Cognitive Therapy and Research

, Volume 42, Issue 2, pp 135–145 | Cite as

Expectancies, Working Alliance, and Outcome in Transdiagnostic and Single Diagnosis Treatment for Anxiety Disorders: An Investigation of Mediation

  • Shannon Sauer-Zavala
  • James F. Boswell
  • Kate H. Bentley
  • Johanna Thompson-Hollands
  • Todd J. Farchione
  • David H. Barlow
Original Article


Patients’ outcome expectancies and the working alliance are two psychotherapy process variables that researchers have found to be associated with treatment outcome, irrespective of treatment approach and problem area. Despite this, little is known about the mechanisms accounting for this association, and whether contextual factors (e.g., psychotherapy type) impact the strength of these relationships. The primary aim of this study was to examine whether patient-rated working alliance quality mediates the relationship between outcome expectancies and pre- to post-treatment change in anxiety symptoms using data from a recent randomized clinical trial comparing a transdiagnostic treatment (the Unified Protocol [UP]; Barlow et al., Unified protocol for transdiagnostic treatment of emotional disorders: Client workbook, Oxford University Press, New York, 2011a; Barlow et al., Unified protocol for transdiagnostic treatment of emotional disorders: Patient workbook. New York: Oxford University Press, 2017b) to single diagnosis protocols (SDPs) for patients with a principal heterogeneous anxiety disorder (n = 179). The second aim was to explore whether cognitive-behavioral treatment condition (UP vs. SDP) moderated this indirect relationship. Results from mediation and moderated mediation models indicated that, when collapsing across the two treatment conditions, the relationship between expectancies and outcome was partially mediated by the working alliance [B = 0.037, SE = 0.05, 95% CI (.005, 0.096)]. Interestingly, within-condition analyses showed that this conditional indirect effect was only present for SDP patients, whereas in the UP condition, working alliance did not account for the association between expectancies and outcome. These findings suggest that outcome expectancies and working alliance quality may interact to influence treatment outcomes, and that the nature and strength of the relationships among these constructs may differ as a function of the specific cognitive-behavioral treatment approach utilized.


Outcome expectancies Working alliance Transdiagnostic Cognitive-behavioral therapy Mediation 


Research has demonstrated that cognitive-behavioral therapies (CBT) are highly effective across a wide variety of problem areas (Lambert 2013; Nathan and Gorman 2015). In addition, the effects of CBT have been shown to be durable (Hollon and Beck 2013) and more cost-effective than alternative treatments for psychological problems, including medication (Chiles et al. 1999; McHugh et al. 2007). Although individuals will, on average, benefit more from CBT than no treatment, there remains significant room for improving both response rates and knowledge about the process of change in CBT (Kazdin and Blase 2011).

Psychotherapy researchers categorize process variables as falling into one of three, non-mutually exclusive, domains: participant, relationship, and technical variables (Castonguay and Beutler 2006). Depending on the context, variables within each of these categories can be considered common or unique to a theoretical orientation or problem area. Common factors are treatment variables that predict and/or are important for understanding therapeutic change, yet are not directly tied to a specific treatment approach or model. For example, the working alliance (generally judged from the client’s perspective) has been considered the “flagship” common factor (Castonguay et al. 2010). The causal importance of the alliance in psychological treatments continues to be a focus of debate (Crits-Christoph et al. 2011); however, the most recent meta-analysis of the working alliance-outcome relationship in psychotherapy demonstrated a consistent small-to-moderate correlation between the alliance and outcome. Specifically, more positive working alliance scores were shown to be associated with better treatment outcomes (Horvath et al. 2011). This association has also been observed across a range of distinct treatment approaches and problem areas.

Drawing the conclusion that the alliance may be an important element of psychotherapy irrespective of the approach or primary diagnosis does not imply that the nature or impact of the alliance is independent of context, nor does it assume that the alliance-outcome association exists in a vacuum. The so-called “next generation” of alliance research has emphasized the importance of interactions between alliance and other potentially important process variables, both common and unique, as well as technical and relational (e.g., attachment; see Zack et al. 2015). Treatment outcome expectancy has been proposed as one such variable that can be theoretically linked to both the working alliance and treatment outcomes and is worthy of further study in this context (Boswell et al. 2016).

Treatment outcome expectancies (or expectations) characterize a prognostic belief about the degree to which one will benefit from a current or forthcoming treatment (Constantino et al. 2011). Outcome expectancies are evoked in any therapeutic context, and early treatment outcome expectancies can be considered an important individual difference variable. Psychotherapy-related expectations, which patients often possess when arriving for treatment, have long been considered potent common factors that can shape experiences, perceptions, motivations, and outcomes (Constantino 2012; Frank 1961; Goldfried 1980; Goldstein 1962; Greenberg et al. 2006; Kirsch 1985; Montgomery and Bovbjerg 2004; Rutherford et al. 2010; Weinberger and Eig 1999).

To systematically examine the nature and strength of the outcome expectation-posttreatment outcome link, Constantino et al. (2011) conducted a meta-analysis of psychotherapy studies published in English through 2009. To be included in this analysis, studies had to investigate a bivariate correlation between client-rated expectation at baseline or session 1 and a posttreatment symptom measure not referenced as a follow-up assessment. The resulting meta-analytic sample included 8016 clients from 46 distinct clinical samples. The analysis revealed a small, but statistically significant association between more optimistic early outcome expectation and more favorable posttreatment outcomes (d = 0.24). Moreover, there was no moderating effect on this association of treatment orientation (CBT or other), primary diagnosis (depression, anxiety, substance, or other), treatment mode (individual, group, or other), study design (comparative trial, open trial, or naturalistic), or publication date (before 2000 or 2000–2009); thus, this association appears to be pantheoretical, pandiagnostic, and pancontextual. The apparent robustness of the outcome expectation-treatment outcome link has been further supported in studies published after 2009 (e.g., Price and Anderson 2012; Thompson-Hollands et al. 2014; Tsai et al. 2014; Webb et al. 2013). For example, Thompson-Hollands et al. (2014) found that more positive outcome expectations were associated with better anxiety, depression, and functioning outcomes in a randomized controlled trial (RCT) of the Unified Protocol (UP; Barlow et al. 2011a, 2017b) for diverse principal anxiety disorders.

There is emerging evidence to suggest that there may be certain conditions for which positive outcome expectancies may be particularly important for outcomes. For example, in an RCT comparing the effectiveness of a transdiagnostic Internet-based maintenance treatment (TIMT) to a treatment-as-usual (TAU) control group following inpatient psychotherapy for mental illness, there was a pronounced advantage across a 3-month follow-up period for TIMT for participants with high positive outcome expectation. Although TIMT was generally beneficial, these results suggest that online interventions may be especially so for clients who believe that treatment will help them.

Despite the relatively consistent outcome expectation—treatment outcome correlation, little is known about the mechanisms accounting for this effect in psychotherapy studies. In what limited empirical work has been conducted on potential mechanisms, the working alliance has emerged as a promising candidate. Several studies (of diverse treatments and patients samples) have demonstrated an association between higher pre- or early-treatment outcome expectation and better alliance quality (e.g., Connolly Gibbons et al. 2003; Constantino et al. 2012, 2005; Joyce and Piper 1998; Patterson et al. 2008; Tsai et al. 2014), and still others have formally demonstrated that alliance quality serves as at least a partial mediator of the association between pre- or early-treatment outcome expectation and posttreatment outcome (Abouguendia et al. 2004; Gaudiano and Miller 2006; Johansson et al. 2011; Joyce et al. 2003; Meyer et al. 2002). These findings suggest that pessimistic outcome beliefs may interfere with successful formation and maintenance of the working relationship, which in turn might relate to poorer treatment outcomes.

For example, one study explored the dynamic relationships between outcome expectancy and working alliance. In the context of CBT for generalized anxiety, session 1 outcome expectation moderated the negative association between client-rated alliance ruptures and post-rupture outcome expectation (Westra et al. 2011). Specifically, ruptures had a more potent adverse influence on post-rupture expectations of therapy for clients who started therapy with a more pessimistic (versus optimistic) efficacy expectation. This finding suggests that early pessimistic outcome expectancy may be a risk factor for demoralization vis-à-vis perceived tensions in the working relationship.

Specific Aims

As noted, Thompson-Hollands et al. (2014) observed positive expectancies to be associated with better symptom and functioning outcomes in a small randomized trial of the UP (Farchione et al. 2012). The present study utilized data from a recently completed, large RCT that compared a transdiagnostic CBT protocol (UP) to published, empirically supported single diagnosis CBT protocols (SDPs) for diverse principal anxiety disorders and comorbid conditions (Barlow et al. 2017a, b). As compared to the earlier RCT referenced above (Farchione et al. 2012), in addition to a larger sample that compared two specified CBT conditions and the assessment of early treatment outcome expectancies, the most recent trial (Barlow et al. 2017a, b) included assessment of the working alliance.

This present study had multiple aims. The primary aim was to test if the association between patient outcome expectancies and post-treatment outcome is mediated by the quality of the working alliance, as rated by the patient. The second aim was to test if the indirect effect of expectancy on treatment outcome through the quality of the working alliance is moderated by treatment condition. That is, is the strength of the observed mediation effect a function, at least in part, of receiving the UP or an SDP? We anticipated that positive outcome expectancies and working alliance ratings would be associated with positive treatment outcome. In addition, we anticipated that the association between expectancies and outcome would be partially mediated by working alliance quality. Given the absence of existing research, we did not propose a priori directional hypotheses regarding the moderating effect of treatment condition.



Participants in the present study were drawn from a large sample (N = 223) of treatment-seeking individuals at the Center for Anxiety and Related Disorders at Boston University (CARD) who participated in a trial comparing two active treatment conditions and a waitlist control condition; only active treatment participants (n = 179) were included in the present study. The study was approved by a university institutional review board (IRB) and written informed consent was obtained prior to any research activity. Recruitment was designed to be broadly inclusive, with few exclusion criteria. Individuals were eligible for the study if they were (1) assigned a principal (most interfering and severe) diagnosis of panic disorder, with or without agoraphobia (PD/A), generalized anxiety disorder (GAD), obsessive–compulsive disorder (OCD), or social anxiety disorder (SOC), as assessed using the Anxiety Disorders Interview Schedule (ADIS; see description below); (2) 18 years or older; (3) fluent in English; and (4) able to attend all treatment sessions and assessments. Following long-standing procedures in our clinical trials, individuals taking psychotropic medications at the time of enrollment were required to be stable on the same dose for at least 6 weeks prior to enrolling in the study, and to maintain these medications and dosages during treatment. Exclusion criteria consisted primarily of conditions that required prioritization for immediate or simultaneous treatment that could interact with the study treatment in unknown ways: specifically, a current diagnosis of bipolar disorder, schizophrenia, schizoaffective disorder, or organic mental disorder; current high suicide risk, or; recent (within 3 months) history of substance abuse or drug dependence, with the exception of nicotine, marijuana, and caffeine. Individuals were also excluded if they previously received at least eight sessions of CBT within the past 5 years.


Anxiety Disorders Interview Schedule (ADIS; Dinardo et al. 1994; Brown and Barlow 2014)

Diagnostic assessments were conducted by study evaluators blinded to condition allocation. Patients were assessed for current DSM diagnoses using the ADIS. The ADIS is a semi-structured clinical interview that focuses on DSM diagnoses of anxiety, mood, somatic symptom, and substance use disorders, with screening for other disorders. Diagnoses are assigned a clinical severity rating (CSR) on a scale from 0 (no symptoms) to 8 (extremely severe symptoms), with a rating of 4 or above (definitely disturbing/disabling) representing the clinical threshold for DSM diagnostic criteria. Due to the introduction of DSM-5 partway through the trial, 168 patients (75%) were assigned diagnoses based on DSM-IV criteria and 55 patients (25%) were assigned diagnoses based on DSM-5 criteria. To standardize clinical severity ratings across these phases, an additional rating was assigned to overall PD/A symptoms for those patients diagnosed according to DSM-5, despite the separation of panic disorder and agoraphobia in DSM-5.

Credibility/Expectancy Questionnaire (CEQ; Devilly and Borkovec 2000)

To assess outcome expectations, patients indicated how much they thought that they would improve by the end of the treatment period on an 11-point scale (from 0 to 100% in 10-point intervals). This cognitively based item, part of the Credibility/Expectancy Questionnaire (CEQ; Devilly and Borkovec 2000), has been used as a measure of outcome expectancy on its own (e.g., Borkovec et al. 2002; Vogel et al. 2006), possesses good face validity, and has been shown to predict treatment outcome (e.g., Borkovec et al. 2002; Price et al. 2008). The CEQ was administered after session 2 to ensure that the treatment rationale and initial introduction of the treatment plan had been delivered.

Working Alliance Inventory-Short Form (WAI-S; Tracey and Kokotovic 1989)

The WAI was developed to assess three dimensions of the therapeutic relationship as conceptualized by Bordin (1979)—client and therapist (a) agreement on goals (goals), (b) agreement on how to achieve these goals (tasks), and (c) affective relationship (bond). The original 36-item questionnaire is rated using a 7-point Likert scale, with items mapping on to a global alliance dimension as well as the goal, task, and bond component subscales. Confirmatory factor analysis by Tracey and Kokotovic (1989) created a 12-item short version of the WAI by taking the four items that loaded most strongly on each of the three factors and retaining the same structure and subscales as the longer version. Busseri and Tyler (2003) evaluated the interchangeability of the WAI and the WAI-S in a direct comparison study and found equally good test–retest reliability, concurrent validity, and predictive validity for therapeutic improvement with both measures. The WAI-S may thus actually be preferable to the WAI, given its greater ease of administration and equally strong psychometric properties. The WAI was administered after sessions 4, 8, and 12 in the trial, and the total score was used in all analyses. Coefficient alpha for this sample was α = 84.

Hamilton Anxiety Ratings Scale (HARS; Hamilton 1959)

The HARS was used to assess anxiety symptoms and was administered in accordance with the Structured Interview Guide for the Hamilton Anxiety (SIGH-A; Shear et al. 2001). This commonly used measure has demonstrated good levels of interrater and test–retest reliability, as well as convergent validity with similar clinician rated and self-report measures of anxiety symptoms (Shear et al. 2001). Independent clinical evaluators received extensive training on the SIGH-A and had to demonstrate acceptable levels of reliability prior to their participation in the trial.


A detailed description of the procedures, including randomization and participant flow, can be found in Barlow et al. (2017a, b). Patients were randomized by principal diagnosis (PD/A, GAD, OCD, and SOC), following a 2:2:1 allocation ratio to UP, SDP, and waitlist control study conditions, respectively. The present study focuses on participants who were randomized to either the UP or SDP condition. Following baseline assessment and randomization, patients in the UP and SDP conditions received between 12 and 16, 50–90 min (see below) sessions of weekly individual treatment.

Unified Protocol (Barlow et al. 2011a, b, 2017a, b)

The UP consists of six core treatment modules: (a) psychoeducation on the adaptive nature of emotions; (b) increasing mindful emotion awareness; (c) increasing cognitive flexibility; (d) identifying and preventing patterns of emotion avoidance; (e) increasing awareness and tolerance of emotion-related physical sensations, and; (f) interoceptive and situational emotion-focused exposures. The six core modules are preceded by a module focused on enhancing motivation, readiness for change, and treatment engagement. A final module consists of reviewing progress over treatment and developing relapse prevention strategies. Treatment and session length of the UP were matched to the SDPs for each principal diagnosis (in accordance with the guidelines described below).

Single Diagnosis Protocols (SDPs)

The SDP treatment protocols included: Managing Social Anxiety: A Cognitive-Behavioral Therapy Approach—2nd edition (MSA-II; Hope et al. 2006); Mastery of Anxiety and Panic—4th edition (MAP-IV; Barlow and Craske 2007); Mastery of Anxiety and Worry—2nd edition (MAW-II; Zinbarg et al. 2006); and Treating Your OCD with Exposure and Response (Ritual) Prevention Therapy—2nd edition (Foa et al. 2008). As recommended by the treatment developers, the MSA, MAW, and OCD protocols were conducted over the course of 16 sessions, whereas the MAP-IV was conducted over 12 sessions. All treatments were administered in an individual format and treatment sessions lasted approximately 50–60 min, with the exception of the OCD treatment protocols, which were 80–90 min for both UP and SDP conditions.

Therapists and Treatment Integrity

Therapists for the study included doctoral students in clinical psychology with 2–4 years of experience, postdoctoral fellows with 5–6 years of experience and licensed psychologists with 10 or more years of experience. Each therapist administered both types of treatment. Initial training and certification in the treatment protocols utilized procedures employed in clinical trials at CARD over the last 20 years (Barlow 2000). Twenty percent of treatment sessions were randomly selected and rated for adherence and competence by raters associated with development of the specific treatments. Treatment fidelity scores were good to excellent (M = UP = 4.44 out of 5; SDPs = 4.09 out of 5).

Data Analysis and Missingness

All analyses were conducted using SPSS version 20.0.0. Pearson correlations were obtained to capture relationships between study variables (outcome expectancy, working alliance, and treatment outcomes). The mediation (Fig. 1) and moderated mediation (Fig. 2) models were evaluated with the PROCESS macro developed by Hayes (2013). In this approach, effects are assessed with bias corrected bootstrap confidence intervals that are significant when the upper and lower bound of the bias corrected 95% confidence intervals (CI) does not contain zero. We began with a simple mediation (PROCESS model 4) to assess for an indirect effect of outcome expectancy (independent variable) on session 4 to post-treatment change in anxiety symptoms (dependent variable) through working alliance (the mediator) within our full sample, controlling for early change in anxiety symptoms (pre-treatment to session 4). Then, we formally tested moderated meditation using PROCESS model 58 (Fig. 2) as a means to explore whether the simple mediation pathways differed as a function of treatment condition. For all analyses, we used bootstrapping with 10,000 samples.

Fig. 1

Simple mediation: The indirect effect of treatment expectancy on symptom improvement through working alliance

Fig. 2

Moderated mediation: Treatment condition moderates the indirect effect of treatment expectancy on symptom improvement through working alliance

Given that any case with a missing value on variables of interest is excluded from analyses using PROCESS, it was important to consider treatment of missing data in this investigation. For scales that utilize a total score (i.e., HARS, WAI-S), mean imputation was used such that the average of completed items was substituted for missing items, as long as no more than 30% of items on a given scale were missing, in the calculation of the total score (Ake 2005; Fox-Wasylyshyn and El-Masri 2005; Roth et al. 1999). Following these procedures, no cases were missing HRS total scores at baseline, 59 cases were missing data on this variable at post-treatment, and 51 cases were missing WAI-S total scores. These imputation procedures could not be implemented for the CEQ, as a single item was used; 26 cases were missing item 4 from the CEQ. Chi square tests revealed that number of missing cases on each study variable did not differ as a function of condition or sex (p > .05). Additionally, independent sample t tests were conducted to examine whether number of missing cases differed as a function of baseline severity (ADIS CSR) and age. There were no significant differences for any of these tests, with the exception of age on the number of missing post-HARS scores (t = 2.3, p = .02); on average, individuals with missing data on this variable tended to be older (M = 33.34 years) than those with complete data (M = 29.34).


Preliminary Analyses

Means and standard deviations for each study variable for the full sample, as well as for UP and SDP participants separately, can be seen in Table 1. To ensure that our outcome variable, change in anxiety symptoms from session 4 to post-treatment, was reliable in the present study, the reliability of this change score was calculated (rDD) using the formula specified by King et al. (2006) and was found to be adequate across groups (full sample: rDD = 0.72, UP: rDD = 0.71, SDP: rDD = 0.71).1 Confidence limits for effect size estimates comparing UP to SDP participants suggested that there were no significant differences between conditions on any study variables or possible covariates (e.g., number of comorbid diagnoses). Additionally, working alliance scores in session 4 were not significantly different from session 8 [ES sg  = −0.14 (−0.31, 0.03)] or session 12 [ES sg  = 0.12 (−0.09, 0.32)] in the full sample, justifying the use of the session 4 measure in subsequent analyses. Observed mean expectancy scores were consistent with values reported in other CBT trials (e.g., Newman et al. 2011). The average WAI total score reflected moderately strong, positive alliances in both conditions. Correlations amongst study variables in the full sample and as a function of treatment condition can be seen in Table 2. All variables were significantly correlated with each other, with the exception of the relationship between working alliance and change in anxiety symptoms from session four to post-treatment in the UP condition.

Table 1

Means and standard deviations for study variables as a function of treatment condition


Full sample



Hedges g (95% CI)

Treatment expectancy

66.80 (16.88)

65.71 (17.87)

67.89 (15.85)

−0.13 (−0.45, 0.19)

Working alliance session 4

69.62 (10.44)

70.81 (09.96)

68.31 (10.87)

0.24 (−0.11, 0.59)

Working alliance session 8

71.65 (09.33)

72.52 (09.26)

70.66 (09.38)

0.20 (−0.15, 0.55)

Working alliance session 12

68.44 (08.67)

69.12 (08.77)

67.55 (08.56)

0.10 (−0.26, 047)

Change in anxiety symptoms

6.62 (07.41)

6.07 (07.80)

7.23 (6.96)

0.16 (−52, 0.21)

Number of comorbid diagnoses

2.27 (1.83)

2.22 (1.69)

2.32 (1.95)

t = 0.37, ns

Treatment expectancy was measured with the C/EQ, working alliance was measured with the WAI, and session 4- to post-treatment change was calculated by creating change scores from session 4 and post-treatment HAS scores. Hedges g effect size compares UP and SDP conditions. Number of participants included in each analysis range from 120 to 153 for the full sample, 63–77 for UP, and 51–76 for SDP, due to deletion of cases with missing values

Table 2

Relationships amongst study variables for the full sample


Full sample









1. Treatment expectancy

2. Working alliance session 4




3. Change in anxiety symptoms




0.21 (ns)



All relationships significant at the p < .01 level except when noted. Treatment expectancy was measured with the C/EQ, working alliance was measured with the WAI, and session 4- to post-treatment change was calculated by creating change scores from session 4 and post-treatment HAS scores. Number of participants included in each analysis range from 100 to 128 for the full sample, 53–77 for UP, and 47–61 for SDP, due to deletion of cases with missing values

Mediation Analyses2

Next, we examined the mediational model depicted in Fig. 1 using the SPSS PROCESS macro (Hayes 2013); all analyses were conducted controlling for early change in anxiety symptoms (from pre-treatment to session 4). First, treatment expectancy (independent variable) significantly predicted session 4 patient-rated working alliance (mediator) [B = 0.21 (.08, 0.33), p = .001]. When treatment expectancy and working alliance were simultaneously included in the model, treatment expectancy significantly predicted session four to post-treatment change in anxiety (outcome) [B = 0.11 (.003, 0.22), p = .008], whereas working alliance did not [B = 0.14 (−0.03, 0.32)], p = .076). This model, with both predictors, explained 18% of variance in pre- to post-treatment change in anxiety (R2 = 0.18, p = .0002). Analysis from a bias-corrected bootstrap with 10,000 resamples (Hayes 2013) revealed a significant indirect effect [B = 0.03, SE = 0.02, 95% CI (.003, 0.09)] of treatment expectancy on change in anxiety, through working alliance. Using the estimate described by Fairchild et al. (2009), the indirect effect accounted for 6% of the variance in change in anxiety symptoms [R 2 med = 0.06 (.004, 0.17)]. Thus, working alliance partially mediated the relationship between positive treatment expectancy and change in anxiety symptoms.

Moderated Mediation Analyses

Next, we used the SPSS PROCESS macro to explore the possibility of moderated mediation; again, all analyses controlled for early change in anxiety symptoms (see Fig. 2). First, we explored whether the path between treatment expectancy and working alliance (“a” path) was moderated by treatment condition. When included simultaneously in the model (controlling for early change in anxiety symptoms), neither treatment expectancy (B = 0.009 [-0.36, 0.38], p = .96), treatment condition [B = −12.01 (−29.00, 4.97), p = .16], nor their interaction term [B = 0.14 (−0.11, 0.39), p = .250] significantly predicted session 4 working alliance. We then explored whether the path between working alliance and change in anxiety symptoms (“b” path) was moderated by treatment condition and a similar pattern of relationships was revealed. When included simultaneously in the model, neither treatment expectancy [B = 0.11 (−0.004, 0.22), p = .06] working alliance [B = −0.07 (−0.59, 0.45), p = .79], treatment condition [B = −9.47 (−33.16, 14.21), p = .43], nor their interaction term [B = 0.14 (−0.18, 0.47), p = .38] were significant predictors of change in anxiety symptoms. Overall, the inferential test of moderated mediation was not statistically significant, B = 0.05, SE = 0.042, 95% CI (−0.01, 0.15). Yet, the separate within group conditional indirect effect tests demonstrated that the conditional indirect effect was only significant for the SDP condition [UP: B = 0.01, SE = 0.02, 95% CI (−0.02, 0.09); SDP: B = 0.07, SE = 0.03, 95% CI (.01, 0.16)].

Given the conflicting evidence supporting the notion that the indirect effect of expectancy predicting outcome through alliance was moderated by treatment condition (e.g., non-significant interaction effects vs. differentially significant indirect effects as a function of condition), we sought to further clarify the nature of these relationships by re-running our analyses using each subscale of the WAI (task, goal, and bond) as our mediator. There did not appear to be a significant indirect effect through WAI task for the UP condition [B = 0.03, SE = 0.03, 95% CI (−0.004, 0.13)] or the SPD condition [B = 0.04, SE = 0.04, 95% CI (−0.02, 0.13)]; further, the interactions terms between expectancy and treatment condition predicting WAI task [B = 0.04 (−0.02, 0.15), p = .12] and between WAI task and treatment condition prediction outcomes [B = − 0.20 (−1.15, 0.75), p = .67] were also not significant. Similarly, for WAI goal, neither interaction term expectancy x treatment condition predicting WAI goal: [B = 0.02 (−0.06, 0.11), p = .57]; WAI goal x treatment condition predicting outcome [B = 0.45 (−0.54, 1.45), p = .37] was significant. In contrast, there was evidence for differential strength of the indirect effect between expectancy and outcome through WAI goal as a function of treatment condition, as this effect was significant in the SDP condition [B = 0.04, SE = 0.03, 95% CI (.0005, 0.12)] but not the UP condition [B = 0.006, SE = 0.02, 95% CI (−0.02, 0.07)]. The strongest case for moderated mediation can be made for the WAI bond subscale which evidenced a trend-level interaction term between WAI bond and treatment condition predicting outcomes [B = 0.67 (−0.10, 1.45), p = .08], as well as a significant indirect effect for the relationship between expectancy and outcomes through WAI bond for the SDP condition [B = 0.07, SE = 0.03, 95% CI (.02, 0.17)], but not the UP condition [B = 0.006, SE = 0.02, 95% CI (−0.03, 0.05)]. The interaction term between expectancy and treatment condition predicting WAI bond was not significant [B = 0.05 (−0.05, 0.15), p = .35].


The present study examined the relationships among patients’ early treatment outcome expectancies, ratings of working alliance, and treatment outcomes in the context of a RCT comparing two types of CBT protocols. Our first hypothesis, that the relationship between treatment expectancy and change in anxiety symptoms would be partially mediated by alliance ratings (collapsed across treatment conditions), was supported. Overall the two predictors accounted for 18% of the variance in anxiety outcomes, with the indirect effect accounting for just over one-third of the explained variance. This finding is consistent with and extends previous research on treatment expectancies, therapeutic alliance, and outcome (e.g., Johansson et al. 2011; Meyer et al. 2002) showing that outcome expectations are associated with a variety of individual characteristics and in-therapy behaviors that may contribute to patients’ overall assessment of the working alliance (e.g., psychological mindedness; Constantino et al. 2016), hostility or affiliative behavior towards the therapist (e.g., Ahmed et al. 2012), and treatment outcome.

Subsequent analyses in the present study provided some evidence that the overall mediation effect was moderated by treatment condition, such that the SDP patients demonstrated a significant conditional indirect effect from expectancies to anxiety outcomes through alliance that was not present among the UP patients. That is, the effect of outcome expectancy on symptom change was observed to be more dependent on early treatment phase working alliance scores in the SDPs compared to the UP. With regard to the three working alliance dimensions (task, bond, goal), the strongest evidence in support of this indirect effect was for the working alliance bond. Given that our analyses examining the role of treatment condition were exploratory, these findings were somewhat surprising and warrants further commentary.

To our knowledge, this is the first such investigation of these particular mechanisms within the context of a controlled transdiagnostic-SDP comparison. As such, our offered interpretations should be considered tentative. Both the UP and SDPs are CBT-based and utilize similar treatment procedures (e.g., objective monitoring of emotional experiences, cognitive restructuring, in vivo and imaginal exposure). In addition, both approaches seek to target comparable mechanisms. However, there are relatively subtle differences in the early treatment rationales. Specifically, the UP is less focused on specific symptoms and more focused on the full range of emotions and factors that cut across common emotional disorders, such as emotion avoidance, affect and distress intolerance, and cognitive rigidity (Boswell 2013). In contrast, SDPs target a more narrowly defined set of symptoms, behaviors, and stimuli (e.g., social interactions, panic sensations, or intrusive thoughts). Although speculative, this has two potential implications for the patient’s experience of the treatment. First, the transdiagnostic cross-cutting rationale is an approach that focuses on the underlying issues that characterize the “whole patient.” Second, comorbidity is the rule rather than the exception and SDPs are, by definition, intended to target a single disorder. Consequently, UP patients may experience the “whole patient” rationale and approach more positively; whereas, SDP patients may struggle more with the relatively narrow focus on specific symptoms or a single diagnostic label and experience the therapist as missing the whole picture at times. This potential distinction may not directly impact early treatment outcome expectations (after session 2), yet it might have implications for the expectancy-outcome association vis-à-vis the development of the working alliance as treatment ensues. Concretely, the strength of the working alliance (and especially working alliance bond) may become more important in single-disorder focused treatment because it buffers against concerns that might be triggered by the relatively narrower, specific symptom focused SDP approach.

It could also be that distinct early treatment procedures influence the observed between-condition differences in the relationship between working alliance (as assessed at session 4) and outcome. As previously noted, 1 of the 8 UP modules (typically delivered during sessions 2 to 3) is dedicated entirely to motivation enhancement, and includes two exercises specifically designed to foster motivation for change (a decisional balance and goal-setting). The other SDPs utilized in this trial either do not incorporate formal motivation enhancement procedures or include a single motivation exercise as one component of a broader introductory session. It is possible that greater explicit emphasis on motivation enhancement in the UP encourages patients to view their potential for progress during treatment as largely contingent on how much effort they put into treatment, thereby potentially reducing the impact of the therapist-client alliance on outcome (see Boswell et al. 2015). During SDP treatment, with less explicit early emphasis on fostering motivation to change, the strength of the therapeutic alliance may be more critical to symptom change. This notion is speculative, however, especially given that skilled therapists are likely to incorporate motivational interviewing techniques to address ambivalence regardless of whether they are using a transdiagnostic or single-disorder approach. It is also possible that the effect of treatment expectancies in the UP is mediated by different variables that may be more relevant to or characteristic of the UP than SDPs. In this study, we intentionally focused on two established common factors in the literature. Further studies to clarify the nature of the relationship between treatment expectancies and outcome during transdiagnostic, emotion-focused CBT are warranted.

This study had several limitations. First, although we used validated and well-established measures of both treatment expectancies and alliance, these measures are both self-report and could therefore be bolstered by observational or audio coding of actual treatment behaviors. Second, the sample consisted only of patients with four principal anxiety disorder diagnoses, and our results may not hold among other treatment-seeking samples. However, patients in the present study did receive a number of comorbid diagnoses at baseline, including other anxiety disorders, depressive disorders, and posttraumatic stress disorder, among others (Barlow et al., under review). Future research should explore the moderating role of comorbid diagnoses in the relationships between treatment condition, outcome expectancy, and working alliance. As noted above, it is possible that that individuals with co-occurring conditions may find SDPs less credible than the UP, driving the importance of working alliance as the mediator through which outcome expectancy exerts its effects on symptom improvement. Third, we focused this study on the relationship between process variables and anxiety outcomes specifically. Further studies examining whether the influence of expectancies and working alliance on outcomes differs as a function of the particular outcome examined are needed. Fourth, we were unable to determine a more precise potential causal link from outcome expectancies and working alliance to outcomes given the lack of experimental manipulation of our designated predictor and mediator variables; research on these constructs that provides information about causality by manipulating expectancies and/or the working alliance is needed. Additionally, given the timing of our assessments, we were unable to explore the mediating effect of dynamic change in working alliance in the relationship between treatment expectancy and outcomes; given that the working alliance between patient and therapist is not static, the impact of trajectories of change on this variable should be explored in future work. Finally, we excluded any cases with missing values on variables of interest in the present analyses. Although we tested for differences in missing data patterns as a function of treatment condition, sex, baseline severity, and age, these procedures for handling missing data are likely not as robust as other well-established imputation procedures (e.g., full information maximum likelihood [FIML]), and the strength of our findings should be tempered accordingly.


This study contributes to the extant literature on expectancy, working alliance, and outcome during CBT for anxiety. Specifically, our results provide further support for the notion that the relationship between outcome expectancies and post-treatment outcome is partially accounted for by the quality of the working alliance during CBT. The results also provide initial evidence that this mediating effect may be stronger during treatment within an SDP framework as compared to a transdiagnostic approach (UP). Future work is needed to shed more light on how and why these important process variables may function and interact differently in distinct treatment approaches for anxiety and related disorders. Findings from this line of research are likely to have meaningful clinical implications for therapists and psychotherapy researchers alike. For example, these results do not imply that the working alliance or that the expectancy-working alliance relationship is unimportant for treatment outcome in transdiagnostic CBT. Rather, the quality of the working alliance appears to be particularly important when following SDP manuals with more complex cases. In addition, this highlights the need to routinely assess outcome expectancies and the quality of the working alliance in routine treatment (Constantino et al. 2013).


  1. 1.

    Although change scores were once considered to be unreliable, recent research has demonstrated that they can be reliable measures of intraindividual change (King et al. 2006).

  2. 2.

    Given the significant difference in number of missing values for post-treatment HARS scores as a function of age, we also ran the models reported in this section controlling for this. As values remained largely unchanged, we report only the uncontrolled models for brevity.



We would like to thank Dr. Michael Constantino for his helpful comments regarding this study and manuscript.


This study was funded by grant R01 MH090053 from the National Institutes of Health.

Compliance with Ethical Standards

Conflict of Interest

Dr. Barlow receives royalties from Oxford University Press (which includes royalties for all five treatment manuals included in this study), Guilford Publications Inc., Cengage Learning, and Pearson Publishing. Grant monies for various projects come from the National Institute of Mental Health, the National Institute of Alcohol and Alcohol Abuse, and Colciencias (Government of Columbia Initiative for Science, Technology, and Health Innovation). Consulting and honoraria during the past several years have come from the Agency for Healthcare Research and Quality, the Foundation for Informed Medical Decision Making, the Department of Defense, the Renfrew Center, the Chinese University of Hong Kong, Universidad Católica de SantaMaria (Arequipa, Peru), New Zealand Psychological Association, Hebrew University of Jerusalem, Mayo Clinic, and various American Universities. Shannon Sauer-Zavala, James F. Boswell, Kate H. Bentley, Johanna Thompson-Hollands, Todd J. Farchione declare that they have no conflict of interest.

Ethical Standard

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. No animal studies were carried out by the authors for this article.

Informed Consent

Informed consent was obtained from all individual participants included in the study. No animal studies were carried out by the authors for this article.


  1. Abouguendia, M., Joyce, A. S., Piper, W. E., & Ogrodniczuk, J. S. (2004). Alliance as a mediator of expectancy effects in short-term group psychotherapy. Group Dynamics, 8, 3–12. doi: 10.1037/1089-2699.8.1.3.CrossRefGoogle Scholar
  2. Ahmed, M., Westra, H. A., & Constantino, M. J. (2012). Early therapy interpersonal process differentiating clients high and low in outcome expectations. Psychotherapy Research, 22, 731–745.CrossRefPubMedGoogle Scholar
  3. Ake, C. F. (2005). Rounding after multiple imputation with non-binary categorical covariates. Paper presented at the annual meeting of the SAS Users Group International, Philadelphia, PA.Google Scholar
  4. Barlow, D. H. (2000). Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory. American Psychologist, 55, 1247–1263.CrossRefPubMedGoogle Scholar
  5. Barlow, D. H., & Craske, M. G. (2007). Mastery of your anxiety and panic: Workbook (4th edn.). New York, NY: Oxford University Press.Google Scholar
  6. Barlow, D. H., Ellard, K. K., Fairholme, C. P., Farchione, T. J., Boisseau, C. L., Allen, L. B., & Ehrenreich-May, J. (2011a). Unified protocol for transdiagnostic treatment of emotional disorders: Client workbook. New York: Oxford University Press.Google Scholar
  7. Barlow, D. H., Farchione, T. J., Fairholme, C. P., Ellard, K. K., Boisseau, C. L., Allen, L. B., & Ehrenreich-May, J. (2011b). Unified protocol for transdiagnostic treatment of emotional disorders: Therapist guide. New York: Oxford University Press.Google Scholar
  8. Barlow, D. H., Farchione, T. J., Sauer-Zavala, S., Latin, H., Ellard, K. K., Bullis, J. R.,…, Cassiello-Robbins, C. (2017a). Unified protocol for transdiagnostic treatment of emotional disorders: Therapist guide. New York: Oxford University Press.CrossRefGoogle Scholar
  9. Barlow, D. H., Sauer-Zavala, S., Farchione, T.J., Latin, H., Ellard, K. K., Bullis, J. R.,…, Cassiello-Robbins, C. (2017b). Unified protocol for transdiagnostic treatment of emotional disorders: Patient workbook. New York: Oxford University Press.CrossRefGoogle Scholar
  10. Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy, Research and Practice, 16, 252–260.CrossRefGoogle Scholar
  11. Borkovec, T. D., Newman, M. G., Pincus, A. L., & Lytle, R. (2002). A component analysis of cognitive-behavioral therapy for generalized anxiety disorders and the role of interpersonal problems. Journal of Consulting and Clinical Psychology, 70, 288–298. doi: 10.1037/0022-006X.70.2.288.CrossRefPubMedGoogle Scholar
  12. Boswell, J. F. (2013). Intervention strategies and clinical process in transdiagnostic cognitive-behavioral therapy. Psychotherapy, 50, 381–386. doi: 10.1037/a0032157.CrossRefPubMedGoogle Scholar
  13. Boswell, J. F., Bentley, K. H., & Barlow, D. H. (2015). Motivation facilitation in the unified protocol for transdiagnostic treatment of emotional disorders. In H. Arkowitz, W. Miller & S. Rollnick (Eds.), Motivational interviewing in the treatment of psychological problems (2nd edn., pp. 33–57). New York: Guilford Press.Google Scholar
  14. Boswell, J. F., Constantino, M. J., & Anderson, L. M. (2016). Potential obstacles to treatment success in adults: Client characteristics. In S. Maltzman (Ed.), The Oxford handbook of treatment processes and outcomes in psychology. New York: Oxford University Press. doi: 10.1093/oxfordhb/9780199739134.013.17.Google Scholar
  15. Brown, T. A., & Barlow, D. H. (2014). Anxiety and related disorders interview schedule for DSM-5—Lifetime version. London: Oxford University Press.Google Scholar
  16. Busseri, M. A., & Tyler, J. D. (2003). Interchangeability of the working alliance inventory and working alliance inventory, Short form. Psychological Assessment, 15, 193–197.CrossRefPubMedGoogle Scholar
  17. Castonguay, L. G., & Beutler, L. E. (Eds.). (2006). Principles of therapeutic change that work. New York, NY: Oxford University Press.Google Scholar
  18. Castonguay, L. G., Constantino, M. J., Boswell, J. F., & Kraus, D. (2010). The therapeutic alliance: Research and theory. In L. Horowitz & S. Strack (Eds.), Handbook of interpersonal psychology: Theory, research, assessment, and therapeutic interventions (pp. 509–518). New York, NY: Wiley.Google Scholar
  19. Chiles, J. A., Lambert, M. J., & Hatch, A. L. (1999). The impact of psychological interventions on medical cost offset: A meta-analytic review. Clinical Psychology: Science and Practice, 6, 204–220.Google Scholar
  20. Connolly Gibbons, M. B., Crits-Christoph, P., de la Cruz, C., Barber, J. P., Siqueland, L., & Gladis, M. (2003). Pretreatment expectations, interpersonal functioning, and symptoms in the prediction of the therapeutic alliance across supportive-expressive psychotherapy and cognitive therapy. Psychotherapy Research, 13, 59–76.CrossRefPubMedGoogle Scholar
  21. Constantino, M. J. (2012). Believing is seeing: An evolving research program on patients’ psychotherapy expectations. Psychotherapy Research, 22, 127–138. doi: 10.1080/10503307.2012.663512.CrossRefPubMedGoogle Scholar
  22. Constantino, M. J., Ametrano, R. M., & Greenberg, R. P. (2012). Clinician interventions and participant characteristics that foster adaptive patient expectations for psychotherapy and psychotherapeutic change. Psychotherapy, 49, 557–569. doi: 10.1037/a0029440.CrossRefPubMedGoogle Scholar
  23. Constantino, M. J., Arnow, B. A., Blasey, C., & Agras, W. (2005). The association between patient characteristics and the therapeutic alliance in cognitive-behavioral and interpersonal therapy for bulimia nervosa. Journal of Consulting and Clinical Psychology, 73, 203–211. doi: 10.1037/0022-006X.73.2.203.CrossRefPubMedGoogle Scholar
  24. Constantino, M. J., Boswell, J. F., Bernecker, S. L., & Castonguay, L. G. (2013). Context-responsive integration as a framework for unified psychotherapy and clinical science: Conceptual and empirical considerations. Journal of Unified Psychotherapy and Clinical Science, 2, 1–20.Google Scholar
  25. Constantino, M. J., Coyne, A. E., Mcvicar, E. L., & Ametrano, R. M. (2016). The relative association between individual difference variables and general psychotherapy outcome expectation in socially anxious individuals. Psychotherapy Research, 11, 1–12.Google Scholar
  26. Constantino, M. J., Glass, C. R., Arnkoff, D. B., Ametrano, R. M., & Smith, J. Z. (2011). Expectations. In J. C. Norcross (Ed.), Psychotherapy relationships that work: Evidence based responsiveness (2nd edn., pp. 354–376). New York, NY: Oxford University Press.CrossRefGoogle Scholar
  27. Crits-Christoph, P., Connolly Gibbons, M. B., Hamilton, J., Ring-Kurtz, S., & Gallop, R. (2011). The dependability of alliance assessments: The alliance-outcome correlation is larger than you think. Journal of Consulting and Clinical Psychology, 79, 267–278. doi: 10.1037/a0023668.CrossRefPubMedPubMedCentralGoogle Scholar
  28. Devilly, G., & Borkovec, T. (2000). Psychometric properties of the credibility/expectancy questionnaire. Journal of Behavior Therapy and Experimental Psychiatry, 31, 73–86. doi: 10.1016/S0005-7916(00)00012-4.CrossRefPubMedGoogle Scholar
  29. Di Nardo, P. A., Brown, T. A., & Barlow, D. H. (1994). Anxiety disorders interview schedule for DSM-IV: Lifetime version (ADIS-IV-L). New York, NY: Oxford University Press.Google Scholar
  30. Fairchild, A. J., MacKinnon, D. P., Toborga, M. P., & Taylor, A. B. (2009). R-suared effect-sie measures of mediation analysis. Behavior Research Methods, 41, 486–498.CrossRefPubMedPubMedCentralGoogle Scholar
  31. Farchione, T. J., Fariholme, C. P., Ellard, K. K., Boisseau, C. L., Thompson-Hollands, J., Carl, J. R.,…, Barlow, D. H. (2012). Unified protocol for transdiagnostic treatment of emotional disorders: A randomized controlled trial. Behavior Therapy, 43, 666–678.CrossRefPubMedPubMedCentralGoogle Scholar
  32. Foa, E. B., Yadin, E., & Lichner, T. K. (2008). Obsessive-compulsive disorder: A cognitive-behavioral therapy approach. New York, NY: Oxford University Press.Google Scholar
  33. Fox-Wasylyshyn, S. M., & El-Masri, M. M. (2005). Handling missing data in self-report measures. Research in Nursing & Health, 28(6), 488–495. doi: 10.1002/nur.20100.CrossRefGoogle Scholar
  34. Frank, J. D. (1961). Persuasion and healing: A comparative study of psychotherapy. Oxford: Johns Hopkins University Press.Google Scholar
  35. Gaudiano, B., & Miller, I. (2006). Patients’ expectancies, the alliance in pharmacotherapy, and treatment outcomes in bipolar disorder. Journal of Consulting and Clinical Psychology, 74, 671–676. doi: 10.1037/0022-006X.74.4.671.CrossRefPubMedGoogle Scholar
  36. Goldfried, M. R. (1980). Toward the delineation of therapeutic change principles. American Psychologist, 35, 991–999. doi: 10.1037/0003-066X.35.11.991.CrossRefPubMedGoogle Scholar
  37. Goldstein, A. P. (1962). Patient-therapist expectancies in psychotherapy. Oxford: Pergamon.Google Scholar
  38. Greenberg, R. P., Constantino, M. J., & Bruce, N. (2006). Are patient expectations still relevant for psychotherapy process and outcome? Clinical Psychology Review, 26, 657–678. doi: 10.1016/j.cpr.2005.03.002.CrossRefPubMedGoogle Scholar
  39. Hamilton, M. (1959). The assessment of anxiety states by rating. British Journal of Medical Psychology, 32, 50–55.CrossRefPubMedGoogle Scholar
  40. Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York: NY: The Guilford Press.Google Scholar
  41. Hollon, S. D., & Beck, A. T. (2013). Cognitive and cognitive-behavioral therapies. In: M. J. Lambert (Ed.), Bergin and Garfield’s handbook of psychotherpay and behavior change (6th edn). Mahwah, NJ: Wiley.Google Scholar
  42. Hope, D. A., Heimberg, R. G., & Turk, C. L. (2006). Managing social anxiety: A cognitive-behavioral therapy approach: Therapist guide. San Antonio, TX: Oxford University Press.Google Scholar
  43. Horvath, A. O., Del Re, A., Flückiger, C., & Symonds, D. (2011). Alliance in individual psychotherapy. In J. C. Norcross (Ed.), Psychotherapy relationships that work: Evidence-based responsiveness (2nd edn., pp. 25–69). New York, NY: Oxford University Press.CrossRefGoogle Scholar
  44. Johansson, P., Høglend, P., & Hersoug, A. (2011). Therapeutic alliance mediates the effect of patient expectancy in dynamic psychotherapy. British Journal of Clinical Psychology, 50, 283–297.PubMedGoogle Scholar
  45. Joyce, A. S., Ogrodniczuk, J. S., Piper, W. E., & McCallum, M. (2003). The alliance as mediator of expectancy effects in short-term individual therapy. Journal of Consulting and Clinical Psychology, 71, 672–679. doi: 10.1037/0022-006X.71.4.672.CrossRefPubMedGoogle Scholar
  46. Joyce, A. S., & Piper, W. E. (1998). Expectancy, the therapeutic alliance, and treatment outcome in short-term individual psychotherapy. Journal of Psychotherapy Practice and Research, 7, 236–248.PubMedPubMedCentralGoogle Scholar
  47. Kazdin, A. E., & Blase, S. L. (2011). Rebooting psychotherapy research and practice to reduce the burden of mental illness. Perspectives on Psychological Science, 6, 21–37. doi: 10.1177/1745691610393527.CrossRefPubMedGoogle Scholar
  48. King, L. A., King, D. W., McArdle, J. J., Saxe, G. N., Doron-Lamarca, S., & Orazen, R. J. (2006). Latent difference score approach to longitudinal trauma research. Journal of Trauma & Stress, 19(6), 771–785.Google Scholar
  49. Kirsch, I. (1985). Response expectancy as a determinant of experience and behavior. American Psychologist, 40, 1189–1202. doi: 10.1037/0003-066X.40.11.1189.CrossRefGoogle Scholar
  50. Lambert, M. J. (2013). The efficacy and effectiveness of psychotherapy. In M. J. Lambert (Ed.), Bergin & Garfield’s handbook of psychotherapy and behavior change (6th edn., pp. 169–218). New York, NY: John Wiley & Sons.Google Scholar
  51. McHugh, R. K., Otto, M. W., Barlow, D. H., Gorman, J. M., Shear, M. K., & Woods, S. W. (2007). Cost-efficacy of individual and combined treatments for panic disorder. Journal of Clinical Psychiatry, 68, 1038–1044.CrossRefPubMedGoogle Scholar
  52. Meyer, B., Pilkonis, P. A., Krupnick, J. L., Egan, M. K., Simmens, S. J., & Sotsky, S. M. (2002). Treatment expectancies, patient alliance and outcome: Further analyses from the National Institute of Mental Health Treatment of Depression Collaborative Research Program. Journal of Consulting and Clinical Psychology, 70, 1051–1055. doi: 10.1037//0022-006X.70.4.1051.CrossRefPubMedGoogle Scholar
  53. Montgomery, G. H., & Bovbjerg, D. H. (2004). Presurgery distress and specific response expectancies predict postsurgery outcomes in surgery patients confronting breast cancer. Health Psychology, 23, 381–387. doi: 10.1037/0278-6133.23.4.381.CrossRefPubMedGoogle Scholar
  54. Nathan, P. E., & Gorman, J. M. (Eds.). (2015). A guide to treatments that work (4th edn.). New York: Oxford University Press.Google Scholar
  55. Newman, M. G., Castonguay, L. G., Borkovec, T. D., Fisher, A. J., Boswell, J. F., Szkodny, L. E., & Nordberg, S. S. (2011). A randomized controlled trial of cognitive-behavioral therapy for generalized anxiety disorder with integrated techniques from emotion-focused and interpersonal therapies. Journal of Consulting and Clinical Psychology, 79, 171–181.CrossRefPubMedPubMedCentralGoogle Scholar
  56. Patterson, C. L., Uhlin, B., & Anderson, T. (2008). Clients’ pretreatment counseling expectation as predictors of the working alliance. Journal of Counseling Psychology, 55, 528–534.CrossRefPubMedGoogle Scholar
  57. Price, M., Anderson, P., Henrich, C., & Rothbaum, B. O. (2008). Greater expectations: Hierarchical linear modeling of expectancies as a predictor of treatment outcome. Behavior Therapy, 39, 398–405. doi: 10.1016/j.beth.2007.12.002.CrossRefPubMedPubMedCentralGoogle Scholar
  58. Price, M., & Anderson, P. L. (2012). Outcome expectancy as a predictor of treatment response in cognitive behavioral therapy for public speaking fears within social anxiety disorder. Psychotherapy, 49, 173–179. doi: 10.1037/a0024734.CrossRefPubMedGoogle Scholar
  59. Roth, P. L., Switzer, F. S., & Switzer, D. M. (1999). Missing data in multiple item scales: A Monte Carlo analysis of missing data techniques. Organizational Research Methods, 2(3), 211–232. doi: 10.1177/109442819923001.CrossRefGoogle Scholar
  60. Rutherford, B. R., Wager, T. D., & Roose, S. P. (2010). Expectancy and the treatment of depression: A review of experimental methodology and effects on patient outcome. Current Reviews in Psychiatry, 6, 1–10. doi: 10.2174/157340010790596571.CrossRefGoogle Scholar
  61. Shear, M. K., Vander Bilt, J., & Rucci, P. (2001). Reliability and validity of a structured interview guide for the Hamilton Anxiety Rating Scale (SIGH-A). Depression and Anxiety, 13, 166–178.CrossRefPubMedGoogle Scholar
  62. Thompson-Hollands, J., Bentley, K. H., Gallagher, M. W., Boswell, J. F., & Barlow, D. H. (2014). Credibility and outcome expectancy in the unified protocol for transdiagnostic treatment of emotional disorders. Journal of Experimental Psychopathology, 5, 72–82. doi: 10.5127/jep.033712.CrossRefGoogle Scholar
  63. Tracey, T. J., & Kokotovic, A. M. (1989). Factor structure of the working alliance inventory. Psychological Assessment, 1, 207–210.CrossRefGoogle Scholar
  64. Tsai, M., Ogrodniczuk, J. S., Sochting, I., & Mirmiran, J. (2014). Forecasting success: Patients’ expectations for improvement and their relations to baseline, process and outcome variables in group cognitive-behavioural therapy for depression. Clinical Psychology and Psychotherapy, 21, 97–107. doi: 10.1002/cpp.1831.CrossRefPubMedGoogle Scholar
  65. Vogel, P. A., Hansen, B., Stiles, T. C., & Götestam, K. (2006). Treatment motivation, treatment expectancy, and helping alliance as predictors of outcome in cognitive behavioral treatment of OCD. Journal of Behavior Therapy and Experimental Psychiatry, 37, 247–255. doi: 10.1016/j.jbtep.2005.12.001.CrossRefPubMedGoogle Scholar
  66. Webb, C. A., Kertz, S. J., Bigda-Peyton, J. S., & Björgvinsson, T. (2013). The role of pretreatment outcome expectancies and cognitive–behavioral skills in symptom improvement in an acute psychiatric setting. Journal of Affective Disorders, 149, 375–382. doi: 10.1016/j.jad.2013.02.016.CrossRefPubMedPubMedCentralGoogle Scholar
  67. Weinberger, L., & Eig, A. (1999). Expectancies: The ignored common factor in psychotherapy. In I. Kirsch (Ed.), How expectancies shape experience (pp. 357–382). Washington, DC: American Psychological Association. doi: 10.1037/10332-015.CrossRefGoogle Scholar
  68. Westra, H. A., Constantino, M. J., & Aviram, A. (2011). The impact of alliance ruptures on client outcome expectations in cognitive behavioral therapy. Psychotherapy Research, 21, 472–481. doi: 10.1080/10503307.2011.581708.CrossRefPubMedGoogle Scholar
  69. Zack, S. E., Castonguay, L. G., Boswell, J. F., McAleavey, A. A., Adelman, R., Kraus, D. R., & Pate, G. A. (2015). Attachment history as a moderator of the alliance-outcome relationship in adolescents. Psychotherapy, 52, 258–267.  10.1037/a0037727.CrossRefPubMedGoogle Scholar
  70. Zinbarg, R., Craske, M., & Barlow, D. H. (2006). Therapist’s guide for the mastery of your anxiety and worry program. New York, NY: Oxford University Press.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2017

Authors and Affiliations

  1. 1.Department of Psychological and Brain Sciences, Center for Anxiety and Related DisordersBoston UniversityBostonUSA
  2. 2.Department of PsychologyUniversity at Albany, State University of New YorkAlbanyUSA
  3. 3.Massachusetts General Hospital/Harvard Medical SchoolBostonUSA
  4. 4.National Center for PTSD, Behavioral Science Division/VA Boston Healthcare System and Boston University School of MedicineBostonUSA

Personalised recommendations