BELIEVABILITY OF ANXIOUS FEELINGS AND THOUGHTS 1
The Believability of Anxious Feelings and Thoughts Questionnaire (BAFT): A Psychometric
Evaluationof Cognitive Fusion in a Non-Clinical and Highly-Anxious Community Sample
Kristin N. Herzberg, Sean C. Sheppard, John P. Forsyth, Marcus Crede, Mitch Earleywine
University at Albany, State University of New York
Georg H. Eifert
Chapman University
Author Note
Correspondence concerning this article should be addressed to Kristin N. Herzberg or John P. Forsyth, University at Albany, SUNY, Department of Psychology, SS399, 1400 Washington Avenue, Albany, New York, 12222; Email: or . Ph: 518-442-4862; Fax: 518-442-4867. 5187294085
Abstract
Cognitive fusion, or the tendency to buy into the literal meaning of unpleasant thoughts and feelings, plays an important role in the etiology and maintenance of anxiety disorders and figures prominently in third generation behavior therapies such as Acceptance and Commitment Therapy (ACT). Nonetheless, there is a lack of validated self-report measures of cognitive fusion/defusion, particularly in the area of anxiety disorders. We attempted to fill this gap with the development and validation of a self-report cognitive fusion measure -- The Believability of Anxious Feelings and Thoughts (BAFT) questionnaire -- in both a healthy undergraduate (N = 437) and highly anxious (N = 503) community sample. Results from evaluations of the factor structure, along with findings from a variety of other psychometric analyses, suggest that the BAFT is a reliable, valid, and useful measure of cognitive fusion.
Keywords: cognitive fusion, defusion, anxiety, fear, Acceptance and Commitment Therapy
The Believability of Anxious Feelings and Thoughts Questionnaire (BAFT): A Psychometric Evaluationof Cognitive Fusion in a Non-Clinical and Highly-Anxious Community Sample
Acceptance and Commitment Therapy (ACT; Hayes, Strosahl & Wilson, 1999) is a newer cognitive-behavior therapy that seeks to undermine various forms of human suffering via the use of (a) acceptance and mindfulness processes linked with (b) commitment and behavior change processes in the service of (c) valued ends. As such, ACT shares a family resemblance with other emerging acceptance and mindfulness-based approaches such as Dialectical Behavior Therapy (DBT; Linehan, 1993), Mindfulness-Based Cognitive Therapy (MBCT; Segal, Teasdale, & Williams 2004), and Metacognitive Therapy (Wells, 2000). ACT is also closely linked with the field of behavior therapy writ large in terms of its empiricism, pragmatism, and use of behavior change technologies (e.g., behavioral activation, exposure) to support movement in valued directions. However, ACT differs from these traditions in terms of its philosophical assumptions (i.e., functional contextualism), roots in a behavior analytic account of language and cognition (i.e., Relational Frame Theory; Blackledge, 2003; Hayes et al., 1999; Hayes & Wilson, 1994), and its process-oriented models of psychopathology and psychological health (i.e., the Hexaflex model; see Hayes, 2004).
Indeed, from an ACT point of view, psychopathology is thought to pivot on processes that narrow behavioral repertoires and get in the way of effective action, chief among them being experiential avoidance (Hayes, Wilson, Gifford, Follette, & Strosahl, 1996) and cognitive fusion (Hayes et al., 1999). Recent work has focused on developing reliable and valid assessment devices to measure these and other ACT process variables (c.f., Hayes, Strosahl et al., 2004;Hayes, Luoma, Bond, Masuda & Lillis, 2006). However, this work is still in its early developmental stages, and many measures of ACT processes tend to be developed idiosyncratically to suit particular research aims. This is particularly true regarding the assessment of cognitive fusion. Indeed, there are relatively few valid and reliable assessment devices measuring cognitive fusion/defusion processes in adults, particularly those suffering with anxiety disorders. This line of work is critical not only to further advance theory regarding the role of putative ACT process variables in psychopathology, but also in terms of basic and applied empirical evaluations of ACT processes and their relation with clinical outcomes. The central aim of this paper, therefore,is to describe two studies focused on the development and validation of a new self-report measure ofcognitive fusion and its intervention counterpart defusion.
Cognitive Fusion and Defusion Processes in ACT
Cognitive fusion and defusion are esoteric terms within ACT, but both refer to basic universal processes entailed in various forms of symbolic communication. For instance, language and cognition enable meaningful verbal communication, in part, because words stand forspecificevents orexperiencesin particular contexts. For instance, the utterance “that’s a piece of cake” can mean different things depending on an individual’s social-learning history, and whether he/sheisat a birthday party or in a situation where he/sheisgiven a task to do. Most people would not mistake words for the events they describe by trying to eat the words “piece of cake.” However, given the right history, individuals may be able to experience all of the stimulus properties and emotional responses that are linked with having a piece of cake or doing a task with ease.
When we speak of cognitive fusion within ACT, we are speaking of a special instance of a verbal process wherein individuals fail to see a distinction between thoughts and the events and experiences that they seem to entail (Hayes et al., 1999; Eifert & Forsyth, 2005). Colloquially, cognitive fusion is a process wherein humans literally “buy into” and become “attached” to thoughts (e.g., reasons, judgments, evaluations, predictions, stories about the past, the self, and the future), evaluations of feelings and somatic sensations (e.g., “anxiety is bad,”or “I can’t do anything until I get a handle on my anxiety”) even when doing so causes significant harm. In short, individuals fail to see the process of thinking itself as just that – historically situated thoughts that need not be listened to or believed in guiding behavioral choices and actions.
Though fusion is a basic property of language and cognition (e.g., it can be part of the experience of intense love, or immersion in a good book), and thus is neither good nor bad, it can become unhelpful when thoughts and feelings are mistaken for actual events or experiences they describe, and thus get in the way of other more useful forms of behavioral regulation (Hayes et al., 1999; Valvidias-Salas, Sheppard & Forsyth, 2010).This is especially true when the content is painful (e.g., “if I get in a plane, I might panic,” or “if I contact germs, I might become a vector for disease and contaminate my kids,” or “I am worthless”). In such cases, painful content (thoughts, feelings, imagery) will quite often exert enormous influence over behavior (e.g., avoidance, escape, struggle for control), and thus behaviors tend to be guided more by what the thoughts seem to entail (e.g., living a life to avoid contact with germs, or avoiding the possibility of sadness, loss) than what the environmentmay offer. Much like a lion entering a room, fusion with painful content grabs attention, the world washes away, and other possibilities are foreclosed. From an ACT perspective, the darker side of fusion leads tosituations where taking thoughts literally narrows behavioral repertoires, organizes behavior in unhelpful ways, diminishes contact with what situations may afford, and thus gets in the way of taking effective action in the service of valued ends (Blackledge, 2007; Blackledge & Hayes, 2001; Eifert & Forsyth, 2005; Luoma, Hayes & Walser, 2007;Valvidia-Salas, Sheppard & Forsyth, 2010).
This can be seen in several forms of psychopathology (e.g., psychosis, anorexia, depression; Bach & Hayes, 2002; Hayes & Pankey, 2002; Gaudiano & Herbert, 2006b; Zettle & Hayes, 1986), including the anxiety disorders (Eifert & Forsyth, 2005; Eifert et al. 2009; Twohig, Masuda, Varra & Hayes, 2005; Twohig, Hayes & Masuda, 2006). For instance, fusion with the thought “I will go crazy when I panic” may result in individuals with panic disorder behaving in specific and often unhelpful ways(e.g., avoiding cues and contexts where panic is likely to occur, suppressing unwanted thoughts, excessive alcohol use). What is often missed here is that the thought “I will go crazy when I panic” is made up of the same stuff (i.e., words, letters) as the thought “I am a banana,” or “I am hungry.” The central ACT question is not whether the thoughts are good or bad, but rather whether they are useful in guiding effective action in the service of valued ends.
In general, when painful private content (e.g., thoughts, feelings) is mistaken for reality itself, human beings willstruggle and act to escape, suppress, or otherwise engage in avoidant behaviors. This avoidance and struggle is often self-limiting, tends not to work in the long run (Wegner, Schneider, Carter & White, 1987; Wegner & Zanakos, 1994) and takes time, energy, and effort that often gets in the way of doing what matters (Amstatder, 2008; Eifert & Forsyth, 2005; Gross, 1998; Gross, 2002). These consequences of fusion points to its toxic nature, and how it relates with other ACT processes such as experiential avoidance (i.e., unwillingness to experience unpleasant private events and efforts to avoid, modify, or control their frequency, form, or situational occurrence; Chawla and Ostafin, 2007; Hayes & Gifford, 1997; Hayes, Strosahl et al., 2004), loss of contact with the present, and lack of value-guided action. Thus, contexts that support cognitive fusion, and similar processes (e.g., experiential avoidance) foster excessive struggle with private content and diminished contact with natural contingencies supporting more flexible, goal-directed actions.
Using a variety of metaphorical and non-literal experiential exercises, ACT teaches cognitive defusion skills in an attempt to undermine cognitive fusionwithout necessarily changing the form of thought directly. The aim is to help clients discriminate between thoughts and the actual events they describe, with the goal of noticing when cognitions are helpful or not relative to actions that would support valued directions (i.e., domains that people tend to associate with quality of life; Blackledge, 2007). For instance, to undermine literality linked with the thought “I may be dying,” a client may learn a defusion strategy that involves describing the experience beginning with the words “I am having the thought” that I may be dying (Eifert & Forsyth, 2005; Luoma et al., 2007), or quickly repeating the word “dying” out loud for 30-40 seconds while noticing how the word’s properties shift into a stream of nonsensical sounds (Masuda, Hayes, Sackett, & Twohig, 2004; Masuda et al., 2009). Ultimately, defusion breaks down the illusion that we are what we think, and gives people greater perspective to notice that what we think need not dictate what we do.
To date, an emerging body of research supports the importance of cognitive fusion/defusion processes in the context of understanding and addressing a variety of forms of human suffering in both healthy (Healy et al., 2008; Hinton & Gaynor, 2010; Masuda et al., 2004; Masuda et al., 2009; Masuda, Feinstein, Wendell, & Sheehan, 2010;Takahashi, Muto, Tada & Sugiyama, 2002) and clinical samples (e.g., depression: Zettle & Hayes, 1986; Zettle & Rains, 1989; tinnitus distress: Hesser, Westin, Hayes, & Andersson, 2009; psychosis: Bach & Hayes, 2002; Gaudiano & Herbert, 2006a; 2006b; pain tolerance:Johnston et al., 2010; McCracken, 2006;eating disorders: Hayes & Pankey, 2002 and substance dependence: Twohig, Schoenberg & Hayes, 2007). Cognitive fusion/defusion processeshave also been emphasized in conceptualizing and alleviating anxious suffering (e.g., PTSD: Twohig, 2009; OCD: Twohig, Hayes, & Masuda, 2006a; Twohig et al., 2010, trichotillomania and skin picking: Twohig & Woods, 2004; Twohig, Hayes & Masuda, 2006b, respectively; social anxiety: Dalrymple & Herbert, 2007, and math anxiety: Zettle, 2003).
Though ACT and other acceptance-based approachesthat incorporate cognitive defusion processes have been found to be effective for a wide range of psychological/emotional difficulties (c.f., Forman, Herbert, Moitra, Yeomans, & Geller, 2007; Hayes et al. 2006; Powers, Vording & Emmelkamp, 2009; Pull, 2008; Ruiz, 2010; Twohig, Masuda, Varra, & Hayes, 2005), there are relatively few validated fusion/defusion assessment instruments, and the ones that do exist have been developed to suit idiographic research purposes. For instance, the Automatic Thoughts Questionnaire (ATQ; Hollon & Kendall, 1980) is a measure of depressogenic thought frequency and is widely used in depression studies to evaluate the impact of cognitive therapy. The ATQ was subsequently modified to evaluate ACT defusion processes, or the believability of depressogenic cognitions, but that modification was not subjected to formal psychometric evaluation (see ATQ-B; Zettle & Hayes, 1986). Similarly, the Stigmatizing Attitudes Believability Scale (SAB, Hayes, Bissett et al., 2004) was developed to measure the believability of stigmatizing thoughts that substance abuse therapists’ hold about their clients (see also Gifford et al., 2004, for a modification of the SAB for use within a smoking cessation program).
Within the domain of pain-related difficulties, the Psychological Inflexibility in Pain Scale (PIPS; Wicksell et al., 2008) was created to measure avoidance and cognitive fusion in a study focused on pain related disability. Other studies examining the utility of cognitive defusion strategies have used believability scales that have tended to be somewhat brief and idiosyncratic, typically measuring believability on a Likert-type scale from 0-100 (for examples see Bach & Hayes, 2002; Masuda et al., 2004; Masuda et al., 2009; Masuda et al., 2010). Interestingly, there is only one measure of fusion and defusion processes that has been subjected to rigorous psychometric evaluation in children, namely the Avoidance and Fusion Questionnaire for Youth (AFQ-Y; Greco, Lambert & Baer, 2008). The AFQ-Y was also recently adapted and psychometrically validated for use with adults (Schmalz & Murrell, 2010).
Despite the utility of existing measures of cognitive fusion, there appears to be relatively few valid and reliable measures of cognitive fusion in the area of anxiety disorders. This is unfortunate because many current outcome studies have found that ACT produces rapid and significant reductions in the believability of unwanted thoughts and feelings, and such decreases may be the strongest indicator of ACT outcomes (see Hayes et al. 2006). Given the interest in assessing ACT's utility as a treatment for anxiety problems (e.g., Dalrymple & Herbert, 2007; Twohig et al., 2006a; 2006b) - and more focally in understanding the processes through which ACT may have its effects - it seems important to develop a valid and reliable measure of the tendency to believe in anxiety-related thoughts and feelings.
The aim of the present set of studies is to report the development and validation of a new measure of cognitive fusion, the Believability of Anxious Feelings and Thoughts (BAFT; Forsyth & Eifert, 2008) questionnaire. This article describes two studies. In Study 1, we describe theinitial item development and preliminary psychometric properties of the BAFT in a healthy undergraduate sample (N = 437). Specifically, we conducted an exploratory factor analysis of the BAFT's factor structure,evaluated its internal consistency, and examined its relationship with other measures, including its convergent and divergent validity with relevant constructs (i.e., anxiety-related outcomes, and keyprocesses in the development, maintenance, and treatment of anxiety disordersfrom the perspective of acceptance and mindfulness-based approaches).
In Study 2, we describe findings from a BAFT cross-validation study using a highly anxious international community sample (N = 503), and further examine its psychometric properties. Specifically, we evaluated the BAFT's factor structure via a confirmatory factor analysis, and assessed its convergent and divergent validity, test-retest reliability, discriminant validity and responsiveness to change in an ACT self-help treatment context.
2.0 Study 1
2.1Method
2.1.1.Initial item generation
The BAFT was originally developed to assess the tendency to fuse with anxious thoughts and feelings in a sample of anxiety sufferers enrolled in a clinical trial for anxiety disorders. Item development was undertaken by two internationally recognized doctoral-level experts in ACT and the anxiety disorders in consultation with a team of advanced graduate students with training and expertise in anxiety disorders, ACT, and more traditional cognitive-behavior therapy. This effort was guided by ACT’s evidence-based model of human suffering and its successful alleviation. An initial pool of 23 rationally derived items was developed so as to be broadly applicable to heterogeneous anxiety-related concerns. The initial item set was then reviewed by the project team with specific attention to item clarity, overlap with other anxiety-related measures (e.g., the Anxiety Sensitivity Index), and consistency with ACT notions of fusion and defusion. Rather than assessing the presence, intensity, or degree of fear symptoms, the BAFT requires participants to indicate how much they believe in each statement on a 7-point Likert scale (1 = not at all believable to 7 = completely believable). In study 1, we used exploratory factor analysis to evaluate the initial item pool, and the outcome of that analysis was used to remove problematic items (see results section 2.2.1).
2.1.2. Participants and procedure
Four hundred thirty-seven healthy undergraduates (52% female, Mage= 19, SD =4.27)from the University at Albany, SUNY participated in this study. The ethnicity of the undergraduates was as follows: 74% Caucasian, 9% African American, 7% Asian, 7% Hispanic, and 3% other. Individuals completed a computerized battery of questionnaires while supervised by a research assistant and were offered course credit in return for participation.
2.1.3. Measures
Participants completed an assessment battery that included the original 23-Item Believability of Anxious Feelings and Thoughts (BAFT) Questionnaire (Eifert & Forsyth, 2008), along with the following theoretically relevant measures (see Table 4 for internal consistency coefficients that were calculated for each measure based on the current sample and depicted in parenthesis on the diagonal):
The Anxiety Sensitivity Index (ASI; Peterson & Reiss, 1993; Reiss, Peterson, Gursky, & McNally, 1986) is a 16-item questionnaire designed to assess for fear of anxiety-related symptoms. Each item is rated on a 5-point Likertscale (0 = very little to 4 = very much). ASI scorespossess excellent internal consistencyboth in previous work (α = .82 to .91; Peterson & Reiss, 1993) and the present study (see Table 4), stable test-retest reliability over a three-year period (r = .71 to .75; Maller & Reiss, 1992), and predict several panic-related variables such as response to panicogenic biological challenge procedures (Holloway & McNally, 1987; Schmidt, Lerew & Jackson, 1997). The ASI appears to have a single higher-order and three lower-order factors, representing physical, social and mental incapacitation concerns (Zinbarg, Barlow, & Brown, 1997). Given our interest in the overarching construct of anxiety sensitivity, we limited our analyses to the total score.