Psychotherapy 2015, Vol. 52, No. 1, 134 –139

© 2014 American Psychological Association 0033-3204/15/$12.00 http://dx.doi.org/10.1037/a0036890

Meta-Analysis of Client Attachment to Therapist: Associations With Working Alliance and Client Pretherapy Attachment Brent Mallinckrodt and JiSun Jeong

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

University of Tennessee A review was conducted of 14 studies published from 1995 to 2013 and 2 thus far unpublished studies that have used the Client Attachment to Therapist Scale (CATS) in research with help-seeking clients. Of these, meta-analyses were conducted for studies that examined the CATS subscales (Secure, Avoidant, Preoccupied) as correlates of client-rated working alliance, client pretherapy general adult attachment, or both (Total k ⫽ 13; k ⫽ 9 for each specific analysis). With regard to pretherapy adult attachment, CATS Secure was negatively associated with client Anxiety and Avoidance. For CATS Avoidant, clients’ attachment avoidance and anxiety were positively associated. For CATS Preoccupied, client adult attachment Anxiety was significantly positively associated. CATS Secure was strongly positively correlated with total working alliance, CATS Avoidant was negatively correlated with total working alliance, and CATS Preoccupied was not significantly associated with working alliance. Implications for models of therapeutic change based on client gains in social competencies and capacity for secure attachment are discussed. Keywords: client attachment to therapist, meta-analysis, working alliance, social competencies Supplemental materials: http://dx.doi.org/10.1037/a0036890.supp

Mallinckrodt (2010) has described how each of the five basic characteristics of a secure infant– caregiver attachment relationship identified by Mikulincer and Shaver (2007) are manifested as elements that can be recognized in a secure psychotherapy relationship, but also how clients with hyperactivating tendencies tend to magnify bids to establish the five basic elements, whereas clients with deactivating tendencies tend to downplay their importance and resist therapists’ efforts to establish these basic elements. For example, the CATS Secure subscale (14 items) taps perceptions of the counselor as a secure base for exploration (“My counselor helps me look closely at the frightening or troubling things that have happened to me”), a safe haven (“My counselor is a comforting presence to me when I am upset”), and as a close, responsive, and dependable figure (“My counselor is sensitive to my needs,” “I feel that somehow things will work out OK for me when I am with my counselor”). In contrast, the CATS Avoidant subscale taps clients’ deactivating resistance to developing safe haven, secure base, or proximity in the relationship (“It’s hard for me to trust my counselor,” “Talking over my problems with my counselor makes me feel ashamed or foolish”). Finally, CATS Preoccupied (10 items) taps clients’ hyperactivating tendencies in the forms of a strong need for proximity (“I wish my counselor could be with me on a daily basis”), a desire to blur professional boundaries (“I think about calling my counselor at home,” “I yearn to be ‘at one’ with my counselor”), and a preoccupation with other clients (“I think about being my counselor’s favorite client”). Thus, although the CATS was developed independently, and well before Mikulincer and Shaver’s (2007) formulation, the three CATS subscales appear to reflect general strategies of primary secure, and secondary deactivating or hyperactivating attachment.

In the 1980s, the rapidly expanding adult attachment literature included Bowlby’s (1988) highly influential book, A Secure Base, which described attachment aspects of the client–therapist relationship. To measure this construct, Mallinckrodt, Gantt, and Coble (1995) developed the Client Attachment to Therapist Scale (CATS). Items were generated by a panel of experienced therapists who were prompted only with descriptions of the three types of infant attachment patterns discovered in Ainsworth, Blehar, Waters, and Wall’s (1978) “strange situation” protocol. Nevertheless, the three subscales that emerged from a factor analysis of 138 clients’ responses appear to parallel contemporary models of adult attachment. (To distinguish these subscales from other measures with similar titles, we refer to them as CATS Secure, CATS Avoidant, and CATS Preoccupied.) When adults who have experienced a history of unavailable attachment figures find their primary bids for secure attachment blocked, they tend to engage in one of two insecure secondary strategies, either hyperactivating or deactivating their attachment behavioral system (Mikulincer & Shaver, 2007).

This article was published Online First November 3, 2014. Brent Mallinckrodt and JiSun Jeong, Department of Psychology, University of Tennessee. This article is dedicated to the career and scientific contributions of Charles Gelso, whose influence as a mentor and psychotherapy researcher is directly responsible for much of the work reviewed in this meta-analysis. We also gratefully acknowledge the assistance of Professor Clara Hill and Joseph Miles in the preparation of this article. Correspondence concerning this article should be addressed to Brent Mallinckrodt, Department of Psychology, University of Tennessee, 1404 Circle Dr., Rm. 305, Knoxville, TN 37996. E-mail: [email protected] 134

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

CLIENT ATTACHMENT TO THERAPIST

Research suggests that adults who experienced very different circumstances of family disruption/dysfunction, but nevertheless share the same memories of emotionally unavailable parents, develop similar problems with attachment security, object relations, and alexithymia (Hadley, Holloway, & Mallinckrodt, 1993; King & Mallinckrodt, 2000). These findings highlight two distinct approaches to measuring adult attachment based on (a) adults’ memories of family experiences, versus (b) self-reports of recent close relationships. Research in the first tradition often relies on the Adult Attachment Interview (AAI, Main & Goldwyn, 1998) to assess current attachment state of mind regarding childhood experiences. The second tradition uses self-report measures such as the Experiences in Close Relationship Scale (ECRS, Brennan Clark, & Shaver, 1998). Roisman et al. (2007) reviewed studies that used both methods with the same subjects, and reported low correspondence, Pearson r ⫽ .09. Lack of agreement might be due to different measurement goals of the two traditions. Researchers use the AAI to assess coherence of narrative memories, and assert that their technique reveals unconscious defensive patterns (Main & Goldwyn, 1998). In contrast, self-report measures like the ECRS assess conscious cognition and affect related to levels of anxiety and avoidance in current relationships. Three recent meta-analyses have contributed to a better understanding of how adult client attachment affects psychotherapy. Diener and Monroe (2011) found a weighted mean effect size across 17 samples of r ⫽ .17 (95% CI ⫽ .10 –.23) for positive associations of client attachment security (or negative associations of insecurity) with stronger therapeutic alliance. In a meta-analysis that included 12 unpublished dissertations as well as 12 published articles, Bernecker, Levy, and Ellison (2013) reported a mean weighted r of ⫺.13 between adult attachment avoidance and working alliance, and r ⫽ ⫺.12 for attachment anxiety and alliance. Levy, Ellison, Scott, and Bernecker (2011), contributed another important advance through meta-analysis of 14 studies that measured associations between pretherapy client attachment and eventual therapy outcome. Attachment anxiety exhibited a mean weighted r with outcome of ⫺.22, whereas the mean weighted r for avoidance was ⫺.01. Thus, a growing body of evidence suggests that clients’ attachment anxiety before therapy is negatively associated with both alliance and outcome, whereas clients’ pretherapy avoidance appears to be related negatively with working alliance, but not outcome. Note that all of the studies reviewed in these three meta-analyses used self-report measures of attachment. Thus, the self-report tradition is much better represented than the AAI tradition in research involving the psychotherapy relationship and therapy outcome. Although the article introducing the CATS (Mallinckrodt et al., 1995) has been cited ⬎100 times to date, we could find no systematic meta-analysis or other review of this research. Such a synthesis could add an important, complementary perspective to the three previous meta-analyses. For example, significant associations between client pretherapy adult attachment and specific CATS subscale scores could be interpreted as empirical evidence of transference. Thus, the purpose of this study was to conduct a meta-analysis of studies that have explored associations of client attachment to therapist (as measured by the CATS), with client pretherapy attachment patterns and with working alliance.

135 Method

The Web of Science and PsycINFO databases were used to identify 106 published studies and five dissertations that cited Mallinckrodt et al. (1995) or were indexed under “client attachment to therapist.” The final search was conducted on January 25, 2014. In addition, a call for unpublished studies was broadcast on the Society for Psychotherapy Research listserv. Two additional studies were identified (Mallinckrodt, Choi, & Daly, in press; Wiseman & Tishby, 2014). Abstracts for all 113 studies were reviewed. Of these, eight were not published in English, 36 were not empirical studies, 38 were empirical but did not use the CATS, one dissertation was subsequently published as an article, 14 empirical studies were qualitative or did not study clients in therapy. After screening, 97 studies were eliminated based on these criteria, leaving 16 empirical, English language studies that used the CATS to study bona fide clients seeking help. Of the 16 studies, five measured client pretherapy attachment and working alliance, four measured working alliance but not pretherapy attachment, and four measured pretherapy client attachment but not working alliance (see Table S1 in the online supplemental). Only these 13 studies were included in meta-analyses. Most studies used the ECRS (Brennan et al., 1998) to assess client pretherapy attachment. The ECRS has two subscales, Avoidance and Anxiety, coded so that higher scores indicate more attachment difficulties. When subscales of other measures indicated more security (e.g., the Close and Depend subscales of the Adult Attachment Scale), for consistency with the ECRS, the valence of these correlations was reversed. The product-moment correlation (r) was adopted to index effect size in these metaanalyses. Three studies did not report CATS zero-order correlations (Janzen, Fitzpatrick, & Drapeau, 2008; Moore & Gelso, 2011; Skourteli & Lennie, 2011). After a personal communication, all provided bivariate correlation data. Following recommendations of Borenstein, Hedges, Higgins, and Rothstein (2009), r correlation coefficients were transformed into Fisher’s Z before aggregation, and then transformed back into r. A random effects model was adopted. Data were analyzed using the Comprehensive Meta-Analysis software Version 2.0 (Borenstein, Hedges, Higgins, & Rothstein, 2013). When more than one effect was reported (e.g., Janzen et al. (2008) collected data after each of the first four sessions), a weighted average r was calculated so that each study contributed only one effect to a given meta-analysis.

Results All ratings of working alliance were provided by clients. A total of 1051 client–therapist dyads were surveyed in these 13 studies (M ⫽ 80.85, SD ⫽ 50.34, range 27–197). Of these clients, 67% were female, and 83% of those whose race/ethnicity was identified were White. In terms of treatment setting, 47.2% of clients were seen at college counseling centers, 6.5% at a Veteran’s Administration hospital, 1.7% at other hospital settings, 14.9% at university training clinics, 5.2% in independent practice, 2.3% in community agencies, and the settings of 22.2% could not be determined (e.g., data collected via Internet). Table 1 presents results of the meta-analyses separately for the three CATS subscales, 95% confidence intervals, Z, and p values for each effect and the weighted mean values for each group of studies. Results suggest that both dimensions of insecure client

MALLINCKRODT AND JEONG

136

Table 1 Studies and Component Effect Sizes Included In Meta-Analyses 95% Confidence interval

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Study CATS Secure and Working Alliance Bachelor et al. (2010) Fuertes et al. (2007) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Martin (2007) Romano et al. (2008) Peterman (2010) Sauer et al. (2010) Weighted Mean CATS Avoidant and Working Alliance Bachelor et al. (2010) Fuertes et al. (2007) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Martin (2007) Romano et al. (2008) Peterman (2010) Sauer et al. (2010) Weighted Mean CATS Preoccupied and Working Alliance Bachelor et al. (2010) Fuertes et al. (2007) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Martin (2007) Romano et al. (2008) Peterman (2010) Sauer et al. (2010) Weighted Mean CATS Secure and Client Pretherapy Avoidance Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean Trim and Fill Weighted Mean CATS Avoidant and Client Pretherapy Avoidance Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean Trim and Fill Weighted Mean CATS Preoccupied and Client Pretherapy Avoidance Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean

N

r

Lower

Upper

80 59 37–50 138 38 197 59 68 93 —

.73 .22 .90 .82 .69 .84 .81 .69 .72 .76

.60 ⫺.04 .82 .76 .48 .79 .70 .54 .60 .69

80 59 37–50 138 38 197 59 68 93 —

⫺.64 ⫺.47 ⫺.81 ⫺.56 ⫺.62 ⫺.64 ⫺.61 ⫺.68 ⫺.52 ⫺.63

80 59 37–50 138 38 197 59 68 93 —

Z

p

.82 .45 .94 .87 .83 .88 .88 .80 .81 .82

8.09 1.67 9.18 13.44 5.02 17.01 8.43 6.84 8.61 22.68

.00 .09 .00 .00 .00 .00 .00 .00 .00 .00

⫺.76 ⫺.65 ⫺.89 ⫺.66 ⫺.78 ⫺.72 ⫺.75 ⫺.79 ⫺.65 ⫺.69

⫺.49 ⫺.24 ⫺.67 ⫺.43 ⫺.38 ⫺.55 ⫺.42 ⫺.53 ⫺.35 ⫺.57

⫺6.70 ⫺3.82 ⫺7.10 ⫺7.35 ⫺4.29 ⫺10.56 ⫺5.30 ⫺6.68 ⫺5.47 ⫺19.40

.00 .00 .00 .00 .00 .00 .00 .00 .00 .00

⫺.14 .24 .32 .07 ⫺.13 ⫺.01 ⫺.29 .16 .01 .02

⫺.35 ⫺.02 .02 ⫺.10 ⫺.43 ⫺.15 ⫺.51 ⫺.08 ⫺.19 ⫺.09

.08 .47 .57 .23 .20 .13 ⫺.04 .38 .21 .14

⫺1.27 1.83 2.08 0.82 ⫺0.77 ⫺0.14 ⫺2.23 1.30 0.10 0.44

.20 .07 .04 .42 .44 .89 .02 .19 .92 .66

30 37–50 133 38 143 59 93 27 54–67 — —

⫺.30 ⫺.02 ⫺.10 ⫺.29 ⫺.13 ⫺.21 ⫺.01 ⫺.23 ⫺.21 ⫺.14 ⴚ.12

⫺.59 ⫺.32 ⫺.27 ⫺.56 ⫺.29 ⫺.44 ⫺.21 ⫺.56 ⫺.44 ⴚ.22 ⴚ.19

.07 .28 .07 .03 .04 .05 .19 .17 .05 ⴚ.06 ⴚ.05

⫺1.59 ⫺0.16 ⫺1.15 ⫺1.77 ⫺1.55 ⫺1.60 ⫺0.10 ⫺1.13 ⫺1.61 ⴚ3.52 ⴚ3.18

.11 .88 .25 .08 .12 .11 .92 .26 .11 .00 .001

30 37–50 133 38 143 59 93 27 54–67 — —

.47 .15 ⫺.13 .38 .27 .13 .05 .29 .36 .20 .18

.13 ⫺.16 ⫺.29 .07 .11 ⫺.13 ⫺.16 ⫺.10 .11 .07 .04

.71 .43 .04 .62 .42 .37 .25 .60 .56 .34 .31

2.66 0.92 ⫺1.49 2.37 3.28 0.98 0.48 1.47 2.80 2.93 2.58

.01 .36 .14 .02 .00 .33 .64 .14 .00 .003 .01

30 37–50 133 38 143 59 93 27 54–67 —

.01 .13 ⫺.07 ⫺.02 ⫺.04 ⫺.03 .00 .22 .17 .01

⫺.35 ⫺.18 ⫺.24 ⫺.34 ⫺.20 ⫺.28 ⫺.20 ⫺.18 ⫺.09 ⴚ.07

.37 .41 .10 .30 .12 .23 .20 .55 .41 .09

0.04 0.80 ⫺0.81 ⫺0.12 ⫺0.47 ⫺0.22 0.00 1.08 1.30 0.25

r adjusted

95% CI

p

No. studies

Q

p

I2

NT





9

59.04

.00

46.78

NT





9

13.18

.11

39.30

NT





9

17.04

.03

53.06

⫺.12

(⫺.19, ⫺.05)

.001

9

4.67

.79

(.04, .31)

.01

9

22.41

.00





9

4.40

.82

.18

NT .97 .42 .42 .91 .64 .82 1.00 .28 .20 .80

0

64.30

0

CLIENT ATTACHMENT TO THERAPIST

137

Table 1 (continued) 95% Confidence interval

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Study CATS Secure and Client Pretherapy Anxiety Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean CATS Avoidant and Client Pretherapy Anxiety Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean Trim and Fill Weighted Mean CATS Preoccupied and Client Pretherapy Anxiety Janzen et al. (2008) Mallinckrodt et al. (in press) Mallinckrodt et al. (1995) Mallinckrodt et al. (2005) Moore & Gelso (2011) Romano et al. (2008) Sauer et al. (2010) Skourteli & Lennie (2011) Wiseman & Tishby (2014) Weighted Mean Trim and Fill Weighted Mean

N

r

Lower

Upper

Z

p

30 37–50 133 38 143 59 93 27 54–67 —

⫺.18 ⫺.26 ⫺.10 ⫺.31 ⫺.04 .20 ⫺.16 ⫺.29 ⫺.21 ⴚ.13

⫺.51 ⫺.52 ⫺.27 ⫺.57 ⫺.20 ⫺.06 ⫺.35 ⫺.60 ⫺.44 ⴚ.22

.19 .05 .07 .01 .12 .43 .04 .10 .04 ⴚ.03

⫺0.96 ⫺1.67 ⫺1.14 ⫺1.90 ⫺0.47 1.52 ⫺1.53 ⫺1.44 ⫺1.62 ⴚ2.62

.34 .10 .25 .06 .64 .13 .13 .15 .11 .009

30 37–50 133 38 143 59 93 27 54–67 — —

.29 .35 .17 ⫺.29 .11 .02 .12 .41 .38 .17 .15

⫺.08 .05 .00 ⫺.56 ⫺.06 ⫺.24 ⫺.09 .03 .14 .05 .03

.59 .59 .33 .03 .27 .28 .32 .68 .58 .29 .27

1.57 2.30 1.96 ⫺1.77 1.31 0.15 1.14 2.11 2.99 2.80 2.52

.12 .02 .05 .08 .19 .88 .25 .04 .00 .005 .012

30 37–50 133 38 143 59 93 27 54–67 — —

.25 ⫺.01 ⫺.01 .24 .08 .12 .06 .44 .31 .14 .07

⫺.12 ⫺.31 ⫺.18 ⫺.09 ⫺.08 ⫺.14 ⫺.14 .08 .06 .04 -.04

.56 .29 .16 .52 .24 .36 .26 .71 .52 .23 .18

1.30 ⫺0.06 ⫺0.11 1.45 0.95 0.90 0.57 2.35 2.43 2.79 1.20

.19 .95 .91 .15 .34 .37 .57 .02 .02 .005 .228

r adjusted

95% CI

p

No. studies

Q

p

I2

NT





9

10.51

.23

23.91

.15

(.03, .27)

.01

9

16.12

.04

50.36

.07

(⫺.04, .18)

.23

9

9.80

.28

18.34

Note. NT ⫽ results indicated no need to trim any studies, and so the observed and adjusted effect sizes are identical.

attachment (Anxiety and Avoidance) were significantly negatively associated with CATS Secure, and positively associated with CATS Avoidant. In contrast, only client pretherapy anxious attachment (but not avoidant) was significantly positively associated with CATS Preoccupied. These findings have significant implications for an understanding of client transference. With regard to working alliance, CATS Secure had a very strong effect (r ⫽ .76, p ⬍ .001), whereas the effect for CATS Avoidant was also strong but in a negative direction (r ⫽ ⫺.63, p ⬍ .001). The effect size for CATS Preoccupied with working alliance was not significant (r ⫽ .02, p ⫽ .70). To assess the effects of possible publication bias, funnel plots were constructed using Comprehensive Meta-Analysis software Version 2.0 (Borenstein et al., 2013), followed by the “trim and fill” procedure (Duval & Tweedie, 2000) available through the metatrim macro (Steichen, 2009) developed for the STATA software package. The trim and fill procedure adjusts for the potential biasing impact of small-sample studies that have extreme results. When publication bias is suggested, this procedure estimates the effect size that might be expected if the hypothetical missing studies had been included in the meta-analyses. In five of the nine analyses, the Metatrim quantitative results indicated there was no evidence of publication bias. Table 1 shows results of the remaining four “Trim and Fill” analyses. The attenuated (i.e. adjusted) values of weighted r, 95% CI, and Z are shown below the standard

values. Note that for three of the four analyses, after accounting for the possible effects of publication bias, the attenuated (i.e. adjusted) r remained significant. Only for the effect of client pretherapy Anxious attachment on CATS Preoccupied scores did the possibility of selective publication reduce the effect below significant levels (i.e., from r ⫽ .14 to r adjusted ⫽ .07). Supplemental online Table S2 presents Q and I2 indices resulting from tests for heterogeneity of effect size. Results suggest that effect sizes for CATS Secure and Preoccupied in connection with working alliance are not uniform across the nine studies, and neither are effects for CATS Avoidant in connection with client pretherapy avoidant attachment.

Discussion Important limitations of these meta-analyses must be recognized. The number of studies is relatively small (Total k ⫽ 13; k ⫽ 9 for each specific analysis). Except for Fuertes et al. (2007) who collected ratings from both clients and therapists for the same construct (Real Relationship), all other studies relied exclusively on self-report data from clients as a single source, which invites the problems of mono-method bias. As with all meta-analyses, this study is susceptible to the effects of publication and selective retrieval biases. A further caution stems from the possibility that beliefs about secure attachment have a strong cultural basis (Wang

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

138

MALLINCKRODT AND JEONG

& Mallinckrodt, 2006). Finally, we encountered problems with incomplete data. Reports in the primary research of partial correlations and composite variables made computing effects sizes difficult. We are grateful to the authors who responded to our requests for their data. The purpose of this study was to conduct a meta-analysis of research that has used the CATS to examine the attachment relationship of clients to their therapist. The current study joins three other meta-analyses of client attachment, two in association with therapeutic alliance (Bernecker et al., 2013; Diener & Monroe, 2011), and one with therapy outcome (Levy et al., 2011). Interestingly, all found roughly the same number of published studies (12–17), although Bernecker et al. (2013) located 12 additional unpublished studies. Despite the modest number of studies, aggregate effects summarized here do provide empirical support for some important theoretical assumptions. First, it appears that clients’ pretherapy attachment tendencies do influence the quality of attachment they establish in the psychotherapy relationship. As attachment theory and the conceptualizations of transference would predict, pretherapy client anxiety is significantly associated with CATS Preoccupied attachment, and thus perhaps hyperactivation in the psychotherapy relationship. Also as predicted, client Anxiety and Avoidance both seem to interfere with development of a Secure therapeutic attachment. An unexpected aspect of these findings is that pretherapy client Anxiety as well as Avoidance were both associated with CATS Avoidance. Predictions based on client transference lead to an expected connection only for client pretherapy Avoidance—not Anxiety—with CATS Avoidance. It should be noted that I2 levels for these latter two effects are among the highest in this study, suggesting that perhaps the setting, type of client, and stage of therapy may serve as moderating variables for how pretherapy client attachment insecurity influences avoidance in the psychotherapy relationship. These findings are consistent with Bowlby’s (1988) secure base conception of attachment in psychotherapy. This model is concerned with the facilitative processes that coevolve with secure attachment in psychotherapy, and the positive changes in clients that follow from the therapeutic tasks Bowlby describes. In contrast, the social competencies attachment model places more focus on the antecedent capacities that clients (and therapists) bring to the relationship (Mallinckrodt, 2000, 2010; Mallinckrodt et al., in press). The two models are complementary. Following closely from interpersonal approaches to therapy (cf. Teyber & McClure, 2010), the social competencies model emphasizes how family dysfunction and maladaptive developmental experiences interfere with adults’ ability to develop the social competencies needed to establish secure attachments. For example, Swanson and Mallinckrodt (2001) found that memories of childhood sexual abuse, as well as a parenting style characterized by withdrawal of affection as a method of discipline were associated with adult attachment avoidance and anxiety. These competency deficits may be at the root of many clients’ presenting problems, and can pose formidable challenges to a therapist’s best efforts to foster a secure attachment (King & Mallinckrodt, 2000). From the social competencies perspective, most of the work of change is accomplished in struggling to build a secure psychotherapy attachment, not after it is established (Daly & Mallinckrodt, 2009; Mallinckrodt, 2010; Mallinckrodt et al., in press).

Findings of this meta-analysis support the social competencies model. Pretherapy client attachment difficulties were associated with poor psychotherapy attachment. Other findings involving the CATS that could not be summarized in these meta-analyses, also support the social competencies model. For example, difficulties in client attachment to therapist have been associated with client interpersonal presenting symptoms (Bachelor et al., 2010), adult object relations and self-efficacy deficits (Mallinckrodt et al., 1995), memories of family dysfunction Mallinckrodt et al., 1998), and poor parenting (Woodhouse et al., 2003). Two studies reported that client-rated Real Relationship was positively associated with CATS Secure, negatively associated with CATS Avoidant, and not significantly related to CATS Preoccupied (Fuertes et al., 2007; Moore & Gelso, 2011). These findings are in close agreement with findings about attachment difficulties and social competencies deficits in mixed samples of clients and nonclients (King & Mallinckrodt, 2000) and nonclient adults (Mallinckrodt & Wei, 2005; Swanson & Mallinckrodt, 2001). Taken together the social competencies model explains how developmental experiences pose barriers to secure attachment in psychotherapy (Mallinckrodt, 2000, 2010), and the secure base model explains how skillful therapists overcome these challenges to promote change (Bowlby, 1988; Daly & Mallinckrodt, 2009). Results of the current metaanalysis provide support for both of these complementary models, but also point to the need for much more research.

References References marked with an asterisk indicate studies included in the meta-analysis. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum. ⴱ Bachelor, A., Meunier, G., Laverdiére, O., & Gamache, D. (2010). Client attachment to therapist: Relation to client personality and symptomatology, and their contributions to the therapeutic alliance. Psychotherapy, 47, 454 – 468. doi:10.1037/a0022079 Bernecker, S., Levy, K. N., & Ellison, W. D. (2013). A meta-analysis of the relation between patient adult attachment style and the working alliance. Unpublished manuscript. Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. Chichester, UK: Wiley. doi:10.1002/ 9780470743386 Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2013). Manual: Comprehensive Meta-Analysis, Ver. 2. Retrieved from http:// www.meta-analysis.com/pages/downloads/Meta-Analysis-Manual.pdf Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. London: Routledge. Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46 –76). New York: Guilford Press. Daly, K. D., & Mallinckrodt, B. (2009). Expert therapists’ approaches to psychotherapy with adult clients who present with attachment avoidance or anxiety. Journal of Counseling Psychology, 56, 549–563. doi:10.1037/a0016695 Diener, M. J., & Monroe, J. M. (2011). The relationship between adult attachment style and therapeutic alliance in individual psychotherapy: A meta-analytic review. Psychotherapy, 48, 237–248. doi:10.1037/ a0022425 Duval, S., & Tweedie, R. (2000). A nonparametric “trim and fill” method of accounting for publication bias in meta-analysis. Journal of the American Statistical Association, 95, 89 –98.

CLIENT ATTACHMENT TO THERAPIST

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.



Fuertes, J. N., Mislowack, A., Brown, S., Gur-Arie, S., Wilkinson, S., & Gelso, C. J. (2007). Correlates of the real relationship in psychotherapy: A study of dyads. Psychotherapy Research, 17, 423– 430. doi:10.1080/ 10503300600789189 Hadley, J. A., Holloway, E. L., & Mallinckrodt, B. (1993). Common aspects of object relations and self-representations in offspring from disparate dysfunctional families. Journal of Counseling Psychology, 40, 348 –356. doi:10.1037/0022-0167.40.3.348 ⴱ Janzen, J., Fitzpatrick, M., & Drapeau, M. (2008). Processes involved in client-nominated relationship building incidents: Client attachment, attachment to therapist, and session impact. Psychotherapy, 45, 377–390. doi:10.1037/a0013310 King, J. L., & Mallinckrodt, B. (2000). Family environment and alexithymia in clients and non-clients. Psychotherapy Research, 10, 78 – 86. doi:10.1080/713663595 Levy, K. N., Ellison, W. D., Scott, L. N., & Bernecker, S. L. (2011). Attachment style. Journal of Clinical Psychology, 67, 193–203. doi: 10.1002/jclp.20756 Main, M., & Goldwyn, R. (1998). Adult Attachment Scoring and Classification Systems. Unpublished manuscript, University of California at Berkeley. Mallinckrodt, B. (2000). Attachment, social competencies, social support and interpersonal process in psychotherapy. Psychotherapy Research, 10, 239 –266. doi:10.1093/ptr/10.3.239 Mallinckrodt, B. (2010). The psychotherapy relationship as attachment: Evidence and implications. Journal of Social and Personal Relationships, 27, 262–270. doi:10.1177/0265407509360905 ⴱ Mallinckrodt, B., Choi, G., & Daly, K. D. (in press). Pilot test of a measure to assess therapeutic distance and its association with client attachment and corrective experience in therapy. Psychotherapy Research. ⴱ Mallinckrodt, B., Gantt, D. L., & Coble, H. M. (1995). Attachment patterns in the psychotherapy relationship: Development of the Client Attachment to Therapist Scale. Journal of Counseling Psychology, 42, 307–317. doi:10.1037/0022-0167.42.3.307 Mallinckrodt, B., King, J. L., & Coble, H. M. (1998). Family dysfunction, alexithymia, and client attachment to therapist. Journal of Counseling Psychology, 45, 497–504. doi:10.1037/0022-0167.45.4.497 ⴱ Mallinckrodt, B., Porter, M. J., & Kivlighan, D. M., Jr. (2005). Client attachment to therapist, depth of in-session exploration, and object relations in brief psychotherapy. Psychotherapy, 42, 85–100. doi: 10.1037/0033-3204.42.1.85 Mallinckrodt, B., & Wei, M. (2005). Attachment, social competence, social support, and psychological distress. Journal of Counseling Psychology, 52, 358 –367. doi:10.1037/0022-0167.52.3.358 ⴱ Martin, J. (2007). Male clients’ perceptions of positive therapeutic alliance. Dissertation Abstracts International. Section B-, 68, 6972.

139

Mikulincer, M., & Shaver, P. R. (2007). Attachment in adulthood: Structure, dynamics, and change. New York, NY: Guilford Press. ⴱ Moore, S. R., & Gelso, C. J. (2011). Recollections of a secure base in psychotherapy: Considerations of the real relationship. Psychotherapy, 48, 368 –373. doi:10.1037/a0022421 ⴱ Peterman, T. L. (2010). Attachment to therapist, the working alliance and emotional processing of traumatic material in session among veterans diagnosed with posttraumatic stress disorder. Dissertation Abstracts International: Section B-, 73, 1852. Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., & Clarke, A. (2007). The Adult Attachment Interview and self-reports of attachment style: An empirical rapprochement. Journal of Personality and Social Psychology, 92, 678 – 697. doi:10.1037/0022-3514.92.4.678 ⴱ Romano, V., Fitzpatrick, M., & Janzen, J. (2008). The secure-base hypothesis: Global attachment, attachment to counselor, and session exploration in psychotherapy. Journal of Counseling Psychology, 55, 495– 504. doi:10.1037/a0013721 ⴱ Sauer, E. M., Anderson, M. Z., Gormley, B., Richmond, C. J., & Preacco, L. (2010). Client attachment orientations, working alliances, and responses to treatment: A psychology training clinic study. Psychotherapy Research, 20, 702–711. doi:10.1080/10503307.2010.518635 ⴱ Skourteli, M. C., & Lennie, C. (2011). The therapeutic relationship from an attachment theory perspective. Counselling Psychology Review, 26, 20 –33. Steichen, T. J. (2009). Nonparametric trim and fill analysis of publication bias in meta analysis. In J. A. C. Sterne (Ed.), Meta-analysis in Stata: An updated collection from the Stata Journal (pp. 165–177). College Station, TX: Stata Press. Swanson, B., & Mallinckrodt, B. (2001). Family environment, love withdrawal, childhood sexual abuse, and adult attachment. Psychotherapy Research, 11, 455– 472. doi:10.1080/713664062 Teyber, E., & McClure, F. (2010). Interpersonal process in psychotherapy: An integrated model. Belmont, CA: Brooks/Cole. Wang, C. C., & Mallinckrodt, B. (2006). Differences between Taiwanese and U.S. cultural beliefs about ideal adult attachment. Journal of Counseling Psychology, 53, 192–204. doi:10.1037/0022-0167.53.2.192 ⴱ Wiseman, H., & Tishby, O. (2014). Client Attachment, Attachment to Therapist and Client-Therapist Attachment Match: How do they relate to Change in Psychodynamic Psychotherapy? Psychotherapy Research, 24, 392– 406. doi:10.1080/10503307.2014.892646 Woodhouse, S. S., Schlosser, L. Z., Crook, R. E., Ligiéro, D. P., & Gelso, C. J. (2003). Client Attachment to Therapist: Relations to transference and client recollections of parental caregiving. Journal of Counseling Psychology, 50, 395– 408. doi:10.1037/0022-0167.50.4.395

Received January 1, 2014 Revision received April 4, 2014 Accepted April 7, 2014 䡲

Meta-analysis of client attachment to therapist: associations with working alliance and client pretherapy attachment.

A review was conducted of 14 studies published from 1995 to 2013 and 2 thus far unpublished studies that have used the Client Attachment to Therapist ...
79KB Sizes 1 Downloads 7 Views