Anthony conducted this project as part of his doctoral studies in counselling psychology, under Doris’s supervision. Although a majority of the work was carried out by Anthony, Doris played an instrumental role in its conception and development. As well as providing a sounding board for Anthony’s developing ideas, Doris engaged with the data to help add breadth to its interpretation and the resulting theory. We will attempt to provide a sense of this collaborative process here. Our use of the word “we” implies that we discussed the work prior to its completion by Anthony. We also specify where either of us contributed to the work individually.

We chose our topic—psychotherapists’ handling of sexual attraction to clients—after Anthony stumbled upon the relevant literature, almost by chance, during a curiosity-driven search for information about the role of gender in psychotherapy. After a cursory review, Anthony realised that the literature on psychotherapists’ sexual attraction to clients was underdeveloped, fragmented and devoid of a cohesive theoretical framework. Given that we both had some experience in conducting qualitative research in the past, we saw an opportunity for the development of a grounded theory that could go some way toward answering what had become our central research question: How do psychotherapists handle experiences of sexual attraction to clients?

To set the groundwork for our project, we clarified our conceptual frameworks by writing individual passages about: our research philosophies; our experiences of, training in and attitudes toward sexual attraction to clients and psychotherapy in general; and our current understanding of the relevant extant literature. The purpose of doing so was fourfold: first, to articulate some preliminary concepts (often referred to as ‘sensitising concepts’), which we would use as points of departure from which to ask questions of and make comparisons among our data; second, to enable us to remain sensitive to and minimise potential areas of bias in our ways of interpreting and analysing data (commonly referred to as ‘bracketing’); third, to begin the process of self-reflection, or ‘reflexivity’, which was to continue throughout the research process; and, fourth, to contribute to the reader’s ability to identify the lens through which we interpreted and analysed our data, and thus determine the “extent to which [he or she] is able to generalize the findings … to his or her own context” (Morrow, 2005, p. 252). To see excerpts of these passages, please visit http://www.psy.mq.edu.au/staff/dmcilwain/arcuri_collab.html.

Given the sensitive nature of our research, we took exhaustive measures to ensure that it would be conducted in an ethically sensitive manner. Nonetheless, we were concerned that an interview study inquiring directly into people’s experiences of erotic desire might prove intolerably alarming for our university’s Human Research Ethics Review Committee. This fear was founded on a previous experience with the Committee in which an anonymous, web-based survey of psychotherapists’ experiences of sexual attraction to clients was held up for over nine months until the Committee was satisfied that the participants’ identities could not conceivably be revealed either to us or to third parties.

To avoid similar delays in our present research, we proposed to explore psychotherapists’ handling of ‘hypothetical’ rather than ‘actual’ experiences of sexual attraction to clients, thus minimising the potential for the participants to be identified as psychotherapists who have indeed experienced such sexual attractions—though of course attraction is not a misdemeanour. Although the Committee was agreeable to this approach, we feared initially that it may compromise the integrity of the qualitative data (a point that remains debatable). We guessed, however, that talking about ‘hypothetical’ rather than ‘actual’ erotic desires would be less confronting for our participants and would thus allow them to engage in more open dialogues.

We also suspected that people would be more likely to volunteer for research involving interviews about ‘hypothetical’ rather than ‘actual’ sexual attractions. But before our suspicions could be put to the test, we needed to reach our target group and bring our project to their attention. In the light of our ‘insider’ knowledge of psychologist networks, Anthony distributed a brief advertisement via two media regularly consumed by psychologists: a quarterly print newsletter sent to members of the Australian Psychological Society (APS) residing in the state of New South Wales, and the APS’s internet classifieds service available to both members and non-members. In an attempt to access unaffiliated psychologists, Anthony also distributed our advertisement via group email to students enrolled in our university’s postgraduate psychology training programmes, and included in all our advertisements a request that the recipient pass on the advertisement to at least one other psychologist (a technique known as ‘snowball sampling’).

Although we were hopeful that our advertisement would be well-received, we had concerns that some psychologists would be reluctant to discuss sexual matters, hypothetical or otherwise, with Anthony, a peer of theirs. Indeed, having seen the advertisement, two of our senior colleagues reacted less than favourably and somewhat indignantly to the nature of our study. But we were not to be discouraged: our recruitment efforts yielded interest from 36 psychologists, a number beyond our expectations. From 35 of these psychologists we were able to collect some basic screening information which we would use later to assist with our sampling decisions. Because of the delicate nature of our topic, we limited this information to the following: age, gender, psychologist registration status, psychology specialisation, length of time practising as a psychotherapist, and theoretical orientation. Please see Table 1 for a summary of this data.

Table 1. 
Characteristics of psychologists interested in participating

Note. ¹ Three participants reported multiple specialisations: counselling/clinical, counselling/organisational, and clinical/forensic. Each is counted separately in the table. ²Participants commonly reported multiple theoretical orientations.

We anticipated that interviewing between 8 and 12 of the 35 psychologists interested in participating would provide us with ample data to generate a grounded theory. We decided to interview first a person who we thought was likely to provide rich data from which to elucidate our sensitising concepts and discover new concepts—an approach known as ‘purposive sampling’. For the purposes of our study, we believed this person to be one who had a psychodynamic or humanistic theoretical orientation and was able to reflect openly on his or her experiences within the therapeutic relationship. To identify such a person, we reviewed the screening information we had collected earlier from our pool of potential participants. Anthony also had the advantage of being personally acquainted with a number of these people, which gave us additional information with which to gauge their suitability for being interviewed first.

Prior to interviewing the first participant, we collaboratively developed an interviewing guide beginning with a series of questions designed to collect demographic information in addition to that collected via the screening process. Our intention behind the inclusion of these additional questions was to provide points from which to ask questions and make comparisons both within the interviews and during data analysis. We followed these brief demographic questions by a non-confrontational inquiry into what it is like being a psychotherapist, which was intended to ease the participants into the interview process. Next, we included a brief passage directing participants to imagine being sexually attracted to a client who has qualities characteristic of people to whom they are usually attracted. In this passage we incorporated the suggestion that the participants could go beyond their imaginations and draw from their actual experiences if they wished, without needing to disclose whether they were doing so.

The remaining items in the interviewing guide evolved over the course of the study. The initial guide included only a few general questions organised around our initial sensitising concepts. We tailored our subsequent interviewing guides to each individual participant, whom we selected on the basis of his or her supposed ability to ‘flesh out’ and clarify the concepts that emerged through our analysis of preceding interviews. To demonstrate this process of ‘theoretical sampling’, we will draw upon a single example from our study that illustrates how concepts arising from the data could form the basis for further sampling decisions about who could be included in the study to maximise a variety of positions with regard to that issue. Many such concepts arose; we use this as illustration.

Early in our analysis, we identified the concepts of ‘therapeutic self-disclosure of the sexual attraction’ and ‘theoretical orientation’. One of our participants had stated that his psychodynamic theoretical orientation would allow him to self-disclose his sexual attraction to a client for therapeutic benefit. In our analysis of this interview, we became curious about how psychotherapists with other theoretical orientations would feel about such a self-disclosure. Using our knowledge of our pool of potential participants, we selected people from other theoretical orientations and posed this question to them.

Through this process, we gathered rich and varied information that allowed us to hypothesise about the role of theoretical orientation in psychotherapists’ decisions about whether to therapeutically self-disclose their sexual attraction to clients. We continued to sample and theorise in this way until we reached ‘theoretical saturation’: the point at which we thought we could learn nothing new and relevant about these concepts or their relationships with each other. After 11 interviews, when we had reached theoretical saturation for all of our concepts and their relationships, we stopped collecting data.

Such a sample size might appear small to a quantitative researcher; however, participants for qualitative research are expected to represent an experience or detailed knowledge of a particular process rather than an entire population. Because one participant can provide a rich and detailed account of an experience, qualitative data from a relatively small number of participants can result in a dense and substantial representation of the experience in question. Adding participants to qualitative research is of value only when that participant can add something new to the explanation of the experience under investigation. Therefore, the importance of sample size in qualitative research is secondary to the quality and richness of the data collected.

Although our data collection focussed primarily on our participants’ narratives, it also involved a record of the interpersonal context of Anthony’s interviews with the participants. For each interview, Anthony created a ‘transcript file’—a document divided into three columns: the middle column contained the verbatim transcript of the interview; the rightmost column provided space for hand coding of the raw data (a process which we will explain in greater detail in the following section); and the leftmost column provided space for corresponding notes about the interview process, including the participant’s and Anthony’s interactions and non-verbal communications, and a critique of Anthony’s questioning techniques.

In addition, Anthony developed a ‘methodological file’ (Browne & Sullivan, 1999) containing detailed methodological information about each interview, cross-referenced with the methodological notes in the corresponding transcript file. This information included: how Anthony came into contact with the participant; how he made arrangements for the interview; the quality of the rapport developed; how and why we selected the participant; a description of the participant; how Anthony may have been perceived by the participant; how Anthony perceived the participant; the suitability of the interview setting; the timing of the interview; Anthony’s interviewing technique; the use of recording equipment and note-taking; and how the interview finished. Within the methodological file, we reflected on how each of the above areas may have impacted on the data collected, and made an overall assessment of the validity of the information obtained in the interview. We then used this methodological file to assist us in our interpretation of the data.

In working with our data, we followed closely the Grounded Theory coding procedures defined by Strauss and Corbin (1998). Their particular approach to Grounded Theory appealed to us because of its clearly articulated and easy-to-follow guidelines, and its inclusion of a paradigm within which to organise our data (which will be discussed later in greater detail). We decided to implement Strauss and Corbin’s (1998) suggested procedures by hand, rather than with the assistance of computer software, primarily because we shared their debatable view that our use of such software would interfere with our organic analysis of the data.

Instead of analysing our data once it had been collected in its entirety, we examined each interview following its completion, ahead of conducting the next interview. To enhance the reliability and reflexivity of our data analysis, we the researchers coded the data independently of each other in the following way. Anthony coded each interview in extreme detail using Strauss and Corbin’s (1998) suggested techniques described later in this section. After Doris had read each transcript until she had reached a point of ‘immersion’ in the data—where she was aware of when and where each comment made by the participant had occurred in the interview—we came together to review Anthony’s specific coding decisions. Doris raised questions about the occasional comment that Anthony might not have seen in a particular light, and about the ways in which he had made theoretical links within and across interviews. We then ensured that the way Anthony had coded the particular comment or hypothesis captured the different perspective Doris had brought to it.

Following his reading of the transcript several times to familiarise himself with the data, Anthony’s first explicit data analytic procedure involved identifying concepts via ‘open coding’. Here, he used the rightmost column of the transcript file (which we introduced in the previous section) to highlight key words and phrases within the participant’s narrative that appeared to represent new or developing concepts. In labelling concepts, Anthony borrowed our participants’ words, wherever possible, but occasionally developed his own paraphrases or drew from existing constructs. During open coding of the early interviews, Anthony scrutinised the data on a phrase-by-phrase or even word-by-word basis. As he became familiar with the emerging concepts through the analysis of several interviews, he tended toward coding whole sentences or even paragraphs.

For each concept identified (see Table 2 for an abbreviated list), Anthony created or added to a series of memos in which he explored the potential meanings of the concept in question and attempted to delineate its properties (or characteristics) and dimensions (the range along which these properties vary). Gaps he found in properties and dimensions indicated the areas remaining to be explored with future participants (whom we would select via theoretical sampling, as described in the previous section). To illustrate open coding with an example, during his analysis Anthony identified the concept, 'intensity of the sexual attraction', which ranged in dimension from low to high, where lower intensity sexual attractions tended to be characterised by enjoyment of the desire, and higher intensity sexual attractions typically involved feelings of anxiety and guilt about the desire.

Once Anthony began to notice relationships among concepts (as early as during coding of the first interview), he incorporated ‘axial coding’ into the analysis strategy. In doing so, he started to link concepts at the level of their properties and dimensions, and identify sub-concepts—those that are subsumed by, and become properties of, higher order concepts (e.g., the concept ‘theoretical orientation’ became a sub-concept and thus a property of ‘psychotherapist characteristics’). To assist Anthony with this intimidating task, he employed an adapted version of Strauss and Corbin’s (1998) ‘paradigm’, which provided a logical method of gathering and ordering concepts. Our adaptation of this paradigm retained the original three components into which concepts are arranged: (1) ‘conditions’, the sets of events or circumstances that create situations relating to a phenomenon; (2) ‘actions/interactions’, the ways in which persons handle the situations encountered; and (3) ‘consequences’, the results of actions/interactions. Because of the ‘hypothetical’ focus of our study, we added to the paradigm: (4) ‘decisions regarding actions/interactions’; and (5) ‘expected consequences’.

Using this paradigm as a base, Anthony created memos in which he made relational statements or ‘provisional hypotheses’ about the phenomenon we were investigating. In these provisional hypotheses, Anthony made direct references to the concepts identified through open coding, and specified the relationships among the properties and/or dimensions of these concepts. For example, relatively early in our data analysis, Anthony developed the following provisional hypothesis (concepts are highlighted in bold, their properties [where included] are underlined, and their dimensions are italicised): “when psychotherapists deem their sexual attraction unmanageable (for example, when it is highly intense and thus characterised by pre-occupation with the client and/or by feelings of anxiety and guilt), they are likely to decide to discontinue psychotherapy with the client, because they expect that continuing psychotherapy in such a situation would lead to negative consequences (such as their inability to remain objective about the client’s issues).” Here, the 'conditions' (highly intense, unmanageable sexual attraction) create the context for a 'decision regarding actions/interactions' (discontinuing psychotherapy), which is made with the intention of avoiding 'expected consequences' (compromised objectivity about the client's issues).

We validated and further elaborated our concepts and provisional hypotheses through our continuing analysis of incoming data, which we collected via theoretical sampling. When we noticed that no new concepts or relationships among concepts seemed to emerge during coding (i.e., the point at which ‘theoretical saturation’ occurred), we stopped collecting data (as discussed in the previous section in the context of theoretical sampling). We acknowledged, however, that we could not possibly reach theoretical saturation for concepts that represented our participants’ ‘actual’ experiences, as we were not permitted ethically to inquire into such experiences. These concepts and their relationships, therefore, remained provisional, but were adequately robust to be integrated into our final theory.

Table 2

Abbreviated list of concepts identified

Our process of arriving at theory involved us cycling somewhat haphazardly through the various coding procedures numerous times. From very early in our analysis, not only did we identify concepts through open coding and articulate relationships among our concepts via axial coding, we also began to discover the ways in which these concepts and their relationships changed over time, thus ‘coding for process’. More specifically, Anthony began to identify sequences of evolving ‘actions/interactions’, which were the result of changes in ‘conditions’, which themselves were the result of either external forces or the ‘consequences’ of ‘actions/interactions’.

For example, during data analysis, Anthony hypothesised that for a sexual attraction perceived as interfering with therapy but deemed manageable nonetheless (the ‘conditions’), a psychotherapist might attempt to minimise the impact of his or her attraction on therapy by mentally distracting himself or herself from the attraction during sessions (the ‘action/interaction’). Anthony theorised further that, when such an attempt is unsuccessful (the ‘consequence’), the psychotherapist is likely to perceive the sexual attraction as less manageable (a change in ‘conditions’), which necessitates a new approach to handling the sexual attraction (the evolving ‘action/interaction’). Finally, Anthony posited that this cyclical process continues until either the sexual attraction is managed successfully or the therapy is terminated. To assist in identifying these processes, Anthony created memos and diagrams explaining their flow.

Once we thought we had comprehensively identified the concepts, sequences and processes relating to the phenomenon under investigation, we were ready to begin our attempt at articulating our grounded theory. Anthony’s first reaction to this impending task, though, was to wonder if it were indeed possible. Having watched the piles of analytic paperwork grow, he had developed an insidious sense of anxiety and dread: how would we ever turn this complicated jumble of words, hypotheses and diagrams into a coherent and whole theory? In Doris’s attempts to manage Anthony’s fears, she encouraged him to draw strength from our knowledge that those before us had found a way; to continue to rely on the data analytic techniques prescribed by those who had developed the grounded theory method; and, humbly, to trust ourselves to bring our theory into being.

‘Selective coding’ allowed us to realise our theory by providing us with a series of techniques for integrating concepts, and their relationships and processes, into a cohesive whole. Anthony’s first task was to identify a central concept that represented the main theme of our research to which all other concepts could relate. Perhaps because of the specific nature of our research, our central concept was, simply, ‘sexual attraction to a client’. Our next step involved a combination of tasks that we cycled through until we arrived at a draft of our overall theory.

Anthony reviewed and sorted our memos and diagrams into physical piles strewn across a rather large floor to provide a topographical sense of where all of our concepts belonged in relation to each other. In constructing diagrams to record this map, Anthony discovered, importantly, that to create a single diagrammatical representation of our theory he needed to pan back far enough to allow all of our hypotheses to be subsumed logically under a relatively uncomplicated yet all-encompassing theoretical scheme. As a result, his diagrammatical snapshot of our overall theory appeared exceedingly general, but invisibly embedded within this diagram was a multitude of detailed and specific hypotheses. To give voice to these hypotheses and place them clearly and appropriately within our overall theory, Anthony wrote an informal yet detailed ‘storyline’ of our grounded theory—a free-flowing narrative incorporating all of our concepts, their relationships with each other, and the ways in which they change over time.

A short excerpt from our rather long storyline follows: “The process of handling sexual attraction to a client involves an appraisal of the sexual attraction, which contributes to the perception of the manageability of the sexual attraction, and provides a context within which to assess the consequences and ethicality of options for handling the sexual attraction. In turn, the appraisal of the sexual attraction, the perceived manageability of the sexual attraction, and the assessment of the consequences and ethicality of options for handling the sexual attraction contribute to the psychotherapist’s decision about how to handle the sexual attraction.”

On multiple occasions we tweaked, altered or even discarded and recommenced our attempts at the tasks described above, but with each reiteration we learned something new about how our concepts did or did not relate to each other as a whole. For example, it seemed to Anthony when he attempted to theorise from a distance that the particular context of a sexual attraction (e.g., its intensity, its timing, the degree to which it interfered with the therapy) contributed directly to the way in which it was handled by the psychotherapist. We learned by reengaging with and discussing our memos and the raw data, however, that the context of the sexual attraction instead contributed indirectly via the psychotherapist’s unique appraisal of its value and relevance to the handling process.

When we had arrived at an adequate draft of our theoretical scheme, we tested its credibility by showing our ‘storyline’ to five of our participants for comment on how well it fitted their experiences of, and attitudes toward, handling sexual attractions to clients. Largely, these participants indicated that our theory was consistent with their experiences and attitudes, but where they suggested that parts of our storyline expressed misinterpretations of their experiences or attitudes, we adjusted it accordingly.

Our final data analytic task involved formally articulating the results of our research in writing. Although it was Anthony’s task to write up this project as part of his doctoral studies (Arcuri, 2008), this writing will be represented here as a joint effort, given Doris’s numerous cycles of reading, feedback and re-reading Anthony’s drafts. To report our project, we drew from each of the various analytic resources we had compiled: our diagrams, storyline, memos, methodological file, verbatim transcripts, and even our memories. Using our diagrams (pinned to a noticeboard) as orientation, we created a series of headings and subheadings reflecting the concepts and sub-concepts we had discovered. Under these headings we pasted the corresponding parts of our ‘storyline’, which we then formalised and elaborated upon by reviewing our memos and methodological file to ensure that all concepts were accounted for and represented appropriately.

In our descriptions of our concepts, we gradually introduced our hypotheses explaining their relationships and evolving processes over time. Following our description of each group of sub-concepts, we summarised their sometimes evolving relationships both with each other and with other groups of sub-concepts that we had introduced previously. We concluded our writing with an overall summary of our theory, incorporating all important concepts, relationships and processes, and providing real-world examples of how these operate. We supplemented our written account with a general diagram of our theory and two additional specific diagrams of particularly complex processes embedded within the theory.

In order to provide our readers with clear evidence for our representation of concepts and the relationships and processes in which they were involved, we wove the participants’ narratives into our writing and quoted at length on occasion, giving prominence to such text via “italics within double quotation marks”. Our desire to preserve smooth grammatical flow in our writing led us to edit our participants’ quotations slightly. Although we made extensive efforts to ensure that we preserved our participants’ original meanings, it is possible that our (or perhaps only Anthony’s) quest for grammatical perfection embellished or suppressed nuances in our participants’ communications. Beyond our idiosyncratic approach to representing their narratives, we were required to alter particularly self-revelatory quotations to protect the identities of our participants.

We continually refined our theory throughout our write-up. Where we discovered that our theory demonstrated poor consistency or logic, we reviewed our memos and diagrams and adjusted our theory accordingly. In a similar vein, where we noticed that our concepts were less well developed than we had thought, we revisited our memos and raw data with the aim of finding overlooked pieces of data that could fill the newly discovered gaps in our concepts. We dropped concepts that we considered extraneous to our theory or poorly developed, with the exception of concepts relating to ‘actual’ experiences, for which we could not reach theoretical saturation (as described in a previous section) but which we retained in our final theory with the proviso that they remain provisional and subject to further investigation. As a final test of the credibility of our theoretical scheme, we compared our theory against the raw data to ensure that it was able to explain all cases. Again, where it was not, we adjusted our theory accordingly.

In our final report, we positioned the majority of the writing described above within our third chapter, the results, which we wrote first. Next, we illustrated our methods (chapter two) by providing a general explanation of the grounded theory approach with reference to the methodological literature, followed by a detailed description of our research process in a manner similar to the report you are reading currently. After we had articulated our method and results, we wrote our opening chapter, in which we comprehensively reviewed the history of and literature associated with our topic, identified its numerous gaps, and eventually arrived at our research question. We concluded our writing with our fourth chapter—a summary and discussion of our theory in the context of the existing literature, of its limitations, and of its implications for psychotherapy practice, supervision, education, and research. This ‘world-backwards’ way of writing fits grounded theory well, even if it seems unfamiliar to those schooled in other approaches.

Because we had the luxury of a self-determined word-limit within which to report our research, we were allowed indulgently to explain and expand upon multiple nuances in our theorising. Later, however, in preparing to present our work at conferences and meetings, we were awakened to the reality of time and space constraints, and were required to deliver our theory in an abridged and concise form. To meet these requirements, we chose to present a very brief rationale for our study, followed by a cursory explanation of the methods employed to achieve it. We decided to dedicate the majority of our time to outlining the overall structure of our theory, and providing several engaging and audience-appropriate anecdotes from our participants’ narratives that were representative of our key hypotheses. We opted to conclude our talks with a summary of the implications of our findings and an opportunity for a reflective discussion with our audience. 

Arcuri, A. (2008). Psychotherapists’ handling of sexual attraction to clients: A grounded theory. Unpublished doctoral dissertation, Macquarie University, Sydney, Australia (available at Macquarie University Library Thesis Collection; Call Number: RC455.2.E8 .A73).

Arcuri, A., & McIlwain, D. (2014). Skilled handling of sexual attraction in therapy: a grounded theory of what makes the difference. In M. Luca (Ed.), Sexual attraction in therapy: clinical perspectives on moving beyond the taboo: a guide for training and practice (pp. 153-172). West Sussex: John Wiley & Sons, Ltd.

Browne, J., & Sullivan, G. (1998). Analysing in-depth interview data using grounded theory. In V. Minichiello, G. Sullivan, K. Greenwood & R. Axford (Eds.), Handbook for research methods in health sciences (pp. 575-611). Frenchs Forest: Addison-Wesley.

Morrow, S. L. (2005). Quality and trustworthiness in qualitative research in counseling psychology. Journal of Counseling Psychology, 52(2), 250-260.

Strauss, A., & Corbin, J. (1998). Basics of qualitative research: Techniques and procedures for developing grounded theory (2nd ed.). Thousand Oaks: Sage.

1. Three years on: research developments

Relevant subsequent research by you or others – where was it done and how different was it to your earlier project?

In the last three years, relevant research is limited to five studies published in a 2014 book edited by Maria Luca and entitled, Sexual attraction in therapy. Clinical perspectives on moving beyond the taboo: A guide for training and practice. Included among the research chapters in this book were:

• Our grounded theory of Australian therapists’ handling of sexual attraction to clients.
• A discourse analysis of British heterosexual males’ experiences of sexual attraction to clients.
• A grounded theory of experienced British therapists’ reactions and attitudes to sexual attraction in therapy.
• An interpretive phenomenological analysis of how British therapists self-disclose their sexual feelings to a client.
• A narrative analysis of British systemic family therapists’ experiences of sexual attraction to clients.

What does it add to or alter in the understanding of your project’s topic?  What is new and why? (Different method? Different theoretical approach? ….)

Given the historical paucity of research in this area, each of the studies mentioned above adds to and broadens our understanding of the phenomenon of sexual attraction in therapy. Because of their qualitative methodologies, rather than altering our understandings, these projects added breadth to our area of study by offering new contexts for exploration (e.g., British therapists, family therapists, self-disclosing therapists). They also provided further depth to the field by offering varying methodological and individual lenses through which therapists’ experiences were interpreted.

2. In hindsight

If you were to design and conduct that project now, what would you do differently? Why? (Different method or location? Different theoretical approach? ….)

If we were to begin our work afresh, Anthony might loosen the structure of his interviews and have freer conversations with the participants, and we might add Doris and others as interviewers on the project. We suspect that this fresh approach would allow for the emergence of a greater range of novel concepts, less susceptible to the potentially restrictive lens of a single person.

We might also write more collaboratively from the start, which was not possible originally given that the work was Anthony’s doctoral thesis. We found recently that when adapting the thesis for the book chapter we mentioned earlier, our very collaborative writing approach brought a sharper yet freer focus on what was important to communicate to our colleagues.

What did the project need to be more satisfying to you, more adequately addressing questions that matter?  (Feel free to grump about what you should have done if you’d had adequate time or money!)

We think it would have been good to include more psychoanalytic therapists and more CBT therapists so that we could explicitly have considered what seemed to be a trend in our data. It seemed to me [Doris] that there are marked differences in the way that they view the occurrence of sexual attraction (from the interviews that we analysed) and also in how ‘in the room’ attraction was. The CBT therapists seemed to think that if only they were thinking or feeling it, that somehow it was not yet in the room unless it was explicitly disclosed. This is an intriguing difference in terms of the ‘privacy of mind’ between CBT therapists and others (like Gestalt therapists, or more psychoanalytically trained therapists). The latter seemed to have more awareness of emotional leakage of expression in terms of micro-momentary facial expressions, and the way that attraction to a client might prompt unconscious shifts in topic to protect the feelings of the therapist rather than what is beneficial for the client, or may even promote the posing of questions that would not be asked but for attraction.  In other words, some therapists seemed very alive to the tacit ways that attraction could ‘enter the therapy’ without explicit verbal disclosure.

3. Software tools

Looking back, could you have used the qualitative software available to you more effectively – and how? (Please do honestly contribute any reflections on your own or team members’ software use and advice to others.)

We chose not to use the available qualitative software because we feared that doing so might interfere with our organic analysis of the data. In hindsight, we might have been more open and committed to learning about how the software could have helped us without interfering with our work. Without having detailed knowledge of the software available at the time, we nonetheless expect that at the very least it could have helped us to organise our masses of cumbersome data.

How would current versions of software tools have helped?  (The answer may be simple – e.g., not at all! – or quite complex, if new tools have appeared to do what you were unable to do three years ago.)

We think that we may have used NVivo to tally themes and subthemes. However, one’s material remains a little trapped within that format once it is generated, and so more malleable forms of software would be welcome.

Forms of software that permit the ready mapping of images and visual linkages between concepts would be helpful.

In the seven years since completing this study, I have worked in cannabis research, utilising a combination of qualitative and quantitative methods to gain a better understanding of people's cannabis use, related problems, and treatment. Additionally, I have taught a seminar series on 'Taboo Desires' in Doris's undergraduate Personality subject, and taught and provided external supervision to some of Doris's qualitative research students. And, throughout all this, I have practiced as a psychotherapist, which is my full-time vocation at present.

Email: info@dranthonyarcuri.com.au

My early training in psychology at Otago, NZ, encouraged me to pursue the questions that interested me and find the methodology that enabled me to ask those questions. I found it hard to get support for qualitative methodology within academe and left unanalysed a full sampling of open-ended questions to Rajneeshi devotees in New Zealand while teaching at Auckland University. By the time I was doing a PhD at University of Sydney, I was driving round the far-flung reaches of NSW administering questionnaires to people involved in New Religious Movements. I left a few open-ended questions, but there was still no ready way to approach this in my mentor network. My fascination with the words as they are spoken (and by whom) in my psychoanalytic training meant that I became a default qualitative supervisor in my position in Psychology at Macquarie University. I learnt from the many students I have worked with who have pursued questions not amenable to laboratory study, or where theory and evidence was too thin on the ground for a testing of details. They asked questions about what therapeutic success might be, how people are prepared to see personality in bodies, how migration from South Africa impacted on individuals and couples, how refugees arrived psychologically in Australia, how letters might enhance therapy, how masculinity is viewed, what expertise is seen to be in therapy now, and, with Anthony, how people try to handle and navigate sexual attraction to clients. We saw how letting the data speak, by bracketing our hunches and hypotheses, is a rich source of theory generation. We found ways of 'cooking' so that the specific findings let more abstract categories emerge, and their inter-relations form the start of a theory. I like the freshness and excitement of such work alongside my more statistical research, teaching, psychotherapy, pottery, poetry and yoga. 

Shortly after writing our ‘Looking Back’ reflections, Doris fell ill and over the following several months, as gracefully as ever, she navigated the undulations and devastations of the illness’s course until, in April 2015, it took her life from ours. Her indelible legacy is very much alive both in her erudite works and in the many lives upon which she left the mark of her ethereal spirit.

Anthony Arcuri