Elsevier

Journal of Anxiety Disorders

Volume 11, Issue 3, May–June 1997, Pages 279-296
Journal of Anxiety Disorders

Beliefs about Worry and Intrusions: The Meta-Cognitions Questionnaire and its Correlates

https://doi.org/10.1016/S0887-6185(97)00011-XGet rights and content

Abstract

This report describes the development of the Meta-Cognitions Questionnaire to measure beliefs about worry and intrusive thoughts. Factor analyses of the scale demonstrated five empirically distinct and relatively stable dimensions of meta-cognition. Four of the factors representing beliefs were: Positive Beliefs About Worry; Negative Beliefs About the Controllability of Thoughts and Corresponding Danger; Cognitive Confidence; and Negative Beliefs about Thoughts in General, including Themes of Superstition, Punishment and Responsibility. The fifth factor represented Meta-Cognitive processes — Cognitive Self-Consciousness — a tendency to be aware of and monitor thinking. The measure showed good psychometric properties on a range of indices of reliability and validity.

Scores on the questionnaire subscales predicted measures of worry proneness, proneness to obsessional symptoms, and anxiety. Regression analyses showed that the independent predictors of worry were: Positive Beliefs about Worry; Negative Beliefs About the Controllability of Thoughts and Corresponding Danger; and Cognitive Confidence. Significant differences in particular MCQ subscales were demonstrated between patients with intrusive thoughts, clinical controls and normals. The implications of these findings for models of worry and intrusive thoughts are discussed. © 1997 Elsevier Science Ltd

Section snippets

INTRODUCTION

Wells (1995) (1997) has advanced a meta-cognitive model of worry in Generalised Anxiety Disorder, in which individuals who develop problematic worrying are thought to hold a range of both positive and negative beliefs about worry. Positive beliefs stimulate the use of worrying as a processing or coping strategy. Normal worry is then transformed into pathological varieties like that found in Generalised Anxiety Disorder (GAD) when worry itself becomes the focus of negative appraisals. In other

Item Selection and Preliminary Factor Analysis

The original items used in constructing the Meta-Cognitions Questionnaire (MCQ), were derived from two sources: a semi-structured interview with 25 (11 female) undergraduate students, and transcripts of cognitive therapy conducted with anxiety outpatients attending for treatment. The patients had diagnoses of; Generalised Anxiety Disorder (n = 4), Obsessive Compulsive Disorder (n = 3), Hypochondriasis (n = 3), Panic Disorder (n = 2). The subjects were questioned about their experience of worry

Full Scale Factor Analysis

The item pool was revised as a few items loaded on more than one factor. These redundant items, and those which loaded less than a conservative 0.40 on their factor were discarded. Additionally, some factors had relatively few items, so new items were generated. The revised scale, consisting of 79 items, was administered to a further set of 306 undergraduate and graduate students (mean age 22, range 18–55, 142 females). These data were analysed using a principal components method of factor

RESULTS

The final five factor version of the scale was constructed from items loading at least .35 on their respective factor for males and for females, and in the combined sex analysis loading at least 0.4 on the factor. Redundant items (i.e., those loading on more than one factor), were dropped, leaving 65 items in total. The five factors obtained were highly similar to those obtained in the previous study, but two of the previous factors collapsed together to form a single factor in this analysis.

Reliability of the Factor Structure

To determine the reliability of the factor structure obtained in study 3, the final version of the MCQ was administered to a new sample of 243 undergraduate and graduate students (age range 18–65, mean 25, 120 females), and we repeated the factor analysis. Sixty of the original 65 items loaded at greater than 0.4. All but one of these items loaded on the same factor as in study 3, suggesting a relatively stable five-factor structure. The proportion of variance accounted for by the factors with

ALPHA RELIABILITIES

Cronbach coefficient alphas for each of the final MCQ subscales were computed using data from the 306 subjects used in study 3. The alpha coefficients were: Positive Beliefs = .87; Uncontrollability and Danger = .89; Cognitive Confidence = .84; SPR3 = .74; Cognitive Self-Consciousness = .72.

ITEM TOTAL CORRELATIONS

Item total correlations were computed using data from the 306 subjects from study 3. A few items were relatively weak, and less representative of their factor. However, at this stage of the analysis, these items were retained in order to offer the most comprehensive measure of dimensions for the subsequent predictive analyses. The item total correlations with each factor (corrected for item overlap), are displayed in Table 5.

SUBSCALE INTER-CORRELATIONS

Pearson correlations and partial correlations (partialling out the effects of overlap between the factors) were computed between the MCQ subscales using the data from the 306 subjects in study 3. The results are displayed in Table 6. Positive Beliefs correlated with Self-Consciousness, and SPR.4 Uncontrollability and Danger correlated with all of the subscales, except for Positive

Test-Retest Reliabilities

In order to test the stability of subscale scores over a period of time, the MCQ was administered to a new sample of 47 subjects drawn from University employees, and postgraduate students, (age range 20–55, 22 females) on two occasions 5 weeks apart. Pearson test-retest correlations for the subscales and total score were as follows: Positive Beliefs = .85 (p < .001); Uncontrollability and Danger = .89 (p < .001); Cognitive Confidence = .84 (p < .001); SPR4 = .76 (p < .001); Cognitive

STUDY 6

Study six sought to identify personality measures associated with individual differences in beliefs about worry. The intention was to validate the factors of the MCQ against existing measures and test specific theoretical predictions. The following measures were chosen:

  • 1.

    1. Spielberger Trait Anxiety Inventory. (Spielberger, Gorsuch, & Lushene, 1983). A 20-item questionnaire measuring anxiety proneness.

  • 2.

    2. The Padua Inventory. A self-report measure of obsessive-compulsive symptoms. (Subscale 1; 16

CONCURRENT VALIDITY: FIRST FINDINGS

It was predicted that the Cognitive Self-Consciousness subscale of the MCQ should correlate with the Private Self-Consciousness subscale. Private Self-Consciousness measures tendencies to focus on one's thoughts, feelings, moods, and attitudes. It, therefore, provides partial concurrent validation of the Cognitive Self-Consciousness Subscale in particular. The Padua Impaired Control of Mental Activities subscale was included as a test of concurrent validity for the MCQ Uncontrollability and

TESTS OF THEORETICAL PREDICTIONS

In this study, individual correlates of worry and intrusive thoughts were examined. Wells' (1994a) theory, predicts that meta-cognitive dimensions encompassing both positive and negative beliefs should correlate positively with worry and intrusions. It was predicted that MCQ subscales should correlate significantly with anxiety proneness as measured by the Spielberger scale, and worry as measured by the Anxious Thoughts Inventory. Correlations between questionnaire measures and MCQ subscales

REGRESSION ANALYSES

Further predictive analyses of general worry proneness, impaired control of mental activities, and obsessional checking were conducted while controlling for the statistical interdependence of MCQ subscales and trait anxiety. Hierarchical multiple regression analyses were run. Trait anxiety was entered on the first step and MCQ subscales on the second step to determine whether the MCQ subscales made independent contributions to scores on the Anxious Thoughts Inventory, Padua Checking, and

Discriminant Validity

The Questionnaire was given to 32 patients (mean age 37, 21 female) with Generalised Anxiety Disorder (GAD), 17 patients (mean age = 33, 7 female) with a diagnosis of Obsessive Compulsive Disorder (OCD), 14 patients (mean age 41, 8 female) with emotional disorders other than GAD or OCD, and a group of 30 nonclinical subjects (mean age 32, 15 female). All patients met DSM-III-R criteria for their respective disorder. Scores on each of the subscales of the MCQ are presented in Table 9. An

DISCUSSION

Factor Analyses of responses to the MCQ demonstrated that this questionnaire measures five empirically distinct categories of beliefs and processes concerned with ones own thoughts, or “meta-cognitions.” The five categories, or subscales are: (1) Positive Beliefs about Worry: this consists of items relating to the belief that worrying helps to solve problems, and to avoid unpleasant situations. It also includes items which suggest that worrying is a necessary feature of a pleasant and normal

Acknowledgements

Both authors are grateful to the Medical Research Council UK for their support. Adrian Wells is also grateful to the Wellcome Trust.

References (16)

There are more references available in the full text version of this article.

Cited by (0)

View full text