INTRODUCTION

Some psychotherapists may insist that their particular techniques are decisive factors for the success of therapy.  In contrast, some other therapists may emphasize that nonspecific factors or common ingredients in diverse therapies are the key for psychotherapeutic success.  This controversy prompted Tschacher, Junghan, and Pfammatter (2014) to investigate common factors shared by various therapeutic approaches.  Tschacher’s team conducted a statistical survey of 68 psychotherapy experts and found that the extent of patient engagement, of affective experiencing and the overall therapeutic alliance were judged as most relevant. These common factors differed with respect to how well they could be theoretically explained by the set of diverse techniques.   A literature review by Ackerman and Hilsenroth (2003) indicated that the therapist’s personal attributes such as being flexible, honest, respectful, trustworthy, confident, warm, interested, and open contribute positively to the therapeutic alliance.

       As pointed out long ago by Kiesler (1971), clinical research on the process of psychotherapy too often implicitly assumes that therapists are a homogeneous group while emphasizing other variables supposedly more directly relevant to outcomes.  For example, patients' perceptions or descriptions of their therapists were not yet sufficiently explored, despite established evidence that the quality of patient therapist relationship or the perceptions of therapists by patients are significantly related to treatment outcomes (Beutler, Johnson, Neville, Elkins, and Jobe, 1975; Lorr, 1965).  The reason for this neglect has been a lack of an adequate research tool. This article describes the theoretical background and characteristics of a multivariate inventory measure developed by Bergold and his team to categorize and quantify patients's descriptions of therapists.

      Bergold's inventory is a measure of verbal behaviour of patients, i.e. not a measure of the actual therapist's behavior.  The inventory only quantifies the patients’s description of their therapists. The special value of Bergold's inventory lies in overcoming some of the drawbacks of the existing inventories such as theoretical restriction of evaluative dimensions to psychoanalytic or Rogerian postulates (see inventories by Barrett-Lennard, 1962, Snyder and Snyder, 1961, Truax and Carkhuff, 1967, and the work by Ashby, Ford, Guerney, Guerney, and Snyder, 1957).

      Some of the older questionnaires led to valuable findings.  For example, Lorr (1965) found that scores on some of his scales descriptive of therapists were significantly associated with improvement over the course of treatment.  At present, there is a need for an instrument with scales more closely related to modern clinical concepts, including those from behavior therapy and those mediated by application of social psychology to clinical realities.

      Thus, in the development of Bergold's inventory, an integration was attempted of Goldstein's (1971) classical work on psychotherapeutic attraction and of behaviorist concepts of social reinforcement and of modelling or anxiety inhibition, with traditional Rogerian and psychoanalytic concepts of therapist's behaviour and of the client's related perceptions (Freud, 1958, 1996, Rogers, 1957), with socio-psychological  perspective on human interaction and on attitude change (e.g., Sears, Freedman, and Peplau, 1985), and also with theories of the European existential-daseinsanalytic school (Condrau, 1965, 1970).  Using this theoretical background, our search of  suitable questionnaire items was simultaneously oriented toward (a) items descriptive of therapist's behaviour, therapy situations which could be under therapist's control, and of patient's feelings about both, and (b) nonredundant dimensions descriptive of therapist variables which have an impact on therapeutic alliance and potentially on outcomes.  One of the major criteria for the choice of dimensions was their stimulus value for further research. Using these various considerations and criteria, the following set of 10 factors was theoretically postulated as worthwhile dimensions that underlie psychotherapeutic relationship, or more specifically, the patient’s perceptions of their therapist: (1) Genuineness, (2) Emotional Resonance, (3) Empathy, (4) Mutual Liking, (5) Directiveness,  (6) Perceived Similarity, (7) Social Reinforcement, (8) Anxiety Inhibition,  (9) Modelling, and (10) Expert Status. 

The descriptions are as follows. 

Genuineness: The therapist is perceived as not feigning a warm approach, as not concealing facts, or secretly manipulating the patient.  The therapist openly expresses his/her various relevant thoughts about the patient.  The therapist does not hide behind a professional role or social status.

Emotional Resonance: The therapist is perceived as being emotionally involved in the therapeutic relation and shows a wide variety of related emotions (e.g., concern, sadness, elation, surprise, or discontent).  The therapist is not perceived as emotionally indifferent.

Empathy: The therapist is perceived as immediately sensing the patient’s emotional processes and can anticipate the patient’s emotional reaction to events.  The therapist seems able to guess at the patient’s private thoughts and intimate feelings.  The therapist skilfully interprets the patient’s nonverbal behavior.  The therapist adjusts his/her verbal behavior carefully to the patient’s immediate emotional states.

Liking: This dimension has been studied extensively by social psychologists.  The patient perceives the therapist as likeable or not likeable.  The patient might report a personal aversion to a particular therapist or a strong enthusiasm for the therapist.  The extent of liking has an impact on interpersonal trust in the therapeutic relationship.

Directiveness: The therapist is perceived as giving firm instructions on what he indicates would be the best for the patient.  The therapist clearly attempts to influence the patient in a certain direction.  Or, at the opposite end of this dimension, the therapist passively waits for the patient to formulate his own or her own decisions.  The patient may perceive the therapist as aloof, not providing any advice, or, at the positive end of this dimension, as constantly steering the patient in some particular direction.

However, in some situations, rather paradoxically, the therapist explicitly forces the patient to take his own decisions, to foster the patient’s personal independence and self-sufficiency, while the patient would prefer being firmly told what to do. 

Perceived Similarity: The therapist is perceived by the patient as having some similarities in personal interests, certain behavioral patterns or idiosyncrasies, free time hobbies or pastimes, or in political opinions, educational values, or in particular cultural beliefs and behavioural standards.  At the opposite end of this dimension, the therapist is perceived as too different from the patient in most relevant respects.

Social Reinforcement: This concept is inspired by behavioral therapies utilizing reinforcement as an explanatory notion.  The therapist’s behavior acquires a reinforcing value.  For example, the therapist’s positive emotions, praise of the patient, or an expressed disappointment or dissent are likely to gradually shape the patient’s behavior in a more healthy and more socially adaptive direction.  The therapist skilfully uses these behavioral contingencies to correctively influence the patient’s behavior and attitudes.  

Anxiety Inhibitor:  This concept is also rooted in theories of behavior therapy.  The therapist is perceived as able to lower the level of the patient’s anxiety with respect to a variety of situations.  At the negative end of this dimension, the patient’s anxiety increases while with the therapist.

Model: In his behavioristic theory of social learning, Bandura (1977) postulated that much of our social behavior is acquired via observational learning in social situations.  Similarly, the patient, often even without consciously noticing, may imitate the behavior of the therapist, in respects such as certain postures or other idiosyncrasies of nonverbal behavior, the choice of clothing or colors, the overt attitudes or even political beliefs, or certain verbal expressions.

Expert Status: The therapist is perceived as having the knowledge, skills, and technical training that allow successful therapy.  The patient sees the therapist as probably more skilled, better trained, and more extensively educated in the field of treatments than other therapists. 

The patient trusts that the therapist is adept to skillfully help.  At the negative end of the dimension, the patient perceives the therapist as incompetent, unable or unlikely to help.

      Some of the items to operationalize these 10 dimensions via questionnaire scales have been borrowed from already published inventories.  The final pool of items was judged by therapists, clinicians, and psychotherapy researchers from several university hospitals and outpatient clinics in Berne (Switzerland), Vienna (Austria), and Munich (Germany) with respect to item content, relevance to the theoretical dimensions and/or therapeutic relationship, ambiguity, and social desirability.  The experts also selected which of the 132 items best represented each of the 10 theoretical scales.

      Early factoranalytic studies with the inventory suggested that the cognitive organization of verbal concepts involved in descriptions of therapists might differ depending on the characteristics of the patient sample, especially on patient's socio-educational level (Cernovsky, Weber, and Bergold, 1974). While factoranalytic dimensions could provide insights into differential organization of percepts or verbal concepts in various subgroups of the patient population, the resulting dimensions might extensively differ from one sample of patients to another.  Such instability was suggested by our early factoranalytical studies.   Furthermore, factoranalytic dimensions are often unwieldy, too unpractical, and not of an adequate clinical interest for the development of an assessment tool. 

       At present, an operationalization in an apriori manner of theoretically derived dimensions might allow a more direct first step for theory testing for clinical purpose than factoranalytic scales.  Not only the latter may fail to transcend the patients' level of experiencing the treatment interactions, but the empirical factors could excessively vary with patients' characteristics. Thus, for the purpose of theory testing, Bergold's inventory is used here with apriori theoretical dimensions, as the first approach.  The dimensional system is used to translate patient's viewpoints into conceptual categories of contemporary psychotherapy research.

      The research on liking as a primary factor in gregarious behavior was extensively investigated not only in social psychology, but its importance within the patient therapist relationship was also stressed by diverse schools of psychotherapy and perceived as a powerful factor in personality change, whether this was formulated by means of the psychoanalytic concept of positive transference or within the behaviorist paradigm of social reinforcement, or in the framework of attitude change theory. Accordingly, the present article statistically analyses correlations of verbal reports of “liking the therapist” by patients to other variables.  Theoretically, we expected the scores on any of our a priori dimensions to be, in some degree, associated with scores on the dimensions of liking the therapist.

      Modelling effects could also potentially trigger positive personality change within the time span of treatment. Therefore, same as in the case of liking, the associations of scores on the modelling scale to other variables are evaluated here to estimate the proportion of variance accounted for by these other variables.  The scores on the modelling scale of Bergold’s inventory could be significantly associated with scores on other scales, especially with those on social reinforcement, genuineness, and expert status scales because they imply both valuing and trusting.  The belief in the efficiency of the therapist’s coping skills is presumably a usual prerequisite for the modelling effect.

      Finally, the correlations of Bergold’s inventory scales to sociodemographic variables such as age, gender, educational status, length of treatment are also examined in the next part of this paper.

METHOD

      About 800 questionnaires were handed out, by therapists and other psychiatric staff, to psychotherapy patients in Berne (Switzerland) and to psychotherapy patients in metropolitan areas of Munich (Germany) and of Vienna (Austria).  Patients with acute psychotic symptoms, or those with low IQs or organic impairment, and those referred for a court requested assessment were excluded from participating in our study.  Only 130 questionnaires were subsequently forwarded to us, directly or in an anonymous manner, by the patients, and of these, only 101 questionnaires were answered in an adequate manner suitable for generating a computerised file.

      The mean age of the 101 patients was 29 yrs (SD=7.9 yrs) with the mode at 31. The sample consistent of 55 men, 45 women, and of a patient who failed to indicate the gender.   About one third of these patients were university students, another third were white collar employees, and the rest were blue collar workers, housewives, unemployed, or retired.  Forty-three, i.e. almost one half of the total sample, were graduates with the “Abitur” (a preparatory school diploma on a level roughly approaching an associate college degree in North American settings) or were further educated beyond this level.  Fifty-seven patients were educated below the level of an Abitur.  The information with respect to education was missing for one patient.  Most of our patients lived in cities with population of more than 30,000 (67 patients). Only 10 lived in villages or rural settlements with less than 1,500 inhabitants.   

     With respect to diagnosis (as reported by the patients), 5 patients were manic depressives, 3 were schizophrenics in remission, the diagnosis of 6 patients was unknown, 10 patients suffered from psychosomatic symptoms, 5 were diagnosed with a personality disorder, 3 patients sought psychiatric help for sexual difficulties, one patient was treated for a suicide attempt, and the remaining 68 patients reported other non-psychotic symptoms. The therapist variable was not controlled or monitored in any manner.  In fact, at the time, the cooperation of some of our therapists was explicitly granted only under this condition.  We estimated that perhaps between 20 to 75 therapists were being described by the patients. The majority of these therapists were working within outpatient setting of universities or were affiliated with a local university.

The German language version of Bergold's inventory was used in our study. The patients were asked to rate their therapist on each of the 132 items as follows: 0 = strongly disagree, 1 = disagree,  2 = agree, and 3 strongly agree.  In the subsequent statistical analyses, the scoring was reversed for some items, as indicated in Table 1 and 2. 

Table 1

Item composition, means (SDs), and score ranges for the 10 main scales of Bergold’s inventory.

Table 2

Additional scales for research use. Plus signs (+) indicate direct scoring, minus signs (-) indicate reverse scoring.

The English version of Bergold’s inventory is reproduced in Appendix 1 and the original German version in Appendix 2.

      As already explained, the major scales of this inventory are as follows: (1) Genuineness,

(2) Emotional Resonance, (3) Empathy, (4) Liking, (5) Directiveness, (6) Perceived Similarity, (7) Social Reinforcement, (8) Anxiety Inhibition, (9) Modelling, and (10) Expert Status.

The item composition of these 10 scales is listed in Table 1, together with mean values, SDs,

and the range from minimum to maximum scores as calculated on our sample of 101 patients.  Additional research scales of Bergold’s inventory suggested for further research are listed in Table 2, i.e.,  Rogerian Warmth, Noninterpretative Analysis, and 6 scales that deal with aspects of directiveness, as follows: (1) Authoritarian Control, (2) Advisory Directing, (3) Client-Centered Steering, (4) Expression Steering Control, (5) Critical Authority, and (6) Verbal Monopoly. 

The Table 2 includes the full list of items operationalizing each of these scales.

RESULTS

      The mean scores, SDs, and the score ranges for each of the 10 main theoretical scales of Bergold’s inventory are listed in Table 1.

     Relationships of Bergold's inventory scales to sociodemographic and therapy variables. Pearson product moment correlation coefficients were computed to assess the relationships of the 18 major scales of Bergold’s inventory to the data on the length of therapy and to the sociodemographic variables including age, gender, patient's education, the educational level of patient's father, population size of the area of the patient's residence, and to a score for geographical mobility (address changes in 5 recent years). The analysis was exploratory and therefore an overall criterion of significance of a p < .01 (2-tailed) was adopted for the correlation matrix to minimize spurious findings. None of the demographic relationships were significant, except for the correlation between the scale that assessed the therapist’s value as an Anxiety Inhibitor and the patients' age: older patients more frequently described the therapists' behaviour as reducing their anxiety level (r=.33).

      We expected that patients with more extensive education would be more prone to rate their therapist as similar to themselves and would also express more liking for the therapist than would patients with lower education.  However, even when this hypothesis was examined by accepting any coefficients at p < .05 (1-tailed) as significant, no support for our assumption was found.  In fact, none of the coefficients computed between patient's education and the 10 major scales reached this level of statistical significance.  Only one correlation computed between education and the additional theoretical scales was significant at p <.05 and suggested a rather too weak trend: less educated patients were slightly more likely to describe their therapists as monopolizing the conversation r=.16.  This coefficient is too small to be of clinical relevance.

      We expected that patients with longer duration of therapy or higher frequency of sessions might develop or report a more positive tie to the therapist and would describe the therapist in a more positive manner along the main 10 Bergold's theoretical scales. Using the significance level of p < .05 (1-tailed) as a criterion, only one significant coefficient was found: patients with more sessions described their therapists as somewhat more expert (r=.18).  This coefficient is also too small to be clinically important.

      It was expected that males might perceive their therapists as more directive in an authoritarian manner. Some weak support for this expectation was found at p < .05 (1-tailed):  of the 7 Directiveness scales, three were significantly correlated with patient's gender. The males were indeed slightly more likely to describe the therapist as Authoritarian-Controlling (r=.19), more Advisory Directing (r=.18), and as exercising more Expression Steering Control (r=.20).

However, all these 3 coefficients are also too small to serve as a meaningful basis for clinical predictions.

      Finally, with respect to population size of the patient's place of residence, we expected that patients from smaller rural settlements would describe their therapists as less similar. This assumption found only weak support (r=.18) at p < .05 (1-tailed), also at a size almost irrelevant for clinical predictions. 

      Predictors of Liking.  Even lay persons often ask, in their conversations, “How do you like your therapist?”  We evaluated how such “liking” is correlated to the other 9 main scales of Bergold’s inventory.  The highest correlations of the Liking scale were to scales of  Social Reinforcement (r = .54) and Emotional Resonance (r = .43).  The scatterplots of relationships of Bergold’s Liking scale to these potential predictor variables did not suggest a presence of strong curvilinear trends.  In a linear multiple regression analysis, the patients’ scores on Emotional Resonance and Social Reinforcement as predictor variables accounted for approximately one fifth of the variance in the scores for liking the therapist (R=.47, R2=.22, F=13.90, p < .001). Adding any other predictors did not lead to significant increments in the R square. A correction for possible overestimates of R2 (Kerlinger and Pedhazur, 1973) resulted in an R2=.20.

      Thus, there is a significant association of reports of  liking or disliking the therapists to reports about the extent of the therapists functioning as a social reinforcer and the extent of emotional resonance within the therapy relationship and this association appears consistent with clinical expectations.

      In the above analysis, special attention was paid to the item overlap between the scale assessing Liking and the two predictor scales. Common items with the Liking scale were found in the Emotional Resonance scale (one item) and Social Reinforcement scale (two items). Their removal from these latter two scales (the scale of Liking being left intact) resulted in a minor decrease of the correlation  coefficient of Liking to Emotional Resonance (decrease from r = .43 to .39) and in a small reduction in the size of the coefficient between Liking and Social Reinforcement (decrease from .54 to .42).

      Predictors of Modelling.  We attempted to determine which theoretical concept could predict if the patient would report using the therapist as a behavioral model.  Several scales of Bergold's inventory were significantly correlated with the Modelling scale.  In order to avoid analyses involving an excessive number of predictor variables (Kerlinger and Pedhazur, 1973), only five strongest (r > .34) were chosen: Emotional Resonance, Anxiety Inhibition, Directiveness, Perceived Similarity, and Social Reinforcement.  Although Rogerian Warmth was also eligible as a potentially useful predictor, its high intercorrelation (r=.96) with Emotional Resonance scale precluded its inclusion in the regression analysis.  There was no item overlap between the five predictor variables and the Modelling scale.  The scattergrams of the relationships did not suggest a presence of strong curvilinear trends.  Thus, linear multiple regression analysis was performed.  A combination of four of the five predictor variables (all but the Social Reinforcement scale) accounted for close to one half of the variance in the scores assessing the reports of Modelling (R=.66, R2=.44, F=18.8, p < .001). The increment in the R square associated with adding the scale of Social Reinforcement as a predictor was not significant (p > .05).  The application of a conservative estimate via R2 shrinkage formula (Kerlinger and Pedhazur, 1973) to correct for small number of cases with a large number of predictors resulted in an R2 of .41.  Thus, more than one third of variance in reports of imitating the behavior of the therapist could be explained by exploiting its association with reports about the extent of emotional exchange in the relationships, the extent of perceived similarity and of the therapist's directiveness, and of the therapist’s role as an anxiety inhibitor.

DISCUSSION

        Although some of the results may seem impressive, they deal only with verbal reports by patients and are only correlational.  They do not necessarily imply causal relationships. These results only map and quantify the patients’ descriptions or ratings of what is happening in therapy sessions.   The most striking feature of Bergold’s inventory seems its relative independence of sociodemographic variables.  The demographic correlations found to the 10 key scales were only of trivial size and reached statistical significance only because the sample was rather large.  In a similar context, Witte (1993) warns researchers that very large samples could lead to significant but excessively weak correlations that lack practical importance.  The lack of strong sociodemographic correlations suggests that it is probably not necessary to establish separate norms (such as means and standard deviations for Bergold’s scales) for subsamples based on categories such as gender or educational level.  In this respect, our theoretical scales of Bergold’s inventory are superior to factoranalytically derived dimensions, those described in Weber, Cernovsky, and Bergold (1974).

         The weakness of this study lies in an underrepresentation of patients with poor socioeducational background in our sample.  Most patients were recruited in university settings and some therapists probably accepted only patients with above average IQs for psychotherapy as opposed to pharmacotherapy.   With the advent of approaches such as the Cognitive Behavior Therapy (CBT), even patients with below average IQs could be included.

        The main contribution of this study lies in providing a viable tool for psychotherapy research.  Much research needs to be carried out in this area.  Statistical data collected in our study are now too old, from the times before the widespread use of CBT.  New data from various countries need to be collected in order to re-evaluate intercorrelation matrix of the key scales of Bergold inventory, calculate the Cronbach’s alpha coefficients of internal consistency for the scales, examine item total correlations within the scales, and to verify if indeed there are no important correlations to sociodemographic variables. 

Furthermore, the studies could examine how the scale scores on Bergold’s inventory change with the progress of treatment.  Of interest might be also investigations of relationship of patient’s diagnosis or particular symptoms such as delusional ideation or of personality characteristics such as low self-esteem to Bergold’s scales.

          The patient’s negative statements about the therapist do not necessarily reflect an underlying incompetence of the therapists.  From the behavioristic point of view, the patient’s behavior towards the therapist may reflect his or her maladaptive individual learning history.  Psychoanalytic terminology refers to the impact of the patient’s learning history on his behavior to the analyst as the transference.  Freudian theory conceptualizes transference as a vehicle of the analytic treatment (Freud, 1958): the lack of activity on the part of analyst intentionally provokes transferential distortions by patients and thus generates a material for analytic work.  These distortions might be as often positive as negative.  The transferential distortion could be measured with instruments such as Gough and Heilbrun’s (2007) Adjective Check List (ACL) which consists of 300 adjectives and adjectival phrases commonly used to describe a person’s attributes. It may be administered to an individual to elicit a self evaluation or a characterization of someone else.  If the patient’s parents were the key suspected transferential source, the patients could be asked to describe their parents via the Adjective Check List and then also the therapist.  These two ACL profiles could be compared to scores on Bergold’s inventory at the onset and towards the end of treatment. 

          Hentschel, Kiessling, and Rudolf (1997) found, in a study of 126 patients, some statistical support for a transference based on relationships to father and mother figures onto the relationship of these patients to their therapist.  An excellent statistical study of transference on 134 patients by Zilcha-Mano, McCarthy, Dinger, and Barber (2014) is of seminal importance in this context.  These authors found that the patient’s pretreatment representations of significant others predicted a substantial part of the alliance with the therapist throughout the course of treatment.  However, these pretreatment representations were not mechanically imposed on the therapist, but rather these projections were also to some extent influenced and modified by the real relationship to the therapist, probably in the same manner as the transference based on relationships with significant others in the childhood is modified by subsequent relationships to close friends. The transferential pattern becomes a collage based on historical and new relationships with significant others.  In behavioristic language, the treatment alters the habitual behavioral patterns of the patient.  The study by Zilcha-Mano’s team (2014) is of particular importance because it is a rather successful attempt at statistically quantifying how the individual’s learning history (the prevailing modus of relating to others, the transference) influences the therapeutic alliance and potentially the treatment outcome.  In Zilcha-Mano’s study, the 23-item Response of Others (ROs) scale from the Central Relationship Questionnaire - Revised (McCarthy, Connolly, Gibbons, & Barber, 2008) was used to assess different behavioral characteristics of others as hurtful, loving, submissive, and independent, separately for each significant other (i.e., father, mother, best friend, romantic partner) from the patient’s pretreatment personal history.  In statistical terms, Zilcha-Mano’s team found that their particular measures of pretreatment transference (in behavioristic terminology, the individual’s past interpersonal habitual behavioral patterns) accounted for about one third of variance in the therapeutic alliance at intake and in the second week of treatment.  Similar approach to study of transference and of its impact on therapeutic relationship could be also carried out with Bergold’s inventory.

         Psychotherapy is often obstructed by the patient’s lack of interpersonal trust.  Greben (1984) pointed out that, in the course of life's experiences, an intense mistrust might develop over the childhood and adulthood years. Clinical manifestations of mistrustfulness in patients obstruct the treatment. Successful psychotherapy has as one of its principal goals the re-establishment of trust within the patient. This takes place through the development of trust toward the therapist. This experience serves as a bridge to the re-establishment of trust in others.  Personal qualities of the therapist or the patient’s perceptions of the therapist are important in this context.  Studies on the extent of the patient’s mistrust or trust towards the therapist and its changes over the course of psychotherapy could be fruitful.  For this reason, our Appendix 3 includes the list of items that constitute the scale of Trust in the therapist.  This scale is a recent addition to Bergold’s inventory.  Some therapists consider trust to be most central issue in counselling.

         The possibilities of future research on correlates of the Similarity scale of Bergold’s inventory also deserve attention.  The impact of similarity on therapist's influence on the patients or on patient's trust in the therapist is presumably not always positive, at least when the similarity is generated by  a pessimistic self disclosure of personal problems such as depression (Curtis, 1981) or by the perception of certain maladaptive attitudes (see also Beutler, Johnson, Neville, Elkins, and Jobe, 1975). In the present study, however, the perceived similarity was operationalized by items dealing with the similarity of interests, free time activities, opinions and reactions, and a similarity in the type of friends and acquaintances.  The results of such therapy studies could be very divergent, depending on how the similarity dimension is conceptualized.

         Of interest could be also data from surveys of patients, via Bergold’s inventory, about their image of an ideal therapist, i.e., about each patient’s particular wishes with respect to the therapist’s behavior, attitudes, and skills.  This is presumably by far not always the image of a therapist that would truly be objectively helpful to the particular patient in a more absolute sense, only an image of what the given patient assumes might be helpful.  Such wishful thinking (the “ideal image”) might change extensively over the course of treatment.  It would also be of interest to evaluate empirical correlations of such measures of a “desired therapist” via Bergold’s inventory to the patients’ baseline personality measures on standard tests. 

          Similarly, surveys of therapists of diverse theoretical backgrounds about what would be an “ideal therapist,” as rated via Bergold’s inventory, could be compared to their alleged theoretical allegiance.

          The various desirable characteristics of the therapist outlined in this study should serve to guide the patient to empowerment, i.e., to an independent mastery within the constantly fluid process between the patient and his environment.  According to Kilian (2008), the goal of modern psychiatry is to foster an independent lifestyle and autonomous dealing with the illness.  To quantify this concept of empowerment, Kilian’s team developed a 33 item questionnaire to measure important aspects of the patient’s life, including financial situation, living arrangements, work activity, social life, participation in therapy, insight, acceptance of illness, medication management, free time activities, political activism, sense of hope, self-efficacy, and family relationships.  The item content of Kilian’s questionnaire was derived from focus groups in several areas of Germany, with the participation of patients with major psychiatric illness, their family members, and experts in psychiatry.  It is of interest to investigate how the treatment outcomes as assessed by Kilian’s empowerment questionnaire (see Kilian, Becker, Schleuning, Welschehold, Hertle,  Hörand, and Matschinger, undated) relate to dimensions of perceived therapist’s behavior as operationalized in Bergold’s questionnaire. 

ACKNOWLEDGEMENT NOTE

Thanks are expressed to Beat Weber (Liz.Phil. in Psychology),  Dr. Marie-Louise Käsermann (professor of psychology), and Dr. Mario von Cranach (professor of social psychology) for their extensive contributions, cooperation, advice, and help in initial stages of this project at the University of Berne, Switzerland.

Thanks are also due to Dr. Jack Ferrari, psychologist, adjunct professor (retired) at the University of Western Ontario, for the help with improving the manuscript. 

Appendix Surveys

The Appendix surveys can be found here: http://ispub.com/IJPSY/5/1/43914