Incongruence Between Implicit Attachment Schemes and Unconscious Attachment Representations

ORIGINAL ARTICLE Incongruence Between Implicit Attachment Schemes and Unconscious Attachment Representations Katja Petrowski, PhD,*† Susan Schurig,* Helmut Kirchmann, PhD,‡ Sashi Singh,‡ Rainer Banse, PhD,§ Roland Imhoff, PhD,|| and Bernhard Strauss, PhD‡ Abstract: Assessments based on reaction time and language-based interviews postulate that unconscious attachment processes be measured. Nevertheless, a possible empirical equivalence of these two approaches has not yet been investigated. To fill this void, the Adult Attachment Interview and the Implicit Association Test were implemented with a group of patients with panic disorder (n = 157, mean age = 29, SD = 2.47) based on the Structured Clinical Interview for DSM-IV, axis I and II disorders and a group of healthy individuals (n = 138). In total, the securely attached individuals showed significantly more positive attitudes toward their mother than the insecurely attached individuals. In the healthy individuals, the secure and disorganized classifications showed significantly more positive attitudes toward the mother in comparison with the insecure attachment classification, as well as the patient group. In summary, implicit attachment patterns based on reaction times are not equivalent to an attachment representation based on language markers. For the disorganized attachment representation, no differences were present between the information processing of the memory/association network and the autobiographic memory function. Key Words: Attachment, adult attachment interview, implicit association test, panic disorder (J Nerv Ment Dis 2019;207: 423–428) I n attachment research, different research lines have established themselves from independently developed methodologies for assessing attachment (interviews versus questionnaires/latency-based methods). In the developmental psychology tradition, detailed language-based analyses of interview transcripts promised insight into the underlying unconscious attachment representation. In contrast, social cognitive research was inspired by latency-based methods for the assessment of automatic and implicit cognitive processes. This parallel development constitutes a major barrier to advancing our understanding of *Department of Psychotherapy and Psychosomatic Medicine, Technische Universität Dresden, University Hospital Carl Gustav Carus, Dresden; †Medical Psychology & Medical Sociology, University Medical Center, Johannes Gutenberg University, Mainz; ‡Institute of Psychosocial Medicine and Psychotherapy, University Hospital Jena, Jena; §Department of Psychology, University of Bonn, Bonn; and ||Department of Psychology, Johannes Gutenberg University, Mainz, Germany. Send reprint requests to Katja Petrowski, PhD, Department of Psychotherapy and Psychosomatic Medicine, Technische Universität Dresden, University Hospital Carl Gustav Carus, Fetscherstraße 74, 01307 Dresden, Germany. E‐mail: katja.petrowski@tu-dresden.de. Sampling statement: We described how the sample size was determined. We also disclose any data exclusions and explained the rationale for these exclusions. Open material statement: We provide information regarding all procedures applied and all measures assessed in this study. Open data statement: The data needed to reproduce the results are open. The data are accessible upon inquiry by contacting the corresponding author. Reproducible script statement: Data analysis scripts allowing reproduction of the reported results are accessible upon inquiry by contacting the corresponding author. Effects statement: We report basic descriptive statistics, effect sizes, exact p values, and 95% confidence intervals. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0022-3018/19/20706–0423 DOI: 10.1097/NMD.0000000000000987 unconscious/implicit relational schemata. Therefore, it is imperative to test whether these two approaches converge, that is, whether they measure the same construct. In developmental psychology, the unconscious aspects of the attachment representation were investigated using the Adult Attachment Interview (AAI; George C, Kaplan N, Main M Attachment interview for adults [unpublished manuscript]). The interview broaches the descriptions of attachment-relevant experiences with parental figures in early childhood. Based on the language analyses of the transcript, it measures the current unconscious state of the cognitive emotional processing of attachment experiences (“states of mind with respect to attachment”). This method takes defense processes such as idealizations into consideration (George C, Kaplan N, Main M Attachment interview for adults [unpublished manuscript]). Main and Solomon (1986) observed a breakdown of defense processes in attachment trauma and proposed the development of a segregated system. Segregated attachment material, then, leads to context-free, uncontrollable, and idiosyncratic language. Thus, the unresolved trauma/disorganized attachment representation can be detected by these language-based methods (Main M, Goldwyn R, Hesse E Adult Attachment scoring and classification system [unpublished manuscript]). In social psychology, the advancement of latency-based approaches was used to measure unconscious attachment aspects, that is, the attachment-specific Implicit Association Test (IAT; Greenwald et al., 1998). Given an automatic association of the partner with positive attributes, the presentation of positive terms with partner characteristics should be easier to validate, that is, responses are faster and less error prone than negative terms. Individual differences in the magnitude of this effect can then be interpreted as indicators of how positive the partner schema is (Banse and Kowalick, 2007; Banse et al., 2013; Zayas and Shoda, 2005). The IAT obtains access to the deeper, “implicit” knowledge structure of relationships (for an overview, see Banse and Imhoff, 2013). The nonaccessibility and unconsciousness of these associations by the participants have recently come under investigation (Hahn and Gawronski, 2014; Hahn et al., 2014). Most people know their implicit attitudes (content awareness) and their respective origin (source awareness) but are less aware of their effect (impact awareness; Gawronski et al., 2006). Therefore, the measurement outcomes are less controllable, less accessible, and more automatic than the plainly selfreported content. The associative connections between implicit schemata or attachment representations are actually related to the secure basescript concept developed by Waters and Waters (2006). Based on the secure-based script, stress leads the partner to seek help and his or her availability recedes. The existence of such a script was investigated in an experimental study by Mikulincer et al. (2009). Attachment style– specific “scriptedness” in narratives and stress-induced dreams could be observed by schema-consistent completion. Additional proof of automatic processes is reflected by a better processing and recollection of attachment-related/consistent material independent of cognitive load manipulation (Mikulincer et al., 2009). These results indicate that the proposed associative connections between representational structures and implicit attitudes do exist. The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 423 The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Petrowski et al. To answer the question concerning an association between implicit and unconscious and attachment representations, the IAT as the latency-based implicit attachment measure and the AAI as the method for determining language-based unaware contents were applied in the current study. This way, the coherence or rather the divergence of both measurements can be specified. In case of a convergence, an economic method (IAT) for differentiating attachment, especially unresolved trauma/disorganized attachment in clinical samples, would be available to research and psychotherapy. Because both methods claim to measure unconscious attachment and an associative connection was suggested by Mikulincer et al., 2009, a) it can be proposed that the secure attachment classification shows positive implicit attitudes toward the partner/ mother and shorter reaction times as compared with the insecure attachment classification. The basis for the unresolved trauma/disorganized attachment is the breakdown of the defense processes and the existence of a segregated system that interferes with language processing. It is unclear whether this system also interferes with the retrieval of associative connections and their reaction times. Based on the notion that an attachment trauma causes disorganized attachment, b) it can be proposed that the unresolved trauma/disorganized attachment representations show negative implicit attitudes toward the partner/mother and longer reaction times than the secure or insecure attachment does. Because of the high prevalence of unresolved trauma/disorganized attachment in a clinical sample (Dozier et al., 2008), the differentiation of psychopathology by the IAT was investigated in a clinical and in a healthy sample. METHODS Design/Procedure The data of this study were obtained within a collaborative study of the Institute of Psychosocial Medicine and Psychotherapy at the University Hospital Jena and the Clinic for Psychotherapy and Psychosomatics at the University Hospital Carl Gustav Carus at the Technische Universität Dresden. The AAI and a latency-based measure, the IAT, were implemented in a clinical and a nonclinical sample (cf. Strauss et al., in prep). The data collection took place between October 2011 and April 2014. For the clinical sample, patients were recruited from an outpatient clinic. Simultaneously, the healthy sample was recruited according to matching age, sex, and educational level. To obtain a very homogeneous sample of patients, only patients between 18 and 65 years of age with a primary panic disorder with or without agoraphobia or with agoraphobia without panic disorder in their history (ICD-10 diagnoses F40.01, F40.1, and F41.0) were included in the study. This specific sample was selected pragmatically because patients with these diagnoses constituted the largest (homogeneous) diagnostic subgroup in the participating centers. The patients were also permitted to have a secondary diagnosis of major depression, dysthymia, alcohol/drug abuse, social phobia, or other specific phobias; however, these diagnoses should not be the primary treatment concern. To ensure the homogeneity of the sample, a lifetime presence of severe and persistent mental disorders (axis I) or personality disorders (axis II) led to the exclusion from the study (in less than 5%). If healthy individuals met the criteria for any current or lifetime axis I or axis II disorders, they were excluded from the study. All the individuals (patients and controls) need to have lived in a partnership before, but need not necessarily have a partner at the present time. Subjects who had never had a partner were not included in the study. All the subjects were required to take part in the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) (Wittchen et al., 1997) to check for the axis I and axis II disorders based on the criteria of the DSM-IV (Saß et al., 2003). All the subjects gave written informed consent. The 424 www.jonmd.com study was performed in agreement with the Declaration of Helsinki and was approved by the University of Dresden Medical Faculty Ethics Review Board (EK 79032011). Sample Description For this study, 318 subjects were recruited. In the total sample, the average age of the participants was 36 years (SD = 11.2, min = 18 years and max = 65 years). Two thirds of the participants were female (66%), and 76% had a steady partner. Of the participants, 42% had graduated with a 10th-grade education from a secondary school with vocational training. The clinical sample comprised 175 participants, and the nonclinical sample comprised 143 participants. With regard to age and sex, there were no significant differences, suggesting that matching had been successful. However, the partner situation (p = 0.028) and education (senior high school graduation with p < 0.001, but not 10th-grade education from a secondary school with vocational training with p = 0.606) showed significant differences. In the clinical sample, over 80% of the subjects had a steady partner in comparison to the healthy individuals (nearly 70%). With regard to education, 64% of the patients had not graduated from senior high school as opposed to nearly 49% of the healthy individuals. The significant differences in terms of inability to work (p < 0.001) were as expected, only 5% of the healthy subjects were unable to work compared with 47% of the patients. The exact values and additional variables are presented in Table 1. Instruments Implicit Association Test The Partner/Mother-IAT is based on the assumption that the evaluation of an object (attitudes) (in the Partner-IAT, the person of the partner; Banse and Kowalick, 2007) could be conceptualized by the strengths of the association between the object category (e.g., partner versus stranger) and the attribute category (e.g., good versus bad) in the semantic network. The method taps into automatic implicit evaluation processes with respect to the partner/mother based on response times for the assignment of terms to categories. The greater the association between the object category and the attribute category in a participant (e.g., partner + loving), the easier it is to react with the same key press to the corresponding attribute and object categories. The Partner/Mother-IAT (Banse and Kowalick, 2007) is constructed analogous to the standard IAT (Greenwald et al., 1998). It consists of three training blocks (1, 2, and 4) and two measuring blocks (3 and 5). The discrimination tasks with positive versus negative (attributes, 40 repetitions) and partner/mother versus stranger (target concept, 160 repetitions) were presented to the subjects. For the positive-negative discrimination, there are four positive and four negative attributes. For the partner/mother-stranger discrimination, four individual items for the partner/mother and four items that are not characteristic of the partner/mother were selected in advance. The sequence of blocks and the randomized presentation of items are constant. For the total score, the D-score, as proposed by Greenwald et al. (2003), was used (see Greenwald et al., 2003, for calculation rules). More positive scores indicate more positive implicit attitudes toward the partner/mother. The reliability (internal consistency) was satisfactory with α = 0.83 for partner (Banse and Kowalick, 2007) and α = 0.83 for mother (Zimmermann and Maier, 2002). In the present study, the internal consistency was satisfactory as well, with α = 0.85 for partner and α = 0.81 for mother. A satisfactory test-retest reliability (partner: r = 0.65–0.69; mother: r = 0.51) was shown (Zayas and Shoda, 2005; Zimmermann and Maier, 2002). In addition, significant positive correlations between the response times and the secure attachment scale of the Relationship Questionnaire could be demonstrated (Bartholomew and Horowitz, 1991), as well as negative correlations © 2019 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Implicit vs Unconscious Attachment TABLE 1. Sociodemographics of the Total Sample (N = 318) and of the Clinical (n = 175) as well as the Healthy Control Sample (n = 143) Total Sample Age Sex (% female) Nationality (% German) Marital status (% single) Partner situation (% steady partner but unmarried) Steady partner (% yes) Educationa (%) Graduation senior high school (% no) Inability to work (% no) GSI PST PSDI BDI Clinical Sample Healthy Controls Differences n Mean (SD)/% n Mean (SD)/% n Mean (SD)/% t/χ2 Significance 318 318 318 318 318 318 318 318 314 318 318 304 310 36.2 (11.02) 66.0 98.4 54.1 42.1 75.8 42.14 57.23 71.7 0.71 (0.63) 57.16 (46.20) 0.63 (0.15) 9.61 (8.35) 175 175 175 175 175 175 175 175 175 175 175 175 168 36.02 (10.54) 65.7 98.9 53.1 41.7 80.6 43.4 64.0 53.1 1.05 (0.61) 81.382 (44.05) 0.68 (0.11) 14.18 (8.16) 143 143 143 143 143 143 143 143 139 143 143 129 142 36.41 (11.61) 66.4 97.9 55.2 42.7 69.9 40.6 49.0 95.0 0.30 (0.34) 27.52 (27.91) 0.57 (0.17) 4.21 (4.44) −0.32 0.02 0.46 0.14 0.03 4.86 0.27 7.28 66.71 13.93 13.24 6.53 13.61 0.752 0.893 0.496 0.708 0.865 0.028 0.606 0.007 <0.001 <0.001 <0.001 <0.001 <0.001 a Tenth-grade education from a secondary school with vocational training. with the subscales avoidance and anxiety (Banse and Kowalick, 2007; Zayas and Shoda, 2005). Beck Depression Inventory The Beck Depression Inventory (BDI) (Beck et al., 1961; Hautzinger et al., 1995) assesses the intensity of depressive symptoms during the previous 7 days. It consists of 21 symptoms rated in terms of intensity from 0 to 3 (range = 0–63). A higher total score indicates a higher symptom severity. Internal consistency of r = 0.88 in clinical samples is regarded as satisfactory. Brief Symptom Inventory The general psychological distress caused by various psychological symptoms was assessed using the Brief Symptom Inventory (BSI) of Derogatis and Melisaratos (1983) (Franke, 2000), a short form of the Symptom Check List-90-R. In this 53-item self-report, inventory participants rate the extent to which they have been troubled by various symptoms during the previous 7 days using a 5-point Likert scale (0 = “not at all” to 4 = “extremely”). The BSI provides information on the severity of symptoms in nine areas and in three global parameters. The Global Severity Index (GSI) indicates the global psychological burden, the Positive Symptom Distress Index (PSDI) provides information on the intensity of the responses, and the Positive Symptom Total (PST) indicates in how many items there is an impairment. The internal consistency of the nine scales ranged between r = 0.63 and r = 0.85 in a sample of 529 patients with renal failure. The test-retest reliability after 1 week in a group of 50 students was between r = 0.73 and r = 0.92. Adult Attachment Interview The AAI (George C, Kaplan N, Main M Attachment interview for adults [unpublished manuscript]), a semistructured interview with 18 questions, detects the current representation of attachment experiences with regard to the past and the present (primary attachment figure, mother, and father). It reflects the current emotional and cognitive states of mind with respect to attachment. The questions predominantly refer to relationships/attachment, separation, illness, and loss. To evaluate the unconscious aspects of the inner working model (mental organization of attachment), the responses of the interviewee were analyzed using verbatim transcripts. The coding process refers to linguistic aspects, such as the degree of coherence. Based on the content analysis, the subjects can be classified into four patterns of attachment representation (Main M, Goldwyn R, Hesse E Adult Attachment scoring and classification system [unpublished manuscript]): secure/autonomous attachment (F), insecure/preoccupied (E), insecure/dismissing (Ds), and disorganized/unresolved trauma (U). Secure (F) and insecure (Ds and E) provide the organized patterns, whereas the unresolved attachment (U) is the so-called disorganized pattern. In the analysis relating to the secure versus insecure classifications, the U classification is counted among the insecure attachment types. The AAI has very good psychometric properties showing high test-retest reliability for the three main classifications with 78% to 90% and κ = 0.63 to κ = 0.79 and high interjudge reliabilities (Bakermans-Kranenburg and van IJzendoorn, 1993; Sagi et al., 1994). In the present study, the interrater reliability of the reliable certified judges (J. B./K. B. and K. N.) was 59% based on 10% of the AAIs for the four attachment representations F, Ds, E, and U (κ = 0.43, p < 0.001). The AAI attachment representation was shown to be independent of nonattachment-related memory and intelligence or verbal fluency (Bakermans-Kranenburg and van IJzendoorn, 1993; Sagi et al., 1994; van IJzendoorn, 1995). The AAI also showed a high predictive validity for the intergenerational transmission of attachment (Benoit and Parker, 1994; van IJzendoorn, 1995). RESULTS Table 2 displays the frequencies of the attachment classification in the AAI. The patients predominantly showed unresolved trauma, and the healthy individuals showed predominantly secure attachment. TABLE 2. Frequencies of AAI Classifications for Patients (n = 175) and Healthy Controls (n = 143) Patients Controls AAI Classification n % n % Secure–F Avoidant–Ds Ambivalent–E Unresolved–U 34 63 10 68 19.4 36.0 5.7 38.9 60 35 6 42 42.0 24.5 4.2 29.4 www.jonmd.com 425 © 2019 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Petrowski et al. TABLE 3. Differences in the Three-Way AAI Classification Secure (n = 92), Insecure (n = 103), and Disorganized (n = 98) and Differences in the Two-Way AAI Classification Organized (n = 206) versus Disorganized (n = 110) for the Mother- and Partner-IAT for the Total Sample IAT-Partner IAT-Mother Secure Mean (SD) Insecure Mean (SD) Disorganized Mean (SD) F df p η2 0.689 (0.377) 0.782 (0.357) 0.688 (0.404) 0.654 (0.395) 0.754 (0.422) 0.694 (0.460) 0.859 2.498 2 2 0.425 0.084 0.006 0.017 Organized Mean (SD) Disorganized Mean (SD) t df p d 0.689 (0.412) 0.715 (0.382) 0.748 (0.423) 0.694 (0.460) 1.208 −0.390 314 165.307 0.228 0.697 0.14 0.05 IAT-Partner IAT-Mother Note. η2 ≥ 0.01 small effect; a post hoc analysis was nonsignificant. Concerning the attachment classification, no significant differences were observable between the secure, insecure, and disorganized (three way) as well as the organized and disorganized attachment (two way) attached individuals for the Mother- and Partner-IAT (Table 3). Concerning the differences in the clinical and the healthy samples (Table 4), there was a significant difference between the patients and the healthy controls in the Mother-IAT. The healthy individuals showed higher values and a shorter reaction time in the Mother-IAT compared with the patients. However, in the Partner-IAT, no differences could be observed. When only healthy individuals were included in the analyses, the differences for the Mother-IAT turned out to be significant (Table 5), with higher values for the securely and disorganized attached individuals (three-way attachment). The two-way attachment classification replicated the significant difference between the securely and the insecurely attached individuals in the Mother-IAT, with higher values and a shorter reaction time for the securely attached individuals (Table 5). When only patients were included in the analyses, no differences were significant for the Mother and Partner-IAT (results not displayed). As shown in Table 6, there were significant correlations between the Mother-IAT and the BDI scores as well as the GSI and PST scores of the BSI. Higher values and shorter reaction times in the Mother-IAT were connected with lower BDI and BSI values. Furthermore, higher values in the Partner-IAT were significantly associated with a lower PSDI score in the BSI. Furthermore, there was a significant moderate correlation between the Mother- and Partner-IAT. DISCUSSION In attachment research, different research lines developed independent methods for detecting the unconscious attachment schemes/ representations. The convergence between the unconscious and implicit attachment representations/schemes is still unclear. For this reason, the IAT as the implicit latency–based method and the AAI as the languagebased method were assessed simultaneously in the present study to investigate the coherence or rather the divergence of unconscious/ implicit attachment. The distribution of the attachment classification in the present study showed a higher frequency of secure attachment in the healthy individuals and unresolved trauma in the patients with panic disorder. This distribution was in accordance with the literature (Dozier et al., 2008) and previous studies of patients with panic disorder at the same center (Petrowski et al., 2011). First, it was proposed that the secure attachment classification would show positive implicit attitudes toward the partner/mother and shorter reaction times than the insecure attachment classification. The present study showed that the attachment classifications (AAI) cannot be clearly differentiated by the implicit relationship schemata (IAT). Comparing the secure, insecure, and disorganized attachment or the secure and insecure attachment, no differences for the implicit mother or partner attitudes could be found even though the power in the cells was high enough (n = 92–103). Because the implicit relationship attitudes differed between the patients and the healthy individuals, separate analyses were implemented. In the Mother-IAT, the healthy individuals showed more positive response attitudes toward the mother than the patient group. However, for the Partner-IAT, no significant differences between the attachment classification and the different samples could be found. This lack of differences might be explicable by the differences in reference. The AAI uses the attachment figure of mother or father as reference. The significant but medium correlation between the implicit attitudes toward the mother and the partner might reflect the discrepancy in the methodological reference. This is in line with earlier studies with low associations between methods with different references (Allen et al., 2005; Crowell and Treboux, 1995; Crowell et al., 2008; Kirchmann and Strauß, 2008; Stein et al., 2002). In the separate analyses of the healthy sample, the three attachment classifications differed significantly. They showed more positive attitudes toward the mother in the secure and the disorganized attachment classifications compared with the insecure attachment classification. This effect is not apparent in the patient sample, which might be explicable by the predominantly disorganized attachment classification in the patients and the very few organized attached patients. Therefore, one can conclude that for differentiating secure-insecure attached TABLE 4. Comparison of the Patients and Healthy Controls in Mother- and Partner-IAT Patients IAT-Partner IAT-Mother Controls n Mean (SD) n Mean (SD) t df p d 174 157 0.727 (0.393) 0.657 (0.423) 142 138 0.688 (0.443) 0.765 (0.386) 0.840 −2.281 314 293 0.402 0.023 0.09 0.27 Note. d ≥ 0.20 small effect. 426 www.jonmd.com © 2019 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Implicit vs Unconscious Attachment TABLE 5. Differences in the Three-Way and Two-Way AAI Classifications Secure, Insecure, and Disorganized for the Mother- and Partner-IAT Only for the Healthy Individuals Secure 3-Way classification 2-Way classification IAT-Partner IAT-Mother IAT-Partner IAT-Mother Insecure Disorganized n Mean (SD) n Mean (SD) n Mean (SD) F/t df p η2/d 60 60 94 92 0.644 (0.387) 0.814 (0.344) 0.700 (0.382) 0.782 (0.357) 39 39 222 203 0.691 (0.444) 0.634 (0.389) 0.714 (0.430) 0.674 (0.427) 38 38 0.807 (0.420) 0.815 (0.424) 1.832 3.157 0.271 −2.111 2 2 314 293 0.164 0.046 0.787 0.036 0.027 0.045 0.03 0.27 Note. Post hoc analyses: secure > insecure (p = 0.067), secure > disorganized (ns), insecure > disorganized (ns); d ≥ 0.20 small effect; η2 ≥ 0.01 small effect. individuals, the IAT for the mother seems to be a valuable instrument to differentiate the attachment classification. For the patients with predominantly disorganized attachment, however, the implicit measure does not detect the segregated attachment material. The second hypothesis was that the unresolved trauma/disorganized attachment representations would show negative implicit attitudes toward the partner/ mother and longer reaction times than the secure or insecure attachment. Based on the present data, the second hypothesis that unresolved trauma/disorganized attachment representations would show negative implicit attitudes toward the partner/mother can be rejected. Concerning the prediction of psychopathology, the symptom severity was associated with the attitudes toward the mother. A more positive attitude toward the mother was present in subjects with a lower depressive symptomatology (BDI), a lower general symptom severity (GSI of the BSI), and a lower number of different impairment symptoms (PST values of the BSI). In contrast, the implicit partner attitudes were exclusively related to the symptom severity (BSI PSDI). Because healthy individuals predominately showed more secure attachment and secure attachment was related to more positive attitudes toward the mother and partner, these association between IAT and psychopathology proved sensitivity to the clinical aspects. The aim of the study was to investigate the differences or similarities of implicit/unconscious attachment schemata evaluated by a developmental and a personality/social psychological approach. First, the present results showed that the same reference had to be measured (AAI-Mother and Mother-IAT). When the same reference of the methodological approach is given, there is a clear overlap between implicit attitudes toward the mother and the unconscious attachment representation of the AAI for secure and insecure attachment representation. However, the unresolved trauma/disorganized attachment (AAI) with their segregated system materials could not be detected by the IAT. These individuals have a similar attitude toward the mother as the secure attached individuals. This effect might be explained by the differences in the levels of awareness of the IAT. Based on Gawronski et al. (2006), the implicit attitude toward the mother has a low impact of TABLE 6. Correlation Between BSI, BDI, and the Mother- and Partner-IAT for the Total Sample IAT-Partner IAT-Mother BDI Score BSI-GSI BSI-PST IAT-Partner IAT-Mother BDI score BSI-GSI BSI-PST BSI-PSDI 1 0.298*** −0.071 −0.057 −0.045 −0.126* 1 −0.158** −0.150* −0.138* −0.071 1 0.797*** 1 0.803*** 0.988*** 1 0.295*** 0.386*** 0.297*** *p < 0.05. **p < 0.01. ***p < 0.001. awareness, as does the attachment representation. These individuals know well that they know something about attachment and why, but the individuals do not know how this knowledge about attachment affects other processes. However, the implicit attitudes toward the mother do have a higher source and content of awareness (“I saw a name”) compared with the attachment representation (AAI) evaluating the current state of mind concerning attachment. The content of awareness and source awareness varies, especially in the case of a segregated system. These differences might be explained by the lack of activation of the attachment system during an implicit association test. During the AAI, the biographical questions and the behavior of the interviewer intend to increase the stress level during the interview to activate the attachment system (Bowlby, 1969–1982; George C, Kaplan N, Main M Attachment interview for adults [unpublished manuscript]). For the latency-based methods, no stress induction and increase in intensity are part of the methodological approach. Therefore, the implicit attributes toward the mother seem to measure the less conscious but aware facets of the attachment construct. The mostly unaware implicit attitudes toward the mother showed an overlap with the attachment representation but not with entirely detached experiences from the memory of the segregated system. Although the sample of the present study is large compared with those of previous research with the AAI for the comparisons of the three-way contrast, the number of individuals in the cells did not reach the required amount of individuals based on the power calculations. In addition, these analyses exclusively relate to patients with anxiety. A similar investigation needs to be conducted with different psychological disorders because the distribution of the attachment classification and the predominant classification differ due to the different etiologies of the disorders. Moreover, the patients with anxiety were recruited exclusively from one university hospital, which might have led to the influence of a sample effect. In conclusion, the IAT and the AAI measured similar attachment-specific characteristics. This presented itself exclusively when a similar point of reference (mother) was evaluated. Besides, the disorganized attachment in the AAI was not detected by the IAT. DISCLOSURE This study was supported by the Deutsche Forschungsgemeinschaft, STR 306/22-1, PE 1804/2-1. The authors declare no conflict of interest. REFERENCES Allen JG, Stein H, Fonagy P, Fultz J, Target M (2005) Rethinking adult attachment: A study of expert consensus. Bull Menninger Clin. 69:59–80. Bakermans-Kranenburg MJ, Van IJzendoorn MH (1993) A psychometric study of the Adult Attachment Interview: Reliability and discriminant validity. Dev Psychol. 29:870–879. © 2019 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. www.jonmd.com 427 Petrowski et al. The Journal of Nervous and Mental Disease • Volume 207, Number 6, June 2019 Banse R, Imhoff R (2013) Implicit cognition and relationship processes. In Simpson JA, Campbell L (Eds), The Oxford handbook of close relationships (pp 475–499). New York: Oxford University Press. Banse R, Imhoff R, Steffens MC, Schramm N, Rösch A, Roberts M, Stangier U (2013) Partner-AMP and well-being: Evidence for an implicit secure base script? Pers Relatsh. 20:140–154. Banse R, Kowalick C (2007) Implicit attitudes towards romantic partners predict wellbeing in stressful life conditions: Evidence from the antenatal maternity ward. Int J Psychol. 42:149–157. Bartholomew K, Horowitz LM (1991) Attachment styles among young adults: A test of a four-category model. J Pers Soc Psychol. 61:226–244. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) An inventory for measuring depression. Arch Gen Psychiatry. 4:561–571. Benoit D, Parker KC (1994) Stability and transmission of attachment across three generations. Child Dev. 65:1444–1456. Bowlby J (1969–1982) Attachment and loss (Vol. 1–3). New York: Basic Books. Crowell JA, Fraley RC, Shaver PR (2008) Measurement of individual differences in adolescent and adult attachment. In Cassidy J, Shaver PR (Eds), Handbook of attachment: Theory, research, and clinical applications (Vol. 2, pp 599–634). New York: Guilford Press. Crowell JA, Treboux D (1995) A review of adult attachment measures: Implications for theory and research. Soc Dev. 4:294–327. Hahn A, Gawronski B (2014) Do implicit evaluations reflect unconscious attitudes? Behav Brain Sci. 37:28–29. Hahn A, Judd CM, Hirsh HK, Blair IV (2014) Awareness of implicit attitudes. J Exp Psychol Gen. 143:1369–1392. Hautzinger M, Bailer M, Worall H, Keller F (1995) Beck-Depressions-Inventar (BDI): Bearbeitung der deutschen Ausgabe. Testhandbuch. Bern: Hans Huber. Kirchmann H, Strauß B (2008) Methoden zur Erhebung von Bindungsmerkmalen. Klinische Diagnostik und Evaluation. 1:293–327. Main M, Solomon J (1986) Discovery of an insecure-disorganized/disoriented attachment pattern. In Brazelton TB, Yogman MW (Eds), Affective development in infancy (pp 95–124). Westport, CT: Ablex Publishing. Mikulincer M, Shaver PR, Sapir-Lavid Y, Avihou-Kanza N (2009) What's inside the minds of securely and insecurely attached people? The secure-base script and its associations with attachment-style dimensions. J Pers Soc Psychol. 97:615–633. Petrowski K, Nowacki K, Pokorny D, Buchheim A (2011) Matching the patient to the therapist: The roles of the attachment status and the helping alliance. J Nerv Ment Dis. 199:839–844. Saß H, Wittchen HU, Zaudig M, Houben I (2003) Diagnostisches und Statistisches Manual Psychischer Störungen–Text Revision—DSM-IV-TR. Göttingen: Hogrefe. Sagi A, van IJzendoorn MH, Scharf M, Koren-Karie N, Joels T, Mayseless O (1994) Stability and discriminant validity of the Adult Attachment Interview: A psychometric study in young Israeli adults. Dev Psychol. 30:771–777. Derogatis LR, Melisaratos N (1983) The Brief Symptom Inventory: An introductory report. Psychol Med. 13:595–605. Stein H, Koontz AD, Fonagy P, Allen JG, Fultz J, Brethour JR Jr., Allen D, Evans RB (2002) Adult attachment: What are the underlying dimensions? Psychol Psychother. 75:77–91. Dozier M, Stovall-McClough KC, Albus KE (2008) Attachment and psychopathology in adulthood. In Cassidy J, Shaver PR (Eds), Handbook of attachment: Theory, research, and clinical applications (Vol. 2, pp 718–744). New York: Guilford Press. van IJzendoorn MH (1995) Adult attachment representations, parental responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychol Bull. 117:387–403. Franke GH (2000) BSI. Brief Symptom Inventory von LR Derogatis–Deutsche Version: Manual. Göttingen: Beltz. Waters HS, Waters E (2006) The attachment working models concept: Among other things, we build script-like representations of secure base experiences. Attach Hum Dev. 8:185–197. Gawronski B, Hofmann W, Wilbur CJ (2006) Are “implicit” attitudes unconscious? Conscious Cogn. 15:485–499. Wittchen HU, Zaudig M, Fydrich T (1997) Skid. Strukturiertes klinisches Interview für DSM-IV. Achse I und II. Handanweisung. Göttingen: Hogrefe. Greenwald AG, McGhee DE, Schwartz JL (1998) Measuring individual differences in implicit cognition: The implicit association test. J Pers Soc Psychol. 74:1464–1480. Zayas V, Shoda Y (2005) Do automatic reactions elicited by thoughts of romantic partner, mother, and self relate to adult romantic attachment? Pers Soc Psychol Bull. 31:1011–1025. Greenwald AG, Nosek BA, Banaji MR (2003) Understanding and using the Implicit Association Test: I. An improved scoring algorithm. J Pers Soc Psychol. 85: 197–216. 428 www.jonmd.com Zimmermann P, Maier M (2002) Bindungs-IAT. Unveröffentlichtes Testmanual. Universität Regensburg. © 2019 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.