Social Anxiety

Examining Positive and Negative Affect as Outcomes and Moderators of Cognitive-Behavioral Therapy and Acceptance and

Commitment Therapy for Social Anxiety Disorder

Amy R. Sewart⁎

University of California, Los Angeles

Andrea N. Niles San Francisco VA Medical Center

Lisa J. Burklund University of California, Los Angeles

Darby E. Saxbe University of Southern California

Matthew D. Lieberman Michelle G. Craske

University of California, Los Angeles

Traditional cognitive-behavioral therapy (CBT) for anxiety disorders has been designed to target reductions in negative affect (NA) associated with defense-related processes. However, a subset of anxiety disorders, including social anxiety disorder (SAD), are also characterized by low positive affect (PA) resulting from separate deficits in appetitive-related processes. In contrast to CBT, “third- wave” approaches, such as acceptance and commitment therapy (ACT), align more consistently with motivational processes and, as a result, PA. However, the differential effect of CBT and ACT on PA and NA has yet to be investigated. Using secondary data from a randomized controlled trial, the present study sought to compare CBT’s (n = 45) and ACT’s (n = 35) effect on PA and NA in SAD. Findings were compared to a wait-list (WL) control

condition (n = 31), as well as normative data from a general adult sample. Baseline PA and NA were also examined as moderators and predictors of theory-relevant treatment outcomes. NA decreased significantly in both CBT and ACT from pre to posttreatment. Although ACT out- performed WL in reducing NA, this effect was not observed for CBT. PA increased significantly in both CBT and ACT from pre to posttreatment, with neither ACT nor CBT outperforming WL in increasing PA. Neither PA nor NA were found to moderate theoretically relevant treatment outcomes. Findings suggest that ACT and CBT share common treatment mechanisms, making them more similar than distinct. Further efforts should be focused on optimiz- ing CBT’s and ACT’s influence on threat and reward learning, and elucidating common processes of change.

Keywords: social anxiety; cognitive-behavioral therapy; acceptance and commitment therapy; positive affect; negative affect

ANXIETY DISORDERS HAVE BEEN CHARACTERIZED by behavioral inhibition, or the overactivation of a proposed defensive system that serves to motivate

Available online at www.sciencedirect.com

ScienceDirect Behavior Therapy 56 (2019) 1112 1124 \

www.elsevier.com/locate/bt

⁎ Address correspondence to Amy R. Sewart, M.A., Department of Psychology, University of California, Los Angeles, 1285 Franz Hall, Los Angeles, CA 90095; e-mail: [email protected]..

0005-7894/© 2019 Association for Behavioral and Cognitive Therapies. Published by Elsevier Ltd. All rights reserved.

–

avoidance of perceived threats and aversive out- comes (Gray, 1994). The defensive system is posited to generate certain forms of negative affect (NA) that are a central feature of anxiety disorders, such as fear and disgust (Davidson, 1994). In alignment with this defensive system, “first-” and “second- wave” cognitive-behavioral approaches for anxiety disorders almost exclusively focused on achieving reductions in NA—including fear and anxiety— and associated functional impairment (Barlow, Ellard, Sauer-Zavala, Bullis, & Carl, 2014; Carl, Soskin, Kerns, & Barlow, 2013; Craske, Meuret, Ritz, Treanor, & Dour, 2016). For example, exposure therapy works to lessen NA through a mechanism principal to the defensive system: fear extinction (Craske, Treanor, Conway, Zbozinek, & Vervliet, 2014). Indeed, these strategies have consistently demonstrated effectiveness in reducing NA and related symptomatology (e.g., Craske et al., in press). These traditional cognitive-behavioral ap-

proaches (referred to as “CBT”), however, have largely disregarded positive affect (PA)—one’s pleasurable engagement with the environment that results in positive emotional states (e.g., joy, interest, and enthusiasm)—an affective dimension that is generated by a separate motivational system related to appetitive and goal-directed behaviors (Davidson, 1994, 1998). Deficits in PA are predic- tive and characteristic of certain anxiety disorders (Khazanov & Ruscio, 2016; Kotov, Gamez, Schmidt, & Watson, 2010), as well as depression (e.g., Mineka, Watson, & Clark, 1998). CBTs have not been found to significantly increase PA for individuals with depression (g = .37, 95% CI [– 0.004, 0.77], p b .07; Boumparis, Karyotaki, Kleiboer, Hofmann, & Cuijpers, 2016). Conceiv- ably, CBT may also be insufficient for PA in anxious individuals. Thus far, the effect of CBT on PA for anxiety disorders has been examined in only one study through person-specific analyses on seven individuals with generalized anxiety disorder (Bosley, Fisher, & Taylor, 2018). PA was found to improve in only two of seven participants, and worsened in five of seven—even as NA improved over the course of treatment. Further investigation of CBT’s effect on PA for anxiety disorders with larger data sets is necessary. Low PA appears to be particularly characteristic of

social anxiety disorder (SAD; Watson & Naragon- Gainey, 2010). A meta-analysis examining cross- sectional and experience sampling studies found a moderate, inverse relationship between social anxi- ety and PA, r = –.36, 95% CI [–.31, –.40], that remained consistent after covarying for depressive symptoms (r = –.21; 95% CI [–.16, –.26]; Kashdan,

2007). These findings are paralleled with results harnessing longitudinal datasets, experimental designs, and structural equation modeling methods (Brown, Chorpita, & Barlow, 1998; Kashdan, Weeks, & Savostyanova, 2011; Khaza- nov & Ruscio, 2016; Sewart et al., 2019). Lack of explicit attention to the positive dimension of affect may in part explain less than optimal treatment response rates for CBT for SAD (see Loerinc et al., 2015). The recent “third wave” in CBT has shifted away

from directly decreasing NA and instead focused on acceptance and relinquishing control over mal- adaptive attempts to regulate internal mood states. The third-wave movement has ushered in modali- ties that emphasize previously unaddressed con- cepts, such as mindfulness of the present moment, acceptance of cognitions and feelings, and behav- iors consistent with personal values (Hayes & Hofmann, 2017). One of the most widespread of these approaches is acceptance and commitment therapy (ACT), which focuses on making contact with the present moment and changing or persisting in behavior to reach value-consistent goals, or “psychological flexibility” (Hayes, Luoma, Bond, Masuda, & Lillis, 2006). When compared with CBT, ACT treatment philosophy aligns more consistently with appetitive and goal-directed be- haviors. For example, the core ACT process of “committed action” requires engaging in activities that align with personal values, which may serve to recalibrate reward learning and increase positive affective states, such as pride, gratitude, and interest. Another core process of ACT, mindfulness, is defined by (a) “the self-regulation of attention so that it is maintained on immediate experience, thereby allowing for increased recognition of mental events in the present moment” and (b) “a particular orientation toward one’s experiences in the present moment, an orientation that is charac- terized by curiosity, openness, and acceptance” (Bishop et al., 2004; Fletcher & Hayes, 2005). While the goal of ACT is not to manipulate affective states, heightened awareness of the present moment may also serve to increase pleasurable engagement with the environment and PA. Given that CBT and ACT appear to target the

defensive and appetitive systems, respectively, it is likely these therapies also possess differential influence on PA and NA for SAD. Understanding how currently disseminated treatments impact affect and how baseline affect may predict treat- ment outcomes has high clinical utility, as such knowledge may be used to guide treatment planning. Furthermore, examining the effect of CBT and ACT on both PA and NA contributes to

sewart et al. 1113

dimensional understanding of mental disorders (Cuthbert, 2015). We first hypothesized that CBT would more

effectively decrease NA and symptoms when compared with ACT. Further, CBT is expected to be most effective for individuals with high NA, given that the defensive system is the primary target of CBT. Specifically, we hypothesized that reduc- tions in NA would be greatest for individuals with high NA receiving CBT compared to ACT. With ACT strategies appearing to more directly target an appetitive reward system, we hypothesized that ACT would more effectively raise PA when compared with CBT and that increases in PA would be greatest for individuals with low PA compared to CBT. To investigate these hypotheses, we utilized an

existing data set to examine the main effects of CBT and ACT on PA and NA. We compared these main effects to a wait-list (WL) control condition to better contextualize pre- to posttreatment gains. Levels of PA and NA for each treatment condition were then compared to a normative sample. Next, we examined whether pretreatment levels of PA and NA acted as moderators of treatment-relevant outcomes for CBT and ACT, symptom reduction, and quality of life, and resulted in differential treatment response.

Method PARTICIPANTS

Participants who met criteria of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV- TR; American Psychiatric Association, 2000) for either principal or co-principal social phobia, gener- alized type, were randomly assigned and stratified by age and gender to ACT, CBT, or WL. Using a modified intent-to-treat approach, all participants who began psychotherapeutic treatment and had available data on variables of interest for at least one time point (n = 111) were included in the present analyses (n = 35 ACT, n = 45 CBT, n = 31 WL). In addition to a diagnosis of generalized social

phobia, individuals were eligible for participation if they were (a) between 18 and 60 years of age, (b) either medication-free or stabilized on psychotropic medication (i.e., 1 month for benzodiazepines and beta blockers, and 3 months for selective serotonin reuptake inhibitors [SSRIs]/serotonin–norepineph- rine reuptake inhibitors [SNRIs] and heterocyclics), (c) psychotherapy-free or stabilized on alternative psychotherapies (not cognitive or behavioral ther- apies) that were not focused on their anxiety disorder for at least 6 months, and (d) English speaking. Exclusion criteria consisted of the fol- lowing: (a) active suicidal ideation; (b) severe

depression (Clinical Severity Rating [CSR] ≥ 6 on a 0- to 8-point scale; Anxiety Disorders Interview Schedule for DSM-IV [ADIS-IV]; Brown, Barlow, & Di Nardo, 1994); (c) history of a psychotic disorder, bipolar disorder, mental retardation, or organic brain damage; (d) substance abuse or dependence within the previous 6 months using DSM-IV-TR criteria (American Psychiatric Association, 2000); and (e) serious medical condi- tions or pregnancy. Given there was an additional neuroimaging component to the study, neuroimag- ing-relevant exclusion criteria were (a) left-handed- ness, (b) metal implants, and (c) claustrophobia (see Craske, Niles, et al., 2014, for further details on inclusion/exclusion criteria). Participants were re- cruited via the Internet, local flyers and newspaper advertisements, and community referrals. Study procedures took place at the Anxiety and Depres- sion Research Center located at the University of California, Los Angeles. The sample was 50% female, with a mean age of

28.3 years (SD = 6.7). The sample was primarily White (46.8%), with the remainder of the sample identifying as Asian/Pacific Islander (20.7%), His- panic/Latinx (18.0%), Black/African American (1.8%), and 12.7% as “other” race/ethnicity. At baseline, 20.0% of those receiving CBT (n = 9) and 22.9% of those receiving ACT (n = 8) reported being stabilized on psychotropic medication.

DESIGN

Random assignment, stratified by age and gender, was used to assign eligible participants to CBT, ACT, and WL. Participants were assessed (a) prior to treatment initiation (pretreatment [pre]), (b) imme- diately after the conclusion of treatment (posttreat- ment [post]), (c) 6 months after pretreatment assessment (6-month follow-up [6MFU]), and (d) 12 months after pretreatment assessment (l2-month follow-up [12MFU]). Assessments comprised self- report questionnaires, laboratory paradigms to assess various moderators and mechanisms of treatment change (see Craske, Niles, et al., 2014), and a semistructured diagnostic interview.

TREATMENTS

Participants received 12 weekly, 1-hour individual CBT or ACT therapy sessions by advanced clinical psychology doctoral student therapists. Further information regarding therapist competence, integ- rity, and training can be found in Craske, Niles, et al. (2014). Treatment protocols used in each condition were based on orientation-consistent manuals (e.g., Eifert & Forsyth, 2005). Given that both ACT and CBT focus to some extent on avoidance reduction, treatments were matched on

p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s1114

the number of sessions dedicated to exposure but provided a different, orientation-consistent ratio- nale. Following the conclusion of treatment, therapists provided follow-up booster phone calls to participants once per 6 months in an effort to reinforce progress consistent with therapy condi- tion. Most participants completed therapy within 12 to 16 weeks (range: 11–18).

Cognitive-Behavioral Therapy CBT treatment modules for generalized SAD were derived from standard CBT protocols (Hope, Heimberg, & Turk, 2010). Session 1 was focused on assessment, self-monitoring, and psychoeduca- tion. Sessions 2–4 emphasized cognitive restructur- ing—specifically, errors of overestimation and catastrophizing regarding negative evaluation— these techniques were combined with strategies of hypothesis testing, self-monitoring, and breathing retraining. Rationale for exposure to socially relevant feared cues was presented in Session 5. Exposure—including in vivo, in vivo augmented with interoceptive exposure, and imaginal—was conducted in Sessions 6–11. Session 12 was focused on relapse prevention.

Acceptance and Commitment Therapy ACT treatment modules for generalized social phobia were developed largely from Eifert and Forsyth (2005). Session 1 was focused on psychoe- ducation and orientation to ACT, which included experiential exercises (e.g., “finger trap”) and a discussion of acceptance and committed action. Sessions 2–3 centered on creative hopelessness and how attempts to control anxiety resulted in a reduction of valued life activities; accepting anxiety, rather than efforts to reduce, was encouraged. Sessions 4–5 directed participants in practices of mindfulness and cognitive defusion, while continu- ing to emphasize acceptance of anxiety. Sessions 6– 11 largely focused on values identification and acceptance, with an additional component focusing on continued practice of acceptance, mindfulness, and defusion. When relevant, exposure—including in vivo, in vivo augmented with interoceptive exposure, and imaginal—was conducted in an effort to have participants practice mindfully observing and accepting anxiety while conducting values-consistent action. Session 12 was focused on what practices were helpful and how to continue moving forward after therapy has concluded.

OUTCOME AND MODERATOR/PREDICTOR VARIABLES

Given that CBT for SAD aims to reduce anxiety- related symptoms, social anxiety symptoms were used as the theory-relevant outcome measure for CBT when examining PA and NA as moderators/

predictors. In contrast to CBT, ACT aims to increase the capacity to experience the present moment (i.e., “psychological flexibility”) and foster a “life worth living.” Thus, quality of life (QOL) was used as the theory-relevant outcome measure for ACT when examining PA and NA as modera- tors/predictors, as QOL captures the importance of, as well as satisfaction with, different life domains.

PA and NA In an effort to create a more valid and reliable index of self-reported PA and NA, scores from two well- validated self-report measures were combined following Z score transformation. PA was mea- sured by the (a) positive affect subscale (PANAS-P; 10-item) of the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988), a measure of dispositional affect in which participants indicate the extent to which they felt like a given affect-related adjective (e.g., “interest- ed,” “distressed”) over the past week using a scale from 1(very slightly/not at all) to 5 (extremely); and (b) high positive emotion items from the anhedonic depression subscale (MASQ-AD; 14-item; Kendall et al., 2016) of the Mood and Anxiety Symptom Questionnaire (MASQ; Watson & Clark, 1991) in which participants are asked to rate the strength of emotional experiences over the past week from 1 (not at all) to 5 (extremely). NA was measured by the (a) negative affect subscale (PANAS-N; 10-item) of the PANAS and (b) the general distress mixed subscale (MASQ-GDM; 15-item) of the MASQ. The PANAS and MASQ have been used frequently in socially anxious samples (see Kashdan, 2007, for a review) and across disorders to assess affective abnormalities (e.g., Brown et al., 1998; Sewart et al., 2019; Watson & Naragon-Gainey, 2010). These measures have also demonstrated both validity and good-to-excellent reliability in clinical and nonclinical populations (Crawford & Henry, 2004; Jolly, Dyck, Kramer, & Wherry, 1994; Keogh & Reidy, 2000; Mehrabian, 1997; Roesch, 1998; Watson, Clark, & Tellegen, 1988).

Social Anxiety Symptoms Three well-validated self-report measures of social anxiety symptoms were also combined following Z score transformation to create a more reliable measurement of disorder-related symptomatology. These self-report measures consisted of the follow- ing: (a) the Liebowitz Social Anxiety Scale—Self- Report (LSAS-SR; Fresco et al., 2001), a 24-item measure designed to assess fear and avoidance in situations involving performance and social inter- action of which each item rated on a 0 (no fear/ never avoid) to 3 (severe fear/usually avoid) Likert scale; (b) the Social Interaction Anxiety Scale (SIAS;

sewart et al. 1115

Mattick & Clarke, 1998), a 20-item measure designed to assess cognitive, affective, or behavioral responses across various socially relevant situations of which each item is rated on a 0 (not at all characteristic or true of me) to 4 (extremely characteristic or true of me) Likert scale; (c) the Social Phobia Scale (SPS; Mattick & Clarke, 1998), a 20-item measure designed to assess anxiety related to observation by others of which each item is rated on a 0 (not at all characteristic or true of me) to 4 (extremely characteristic or true of me) Likert scale.

Quality of Life The Quality of Life Inventory (QOLI; Frisch, 1994) was used to assess QOL, an outcome that corresponds with the overall goals of psychological flexibility and valued living within ACT. The QOLI assesses well-being and satisfaction with life across 16 domains (e.g., goals and values, friendships, health). This measure has demonstrated good internal validity and test–retest reliability (Frisch et al., 2005).

STATISTICAL ANALYSES

Analyses were performed in Stata 14.0 using the xtmixed command. We used multilevel modeling (MLM), a superior statistical approach for data with nested sources of variability, as it allows for examination of within- and between-subject change across time (pre, post, 6MFU, 12MFU) and by group (ACT, CBT, and WL). MLM accommodates unequal observations across individuals, allowing for the inclusion of participants missing one or more assessments.

Main Effects For the main effect of treatment on PA and NA, time was modeled at Level 1 (pre, post, 6MFU, 12MFU) using a piecewise approach (e.g., Roy- Byrne et al., 2005) that examines two separate linear segments of change: Pre and post (S1) and post, 6MFU, and 12MFU (S2). This approach captures typical trends in treatment studies, where- in the greatest effect of treatment occurs by the end of treatment and is then subsequently maintained, or lessened, during follow-up assessments. Treat- ment group (CBT, ACT, and WL) was included at Level 2. All models included random effects of the intercept. Between-group differences were assessed via marginal means at each time point, and by comparing S1 in ACT, CBT, and WL, and S2 in ACT and CBT. The MLM equation for a model with random effects for the intercept can be referenced in Craske, Niles, et al. (2014).

Moderated Effects To examine the moderating effects of baseline PA and NA on theory-relevant outcomes for ACT and CBT, time was modeled at Level 1 and condition and moderators at Level 2. Using a similar statistical approach as the main effects, we examined pretreat- ment to posttreatment and posttreatment to follow- up as two separate linear segments as described above. Separate moderator models were run for PA and NA. Omnibus three-way interaction effects were tested first (i.e., Time × Condition × Moderator). If moderation was not observed, we followed up with further testing to determine whether PA and NA functioned as a nonspecific predictor of treatment response to both CBT and ACT.

Table 1 Baseline Sample Characteristics and Bivariate Correlations

Correlations

CBT M(SD)

ACT M(SD)

WL M(SD)

F value p 1 2 3 4

1. Negative affect – PANAS 15.58 (7.60) 12.91 (8.18) 13.08 (6.91) 1.48 .23 MASQ 41.03 (9.89) 39.59 (12.69) 37.11 (11.20) 1.07 .35 Composite 0.18 (0.88) -0.05 (1.01) -0.17 (0.84) 1.39 .25

2. Positive affect -.117 – PANAS 16.43 (6.93) 15.85 (8.07) 15.10 (8.08) .26 .77 MASQ 34.39 (10.38) 35.21 (11.36) 32.15 (10.18) .70 .50 Composite 0.06 (0.82) 0.06 (0.99) -0.13 (0.93) .54 .59

3. Social anxiety symptom composite 0.01 (0.70) 0.10 (0.86) -0.14 (0.92) .64 .53 .517** -.217* – 4. Quality of life 0.15 (1.80) 0.08 (2.43) -0.09 (1.48) .13 .88 -.392** .635** -.435** –

Note. CBT = cognitive-behavioral therapy; ACT = acceptance and commitment therapy; WL = wait-list; PANAS = Positive and Negative Affect Schedule; MASQ = Mood Anxiety Symptom Questionnaire; Composite = composite of PANAS and MASQ; Social anxiety symptom composite = composite of Liebowitz Social Anxiety Scale—Self-Report, Social Interaction Anxiety Scale, and Social Phobia Scale; Quality of life = Quality of Life Inventory. **p N .01, *p N .05.

p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s1116

Results Means and standard deviations of PA and NA, along with theory-relevant outcomes examined in moderator analyses are presented in Table 1. No significant pretreatment group differences were found between CBT, ACT, and WL for variables of interest (ps N .23).

PA AND NA AS OUTCOME VARIABLES

Changes in PA and NA were examined by determining whether each slope differed signifi- cantly from zero from pre- to posttreatment (S1) for CBT, ACT, and WL and from 6MFU to 12MFU (S2) for CBT and ACT. Degree of slope change in these segments was then compared between groups. For Group × Time interactions, we present beta values for pairwise slope com- parisons (i.e., simple effects) and associated p values. Estimated means and 95% CI from the piecewise model for each group are displayed in

Table 2. Estimated slopes and effect sizes for pairwise group slope comparisons are further detailed in Table 3. Results are visually displayed in Figure 1.

Negative Affect From pre- to posttreatment (S1), NA decreased significantly for both CBT (b = –0.5, 95% CI [–0.8, –0.3], p b. 001) and ACT (b = –0.7, 95% CI [–0.9, – 0.3], p b. 001), but not WL (b = –0.2, 95% CI [–0.5, –0.1], p = .18). For S1 pairwise slope comparisons, ACT demonstrated larger reductions in NA than WL (b = –0.5, 95% CI [–0.9, –0.1], p = .01). CBT, however, did not outperform WL in reducing NA, although the observed effect approached signifi- cance (b = 0.3, 95% CI [–0.04, 0.7], p = .08). Inconsistent with our hypothesis, CBT and ACT S1 slopes did not significantly differ from each other (b = –0.2, 95% CI [–0.5, 0.2], p = .39). S2 change slopes were not significant in CBT (p = .19) or ACT (p = .37).

Table 2 Estimated Means and Confidence Intervals at Each Time Point for Cognitive-Behavioral Therapy (CBT), Acceptance and Commitment Therapy (ACT), and Wait-List (WL)

CBT (95% CI)

ACT (95% CI)

WL (95% CI)

Negative affect Baseline 0.17 [-0.08, 0.41] -0.06 [-0.34, 0.22] -0.15 [-0.45, 0.14] Post -0.34 [-0.61, -0.07] -0.72 [-1.02, -0.42] -0.34 [-0.64, 0.03] 6MFU -0.44 [-0.69, -0.19] -0.65 [-0.92, -0.39] – 12MFU -0.55 [-0.85, -0.24] -0.59 [-0.90, -0.28] –

Positive affect Baseline 0.1 [-0.2, 0.3] 0.1 [-0.2, 0.4] -0.1 [-0.5, 0.2] Post 0.4 [0.1, 0.7] 0.3 [0.0, 0.7] 0.0 [-0.4, 0.3] 6MFU 0.4 [0.2, 0.7] 0.4 [0.1, 0.7] – 12MFU 0.5 [0.1, 0.8] 0.4 [0.0, 0.7] –

Note. CBT = cognitive-behavioral therapy; ACT = acceptance and commitment therapy; WL = wait-list; 6MFU = 6-month follow-up; 12MFU = 12-month follow-up.

Table 3 Estimated Slopes (Change) From Pre- to Posttreatment (S1) and Post to 12MFU (S2), and Effect Sizes (Cohen’s d) for Pairwise Group Slope Comparisons for CBT Versus ACT

CBT beta (95% CI)

ACT beta (95% CI)

WL beta (95% CI)

CBT vs. ACT d

ACT vs. WL d

CBT vs. WL d

Negative affect S1 -0.5 [-0.8, -0.3]*** -0.7 [-0.9, -0.3]*** -0.2 [-0.5, 0.1] 0.2 0.5 0.3 S2 -0.1 [-0.3, 0.1] -0.1 [-0.1, 0.2] – 0.0 – –

Positive affect S1 0.3 [0.1, 0.6]** 0.3 [0.01, 0.5]✝ 0.1 [-0.1, 0.4] 0.0 0.2 0.2 S2 0.04 [-0.1, 0.2] 0.01 [-0.1, 0.2] – 0.03 – –

Note. CBT = cognitive-behavioral therapy; ACT = acceptance and commitment therapy; WL = wait-list. ***p N .001, **p N .01, *p N .05, ✝p = .053.

sewart et al. 1117

Positive Affect From pre- to posttreatment (S1), PA increased significantly for CBT (b = 0.3, 95% CI [0.1, 0.6], p = .009) and ACT (b = 0.3, 95% CI [0.01, 0.5], p = .05), but not WL (b = 0.1, 95% CI [–0.1, 0.4], p = .37). For S1 pairwise slope comparisons, neither ACT nor CBT demonstrated larger increases in PA than WL (ACT b = 0.1, 95% CI [–0.2, 0.5], p = .48; CBT b = –0.2, 95% CI [–0.6, 0.2], p = .26).

Inconsistent with our hypothesis, ACT and CBT S1 slopes did not significantly differ from each other (b = –0.01, 95% CI [–0.4, 0.3], p = .70). S2 change slopes were not significant in CBT (p = .62) or ACT (p = .88).

Nonclinical Sample Comparison To further evaluate the degree to which PA and NA compared to normative levels across treatment, we referenced PANAS normative data from a general

FIG. 1 Main effects of cognitive-behavioral therapy (CBT), acceptance and commitment therapy (ACT), and wait-list (WL) on (a) negative affect and (b) positive affect. Note. 6MFU = 6- month follow-up; 12MFU = 12-month follow-up.

p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s1118

adult sample recruited from various commercial and public service organizations (NA M = 16.00 SD = 5.90, PA M = 31.31, SD = 7.65; Crawford & Henry, 2004). We compared these nonclinical sample values to our sample’s PANAS scores for CBT and ACT treatment completers. At pretreatment, levels of NA were notably above

normative values as indicated by average values in the 90th and 84th percentile for CBT and ACT, respectively. At posttreatment, NA had reduced to the 78th percentile for CBT and 63rd percentile for ACT. Twelve months following treatment, NA of those who had received CBT were averaged at the 74th percentile, while those receiving ACT were at the 69th percentile. Thus, NA scores approached the population average at posttreatment and reductions were generally maintained through 12MFU for both CBT and ACT. PA was below normative values at pretreatment

as indicated by average values in the 24th and 28th percentile for CBT and ACT, respectively. At posttreatment, PA had increased to the 41st percentile for both CBT and ACT. Twelve months following treatment, PA of those who had those received CBT were averaged at the 41st percentile, while those receiving ACT were at the 36th percentile. Thus, PA scores approached the popu- lation average at posttreatment and improvements were generally maintained through 12MFU for both CBT and ACT.

PA AND NA AS MODERATOR VARIABLES

PA and NA were evaluated as moderators of self- reported composite symptoms of social anxiety and QOL for CBT and ACT. Omnibus three-way interaction effects were tested first (Time × Condi- tion × Moderator) for S1 and S2, and if not significant, we then tested PA and NA as nonspe- cific treatment predictors.

Social Anxiety Symptoms NA was not found to differentially moderate social anxiety symptoms for CBT or ACT for S1 (b = 0.4, 95% CI [–0.1, 0.8], p = .09) or S2 (b = 0.1, 95% CI [–0.2, 0.2], p = .85). Nor did PA moderate social anxiety symptoms for CBT or ACT for S1 (b = 0.04, 95% CI [–0.4, 0.5], p = .88) or S2 (b = 0.2, 95% CI [–0.1, 0.4], p = .30). Further testing revealed that regardless of treatment condition neither NA nor PA significantly predicted outcome for S1 (NA b = 0.1, 95% CI [–0.1, 0.3], p = .50; PA b = –0.1, 95% CI [–0.4, 0.1], p = .28) or S2 (NA b = 0.01, 95% CI [–0.1, 0.1], p = .88; PA b = –0.01, 95% CI [–0.1, 0.4], p = .27).

Quality of Life NA was not found to differentially moderate QOL for CBT or ACT for S1 (b = 0.1, 95% CI [–0.7, 0.8], p = .81) or S2 (b = –0.01, 95% CI [–0.5, 0.4], p = .96). Nor did PA moderate QOL for CBT or ACT for S1 (b = –0.2, 95% CI [–1.0, 0.5], p = .54) or S2 (b = 0.2, 95% CI [–0.2, 0.7], p = .28). Regardless of

FIG. 2 Nonspecific treatment prediction of quality of life by positive affect. Note. 6MFU = 6- month follow-up; 12MFU = 12-month follow-up; CBT = cognitive-behavioral therapy; ACT = acceptance and commitment therapy; WL = wait-list.

sewart et al. 1119

treatment condition, NA did not significantly predict outcome for S1 (b = –0.01, 95% CI [–0.4, 0.4], p = .97) or S2 (b = –0.1, 95% CI [–0.3, 0.1], p = .48). In contrast, as shown in Figure 2, PA predicted outcome for S1 (b = –0.6, 95% CI [–1.0, –0.2], p = .002), but not for S2 (b = 0.1, 95% CI [–0.2, 0.2], p = .99). Tests of simple effects revealed that among participants with low (1 SD below the mean) and average (mean) PA, the slope change in QOL was significant (low PA slope = 1.1, 95% CI [0.6, 1.6], p b .001; average PA slope = 0.5, 95% CI [0.3, 0.9], p b .001). Slope change was not found to be significant for individ- uals with high PA (1 SD above the mean; slope = 0.05, 95% CI [–0.4, 0.5], p = .82). However, high PA individuals already demonstrated high QOL at pretreatment.

Discussion Anxiety disorders have long been characterized by abnormal activation of a theorized defensive system that gives rise to elevated levels of NA (e.g., fear; Gray, 1994). As such, traditional CBT for anxiety disorders has been designed to target reductions in NA associated with defense-related mechanisms. However, a subset of anxiety disorders, including SAD, are also characterized by low PA resulting from separate deficits in appetitive learning pro- cesses (e.g., Brown et al., 1998). In contrast to CBT, “third-wave” approaches, such as ACT, have focused on new treatment targets that—although not a direct goal of these treatments—align more consistently with and facilitate appetitive and goal- directed behaviors and, as a result, PA. However, whether or not these treatment approaches actually exercise differential influence on appetitive and defensive systems has yet to be investigated. Thus, using secondary data from a previously conducted randomized controlled trial, the present study sought to compare CBT’s and ACT’s effect on PA and NA in SAD. The findings were compared to a WL control condition, as well as normative data from a general adult sample. In addition, we sought to investigate baseline PA and NA as moderators and predictors of theory-relevant treatment out- comes. The present study has several main findings,

some of which contrast with our initial hypotheses. First, NA decreased significantly in both CBT and ACT from pre- to posttreatment with these reductions maintained through follow-up assess- ments. However, contrary to our hypothesis, CBT was not found to outperform ACT in decreasing NA pre- to posttreatment. Moreover, whereas ACT outperformed WL in reducing NA, the difference in slopes between CBT and WL only approached significance (p = .08). Comparing posttreatment

NA to population average data, as reported in one reference study comprising an adult sample collect- ed through various commercial and public service organizations, levels of NA approached normative values for CBT and ACT. Second, PA increased significantly in both ACT and CBT from pre- to posttreatment with these elevations maintained through follow-up assessments and approaching normative values for CBT and ACT. Again, contrary to our hypothesis, ACT was not found to significantly outperform CBT in increasing PA pre- to posttreatment. Neither ACT nor CBT were found to outperform WL in elevating PA pre- to posttreatment. Third, pretreatment levels of neither PA nor NA were found to moderate theoretically relevant treatment outcomes for either CBT or ACT. In contrast with our hypothesized differential

moderation, pretreatment levels of PA functioned as a nonspecific predictor of QOL in both ACT and CBT. While individuals with low to average levels of pretreatment PA experienced significant im- provements in QOL, individuals with high pre- treatment PA did not. As displayed in Figure 2, this appears to be a ceiling effect due to individuals with high pretreatment PA reporting high levels of QOL prior to initiation of ACT and CBT, thus giving less potential for improvement. These results may be further explained by the “broaden-and-build” theory of positive emotion, wherein positive emo- tions enhance QOL through encouraging enduring personal resources and relationships and promot- ing exploratory and novel thoughts and actions— thus broadening one’s behavioral range (Fredrick- son, 1998, 2004). Overall, our PA predictor findings corroborate a larger body of literature that implicates PA as a significant, positive predic- tor of various QOL dimensions (e.g., Brennan, Singh, Spencer, & Roberts-Thomson, 2006; Headey, Kelley, & Wearing, 1993; Louro, Fernán- dez-Castro, & Blasco, 2015; Stauber et al., 2013) and highlights the importance of further identifying psychotherapeutic strategies that more effectively increase PA. One of the most notable findings is that although

CBT significantly decreased NA pre- to posttreat- ment, this effect was not found to differ from WL. There may be a number of explanations. One possibility is that the treatment strategies (e.g., cognitive restructuring, exposure, and breathing retraining) targeted specific threat responses (e.g., fears of specific stimuli) without influencing general negative emotions. Furthermore, while ACT and CBT were not

observed to exercise a differential effect on NA, ACT did outperform WL with a moderate effect

p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s1120

size (d = 0.5). Although reducing NA is not a focus of ACT, it is likely that this reduction is a by- product of ACT-based strategies. For example, carrying out committed action for an individual with SAD may include spending more time socializing with friends, which may serve to lessen negative emotions. Why ACT outperformed WL and CBT did not, is unclear. Based on these contradictory findings, further research is needed to examine changes in threat-related processes as a result of CBT and ACT for anxiety disorders. Another striking finding is that while ACT and

CBT significantly increased PA, neither treatment outperformed WL pre- to posttreatment. This converges with previous findings that cognitive- behavioral and mindfulness-based psychotherapies are largely ineffective in heightening PA for depression (Boumparis et al., 2016; Dichter et al., 2009; Moore et al., 2013). Our results should be interpreted with the acknowledgment that CBT for SAD is not designed to target appetitive and goal- directed learning in the same manner as CBT for depression. This is not to say that strategies in CBT for SAD (e.g., exposure) should not also serve to indirectly increase PA through encouraging contact with feared situations that are likely to also be rewarding (e.g., social gatherings). ACT, however, seeks to improve QOL—a construct inherently related to reward attainment—and yet it does not outperform WL in increasing PA. These findings call for clinical scientists to identify psychothera- peutic strategies that more effectively target the appetitive approach system and—as a result— increase PA for anxiety disorders that demonstrate deficits, such as SAD and generalized anxiety disorder (Prenoveau et al., 2010). As with NA, further research is needed to examine changes in in reward-related processes as a result of CBT and ACT for anxiety disorders. These results corroborate and extend our initial

outcome and moderator findings (Craske, Niles, et al., 2014). Comorbid depression previously failed to predict composite symptoms, which is consistent with our PA finding given that low PA is characteristic of depression (Watson, Clark, & Carey, 1988). Craske, Niles, et al. (2014) found extraversion to be a nonspecific significant linear predictor of composite symptoms (z = 1.96, p = .05), wherein participants higher in extraversion reported significantly fewer symptoms collapsed across follow-up assessments (Craske, Niles, et al., 2014). There is both self-report and psychophysi- ological evidence to suggest that PA forms the “conceptual core” of extraversion (Hermes, Hage- mann, Naumann, & Walter, 2011; Lucas & Baird, 2004; Watson & Clark, 1997). This finding

diverges from our current analyses that found that pretreatment PA was neither a moderator nor a nonspecific predictor of symptom reduction. One potential explanation for these differing findings is that extraversion is a socially relevant trait, whereas PA encompasses more pleasurable engagement broadly. The ability to derive positive emotional states from social experiences may be particularly beneficial for enhancing treatment response to both CBT and ACT for individuals with SAD. Future research should investigate the dynamics between PA and social interaction in SAD and how their relationship may affect treatment outcome. As the field moves more toward a dimensional

understanding of psychopathology, it is essential that we begin to examine our standard psychother- apeutic treatments from this perspective (Brown & Barlow, 2009; Insel et al., 2010). This is the first study to our knowledge that examines changes in PA and NA as a result of CBT and ACT for anxiety disorders. Our findings raise important questions regarding the effectiveness of these interventions with respect to defense and reward systems. Furthermore, our findings suggest that individuals with low PA may benefit from augmenting CBT and ACT with strategies specifically designed to regulate positive emotion. For example, individuals with low PA may benefit from adjunctive training in how to “savor” past, current, or future positive experiences (Bryant, 2003) or “capitalize” on success (Carl et al., 2013; Langston, 1994). Indeed, protocols with such strategies have recently been shown to significantly improve PA, although results were not compared to WL (e.g., Craske et al., in press). Our study possessed some limitations. Given that

these two therapeutic approaches focus on avoid- ance reduction, exposure was employed in both treatment conditions for an equal number of sessions with an orientation-consistent rationale. As a result, it is possible that exposure may be driving effects in both ACT and CBT, making the differential impact of these approaches less detect- able and more difficult to assess—or less present than suspected. Optimizing QOL is also a focus of CBT that has more recently been examined as a critical outcome of treatment success (e.g., Hofmann, Wu, & Boettcher, 2014). As a result, contending that QOL is not theoretically relevant to CBT as well is debatable. Given that PA and NA were not assessed throughout treatment, we were unable to examine these factors through media- tional analyses that may have provided further insight into the mechanistic nature of threat and reward-related processes in CBT and ACT. Partic- ipants were deemed ineligible if they presented with

1121sewart et al.

severe depression (CSR ≥ 6 on a 0- to 8-point scale; Brown et al., 1994). While this is not an uncommon procedure for anxiety psychotherapy studies, this likely excluded individuals with anhedonia and as a result may have indirectly affected our PA findings. To further examine treatment change, we com- pared levels of PA and NA to a nonclinical sample collected in the United Kingdom with a mean age (M = 42.9) and demographic composition that was substantially different from our sample. Thus, these normative values do not generalize to our sample in an ideal manner. Last, we relied on self-report measures as the exclusive modality of treatment response, which may have inflated ratings (Loerinc et al., 2015). However, our treatment’s main effects more often than not did not outperform WL. Future studies may address this limitation by employing experimental paradigms that more directly measure change in purported reward and threat-learning mechanisms during treatment. Overall, these findings suggest that while ACT

and CBT may approach psychotherapeutic treat- ment with different methods, they share common treatment mechanisms making them more similar than distinct. Still further research is needed to elucidate common processes of change in CBT and ACT. Future research should assess PA and NA during CBT and ACT for a more rigorous examination of these factors as potential mecha- nisms. Most importantly, efforts should be focused on enhancing CBT’s and ACT’s influence on threat and reward learning in an effort to better recali- brate PA and NA.

Conflict of Interest Statement The authors declare that there are no conflicts of interest.

References

American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. https://doi.org/10.1176/appi. books.9780890425596

Barlow, D. H., Ellard, K. K., Sauer-Zavala, S., Bullis, J. R., & Carl, J. R. (2014). The origins of neuroticism. Perspectives on Psychological Science, 9(5), 481–496. https://doi.org/10. 1177/1745691614544528

Bishop, S. R., Lau, M., Shapiro, S., Carlson, L., Anderson, N. D., Carmody, J. … Velting, D. (2004). Mindfulness: A proposed operational definition. Clinical Psychology: Science and Prac- tice, 11(3), 230–241. https://doi.org/10.1093/clipsy.bph077

Bosley, H. G., Fisher, A. J., & Taylor, C. B. (2018). Differential responses of positive affect, negative affect, and worry in CBT for generalized anxiety disorder: A person-specific analysis of symptom course during therapy. Psychotherapy Research, 28(4), 630–642. https://doi.org/10.1080/ 10503307.2016.1233366

Boumparis, N., Karyotaki, E., Kleiboer, A., Hofmann, S. G., & Cuijpers, P. (2016). The effect of psychotherapeutic

interventions on positive and negative affect in depression: A systematic review and meta-analysis. Journal of Affective Disorders, 202, 153–162. https://doi.org/10.1016/J.JAD. 2016.05.019

Brennan, D. S., Singh, K. A., Spencer, A. J., & Roberts- Thomson, K. F. (2006). Positive and negative affect and oral health-related quality of life. Health and Quality of Life Outcomes, 4(1), 83. https://doi.org/10.1186/1477-7525-4- 83

Brown, T. A., & Barlow, D. H. (2009). A proposal for a dimensional classification system based on the shared features of the DSM-IV anxiety and mood disorders: Implications for assessment and treatment. Psychological Assessment, 21(3), 256–271. https://doi.org/10.1037/ a0016608

Brown, T. A., Barlow, D. H., & Di Nardo, P. A. (1994). Anxiety Disorders Interview Schedule for DSM-IV (ADIS- IV): Client Interview Schedule. Boulder, CO: Graywind.

Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107(2), 179–192. https://doi.org/10. 1037/0021-843X.107.2.179

Bryant, F. (2003). Savoring Beliefs Inventory (SBI): A scale for measuring beliefs about savouring. Journal of Mental Health, 12(2), 175–196. https://doi.org/10.1080/ 0963823031000103489

Carl, J. R., Soskin, D. P., Kerns, C., & Barlow, D. H. (2013). Positive emotion regulation in emotional disorders: A theoretical review. Clinical Psychology Review, 33(3), 343–360. https://doi.org/10.1016/j.cpr.2013.01.003

Craske, M. G., Meuret, A. E., Ritz, T., Treanor, M., & Dour, H. J. (2016). Treatment for anhedonia: A neuroscience driven approach. Depression and Anxiety, 33(10), 927–938. https://doi.org/10.1002/da.22490

Craske, M. G., Meuret, A. E., Ritz, T., Treanor, M., Dour, H., & Rosenfield, D. (2019). Positive affect treatment for depression and anxiety: A randomized clinical trial for a core feature of anhedonia. Journal of Consulting and Clinical Psychology, 87(5), 457–471.

Craske, M. G., Niles, A. N., Burklund, L. J., Wolitzky-Taylor, K. B., Vilardaga, J. C. P., Arch, J. J. … Lieberman, M. D. (2014). Randomized controlled trial of cognitive behavioral therapy and acceptance and commitment therapy for social phobia: Outcomes and moderators. Journal of Consulting and Clinical Psychology, 82(6), 1034–1048. https://doi.org/ 10.1037/a0037212

Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. https://doi.org/10.1016/j.brat.2014. 04.006

Crawford, J. R., & Henry, J. D. (2004). The Positive and Negative Affect Schedule (PANAS): Construct validity, measurement properties and normative data in a large non-clinical sample. British Journal of Clinical Psychology, 4 3 ( 3 ) , 2 4 5 – 2 6 5 . h t t p s : / / d o i . o r g / 1 0 . 1 3 4 8 / 0144665031752934

Cuthbert, B. N. (2015). Research domain criteria: Toward future psychiatric nosologies.Dialogues in Clinical Neuro- science, 17(1), 89–97. Retrieved from , http://www.ncbi. nlm.nih.gov/pubmed/25987867

Davidson, R. J. (1994). Asymmetric brain function, affective style, and psychopathology: The role of early experience and plasticity. Development and Psychopathology, 6(4), 741–758. https://doi.org/10.1017/S0954579400004764

1122 p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s

Davidson, R. J. (1998). Affective style and affective disorders: Perspectives from affective neuroscience. Cognition and Emotion, 12(3), 307–330. https://doi.org/10.1080/ 026999398379628

Dichter, G. S., Felder, J. N., Petty, C., Bizzell, J., Ernst, M., & Smoski, M. J. (2009). The effects of psychotherapy on neural responses to rewards in major depression. Biological Psychiatry, 66(9), 886–897. https://doi.org/10.1016/j. biopsych.2009.06.021

Eifert, G. H., & Forsyth, J. P. (2005). Acceptance and commitment therapy for anxiety disorders: A practitioner’s treatment guide to using mindfulness, acceptance, and values-based behavior change. Oakland, CA: New Harbinger.

Fletcher, L., & Hayes, S. C. (2005). Relational frame theory, acceptance and commitment therapy, and a functional analytic definition of mindfulness. Journal of Rational- Emotive and Cognitive-Behavior Therapy, 23(4), 315–336. https://doi.org/10.1007/s10942-005-0017-7

Fredrickson, B. L. (1998). What good are positive emotions? Review of General Psychology, 2(3), 300–319. https://doi. org/10.1037/1089-2680.2.3.300

Fredrickson, B. L. (2004). The broaden-and-build theory of positive emotions. Philosophical Transactions of the Royal Society B: Biological Sciences, 359(1449), 1367–1378. https://doi.org/10.1098/rstb.2004.1512

Fresco, D. M., Coles, M. E., Heimberg, R. G., Liebowitz, M. R., Hami, S., Stein, M. B., & Goetz, D. (2001). The Liebowitz Social Anxiety Scale: A comparison of the psychometric properties of self-report and clinician-administered formats. Psychological Medicine, 31(6), 1025–1035. https://doi.org/ 10.1017/S0033291701004056

Frisch, M. B. (1994). Quality of Life Inventory (QOLI). Journal of Consulting and Clinical Psychology, 4, 92–101. https:// doi.org/10.1007/978-94-007-0753-5

Frisch, M. B., Clark, M. P., Rouse, S. V., Rudd, M. D., Paweleck, J. K., Greenstone, A., & Kopplin, D. A. (2005). Predictive and treatment validity of life satisfaction and the Quality of Life Inventory. Assessment, 12(1), 66–78. https:// doi.org/10.1177/1073191104268006

Gray, J. A. (1994). Framework for a taxonomy of psychiatric disorder. In S. H. M. van Goozen, N. E. Van de Poll, J. A. Sergeant, & J. A. (Eds.), Emotions: Essays on emotion theory (pp. 29–59). New York, NY: Psychology Press.

Hayes, S. C., & Hofmann, S. G. (2017). The third wave of cognitive behavioral therapy and the rise of process-based care. World Psychiatry, 16(3), 245–246. https://doi.org/10. 1002/wps.20442

Hayes, S. C., Luoma, J. B., Bond, F. W., Masuda, A., & Lillis, J. (2006). Acceptance and commitment therapy: Model, processes and outcomes. Behaviour Research and Therapy, 44(1), 1–25. https://doi.org/10.1016/J.BRAT.2005.06.006

Headey, B., Kelley, J., & Wearing, A. (1993). Dimensions of mental health: Life satisfaction, positive affect, anxiety and depression. Social Indicators Research, 29(1), 63–82. https://doi.org/10.1007/BF01136197

Hermes, M., Hagemann, D., Naumann, E., & Walter, C. (2011). Extraversion and its positive emotional core— further evidence from neuroscience. Emotion, 11(2), 367–378. https://doi.org/10.1037/a0021550

Hofmann, S. G., Wu, J. Q., & Boettcher, H. (2014). Effect of cognitive-behavioral therapy for anxiety disorders on quality of life: A meta-analysis. Journal of Consulting and Clinical Psychology, 82(3), 375. https://doi.org/10.1037/ a0035491

Hope, D. A., Heimberg, R. G., & Turk, C. L. (2010). Managing social anxiety: A cognitive-behavioral therapy approach

(2nd ed.). New York, NY: Oxford University Press. https:// doi.org/10.1093/med:psych/9780195336689.001.0001

Insel, T., Cuthbert, B., Garvey, M., Heinssen, R., Pine, D. S., Quinn, K. … Wang, P. (2010). Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychi- atry, 167(7), 748–751. https://doi.org/10.1176/appi.ajp. 2010.09091379

Jolly, J. B., Dyck, M. J., Kramer, T. A., & Wherry, J. N. (1994). Integration of positive and negative affectivity and cognitive content-specificity: Improved discrimination of anxious and depressive symptoms. Journal of Abnormal Psychology, 103 (3), 544–552. https://doi.org/10.1037/0021-843X.103.3. 544

Kashdan, T. B. (2007). Social anxiety spectrum and diminished positive experiences: Theoretical synthesis and meta-analy- sis. Clinical Psychology Review, 27(3), 348–365. https://doi. org/10.1016/j.cpr.2006.12.003

Kashdan, T. B., Weeks, J. W., & Savostyanova, A. A. (2011). Whether, how, and when social anxiety shapes positive experiences and events: A self-regulatory framework and treatment implications. Clinical Psychology Review, 31(5), 786–799. https://doi.org/10.1016/J.CPR.2011.03.012

Kendall, A. D., Zinbarg, R. E., Bobova, L., Mineka, S., Revelle, W., Prenoveau, J. M., & Craske, M. G. (2016). Measuring positive emotion with the Mood and Anxiety Symptom Questionnaire: Psychometric properties of the Anhedonic Depression Scale. Assessment, 23(1), 86–95. https://doi.org/ 10.1177/1073191115569528

Keogh, E., & Reidy, J. (2000). Exploring the factor structure of the Mood and Anxiety Symptom Questionnaire (MASQ). Journal of Personality Assessment, 74(1), 106–125. https:// doi.org/10.1207/S15327752JPA740108

Khazanov, G. K., & Ruscio, A. M. (2016). Is low positive emotionality a specific risk factor for depression? A meta- analysis of longitudinal studies. Psychological Bulletin, 142 (9), 991–1015. https://doi.org/10.1037/bul0000059

Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking “big” personality traits to anxiety, depressive, and substance use disorders: A meta-analysis. Psychological Bulletin, 136(5), 768–821. https://doi.org/10.1037/ a0020327

Langston, C. A. (1994). Capitalizing on and coping with daily- life events: Expressive responses to positive events. Journal of Personality and Social Psychology, 67(6), 1112–1125. https://doi.org/10.1037/0022-3514.67.6.1112

Loerinc, A. G., Meuret, A. E., Twohig, M. P., Rosenfield, D., Bluett, E. J., & Craske, M. G. (2015). Response rates for CBT for anxiety disorders: Need for standardized criteria. Clinical Psychology Review, 42, 72–82. https://doi.org/10. 1016/j.cpr.2015.08.004

Louro, A. C., Fernández-Castro, J., & Blasco, T. (2015). Is there a relationship between positive affect and quality of life in colorectal cancer patients? Annals of Psychology, 31(2), 404–413. https://doi.org/10.6018/analesps.31.2.160331

Lucas, R. E., & Baird, B. M. (2004). Extraversion and emotional reactivity. Journal of Personality and Social Psychology, 86(3), 473. https://doi.org/10.1037/0022- 3514.86.3.473

Mattick, R. P., & Clarke, J. C. (1998). Development and validation of measures of social phobia scrutiny fear and social interaction anxiety. Behaviour Research and Therapy, 36(4), 455–470. https://doi.org/10.1016/S0005-7967(97) 10031-6

Mehrabian, A. (1997). Comparison of the PAD and PANAS as models for describing emotions and for differentiating anxiety from depression. Journal of Psychopathology and

1123sewart et al.

Behavioral Assessment, 19(4), 331–357. https://doi.org/10. 1007/BF02229025

Mineka, S., Watson, D., & Clark, L. A. (1998). Comorbidity of anxiety and unipolar mood disorders. Annual Review of Psychology, 49(1), 377–412. https://doi.org/10.1146/ annurev.psych.49.1.377

Moore, R. C., Chattillion, E. A., Ceglowski, J., Ho, J., von Känel, R., Mills, P. J. … Mausbach, B. T. (2013). A randomized clinical trial of behavioral activation (BA) therapy for improving psychological and physical health in dementia caregivers: Results of the Pleasant Events Program (PEP). Behaviour Research and Therapy, 51(10), 623–632. https://doi.org/10.1016/j.brat.2013.07.005

Prenoveau, J. M., Zinbarg, R. E., Craske, M. G., Mineka, S., Griffith, J. W., & Epstein, A. M. (2010). Testing a hierarchical model of anxiety and depression in adolescents: A tri-level model. Journal of Anxiety Disorders, 24(3), 334–344. https://doi.org/10.1016/j.janxdis.2010.01.006

Roesch, S. C. (1998). The factorial validity of trait positive affect score: Confirmatory factor analyses of unidimensional and multidimensional models. Educational and Psycholog- ical Measurement, 58(3), 451–466. https://doi.org/10.1177/ 0013164498058003006

Roy-Byrne, P. P., Craske, M. G., Stein, M. B., Sullivan, G., Bystritsky, A., Katon, W. … Sherbourne, C. D. (2005). A randomized effectiveness trial of cognitive-behavioral ther- apy and medication for primary care panic disorder. Archives of General Psychiatry, 62(3), 290–298. https:// doi.org/10.1001/archpsyc.62.3.290

Sewart, A. R., Zbozinek, T. D., Hammen, C. L., Mineka, S., Zinbarg, R. E., & Craske, M. G. (2019). Positive affect as a buffer between chronic stress and symptom severity of emotional disorders. Clinical Psychological Science. https:// doi.org/10.1177/2167702619834576

Stauber, S., Schmid, J. -P., Saner, H., Znoj, H., Saner, G., Grolimund, J., & von Känel, R. (2013). Health-related quality of life is associated with positive affect in patients with coronary heart disease entering cardiac rehabilitation. Journal of Clinical Psychology in Medical Settings, 20(1), 79–87. https://doi.org/10.1007/s10880-012-9311-6

Watson, D., & Clark, L. A. (1991). The Mood and Anxiety Symptom Questionnaire. Unpublished manuscript, Depart- ment of Psychology, University of Iowa, Iowa City.

Watson, D., & Clark, L. A. (1997). Extraversion and its positive emotional core. In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp. 767–793). San Diego, CA: Academic Press. https:// doi.org/10.1016/B978-012134645-4/50030-5

Watson, D., Clark, L. A., & Carey, G. (1988). Positive and negative affectivity and their relation to anxiety and depressive disorders. Journal of Abnormal Psychology, 97 (3), 346. https://doi.org/10.1037/0021-843X.97.3.346

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063. https://doi.org/10.1037/0022- 3514.54.6.1063

Watson, D., & Naragon-Gainey, K. (2010). On the specificity of positive emotional dysfunction in psychopathology: Evidence from the mood and anxiety disorders and schizophrenia/schizotypy. Clinical Psychology Review, 30(7), 839–848. https://doi.org/10.1016/j.cpr.2009.11. 002

RECEIVED: October 22, 2018 ACCEPTED: July 3, 2019 AVAILABLE ONLINE: July 11, 2019

1124 p os i t i v e a n d ne g at i v e af f e ct a s ou t co m e s a nd m o d e r at or s