The specific content of core beliefs and schema in adolescent girls high and low in eating disorder symptoms

Eating Behaviors 7 (2006) 27 – 35

The specific content of core beliefs and schema in adolescent girls high and low in eating disorder symptoms Myra J. Cooper a,T, Kathryn S. Rose a, Hannah Turner b a

Isis Education Centre, University of Oxford, Warneford Hospital, Oxford OX3 7JX, UK b University of Southampton, UK Received 9 February 2005; accepted 26 May 2005

Abstract The specific content of core beliefs and schema in adolescent girls high and low in eating disorder symptoms was investigated using the Eating Disorder Belief Questionnaire (EDBQ)-negative self-beliefs sub-scale and Young Schema Questionnaire (YSQ). Girls with high Eating Attitude Test (EAT) scores had higher scores on both measures (using total and sub-scale scores) than girls with low EAT scores. However, all scores were related to eating disorder as well as depressive symptoms. Analysis of individual items on the EDBQ and YSQ found that scores on all EDBQ, but not all YSQ items, differed between the two groups, with higher scores in the high EAT group. However, a number of individual items (on both measures) in the high EAT group, as well as in the low EAT group, were related specifically to eating disorder symptoms and not also to depressive symptoms. D 2005 Elsevier Ltd. All rights reserved. Keywords: Cognition; Core beliefs; Schema; Eating disorders

A number of empirical studies indicate that core beliefs and schema are important cognitions to be considered if a comprehensive understanding of eating disorders is to be developed (for reviews see: Cooper, 1997; Cooper, 2005; Waller, Kennerley, & Ohanian, 2005). Cognitive theories of eating disorders have begun to be developed which have a role for core beliefs and schema, in both the development and maintenance of these disorders (Cooper, Wells, & Todd, 2004; Waller et al., 2005). However, it has been noted that it is unclear how far these core beliefs and schema are specific to eating disorder symptoms rather than, for example, depressive symptoms (Waller, Shah, Ohanian, & T Corresponding author. Tel.: 44 1865 226431; fax: 44 1865 226364. E-mail address: [email protected] (M.J. Cooper). 1471-0153/$ - see front matter D 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.eatbeh.2005.05.007

28

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

Elliott, 2001). Eating disorder patients often have high levels of depression and, moreover, where this has been investigated, studies find that depressed patients also have high levels of the core beliefs and schema identified in those with eating disorders (Cooper & Hunt, 1998; Waller et al., 2001). Clinically, it is unclear whether or not there are differences between the content of core beliefs and schema associated with eating disorder symptoms and that associated with depressive symptoms. Bruch (1973), for example, noted a number of features typical of the disturbance in the way eating disorder patients see themselves. The list is lengthy, and includes feelings of self dislike and self hatred, failure, unworthiness, a high need for achievement and perfectionism, compliance, excessive concern for others, reluctance to express emotion, a need to please others, lack of self assertion, a need to take care of others, and a feeling that they are, or wish to be, special and different. Some of these features clearly overlap with features observed in depressed patients. Beck (1967), for example, generated a lengthy list of the typical ways depressed patients see themselves. This also included self dislike and self hatred, as well as unworthiness, feelings of disappointment in the self, failure, self disgust, self criticism, self blame, inferiority, and inadequacy. More specifically, Guidano (1987); Guidano and Liotti (1983) have suggested that while the eating disorder (ED) self is organised around an absolute need for approval by significant others, and a fear of being intruded upon or disconfirmed by them, the depressive self is organised around a sense of loss. Two self report measures of core beliefs and schema have been used in studies investigating differences and similarities between eating disorders and depression. One measure (the Eating Disorder Belief Questionnaire: EDBQ) was developed specifically for eating disorder patients and includes an assessment of negative self-beliefs (Cooper, Cohen-Tovee, Todd, Wells, & Tovee, 1997). It has good psychometric properties (Cooper et al., 1997). The other measure (the Young Schema Questionnaire: YSQ) was developed as a generic measure of early maladaptive schemas. Studies with the 75 item version (Young, 1998) indicate that it has acceptable psychometric properties in bulimic women (Waller, Meyer, & Ohanian, 2001). However, given that that the YSQ was not developed for eating disorder patients it is possible that sub-scale totals may obscure patterns in individual items that do distinguish eating disorder from depressed patients in meaningful ways. The EDBQ negative self-belief sub-scale (EDBQ-nsb), although developed specifically for ED patients, was not developed from a detailed analysis of the specific differences that might characterise ED patients and not depressed patients, but rather to capture the broad range of core beliefs and schema relevant to EDs. However, it is possible that individual items may be differentially related to eating disorder and depressive symptoms. Thus a more fine-grained analysis of the individual items on both measures might be useful in identifying particular and specific differences between eating disorder and depressive symptoms in core beliefs and schema. To date, such an analysis has not been reported in the literature. A clinical sample is not necessarily ideal for the current purpose, as it is often difficult to match ED and depressed patients on potential confounding variables. Non-clinical matching also presents problems (see Cooper, Rose, &, Turner, in press). We therefore selected two groups of women to study for the purposes of the current paper. One group had high EAT scores (above a clinical cut off, Garner & Garfinkel, 1979) and the other group had low EAT scores. The two groups were matched on relevant demographic variables. While we were particularly interested in the relationships between beliefs, schema and symptoms in a group that might be considered an analogue for a clinical group, we included the second group to determine if any relationships uncovered might also generalise to those who had few eating disorder related symptoms. After analysis of total scores we investigated whether individual items

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

29

on the EDBQ-nsb sub-scale and YSQ were related to eating disorder symptoms in the two groups separately. We then used partial correlations to control for depressive symptoms, in order to examine whether any links between individual core beliefs and schema, eating disorder symptoms and depressive symptoms were retained or lost when the effects of either eating disorder or depressive symptoms were controlled for. This analysis permitted us to draw some preliminary conclusions about which individual core beliefs and schema might be specifically related to eating disorder symptoms, and not also to depression. We were also able to determine whether any relationships were specific to high EAT scorers or not. The aim of the study, therefore, was to identify the specific individual core beliefs and schema associated with eating disorder symptoms and that were not also associated with depressive symptoms in high and low EAT scorers.

1. Method 1.1. Participants 1.1.1. High scorers These were 52 female adolescents who scored 30 or more on the Eating Attitudes Test (EAT: Garner & Garfinkel, 1979). A score of 30 is the recommended clinical cut off for an ED diagnosis in the original validation study. The group comprised the top 14% of a larger sample of adolescents recruited through their schools. 1.1.2. Low scorers These were 52 female adolescents who comprised the bottom 14% of the larger sample. They all scored 3 or less on the EAT. 1.2. Measures Participants were recruited through schools. The study was presented to potential participants during class time, and those who volunteered subsequently completed the measures individually and anonymously, and returned them in an appointed collection box at their school. Demographic information was collected on age, weight and height. Each participant completed four self report questionnaires. These were the EAT, Beck Depression Inventory (BDI: Beck & Steer, 1993), EDBQ-nsb sub-scale and YSQ. Both the EAT and BDI are well validated measures that have been widely used in the study of eating disorders in clinical and non-clinical populations. The psychometric properties of the EDBQ-nsb sub-scale and YSQ have been less extensively investigated, but are both promising (EDBQnsb, Cooper et al., 1997; YSQ, Waller et al., 2001).

2. Results The two groups were similar in mean age (low EAT: mean = 17.8 years, SD = .42; high EAT: mean = 17.6 years, SD = .51), and in mean Body Mass Index (low EAT; mean = 20.2, SD = 2.2: high EAT:

30

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

mean = 21.9, SD = 2.1). They scored significantly differently on the EAT (low EAT: mean = 2.1, SD = 0.8; high EAT: mean = 41.0, SD = 8.9, t = 30.9, p b .0001), and also on the BDI (low EAT: mean = 4.56, SD = 3.6; high EAT: mean = 20.2, SD = 10.2). As expected the high EAT group had higher scores than the low EAT group on the EAT and also on the BDI, indicating higher levels of eating disorder and depressive symptoms in this group. 2.1. Analysis of total scores Mean scores for the two groups on the EDBQ negative self-beliefs sub-scale and YSQ total and individual sub-scales can be seen in Table 1. The high EAT group scored significantly higher than the low EAT group on the negative self-beliefs sub-scale of the EDBQ, on total YSQ score and on all but one of the individual YSQ sub-scales (selfsacrifice). The number of clinically significant schema endorsed on the YSQ by the high EAT group (mean 13.8, SD = 11.8) was also significantly higher than that of the low EAT group (mean 3.0, SD = 3.0). The relationship of the EDBQ negative self-beliefs sub-scale (total score) and YSQ total and individual sub-scales (total scores) to levels of eating disorder and depressive symptoms was then investigated. Six correlations obtained in the low EAT group with the BDI were significant ( p b .05 or Table 1 Scores for the two groups on the eating disorder belief questionnaire, negative self-beliefs subscale, and Young Schema Questionnaire Variable

Low EAT group

High EAT group

T value

Significance level

Mean

SD

Mean

SD

EDBQ Negative self-beliefs

8.9

9.1

36.3

21.9

8.2

.0001

YSQ Emotional deprivation Abandonment Mistrust/abuse Social isolation Defectiveness/shame Failure to achieve Dependence/incompetence Vulnerability to harm Enmeshment Subjugation Self-sacrifice Emotional inhibition Unrelenting standards Entitlement Insufficient self control Mean total score Clinically significant score

1.2 1.5 1.4 1.6 1.2 1.7 1.5 1.4 1.4 1.5 2.7 1.6 3.1 1.9 2.1 1.7 3.0

0.5 0.4 0.5 0.7 0.3 0.9 0.5 0.4 0.5 0.5 0.9 0.8 1.1 0.8 0.9 0.3 3.0

2.2 2.8 2.8 2.6 2.8 3.6 2.5 2.3 1.8 2.5 3.0 2.4 3.7 2.3 3.3 2.7 13.9

1.1 1.6 1.2 1.4 1.4 1.7 1.1 1.1 1.0 1.2 1.0 1.1 1.4 0.8 1.3 0.8 11.9

6.0 6.0 7.0 4.8 7.9 6.5 6.2 5.4 3.0 5.9 1.7 4.5 2.5 2.4 5.5 8.4 6.2

.0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .004 .0001 .09 .0001 .02 .02 .0001 .0001 .0001

EDBQ=Eating Disorder Belief Questionnaire. YSQ=Young Schema Questionnaire. EAT=Eating Attitudes Test.

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

31

p b .01). All correlations with the BDI in the high EAT group were significant ( p b .05 or p b .01), except for one. Two correlations with the EAT were also significant in this group ( p b .05). When BDI score was partialled out of the equation, however, all the previously significant correlations became non-significant (all p N .05). When EAT score was partialled out, neither significant correlation became non-significant.

Table 2 Results of partial correlations between EDBQ negative self-beliefs subscale and YSQ and EAT and BDI scores for low EAT scorers Negative self-belief items related to EAT when BDI score is partialled out

Negative self-belief items no longer related to BDI when EAT score is partialled out

I’m stupid ( p = .059)

I’m a failure I’m all alone I don’t like myself very much I’m dull

YSQ items related to EAT when BDI score is partialled out

YSQ items no longer related to BDI when EAT score is partialled out

Mistrust and abuse I feel that people will take advantage of me ( p = .054) Emotional inhibition I find it embarrassing to express my feelings to others (negative correlation) Insufficent self control I can’t force myself to do things I don’t enjoy, even when I know it’s for my own good (negative correlation)

Emotional deprivation Most of the time, I haven’t had someone to nurture me, share him/herself with me, or care deeply about everything that happens to me In general, people have not been there to give me warmth, holding, and affection For much of my life I haven’t felt that I was special to someone For the most part, I have not had someone who really listens to me, understands me, or is tuned into my true needs and feelings Abandonment I worry that people I feel close to will leave me or abandon me Sometimes I am so worried about people leaving me that I drive them away Mistrust and abuse I feel that people will take advantage of me I am quite suspicious of other people’s motives Defectiveness I am too unacceptable in very basic ways to reveal myself to other people Dependence/incompetence My judgement cannot be relied upon in everyday situations Enmeshment My parent(s) and I tend to be overinvolved in each otherTs lives and problems Insufficient self control I can’t seem to discipline myself to complete routine or boring tasks

EDBQ=Eating Disorder Belief Questionnaire. YSQ=Young Schema Questionnaire. EAT=Eating Attitudes Test.

32

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

Table 3 Results of partial correlations between EDBQ negative self-beliefs subscale and YSQ and EAT and BDI scores for high EAT scorers Negative self-belief items related to EAT when BDI score is partialled out

Negative self-belief items no longer related to BDI when EAT score is partialled out

I’m not a likeable person I don’t like myself very much I’m dull

I’m a failure I’m all alone I don’t like myself very much I’m dull

YSQ items related to EAT when BDI score is partialled out

YSQ items no longer related to BDI when EAT score is partialled out

Abandonment When I feel someone I care for pulling away from me, I get desperate Dependence/independence I do not feel capable of getting by on my own in everyday life ( p = .053) Vulnerability to harm I worry I’ll lose all my money and become destitute Emotional inhibition I find it hard to be warm and spontaneous ( p = .052) I control myself so much that people think I am unemotional Unrelenting standards I must be the best at most of what I do, I can’t accept second best I have a lot of trouble accepting bnoQ for an answer when I want something from other people I’m special and shouldn’t have to accept many of the restrictions placed on other people ( p = .057)

Self sacrifice I’m so busy doing for the people I care about, that I have little time for myself Insufficient self control If I can’t reach a goal, I become easily frustrated and give up I have a very difficult time sacrificing immediate gratification to achieve a long-range goal I have rarely been able to stick to my resolutions

EDBQ=Eating Disorder Belief Questionnaire. YSQ=Young Schema Questionnaire. EAT=Eating Attitudes Test.

2.2. Analysis of individual items1 To investigate which individual negative self-belief and schema items differentiated the two groups best, the scores of the two groups were compared for each individual EDBQ-nsb and YSQ item. All EDBQ negative self-belief item scores were significantly greater in the high EAT compared to the low EAT group (all p b .05). Sixty out of 75 YSQ item scores were also significantly greater in this group (all p b .05). The relationship of the individual negative self-belief and schema items to levels of eating disorder and depressive symptoms was then investigated in each of the two groups.

1

Given that the study was exploratory, corrections were not made for the number of statistical tests conducted. It has been very persuasively argued that this reduces the meaning of exploration (Cohen, 1992).

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

33

2.2.1. Low EAT group When BDI score was partialled out of the equations containing the EDBQ negative self-belief items four out of 10 items remained significantly associated with EAT score (all p b .05, except one p = .059, one p = .054). These items are listed in Table 2. When EAT was partialled out of these same equations 4 previously significant correlations became non significant (all p N .05). These items are listed in Table 2. When BDI score was partialled out of the equations containing the YSQ items only 3 correlations remained significant (all p b .05, with one p = .034). These items are shown in Table 2. When EAT was partialled out, however, 12 previously significant correlations became non-significant (all p N .05). These items are also shown in Table 2. 2.2.2. High EAT group When BDI score was partialled out of the equations containing the EDBQ negative self-belief items three out of 10 items remained significantly associated with EAT score (all p b .05). These items are listed in Table 3. When EAT was partialled out of these same equations no previously significant correlations became non-significant. When BDI score was partialled out of the equations containing the YSQ items only 8 correlations remained significant (all p b .05, except one p = .054, one p = .052, and one p = .057). These items are shown in Table 3. When EAT was partialled out, however, 4 previously significant correlations became non-significant (all p N .05). These items are also shown in Table 3.

3. Discussion Total and sub-scale scores on core beliefs and schema, as assessed by the EDBQ negative selfbeliefs sub-scale and YSQ total and sub-scale scores, discriminated a community sample of female adolescents high and low on ED symptoms. Those with high EAT scores had significantly higher scores than those with low EAT scores. This replicates previous findings in patient samples (Cooper & Hunt, 1998; Waller et al., 2001). It provides further validity for the use of the EDBQ-nsb and YSQ in those with eating disorders and related symptoms. All those in the high EAT group scored about the bcut offQ for caseness on the EAT (Garner & Garfinkel, 1979). Thus a non-patient group at high risk of an eating disorder, compared to a low risk control group seem, like patients, to have elevated scores on these measures (e.g. Cooper & Turner, 2000; Waller, Ohanian, Meyer, & Osman, 2000). This represents a new finding, as no previous studies with these measures have used such an bat riskQ group. Analysis of the relationship between core beliefs, schema and eating disorder and depressive symptoms, using both EDBQ-nsb and YSQ total and sub-scale scores, suggested that core beliefs and schema, as assessed by these measures, were strongly related to depressive symptoms, and not independently related to eating disorder symptoms (except via their relationship to symptoms of depression). This finding is also consistent with those of previous studies, which have used the EDBQnsb and YSQ total and sub-scale scores to assess core beliefs and schema in eating disorders (e.g. Cooper & Hunt, 1998; Waller et al., 2001). However, a more detailed, fine-grained analysis investigating the individual items on both the EDBQnsb and YSQ found that while all the individual EDQB-nsb items were scored more highly by the high EAT group, compared to the low EAT group, some of the individual items on the YSQ were not rated

34

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

differently by the two groups. Moreover, when partial correlations were conducted, some specific links between core beliefs and schema and symptoms of eating disorders were uncovered, that did not appear to be explained by any link to depressive symptoms. The exact relationships were not the same in the low and high EAT groups but there was, nevertheless, considerable overlap. All but two of the EDBQ negative self-beliefs related to EAT score were found to be related in both the high and low EAT groups. Similarly, only three YSQ sub-scales were not related to EAT in both groups. These were vulnerability to harm, self sacrifice and unrelenting standards, which were related to EAT score in the high but not low EAT group. Eating disorder symptoms thus appeared to be associated with a rather different profile of core beliefs and schema than depressive symptoms. The items that were most closely related to EAT score, rather than to BDI, when the EDBQ-nsb was investigated were as follows: not likeable, failure, all alone, don’t like myself, dull, and stupid. The items and themes that appeared to be most closely related to EAT score, rather than to BDI, in analyses involving the YSQ sub-scales were: mistrust (2 items), dependency/incompetence (2 items), enmeshment (one item), self sacrifice (one item), emotional inhibition (3 items), insufficient self control (5 items), emotional deprivation (3), defectiveness (one item), abandonment (one item), vulnerability to harm (one item) and self sacrifice (one item). Inspection of these items suggests a picture in which a person with high levels of eating disorder symptoms expect to be betrayed, are suspicious of people, feel not intelligent, dependent, not separate from parents (including in maintaining emotional boundaries), feel that no one has ever cared deeply for them, think of others more than themselves, listen to others problems, do too much for others and not enough for themselves, are emotionally self controlled, striving to reach high standards and to meet their responsibilities, unable to accept being bgood enoughQ, and have a need to feel special. This is very consistent with the clinical picture painted, for example, by Bruch and also Guidano and Liotti. Interestingly, the theme of blossQ identified as particularly characteristic of depression by Guidano and Liotti, and also apparent as a general theme in much cognitive theory writing on depression, does not appear to overlap significantly with the current themes. Interestingly too, high EAT score is associated with perceived lack of self control — a feature of bulimia nervosa that is also noted to increase vulnerability to anorexia nervosa (Bruch, 1973). This study provides a first step towards outlining a specific core belief and schema profile for those with high levels of ED related symptomatology and, significantly, one that is independent of any associated depressive symptoms. Limitations of the study include use of a non-clinical group (although see note in the introduction about the difficulties inherent in achieving this), and failure to separate AN symptoms from BN symptoms. Use of a cross sectional design also makes it impossible to attribute a causal role to the beliefs and schema identified. Interestingly, we found that many of the relationships appear to hold in a low EAT as well as a high EAT group, thus they may not only be typical of those with clinically significant symptoms. Future research might usefully investigate AN and BN separately to see if any beliefs and schema are differentially characteristic of different EDs or of key behaviours such as bnot eatingQ and bingeing, including in both clinical and non-clinical samples. The current study highlights the limitations of pre-formulated self report measures, particularly in generating new knowledge, and also the danger of premature conclusions based on global analysis of the data. Future research in this area might also usefully benefit from more fine-grained, exploratory analyses such as these, and should be wary of relying only on global scores. Detailed analysis of the kind conducted here might be of particular value in an area such as eating disorders where differences between normal and sub-clinical, and between the different eating disorders, particularly in cognition, are currently poorly understood.

M.J. Cooper et al. / Eating Behaviors 7 (2006) 27–35

35

References Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. New York7 Hoeber Medical Division, Harper and Row. Beck, A. T., & Steer, R. A. (1993). Beck depression inventory II: Manual. San Antonio7 Psychological Corporation. Bruch, H. (1973). Eating disorders. New York7 Basic Books. Cohen, J. A. (1992). A power primer. Psychological Bulletin, 112, 155 – 159. Cooper, M. J. (2005). Cognitive theory of anorexia nervosa and bulimia nervosa: Progress, development and future directions. Clinical Psychology Review, 25, 511 – 531. Cooper, M. J. (1997). Cognitive theory of anorexia nervosa and bulimia nervosa: A review. Behavioural and Cognitive Psychotherapy, 25, 113 – 145. Cooper, M. J., Cohen-Tovee, E., Todd, G., Wells, A., & Tovee, M. (1997). A questionnaire to assess assumptions and beliefs in eating disorders: Preliminary findings. Behaviour Research and Therapy, 35, 381 – 388. Cooper, M. J., & Hunt, J. (1998). Core beliefs and underlying assumptions in bulimia nervosa and depression. Behaviour Research and Therapy, 36, 895 – 898. Cooper, M. J., Rose, K. S. & Turner, H. (in press). Core beliefs and the presence or absence of eating disorder symptoms and depressive symptoms in adolescent girls. International Journal of Eating Disorders. Cooper, M. J., & Turner, H. (2000). Underlying assumptions and core beliefs in anorexia nervosa and dieting. British Journal of Clinical Psychology, 39, 215 – 218. Cooper, M. J., Wells, A., & Todd, G. (2004). A cognitive theory of bulimia nervosa. British Journal of Clinical Psychology, 43, 1 – 16. Garner, D. M., & Garfinkel, P. E. (1979). The eating attitudes test: An index of the symptoms of anorexia nervosa. Psychological Medicine, 9, 273 – 279. Guidano, V. F. (1987). Complexity of the self. New York7 Guilford. Guidano, V. F., & Liotti, G. (1983). Cognitive processes and emotional disorders: A structural approach to psychotherapy. New York7 Guilford Press. Waller, G., Kennerley, H. & Ohanian, V. (2005). Schema-focussed cognitive behaviour therapy with the eating disorders. In L. P. Riso, P. T. du Toit & J. E. Young (Eds.), Cognitive schemas and core beliefs in psychiatric disorderss: A scientist practitioner guide. New York: American Psychiatric Association. Waller, G., Meyer, C., & Ohanian, V. (2001). Psychometric properties of the long and short versions of the Young Schema Questionnaire: Core beliefs among bulimic and comparison women. Cognitive Therapy and Research, 25(2), 137 – 147. Waller, G., Ohanian, V., Meyer, C., & Osman, S. (2000). Cognitive content among bulimic women: The role of core beliefs. International Journal of Eating Disorders, 28, 235 – 241. Waller, G., Shah, R., Ohanian, V., & Elliott, P. (2001). Core beliefs in bulimia nervosa and depression: The discriminant validity of Young’s Schema Questionnaire. Behaviour Therapy, 32, 139 – 153. Young, J. E. (1998). Young schema questionnaire — short form. New York7 Cognitive Therapy Centre. Available at http:// www.schematherapy.com