Encompassing subjective, perceptual, and attitudinal experiences about one’s body and physical appearances (Cash, 2004; Thompson et al., 2012), body image concerns have been referred to as a contemporary norm (Hosseini et al., 2020; Rosen et al., 1995). Body image distress is a risk factor for a range of negative outcomes including low self-esteem (Mellor et al., 2010), depression (Stice, 2002;), eating disorders (Sadibolova et al., 2019), and body dysmorphic disorder (Cooper & Osman, 2007; Veale, 2004). While there is considerable evidence for the efficacy of cognitive behavioural therapy (CBT) in decreasing body image concerns (e.g., Farrell et al., 2006; Lewis-Smith et al., 2016), there is a dearth of research looking at how specific cognitive processes may contribute to negative appearance evaluation (Linardon et al., 2019). Metacognitive beliefs (i.e., beliefs people have about their cognitive processes) concerning the value of worry and rumination may be a useful addition, given that past research has identified such beliefs as a transdiagnostic factor in psychopathology, including body dysmorphia and eating disorders (e.g., Cooper & Osman, 2007; Donyavi et al., 2015; Waine et al., 2009; Wells, 2019). Accordingly, this study developed a questionnaire designed to assess metacognition regarding negative appearance evaluation; that is, one’s metacognitive beliefs about physical appearance.
Theoretical Framework
Elevated metacognitions have been identified across a variety of psychological conditions and problem behaviours (Normann & Morina 2018; Wells 2019). Metacognition can be described as any knowledge or cognitive processes involved in the control, appraisal, monitoring and regulation of cognition (Wells & Matthews 1994, 1996). Wells and Matthew (1994, 1996) developed the transdiagnostic Self-Regulatory Executive Function Model (S-REF), which accounts for the universal role of metacognitive beliefs (i.e., metacognitions) in psychopathology. Beliefs, as outlined by Fodor (1983) and Quine and Ullian (1970), do not exist in isolation but rather form interconnected networks that shape cognitive processes (Connors & Halligan, 2015). These conceptualisations align with the S-REF model, which posits that metacognitive beliefs play a central role in maintaining cognitive attentional syndromes (CAS). While Fodor emphasizes the complexity of studying high-level beliefs, the S-REF model offers a framework for operationalising these beliefs by examining their influence on cognitive processes including forms of repetitive negative thinking, such as worry and rumination (Wells, 2009; Wells & Capoianco; 2020). Central to the S-REF is the principle that psychological problems/distress are caused by Cognitive Attentional Syndrome (CAS), an inflexible attentional processing style that comprises worry, rumination, threat monitoring, and self-regulation strategies that maintain perseverative negative thinking patterns. The CAS results from biased positive and negative metacognitions (Wells & Matthew, 1994, 1996). Positive metacognitive beliefs concern the usefulness of the CAS and the need to engage in cognitive activities such as worry or rumination (e.g., “If I worry, I will be prepared”, “Focusing on danger will keep me safe”). Conversely, negative metacognitive beliefs focus on the dangerousness, meaning, importance, or uncontrollability of thoughts or cognitive activities, such as worry or rumination (e.g., “I have no control over my thoughts”, “Worrying can harm me”) (Wells, 2009). Metacognitive beliefs about appearance play a dual role in both the aetiology and maintenance of psychopathologies such as Body Dysmorphic Disorder (BDD) and eating disorders. In the early stages of these disorders, positive metacognitive beliefs (e.g., ‘Worrying about my appearance will help me feel prepared or in control’) may contribute to the onset of maladaptive behaviours, while negative metacognitive beliefs (‘I cannot control my thoughts about my appearance’) are more likely to maintain these behaviours by perpetuating rumination and distress.
Previous Measures
Traditionally, frequently used measures for assessing metacognition in problem behaviours and psychological disorders are the self-report Metacognitions Questionnaire (MCQ; Cartwright-Hatton & Wells, 1997) and the shorter MCQ-30 (Wells & Cartwright-Hatton, 2004). Both MCQs are based on the S-REF model and assess five dimensions of dysfunctional metacognitive beliefs namely positive beliefs about worry; negative beliefs about uncontrollability of thoughts and corresponding danger; cognitive confidence; negative beliefs about thoughts in general; and cognitive self-consciousness. Research has supported strong links between metacognitive beliefs, as measured by the MCQs, and psychopathologies. For example, in their metanalysis, Sun et al. (2017) looking at 47 studies found increased levels of metacognitive beliefs in individuals (N = 3772) with eating disorder, generalised anxiety disorder, major depressive disorder, obsessive-compulsive disorder, and schizophrenia compared to healthy control individuals (N = 3376). These findings suggest that metacognitive beliefs are transdiagnostic and that negative evaluation of one’s mental state is an important consideration when looking at problem behaviours and psychological disorders (Sun et al. 2017).
A growing number of metacognitive measures have been developed and validated for specific disorders or problem behaviours (e.g., gambling, Spada et al., 2015; alcohol use, Spada & Wells, 2008; rumination, Kolubnski et al, 2017; Papageorgiou & Wells, 2003; health anxiety, Bailey & Wells, 2015; anger, Moeller et al., 2019; and insomnia, Waine et al., 2009) to enable sensitive assessment of unique cognitive processes, improve validity, and support targeted intervention (Spada & Wells, 2008; Wells, 2000, 2009). While metacognition has been explored in conditions with a body image component such as eating disorders (Aloi et al., 2020), obesity (Woolrich et al., 2008), and body dysmorphic disorder (Donyavi et al., 2015; Nikfarjam et al., 2015), no metacognitive measure relating specifically to physical appearance currently exists.
To date, psychometric tools have focused primarily on three dimensions of body image: subjective and affective measures of body (dis)satisfaction or disturbances relating to global and/or specific self-evaluation of appearance (Shroff et al., 2009; Thompson et al., 1999); behavioural aspects of body image such as avoidance of situations or body checking behaviours (Kling et al., 2019; Shroff et al., 2009); and cognitive aspects of body image that aim to evaluate thoughts and beliefs about one’s appearance (Shroff et al., Kling et al., 2019). The focus has largely been on the content of thoughts about one’s appearance rather than actual beliefs about cognitive processes (i.e., metacognitive beliefs) such as needing to engage in worry about one’s appearance, despite the latter playing an important role in appearance-related concerns. Development of a measure which specifically taps into metacognitive beliefs regarding appearance will assist researchers and clinicians to understand core factors related to body image distress, and can be used with existing clinical measures as part of a comprehensive battery of scales.
The Current Study
Appearance beliefs are an important contemporary issue due to the impact negative body image has on self-perception and psychological wellbeing. To effectively research expanding conceptualisations of appearance and body image, reliable and valid assessment measures are required, particularly in the context of assessing and treating psychopathology associated with the transdiagnostic construct of physical appearance concerns (e.g., body dysmorphia, eating disorders). Accordingly, the overarching goal of the current study was to develop and evaluate the psychometric properties of a new measure, the Metacognitions about Physical Appearance Questionnaire (MPAQ), which assesses metacognitive beliefs about worry or rumination on physical appearance and the importance of such thoughts and cognitive processes, across two independent samples. To meet this aim, the following steps were planned: (1) thematic analysis of reports from clinicians regarding metacognitive beliefs about appearance to assist with generating an item pool before obtaining feedback from experts and laypeople; (2) exploratory factor analyses to gauge the underlying structure of the MPAQ and assess items for reduction; (3) conduct confirmatory factor analysis and assess the suitability of fit indices; (4) evaluate the reliability of the MPAQ using Cronbach’s alpha, composite reliability, and McDonald’s ω; (5) investigate the convergent validity of the MPAQ by examining possible associations with existing measures of metacognitive beliefs, body image concern, rumination, appearance orientation, and frequency of worry; (6) investigate discriminant validity against theoretically distinct constructs including cognitive confidence and positive appearance evaluation; and (7) examine the incremnetal validity of the MPAQ by exploring predictive relationships with depression, anxiety, and stress.
Method
Item Generation for the MPAQ
To generate initial items to measure the presence and severity of positive and negative metacognitive beliefs about physical appearance, the research team (a) reviewed existing metacognitive and body image literature and measures; (b) sought expert opinion from experienced clinicians working within the field of body image disorders and associated conditions; and (c) conducted thematic analysis following steps outlined by Braun and Clarke (2006) of the information provided by clinicians to generate items before obtaining feedback from laypeople. That is, two of the researchers (LKD & JC), with expertise in metacognition research and therapy, independently familiarised themselves with the clinicians’ responses and generated initial themes. Then they met to discuss and mutually agree on themes to inform item wording. A convenience sample consisting of 23 professionals (Mage = 46, range 36-62 years) completed a brief online qualitative survey. These professionals were known to the researchers through their clinical work with individuals with body image concerns or were identified from health services that worked with body image issues. The expert panel consisted of psychologists, social workers, psychotherapists, dieticians, and mental health nurses. The clinical experience ranged from 2 to 35 years (M = 13) with predominant areas of experience including eating disorders, body dysmorphia, anxiety disorders, mood disorders, self-esteem, and oncology.
The survey presented participants with a modified version of the metacognitive profiling template (Wells & Mathews, 1994) to obtain expert professional opinion of metacognitions occurring in clients who present clinically with body image concerns. The modifications of the metacognitive profiling template included narrowing in on beliefs clients may have about their physical appearance and constructing wording to identify clients’ metacognitive beliefs about their physical appearance. For example, “What thoughts or beliefs have clients expressed regarding their worry or rumination about body image?”. Specifically, qualitative questions related to clients’ (1) thoughts, beliefs, and attitudes about their worry or rumination about body image, (2) advantages to worry or ruminations about their body image, (3) disadvantages to worry or ruminations about their body image, (4) goals for their worrying or rumination about their body image, (5) strategies to control or cope with their worry or rumination about their body image, and (6) any further comments. Subsequent thematic analysis, as noted above, of participants’ responses identified seven themes regarding positive beliefs about worrying/ruminating about physical appearance and six themes about negative beliefs on worrying/ruminating about physical appearance (see Table 1).
Table 1
Themes identified relating to positive and negative beliefs on worrying/ruminating about physical appearance
Positive beliefs about worry/rumination on physical appearance | Negative beliefs about worry/rumination on physical appearance |
|---|---|
1. Maintains control of my weight/size | 1. Makes me anxious / depressed |
2. Maintains control of my eating | 2. Makes me exhausted |
3. Controls my emotions | 3. Interferes with life |
4. Controls eating healthy | 4. Can’t control thinking about it |
5. Makes/motivates me to exercise | 5. Its harmful to my self-worth |
6. Helps my appearance to be attractive and controls how I am perceived by others | 6. Interferes with my relationships with others/social |
7. Maintains control of myself and my self-expectations |
The themes were used to guide initial item construction, jointly formulated by the researchers. The general term 'physical appearance' was used rather than specific body parts or features (e.g., ‘muscles,' 'weight,' 'stomach') to ensure that the MPAQ captured metacognitive beliefs applicable across a wide range of appearance-related concerns, consistent with literature that suggests individuals often form general impressions of their appearance rather than isolating their self-assessment to specific body parts (Cash & Pruzinsky, 2002). Focusing on the term ‘physical appearance’ also allows for inclusivity, ensuring that the measure is relevant for individuals with varied appearance issues while limiting construct fragmentation (Cash & Pruzinsky, 2002). To ensure face validity and provide feedback on the clarity of the instructions and items, 10 laypeople who were not experts or clinicians in body image reviewed the items independently while noting any difficulty understanding what the items were asking about. Feedback was used to improve the clarity and wording of the final pool of 30 items. The 30 items consisted of statements relating to beliefs that people may have when thinking about their physical appearance using a 5-point Likert response scale scored from 1 (strongly disagree) to 5 (strongly agree). Higher scores indicate a stronger presence of either positive or negative metacognitions about physical appearance with a total score obtained from the sum of the two subscales.
Procedure and Participants
Ethics approval was sought and granted by a university human research ethics committee for all phases of the study. A cross-sectional survey design was used for two independent samples, 12 months apart. Participants were invited to complete the anonymous survey via SurveyMonkey. The survey took approximately 25–30 min to complete.
Sample 1
A total of 454 participants were recruited via snowballing, advertisement on social media, in gyms, and recruitment of students from the home institution. Twenty-eight participants were deleted due to no response after consent and a further 57 deleted as they only responded to demographic items. The final sample consisted of 370 participants (Mage = 39.36; SD = 11.29; MBMI = 27.55; SD = 7.28) with 293 females and 77 males. Participation was voluntary and open to anyone ≥18 years. As an incentive, first-year university students at the home institution were offered course credits while other participants were invited to enter a prize draw to win one of two Amazon gift vouchers to the value of AUD$40.
Sample 2
In total, 540 participants were recruited via snowballing on online media platforms (e.g., Facebook) and advertisements in staff and student news bulletins at the home institution. Twenty-five participants were deleted due to no response after consent and a further 51 deleted as they only responded to demographic items. The final sample was composed of 463 participants (Mage = 30.88; SD = 12.67; MBMI = 25.88; SD = 6.26) with 377 females and 85 males. Participation was voluntary and open to anyone 18 years or above. First-year psychology students at the home institution were offered course credits as an incentive to participate in the study (see Table 2 for demographic information relating to Sample 1 and Sample 2).
Table 2
Demographic information of final sample 1 (N = 370) and sample 2 (N = 463)
Characteristics | Sample 1 | Sample 2 | Sample 1 | Sample 2 | |
|---|---|---|---|---|---|
(N) | (N) | (N) | (N) | ||
Ethnicity | Education Level | ||||
Australian | 272 | 365 | University degree | 213 | 107 |
New Zealander | 29 | 14 | TAFE/College diploma | 99 | 141 |
European | 44 | 35 | Year 12 or equivalent | 38 | 196 |
North American | 13 | 8 | < year 12 or equivalent | 20 | 18 |
Aboriginal/Torres Strait | 4 | 11 | Sexual Orientation | ||
Islander | Heterosexual | 326 | 388 | ||
Asian | 3 | 8 | Homosexual | 13 | 10 |
Maori | 1 | 0 | Bisexual | 26 | 59 |
Other | 4 | 13 | Other | 5 | 6 |
Country of Residence | Mental Health | ||||
Australia or NZ | 330 | 461 | Anxiety | 55 | 119 |
UK | 12 | – | Depression | 42 | 61 |
USA | 12 | – | Bipolar | 6 | 2 |
Other | 16 | 2 | PTSD | 4 | 17 |
Employment Status | Eating disorder | 5 | 2 | ||
Full or part time | 274 | 165 | BPD | 4 | 5 |
Casual | 29 | 105 | Gender dysphoria | 2 | 1 |
Full time student | 26 | 168 | Schizoaffective disorder | 1 | 1 |
Not in paid work | 41 | 25 | OCD | 2 | 6 |
BDD | 1 | 2 | |||
ADHD | 1 | 5 |
Measures
Demographics
Information regarding age, gender, ethnicity, country of residence, sexual orientation, employment status, education, mental health (i.e., do you consider yourself to be currently experiencing mental illness: yes/no; if yes, please specify), and weight and height (used to calculate BMI index) was recorded. BMI was recorded as part of the demographic information due to its association with body image concerns (e.g., Weinberger et al., 2017).
The Body Image Concern Inventory (BICI)
The BICI (Littleton et al., 2005) is comprised of 19-items assessing dysmorphic concerns including body dissatisfaction, checking behaviours and avoidance. Items include “I am ashamed of some part of my body” and “I examine flaws in my appearance”. Participants respond using a 5-point Likert scale from 1 (never) to 5 (always). Scores are summed (range 5–95) with higher scores indicative of greater body image concern. The BICI has demonstrated sound psychometric properties in prior research including for validation against other body image disorder measures (r = 0.76; Collison & Mahlberg, 2019) and discriminant validity with the Rosenberg self-esteem scale (r = −0.65; Collison & Mahlberg, 2019). The BICI has excellent internal consistency (α =0.91–0.96; Littleton et al., 2005). The BICI was used to assess the MPAQ’s convergent validity.
The Multidimensional Body Self-Relations Questionnaire (MBSRQ)
The MBSRQ (Brown et al., 1990) consists of 69 items and seven subscales measuring body image attitudes. For the current study, only the 12-item appearance orientation and the 9-item appearance evaluation subscales were used to measure evaluative, cognitive, and behavioural components toward physical appearance. Appearance orientation (AO) was selected for convergent validity, while appearance evaluation (AE) was selected for discriminate validity. Items include “My body is sexually appealing” (appearance evaluation) and “It is important that I always look good” (appearance orientation). Participants respond using a 5-point Likert scale from 1 (definitely disagree) to 5 (definitely agree). Two items on the appearance evaluation subscale and four items on the appearance orientation subscale are reversed scored. Subscale scores are obtained by averaging scores with a range from 1 to 5. Higher scores indicate more importance on how they look, attention to appearance, and extensive grooming (appearance orientation), and feeling mostly positive and satisfied with their appearance (appearance evaluation). Both subscales have demonstrated good internal consistency in norm samples (α = 0.85–0.88; Cash, 2000) and test–retest reliability (r = 0.81–0.91; Cash, 2000).
The Metacognitions Questionnaire-30 (MCQ-30)
The MCQ-30 (Wells & Cartwright-Hatton, 2004), is comprised of 30 items across 5 subscales capturing different aspects of metacognition including (1) positive beliefs about worry, (2) negative beliefs about uncontrollability of thoughts and danger, (3) cognitive confidence, (4) beliefs about need to control thoughts and (5) cognitive self-consciousness.
Items include “my worrying is dangerous for me” and “I monitor my thoughts”. Participants respond using a 4-point Likert scale from 1 (do not agree) to 4 (agree very much). Subscale scores and overall score are obtained by summating items on subscales and subscale totals respectively. Total scores range from 30 to 120 with higher scores indicating greater metacognitive beliefs, judgements and monitoring tendencies. The MCQ-30 has good to excellent internal consistency (α = 0.72–0.93; Wells & Cartwright-Hatton, 2004) and sound convergent validity with measures of compulsion and worry (r = 0.30–0.54 respectively; Wells & Cartwright-Hatton, 2004). This measure was selected because it is the shortened version of the gold standard in metacognitive belief measures (Wells & Capobianco, 2020); subscales and total score were used to evaluate convergent validity, discriminant validity, and incremental validity.
The Rumination-Reflection Questionnaire (RRQ)
The RRQ (Trapnell & Campbell, 1999) contains 24 items across two subscales (rumination and reflection). For the current study, only the 12-item rumination subscale was used to measure dispositional self-focus on negative attributes. Given the maladaptive nature of rumination, like worry, and its relationship to positive and negative metacognitive beliefs, the rumination subscale was used to assess the convergent validity of the MPAQ. Items include “I often find myself re-evaluating something I’ve done” and “I often reflect on episodes in my life that I should no longer concern myself with”. The scale uses a 5-point Likert scale with response options from 1 (strongly disagree) to 5 (strongly agree). There are three reversed scored items. Scores were obtained by averaging subscale items (range 1 to 5) with higher scores suggestive of greater rumination. The scale has demonstrated sound criterion and discriminant validity in prior research with self-rumination positively correlated with personal distress and negatively correlated with perspective-taking (e.g., Joireman & Hammersla, 2002). The scale also has good to excellent internal consistency (α = 0.88–0.90; Trapnell, & Campbell, 1999).
The Depression, Anxiety and Stress Scale–21 (DASS-21)
The DASS-21 is a brief version of the original 42-item DASS developed by Lovibond and Lovibond (1995). Each subscale is composed of seven items measuring depression (e.g., “felt down-hearted and blue”), anxiety (e.g., “I was aware of dryness of my mouth”), and stress (e.g., “I found it hard to wind down”) over the past week. Participants respond using 4-point Likert scale from 0 (did not apply to me at all) to 3 (applied to me very much, or most of the time). Scores on each subscale are summed with a range from 0 to 21, with higher scored indicative of greater levels of depression, anxiety and/or stress. The DASS-21 has demonstrated sound convergent and discriminant validity, and good to excellent internal consistency in prior research with Cronbach’s alphas on the depression, anxiety and stress subscales ranging from 0.88–0.94, 0.80–0.87 and 0.84–0.91 respectively (Sinclair et al., 2012). The DASS was included to assess convergent and incremental validity.
Frequency of Worry About Physical Appearance
A single item adapted by Bennett and Wells (2010) was included to capture frequency of worry about physical appearance, reflecting elevated symptoms rather than a formal diagnosis. Participants were asked to respond to the following item: “I am often preoccupied with thinking about (worrying, dwelling on) my physical appearance.” Responses were recorded using a 5-point Likert scale from 1 (strongly agree) to 5 (strongly disagree), with higher scores indicating less frequent worry about physical appearance. Because worry, like rumination, are types of maladaptive repetitive thinking, maintained by the CAS, including both negative and positive metacognitive beliefs, this measure was used to evaluate the convergent validity of the MPAQ.
Data Analysis
There is lack of consensus in the literature on sample size needed for exploratory factor analysis (EFA), with some indicating a total of 300 participants or over as good (e.g., Comfrey & Lee, 1992) while others recommend a participant-to-item ratio of more than 10:1 (e.g., Nunnally, 1978). Based on the literature (e.g., Mundfrom et al., 2009), the sample size of 369 was considered adequate for an EFA with a two-factor solution for Sample 1. Using a Monte Carlo approach to determining sample size for confirmatory factor analysis (CFA) based on a theoretical model, Myers and colleagues (2011) developed a rule of thumb of N ≥ 200. Considering this notion, Sample 1 and 2 were deemed sufficient for CFA. IBM SPSS Statistics version 26 and the lavaan package in R (R Core Team, 2000; Rosseel, 2012) were used to run data analyses.
Results
Assumption Testing
For the MPAQ, Kolmogorov-Smirnov and Shapiro-Wilk tests identified violations of normality (p < 0.05) on all items, but these tests are sensitive in large samples, and principal axis factoring is robust to such violations (Field, 2009; Tabachnick & Fidell, 2013). No missing cases were found in either sample. Skewness and kurtosis were evaluated using a ratio of skew or kurtosis to SE > ±3.29. Minor skewness occurred in 12 items for both samples, and minor kurtosis in 22 items for Sample 1 and 23 items for Sample 2. Three univariate outliers were identified in Sample 1 and two in Sample 2 (z scores > ±3.29). Mahalanobis distance identified 13 multivariate outliers in Sample 1 and 10 in Sample 2. All outliers were examined and retained as true scores.
For other measures, assumption testing was done at the scale/subscale level, as item-level data were not used. Little's MCAR test confirmed that missing data were random in both samples (Sample 1: 2.4–8.9%, X2(38) = 7.69, p = 1.00; Sample 2: 1.9%–7.8%, X2(28) = 7.69, p = 1.00). Pairwise deletion was used for statistical analyses, with Full Information Maximum Likelihood estimation applied for missing data in CFA. Univariate outliers (z scores > ±3.29) were found in some DASS and MCQ-30 subscales, but no extreme outliers were identified in box-plot inspections. Given the large sample size, no transformations were applied, congruent with the central limit theorem (Wilcox, 2005). Mahalanobis distance identified no multivariate outliers for either sample.
Factor Analyses
Sample 1
To identify underlying latent constructs in the items for the MPAQ, EFA was conducted (Fabrigar & Wegener, 2012; Hair et al., 2010). Given the presence of both skewness and kurtosis in addition to multivariate outliers, an EFA with principal axis factoring, using varimax rotation to improve interpretability while identifying correctness, was identified as most appropriate (Brown, 2015; Howard, 2016). To preserve validity, decisions regarding item retention and elimination were made following relevant psychometric (e.g., factor loadings, cross-loadings) and theoretical considerations (DeVellis, 2017; Worthington & Whittaker, 2006).
The correlation matrix found that the majority of correlations were above 0.3 with a determinant of 2.36 x 10−9. Further, Kaiser-Meyers-Olkin (KMO) measure of sampling accuracy was 0.921 and Bartlett’s test of sphericity p <0.001, suggesting the data was suitable for EFA (Howard, 2016). Eigenvalues above 1 indicated a possible five factor solution accounting for 64.77% of variance. Considering that two eigenvalues were close to 1 and the ongoing debate about the Kaiser (1960) criterion (i.e., all factors with eigenvalues above 1 should be retained), visual inspection of the scree plot suggested a possible two or three factor solution. Compared to 1000 randomly generated samples, parallel analysis supported the existence of a three-factor solution (Green et al., 2015; Horn, 1965).
Following literature regarding item retention for EFA (e.g., Howard, 2016), the rotated factor matrix was used to identify small factor loading (<0.3) and high cross-loadings to reduce the initial pool of 30 to 20 items across factors. Item reduction was performed through consideration of multiple criteria. This included removal of items with small factor loadings and removal of items with high-cross loadings (to avoid single items that may align to more than one factor). For some cross-loaded items, judgment was required to decide which item to retain or remove. This involved consideration of alignment to theoretically meaningful factors, and face validity of items (DeVellis, 2017; Pett et al., 2003; Williams et al., 2005; Worthington & Whittaker, 2006). For example, items with similar loadings, such as items 9 and 12, were differentiated based on their conceptual alignment with the intended constructs. Items that exhibited higher cross-loadings, like item 5, were retained due to their theoretical relevance (e.g., the importance of interpersonal appraisals on body image) and because item wording was more reflective of metacognitive belief item wording on measures such as the MCQ-30.
No specific cutoff was used for defining high-cross loadings, except that items were closely inspected when they loaded more than 0.3 onto two factors. Items 6, 11, 12, 13, 16, 23, and 26 were removed as they had high cross-loadings on Factor 3, items 19 and 24 were removed as they were the only items loading on Factor 4, while item 14 was removed due to low loading on Factor 2 (<0.5). A second EFA was re-run with the retained 20 items. Eigenvalues suggested that a two-factor model accounted for 54.81% of the total variance. A possible third factor identified on the scree plot was not pursued due to an eigenvalue close to 1 (1.19) and cross-loadings on both Factor 1 and Factor 2 (see Table 3). Parallel analysis confirmed the existence of a two-factor structure compared to 1000 randomly generated samples (Green et al., 2015; Horn, 1965).
Table 3
Exploratory factor solutions (principal axis factoring with varimax rotations) for initial pool of 30 items for the MPAQ in sample 1 (N = 370)
Initial EFA – 30 Items | Second EFA – 20 items | |||||||
|---|---|---|---|---|---|---|---|---|
F1 | F2 | F3 | F4 | F5 | F1 | F2 | F3 | |
1. Worrying about my physical appearance helps me to control my shape and weight. | 0.770 | 0.763 | ||||||
2. I must think about my physical appearance to help me to relax. | 0.507 | 0.335 | 0.503 | 0.321 | ||||
3. Worrying about my physical appearance helps me to control what I eat. | 0.807 | 0.763 | ||||||
4. Worrying about my physical appearance helps me control my emotions. | 0.555 | 0.354 | 0.506 | |||||
5. I need to worry about my physical appearance so that I look good to others. | 0.555 | 0.452 | 0.512 | 0.312 | ||||
6. I need to think a lot about how my physical appearance compares to others, so I know what needs to be improved. | 0.401 | 0.649 | – | |||||
7. Worrying about my physical appearance helps me to control my size. | 0.865 | 0.852 | ||||||
8. Worrying about my physical appearance motivates me to exercise. | 0.684 | 0.657 | ||||||
9. Worrying about my physical appearance helps me to keep looking good to others a priority. | 0.622 | 0.446 | 0.559 | 0.318 | ||||
10. Worrying about my physical appearance will help me to maintain control of myself. | 0.738 | 0.729 | ||||||
11. To improve my physical appearance, I need to analyse how it compares to others. | 0.407 | 0.570 | – | |||||
12. Dwelling on my physical appearance helps me to figure out how to achieve the body I want. | 0.642 | 0.413 | – | |||||
13. Dwelling on my past physical appearance, will help me achieve the physical appearance I want in the future. | 0.617 | 0.312 | – | |||||
14. Whenever I feel bad about my physical appearance, thinking about it will help me to understand why. | 0.469 | – | ||||||
15. Worrying about my physical appearance will make me anxious. | 0.661 | 0.650 | ||||||
16. I can’t stop thinking about my physical appearance. | 0.545 | 0.562 | – | |||||
17. Worrying about my physical appearance will make me exhausted. | 0.776 | 0.770 | ||||||
18. Worrying about or dwelling on my physical appearance will interfere with my life | 0.779 | 0.809 | ||||||
19. Worrying about or dwelling on my physical appearance means I am a self-centred person | 0.617 | 0.754 | – | |||||
20. Worrying about my physical appearance will harm my self-worth | 0.642 | 0.689 | ||||||
21. Worrying about or dwelling on my physical appearance will interfere with my relationships with others. | 0.749 | 0.674 | ||||||
22. Dwelling on my physical appearance will make me depressed | 0.526 | 0.772 | ||||||
23. No matter how hard I try, I can’t avoid thinking about my physical appearance | 0.572 | – | ||||||
24. Worrying about or dwelling on my physical appearance means that I am superficial or an unpleasant person | 0.708 | – | ||||||
25. I can’t focus on anything else when I am worrying about my physical appearance. | 0.580 | 0.527 | 0.438 | |||||
26. Worries or images of my physical appearance persist no matter what I do | 0.506 | 0.528 | – | |||||
27. I could make myself go crazy if I keep worrying about my physical appearance | 0.733 | 0.670 | 0.311 | |||||
28. I could make myself sick worrying about my physical appearance | 0.762 | 0.686 | 0.332 | |||||
29. Once I start thinking about my physical appearance, I can’t control it, no matter what I do | 0.590 | 0.354 | 0.502 | 0.515 | 0.662 | |||
30. Whenever I feel bad about my physical appearance, I can’t stop dwelling on it | 0.598 | 0.385 | 0.403 | 0.527 | 0.583 | |||
Eigenvalues | 9.76 | 5.54 | 1.76 | 1.28 | 1.09 | 6.50 | 4.40 | 1.19 |
(& variance explained) | 32.55 | 18.45 | 5.86 | 4.78 | 3.64 | 32.5 | 21.99 | 5.95 |
The remaining 20 items underwent a CFA to confirm the factor structure of the MPAQ using the lavaan package in R statistical software. The factor structure was based on the results of the EFA and was consistent with theory-driven expectations (i.e., positive and negative metacognitive beliefs; Wells, 2000). Conventional standards of model fit indices (e.g., Hu & Bentler, 1999) were used to evaluate the CFA measurement models. Initial analysis of the two-factor structure yielded a poor fit (X2(158) = 431.26, p < 0.001, CFI = 0.937, TLI = 0.925, RMSEA = 0.068, SRMR = 0.089). Due to shared residual variance and cross-loadings, items 2 and 9 were removed from the latent factor measuring positive beliefs about worry or rumination on physical appearance (i.e., MPAQ-PB), along with items 20, 29, and 30 from the latent factor measuring negative beliefs about worry or rumination on physical appearance (i.e., MPAQ-NB). According to Brown (2015), the default assumption of uncorrelated residuals can be a source of misfit, as such residuals were allowed to covary for retained items of similar wording and/or theme (i.e., items: 27 and 28; sick or crazy wording); 17 and 18 (wording related to worry exhausts or interferes); 4 and 10 (both relating to control); 18 and 21 (interference with life or relationships); 21 and 22 (harm to self and relationships). A final CFA model indicated adequate model fit with a two-factor structure in Sample 1 (X2 (84) = 209.76, p < 0.001, CFI = 0.958, TLI = 0.948, SRMR = 0.061, RMSEA = .064).
Sample 2
The final CFA model described above for Sample 1 was replicated and confirmed as yielding an adequate fit to the data in Sample 2 (X2 (84) = 236.19, p < 0.001, CFI = 0.952, TLI = 0.940, SRMR = 0.072, RMSEA = 0.063). Item wording and standardised factor loadings from the CFAs relating to both samples are presented in Table 4. Significant inter-item correlations ranged from −0.15 to 0.80 (0.38–0.74, MPAQ-PB; 0.39–0.80, MPAQ-NB).
Table 4
Standardised factor loadings in confirmatory factor analyses for the MPAQ in Sample 1 (N = 370) and Sample 2 (N = 463)
Sample 1 | Sample 2 | |
|---|---|---|
MPAQ-PB | ||
1. Worrying about my physical appearance helps me to control my shape and weight. | 0.801 | 0.768 |
3. Worrying about my physical appearance helps me to control what I eat. | 0.808 | 0.747 |
4. Worrying about my physical appearance helps me control my emotions. | 0.494 | 0.413 |
5. I need to worry about my physical appearance so that I look good to others. | 0.571 | 0.381 |
7. Worrying about my physical appearance helps me to control my size. | 0.898 | 0.870 |
8. Worrying about my physical appearance motivates me to exercise. | 0.685 | 0.612 |
10. Worrying about my physical appearance will help me to maintain control of myself. | 0.730 | 0.663 |
MPAQ-NB | ||
15. Worrying about my physical appearance will make me anxious. | 0.715 | 0.711 |
17. Worrying about my physical appearance will make me exhausted. | 0.760 | 0.804 |
18. Worrying about or dwelling on my physical appearance will interfere with my life. | 0.728 | 0.808 |
21. Worrying about or dwelling on my physical appearance will interfere with my relationships with others. | 0.588 | 0.565 |
22. Dwelling on my physical appearance will make me depressed. | 0.760 | 0.744 |
25. I can’t focus on anything else when I am worrying about my physical appearance. | 0.650 | 0.603 |
27. I could make myself go crazy if I keep worrying about my physical appearance. | 0.701 | 0.709 |
28. I could make myself sick worrying about my physical appearance. | 0.737 | 0.722 |
Independent samples t-tests, equal variances unassumed, were used to gauge possible gender (i.e., male and female) differences across the MPAQ subscales, one participant in Sample 2 identified as “neither/other” and was not included in the analysis. While MPAQ-PB scores were non-significant in both samples, t(155.03) = 1.16, p = 0.250, d = 0.15 (Sample 1) and t(117.40) = − 0.208, p = 0.835, d = − 0.03 (Sample 2), MPAQ-NB scores were significantly higher for females in both samples, t(116.43) = − 4.83, p <0.001, d = − 0.63 (Sample 1) and t(123.23) = − 5.03, p <0.001, d = 0−0.61 (Sample 2).
Reliability Analyses
Internal Consistency
The final 15-item MPAQ scale (see supplementary material) and subscales exhibited good to excellent internal consistency in both samples. The composite reliability (ρ), Cronbach’s alpha (α), and McDonald’s (ω) for the scale and subscales can be found in Table 5. Composite reliability coefficients were calculated using average variance extracted after factor analysis in SPSS and Excel Omega (total) ω was calculated according to McNeish’s (2017) guidelines. Given the non-normality of items, we interpreted reliability based on McDonald’s omega, whereby 0.70 or higher is indicative of acceptable internal consistency (McNeish, 2017).
Table 5
Reliability statistics for the MPAQ in Sample 1 (N = 370) and Sample 2 (N = 463)
Sample 1 | Sample 2 | |||||
|---|---|---|---|---|---|---|
ρ | ω | α | ρ | ω | α | |
MPAQ-15 | 0.95 | – | 0.84 | 0.95 | – | 0.83 |
MPAQ-PB | 0.91 | 0.89 | 0.88 | 0.87 | 0.84 | 0.83 |
MPAQ-NB | 0.91 | 0.90 | 0.90 | 0.92 | 0.90 | 0.90 |
Convergent Validity
Convergent validity of the MPAQ-15 was tested in both Sample 1 and Sample 2 by running a series of Pearson’s bivariate correlations. Correlation magnitudes have been interpreted based on Cohen (1988), with small correlations at 0.10, moderate correlations at 0.30, and large correlations at 0.50. As anticipated, moderate positive correlations were found between MPAQ-PB and the MCQ-30 positive beliefs about worry; MPAQ-NB and the MCQ-30 negative beliefs about uncontrollability and danger of worry; and MPAQ with the MCQ-30. Further, the MPAQ demonstrated moderate positive correlations with the BICI, RRQ, the appearance orientation subscale of the MBSRQ, frequency of worry, and the DASS, supporting convergent validity (see Table 6).
Table 6
Correlations and 95% Confidence Intervals Between the MPAQ-15, MPAQ-PB, MPAQ-NB and Other Conceptually Related Measures in Sample 1 (N = 370) and Sample 2 (N = 463)
MCQ-30 (POS) | MCQ-30 (NEG) | MCQ-30 | BICI | RRQ | AO | FoW | Depression | Anxiety | Stress | M(SD) | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
MPAQ-PB | S1 | 0.43** [0.34, 0.51] | – | – | – | – | – | – | – | – | – | 19.52(6.10) |
S2 | 0.36** [0.27, 0.43] | – | – | – | – | – | – | – | – | – | 20.64(7.19) | |
MPAQ-NB | S1 | – | 0.52** [0.44, 0.59] | – | – | – | – | – | – | – | – | 23.83(7.30) |
S2 | – | 0.57** [0.50, 0.63] | – | – | – | – | – | – | – | – | 24.86(7.19) | |
MPAQ | S1 | – | – | 0.54** [0.46, 0.61] | 0.62** [0.55, 0.68] | 0.54** [0.46, 0.61] | 0.47** [0.39, 0.55] | 0.50** [0.38, 0.55] | 0.38** [0.29, 0.47] | 0.40** [0.30, 0.48] | 0.47** [0.38, 0.55] | 43.35(9.81) |
S2 | – | – | 0.46** [0.39, 0.53] | 0.64** [0.58, 0.69] | 0.41** [0.32, 0.48] | 0.50** [0.42, 0.56] | 0.50** [0.41, 0.55] | 0.38** [0.29, 0.45] | 0.38** [0.29, 0.46] | 0.42** [0.34, 0.49] | 45.50(9.29) | |
M(SD) | S1 | 10.00(4.01) | 11.97(4.95) | 58.11(16.43) | 53.78(15.59) | 3.47(.89) | 3.29(.68) | 3.35(1.17) | 11.99(4.85) | 11.00(3.82) | 14.00(4.37) | – |
S2 | 11.11(4.19) | 13.23(5.20) | 63.46(16.11) | 56.19(15.91) | 3.67(.82) | 3.40(.67) | 3.45(1.12) | 13.29(4.75) | 12.44(4.46) | 14.78(4.35) | – | |
α | S1 | 0.90 | 0.90 | 0.94 | 0.95 | 0.94 | 0.87 | – | 0.93 | 0.85 | 0.88 | – |
S2 | 0.89 | 0.90 | 0.92 | 0.95 | 0.93 | 0.87 | – | 0.91 | 0.87 | 0.87 | – |
Discriminant Validity
Discriminate validity was tested in both samples using a Pearson’s bivariate correlations between the MPAQ-15, its subscales, and measures hypothesised to assess distinct constructs, including the MCQ-30 subscales of cognitive confidence (MCQ-CC) and self-consciousness (MCQ-CSC), and the appearance evaluation (AE) subscale of the MBSRQ. Weak correlations were found between MPAQ-PB and MPAQ-NB and cognitive confidence across both samples, confirming distinct constructs. In Sample 1, a weak correlation with cognitive self-consciousness (r = 0.29), suggested some shared cognitive processes but was weaker than those used to support convergent validity (e.g., MPAQ-PB and MCQ-30 POS, r = 0.36), suggesting that the constructs are related but distinct. A moderate negative correlation with the appearance evaluation subscale (r = -0.47 in Sample 2), indicated that metacognitive beliefs about appearance differ from feelings of physical attractiveness, supporting the discriminant validity of the MPAQ (see Table 7).
Table 7
Correlations and 95% confidence intervals between the MPAQ-15 and conceptually different measures in Sample 1 (N = 370) and Sample 2 (N = 463)
MCQ-30 (CC) | MCQ-30 (CSC) | AE | ||
|---|---|---|---|---|
MPAQ-PB | S1 | 0.15** [−0.04, 0.25] | 0.21** [−0.10, 0.31] | – |
MPAQ-NB | S2 S1 | −0.01[−0.11, 0.08] 0.27** [0.17, 0.37] | 0.07 [−0.02, 0.17] 0.29**[0.19, 0.40] | -.01 [−0.10, 0.08] – |
S2 | 0.28** [0.19, 0.37] | 0.18**[0.09, 0.27] | -.47** [−0.54, −0.38] | |
MPAQ-15 | S1 | 0.29** [0.19, 0.39] | 0.35**[0.25, 0.44] | – |
S2 | 0.21** [0.12, 0.30] | 0.18**[0.9, 0.27] | -.37** [−0.45, −0.28] | |
M(SD) | S1 | 10.95(4491) | 14.91(4.66) | – |
S2 | 11.42(4.59) | 16.20(4.38) | 2.87(.96) | |
α | S1 | 0.90 | 0.88 | – |
S2 | 0.90 | 0.86 | 0.91 | |
Incremental Validity
A series of hierarchical regression analyses were conducted to examine the incremental validity of the MPAQ-15 in cross-sectionally predicting the DASS subscales of depression, anxiety, and stress in both samples in addition to variance accounted for by the MCQ-30 (see Table 8). The MCQ-30 demonstrated significant relationships with depression, anxiety, and stress in both samples, accounting for 30.4 to 40.9% of the variance in participants' scores. The addition of the MPAQ resulted in small but significant model changes for depression and stress in sample 1, and depression, stress, and anxiety in sample 2, accounting for 1-2% of the variance in participants’ scores. Overall, the results provide initial evidence of incremental validity for the MPAQ-15.
Table 8
Unstandardised (B) and standardised (β) regression coefficients, and squared semi-partial correlations (sr2) for the MCQ-30 and MPAQ-15 in Both Sample 1 (N = 337) and Sample 2 (N = 427) in hierarchical regression models predicting DASS subscales of depression, anxiety and stress
Depression | Anxiety | Stress | |||||||
|---|---|---|---|---|---|---|---|---|---|
β | sr2 | B [95% CI] | β | sr2 | B [95% CI] | β | sr2 | B [95% CI] | |
S1: Step 1 | |||||||||
MCQ-30 | 0.55 | 0.449 | 0.164 [0.137, 0.190]** | 0.61 | 0.509 | 0.141 [0.121, 0.161]** | 0.64 | 0.520 | 0.170 [0.148, 0.192]** |
R2 | 0.308 | 0.368 | 0.409 | ||||||
Adjusted R2 | 0.306 | 0.366 | 0.408 | ||||||
S1: Step 2 | |||||||||
MCQ-30 | 0.49 | 0.449 | 0.145 [0.114, 0.177]** | 0.55 | 0.509 | 0.129 [0.105, 0.152]** | 0.54 | 0.520 | 0.145 [0.119, 0.170]** |
MPAQ-15 | 0.12 | 0.117 | 0.057 [0.005, 0.109]* | 0.10 | 0.103 | 0.038 [0.001, 0.077]† | 0.18 | 0.195 | 0.079 [0.036, 0.122]** |
R2 | 0.318 | 0.375 | 0.432 | ||||||
Adjusted R2 | 0.314 | 0.371 | 0.429 | ||||||
∆R2 | 0.01* | 0.01† | 0.02** | ||||||
S2: Step 1 | |||||||||
MCQ-30 | 0.55 | 0.459 | 0.164 [0.141, 0.188]** | 0.61 | 0.537 | 0.172 [0.151, 0.193]** | 0.63 | 0.549 | 0.173 [0.153, 0.193]** |
R2 | 0.304 | 0.379 | 0.403 | ||||||
Adjusted R2 | 0.302 | 0.377 | 0.402 | ||||||
S2: Step 2 | |||||||||
MCQ-30 | 0.48 | 0.459 | 0.143 [0.117, 0.170]** | 0.56 | 0.537 | 0.157 [0.133, 0.180]** | 0.56 | 0.549 | 0.153 [0.131, 0.176]** |
MPAQ-15 | 0.15 | 0.163 | 0.079 [0.033, 0.125]** | 0.12 | 0.133 | 0.058 [0.017, 0.098]* | 0.16 | 0.180 | 0.074 [0.035, 0.113]** |
R2 | 0.322 | 0.390 | 0.423 | ||||||
Adjusted R2 | 0.319 | 0.387 | 0.420 | ||||||
∆R2 | 0.02** | 0.01* | 0.02** | ||||||
Discussion
The present study aimed to develop and test the psychometric properties of a scale measuring metacognitions about physical appearance, the MPAQ. Overall, results revealed that the MPAQ had a clear factor structure and two theoretically driven subscales. Reliability and validity of the scale were supported, suggesting the MPAQ is a psychometrically sound measure. A panel of experts, a review of the existing literature and current metacognitive measures (e.g., Cartwright-Hatton & Wells, 1997; Wells & Cartwright-Hatton, 2004; Cooper & Osman, 2007; Nikfarjam et al., 2015), and thematic analysis (see Wells & Matthews, 1994) was used to develop the initial 30-item MPAQ. As supported by the literature (e.g., Bornstein, 1996; Matsunaga, 2010) the item pool was purposely expansive to maximise face validity and avoid missing important aspects of metacognitive beliefs related to physical appearance. Congruent with the initial thematic analysis, the final 15-item, two-factor solution of the MPAQ aligned with theory-driven expectations (i.e., positive and negative metacognitive beliefs; Wells, 2000). However, upon review of the S-REF model (Wells & Matthews, 1996) and the MCQ and MCQ-30 (Cartwright-Hatton & Wells, 1997; Wells & Cartwright-Hatton, 2004), it is plausible that many of the removed items represented metacognitive factors other than positive and negative beliefs such as cognitive self-consciousness or cognitive control. Considering that the items were generated from thematic analysis of clinicians’ reports regarding common metacognitive beliefs for those with body image concerns, it is plausible that positive and negative metacognitive beliefs about physical appearance are pertinent dysfunctional forms of cognition for this population. This idea is mirrored in a meta-analysis by Sun and colleagues (2017) who found moderate-to-large effects for negative and positive metacognitions across psychopathologies. It remains unclear if positive or negative metacognitive beliefs about appearance may vary in strength and dysfunction across different types of body image and eating disorder difficulties, highlighting avenues for future research.
Internal reliability of the MPAQ was tested with Cronbach’s alpha. Cronbach’s alpha has been criticised as providing the lower-bound estimate for the composite score reliability, thus underestimating true reliability (Hayes & Coutts, 2020; Peterson & Kim, 2013). Hence, composite reliability coefficients and McDonald’s Omega were also included to test the internal reliability of the scale. All three tests confirmed good to excellent internal reliability of the MPAQ and its’ subscales.
Findings also supported the convergent validity of the MPAQ. Specifically, convergent validity of the MPAQ and its subscales was supported in both samples with moderate correlations demonstrated between the MPAQ and the MCQ-30, MPAQ-PB and MCQ-30 positive belief about worry, and MPAQ-NB and MCQ-30 negative beliefs about uncontrollability and danger of worry. Significant associations were also observed between the MPAQ and BICI, RRQ, appearance orientation subscale of the MBSRQ, frequency of worry, and the DASS. Strong correlations between the BICI and MPAQ were anticipated due to their conceptual overlap. While the MPAQ focuses on metacognitive beliefs about appearance-related worry, the BICI assesses concerns related to physical dissatisfaction and dysmorphic features. Correlations were selectively run between the full MPAQ and some measures, and between subscales and others, based on the conceptual relevance of the constructs. For instance, the positive beliefs subscale of the MPAQ was correlated with the MCQ-30 (POS) due to their shared focus on positive metacognitive beliefs. Similarly, the full MPAQ was compared to the BICI and the DASS, given that both measures capture broader concerns about physical appearance. This approach was intended to reflect the alignment between specific aspects of the MPAQ and corresponding constructs in other measures. The findings are collectively supported by literature which has found that negative body image is strongly linked to the development of body dysmorphia (e.g., Rosen et al., 1995; Veale, 2004), rumination (Etu & Gray, 2010), and appearance evaluation (Cash, 1994).
Discriminant validity was demonstrated across both samples. The MPAQ-PB and MPAQ-NB were both distinct from cognitive confidence (MCQ-CC), cognitive self-consciousness (MCQ-CSC), and appearance evaluation. In Sample 1, the weak correlation with cognitive self-consciousness suggests some overlap in cognitive processes such as self-awareness. However, the MPAQ remains distinct in its focus on metacognitive beliefs related to appearance. In Sample 2, the MPAQ demonstrated a moderate negative correlation with appearance evaluation, indicating that metacognitive beliefs about physical appearance are distinct from positive evaluations of one's appearance; and that individuals with higher appearance-related metacognitive concerns tend to have less positive evaluations of their appearance. The similarity in content between MPAQ and appearance evaluations may indicate that such appraisals are more ego-syntonic and self-focused, consistent with metacognitive theory, where negative beliefs are more prominent than positive ones. Given the smaller correlations observed with the MPAQ-PB subscale, these findings should be interpreted cautiously, as they may reflect less robust discriminant validity for this subscale. Further research could investigate the specific metacognitive processes captured by the MPAQ-PB in both clinical and non-clinical populations. No data was available for these constructs in Sample 1 and thus this relationship should be replicated in future studies.
Finally, incremental validity was demonstrated as the MPAQ explained additional variance in depression, anxiety, and stress across both samples, beyond what was accounted for by the MCQ-30. This supports the notion that appearance-related metacognitive beliefs have a significant impact on mental health outcomes. This finding aligns with past research, which has shown that negative body image and appearance-related concerns are closely tied to poor mental health outcomes (e.g., Cash et al., 2004; Etu & Gray, 2010; Ivarsson et al., 2006; Satghara et al., 2019; Stice, 2002). For example, Cash and colleagues (2004) found that negative body image was associated with social-evaluation anxiety and intimacy anxiety in women. Similarly, stress has been linked to negative body image in adolescents (Murray et al., 2013), and numerous studies have demonstrated a strong connection between negative body image and depression (e.g., Manaf et al., 2016; Rierdan et al., 1989; Stice, 2002; Stice et al., 2000). Importantly, the MPAQ showed incremental validity by explaining additional variance in depression, anxiety, and stress, beyond that explained by the MCQ-30. This suggests that the MPAQ captures unique aspects of metacognitive beliefs about appearance that contribute to mental health outcomes, making it a valuable tool for understanding how appearance-related worries specifically impact psychological well-being.
Theoretical and Practical Implications
The current study suggests that the MPAQ is a reliable and valid measure. The MPAQ may provide further insight concerning metacognitive processes implicated in appearance related concerns and guide future research, clinical assessments, and therapeutic interventions in those populations where appearance related concerns contribute to psychological distress. Research suggests that metacognitions are central in understanding appearance related disorders such as body dysmorphic disorder (e.g., Cooper & Osman, 2007; Donyavi et al., 2015) and eating disorders (e.g., Olstad et al., 2015). Individuals scoring high on the MPAQ may benefit from metacognitive therapy (Wells, 2009); however, this needs further research.
The findings of the current study also have theoretical implications. The two subscales of the MPAQ measuring positive and the negative beliefs about physical appearance corroborates the S-REF theory (Wells, 2000). Particularly, positive beliefs can be seen as activating the CAS and the need to engage in worry or ruminations while negative beliefs maintain and exacerbate the CAS by focusing on the danger or uncontrollability or worry and rumination (Spada et al., 2015; Wells, 2000). Collectively this suggests that metacognitive beliefs are present and relevant in physical appearance concerns.
Strengths, Limitations, and Future Directions
The current study was strengthened by two large independent samples, which had sufficient power, minimising the margin of error around estimates (Faber & Fonseca, 2014). To the authors’ knowledge, this is also that first known measure of metacognition about physical appearance that may be a valuable transdiagnostic measure in disorders where appearance related metacognitions are implicated. The MPAQ is a brief and easy to administer, free self-report measure which could be used by clinicians, researchers, and other health professionals to inform assessments related to individuals’ physical appearance concerns.
This study was limited by using two cross-sectional homogenous samples composed largely of Caucasian females over the age of 18 years. Consequently, caution should be exercised when generalising to different ethnic groups and males. Further validation of the MPAQ is recommended in male samples, diverse ethnicities, and clinical samples (e.g., individuals diagnosed with eating disorder or body dysmorphic disorder). The use of a cross-sectional correlational design also meant we were unable to examine test-retest reliability of the MPAQ and the temporal precedence among related constructs, establishing concurrent rather than predictive validity. It is recommended that future studies include a longitudinal element to examine both the consistency of the MPAQ over time and predicative validity with additional ancillary measures of metacognition, body image concern, and psychological distress. Unfortunately, measurement invariance was not considered in the current study and represents a consideration for future research.
Although the expert panel was not consulted again after the items were generated, the items were grounded in robust qualitative feedback from experienced clinicians, ensuring relevance and alignment with clinical themes. To further validate the content of the items, future research could engage expert panels to review and refine the generated items, ensuring that they fully capture the identified themes. Finally, the study was limited by using a long survey packed without attention checks which may have caused responder fatigue and thus impacted the internal validity of the study.
Conclusion
The MPAQ is a valid and reliable measure of metacognition, that is metacognitive beleifs, about physical appearance in nonclinical smaples. Preliminary evidence for its psychometric properties was found across two independent samples. In accordance with metacognitive theory, the MPAQ measure positive and negative metacognitive beliefs related to worry and rumination about physical appearance. Although further validation is needed in clinical populations, the MPAQ shows promise as a valuable tool for researchers and health professionals investigating the role of metacognitive beliefs in body image and eating disorders.
Acknowledgements
The authors wish to thank the health professionals and laypeople who contributed to the item development phase, the many participants who engaged in the psychometric properties testing phase, and Ezyfit Health Club and its members (Sunshine Coast, Australia).
Declarations
Conflict of interest
The authors Lee Kannis-Dymand, Karina Rune, Nicola Mitchell, Janet D. Carter, Jacob Keech, Andrew Allen, Kate Mulgrew, David Tod, and Geoff P Lovell report that there are no competing interests to declare.
Ethical Approval
The project received ethical approval from a University Human Research Ethics Committee.
Infomed Consent
Informed consent was obtained from participants involved in the research project.
Animal Rights
No animals were used in the presented studies.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.