Transactional Analysis as Psychotherapy Method – A Discourse Analytic Study

Operational definitions of categorisations by McNeel (1975) were developed and applied by the author and an independent assessor to complete discourse analysis of 72 hours of transactional analysis group therapy in the style of Goulding & Goulding (1976, 1979) conducted during 1984/85. Results showed that the therapist used an average of 42% of the discourse space and that the therapy did indeed contain TA components, with the two main categories being ‘Feeling Contact’ and ‘Contracts’, and with particular use of TA techniques of ‘talking to Parent projections’, ‘make feeling statement’, ‘mutual negotiation’ and ‘specificity/clarity’. Inter-rater reliability was 46.2% (Araujo & Born 1985), Cohen’s (1960) kappa coefficient shows a spread from slight to moderate agreement, and the Odds Ratio (Viera, 2008) is above 1.0 for most categories.


Introduction
This study of TA group therapy focuses primarily on the discursive strategies, i.e. the therapist's categorised interventions.This means identifying changes and repetitions of categorised conversation processes and codings (identification) of when and how often they occur in the conversations.

Discourse Analysis
A discourse is a specific way to talk about and understand the world.It specifies the manner or pattern we use when we interact and express ourselves in different social contexts or discuss certain phenomena Discourse analysis is an analysis of emerging patterns and regularities focusing on these social exchanges of words (Foucault, 1993).
The current view that discourse is something fairly regular follows Michel Foucault (1972Foucault ( , 1993)); he assumes a social constructionist perspective, where the truth is a discursive construction.Different knowledge regimes, such as transactional analysis, indicate what is true and false.This defines the theoretical and practical frame of reference, thereby creating conditions for the study of repetitive interventions and opportunities for categorisation and coding of therapy evidence.Discourse analysis is primarily interested in the discursive practice rather than in the individual experience.
According to discourse analysis, the client and the therapist identification are determined by the patterns that emerge here-and-now in the conversation and not by the patterns that individuals historically carry with them.It is said that the subject will become fragmented or decentred (Winther-Jorgensen & Phillips, 2000) with an increasing number of identities, depending on what discourses they are part of.The identity is changing, being represented by the position selected in the discursive context.To speak is the same as to construct an identity, according to Potter, String & Wetherell (1984).
Discursive psychology (DIP) was developed in England by Billig (1987Billig ( , 1996)), Edwards (1997) and Potter (1997).When applied to analysing therapy sessions and authentic conversations in different contexts it is called conversation analysis, CA (Sacks, 1992).In her research at Linköping University in Sweden, Karin Aronsson (Aronsson, 1996(Aronsson, , 1998;;Aronsson & Cederborg, 1996) focuses on 'identity-ininteraction', where the social order is an important factor.This has stimulated studies of institutional 2 (2), 3-18 https://doi.org/10.29044/v2i2p3contexts such as court trials and family therapies; she also makes analyses of 'Social Choreography', where studies of the social space (Bakhtin, 1984) in the 'communicative dance' develop in an on-going dialectical process.The positioning does not proceed from a predetermined social order but from what happens in the conversation process.
The Linköping group is interested in concepts like discourse space (Aronsson & Rundström 1988, 1989), allocation of discourse space, turn-taking control, direction of conversation, orchestration (Aronsson, 1999), allocation and definition of turns in interaction and preferential right of interpretation (Peyrot 1987;Buttny, 1990;Aronsson & Cederborg, 1996), control of choice, change and summation of topics.These different aspects on what determines communicative exchange affect balance and influence the conversation, for good and bad.

Transactional Analysis
The creator of TA, Eric Berne, was interested in group therapy long before he developed the TA method.In some early publications (1953,1954,1955,1958) he presented TA as a group therapy, exhibiting a preference for group over individual therapy because the process in the group offered a practical tool for understanding how interactions between individuals in the present moment (transactions and games) are linked to the individual and their underlying patterns (ego states and script).Berne's group therapy differs from psychoanalytically oriented groups (Bion, 1974) that see the group as a systemic, separate whole, which affects the individual's unconscious needs.One of these directions (Yalom, 1995;Rogers, 1951;Slavson, 1947, Wolf et al, 1993) emphasises the interaction between group members and the therapist as a facilitator.The psychological forces may operate freely with few therapist interventions, increasing anxiety, projection and acting out, which are then interpreted by the therapist.In Berne's group therapy, however, the therapist is an active and visible leader in every transaction (Berne 1970).This fact makes TA an adequate method to be studied with a discursive approach, where the therapist's interventions can be categorised and identified.
Discourse psychology focuses on language as social practice in interaction with others.The discursive approach differs from transactional analysis in the perception of identity stability.TA emphasises that the discursive practice, the therapy, should lead to a change by making new decisions on an emotional and cognitive level.The assumption is that, for example, a negative sense of identity has its origin in locked adaptive patterns developed in childhood, known as scripts.Both TA and discourse analysis share the basic interactionist view, but have different views of self and identity changeability.TA emphasises instead that the exchange between people, the transaction, has a potential for change, but the underlying mental structures, as a script, limit the individual's choice and possibilities for change.
One direction in TA that developed in the 1970s was Redecision Therapy (Goulding & Goulding, 1976, 1979).Goulding's group therapy is focused on intrapsychic change in clients (redecision).Interactions in the group are toned down in favour of individual therapy in group (my emphasis).The main exchange is between therapist and client, and the group acts as a resonance and support in their individual work.The group therapy in the study follows this direction.

Categorisations used for Discourse Analysis
In a PhD dissertation by the American psychologist and TA therapist John McNeel, (1975), the major elements of this therapy were categorised.The thesis was primarily an effectiveness study.In a comparative t-test before and then three months after therapy he stated, using Shostrom's (1964) Personal Orientation Inventory (POI), that intensive therapy over a weekend (a so-called marathon) resulted in significant changes in clients in 10 of the 12 personal orientations (e.g.self-acceptance, spontaneity).
McNeel's secondary interest was to see what factors/categories of the therapy led to changes in the client.The seven main categories with their 42 subcategories (components) form the basis for this study.These categorisations, modified and operationally defined by the author and an independent observer, are thought of as requirements of the TA method that will be met, coded and compared as a measure of TA consistency.

Aims of the study and questions posed
The aim of this study was to examine whether the psychotherapy conducted was consistent with what the TA method requires.According to the aim the following questions have been posed:

1.
Can essential components of a transactional analysis group therapy be found in the study?

2.
Are there agreements between codings of the independent assessor and the author in that both identify the elements constituting a TA group therapy?

3.
What is the difference in agreements between the coding of sub-categories and main categories and what does this entail?

4.
Do the categories describe what is defined as typical or specific to a transactional analysis group therapy?

5.
What categories are in this case TA-specific?

Discourse analytical study design
The following was applied to recordings of a one-year therapy group:

Source material
The source material consisted of 24 videotaped therapy sessions from the year 1984/85 with 10 clients and one therapist, with session lasting three hours including a coffee break.The therapist (the author) was a certified psychologist as well as a Certified Transactional Analyst (Psychotherapy) (International Transactional Analysis Association, 2004).
The therapy sessions were recorded by a sound engineer.Due to technical problems only 66 sixtyminute tapes were available out of a total of 75 therapy tapes.The transcriptions of these sessions comprised 813 pages, with an average of 65 pages per session.
A sample of 11 sessions was made so that all phases of the therapy were represented.Ten sessions made up the basic data set for the regular part of the study, Sessions 2, 4 and 6 from the beginning, 9, 11, 12 and 16 the middle and 19, 23 and 24 the end.Session 22 was used as a pilot study.

Clients
The clients were eight women and two men.The average age was 35 with a variation between 27 and 55.Half of them were single.Six clients had academic backgrounds.Clients were volunteers who had requested therapy at a private clinic in Malmö, Sweden (Institute of Life Therapy -IFL), included consecutively from a waiting list.A secretary managed written and verbal information about the therapy and notifications to the group.Before the therapy began the clients were contacted via telephone by the therapist.In an individual meeting a short check of the conditions for the therapy was made.Only clients with severe disorders such as psychoses were rejected.All the first ten clients on the waiting list were accepted.
Their therapy was self-funded and they had given their written consent to video-recording the therapy for research purposes.
The Independent Observer/Assessor The independent assessor participated in the study from once the transcripts had been prepared.He was a 30-year-old psychology student with nothing but the written exam work left to be awarded his psychologist degree.He had no previous knowledge of TA.He coded from the transcripts without listening to the recordings, to achieve a level of blind review.

The pilot study
Categorisation of the pilot session revealed that the assessors had different opinions so the category contents and definitions were made more robust.We realised that the author/therapist's inside perspective and the independent assessor's outside perspective influenced the content definition of the categories.In order to achieve a good consistency in identifying the categories, but without reducing the differences in perspectives too much, we decided to begin the independent coding after our third coordination meeting.Appendix A shows the final definitions.These definitions are different from McNeel's, who used TA terminology and examples instead of definitions.All the main categories were restructured with new headings.Some categories were added, such as the main category 'Relations' with its four subcategories.Common psychological terminology was used to define the categories.

The main study
The 10 sessions were then analysed.The author (A) and the independent assessor (I) coded independently.A total of 8452 codings were made; 3731 by A and 4721 by I.

Calculation of inter-assessor reliability
The two assessors' codings were compared and the percent agreement and kappa ratio were calculated.Full details are provided in the Results section (Appendix A).

The Categories and Coding Principles
The extensive pilot study was carried out before the main study to enable the assessors to obtain a mutual understanding of the meaning of the 42 categories.
The assessors first used an individual interpretation procedure followed by a consensus discussion and an agreement decision.A high number of coding options (42 categories) complicated the coding, and therefore the assessors were prepared for and trained in the use of a simplified computerised procedure ('a pop-up menu').
Two general principles for coding were formulated: 1.The therapist's statements or interventions are coded, based on the assumption that it is the therapist in interaction with the client who contributes to the therapy, following a line that is specific to the psychotherapeutic direction.

2.
Each statement could be allocated to a maximum of three categories, although in most cases only one coding was used.
Statements containing "xxx" (i.e. an inaudible fragment) are excluded from coding.As the material is extensive, the loss is deemed acceptable and viewed as random.
When one, two or three codings were exactly alike, it was assessed as full agreement.When at least one of the assessors used more than one coding and this matched one coding by the other, it was considered partial agreement.

Discourse space
Although not the main interest in the study, when reading the therapy transcripts it was noted that a communication structure emerged in which the therapist has great influence on the arrangement of the therapy session.The therapist controls the initial and final discussion, directing double-chair work, has the largest share of and influence on the discourse space as well as greatest control of choices, changes and topic summaries.
The clients' discourse space in their own therapy work is counted in the transcripts as a dialogue with the therapist.The assessors identified that the therapist used an average of 41.7% of the total discourse space (Table 1).From the remaining 58.3%, each client's allocation of discourse space varies between 3.8 and 8.3 % of the entire therapy.
The framework and contract procedures in the TA therapy create opportunities for communication that are both controlling and permissive.The process follows a democratic dialogue methodology based on mutual negotiation, where the client's influence is supposed to be equal to that of the therapist.In the study the therapist dominates the discourse space to a fairly great extent, which might reduce the client's potential for spontaneous contributions.In the therapist role, according to TA, a combination of an active and democratic leadership with a strengthening of client power is preferable.

Individual and matching codings
In Table 2 the two assessors' individual and matching codings are shown for both main and sub-categories.The difference between matching codings calculated from all the main and sub-categories also appear.Of the independent assessor's (I) 4721 and the author's (A) 3731 codings there is agreement in 1419 codings in the sub-categories and 1953 codings in the main categories, as shown in Table 3.It also shows that the agreement is generally higher (534 + codings) in the main categories, which is natural, given the more general basis.The calculated difference is large for the main categories Feeling Contact (+144) and Reality Testing (+124), but Language Usage (+93) and Strokes (+80) also show a significant difference.The lowest difference is in Contract (+18) and Relations (+5).This means that the Feeling Contact and Reality Testing categories have higher agreement (priority) in calculations made on the basis of main categories, while the Contract category has priority in the sub-category calculation.Both methods of calculation rank Language Usage the highest and Relations the lowest.Note.I = Independent assessor, A = Author, Agree = Agreement between the assessors

Sub-category frequency
Some sub-categories are coded as more frequent than others, both in terms of the assessors' individual and of their jointly matching codings.A high frequency in one single assessor means that he believes that the category is commonly used in therapy.A similar high frequency agreement with the other assessor increases the reliability of one category being TA-typical.An overview of the nine most frequent assessments (Table 4) shows that the two sub-categories 'specificity/clarity' and 'mutual negotiation' are clearly the most frequent in the assessors' matchings but also in individual codings.'Make feeling statement', 'talking to Parent projections' and 'responsibility' also have a high correspondence between assessors' matching and individual codings.However, the 'train Adult' category differs from this, as it is a highly matching coding but shows a big difference in the individual codings.A has, in relative terms, given higher priority to this category, as compared with I.The following 15 categories may be prioritised individually or jointly, but there is no clear priority for both.The other 18 categories have low priority.
The assessors' individual and matching codings for the 7 main categories and their sub-categories.
The results of the codings for each of the seven main categories and their sub-categories were reviewed (see example in Table 5).The table shows the assessors' individual and matching codings for the sub-categories in each session.The results of the main category were also reported, calculated both on the basis of the total sub-categories and on the total main categories.Moreover, it is shown in which phase of the therapy the main category is the most frequent.The main Contract category and its 4 sub-categories.Main Language Usage category and its 5 sub-categories

In the example in
As many as 87.5% of the codings in the main category of Language Usage derive from the 'specificity/clarity' sub-category.This is the most frequent jointly coded sub-category in the study, representing 19.8% of all categories.The other sub-categories in this main category were rare with the exception of 'word confrontation/word change'.Language Usage appears to be relatively evenly distributed throughout the therapy.
Main Pattern category and its 5 sub-categories Of all the codings 'expose myth and magical thinking' occurred most frequently accompanied by 'connect past scenes with present conflict'.In relation to all codings in the study, these two categories occurred relatively seldom, only 2.5 and 1.5 %, respectively.Interventions related to Pattern were more frequent at the beginning of therapy (Sessions 2-9).

Main Reality Testing category and its 7 sub-categories
The main category of Reality Testing had three subcategories with the highest frequency: 'responsibility', 'train Adult' and 'use humour'.In relation to all the subcategories in the study they represent together 16 %.
The other four sub-categories are insignificant.
Interventions related to this main category are common in the middle of the therapy (sessions 6-19).

Main Feeling Contact category and its 9 sub-categories
The coding rate for the main category of Feeling Contact with its nine sub-categories had high rates for the groups 'make feeling statement' and 'express feelings'.In this main category these two sub-categories had 69.6% of all codings.In the study they represent 9.2% of all codings agreed upon.In the therapy process this kind of intervention occurs evenly throughout the sessions with a slight increase towards the end (Sessions 19-23).

Main Relations category and its 4 sub-categories
The last main category, Relations, had no consistent codings at all.The few existing ones had been coded as 'transference' by both examiners.I had also coded 'alliance rupture' on 43 occasions when A did not code that category at all.This coding was most frequent in the middle of the therapy process but may be considered of minor importance compared to all the categories included in the study.

Summary of the most frequent sub-categories under each main category
A summary of the most frequent sub-categories under each main category is presented in Table 6.Compared with the ranking of the most frequent sub-categories (Table 4) it shows that the Contract interventions 'mutual negotiation' and 'specificity/ clarity' are used the most, while the Pattern interventions 'expose myth and magical thinking' and 'separate old scene from present impasse' are used the least.

Summary of the most frequent categories in different phases of therapy
A summary of the spread of codings over the different phases of therapy provides the results in Table 6.The therapy process broadly follows what is indicated in the transactional analysis literature (Berne, 1966, Goulding & Goulding, 1979;Ohlsson, Birch & Johnson, 1992;Hewitt, 1995).The Contract phase is most important in the beginning to create the alliance and the goals of therapy ('mutual negotiation').Then comes a clarification phase where the level of awareness of the general Pattern (script) is raised and the Pattern (script) processed (the regressive phase).The client's own 'responsibility' for the problem increases in the middle of the therapy (Erskine, 1975).In the termination phase, changes are anchored with support and encouragement from the therapist.'Feeling Contact' (Johnson & Stenlund, 2010) and 'specificity/clarity' are key elements throughout the therapy.

The assessors' agreements
Table 7 shows that the assessors mainly used one code for each therapist intervention even if there was some diversity between them (A = 92.6%,I = 73.6%).Assessor I needed more often (+18.3%)than A (+7.1%) to encode two categories.This is understandable, given that A has an inside perspective and is familiar with the material.A third coding was used less frequently.

Percentage of agreement for sub-and main categories
Inter-assessor reliability was calculated partly on the basis of sub-categories and partly by main categories.Marques & McCall (2005) consider that different reliability measurements of assessor accordance create stability in qualitative research.The inter-assessment reliability percentage was calculated using a formula from Araujo & Born (1985), supplemented with Cohen's (1960) kappa coefficient and Viera's (2008) Odds Ratio.Of all 1419 assessors' matching codings (full + partial) full compatibility was coded 795 times and partial agreement 624 times.The individual codings are relatively even for all sessions except Session 24 with its distinctly low number of codings.The mean of the percentage agreement (full + partial) was estimated at 33.5% with a relatively even distribution across all 10 sessions.When the matches from the sub-categories were recalculated to the main categories the number of consistent assessments increased by 534 to a total of 1,953.The percentage then increased to 46.2%.The calculations followed similar trends to the sub-categories.

Complementary calculations of agreements for all suband main categories
All the kappa coefficients (κ) and Odds Ratios (OR) were calculated as a complement to the average percentage agreement (%), as shown in Tables 8a and 8b.OR is a standardised measure of effect that indicates the odds or the chance that agreement between assessors I and A in the coded category is more likely than disagreement.The quota (ratio) indicates the possibility that they agree versus the possibility that they do not agree.Odds Ratios above 1.0 strengthen the connection (association) between the assessors' matching codings and thus the probability that the assessed category is present.The tables show that all categories except six (which have 0.0) and five that could not be calculated have an OR that is above 1.The average percentage agreement broadly follows the kappa and OR values.

Ranking of agreements for main categories
Ranking the main categories (Table 9) enables the study of the main categories which are the highest on all measurements.Feeling Contact and Contract are at the top and both have 'moderate' agreement.These sub-categories also have high coding frequencies from both assessors.Language Usage has a high frequency but slightly lower reliability.

Ranking of sub-category agreement
Of the 21 categories, six have 'moderate' agreement (Table 10).They are 'talking to Parent projections'', 'active use of TA terminology', 'make feeling statement', 'mutual negotiation', 'refer to contract' and 'discrepancy in body language'.The other 15 categories have slight agreement.

Comparison of frequency and agreement of subcategories
If one weighs up the sub-categories with the highest coding frequencies and reliability, the following categories are specifically important and also specific for TA: 'mutual negotiation', 'making feeling statement', 'talking to Parent projections' and 'specificity/clarity'.

Discussion and Conclusions
The issues are related to the overall aim to investigate whether the psychotherapy the assessors have analysed is in accordance with what is considered to be transactional analysis group therapy.The first three questions are linked to a general conclusion about what constitutes TA therapy, while the remaining two are concerned with the specificity of the method.In order to study this issue a modified discourse analytic approach was applied, where high overall agreement between the assessors' category codings was supposed to show that the psychotherapy conducted follows what is considered as constituting transactional analysis psychotherapy.
The categorisation, which acted as the assessors' coding key, was based on McNeel's thesis (1975), which was revised to create operational definitions with a general psychological content.In the pilot study the assessors made an experimental control of how well the instrument worked, and afterwards it was found that the validity of the updated coding key was good.In the ordinary study the assessors then used the coding key in order to examine how consistent and reliable the encodings could become and whether the two assessors could achieve the same results regardless of who carried out the measurement.
The main question was whether they could agree that the categories they considered to describe TA in practical work could be observed in the transcribed sessions.The answer was a calculated agreement of 33.5 and 46.2% based on the sub-and main categories, respectively.The first calculation involving the 42 sub-categories had higher precision and richness of detail than the seven generalised main categories, and gave a deeper understanding of the TA therapy components.Since sub-categories are included as aspects of the main categories, the latter reliability measure of 46.2% and a kappa coefficient of 0.32 indicate that these should still mainly describe what the assessors jointly considered to be transactional analysis group therapy.
Among the results should also be mentioned that the 20 sub-categories and two main categories (Relations and Pattern) were not coded at all or very little.The assessors had mostly one code for each intervention, which underpins the stability of agreement.That assessors on the whole used only half of the categories may be due to the difficulty of distinguishing and using a relatively large number of categories.It can of course also mean that these were representative of the therapy.A future study with codings based on the more frequent categories might provide better evidence for this.
The therapist's adherence to a method forms an important part of the result, because interventions are linked to a categorised method.The crucial issue is how purely the therapist manages to stick to 'official' theory.Canestri (2006) argues that there is a possibility that therapists develop, through further education, practical applications and personal experience, a 'private' application of the 'official' method.From the perspective of a methodological appraisal this is a problem.Nevertheless, it may be assumed as likely that the 'official' method forms the background to any new developments that can be observed and identified.Analytical (inferential) statistics have been used to make a correlational analysis of the assessors' agreements (inter-assessment reliability).
Primarily, the percentage agreement has been specified, but kappa coefficients and Odds Ratio were also calculated in order to compensate for the randomness.The significance of the measures will depend on how well the assessors can apply the previously agreed coding alternatives.The training of assessors may have led to forced consensus, which reduces their independence and thus threatens the validity of the coding categories.The validity was based on previous studies (McNeel, 1975) in which different categories were induced from an observed TA therapy.These categories were improved by operational definitions and practical evaluations of the application.
The subsequent coding and data collection were thus linked to the chosen problem and research questions as well as the outcome, which by Holme & Solvang (1997) is considered essential for validity.
The specific conclusion is linked to differences in the coding rate for different categories.Categories with high reliability and a high individual frequency in the assessments show which ones are typical of or specific to TA therapy.Two sub-categories are clearly the most frequent, namely 'specificity/clarity' and 'mutual negotiation'.They are included in the main categories of Language Usage and Contract.The first main category is also individually and jointly the most common one coded with the calculation based on main categories, while the Contract category acquires much less emphasis in the more general assessment.The 'responsibility' and 'train Adult' categories, which both belong to the main Reality Testing category, were coded frequently by both evaluators and have also received many joint markers.'Make feeling statement' obtains a great deal of agreement in the coding, even though no assessor has coded it individually to the same extent.The main Strokes category is the third most frequent and has high representation within the sub-categories 'talking to Parent projections', 'on the side of the Child' and 'support/permission'.
A number of sub-categories have been clearly emphasised in different degrees by the author and the independent assessor.'Connect past scenes to present impasse' is marked more frequently by A, while I has coded 'confront racket feelings', and 'game analysis' more frequently.This is probably due to different perceptions of the content of the categories, since these three are transactional analysis knowledge categories, where individual knowledge and experiences have gained greater importance.
In conclusion, one can assume that the nine most frequent sub-categories show the TA categories that are most likely specific to transactional analysis psychotherapy.The question, however, remains whether these categories can also be found in other therapies and can be excluded because they may be assessed as being non-specific or 'common factors'.The therapeutic alliance is usually mentioned in this connection along with the therapist's acceptance, understanding, rational explanations and encouragement.Holmqvist (2006) and Lundh (2006) have discussed the difficulty in psychotherapy effect research of distinguishing the characteristic theory-related ingredients from common and temporary ones.Messer & Wampold (2002) as well as Luborsky et al (2002) showed that the differences between methods were small and that many 'psychotherapy interventions' are shared by most therapies.The TA method has an integrated or eclectic approach, which complicates making a clear distinction from other therapies.
Although a great many therapeutic techniques and approaches are shared, they may be practised in a way that is specific to the therapy form.Since this is not an effect study, I will confine myself to discussing what may be specific to TA, regardless of whether it is effective or not.Starting from the operational definitions, one can see that the most frequent and reliable category, 'specificity/clarity', is available in all therapies.Another highly frequent and reliable category, 'make feeling statement', can be regarded as a recurrent element in most therapies.However, 'mutual negotiation', which is often coded in agreement by the assessors, is considered to be TAspecific.TA is a contractual therapy form where mutual negotiation is an important ingredient in therapeutic cooperation.
The idea of a contract is also available in cognitive behavioural therapy (Beck, 1976(Beck, , 1995) ) but does not permeate this form of therapy and the therapist's approach as profoundly as in TA.Another equally preferred category is 'talking to Parent projections'.It was coded in the so-called double-chair work, which is a technique originating from Gestalt therapy but is developed as a special technique in TA therapy.The therapist is schooled in this specific TA direction, named Redecision Therapy (Goulding & Goulding, 1975), which is a likely explanation of the category's high priority.
With 'slight' (Landis & Koch (1977) reliability the two assessors have agreed about having observed transactional analysis psychotherapy in a group.This means that a description of transactional analysis psychotherapy in general terms could be made.
A large number of categories were coded a little or not at all, while a few were coded a great deal by the two assessors.Among the most frequently coded, 'mutual negotiation' is considered to be the most specific category in the TA method.
In the light of all the TA concepts and techniques that are highlighted in the study TA practitioners will find scientific support in their application of TA.
Roland Johnsson, lic.psychologist, lic.Psychotherapist, MSc, MSW, MA, Teaching & Supervising Transactional Analyst (Psychotherapy) can be contacted on roland@livsterapi.seAppendix A: Operational definitions of the study's seven main categories and 42 sub-categories

I. Contract
The client and therapist mention, quote and/or negotiate treatment contracts in some form.

'Mutual negotiation'
The therapist starts a contract-related negotiation or responds to a negotiation initiated by the client.

'Behavioural description'
The therapist defines and substantiates a contract in behavioural terms.

'Confront Parent' contract
The therapist confronts the communication from clients in which they express their goals from a Parent position instead of listening to their own natural needs.

'Refer to contracts'
The therapist refers to the original written treatment contract or a daily contract.

II. Strokes
The therapist draws attention to a statement which testifies to the client's resources or confronts a self-devaluating statement.The therapist requests the client's active stance.

'Stroking strength and health'
The therapist draws attention to new salutogenic behaviours and emotions in the client.

'Repetition of positive strokes'
The therapist repeats a positive assessment of the client, since it seems not to have been understood.

'Change self-harassment to a positive fantasy'
The therapist invites the client to replace self-torture with an enjoyable and positive imagination.
8. 'Careful use of "Will you?"'The therapist asks, "Will you ...?" in order to help clients to actively make their own decisions regarding a behaviour or a life situation.9. 'Not laughing at gallows humour'.
The therapist recognises and confronts a self-devaluating statement from the client disguised as humour.

'Talking to Parent projections'
The therapist speaks during double-chair work with the client while the client is playing the role of mother or father, as though the client were the parent at that present moment.

'Support/permission'
The therapist expresses himself non-judgmentally and encouragingly to help the client dare to express forbidden feelings and thoughts.
12. 'On the side of the Child' The therapist supports the client unconditionally in an attempt to express the needs, hopes and disappointments directed at authority figures from childhood.

III. Language Usage
The therapist asks for or makes a clarification or reformulation in terms of the here-and-now.

'Hearing literally'
The therapist repeats a statement from the client which expresses destructive beliefs.

IV. Pattern
The therapist questions contamination or confusion, or helps the client to formulate a connection between the client's history and the here-and-now situation.

'Separate self from others'
The therapist challenges the client to create a self-image as separate and autonomous rather than inseparably paired with someone else.
19. 'Separate old scene from present impasse' The therapist helps the client to distinguish how current conflict situations reflect similar scenes from childhood.

'Expose myth and magical thinking'
The therapist points out to the client unconscious and early established notions, which continue to control the client in an inappropriate way.

'Use fantasy'
The therapist invites the client to use fantasies and metaphors to playfully get an emotional image of self and own practices.

'Game analysis'
The therapist makes clear to the client in TA terms the destructive social processes the client chooses to follow.

V. Reality Testing
The therapist challenges clients to examine a belief about themselves, others or the world.

'Use of intuition'
The therapist uses inspiration or an intuitive notion as a hypothesis from which clients can explore their actions.

'Train Adult'
The therapist invites the client to reflect upon and evaluate information and identify options for action.

'Responsibility'
The therapist invites clients to accept and take the consequences of the ability to affect their lives.

'Own personal power'
The therapist invites clients to accept the importance of their own choices to achieve a specific goal.

'Own projections'
The therapist invites clients to take in statements on a personal level which refer to something outside of them (e.g."What a nice day!" to "I look nice").

'Use of video'
The therapist plays a video clip to enable the client to hear and see what took place during the therapy.

'Use humour (distancing)'
The therapist uses humour to create a distance to a subject or a situation, which is of advantage for the therapeutic process.

VI Feeling Contact
The therapist makes clients aware of the emotional content in client communications.The therapist stimulates and makes room for living out emotions.

'Make feeling statement'
The therapist invites clients to express themselves verbally about their emotional state.

'Make feeling comments'
The therapist comments on the client's state of mind.

'Express feelings'
The therapist invites clients to express and show their feelings.

'Distinction Between feeling/thinking'
The therapist makes clear to clients that a feeling was asked for but a thought received in response.

'Confront the racket feeling'
The therapist confronts clients when they fall back on habitual negative emotional expressions rather than allowing themselves genuine underlying feelings.

'Discrepancies in body language'
The therapist invites clients to pay attention to the incongruence between what is said in words and what is expressed non-verbally and then asks them to express themselves congruently.

'Use bataca'
The therapist invites the client to use a padded bat (bataca) to stimulate contact with and living out of anger.

'Use present tense'
The therapist stimulates clients to a more intensive feeling contact by encouraging them to use the present tense in descriptions.

VII. Relations
The communication is disturbed or interrupted by here-and-now-inadequate responses by one of the people involved.The therapist fails to pick up and respond to the client's direct or indirect appeal for help.

'Boundary violation'
The therapist or the client goes beyond the limits agreed for the therapy.
. The social context thus consists of what we are saying, what we accomplish with what we are saying, and what impact what is said has on us.
36.'Double-chair work'The therapist invites the client to do what is called double-chair work.(The client improvises under the therapist's guidance a real or imagined situation usually taken from the client's history, where childhood authority figures are included and where the client may act in all of the roles. 39. 'Transference' (crossed Parent-Child transactions from client) The therapist confronts clients when they express feelings and beliefs towards the therapist, which originate in their relationship to authority figures from childhood.40.'Counter transference' (crossed Parent-Child transaction from therapist) The therapist expresses feelings and beliefs towards the client which belong to the therapist's own relationship to authority figures from childhood.41. 'Alliance rupture' (crossed transaction on an Adult-Adult transaction) There is both an interest in what can generally be considered to describe TA and also in what is specific to the method.
from the Ethical Research Committee of Lund Universities meeting 20 March 2002) confirming ethical permission to use the clinical material for research.

Table 1 :
Discourse space for the clients and the therapist (the number of conversational turns/interventions) for each therapy session and in total as well as the therapist's percentage of discourse space Note.M = mean in percentage

Table 2 :
Ranking based on the number of matching codings by main and sub-category calculation, respectively.Priority 1-7

Table 3 :
Ranking of the frequency of the assessors' individual and matching codings of the sub-categories.Priority 1-9

Table 4 :
Examples of the main Contract category and 4 sub-categories with assessors' individual and matching codings Note.I = Independent assessor, A = Author, Agree = Agreement

Table 5 :
The most frequent sub-categories under each main category

Table 6 :
The most frequent codings of the different main categories during different phases of therapy

Table 7 :
Percentage distribution of the number of assessor codings for each therapist intervention, at first, second and third codings Note.I = Independent assessor, A= Author

Table 8a :
Calculation of kappa quotient (k) and OR quotients (OR) for 4 main and 22 sub-categories Note.Discordant = not in agreement, NE = not estimated, OR = Odds Ratio, kappa (κ) = Cohen's kappa coefficient, A = Author, I = Independent assessor

Table 8b :
Calculation of kappa quotient (k) and OR quotients (OR) for 3 main categories and 20 sub-categories Based on the main categories, four quotas have 'slight agreement', while one had 'poor', and two had 'moderate agreement'.The distribution in the sub-categories was 15 quotas each in 'poor' and 'slight agreement', and six quotas with 'moderate agreement'.

Table 9 :
Ranking of main category agreement according to kappa (κ) and OR values, as well as percentage agreement (%)