there will be very little generalization
to wine and whiskey. Avoidance to wine and
whiskey must be treated separately. Sometimes
I combine a covert sensitization trial for wine,
beer and whiskey by having the client see a
glass of wine, a glass of beer, and a glass of
whiskey on a table. As in the manner described
above, he is told that he is sick and he vomits
over all three beverages. (p. 40)
(7) Is Suggestibility or Vividness of Imagery a
Necessary Condition for the Effectiveness of Covert
Sensitization? An interesting result found by Manno and
Marston (1970) was that vividness of imagery did not corre
late with weight loss, although weight loss did appear to
be treatment specific. One might interpret this as meaning
that stimulus clarity need only reach a minimal level for
recognition, beyond which it is unrelated to performance.
Since all subjects reported their ability to imagine the
scenes as directed, there was no chance to assess weight
loss for subjects who were unable to do so. Without further
empirical evidence, then, we are left with this question
unanswered.
21
(8) At What Place, in the Higherarchy of Target
Behaviors, Should One Begin Covert Sensitization Treatment?
Unfortunately, no studies in the covert sensitization
reports have answered this question. It has been the prac
tice, however, to begin with the behaviors at the top of
the hierarchy. Cautela suggests some intuitive preference
for this custom:
A practical problem still exists concerning
whether to proceed first with the kind of
drinking he does most often in the most usual
situations or to begin covert sensitization
with the type of drinking and its situations
which occur the least often. For the most
part, I have used the first method. The primary
advantage of the second method, however, is the
provision of some measure of success since it
involves the least amount of habit strength
and will make the client more eager to continue
treatment. (Cautela, 1967, p. 461)
(9) Where, in the Chain of Events, is the Most
Effective Placement of the Aversive Imagery in a Covert
Sensitization Procedure? This is perhaps the most interest
ing question to approach, since its answer must ultimately
deal with the host of criticisms which have been leveled
against behavior modification in general and aversion
therapy in specific. The question arises, in part, because
the temporal placement of the negative stimulus varies for
different experimenters. Anant (1966, 1967, 1968) reported
that in his treatment of alcoholism he asked the subject to
imagine taking several drinks which ultimately lead to
22
nausea. On the other hand, he employed trials in which
the mere image of alcohol was paired with aversion. Alter
nately, Cautela (1966, 1967) paired nausea with the image
of liquor brought to the mouth, rather than the consummatory
behavior while Ashem and Donner (1968) reported using the
former method in a test of covert sensitization.
Additionally, Cautela interspersed punishment
trials with trials in which punishment was absent, that is,
partial reinforcement. It would not be difficult to con
ceive of non-reinforcement trials as trials in the practice
of "self-control" reinforced by the absence of punishment.
Alternatively, these trials might be thought of as trials
in which substitute responses for mal-adaptive overt
behavior are strengthened by the secondary reinforcement
qualities of the image. If, indeed, non-reinforced trials
offer a significant advantage over trials in which punish
ment occurs, a serious theoretical problem would have to
be dealt with.
CHAPTER II
THE RESEARCH PROBLEM
Since the issue of the timing of punishment has
not been dealt with in the covert sensitization literature,
this study primarily addressed itself to that problem. The
study was designed to allow for evaluation of several
questions: (1) Is covert punishment, instituted early in
an imagined response chain of eating behaviors, more or
less effective in producing suppression of mal-adaptive
eating behaviors than punishment applied after the occur
rence of the consummatory response; (2) Will punishment
of both anticipatory and consummatory eating behaviors be
more effective than the punishment of either response
alone; (3) Will the occurrence of images followed by
punishment be more effective in producing response suppres
sion than the same images occurring in the absence of
punishment. (Effectiveness was defined as weight loss
among a group of subjects who defined themselves as over
weight and who participated in the program labeled as a
weight reduction clinic).
The secondary purpose of this study was to allow
for correlational analyses of various indeces with initial
weights and post-treatment weight losses.
23
24
In terms of the primary research problems,
Aronfreed's (1968) conceptualization of symbolic control
is relevant. Essentially, he suggested that behavior is,
in part, determined by its anticipatory consequences. That
is, through a variety of cues, such as the prior conse
quences of a particular behavior, an individual cognitively
evaluates the probability that a given response will be
punished or rewarded. These symbolically produced antici
pations mediate behavior to inhibit or facilitate the
response in question. In the case of behavior resulting in
aversive consequences, it was assumed that anxiety is
classically conditioned to intrinsic proprioceptive cues
and their symbolic representations. Anxiety, then, is
aroused by stimuli which accompany anticipatory movements.
The behavior is inhibited, resulting in the self-reinforc
ing reduction of anxiety.
According to this interpretation then, punishment
occurring early in the behavioral chain will condition
anxiety to cues which accompany anticipatory responses;
therefore, behavior will be inhibited at an early stage.
On the other hand, where punishment follows a transgression,
anxiety will not be aroused until the act is completed,
resulting in less effective internalized response
suppression.
25
Mower's (1960) discussion of conditioned emotional
responses offers a similar interpretation of the timing of
punishment. In the active avoidance learning paradigm, a
formerly neutral external cue which elicits an avoidance
response, does so by its capacity to arouse fear. In the
case of passive avoidance learning, the conditioned stimuli
are proprioceptive cues. With the occurrence of a trans
gression, there is a sequence of responses, each accompanied
by response-produced feedback. If punishment follows the
completion of the act, fear will be associated with the
stimuli surrounding the consummatory behavior. If punish
ment occurs at the initiation of the transgression, fear
will be associated with the response-produced cues occurr
ing at that time.
On the other hand, Solomon (1964) suggested an
alternative relationship between the placement of reinforce
ment and response suppression, that suppression will be
increased when punishment is applied to the consummatory,
rather than the instrumental act:
Perhaps the intimate temporal connection
between drive, incentive, and punishment
results in drive or incentive becoming con-
ditioned-stimulus (CS) patterns for aversive
emotional reactions when consummatory acts
are punished. Perhaps this interferes with
vegetative activity: i.e., does it 'kill the
appetite1 in a hungry subject? But one may
ask why the same punisher might not appear
to be as effective when made contingent on an
26
instrumental act as contrasted with the con
summatory act. Perhaps the nature of operants
is such that they are separated in time and
space and response topography from consummatory
behavior and positive incentive stimuli, so that
appetitive reactions are not clearly present
during punishment of operants. (p. 242)
Perhaps the reason for these conflicting interpre
tations is that the effects of punishment are determined by
a wide variety of factors, only one of which is its place
ment in the response chain. Solomon (1964) has shown that
punishment can either partially or completely suppress a
response previously established by positive reinforcement,
depending upon the intensity of the punishing stimulus,
while Azrin (1959) demonstrated a response increase when
operants were punished. However, given that stimulus inten
sity remains constant across conditions, there would appear
to be sufficient evidence to support the former predictions.
Aronfreed and Reber (1965) found, with children, that
punishment at the initiation of an act was more effective
in producing internalized response suppression than punish
ment at the completion of the act. Similar results have
been reported by Bixenstine (1956) and Kamin (1959).
Solomon, Turner, and Lessac (1958) studied the effects of
delay of punishment on resistance to temptation in dogs
and found that long delays of punishment produced much
weaker resistance to temptation to eat than short delays.
Finally, Walters and Demkow (196 3) punished children with
27
a loud, unpleasant noise either early in a deviant response
sequence or after the deviation had occurred and found the
earlier punishment to be more effective in inhibiting the
response.
The present study, in part, tested this relationship
in an adaptation of the covert sensitization procedure
employing five groups of overweight subjects:
Group I: (CA), Consummatory Aversion, was punished
covertly following consummatory eating behaviors
in imagination;
Group II: (AA), Anticipatory Aversion, was punished
covertly following anticipatory eating behaviors in
imagination;
Group III: (ACA), Antipatory and Consummatory
Aversion, was punished covertly following both
anticipatory and consummatory eating behaviors in
imagination;
Group IV: (Cl), Consummatory Imagery, received con
summatory imagery with no aversive consequences and
served as the control for the Consummatory Aversion
group;
Group V: (AI), Anticipatory Imagery, received anti
cipatory imagery with no aversive consequences and
28
served as the control for the Anticipatory Aversion
group.
The hypotheses of this study led to the predictions
that the Anticipatory Aversion (AA) group would lose signi
ficantly more weight than the Consummatory Aversion (CA)
group, since the preceeding evidence indicated that punish
ment occurring at the initiation of a transgression tends
to suppress that behavior more effectively than punishment
following the occurrence of the consummatory response.
Secondly, it was predicted that the Anticipatory
and Consummatory Aversion (ACA) group would be superior to
both the Anticipatory Aversion (AA) and Consummatory
Aversion (CA) groups, since (ACA) was punished throughout
the behavioral chain. It was thus assumed that any response
inhibition resulting from both anticipatory and consumma
tory punishment would be additive.
Finally, since there was no a priori basis for pre
dicting treatment-specific behavior changes in the two
imagery controls (AI, Anticipatory Imagery and Cl, Consum
matory Imagery), it was predicted that each of the aversive
groups would show significantly greater weight losses
relative to these control groups with no predicted differ
ence between the latter two.
CHAPTER III
METHOD
Subjects
The subjects for this experiment were nine males
and 67 females ranging in age from 17 to 46 years. Of the
76 subjects, 62 were part or full-time students at the
University of Southern California and 14 were University
personnel. A notice appeared in the campus newspaper
informing interested people of a research and treatment
program for weight reduction. All applicants who met the
following criteria were selected for the program: (1) A
minimum desired weight-loss of 15 lbs.; (2) Good physical
condition and not under the care of a physician for a
medical or weight-related purpose during the course of the
program; (3) A deposit of $15.00 refunded at the conclu
sion of the program. In addition, no applicant was
selected who had participated in a previous weight-reduc-
tion program at the University (Manno and Marston, 1970).
Materials
Materials consisted of: (1) Standard Detecto
Bathroom Scale; (2) Personal History Questionnaire;
29
30
(3) Body Image Form; (4) Food Approach Questionnaire;
(5) Food Preference Scale; (6) 900-calorie diet; (7) I-E
Scale (Rotter, 1966); (8) Achievement Scales for Males and
Females (Mehrabian, 196 9).
The Body Image Form consisted of a group of draw
ings of body contours from which subjects were required to
identify the one that matched their Ideal body and the one
that matched their Real body. This provided an index of
change in Real or Ideal body image from pre to post-treat
ment in addition to the discrepancy between Real and Ideal
body image. The Discrepancy Score, then, is the difference
between the Real and Ideal body scores.
The Food Approach Questionnaire is a rating scale
adapted from items on the Smoker's Self-Testing Kit
developed by the Los Angeles County Interagency Council on
Smoking and Health. The Food Approach Questionnaire was
designed to assess responses to eating. It yields two
scores. The Habit Score measures the tendency of the sub
ject to reach for food "automatically." That is, eating
behaviors are often under the control of a variety of
environmental stimuli. It is not unusual to find, that as
a person watches T.V. , she consumes half a box of candy
without realizing it. In fact, treatment approaches often
are aimed at bringing eating under limited, highly specific
stimulus control (Goldiamond, 1965). The Habit Score
31
measures the extent of these "automatic" behaviors by having
the subject rate the frequency with which they occur. An
example of one of the items is: "I begin to reach for more
food without realizing that I still have some left on my
plate."
The second score of the Food Approach Questionnaire
is the Total Food Approach Score which measures such vari
ables as relaxation afforded by food, craving for food and
so forth. Again, subjects are asked to rate the frequency
with which specific behaviors occur. Examples of some of
these items are: "I eat in order to prevent myself from
getting bored."; "When I feel 'blue' or want to take my
mind off cares and worries, I eat."; "I eat when I feel
angry about something." The Total Food Approach Score
includes the Habit Score as one of its indeces.
The Food Preference Scale was designed to assess
the subject's most problematic foods, the physical environ
ments in which the subject was most prone to eat, the nega
tive consequences of being overweight and the positive
consequences of weight loss.
Procedure
All pre-treatment data were obtained during the
enrollment period, at which time subjects, seen individually,
were given the following information:
32
The Psychological Research and Service Center
of U.S.C, is now conducting a treatment and
research program for weight reduction. The
program is approximately three weeks in length
and consists of two treatment sessions per week.
Each treatment session will run between 30-45
minutes. The weight reduction program is
based upon new techniques developed in clinical
psychology that treat eating as a habit which
follows the laws of learning. Therefore, eating
behaviors can be brought inder control through
the appropriate use of behavioral principles
and methodology. No drugs or medication will
be used.
Subjects were then assigned on a random basis to
one of five groups (the male subjects and staff personnel
were distributed evenly among the groups): Group I: (CA),
Consummatory Aversion (covert punishment following consum
matory eating behaviors in imagination); Group II: (AA),
Anticipatory Aversion (covert punishment following antici
patory eating behaviors in imagination); Group III: (ACA),
Anticipatory and Consummatory Aversion (covert punishment
following both anticipatory and consummatory eating behavi
ors) ; Group IV: (Cl), Consummatory Imagery (consummatory
imagery in the absence of aversive consequences); Group V:
(AI), Anticipatory Imagery (anticipatory imagery in the
absence of aversive consequences).
Each group was divided into two sections. Sessions
one-six of each section were run alternately by two
experimenters (Es). Thus, each treatment group was led for
three sessions by one E, and three by the other E. Both
33
experimenters were male graduate students in clinical
psychology with essentially similar backgrounds in group
therapy and research. Each E was given a therapy manual
which consisted of introductory remarks, treatment proce
dures (which were followed verbatim), and sample questions
and answers. Pre and post-treatment data were collected
by the author who otherwise did not participate in the
study.
Each section was seen, as a group, for six sessions,
two each week. The sessions ran approximately 30-45
minutes. A four-week follow-up was conducted, at which
time subjects were weighed and deposits returned, provided
subjects had completed all phases of the program and had
not missed more than two treatment sessions.
Group I: (CA) Consummatory Aversion.
Session One.— The following introductory remarks
were made:
As you know, this is both a treatment and
research program. We want you to follow the
diet you were given, or one of your own choos
ing, as rigorously as possible. It is extremely
important that you attend all of the treatment
sessions. We believe that eating is a habit
like any other habit. It's been found that
habits can be controlled through various kinds
of imagery. In fact, imagining a response can
serve as a substitute for making the response.
In the past you probably imagined eating and
then went out to eat. Here we will learn to
stop that sequence of events. That is, you
will learn to imagine food without going out
34
and actually eating. We will do this by
associating eating with negative reinforcement.
You will imagine eating certain foods and this
will become associated with an unpleasant
experience. You will learn to have less desire
to eat certain fattening foods. This procedure
will not make you stop eating altogether, but
will prevent you from eating foods you shouldn't
eat. Are there any questions?
(These instructions were given to all groups. For
the two non-aversive groups, Anticipatory Imagery and Con
summatory Imagery, however, the underlined sentences were
omitted, since these groups received no punishment.)
Following the instructions, subjects were asked to
keep a record of how often they used the procedures at
home; they were told that this would be collected at the
end of the program. Diets were distributed and subjects
were weighed. Individuals1 weights were not disclosed to
the rest of the group. Subjects were then given simple
relaxation instructions in which they were asked to close
their eyes, get as comfortable as they could and imagine
pleasnat, non-arousing images. They were further asked to
raise their hand if, at any time during the course of
treatment, they were not experiencing the instructions as
directed. The group was then given the following treatment
instructions:
I want you now to imagine that you have a
specific food in your mouth; the food you
listed on the questionnaire last week as being
most problematic. Imagine the food in your
35
mouth. Is there anyone who cannot imagine
this? I want you to imagine yourself in your
living room watching T.V. with your friend
sitting next to you. Imagine yourself seated
in the room; imagine your friend. Now imagine
eating the food; you're chewing the food; you
can taste it; you can feel your mouth salivat
ing as you eat it. Eat the food; chew it;
feel the food in your mouth; taste it; smell
the aroma of the food. Now you're beginning
to feel sick; you're starting to feel sick to
your stomach. Is there anyone who doesn't
feel ill? The vomit is beginning to come up
in your mouth; you're nauseaus; you can feel
the mucousy vomit rolling around in your mouth.
It's oozy and it smells. Is there anyone who
doesn't feel ill? You've still got the food
in your mouth. You can't hold the vomit in any
longer; it pours out all over you, all over
your hands and clothes, all over your friend
sitting next to you. Imagine the vomit is
still coming out of your mouth and nose. You're
puking all over everything. Feel the puke,
feel how embarrassed you are.
Four more treatment trials were given, each trial
separated by a 30-second interval. Eyes remained closed
throughout the entire procedure. The session concluded
with a discussion period which had primarily to do with
questions related to the procedure and the home practice of
the treatment imagery. All questions asked were answered
with reference to the standard question and answer sheets
given to each E.
Sessions 2-6.— These sessions were essentially the
same as the first with the following exceptions: (1) Sub
jects imagined other foods which they had previously
listed on the hierarchy as being problematic; (2) Subjects
36
imagined eating in different environments. These environ
ments were selected from the Food Preference Scale and con
stituted the most often reported rooms in which eating
occurred.
With the exceptions of the underlined sentences in
the introductory remarks, and the specific treatment imagery,
all other aspects of the procedure remained the same for
the five groups during each treatment session. Therefore,
only the differential treatment instructions will be
listed here.
Group II: (AA) Anticipatory Aversion.
I want you now to think of a specific class
of foods, the one you listed on the questionnaire
last week as being most problematic. I want you
to see the food. Is there anyone who cannot
see it? I want you now to see yourself in your
living room watching T.V. with your friend sit
ting next to you. Imagine yourself seated in the
living room; imagine your friend; now imagine
the food. Now imagine wanting the food. You
want the food; you want to eat it. Reach for the
food, pick it up and bring it to your lips.
Imagine the food at your lips. Is there anyone
who cannot imagine this? Now you're beginning
to feel sick; you're starting to feel sick to
your stomach. The vomit is beginning to come up
in your mouth; you're nauseaus; you can feel the
mucousy vomit rolling around in your mouth.
Feel the vomit in your mouth. It's oozy and it
smells. Is there anyone who doesn't feel ill?
You're still holding the food in your hand.
You can't hold the vomit in any longer; it pours
out all over the food, all over your hands and
clothes, all over your friend sitting next to
you. Imagine the vomit gushing out of your
mouth. The yellow, mucousy vomit is still
coming out of your mouth and nose. You're
37
puking all over everything. Feel the puke,
feel how embarrassed you are.
Group III: (ACA) Anticipatory and Consummatory
Aversion.
I want you now to think of a specific class of
foods, the one you listed on the questionnaire
last week as being most problematic. I want you
to see the food. Is there anyone who cannot see
it? I want you now to see yourself in your
living room watching T.V. with your friend sit
ting next to you. Imagine yourself seated in
the room; imagine your friend; now imagine the
food. Now imagine wanting the food. You want
the food; you want to eat it. Reach for the food;
pick it up and bring it to your lips. Imagine
the food at your lips. Now you're beginning to
feel nauseaus; you're starting to feel sick.
Your stomach is starting to turn. Is there any
one who doesn't feel nauseaus? Now imagine
eating the food; you're chewing the food; you
can taste it. Eat the food, chew it; feel the
food in your mouth; taste it; smell the aroma of
the food. You're feeling sick to your stomach.
The vomit is coming up in your mouth; you're
nauseaus; you can feel the mucousy vomit rolling
around in your mouth. Feel the vomit in your
mouth. Is there anyone who cannot imagine this?
It's oozy and it smells. You've still got the
food in your mouth. You can’t hold the vomit
in any longer; it pours out all over you; all
over your hands and clothes, all over your
friend sitting next to you. Imagine the vomit
gushing out of your mouth. The yellow mucousy
vomit is still coming out of your mouth and nose.
You're puking all over everything; feel how
embarrassed you are.
Group IV: (Cl) Consummatory Imagery.
I want you now to imagine that you have a
specific food in your mouth; the food you listed
on the questionnaire last week as being most
problematic. Imagine the food in your mouth.
Is there anyone who cannot imagine this? I want
38
you to imagine yourself in your living room
watching T.V. with your friend sitting next
to you. Imagine yourself .seated in the living
room; imagine your friend. Now imagine eating
the food; you're chewing the food; you can
taste it; you can feel your mouth salivating as
you eat it; eat the food; chew it; feel the
food in your mouth; taste it; smell the aroma
of the food. Is there anyone who cannot imagine
this? Chew the food slowly; feel the texture of
the food. Imagine yourself eating the food.
Imagine the food in your mouth. You can taste
its flavor as you're chewing it. Feel the
saliva in your mouth. Eat more of the food.
Bite into it; see the food; see its color;
smell it; taste it.
Group V: (AI) Anticipatory Imagery.
I want you to think of a specific class of
foods; the one you listed on the questionnaire
last week as being most problematic. I want you
to see the food. Is there anyone who cannot see
the food? I want you now to imagine yourself in
your living room watching T.V. with your friend
sitting next to you. Imagine yourself seated in
the living room; imagine your friend. Now imagine
the food. Now imagine wanting the food. You
want the food; you want to eat it. Walk over to
the food. Imagine walking over to the food.
Look at its color, look at its texture. Imagine
that you want to eat it. Imagine yourself going
into the kitchen to get a plate and a napkin.
Go to the kitchen and get a platej get a napkin.
Now come back to the food. Is there anyone who
cannot imagine this? Sit down. Imagine yourself
seated; see the food. Now reach for the food.
Pick it up. It's now in your hand. Look at the
food. Smell its aroma; feel the texture of the
food. Bring the food to your lips. Imagine the
food at your ]ips.
CHAPTER IV
RESULTS
Of the 88 subjects initially enrolled in the prog
ram, 11 subjects were lost. Of these 11 subjects, two had
enrolled, but never began treatment, while the remaining
seven dropped out after the first or second session of
treatment. The differential drop-out by treatment was as
follows: The Consummatory Aversion (CA) group lost three
subjects? Anticipatory Aversion (AA), three; Anticipatory
and Consummatory Aversion (ACA), two; Consummatory Imagery
(Cl), one? Anticipatory Imagery (AI), two. In spite of the
loss of subjects, no significant difference was found
between the groups' mean weights (F^ ^ = .224) (Table 2).
All data, with the exceptions of specified corre
lations, were analyzed in terms of the arc sines of percent
age weight loss. That is, weight loss (in lbs.) was
converted into percentage weight loss, which in turn was
transformed into arc sine scores.
Weight Loss
A. Test of Hypotheses: Treatment Differences
A treatment by trials analysis of variance was per
formed on the arc sines of percentage weight loss. The
analysis yielded a non-significant group main effect
39
40
(F = 1.285, df = 4,71), a non-significant interaction
effect (F = .926, df = 20,355) and a significant trials
effect (F = 6.593, df = 5,355, p. < .01) (Table 4). A
Newman-KueIs test revealed that weight losses recorded on
Trial Trial 2 were significantly less than cummulative
weight losses for Trails 3 through follow-up. The differ
ence between weight loss at Trial 2 and Trial 3 was 6.693;
Trial 2 and Trial 4 was 5.97 6; Trial 2 and Trial 5 was
10.268; Trial 2 and Trial 6 (post-treatment) was 6.618;
Trial 2 and Follow-up was 9.800. All values were signifi
cant beyond .01. The difference in weight loss between
Trials 3 and 4 was .717; Trials 3 and 5 was 3.575; Trials 3
and 6 was .075; Trial 3 and Follow-up was 3.107; Trials 4
and 5 was 4.292; Trials 4 and 6 was .642; Trial 4 and
Follow-up was 3.824; Trials 5 and 6 was 3.650; Trial 5 and
Follow-up was .468; Trial 6 and Follow-up was 3.182. None
of these values were significant at the .05 level.(Table 5).
Although the analysis of variance did not yield a
significant treatment effect, individual t tests were per
formed to test whether or not punishment, following the
treatment imagery, was more effective in suppressing mal
adaptive eating responses than the same imagery without
punishment. When the three aversive groups were combined
and the two non-aversive groups combined, a t test on mean
post-treatment weight loss (arc sines of percentage weight
41
loss) revealed that the non-aversive groups lost signifi
cantly more weight than the aversive groups (t = 2.007,
df = 74, p. < .05). This was not the case at follow-up
where no significant difference between the two was found
(t = .5197, df = 74). When the two non-aversive groups were
tested against each other at post-test, the difference in
weight loss was not significant (t = .758, df = 32), nor
was the difference significant at Follow-up (t = 1.332,
df = 32). When the three aversive groups were tested
against each other at post-treatment, the differences in
weight Ddss were not significant: Anticipatory Aversion (AA)
and Anticipatory and Consummatory Aversion (ACA) was
t = 1.048, df = 28; Anticipatory Aversion (AA) and Consumma
tory Aversion (CA) was t = 1.283, df =26; Anticipatory and
Consummatory Aversion (ACA) and Consummatory Aversion (CA)
was t = 1.295, df = 24. At Follow-up there were again no
significant differences in weight loss between these groups:
Anticipatory Aversion (AA) and Anticipatory and Consummatory
Aversion (ACA) was t = .9474, df = 28; Anticipatory Aver
sion (AA) and Consummatory Aversion (CA) was t = .686,
df = 26; Anticipatory and Consummatory Aversion (ACA) and
Consummatory Aversion (CA) was t = .206, df = 24.
B. Correlations Related to Weight Loss
Correlations between post-treatment weight loss and
follow-up weight loss were all significant with the excep-
42
tion of the Anticipatory Aversion group where a strong
trend towards significance was found (r = .3959, df = 15).
Correlations for the other groups were as follows: AI:
r = .5799, df = 15, p. < .01; Cl: r = .4942, df = 17,
p. < .05; ACA: r = .7260, df = 13, p. < .005; CA: r = .5263,
df = 11, p. < .05. That is, subjects who lost the most
amount of weight at post-treatment tended to lose the most
weight at follow-up (Table 19).
Post-treatment weight loss was also correlated with
desired weight loss reported in Session 1. This correlation
was significant for the Consummatory Aversion group:
r = -.6233, df = 11. Subjects who reported the largest
desired weight loss lost the least at post-treatment. This
relationship did not exist for the other groups: AI:
r = -.0479, df = 15; Cl: r = -.2159, df = 17; ACA:
r = .4131, df = 13; AA: r = .0122, df = 15 (Table 18).
Eating Behaviors
1. Total Food Approach Scores
A. Treatment Differences
A treatment by trials analysis of variance was per
formed on Total Food Approach Scores. The results indicated
that only the trials main effect was significant (F = 10.633,
df =1,61, p. < .01). That is, there was a significant
43
decrease in Total Food Approach Scores over the course of
treatment, across groups. The group main effect and inter
action effect were as follows: F = 1.592, df = 4,61;
F = .699, df = 4,61, respectively (Table 6, Fig. 2).
B. Correlations Related to Total Food Approach
Total Food Approach cores recorded in Session 1
(T^) did not significantly correlate with pre-treatment
weights (r* = -.7283, df = 13, p. < .005). That is, for
this group, subjects who had the greatest tendency to eat
prior to treatment lost least at the end of the program.
Correlations for the other groups were not significant:
AI: r = .2398, df = 15; Cl: r = .0327, df = 17; AA:
r = -.1662, df = 15; CA: r = -.1378, df = 11 (Table 13).
At follow-up, there was no significant relationship between
T^ and weight loss (AI: r = .3914, df = 15; Cl: r = -.0478,
df = 16; ACA: r = -.2960, df = 13; AA: r = -.0690,
df = 15; CA: r = .0190, df = 11) (Table 14).
Total Food Approach cores recorded on Session 6
(T2 ) did not significantly correlate with pre-treatment
weights (r = .2131, df = 74). T2 highly correlated with
post-treatment weight loss for the Anticipatory and Consum
matory Aversion group (ACA) (r = -.8654, df = 13, p. < .005).
*
Unless otherwise indicated, r is across all groups.
44
That is, for this group, subjects who had the greatest
tendency to eat at the end of the program lost least.
Correlations for the other groups were not significant:
AI: r = -.1464, df = 15; Cl: r = .3001, df = 17;
AA: r = -.2176, df = 15; CA: r = -.1783, df = 11 (Table
13). At follow-up, there was no significant relationship
between T2 and weight-loss: AI: r = .2389, df = 14;
Cl: r = .2176, df = 13; ACA: r = -.4961, df = 10;
AA: r = -.3277, df = 13; CA: r = .1552, df = 10 (Table 14).
Changes in Total Pood Approach Scores from Session
1 to Session 6 (Ti”T2^ highly correlated with pre-treatment
weights (r = .3913, df = 65, p. < .005). That is, subjects
who weighed the most prior to treatment had the least
decrease in their tendency to approach food over the course
of the program. T^-T2 did not correlate with post
treatment weight loss within the five groups: AI: r = .4226,
df = 14; Cl: r = -.2594, df = 15; ACA: r = .1987, df = 10;
AA: r = .0954, df = 13; CA: r = .0670, df = 10 (Table 13).
At follow-up, again there was no significant relationship:
AI: r = .1210, df = 14; Cl: r = -.2318, df = 13; ACA:
r = .1353, df = 10; AA: r = .3536, df = 13; CA: r = -.1548,
df = 10 (Table 14).
Correlations between Total Food Approach scores
recorded in Session 1 (T^) and Total Food Approach Scores
recorded in Session 6 (T2) were significant for groups AI
45
(Anticipatory Imagery); Cl (Consummatory Imagery); and ACA
(Anticipatory and .Consummatory Imagery). (r = .5982,
df = 15, p. < .01; r = .5541, df = 15, p. .05; r = .9344,
df = 10, p. < .005, respectively). That is, for these
groups, subjects who had a great tendency to approach food
prior to treatment had a great tendency to approach food at
the end of treatment. A similar trend existed for AA
(Anticipatory Aversion) and CA (Consummatory Aversion),
although these correlations did not reach significance
(r = .3370, df = 11; r = .4681, df = 10, respectively)
(Table 19).
2. Habit Scores
A. Treatment Differences
A treatment by trials analysis of variance was done
on Habit Scores. The treatment effect was not significant
at the .05 level (F = .498, df = 4,61). The trials effect
(Session 1-Session 6) revealed a strong trend towards sig
nificance at the .05 level (F = 3.97, df = 1,61). Signifi
cance would have required that F = 4.0 for df = 1,61
(Table 7, Fig. 3).
B. Correlations Related to Habit Scores
Habit Scores recorded on Session 1 (H^) did not
significantly correlate with pre-treatment weights
46
(r = .1860, df = 74), nor did they correlate with post-
treatment weight loss within each of the five groups:
AI: r = .0412, df = 15; Cl: r = .3206, df = 17; ACA:
r = -.3746, df = 13; AA: r = -.2160, df = 15; CA:
r = -.3350, df = 11 (Table 13). At follow-up, again there
were no significant relationships between and weight loss
AI: r = .2974, df = 15; Cl; r = .0671, df = 16; ACA:
r = -.0547, df = 13; AA: r = -.2141, df = 15; CA:
r = -.1117, df = 11 (Table 14).
Habit Scores recorded on Session 6 (H2) did not
significantly correlate with pre-treatment weights
(r = .2023, df = 65). H2 did, however, correlate with
post-treatment weight loss for group ACA and group CA
(r = .5955, df = 10, p. < .025; r « -.4978, df = 10,
p. < .05, respectively). That is, subjects with the
greatest food Habit Score at post-treatment lost least at
post-treatment. Correlations for the other groups at post
treatment were not significant: AI: r = -.3777, df = 14;
Cl: r = .3062, df = 15; AA: r = .0011, df = 13 (Table 13).
At follow-up, only the Anticipatory Aversion group (AA)
showed a significant correlation between H2 and follow-up
weight loss (r = .4477, df = 13, p. < .05). That is,
subjects in this group who had the greatest eating habit
score at the end of treatment lost the least at follow-up.
This relationship did not, however, exist for the other
47
groups: AI: r = .1508, df = 14; Cl: r = -.2437, df = 13;
ACA: r = -.1738, df = 10; CA: r = -.2338, df = 10
(Table 14).
Changes in Habit Scores from Session 1 to Session 6
(H^-H2) did not correlate with pre-treatment weights
(r = 0.2135, df = 65) nor did they correlate with post
treatment weight loss for the five groups: AI: r = .3701,
df = 14; Cl: r = .1751, df = 15; ACA: r = .4182, df = 10,
AA: r = -.0811, df = 13; CA: r = .2135, df = 10 (Table 13).
At follow-up, there were again no significant relationships
between H^-H2 and follow-up weight loss: AI: r = .1322,
df = 14; Cl: r = .3907, df = 13; ACA: r = .1168, df = 10,
AA: r = .2612, df = 13; CA: r = .1618, df = 10 (Table 14).
Correlations between Habit Scores, Session 1 (H-^)
and Habit Scores, Session 6 (H2) were significant for each
of the five groups: AI: r = .4606, df = 15, p. < .05;
Cl: r = .7060, df = 15, p. < .005; ACA: r = .8766, df = 10,
p. < .005; AA: r = .5681, df = 13, p. < .05; CA: r = .7580,
df = 10, p. < .005. (Table 19). That is, subjects who had
relatively high habitual eating scores prior to treatment
had relatively high habitual eating scores at the end of
treatment.
48
Body Images
1. Real Body Image
A. Treatment Differences
A treatment by trials analysis of variance was per
formed on Real Body Image Scores. Only the Trials effect
(Session 1-Session 6) was significant (F = 12.4, df = 1,62,
p. < .01). The treatment and interaction effects were:
F = 1.047, df = 4,62; F = .2017, df = 4,62). Thus, there
was a significant decrease in Real Body Image from Session
1 to Session 6 across groups (Table 8, Fig. 4).
B. Correlations Related to Real Body Image
The Real Body Image recorded in Session 1 (R^) sig
nificantly correlated with pre-treatment weights (r = .4854,
df = 74, p. < .005). That is, subjects who weighed the
most prior to treatment had the largest real body image
scores in Session 1. R-^ did not, however, correlate with
post-treatment weight loss: AI: r = .1074, df = 15;
Cl: r = .1020, df = 17; ACA: r = .0787, df = 13;
AA: r = -.1958, df = 15; CA: r = -.2739, df = 11 (Table 15).
At follow-up, there were again no relationships between
R^ and weight loss: AI: r = .0663, df = 15; Cl:
r = -.1964, df = 16; ACA: r = .3900, df = 13;
49
AA: r = -.0895, df = 15; CA: r = .4033, df = 11
(Table 16) .
The Real Body Image recorded on Session 6 (R2 ) sig
nificantly correlated with pre-treatment weights (r = .5217
df = 66, p. < .005). That is, subjects who weighed the
most prior to treatment had the largest real body images
at the end of treatment. R2 did not, however, correlate
with post-treatment weight-loss: AI: r = -.0208, df =15;
Cl: r = .0111, df = 15? ACA: r = -.2870, df = 10;
AA: r = -.2291, df = 13; CA: r = -.4595, df = 10 (Table 15)
At follow-up, there was again no significant relationship
between the two: AI: r = .1815, df = 15? Cl: r = -.0611,
df = 13; ACA: r = -.1800, df = 10; AA: r = -.1409,
df = 13; CA: r = .4084, df = 10 (Table 16).
Changes in Real Body image from Session 1 to
Session 6 (R^-R2 ) did not correlate with pre-treatment
weights (r = -.1159, df = 66). Ri“R2 however, corre
late with post-treatment weight loss for the Consummatory
Aversion group (CA): (r = .5569, df = 10, p. < .05). That
is, subjects in this group who had the greatest decrease
in eal ody mage over the course of treatment, lost the
most. Correlations for the other groups were not signi
ficant; AI: r = .1338, df = 15; Cl: r = .1268, df = 15;
ACA: r = .2741, df = 10; AA: r = .3899, df = 13 (Table 15)
At follow-up, R^-R2 did correlate with weight loss for the
50
Anticipatory and Consummatory group (r = .6302, df = 10,
p. < .05). Subjects who showed the greatest change in Real
Body Image (decrease) had the greatest weight loss at
follow-up. This relationship did not exist for the other
groups: AI: r = -.2238, df = 15; Cl: r = -.1541, df = 13;
AA: r = -.1810, df = 13; CA: r = .0724, df = 10
(Table 16) .
2. Ideal Body Images
A. Treatment Differences
A treatment by trials analysis of variance was per
formed on Ideal Body Image Scores. Neither the group main
effect, trials effect, nor interaction effect were signifi
cant (F = .536, df = 4,62; F = 2.389, df = 1,62;
F = 1.637, df = 4,62; respectively) (Table 9, Fig. 5).
B. Correlations Related to Ideal Body Image
The Ideal Body Image recorded on Session 1
highly correlated with pre-treatment weights (r = .7367,
df = 74, p. < .005). That is, subjects who weighed the
most prior to treatment had the largest Ideal Body Scores.
1^ was also found to correlate with post-treatment weight
loss for the Anticipatory and Consummatory Aversion group
51
(ACA): r = .4542, df = 13, p. < .05. That is, for this
group, subjects who had the largest Ideal Body Image score
prior to treatment lost the most at the end of treatment.
Correlations for the other groups were not significant:
AI: r = .0493, df = 15; Cl: r = .2554, df = 17; AA:
r = .1271, df = 15; CA: r = -.3563, df = 11 (Table 15).
At follow-up, again weight loss correlated with 1^ for the
Anticipatory and Consummatory group: r = .5327, df = 13,
p. < .05. Subjects who had the largest ideal Body Image
at pre-treatment lost the most at follow-up. This relation
ship again did not exist for the other groups: AI:
r = -.2086, df = 15; Cl: r = -.0724, df = 16; AA:
r = .0916, df = 15; CA: r = .3174, df = 11 (Table 16).
The Ideal Body Image recorded on Session 6 (I2)
highly correlated with pre-treatment weights (r = .5636,
df = 66, p. < .005). That is, subjects who weighed the
most prior to treatment had the largest Ideal Body Image
Scores at the end of treatment. I2 did not, however, corre
late with post-treatment weight loss (AI: r = .0529,
df = 15; Cl: r = .2935, df = 15; ACA: r = .3750, df = 10;
AA: r = .0944, df = 13; CA: r = -.2613, df = 10 (Table 15).
At follow-up, I2 did not correlate with weight loss in any
of the groups: AI: r = -.1559, df = 15; Cl: r = .2991,
df = 13; ACA: r = .2551, df = 10; AA: r = .2996, df = 13;
CA: r = .3057, df = 10 (Table 16).
52
Changes in Ideal Body Image from Session 1 to
Session 6 (I-^-^) did not correlate with pre-treatment
weight {r = .0145, df = 66), nor did they correlate with
post-treatment weight loss: AI: r = -.0085, df = 15;
Cl: r = -.2888, df = 15; ACA: r = -.0513, df = 10;
AA: r = -.1067, df = 13; CA: r = -.1100, df = 10
{Table 15) . At follow-up, c ^ c ^' however, correlate
with weight loss for the Consummatory Imagery group:
(r = -.5723, df = 13). Subjects who showed the greatest
change in Ideal Body Image at post-treatment lost the least
amount of weight at follow-up. Correlations for the other
groups were not significant: AI: r = -.1353, df = 15;
ACA: r = .1988, df = 10; AA; r = -.3002, df = 13;
CA: r = .0731, df = 10 (Table 16).
3. Body Image Discrepancy
A. Treatment Differences
The Body Image Discrepancy Score is the difference
between the Real Body Image Score and the Ideal Body Image
Score. A treatment by trials analysis of variance was per
formed on these scores. Only the Trials effect was signi
ficant (F = 20.060, df = 1,62, p. < .01). That is, there
was a significant decrease in Body Image Discrepancy Scores
i _ j
53
over the course of treatment across groups. Group and
interaction effects were not significant (F = .750,
df = 5,62; F = 1.946, df = 4,62; respectively) {Table 10,
Fig. 6) .
B. Correlations Related to Body Image Discrepancy
The difference between the Real and Ideal Body
Images recorded on Session 1 (h-^) did not significantly
correlate with pre-treatment weights (r = .0942, df = 74)
nor did it correlate with post-treatment weight loss
(AI: r = .0928, df = 15; Cl: r = -.0694, df = 17;
ACA: r = -.0633, df = 13; AA: r = -.3133, df = 15;
CA: r = -.0282, df = 11) (Table 15). At follow-up, there
were again no significant relationships between and
weight loss: AI: r = .2018, df = 15; Cl: r = -.1631,
df = 16; ACA: r = .2627, df = 13; AA: r = -.1671, df = 15,
CA: r = .3036, df = 11 (Table 16).
The difference between the Real and Ideal Body
Images recorded on Session 6 (D2 ) significantly correlated j
with pre-treatment weights (r = .2547, df = 66). That is,
subjects who weighed the most prior to treatment had the
largest Discrepancy between Real and Ideal Scores, at the
end of treatment. was also significantly correlated |
!
with post-treatment weight loss for the Anticipatory and ;
54
Consummatory Aversion group (ACA) and the Anticipatory
Aversion group (AA), (r = -.7601, df = 10, p. < .005;
r = -.4599, df = 13, p. < .05, respectively). That is,
subjects in these groups who had the largest Real-Ideal
Discrepancy Score at the end of treatment, lost least at
the end of treatment. Although correlations for the other
groups were not significant, a similar trend towards signi
ficance appeared for the Consummatory Aversion group (CA),
(r = -.4141, df = 10). The correlations for AI and Cl
were: r = -.0552, df = 15 and r - -.3463, df = 15,
respectively (Table 15). At follow-up, D^ significantly
correlated with weight loss for the Anticipatory Aversion
group: r = -.5270, df = 13. Subjects who had the smallest
Body Image Discrepancy at post-treatment lost the most at
follow-up. This did not hold true, however, for the other
groups: AI: r = .3145, df = 15; Cl: r = -.4328, df = 13;
ACA: r = -.3056, df = 10; CA: r = .3131, df = 10
(Table 16).
Changes in Discrepancy scores over the course of
treatment (D^-D2 ) did not significantly correlate with
pre-treatment weights (r = -.0691, df = 66) nor did they
correlate with post-treatment weight loss (AI: r = .3160,
df = 15; Cl: r = .2347, df = 15; ACA: r = .3053, df = 10;
AA: r = .0307, df = 13; CA: r = .2159, df = 10 (Table 15).
AT follow-up, there were again no significant relationships]
55
between the two: AI: r = -.24 63, df = 15; Cl: r = -.1056,
df = 13; ACA: r = .3992, df = 10; AA: r = .0740, df = 13;
CA: r = .0289, df = 10 (Table 16)
Motivation (Mehrabian Scale)
Motivation scores did not significantly correlate
with pre-treatment weights (r = .0355, df = 75) nor did
they correlate with post-treatment weight loss within the
five groups: AI: r = .3510, df = 15; Cl: r = -.0676,
df = 17; ACA: r = -.0994, df = 13; AA: r = -.0966,
df = 15; CA: r = .1240, df = 11 (Table 17). At follow-up,
there were again no significiant relationships between
motivation score and weight loss: AI: r = .0049, df = 15;
Cl: r = .0358, df = 16; ACA: r = -.3076, df = 13;
AA: r = .0181, df = 15; CA: r = .0858, df = 11 (Table 17).
Rotter’s Internal-External Scale
I-E scores did not significantly correlate with
pre-treatment weights (r = .0654, df = 75) nor did they
correlate with post-treatment weight loss: AI: r = -.2108,
df = 15; Cl: r = -.0387, df = 17; ACA: r = .0692, df = 15;
CA: r = .1084, df = 11 (Table 17) At follow-up, correla
tions were again non-significant: AI: r = -.2160, df = 15;
56
CIs r = -.3404, df = 16; ACA: r = .3386, df = 13;
AA: r = -.3067, df = 15; CA: r = -.2051, df = 11
(Table 17).
Program Value
A. Treatment Differences
An analysis of variance was performed on Program
Evaluation Scores. These scores consisted of self-ratings
on three items: responses to experimenters, responses to
treatment images, and responses to the total program).
There was no significant difference between groups in
terms of how valuable subjects felt the program had been
F = 2.026, df = 4,62 (Table 11).
B. Correlations Related to Program Evaluation
Program evaluation scores did not correlate signi
ficantly with post-treatment weight loss for any of the
five groups. The correlations were as follows:
AI: r = .1256, df = 15; Cl: r = .3314, df = 14; ACA:
r = .1372, df = 10; AA: r = .1213, df = 13; CA: r = -.2180,
df = 10 (Table 18).
57
Home Practice of Treatment Imagery
A. Treatment Differences
An analysis of variance was done on the frequency
of home practice of treatment imagery. It was found that
there were no significant differences between groups
(R = 1.63, df = 4,62) (Table 12).
B. Correlations Related to Home Practice of Treatment
Imagery
The number of times subjects practiced the treatment
procedure at home did not significantly correlate with
post-treatment weight loss: AI: r = -.1223, df = 15;
Cl: r = 0.1386, df = 14; ACA: r = -.3681, df = 10;
AA: r = .0109, df = 13; CA: r = .0033, df = 10 (Table 18).
Weight History
The length of time subjects were overweight prior
to the program did not significantly correlate with post
treatment weight loss: AI: r = .1523, df = 15;
Cl: r = -.2780, df = 17; ACA: r = .0393, df = 13;
AA: r = -.2841, df = 15; CA: r = -.1016, df = 11
(Table 18),
58
The number of times subjects had dieted in the past
significantly correlated with post-treatment weight loss
for the Anticipatory Aversion group: r = -.6341, df = 15,
p. < .005. That is, in this group, subjects who had
dieted most often in the past lost least during the prog
ram. Correlations for the other groups were not signifi
cant: AI: r = .1349, df = 15; Cl: r = .0644, df = 17;
ACA: r = -.3173, df = 13; CA: r = .1305, df = 11
(Table 18).
CHAPTER V
DISCUSSION
The results obtained provide only partial support
for the hypothesis that punishment occurring early in the
chain of events will be more effective in decreasing
response probability than punishment made contingent upon
the occurrence of the consummatory response. Although a
significant trials effect existed, the five treatment groups
did not significantly differ in weight loss at either post
treatment or follow-up (Tables 4 and 5). However, certain
trends appear to be evident. Of the three aversive groups,
the Anticipatory and Consummatory (ACA) group lost more
weight post-treatment than the other two, while at the same
time maintaining a consistent increase in weight loss
throughout the program. On the other hand, the Anticipatory
Aversion (AA) group was superior to the Consummatory Aver
sion (CA) group on all six trials and, at follow-up, showed
the greatest weight loss of the three groups (Fig. 1,
Table 3). To the extent that these trends reflect an under
lying phenomenon, Aronfreed’s (1968) predictions are born
out. Why then were there no statistically significant
differences between these three aversive groups?
In the traditional study of the timing of punish
ment (Aronfreed and Reber, 1965), reinforcement occurs
60
either following preparatory behaviors or following the
consummatory response which culminates an entire behavioral
sequence of anticipatory behaviors. In Aronfreed's (1968)
conceptualization, behavior is mediated by symbolically
produced anticipations of reward or punishment. Where
punishment has been the prior consequence of a particular
response, anxiety, which was classically conditioned to
proprioceptive cues, is then aroused when anticipatory
responses are made. That is, punishment following antici
patory responses acts to inhibit behavior through the
elicitation of anxiety. Accordingly, it was hypothesized
that the Anticipatory Aversion group (AA), where covert
nausea followed anticipatory eating behavrios in imagina
tion, would show a significant weight loss relative to
the Consummatory Aversion group (CA). Indeed there did
appear to be a trend in that direction (Table 3, Fig. 1).
Consummatory punishment is assumed to be less effective
because the preceeding anticipatory behaviors occur in
the absence of inhibiting symbolic cues. That is,
anxiety will not be aroused until the behavior has already
occurred. Indeed, it might be assumed that, in the
absence of punishment, anticipatory behaviors are posi
tively reinforced through secondary reinforcement
thereby strengthening the response chain leading to the
consummatory act.
61
In the present study, subjects in the Consummatory
Aversion group were asked to imagine eating which was then
paired with nausea. All preparatory behaviors were excluded
from the imagery instructions. That is, any facilitating
effect of the prior occurrence of anticipatory behaviors was
absent. If indeed the relative ineffectiveness of consumma
tory punishment hinges upon the occurrence of these
responses and their accompanying self-reinforcement, then
this ineffectiveness should decrease as the anticipatory
behaviors decrease. This hypothesis might account for the
less than predicted difference between the Anticipatory
and Consummatory Aversion groups.
On the other hand, it would follow that if punish
ment were applied at both anticipatory and consummatory
points in the sequence, behavioral suppression should
increase all the more given the cue function of anxiety
arousal postulated by Aronfreed (1968), In fact the Anti
cipatory and Consummatory Aversion group (ACA) did perform
better than either of the other two at post-treatment
(Table 3, Fig. 1).
Given this assumption of second reinforcement, it
would be expected that the Anticipatory Imagery group (AI)
would lose the least amount of weight since each behavior
in the chain would be reinforced and lead to the next
ibehavior. Additionally, it would be expected that the
62
Consummatory Imagery group (Cl) would also do poorly in
terms of weight loss because they are receiving secondary
reinforcement for eating. In fact, at post-treatment, the
trend is in the opposite direction; the Anticipatory Imagery
(AI) group lost the most weight, followed by the Consumma
tory Imagery (Cl) group (Table 3). The only tenable hypo
thesis would seem to be that these groups did indeed
receive secondary reinforcement, reinforcement which served
to strengthen the learning of substitute responses for
eating. On the other hand, it was assumed that, for the
aversive groups, what was punished was in fact the learning
of those substitute responses. If this interpretation is
correct, there should be differential weight losses if we
compare aversive groups to non-aversive groups in combina
tion. When the two non-aversive groups were combined at
post-treatment and compared with the two aversive groups,
the non-aversive groups did show a significantly greater
weight loss than the aversive groups (t = 2.007, df = 74,
p. < .05).
Because we are dealing with cognitive phenomena,
and because we are far from able to evaluate the many
•mediational processes which exist, an analysis was done on
the frequency of home practice of the treatment images
;between groups, in an attempt to strengthen the discussion
iof treatment differences. In fact, it was found that there
63
were no significant differences between groups in terras of
the number of times subjects practiced the treatment imagery
(F = 1.63, df = 4,62) (Table 12). Therefore, the hypothesis
concerning the role of punishment in the suppression of
eating behaviors becomes somewhat more tenable.
Referring back to the post-treatment data, then, it
has been suggested that the non-aversive groups learned a
substitute response for eating which was self-reinforced via
the secondary reinforcement qualities of the imagery. On
the other hand, it was suggested that the aversive groups
were punished for the learning of these responses. Since
all groups lost weight, it might be assumed that the posi
tive reinforcement inherent in the imagery was more potent
than the negative reinforcement in affecting response
strength. If this is correct, it might account for the fact
that the Anticipatory and Consummatory Aversion (ACA) groups
performed better than either of the other two aversive
groups.
At follow-up, there was no significant difference
between aversive and non-aversive groups (t = .519-, df = 74)
nor any significant differences between the groups indivi
dually (Table 4). However, the groups generally continued
to lose weight as they had done during the program (Table 1 :
and 3, Fig. 1). Apparently, any differential effects due to
negative reinforcement were extinguished during the three
64
and one-half week interval from post to follow-up sessions.
It is interesting, however, that at follow-up, the groups
showing the greatest weight loss were the Anticipatory
Imagery (AI) and Anticipatory Aversion (AA) groups
(Table 3, Fig. 1).
The attempt to find indeces which might predict
weight loss or residivism was quite unsuccessful. In look
ing at Food Approach Scores from the beginning of treatment
to the end of treatment, there was a significant decrease in
Total Food Approach Scores across groups, although no sig
nificant difference between groups (Table 6, Fig. 2). In
addition, there were high correlations between scores
reported in Session 1 and scores in Session 6 (Table 19).
That is, although Total Food Approach Scores decreased
through the program, subjects who had relatively high
scores at the beginning had relatively high scores at the
end of treatment. Similar correlations were found for Habit
Scores (Table 19), although Habit scores neither differed
significantly between groups nor over trials (Table 7,
Fig. 3). However, Fig. 3 reveals that there was a trend in;
the direction of decreasing Habit Scores from pre to post- ;
treatment for all groups except the Consummatory Aversion
(CA) group, which increased its Habit Scores over trials.
This is not too surprising since this group lost the least
amount of weight during treatment (Table 3). ;
65
In spite of the fact that subjects having high
Total Food Approach and Habit scores prior to treatment had
similarly high scores at the end of treatment, these scores
were not related in any consistent fashion to either post
treatment or follow-up weight loss. Although there were
occasional significant correlations, no trends appeared to
exist (Tables 13 and 14). It was found that subjects who
weighed the most prior to treatment had the smallest
decrease in Total Food Approach Scores during the program.
This correlation was highly significant (p. < .005)
(Table 13). Again, however, neither the change in Food
Approach nor pre-treatment weights correlated with weight
loss (Table 19).
An examination of Body Image Scores revealed quite
similar results. There was a significant change in Real
Body Image from pre to post-treatment across groups,
although there were no significant differences between
groups (Table 8, Fig. 4). Again, Fig. 4 reveals that the
Consummatory Aversion (CA) group increased, rather than
decreased, its Real Body Image Scores through treatment.
Ideal Body Images neither changed significantly through
treatment nor differed between groups (Table 9, Fig. 5).
When Ideal and Real Body Images were correlated with pre
treatment weights, some interesting relationships were
revealed. Subjects who weighed the most prior to treatment !
66
had the largest Ideal Body Images both before and after
treatment, again, regardless of weight loss (Table 15). The
Body Image Discrepancy Score (the difference between the
Real and Ideal Body Images Scores) recorded at pre-treatment
did not correlate with pre-treatment weight, post-treatment
weight loss or follow-up weight loss. On the other hand,
the Discrepancy Scores recorded at post-treatment (D2 )
highly correlated with pre-treatment weights and post
treatment weight loss. That is, subjects who weighed the
most before treatment had the largest discrepancy between
Ideal and Real Body Images at the end of treatment. At the
same time, subjects with the largest Discrepancy scores
post-treatment lost the least amount of weight. This
relationship seemed fairly consistent through post-treatment
(Table 15). A possible explanation of these results would
hold that subjects who were most overweight, but who were
unsuccessful in losing weight, saw their Real Bodies as
very much larger than their Ideal Bodies. Body Image
Discrepancy Scores did however, significantly decrease over
trials, although there were no significant differences
between groups (Table 10, Fig. 6). Again, while all groups;
decreased the discrepancies between their Real and Ideal
Body Images during treatment, the Consummatory Aversion
!(CA) group increased these discrepancies). The amount of 1
67
change in Real or Ideal Body Image was not consistently
related to pre-treatment weights or weight loss (Table 15).
Several other variables were studied in terms of
their potential ability to predict weight loss: length of
time subjects had been overweight, number of times subjects
dieted in the past, pre-treatment weights, amount of
desired weight loss, motivation and I-E score. None of
these indeces correlated in any consistent fashion with
post-treatment or follow-up weight loss (Tables 17, 18, and
19). In addition, the amount of time subjects practiced
treatment images at home was unrelated to weight loss
(Table 18), a result that might certainly present some prob
lems for instigation therapies. Finally, subjects evalua
tion of the treatment program did not differ between groups
(Table 11) nor did these evaluations correlate with weight
loss (Table 18). The only variable found capable of pre
dicting follow-up weight loss was post-treatment loss
(Table 19). Subjects who lost the most at post-treatment
lost the most at follow-up.
In summary, then, no pre-treatment measure studied
was able to predict post-treatment weight loss, while the
only measure found to predict follow-up loss was, in fact,
post-treatment loss. These results are not inconsistent
with other reports in the literature. Stunkard and
McLaren-Hume (1959) studied weight loss as a correlate of
68
anxiety, prior weight history and the sex of the patient and
found only that the sex of the patient related to weight
loss, males having significantly greater losses than females.
On the other hand, Marston (personal communication) is in
the process of devising a questionnaire which appears to
offer much promise as a possible predictor of self
controlling behaviors. The Conflict Resolution Inventory
was designed to assess self-control via the discrepancies
between likelihood of a response, effortfulness of the
response and the responses a subject would ideally prefer
to make.
Perhaps the most surprising result of this study
was the fact that weight losses, in general, were so small
(Tables 1 and 3), less than that typically reported in the
literature (Cautela, 1966). However, several issues stand
out which might account for these results. In the first
place, most of the results reported are based on single case
studies, where it is difficult to evaluate the extent to
which non-experimental variables contributed to weight loss.
Secondly, of the group studies, group support and group
interaction have usually not been controlled. In the present
study, individual weights and weight losses were not dis
closed to the group, while group interaction was limited to
procedure and treatment-related discussions. Thirdly,
69
experimenter biases are often uncontrolled while in the
present study, this was not the case.
Finally, the procedure used was not identical to
the covert sensitization procedure employed by Cautela
(1966, 1967) and Anant (1967, 1968). Traditionally, the
pairing of nausea with food-related imagery is followed by
aversion relief, in which the subject imagines recovering
from nausea and feeling wonderful as he turns away from
food; that is, positive reinforcement for refraining from
eating. In the Manno and Marston (1970) study, a covert
sensitization group was compared with a group that received
only positive reinforcement for not eating, in imagination.
Both methods were equally effective in controlling eating
behavior. It is therefore postulated that the reported
success of covert sensitization procedures is due, in part,
to the combination of punishment for food approach and
reward for food denial, the total effect being perhaps some
combined function of the two. In the present study, aver
sion relief was eliminated from the program in terms of the
treatment instructions, and it is suggested that this might
explain the decreased effectiveness of the procedure.
On the other hand, the results of this study do
present some problems for the covert sensitization theorist,
problems which perhaps have implications for aversion
therapy in general. That is, in the present study, there
70
was a significant difference between aversive and non-
aversive groups at post-treatment; the non-aversive groups
showing a significantly greater weight loss. If the assump
tion is correct that, in part, the effectiveness of covert
sensitization is due to the elements of positive reinforce
ment, then it might be worthwhile to assess the advantages
of each in an empirical test. To this author's knowledge,
no such study has of yet been done. It might, in fact, be
found that the aversive components detract from treatment
effectiveness, a hypothesis offered to account for the
present results. Additionally, there is evidence from
Cautela (1970) himself, for the efficacy of covert positive
reinforcement (COR) which he has found to be effective in
controlling a variety of mal-adaptive responses.
In conclusion, this study lends some support to the
notion that when punishment is used, it is most effectively
applied early in the chain of events. On the other hand,
however, there is some indication that positive reinforce
ment is a more advantageous method of controlling eating
behaviors than aversive conditioning, although more defini
tive statements will have to await further studies in this
area.
REFERENCES
Anant, S. S. The use of verbal aversion technique with a
group of alcoholics. Saskatchewan Psychologist,
1966, 28-30.
Anant, S. S. A note on the treatment of alcoholics by a
verbal aversion technique. Canadian Psychologist,
1967, 8a, 1, 19-22
Anant, S. S. The use of verbal aversion (negative condi
tioning) with an alcoholic: A case report.
Behavior Research and Therapy, 1968, 6, 395-396.
Anant, S. S. Treatment of alcoholics and drug addicts by
verbal aversion techniques. The International
Journal of Addictions, 1968, 3, 2, 381-387. (a)
Aronfreed, J. Conduct and conscience. New York: Academic
Press, 1968.
Aronfreed, J., and Reber, A. Internalized behavioral
suppression and the timing of social punishment.
Journal of Personality and Social Psychology, 1965,
1, 3-16.
Ashem, B., and Donner, L. Covert sensitization with alco
holics: A controlled replication. Behavior
Research and Therapy, 1968, 6, 7-12.
Azrin, N. H. Punishment and recovery during fixed-ratio
performance. Journal of the Experimental Analysis
of Behavior, 1959, 2, 301-305.
71
72
Barlow, D. H., Leitenberg, H.r and Agras, W. S. Experimen
tal control of sexual deviation through manipula
tion of the noxious scenes in covert sensitization.
Journal of Abnormal Psychology, 1969, 74, 597-601.
Bixenstine, V. E. Secondary drive as a neutralizer of time
in integrative problem-solving. Journal of Compara
tive and Physiological Psychology, 1956, 49,
161-166.
Bruch, H, Transformation of oral impulses in eating
disorders: A conceptual approach. Psychiatric
Quarterly, 1961, 35, 3, 458-481.
Cannon, W. B. Bodily changes in pain, hunger, fear and rage.
New York: Appleton, 1915.
Cautela, J. R. Treatment of compulsive behavior by covert
sensitization. Psychological Record, 1966, 16,
33-41.
Cautela, J. R. Covert sensitization. Psychological Reports,
1967, 20, 459-468.
Cautela, J. R. Covert reinforcement. Behavior Therapy,
1970, 1, 33-50.
Cautela, J. R., and Kastenbaum, R. A. A reinforcement
survey schedule for use in therapy, training and
research. Psychological Reports, 1967, 20, 1115-
1130. (a)
73
Cautela, J. R., and Wisocki, P. A. The use of imagery in
the modification of attitudes towards the elderly:
A preliminary report. The Journal of Psychology,
1969, 73, 193-199.
Davison, G. C. Elimination of a sadistic fantasy by a
client-controlled counterconditioning technique:
A case study. Journal of Abnormal Psychology, 1968,
73, 84-90.
Eysenck, H. J. Handbook of abnormal psychology. New York:
Basic Books, 1960.
Eysenck, H. J., and Rachman, S. Causes and cures of
neurosis. London: Routledge and Kegan Paul, 1965.
Feinstein, R., Dole, V. P., and Schwartz, I. L. The use of
a formula diet for weight reduction of obese out
patients. Annals of Internal Medicine, 1958, 48,
330-343.
Ferster, C. B., Nurnberger, J. X., and Levitt, E. B. The
control of eating. Journal of Mathetics, 1962, 1,
87-109.
Garcia, J., and Koelling, R. A. Relation of cue to conse
quences in avoidance learning. Psychonomic Science,
1966, 4, 123-124.
Garcia, J., McGowan, B. K., Ervin, F. R., and Koelling, R. A.
Cues: Their relative effectiveness as a function of
the reinforcer. Science, 1968, 160, 794-795.
74
Gold, S., and Neufeld, I. L. A learning approach to the
treatment of homosexuality. Behavior Research and
Therapy, 1965, 2, 201-204.
Goldiamond, I. Self-control procedures in personal behavior
problems. Psychological Reports, 1965, 17, 851-868.
Harris, M. B. Self-directed program for weight control:
A pilot study. Journal of Abnormal Psychology,
1969, 74, 2, 263-270.
Hashim, S. A., and Van Itallie, T. B. Studies in normal
and obese subjects with a monitored food dispensing
device. Annals of the New York Academy of Sciences,
Homme, L. E. Perspectives in psychology: XXIV. Control of
coverants, the operants of the mind. Psychological
Record, 1965, 15, 501, 511.
Kamin, L. J. The delay-of-punishment gradient. Journal of
Comparative and Physiological Psychology, 1959, 52,
434-437.
Lazarus, A. A. Aversion therapy and sensory modalities:
Clinical impressions. Perceptual and Motor Skills,
1968, 27, 178.
Mahoney, M. J. Toward an experimental analysis of coverant
control. Behavior Therapy, 1970, 1, 510-521.
Manno, B. I., and Marston, A. R. Weight reduction as a
function of negative covert sensitization versus
positive covert reinforcement. Unpublished manu
script, University of Southern California, 1970.
75
Marston, A. R. Personality variables related to self
reinforcement. The Journal of Psychology, 1964,
58, 169-175.
Marston, A. R. Variables affecting incidence of self-
reinforcement. Psychological Reports, 1964, 14,
879-884. (a)
Marston, A. R. Imitation, self-reinforcement, and reinforce
ment of another person. Journal of Personality and
Social Psychology, 1965, 2, 2, 255-261.
Marston, A. R. Self-reinforcement: The relevance of a con
cept in analogue research to psychotherapy.
Psychotherapy: Theory, Research and Practice, 1965,
2, 1, 1-5. (a)
Marston, A. Self-reinforcement and external reinforcement
in visual-motor learning. Journal of Experimental
Psychology, 1967, 74, 1, 93-98.
Mayer, J. Exercise does keep the weight down. Atlantic
Monthly, 1955, 196, 63-66.
Mehrabian, A. Measures of achieving tendency. Educational
and Psychological Measurement, 1969, 29, 445-451.
Meyer, V., and Crisp, A. H. Aversion therapy in two cases
of obesity. Behavior Research and Therapy, 1964,
2, 143-147.
76
Morgenstern, F., Pearce, J., and Davies, B. The application
of aversion therapy to transvestism. Paper pre
sented at the meeting of the Reading Conference of
the British Psychological Society, 1963.
Mower, 0. H. Learning theory and behavior, New York: Wiley,
1960.
Mullen, F. G. The effect of covert sensitization on smoking
behavior. Unpublished study, Queens College,
Charlottesville, North Carolina, 196 8.
Nisbett, R. E. Taste, deprivation and weight determinants
of eating behavior. Unpublished doctoral disserta
tion, Columbia University, 1966.
Premack, D. Reinforcement theory. In D. Levine (Ed.),
Nebraska Symposium on Motivation: 1965. Lincoln:
University of Nebraska Press, 1965, pp. 123-180.
Rachman, S. Aversion therapy: Chemical or electrical?
Behavior Research and Therapy, 1965, 2, 289-299.
Rehm, L. P., and Marston, A. R. Reduction of social
anxiety through modification of self-reinforcement:
An instigation therapy technique. Journal of
Consulting and Clinical Psychology, 1968, 32, 5,
565-574.
; Rotter, J. B. Generalized expectancies for internal versus
external control of reinforcement. Psychological
Monograph, 1966, 80, 1, 1-28.
77
Schachter, S. Cognitive effects on bodily functioning:
Studies of obesity and eating. In D. C. Glass (Ed.),
Neurophysiology and Emotion. New York: Rockefeller
University Press and Russell Sage Foundation, 1967,
pp. 117-144.
Schachter, S., and Gross, L. Manipulated time and eating
behavior. Unpublished manuscript, 1967.
Schachter, S., Goldman, R., and Gordon, A. The effects of
fear, food deprivation, and obesity on eating.
Unpublished manuscript, 1967.
Skinner, B. F. Science and human behavior. New York:
Macmillan, 1953.
Solomon, R. L. Punishment. American Psychologist, 1964,
19, 239-253.
Solomon, R. L., Turner, L. H., and Lessac, M. S. Some
effects of delay of punishment on resistance to
temptation in dogs. Journal of Personality and
Socia1 Psycho1ogy, 1968, 8, 233-238.
Stollak, G. E. Weight loss obtained under various experi
mental procedures. Paper presented at the meeting
of the Midwestern Psychological Association,
Chicago, May, 1966.
Stunkard, A. Obesity and the denial of hunger. Psychoso
matic Medicine, 1959, 21, 4, 281-289.
78
Stunkard, A., and Koch, C. The interpretation of gastric
motility: I. Apparent bias in the reports of hunger
by obese persons. Archives of General Psychiatry,
1964, 11, 74-82.
Stunkard, A., and McLaren-Hume, M. The results of treat
ment for obesity. Archives of Internal Medicine,
1959, 103, 79-85. (a)
Suczek, R. F. The personality of obese women. American
Journal of Clinical Nutrition, 1957, 5, 2, 197-202.
Thorpe, J. G., Schmidt, E., Brown, P. T., and Castell, D.
Aversion-relief therapy: A new method for general
application. Behavior Research and Therapy, 1964,
2, 71-82.
Walters, R. H., and Demkow, L. Timing of punishment as a
determinant of resistance to temptation. Child
Development, 1963, 34, 207-214.
Winer, H. Real and imagined cost effects upon human fixed-
interval responding. Psychological Reports, 196 5,
17, 659-662.
Wilson, G. T., and Davison, G. C. Aversion techniques in
behavior therapy: Some theoretical and metatheoreti-
cal considerations. Journal of Consulting and
Clinical Psychology, 1969, 33, 3, 327-329.
Wisocki, P. A. The use of covert sensitization and covert
reinforcement in the treatment of obsessive-compul-
sion behavior: A new approach. Unpublished study,
Boston College, Chestnut Hill, Massachusetts, 1969.
79
Wollersheim, J. P. Effectiveness of group therapy based
upon learning principles in the treatment of over
weight women. Journal of Abnormal Psychology, 1970,
76, 3, 462-474.
Wolpe, J. Psychotherapy by reciprocal inhibition.
Stanford: Stanford University Press, 1958.
80
TABLE 1
Mean Weights (lbs.) Across Groups At
Pre-treatment, Post-treatment, and Follow-up
Mean Mean Mean
Pre-treatment Post-treatment Follow-up
CONDITION Weight (lbs.) Weight (lbs.) Weight (lbs.)
Group I
Consummately Aversion 154.25 154.25 151.83
(CA) n = 12 n = 12 n = 12
Group II
Anticipatory Aversion 147.25 145.13 142.69
(AA) n = 16 n = 16 n = 16
Group III
Anticipatory and
Consummately Aversion 145.14 142.57 141.64
(ACA) n = 14 n = 14 n = 14
Group IV
Consummately Imagery 151.06 148.39 148.39
(Cl) n = 18 n = 18 n = 18
Group V
Anticipatory Imagery 150.31 147.31 145.69
(AI) n = 16 n = 16 n = 16
81
SOURCE
TABLE 2
Analysis of Variance
Pre-treatment Weights (lbs.)
df MS
Between Groups
S's Within Groups
4
62
197
881
.224
82
TABLE 3
Mean Weight Loss Across Groups At
Post-treatment and Follow-up
MEAN MEAN
POST-TREATMENT FOLLOW-UP
WEIGHT LOSS WEIGHT LOSS
CONDITION lbs. Arc Sines Lbs. Arc Sines
Group I
Consumnatory Aversion 0 1.004 2.417 1.153
(CA)
Group II
Anticipatory Aversion 2.12 1.129 4.560 1.234
(AA)
Group III
Anticipatory and
Ccnsummatory Aversion 2.57 1.140 3.500 1.126
(AGA)
Group IV
Cansunmatory Imagery 2.67 1.182 2.670 1.146
(Cl)
Group V
Anticipatory Imagery 3.00 1.234 4.620 1.270
(AI)
83
TABLE 4
Weight Loss*
Treatment X Trials Analysis of Variance
SOURCE df MS F
B (Groups) 4 .302 1.285
S's (B) 71 .235
T (Trials) 5 .178 6.593**
B x T 20 .025 .926
S’s (B) x T 355 .027
** p. < .01
* Arc Sine transformations of percentage weight loss.
84
TABLE 5
Newman-Kuels Analysis
Differences in Weight Loss Over Trials
TRIAL
2
TRIAL
4
TRIAL 6
Post-Tr.
TRIAL
3
TRIAL 7
Follow-up
TRIAL
5
Value Needed
For Signifi
cance at
.01
80.512 86.488 87.183 87.205 90.312 98.780
5.976* 6.618* 6.693* 9.800* 10.268*
6.816
.642 .717 3.824 4.292 6.59
.075 3.182 3.650 6.30
3.107 3.575 5.90
.468 5.21
* p. < .01
85
TABLE 6
Analysis of Variance
Total Food Approach
SOURCE df MS F
B (Groups) 4 256.000 1.592
S's (B) 61 160.721
T (Trials) 1 452.000 10.633*
B x T 4 29.750 .699
S's (B) x T 61 42.508
* p. < .01
86
TABLE 7
Habit Scores
Analysis of Variance
SOURCE df MS
B (Groups)
S's (B)
T (Trials)
B x T
S's (B) x T
61
1
4
61
5.38
10.81
8.9
3.38
2.24
.498
3.970
1.51
87
TABLE 8
Real Body Image Scores
Analysis of Variance
SOURCE df MS F
B (Groups) 4 2.200 1.047
S's (B) 62 2.100
T (Trials) 1 4.300 12.400*
B x T 4 .075 .217
S's (B) x T 62 .345
* p. < .01
88
TABLE 9
Ideal Body Image Scares
Analysis of Variance
SOURCE d£ MS
B (Groups)
S's (B)
T (Trials)
62
.395
.737
.270
.536
2.389
B x T 4 ' .185 1.637
S’s (B) x T 62 .113
89
TABLE 10
Body Image Discrepancy Scores
Analysis of Variance
SOURCE df MS F
B (Groups) 4 .900 .750
S's (B) 62 1.165
T (Trials) 1 6.700 20.060*
B x T 4 .650 1.946
S's (B) x T 62 .334
* p. < .01
90
TABLE 11
Program Evaluation
Analysis of Variance
SOURCE df MS F
Between Groups 4 5.306 2.026
S's Within Groups 62 2.619
91
TABLE 12
Home Practice of Treatment Imagery Scores
Analysis of Variance
SOURCE df MS
Between Groups
S's Within Groups
4
62
2.50
1.53
1.630
92
Table 13
Pearson Correlations
Food Approach Scores
Pre-treatment
Weight (lbs.)
Post-treatment Loss a
(arc sines of percentage
weight loss)
AI
CI
ACA AA CA
T
1
Total Food Approach
Session 1
-.1432
df=74
.2398
df=15
.0327
df=17
***
-.7283
df=13
-.1662
df=15
-.1378
df=ll
T
2
Total Food Approach
Session 6
.2131
df=74
-.1464
df=15
.3001
df=17
***
-.8654
df=13
-.2176
df=15
-.1783
df=ll
H1
Habit Score
Session 1
.1860
df=74
.0412
df=15
.3206
df=17
-.3746
df=13
-.2160
df=15
-.3350
de=ll
H2
Habit Score
Session 6
.2023
df=65
-.3777
df=14
.3062
df=15
**
-.5955
df=10
.0011
df=13
-.4978
df=10
Tl - T2
***
-.3913
df=65
.4226
df=14
-.2594
df=15
.1987
df=10
,0954
df=13
.0670
df=10
hi - h2 -.2136
df=65
.3701
df=l4
.1751
df=15
.4182
df=10
-.0811
df=13
.2135
df=10
aGroup Abbreviations:
AI = Anticipatory Imagery
Cl = Consunrnatory Imagery
ACA = Anticipatory and Consummatory Aversion
AA = Anticipatory Aversion
CA = Consunroatory Aversion
* p <.05
** p <.025
*** p <.005
93
Table 14
Pearson Correlations
Food Approach Scores
Follcw-up
a
(arc sines of percentage weight loss)
AI Cl ACA AA CA
T1
Total Food Approach
Session 1
.3194
df=15
-.0478
df=16
-.2960
df=13
-.0690
d£=15
.0190
df=ll
T
2
Total Food Approach
Session 6
.2389
df=l4
.2176
df=13
-.4961
df=10
-.3277
df=13
.1552
df=10
Hi
l
Habit Score
Session 1
.2974
df=15
.0671
df=16
-.0547
df=13
-.2141
df=15
-.1117
df=ll
H„
z
Habit Score
Session 6
.1508
df=14
-.2437
df=13
-.1737
df=10
-.4477
d£=13
-.2338
df=10
T1 " T2
.1210
df=14
-.2318
df=13
.1353
df=10
.3536
df=13
-.1548
df=10
H1 " H2
.1322
df=14
.3907
df=13
.1168
df=10
.2612
df=13
.1618
df=10
: * p < .05 Grorp Abbreviations:
;** p < .025
AI = Anticipatory Imagery
*** p < .005 Cl = Consurtmatory Imagery
ACA = Anticipatory and Consunrnatory Aversion
AA = Anticipatory Aversion
CA = Consunrnatory Aversion
94
Table 15
Pearson Correlations
Body Image Scores
Pre- treatment
Weight (lbs.)
Post-treatment Loss
(arc sines of percentage
weight loss)
a
AI Cl ACA AA CA
Real Body Image
Session 1
**
.4854
df=74
.1074
df=15
.1020
df=17
.0787
df=13
-.1958
df=15
-.2739
df=ll
I1
Ideal Body Image
Session 1
**
.7367
df=74
.0493
df=15
.2554
df=17
*
.4542
df=13
.1271
df=15
-.3563
df=ll
D1
(R1 " Il)
.0942
df=74
.0928
df=15
-.0694
df=17
-.0633
df=13
-.3133
df=15
-.0282
df=ll
*2
Real Body Image
Session 6
**
.5217
df=66
-.0208
df=15
.0111
df=15
-.2870
df=10
-.2291
df=13
-.4595
df=10
I2
Ideal Body Image
Session 6
.5636
df=66
.0529
df=15
.2935
df=15
.3750
df=10
.0944
df=13
-.2613
df=10
°2
(R2 " I2)
*
.2547
df^66
-.0552
df=15
-.3463
df=15
**
-.7601
df^=10
**
-.4599
df=13
-.4141
df=10
*1 ~ R2
-.1159
df=66
.1338
df=15
.1268
df=15
.2741
df=10
.3899
df=13
.5569
df=10
.0145
df=66
-.0085
df=15
-.2888
df=15
-.0513
df=10
-.1067
df=13
-.1100
df=10
D1 - D -.0691 .3160 .2347 .3053 -0307 .2159
df=66 df=15 df=15 df=10 df=13 df=10
aGroup Abbreviations:
AI - Anticipatory Imagery
Cl = Cons ummatory Imagery
ACA = Anticipatory and Cons ummatory Aversion
AA = Anticipatory Aversion
CA = Consunrnatory Aversion
* p < .05
** p < .005
95
Table 16
Pearson Correlations
Bod/ Image Scores
Pollow-up Loss
(arc sines of percentage weight loss)
a
AI Cl ACA AA CA
®1
Real Body Image:
Session 1
.0663
df=15
-.1964
df=16
.3900
df=L3
-.0895
df=15
,4133
df=ll
I1
Ideal Body Image:
Session 1
-.2086
d£=15
-.0724
df=16
.5327
df=13
.0916
df=15
.3174
df=ll
D1
(Rx - Ix)
.2018
df=L5
-.1631
df=16
.2627
df=13
-.1671
df=15
.3036
df=ll
Real Body Image:
Session 6
.1815
df=15
-.0611
df=13
-.1880
df=10
-.1409
df^l3
.4184
df=10
I2
Ideal Body Image:
Session 6
-.1559
df=15
.2991
df=13
.2551
df=10
.2926
df=13
.3057
d&=10
D2
(R2 " I2)
.3145
df==15
-.4328
df=13
-.3056
df=10
-.5270
df=13
.3131
df=10
*1 ~ ®2
-.2238
df=15
-.1541
df=13
.6302
df=10
-.1810
df=13
.0724
df=10
Il “ *2
-.1353
df=15
-.5723
df=13
.1988
df=10
-.3002
d:&=13
.0731
df=10
D1 ~ D2
-.2463
df=15
-.1056
df=13
.3992
df=10
.0740
df=13
.0289
df=10
a
Group Abbreviations:
AI = Anticipatory Imagery
Cl = Cons ummatory Imagery
ACA = Anticipatory and Consunrnatory Aversion
AA = Anticipatory Aversion
CA = Consunrnatory Aversion
Table 17
Pearson Correlations
Pre-treatment
Weight (lbs.)
AI
Post-treatment Loss
(arc sines of percentage
weight loss)
CC ACA AA
a
CA AI
Follcw-up
(arc sines of percentage
weight loss)
Cl ACA AA CA
Motivation .0355 .3510 -.0676 -.0994 -.0966 .1240 .0049 .0358 -.3076 -.0181 ,0853
Score df=75 df=15 df=17 df=13 df=15 df=ll df=15 df=16 df=13 df=15 df=ll
I - E .0654 -.2108 -.0387 .0692 -.3876 .1084 -.2160 -.3404 .3386 -.3067 ■ -.2051
df=75 df=15 df=17 df=13 df=15 df=ll df=15 d&=16 df=13 df=15 df=ll
aGroup Abbreviations:
AI = Anticipatory Imagery
CL = Consunrnatory Imagery
ACA = Anticipatory and Consummately Aversion
AA = Anticipatory Aversion
CA = Cons ummatory Aversion
97
Table 18
Pearson Correlations
Weight History and Program Evaluation
Groups
a
AI Cl ACA AA CA
Length of time subject was
overweight x post-treatment
weight loss
.1523
df=15
-.2780
df=17
.0393
df=13
-.2841
df=15
-.1016
df=ll
Number of times subject
dieted in the past x post
treatment weight loss
.1349
df=15
.0644
df=17
-.3173
df=13
**
-.6341
df=15
-.1305
df=ll
Number of times subject
practiced procedure at home
x post-treatment weight loss
-.1223
df-15
-.1386
dfr=14
-.3681
df=10
.0109
df=13
.0033
df=10
How valuable subject felt
the program had been x
post-treatment weight loss
.1256
df=15
.3314
df=14
.1372
df=10
.1213
df=13
-.2180
df=10
Desired weight Ices:
Session 1 x Weight loss:
Session 6
-.0479
df=15
-.2159
df=17
.4131
df=13
.0122
df=15
-.6233
df=ll
* p. < .05 aGroup Abbreviations:
** p. < .005 AI = Anticipatory Imagery
d = Consunrnatory Imagery
ACA = Anticipatory and Consunrnatory Aversion
AA = Anticipatory Aversion
CA = Consunrnatory Aversion
98
Table 19
Pearson Correlations
Groups
a
AI Cl ACA AA CA
Pre-treatment weight x
Post-treatment weight loss
.1783
df=15
.2649
df=17
.3995
df=13
.3045
df=15
-.2817
df=ll
Habit Score: Session 1 x
Habit Score: Session 6
.4606
df=15
***
.7060
df=15
***
.8766
df=10
.5681
df=13
***
.7580
df=10
Total Food Approach :
Session 1 x
Total Food Approach :
Session 6
**
.5982
df=15
.5541
df=15
***
.9344
d^=10
.3370
df=ll
.4681
df=10
Weight loss: Post-treatment
x weight loss: Follow-up
**
.5799
df=15
.4942
df=17
***
.7260
db=13
.3959
df=15
.5263
d^ll
c L
* p. < .05 Group Abbreviations:
** p. < .01 AI = Anticipatory Imagery
*** p. < .005 Cl = Consummatory Imagery
ACA = Anticipatory and Gonsummatory Aversion
AA = Anticipatory Aversion
CA = Cons ummatory Aversion
MEAN WEIGHT LOSS
(arc sine percentage of weight loss)
1.30
1.25
1.20
Q-
1.15
1.10
o —
1.05
1.00
Figure 1
Maan weight loss (arc sines of percentage weight loss) across trials for Consunrnatory Imagery (Cl;,
Anticipatory Imagery (AI), Anticipatory and Cons ummatory Aversion (ACA) . Consunrnatory Aversion (CA),
ana Anticipatory Aversion (Aa) Groups.
MEAN TOTAL FOOD APPROACH SCORES
57
56
55
54
53
52
51
50
49
48
47
46
o— o
O D
O O
A A
ci
A±
ACA
CA
AA
100
X
\
X
X
X
C P
X
_L
Pre- tre atment Post-treatment:
TRIALS
Figure 2
Changes in Mean Total Food Approach Scores frcm pre to post-treatment :
For Consuxnnatory Imagery (CI) , Anticipatory Imagery (AI),
Anticipatory and Cons ummatory Aversion (ACA), Consummatory Aversion (CA)
and Anticipatory Aversion (AA) groups.
MEAN HABIT SCORES
O--- -O CI
A - - - A AI
D ----□ ACA
O--O CA
A — A AA
101
8.0
7.5
7.0
6.5
6.0
5.5
Pre-treatment Post-treatment
TRIALS
Figure 3
Changes in Mean Habit Scores frcm pre to post-treatment for
Consurtmatory Imagery (CI) , Anticipatory Imagery (AI) , Anticipatory
and Consumnatory Aversion (ACA) , Consunrnatory Aversion (CA) and
Anticipatory Aversion (AA) groups.
MEAN REAL BODY IMAGE SCORES
0- -O
CI
A- -A
AI
Q-— Q
ACA
0-— 0
CA
A-— A
AA
3.9
3.6
3.5
s
3.3
3.2
3.0
2.9
Pos t-treatment Pre-treatment
TRIAIB
Figure 4
Changes in Mean Real Body Image Scores from pre to post-treatment for ;
Consunrnatory Imagery (CI) , Anticipatory Imagery (AI) , Anticipatory !
and Cons ummatory Aversion (ACA), Consunrnatory Aversion (CA) and j
Anticipatory Aversion (AA) groups,
.... ......... . . . . . ..........._ J
MEM IDEAL BODY IMAGE SCORES
O----O CI 103
A---A ai
□ --- Q ACA
O--O CA
A--A AA
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
Post-treatment Pre-treatment
TRIALS
Figure 5
Changes in Mean Ideal Body Image Scores from pre to post-treatment for :
Consunrnatory Imagery (CI), Anticipatory Imagery (AI), Anticipatory:
and Consunrnatory Aversion (ACA) , Consunrnatory Aversion (CA) and
Anticipatory Aversion (AA) groups.
MEM BODY IM2GE DISCREPANCY SCORE
O O CI 104
A A ai
D ----- Q ACA
O ------O CA
A-----A AA
2.5
2.4
2.3
2.2
2.0
1.9
1.8
1.7
1.6
1.5
Pre- treatment Post- treatment
TRIALS
Figure 6
Changes in Mean Body Image Discrepancy Scores from pre to post-treatrrent
For Consuirmatory Imagery (CI) , Anticipatory Imagery (AI),
Anticipatory and Consurrmatory Aversion (ACA) , Consunrnatory Aversion
(CA) and Anticipatory Aversion (AA) groups.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
An Examination Of Positive And Negative Reinforcement In Classical And Operant Conditioning Paradigms In The Primary Psychopath
PDF
Attention, retention, and incentive processes in observational learning
PDF
The Development And Evaluation Of Three Therapeutic Group Interventions For Widows
PDF
Death Anxiety In Leukemic Children
PDF
Modification Of Low Self-Confidence In Elementary-School Children By Reinforcement And Modeling
PDF
Experimenter Expectancy Effect Examined As A Function Of Task Ambiguity And Internal Versus External Control Of Reinforcement
PDF
The Effect Of Conditions Of Risk, Internal Versus External Control Of Reinforcement, And Sex On Binary Choice Probability Learning
PDF
The Effect Of Discriminability On The Partial Reinforcement Effect In Human Gsr Conditioning
PDF
Mental Imagery As A Function Of Muscular Tension And Suggestion
PDF
Stimulus and response generalization of classes of imitative and nonimitative behavior as a function of reinforcement, task, cues, and number of therapists
PDF
Looking Back After Coming Down: Conformity And Commitment In Campus Protest
PDF
The Effects Of Diphenylhydantoin On The Galvanic Skin Responses Of Psychopathic And Normal Prisoners
PDF
The Effects Of A Self Shock Procedure On Hallucinatory Activity In Hospitalized Schizophrenics
PDF
The Effect Of Personalized Emotional Stimuli On Asthmatic Reactions
PDF
The Effects Of Anxiety And Threat On Self-Disclosure
PDF
The Differential Effectiveness Of External Versus Self-Reinforcement On The Acquisition And Performance Of Assertive Responses
PDF
A Temporal Approach-Avoidance Conflict In An Academic Test Situation
PDF
Self-Sacrifice, Cooperation, And Aggression In Women Of Varying Sex-Role Orientations
PDF
A Cognitive Dissonance Analysis Of Conformity Behavior, As Applied To The full Denture Patient
PDF
The Relationship Of Teacher Empathy And Student Personality To Academic Achievement And Course Evaluation
Asset Metadata
Creator
Manno, Beatrice Ila Scheinbaum
(author)
Core Title
Weight Reduction As A Function Of The Timing Of Reinforcement In A Covertaversive Conditioning Paradigm
Degree
Doctor of Philosophy
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,Psychology, clinical
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Marston, Albert R. (
committee chair
), Frankel, Andrew Steven (
committee member
), Tiber, Norman (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c18-549290
Unique identifier
UC11362478
Identifier
7203789.pdf (filename),usctheses-c18-549290 (legacy record id)
Legacy Identifier
7203789
Dmrecord
549290
Document Type
Dissertation
Rights
Manno, Beatrice Ila Scheinbaum
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA