Restraint, disinhibition, hunger and negative affect eating

Restraint, disinhibition, hunger and negative affect eating

Addictive Behaviors, Vol. 13, pp. 369-377, 1988 Printed in the USA. All rights reserved. RESTRAINT, DISINHIBITION, MICHAEL 0306-4603/88 $3.00 + .O...

754KB Sizes 28 Downloads 79 Views

Addictive Behaviors, Vol. 13, pp. 369-377, 1988 Printed in the USA. All rights reserved.

RESTRAINT,

DISINHIBITION,

MICHAEL

0306-4603/88 $3.00 + .OO Copyright 0 1988 Pergamon Press plc

HUNGER EATING

AND NEGATIVE

AFFECT

R. LOWE and BARBARA MAYCOCK

Department

of Psychology,

Rutgers University

Abstract - The Three-Factor Eating Questionnaire (Stunkard & Messick, 1985) contains factors measuring dietary restraint, disinhibition, and hunger. In this study, the ability of these factors to predict eating responses to negative affect was tested. The Velten mood induction procedure was used to produce neutral or depressed moods in normal weight college students. Subjects were encouraged to sample candy which was made available during the mood induction procedure. The results indicated that the Hunger factor was the only significant predictor of negative affect eating. Depressed, high-hunger subjects were more likely to eat than subjects in other conditions and, in one of two analyses of amount of candy consumed, were found to eat the most candy as well. The possibility that these results could be explained by an overlap between the Hunger factor and the construct of external responsiveness was considered. Implications of the overall findings for theories of emotional eating were briefly discussed.

Individuals differ markedly in their tendency to eat when emotionally distressed. Obese individuals have traditionally been viewed as eating more when distressed (Bruch, 1973; Lowe & Fisher, 1983) and normals as eating less (Schachter, Goldman, & Gordon, 1968). However, restraint theorists have argued that obese persons may be more prone to emotional eating than normals only because they are more likely to be actively restraining their food intake (Herman & Polivy, 1975). Restraint could predispose an individual to negative affect eating because stress may overwhelm the inhibition of eating that restrained eaters normally maintain. With their customary dietary controls temporarily undermined by stress, the cognitive and/or physiological characteristics associated with restraint may cause restrained eaters to eat with abandon (Herman & Polivy, 1980; Ruderman, 1985). Research has supported the theory that restraint is a powerful predictor of eating responses to distress. Herman and Polivy (1975) found that normal weight restrained eaters ate somewhat more when fearful than when calm, whereas normal weight unrestrained eaters ate significantly less when fearful. These results parallel those found by Schachter et al. (1968) for obese and normal weight persons, respectively. Results similar to Herman and Polivy’s have been found by Frost, Goolkasian, Ely, and Blanchard (1982) and by Ruderman (1985) for restraint and by Baucom and Aiken (1981) for dieting. In addition, two studies have reported that restraint level predicts the direction of weight change in clinically depressed individuals (Polivy & Herman, 1976; Zielinski, 1978). It is unclear, however, exactly why restrained eaters are prone to negative affect eating. Restrained eaters may differ from unrestrained eaters in terms of cognitive style, current dieting status and/or history of dieting and weight loss (Lowe, 1984, 1986, This research was supported by funds from the Rutgers Research Council and the Charles and Johanna Busch Memorial Fund. Portions of this article were presented at the annual convention of the Eastern Psychological Association, Boston, Mass., in March, 1985. We wish to thank Susan Sweet for her help in conducting this study and Bill Whitlow for his comments on an earlier version of this paper. Requests for reprints should be sent to Michael R. Lowe, who is now at General Internal Medicine, Broad and Tioga Sts., 2nd Floor, Temple University Medical School, Philadelphia, PA 19140. 369

370

MICHAEL

R. LOWE and BARBARA

MAYCOCK

1987). Frost et al. (1982) reported that the restraint x mood interaction they found was due to the weight fluctuation factor of the Restraint Scale, whereas Ruderman (1985), while finding the same restraint x mood interaction, determined that it was due to the dietary concern factor of the Restraint Scale. A recently developed measure of eating behavior, derived in part from Herman and Polivy’s (1980) Restraint Scale, may be helpful in identifying factors associated with negative affect eating. This measure, called the Three-Factor Eating Questionnaire (TEQ), contains factors measuring dietary restraint, disinhibition and hunger. The nature of these factors, and their relevance for understanding negative affect eating, is described below. The Dietary Restraint (DR) factor contains several items from the dietary concern factor of the Restraint Scale and a number of additional items describing concern with controlling weight and strategies for doing so. The DR factor appears to describe many characteristics of restrained eaters (Herman & Polivy, 1980); furthermore we (Lowe & Kleifield, 1988) found that the DR factor, with redundant items removed, was moderately correlated with the Restraint Scale (r = .54, p < .OOl). Since DR contains no items about weight or weight fluctuation, it represents one way of testing the relationship between restraint and negative affect eating while avoiding the potentially confounding influence of weight history (Drewnowski, Riskey, & Desor, 1982). The Disinhibition (DI) factor describes the tendency to eat because palatable foods are available, because others are eating or because of emotional distress. DI is correlated with the severity of binge eating in overweight women (Marcus, Wing, & Lamparski, 1985), with the amount of weight gained by individuals who became clinically depressed (Weissenberg, Rush, Giles, & Stunkard, 1986), and with the Restraint Scale (r = .62, p < .OOl; Lowe & Kleifield, 1988). Thus DI obviously has potential as a predictor of negative affect eating. The Hunger factor describes the intensity with which hunger sensations are perceived and the extent to which such sensations evoke eating. (To capture this meaning the Hunger factor will be referred to as “hunger sensitivity” (HS) here.) Thus there should be a main effect of HS on eating, with high-HS individuals eating more than low-HS individuals. Although the relevance of the HS factor for understanding negative affect eating is unclear, HS predicted binge eating severity in obese women (Marcus et al., 1985) and weight gain in depressed patients (Weissenberger et al., 1986). The purpose of the study described here is to experimentally evaluate the joint effects of negative affect and all three TEQ factors on eating. For each factor, it was predicted that the combination of high scores and negative affect should be associated with more eating than combinations involving either low scores or neutral affect. METHOD

Subjects

Seventy-one normal weight, female undergraduate students were solicited for this study. Eleven subjects had to be dropped from consideration for various reasons (seven guessed the study’s true purpose, three reported they could not eat chocolate, and one was overweight). The remaining 60 subjects were all < 15% overweight according to the Metropoli-

Negative

affect

eating

371

tan Life Insurance Tables (1959). The subjects’ mean weight was close to ideal (mean = - .98%). Subjects received either $5 or course credit for their participation. Materials

Undergraduate subjects enrolled in psychology courses completed an “eating questionnaire” which contained the TEQ and several other questions (height, weight and phone number). At the same time they were given the opportunity to sign up for a purportedly unrelated study on “personality and mood,” which in fact was the study described here. Respondents’ DR, DI and HS scores were calculated using the method described by Stunkard and Messick (1985). T-tests run on these scores for students who did and did not volunteer to participate in the second study revealed no differences between these groups on any factor (allps > .15). Subjects who participated in the current study were assigned to low and high groups on DR, DI and HS based on a median split of their raw scores on these factors.’ Mood manipulation

and measurement

Depressed and nondepressed moods were induced using Velten’s (1968) mood induction method. In the depressed condition, this procedure involved having subjects read a series of 45 increasingly depressing self-statements (e.g., “I’m not very alert; I feel listless and vaguely sad”), with instructions to try to induce in themselves the mood suggested to them by the statements. The neutral condition involved reading an equal number of bland descriptive statements (e.g., “Many states supply milk for grammar school children.“). The Velten procedure has been shown to be an effective means of inducing depressive reactions in the laboratory (e.g., Coleman, 1975; Hale & Strickland, 1976; Velten, 1968). Prior to and immediately following the mood induction procedure, subjects completed the Personal Feeling Scale (PFS) and the Depression subscale of the Multiple Affect Adjective Check List (MAACL-D; Zuckerman & Lubin, 1965). The PFS contains eight bipolar ratings relevant to depression (e.g., optimistic-pessimistic; hopefulness-hopelessness), each of which is rated on a lo-point scale. Both the PFS and the MAACL-D were also used by Frost et al. (1982) in a study similar to the present one. Procedure

Subjects were run in the early afternoon (at either 12:30, 1:00 or 1:30) in a comfortably furnished office. They were seated behind a desk which contained various accoutrements (in-out trays, tissues, books, etc.) as well as an opaque bowl containing 300 grams of M & M candies. The M & MS were initially shielded from the subjects’ view so that the presence of the candy would not influence subjects’ responses to the pretest questionnaires. Subjects were given the PFS and MAACL-D and were left alone for three minutes to complete them. Just prior to returning to the subject, the experimenter determined the subject’s assignment to either the neutral or depressed mood condition via a table of random numbers and brought the appropriate set of Velten cards to the subject. The subject was told that we were interested in “people’s moods and their ability to ‘Because of tied scores, the median splits based on raw scores resulted in fewer subjects in the low than in the high groups for DI and HS. Therefore, factor scores were calculated for subjects falling at the median for these two factors (using factor loadings originally reported by Stunkard and Messick). Of these subjects, those with the lowest factor scores were assigned to the low (DI or HS) group until there were 30 subjects in both low and high (DI and HS) groups.

372

MICHAEL

R. LOWE and BARBARA

MAYCOCK

change their moods.” The experimenter explained that the subject should read the cards one at a time and try to bring on in themselves the mood suggested by each card. The subject was also told that the task could get somewhat monotonous, and, in an apparent gesture of friendliness, said that the subject should “help yourself to the candy while completing the task.” While saying this, the experimenter moved the bowl to a position approximately 18” in front of the subject’s right hand while casually eating two M & MS at the same time. The experimenter then explained that instructions for moving from card to card would be given from a tape recorder that was on the desk. This recording instructed subjects to move to a new card every 15 seconds. Subjects were told that the tape would also instruct them to complete two measures (the second PFS and MAACL-D scales) in a manilla folder in front of them after they had finished with all of the cards. The tape recorder was then turned on and the subject was left alone for 18 minutes to complete the Velten cards and the two posttest measures. At the end of this procedure, the subject was brought to a second room, asked what she thought the purpose of the study was, measured for height and weight, and debriefed. Subjects who came close to naming the true focus of the study (i.e., on amount eaten) were eliminated from consideration and replaced by another subject. The overall procedure described above represents a close replication of the procedure used by Frost et al. (1982) in a study described earlier. The experimenters were the two authors (a Ph.D. clinical psychologist and an undergraduate psychology major) and a third undergraduate psychology major. All experimenters were blind as to subjects’ TEQ scores and were unaware of subjects’ mood assignments until just prior to the administration of the Velten cards. Each experimenter ran approximately one third of the subjects. RESULTS

Check on mood manipulation Two x two analyses of variance were conducted on the pre-post difference scores on the PFS and the MAACL-D to see if the mood manipulation had its intended effect. Separate analyses were conducted for each TEQ factor. For each factor and for both the PFS and the MAACL-D, there were significant main effects for mood (allps < .OOl), indicating that subjects in the depressed mood condition experienced a greater negative mood shift than subjects in the neutral mood condition for each comparison. The only other significant effect was a HS level x mood interaction on the MAACL-D (F(1, 56) = 4.11, p < .05). There was a weak trend toward the same interaction on the PFS (p < .15). The interaction on the MAACL-D stemmed from high-HS subjects becoming more dysphoric than low-HS subjects in the depressed condition (means of 11.35 and 8.18, respectively), and less dysphoric than low-HS subjects in the neutral condition (means of - 1.37 and 2.90, respectively). Consumption data An examination of the within-cell distribution of eating scores for all three factors indicated that most distributions were positively skewed. Furthermore, within-cell variances were heterogeneous for the comparisons involving HS (using Levene’s (1960) test, F(1, 56) = 14.77, p < .OOl). To help normalize the skew and reduce the heterogeneity of variance in these distributions, the eating data were log transformed.

Negative

Table

1. Consumption Cognitive

affect

data as a function Restraint

eating

of mood

313

and TEQ factor

assignment Hunger

Disinhibition

Sensitivity

TEQ Mood Neutral

Depressed

Low

High

Low

High

12.38

10.57

10.69

SD N

8.64’ [.6112 13.99 14

B

19.06

SD N

L.911 23.03 16

Factor x

i.721

[.621

1.721

Low 7.43

L.621

High 13.44

t.711

13.96 16

15.34 14

12.95 16

10.32 14

16.17 16

8.07

13.81

I.511 15.78 14

l.811 15.48 16

14.07 r.631 25.57 14

4.31 ].361 8.03 16

24.93 [1.14] 24.73 14

‘Unbracketed figure is amount consumed in grams. ‘Bracketed figure is log transformation of grams consumed.

The mean amount consumed by subjects as a function of mood and each of the three TEQ factors is shown in Table 1. When the effects of DR and mood on consumption were analysed in a two x two ANOVA, no significant effects emerged. When subjects were assigned to groups according to their DI scores, there was again no significant effects of DI, mood, or their interaction. However, the analysis based on subjects’ HS scores produced a significant main effect due to HS, F(1,56) = 7.73, p < .Ol, and a significant HS x mood interaction, F(l,56) = 4.85,~ < .05 (see Fig. 1). Simple effects tests of this interaction indicated that the depressed, high-HS subjects ate significantly more candy than depressed, low-HS subjects, t(28)= 3.18, p < .005, and neutral, low-HS subjects, t(28) = 2.37, p =C.05, and marginally more than neutral, high-HS subjects, t(28) = 1.75, p < .lO. Since the results for the HS factor could be due to the proportion of subjects who ate at least something, as well as to the amount those subjects ate, the percentage of subjects who ate at least some candy was also analyzed. These percentages were as follows: neutral, low-HS, 64%; neutral, high-HS, 56%; depressed, low-HS, 38%; and depressed, high-HS, 86%; x2(1) = 4.92,~ < .05. When subjects who ate nothing were dropped from consideration and the two (HS) x two (mood) ANOVA was repeated, only a main effect for the HS factor emerged (F(1, 32) = 6.52, p < .02). One additional comparison of interest was that between mood change and consumption for the analysis involving HS. Since depressed, high-HS subjects tended to both get more upset than depressed, low-HS subjects, and ate more candy as well, it is possible that this occurred not because high-HS subjects were more hunger sensitive than low-HS subjects, but because they were more depressed.2 This possibility was evaluated by covarying out extent of mood change in the analysis of the joint effects of mood and HS on eating. Two 2 (HS level) x 2 (mood) ANOVAs were therefore conducted, using pre-post changes in the PFS and the MAACL-D as covariates. In both ANOVAs the original mood x HS interaction was retained in slightly weakened form (for the PFS, F(1, 55) = 3.74, p < .06; for the MAACL-D, F(1, 55) = 4.44,p < .05. ?t might be argued that the HS x mood interaction was due to high-HS, depressed subjects becoming more upset because they ate more food. However, if true, then high-HS, neutral subjects, who ate more than low-HS, neutral subjects, should have become more depressed than this latter group. Instead, they became less depressed.

374

MICHAEL

R. LOWE

and BARBARA

MAYCOCK

Low

High Hunger

Fig. 1. Interaction

between

Sensitivity

mood and hunger

sensitivity

for amount

of candy

eaten.

DISCUSSION

While there is empirical evidence to suggest that ah three factors of the TEQ might predict the tendency to eat when distressed, only the Hunger Sensitivity factor actually predicted this eating pattern. A chi-square analysis indicated that depressed, high-HS subjects were more likely to eat at least some candy than were subjects in the other three groups. If subjects who ate nothing are included in the analysis of amount eaten, then the same pattern held - that is, depressed, high-HS subjects ate more candy than subjects in the other groups. However, since noneating subjects could be viewed as unwilling to eat, rather than as eating zero grams of candy, a second analysis was done eliminating noneating subjects from consideration. This analysis again produced a main effect of hunger sensitivity on eating (with high-HS subjects eating more candy than low-HS subjects), but no HS x mood interaction. This analysis introduces a potential bias, however, since unequal proportions of subjects were dropped from different conditions. Perhaps the most appropriate conclusion is that emotional distress increases the probability that hunger sensitive individuals will eat when food is available, and may increase the amount eaten as well. Theoretically, it is unclear why hugger sensitivity was related to negative affect eating in this study and to binge eating (Marcus et al., 1985) and depression-induced weight change (Weissenberger et al., 1986) in others. Since HS and the Restraint

Negative affect eating

375

Scale are uncorrelated (r = .22, ns; Lowe & Kleifield, 1988), this relationship does not appear to stem from any overlap in the constructs assessed by these measures. The results obtained for the HS factor might be explained by the revised version of externality theory described by Rodin (1981). Rodin notes that distinguishing normal and obese individuals on external responsiveness is not always possible and suggests that this occurs in part because there are many normal weight people who are externally responsive. Nonetheless, while externality may not be highly predictive of body weight, it predicts emotionality (Rodin & Schachter, 1974) and sensitivity to food cues (Rodin 8z Slochowei, i976). Furthermore, as Rodin (1978) has shown, individuals who are externally responsive show greater insulin release in response to the presentation of palatable foods. This latter observation points to a possible convergence between externality and hunger-related responses to palatable foods. Thus it is possible that the Hunger Sensitivity factor examined in the current study may reflect enhanced internal responses to palatable foods (i.&, “hunger”) also shown by externally responsive individuals (cf. Van Strien, Frijters, Roosen, Knuiman-Hill, & Defares, 1985). If the HS factor taps a dimension similar to that tapped by measures of externality, then both the emotionality and eating results obtained for this factor may be more readily understood. That is, just as externally sensitive people are more emotional (Rodin & Schachter, 1974), high;HS subjects In the present study reacted more dramatically to the depressed mood induction, The tendency for high-HS subjects to eat more when depressed may reflect both high-HS subjects’ sensitivity to food cues and an enhancement in the saliency of these cues produced by their aroused emotional state (cf. Spitzer & Rodin, 1983). The absence of a relationship between the DR factor and negative affect eating suggests that, despite their similarities, this factor and the Restraint Scale are measuring two somewhat different constructs. Other findings also support this conclusion. For instance, while the Restraint Scale predicts salivary responses (Klajner Herman, Polivy, & Chhabra, 1981) and binge eating (Wardle, 1980), the DR factor is not related to salivation (Sahakian, Lean, Robbins, & James, 1981) or to bingeing (Marcus et al., 1985). There are at least two differences between the Restraint and DR measures which may account for these contrasting findings. The Restraint Scale appears to emphasize an emotional involvement with eating more than the DR factor, and the Restraint Scale, unlike DR, contains reference to weight fluctuations. The lack of relationship between DI and negative affect eating was even more surprising. The DI factor actually contains two items which describe emotional eating, but most of the items describe overeating which occurs because of an inability to stop eating once it has commenced. Thus the fact that many subjects in the present study chose not to eat any candy at all may have worked against the emergence of DI as a predictor of negative affect eating. The relevance of the current study for naturally-occurring emotional eating and for eating disorders remains to be determined. Since the validity of the Velten mood induction procedure has been questioned in several studies (e.g., Buchwald, Strack, & Coyne, 1981; Polivy & Doyle, 1981), it would be desirable to reexamine the relationship between the TEQ and negative affect eating using other mood induction techniques. Also, since the subjects in the current study were normal weight and (presumably) not eating disordered, the relevance of the current findings for obese and bulimic individuals also requires further study. As suggested previously, it would also be desirable in future studies on the TEQ

376

MICHAEL R. LOWE and BARBARA MAYCOCK

and negative affect eating to use eating tests which require at least some food consumption. Most previous studies of restraint-related determinants of emotional eating have done so (Baucom & Aiken, 1981; Ruderman, 1985). In addition, the nature of the DI factor suggests that it may successfully predict eating behavior only when subjects are required to taste some food. Finally, requiring subjects to eat some food in a taste test avoids the problems created here and elsewhere (Frost et al., 1982) of how to treat subjects who do not eat any food. REFERENCES Baucom, D.H., & Aiken, P.A. (1981). Effect of depressed mood on eating among obese and nonobese dieting and nondieting persons. Journal of Personality and Social Psychology, 41, 3 17-32 1. Bruch, H.-(1973). Eating disorders. New York: Basic Books. Buchwald, A.M., Strack, S., & Coyne, J.C. (1981). Demand characteristics and the Velten mood induction procedure. Journal of Consulting and Clinical Psychology, 49, 478-479. Coleman, R.E. (1975). Manipulation of self-esteem as a determinant of mood of elated and depressed women. Journal of Abnormal Psychology, 84, 693-700. Drewnowski, A., Riskey, D., & Desor, J.A. (1982). Feeling fat yet unconcerned: Self-reported overweight and the Restraint Scale. Appetite 3, 273-279. Frost, R.O., Goolkasian, G.A., Ely, R.J., & Blanchard, F.A. (1982). Depression, restraint, and eating behavior. Behaviour Research and Therapy, 20, 113-121. Hale, W.D., & Strickland, B.R. (1976). The induction of mood states and their effect on cognitive and social behavior. Journal of Consulting and Clinical Psychology, 44, 155. Herman, C.P., & Polivy, J. (1975). Anxiety, restraint and eating behavior. Journal ofAbnormal Psychology, 84, 666-672.

Herman, C.P., & Polivy, J. (1980). Restrained eating. In A.J. Stunkard (Ed.), Obesity. Philadelphia, PA: Saunders. Isen, A.M., & Gorgoglione, J.M. (1983). Some specific effects of four mood induction procedures. Personality

and Social Psychology

Bulletin,

9, 136-143.

Klajner, F., Herman, C.P., Polivy, J., & Chhabra, R. (1981). Human obesity, dieting and the anticipatory salivation to food. Physiology and Behavior, 27, 195-198. Levene, H. (1960). Robust tests for equality of variances. In I. Olkins (Ed.), Contributions to Probability and Statistics. Stanford, CA: Stanford University Press. Lowe, M.R. (1984). Dietary concern, weight fluctuation and weight status: Further explorations of the Restraint Scale. Behaviour Research and Therapy, 22, 243-248. Lowe, M.R. (1986). Dieting and Binging: Some unanswered questions. American Psychologist, 41, 326 327. Lowe, M.R. (1987). Set point, restraint, and the limits of weight loss: A critical analysis. In W. Johnson (Ed.), Advances in Eating Disorders, Vol. I: Treating and preventing obesity. Greenwich, CT: JAI

Press. Lowe, M.R., & Fisher, E.B., Jr. (1983). Emotional reactivity, emotional eating, and obesity: A naturahstic study. Journal of Behavioral Medicine, 6, 135-149. Lowe, M.R., & Kleifield, E. (1988). Cognitive restraint, weight suppression, and the regulation of eating. Appetite, in press. Marcus, M., Wing, R.R., & Lamparski, D.M. (1985). Binge eating and dietary restraint in obese patients. Addictive Behaviors, 10, 163-168. Metropolitan Life Insurance Co. (1959). New weight standards for men and women. Statistical Bulletin, 40, l-4.

Polivy, J., & Doyle, C. (1980). Laboratory inductions of mood states through the reading of self-referent mood statements: Affective changes or demand characteristics? Journal of Abnormal Psychology, 89, 286290.

Polivy, J., & Herman, C.P. (1976). Clinical depression and weight change: a complex relation. Journal of Abnormal

Psychology,

85, 331340.

Rodin, J. (1978). Has the distinction between internal vs. external control of feeding outlived its usefulness? In G.A. Bray (Ed.), Recent Advances In Obesity Research (Vol. 2). London: Newman. Rodin, J. (1981). The current status of the internal-external obesity hypothesis: What went wrong. American Psychologist,

36, 361-372.

Rodin, J., & Slochower, J. (1976). Externality in the non-obese: The effects of environmental responsiven&s on weight. Journal of Personality and Social Psychology, 29, 557-565. Ruderman, A.J. (1985). Dysphoric mood and overeating: a test of restraint theories disinhibition hypothesis. Journal of Abnormal Psychology, 94, 78-85.

Negative affect eating

377

Sahakian, B.J., Lean, M.E., Robbins, T.W., &James, W.P.T. (1981). Salivation and insulin secretion in response to food in non-obese men and women. Appetite. 2, 209-216. Schachter. S., Goldman, R., &Gordon, A. (1968). Effects of fear, food deprivation, and obesity on eating. Journal of Personality and Social Psychology, 10, 91-97. Schachter. S.. & Rodin. J. (1974). Obese humans and rafs. Washington, DC: Erlbaum/Halsted Spitzer, L:, a Rodin, J.‘(1983). Arousal induced eating: Convention> wisdom or empirical findings. In J. Cappioco and R. Petty (Eds.), Social Psychophysiology. New York: Guilford. Stunkard, A.J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition, and hunger. Journal of Psychosomatic Research, 29, 71-83. Van Strien, T., Frijters, J.E.R., Roosen, R.G.F.M., Knuiman-HijI, W.J.H., & Defares, P.B. (1985). Eating behavior, personality traits, and body mass in women. Addictive Behaviors, 10, 333-343. Velten, E. (1%8). A laboratory task for induction of mood states. Behaviour Research and Therapy, 6, 473-482. Weissenberger, J., Rush, A.J., Giles, D.E., Kunetz, N., & Stunkard, A.J. (1986). Determinants ofweight change in depressed patients. Psychiatry Research, 17, 275-283. Zielinski, J. (1978). Depressive symptomatology: Deviation from a personal norm. Journal of Community Psychology, 6, 163-167. Zuckerman, M., & Lubin, B. (1965). Manual for the Multiple Affect Adjective Checklist. San Diego: Education and Industrial Testing Service.