Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 1
THIS MANUSCRIPT IS A PREPUBLICATION DRAFT OF THE PAPER:
Sánchez Prada, A., y Beyebach, M. (2014). Solution-focused responses to “no improvement”: A
qualitative analysis of the deconstruction process. Journal of Systemic Therapies, 33, 78-91..
Available online at: http://guilfordjournals.com/doi/pdfplus/10.1521/jsyt.2014.33.1.48
Solution-focused Responses to “No Improvement”:
A Qualitative Analysis of the Deconstruction Process
Pontifical University of Salamanca
Universidad Pública de Navarra
Andrés Sánchez-Prada, Department of Psychology, Pontifical University of Salamanca,
Salamanca, Spain; Mark Beyebach, Partners for Collaborative Change, Miranda de Azán, Spain.
This study was completed as part of the first author’s doctoral dissertation at the Pontifical
University of Salamanca. We gratefully acknowledge the patient editing of Janet Bavelas, as
well as the suggestions of the other contributors to this special section.
This article, as those included in this special section, focuses on the relevance of research
for practitioners. Therefore, many specialized research details were omitted because of space
limitations. Complete methodological information is available from the first author: Andrés
Sánchez-Prada, Department of Psychology, Pontifical University of Salamanca; Compañía 5,
37002 Salamanca, Spain; email@example.com.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 2
When a client reports “no improvement” since the previous session, one response for the
therapist can be to deconstruct this description and seek improvements, however small. In this
study, a qualitative, discovery-oriented method examined the process of deconstruction in eight
solution-focused brief therapy sessions where clients had initially reported no improvement. The
findings suggest that the deconstruction of initial reports of no improvement is a complex
process and that therapists do not follow a single path, but respond in a flexible way to their
clients´ discourse: They may move directly into deconstruction, elaboration, and consolidation
or may begin indirectly by first connecting with the negative report and preparing for
deconstruction. Overall, maintaining positive (versus negative) topics in the conversations is
important, but other therapeutic topics can be helpful at some points. It may also be useful to
move systematically along a continuum from specific episodes to general evaluations (or the
Keywords: solution focused therapy, deconstruction processes, qualitative research,
clinical decision-making, psychotherapy conversation, solution-focused interventions
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 3
Solution-focused Responses to “No Improvement”:
A Qualitative Analysis of the Deconstruction Process
Since the late 1980’s, many theorists and practitioners in the field of family therapy have
embraced a social constructionist perspective, which includes the proposal that therapeutic
change occurs through language during the process of therapeutic conversations (e.g., Anderson
& Goolishian, 1988; Sluzki, 1992; Tomm, 1987a, 1987b). In this perspective, solution focused
brief therapy (SFBT) can be described as a linguistic process in which therapist and clients co-
construct new realities in order to “dissolve” what clients initially perceive as their problems (de
Shazer, 1994). That is, they co-construct a new conversational context that focuses on the details
of solutions rather than the details of problems.
From the first session, SFBT therapists focus their conversations with clients on goals,
resources, and exceptions to the problem. Indeed, their preferred opening of second and
subsequent sessions is to ask “What is better?” When clients describe some improvement, the
therapist follows this theme, trying to get a detailed description of the improvements and to co-
construct them as something clients have brought about deliberately (O’Hanlon & Weiner-Davis,
1989). However, sometimes clients answer the initial “What´s better?” question by reporting
that there is “no improvement,” that nothing is better, or even that things are worse. In these
cases, SFBT therapists try to deconstruct this initial response and to generate a new description
that includes some kind of improvement (de Shazer, 1988; de Shazer & Berg, 1992). De Shazer
defined this process of deconstruction as “developing doubts about global frames” (de Shazer, p.
101). Several authors have proposed possible questions that might help to deconstruct the initial
global frame of no improvement (Berg & Miller, 1992; de Shazer, 1994; Beyebach, 2006):
Question the initial report: “Are you sure? Is it possible that nothing is better?”
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 4
Ask for smaller changes: “So what is a little bit better?”
Change the time frame: “So last week was awful; what was better the first week?”
Change the context: “So things at school have been rocky. What about home?”
Change the perspective: “What would your wife say is better?”
Reframe improvements: “How come things are not worse?”
Use coping questions: “With things being that bad, how are you coping?”
Use Scaling Questions: “On a scale from 0 to 10, where 10 stands for…”
A Dilemma not Completely Solved
So far, only two studies have examined the effect of the SFBT approach to clients' reports
of no improvement. Reuterlov, Lofgren, Nordstrom, Ternstrom, and Miller (2000) analyzed the
opening and final phases of 93 SFBT sessions. Their main finding was that only 13% of the
clients who had said at the beginning of the session that nothing was better were categorized as
improved by the end of the session. The other 87% confirmed their initial negative report by
giving the same or even a lower number on the progress-scaling question. Reuterlov et al.’s
interpretation was that, when facing a client’s initial report of no improvement, it might be more
useful for therapists not to persist in following a solution-focused approach, and to change to a
Herrero de Vega and Beyebach’s (2004) replication of the Reuterlov et al. (2000) study had
overall similar but somewhat different findings. In a sample of 96 SFBT sessions, they found
that 37.5 % of the clients who had initially reported no improvement scored higher on the
progress scale by the end of the interview. Beyond possible explanations for the difference
between the two studies (e.g., samples, treatments implementation, cultural contexts, therapists’
experience), their results create some uncertainty. If deconstruction were only successful on one
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 5
occasion out of every eight (Reuterlov et. al’s 13%), then using the session to deconstruct clients'
initial negative reports could be seen as a useless practice. However, when almost four out of
every 10 (37.5%) of these initially negative reports changed to progress by the end of the session,
“maybe, instead of dismissing it [deconstruction] as an useless procedure, it becomes worthwhile
to study more carefully under what circumstances it does work, and under what circumstances it
does not” (Herrero de Vega & Beyebach, p. 23). In this sense, prior to face the dilemma to
deconstruct or not to deconstruct? it might be useful to have a closer look at what actually
happens in those sessions where deconstruction is performed, trying to shed light on the
conditions that make it helpful.
A Qualitative, Discovery-oriented Design
In line with the increasing demand for studies that come closer to real clinical practice and
that narrow the research-practice gap (Trepper, Dolan, McCollum, & Nelson, 2006), this study
analyzed therapeutic conversations that took place after clients had reported that their situation
had not improved. The method followed a qualitative, discovery-oriented approach, which in
recent years has provided interesting insights into different aspects of SFBT practice (e.g.,
Franklin, 1996; Gale & Newfield, 1992; Lloyd & Dallos, 2006; Nau & Shilts, 2000). The main
method used was Greenberg´s Task Analysis method (1984).
The research question guiding this study was what happens in those SFBT sessions that
start with a report of no improvement and, after deconstruction, finish with the client
acknowledging improvements? Following task analysis’ terminology (Rhodes & Greenberg,
1994), the initial rational model, that is, the starting hypothesis about how the process of
deconstruction may unfold, was the general description that de Shazer (1988) proposed: a first
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 6
phase of deconstructing the global frame of no change, followed by a second phase of amplifying
and developing the improvements identified in the first phase. Hereafter, this paper presents a
detailed analysis of actual therapist-client dialogues from a microanalytic perspective (Bavelas,
McGee, Phillips, & Routledge, 2000), which focused on how solution-focused therapists try to
deconstruct, and on how deconstruction works. The ultimate aim of this research was to
generate a model of deconstruction that, as a clinical heuristic, might help therapists to
successfully shape their conversation with clients in a solution-focused way.
The search for a set of suitable data started by examining videotapes of second or later
individual sessions conducted between 1998 and 2006 at the Family Therapy Service of the
Pontifical University of Salamanca (part of a postgraduate program in SFBT). According to the
criteria required by Greenberg’s (1984) method of task analysis, the sessions had to contain an
excerpt with three components (definitions and examples are available from the first author):
1. an initial marker: the client’s initial report that there had been no improvement
2. a task environment: a period in which the therapist performed deconstruction
3. a final marker: the client’s final answer to the progress-scaling question
From an initial pool of 42 possible sessions, eight sessions were selected that had these
three components as well as the following additional criterion: There had to be a client’s
numerical answer to the progress-scaling question in the previous session. Compared to their
respective previous sessions, four of these excerpts had an increase on the scaling question at the
end of the session (i.e., the final marker was improvement). The other four had no increase on
the scaling question (i.e., the final marker was no improvement). In half of the excerpts (two that
improved and two that did not), the therapist was a trainer, and in the other half, the therapist was
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 7
a trainee. The trainers were PhD psychologists, with an average of 12 years clinical experience.
The trainees had participated for at least one year in an intensive SFBT training program. The
sessions covered a variety of individual presenting problems: anxiety, depression, obsessions,
grieving, and problems with children. The mean duration of the eight selected excerpts was 33
min, 46 s, with no statistically significant difference between improved and not-improved.
These eight excerpts were transcribed from the initial report of no improvement to the final
scaling question. Two independent judges and the first author analyzed the transcripts to confirm
that the eight excerpts met all of the criteria. The initial report of no improvement and the final
scaling question were confirmed with high inter-rater reliability (kappa = 1.0 for both markers).
Inter-rater kappa’s for identifying deconstruction interventions were .64 to .79. In spite of these
low kappa’s, both judges confirmed that in all eight excerpts deconstruction was performed.
Analysis Procedure. Stage I: Developing Relevant Categories
In the first step of analysis, the goal was to create a common vocabulary with which all
deconstruction processes could be described. The first author followed an iterative procedure to
find a vocabulary that would describe the eight excerpts: He watched the first excerpt and
described its key features, then he moved on to the second excerpt to confirm or modify this
vocabulary, then he moved on to the third excerpt, and so on. The Atlas-ti 5.0 software (Mühr,
1996) was used to facilitate this analytic process.
This procedure yielded three useful ways to capture the noteworthy differences in how
therapists and clients talked during the deconstruction process:
Topic. There were four ways to describe the therapist’s or client’s topics: (a) a positive
topic was any description of improvements or accomplishment of therapeutic goals (e.g., “I am
more relaxed these days”); (b) a negative topic was any description of no improvement or even
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 8
deterioration (e.g., “I have been more obsessed these days, I just cannot get rid of those
thoughts”); (c) an alternative topic was related to the therapeutic process, such as future goals,
potential resources, or hypothetical scenarios but was not about improvement (e.g., “I wish I
could simply turn the thoughts off, to have a way to simply make it stop”); (d) a tangential topic
was unrelated to the therapy (e.g., “It’s pretty cold today”).
Generality/specificity. There were three levels of generality/specificity: (a) an evaluation
was a general appraisal of an experience (e.g., “Things have been better this week”); (b) an
indicator described somewhat more specific thoughts, emotions, or behaviors (e.g., “I have been
eating more this week”); (c) an episode described something that had happened at a specific time
and in a specific situation (e.g., “Yesterday I was able to have full breakfast, I ate with my
husband and I even had some bacon and eggs”).
Action. There was a different vocabulary for the actions of therapists and clients. The
therapist could (a) introduce a proposition, which was new information, not previously
mentioned by the client, or (b) ask a question, which instead requested new information from the
client. A client, on the other hand, could (a) make a contribution that offered new information,
either spontaneously or on request, (b) indicate acceptance, explicitly agreeing with something
proposed by the therapist, (c) indicate rejection, explicitly disagreeing with something proposed
by the therapist, or instead (d) express doubt about what the therapist had proposed.
Each speech turn was described in terms of topic, generality/specificity, and action,
and a single speech turn could have two different descriptions. For example,
Therapist: "Tell me, on what occasions have you been more relaxed this last week?"
Some “occasions” being “more relaxed” is a positive topic.
Being “more relaxed last week” is an indicator (middle level of specificity).
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 9
The therapist’s action is a question.
Client: “Never. This has been a terrible week.”
“Never” and “terrible week” are negative topics.
“Never” refers to a requested indicator. “Terrible” is a general evaluation.
Both “never” and “terrible week” are the client’s new contributions.
Therapist: “How come?”
The therapist’s topic is about the client’s negative report.
The request is very general, asking for an evaluation.
The action is again a question.
Client: “Anxiety, this week it has skyrocketed. Yesterday I had a fight with my sister and
told her terrible things.”
The topic continues being negative.
“Anxiety, this week” is an indicator. “A fight” refers to a specific episode.
The client’s action is again a new contribution.
Analysis Procedure. Stage II: Finding a Model of Deconstruction Processes
Having developed a common vocabulary for describing all of the excerpts, in the second
step of analysis the authors set out to discover the possible differences between those that ended
with improvement and those that did not, in order to develop a model of deconstruction that
works. Through an intensive qualitative analysis of each excerpt, the emerging hypotheses about
what makes the difference in successfully deconstruction processes were constantly compared
against the data and refined, along an iterative process that resulted in a progressive adjustment
and development of the initial rational model (Greenberg, 1984; Rhodes & Greenberg, 1994).
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 10
Five Phases of Deconstruction
The analyses led to a tentative model of deconstruction that includes up to five
conversational phases: connection, preparation, deconstruction, elaboration, and consolidation.
As shown in Figure 1, these phases could occur in either an indirect or a direct approach.
The indirect approach (on the right side of Figure 1) began with a connection phase, during
which the conversation explored the negative topics that clients had contributed. These topics
were often discussed specifically, as episodes (i.e., specific incidents at a specific time;
sequence). There was also some discussion at the general level of evaluation (e.g., in relation to
possible explanations and causes; attribution).
The next phase of the indirect approach was the preparation phase, in which the therapist
shifted to alternative therapy topics, not directly related to either positive or negative changes.
These alternative topics could be specific future goals, potential resources, and hypothetical
scenarios (i.e., episodes and indicators; opening). They could also be general evaluations (e.g.,
that things could be better; information), which created a framework within which differences
and improvements in the present might later be noticed. Only then did the conversation move
into a deconstruction phase.
The direct approach (on the left side of Figure 1) began with deconstruction immediately,
without the two earlier phases. The deconstruction phase consisted of conversations that focused
on present or past experiences that could be construed as positive. The therapist tried to generate
positive indicators and episodes (e.g., specific incidents or feelings related to the therapeutic
goals; persistence). The therapist might also focus on alternative indicators and episodes that,
without being clearly positive, were at least different from the initial negative report (probe).
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 11
The elaboration phase built on the deconstruction phase. The therapist elaborated on the
positive and alternative topics by asking for details (i.e., indicators and episodes; specification)
and also elaborated by discussing the personal and interpersonal effects of these specifics (e.g.,
“So, in those occasions when you were able to keep anxiety down at home, what was different
between you and your sister?”; expansion). The therapist also sought explicit confirmation of
these improvements from the client and tried to highlight the client’s active role in bringing
about the changes (anchoring).
Finally, in the consolidation phase the therapist worked more at the evaluation level,
relating the now-elaborated positive topics to the client’s initial complaint and goals. In other
words, the conversations at this stage focused on the therapeutic relevance of the positive topics
discussed so far.
Therapists´ Perspective: Clinical Decision-making
Although we have described two patterns in the data, it is important to emphasize that the
therapists did not rigidly follow a fixed sequence. From what we observed, the therapists were
tracking their client’s responses closely. For example, if the client followed the therapist’s
positive topic in the deconstruction phase (i.e., if the client accepted it or even contributed his or
her own positive topic), then the therapist proceeded in the direction of the continuous arrow in
Figure 1 into the elaboration phase. If the client rejected or questioned the proposed topic, then
the therapist went back one step (e.g., into the preparation phase) before getting on track again.
The step-by-step sequence could also be shortened during the deconstruction phase. If the
client accepted or contributed positive topics without hesitation, then therapists would continue
elaborating them or even going directly into the consolidation phase. If at any point in the
consolidation phase the client showed doubts or rejected the relevance of those topics, the
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 12
therapist would stop to elaborate them more thoroughly or go back to deconstruction phase in
order to get a broader pool of positive (or alternative) topics on which the global improvement
could be built in the elaboration and consolidation phases. Even topics tangential to therapy
could sometimes be useful for balancing the conversation (Beyebach & Carranza, 1997).
In other words, the interviewing patterns were recursive–the clients responded to their
therapists, who then selected aspects of the interview to focus on. The therapists stayed flexible
and open to clients´ feedback. In this line, the broken arrows in Figure 1 represent possible
sidesteps in this process: During elaboration, a therapist sometimes went back to acknowledging
problems (i.e., to the connection phase) or to discussing hypothetical alternatives (the preparation
phase). Other times, the conversation in the connection or preparation phase generated
improvements that could be directly elaborated upon, without any need for deconstruction.
Clients´ Perspective: Hypothetical Change Mechanisms
The model of the deconstruction process developed here led to speculation about possible
change mechanisms on the clients’ part. Over the course of the project, it was observed that the
main aspect that differentiated the excerpts that ended with improvement from those that did not
was that the client acknowledged the relevance of the positive topics discussed so far. Therefore,
it became clear that, from the client’s perspective, it was important that the topics remained
directly relevant to his or her own goals within the global therapeutic process. And topic’s
relevance might be also constructed along the therapeutic conversation.
In this sense, movement along the generality/specificity levels of the same topic might play
a role in the construction of perceived relevance. Using Pearce and Cronen´s (1980)
Coordinated Management of Meaning framework, a hierarchical ordering of the three levels of
generality/specificity can be proposed, from the evaluation level (highest) to the indicator level
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 13
(intermediate) and the episode level (lowest). It is then possible to speculate that the evaluation
level exerts what Pearce and Cronen called a downward contextual force, that is, a general
evaluation creates a context for providing more specific indicators and episodes that illustrate the
topic of that evaluation. For example, one client started by giving a negative initial report (“not
better”; i.e., a general evaluation), illustrated by feelings of loneliness and helplessness
(indicators), which in turn led her further downward to talk about specific examples (at the
episode level) of being tense and feeling isolated with her group of friends, of not knowing what
to do with her kids, and of being staying home, crying, “without doing anything good for me.”
The other direction is also possible: Specific episodes can exert what Pearce and Cronen
describe as an upward implicative force on indicators. A single specific episode implies that
there may be more on the same topic, which implies something more general may be going on
(an indicator), which could lead to a more general evaluation. The above client discovered one
specific, small but positive episode in the deconstruction phase: one day she had “forced” herself
to cook for a visitor and had had a good time. She then also recalled that she had been buying
Christmas presents for her kids and preparing Christmas decorations with them and she had
bought some nice clothing for herself. The elaboration of these exceptions provided an upward
implicative force that supported a more general indicator of her “overcoming difficulties,” which
eventually transformed into the overall evaluation that she was valuing small things that she had
given up in the past and that now (in the consolidation phase) represented a “difference that
makes a difference.” Figure 2 illustrates this example, in which the therapist initially intervened
at the middle level (indicators) and from there sought specific episodes, which, along the session,
become relevant signs of “things getting better” (evaluation).
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 14
A therapist could also intervene initially at the most specific level (episodes). For
example, a client said things were “not better” (an evaluation) and illustrated this with negative
indicators (e.g., obsessions and “crazes” that overwhelm her) and specific episodes (“becoming
annoying” and not letting her Mum sleep; “almost cutting herself” for having lost a job). In the
deconstruction phase, the therapist began at the level of these episodes, working with her to
construct several “small differences” and resources that might have been useful (e.g., being “less
annoying” with her Mum on certain occasions; things that helped her to calm down when she
lost a job and prevented her from cutting herself). The collection and elaboration of a series of
episodes that were exceptions—and their contrast with similar situations in the past—seemed to
exert an upward implicative force on the upper levels. The “overwhelming obsessions” indicator
shifted to a “feeling of certain self-control,” which in the consolidation phase resulted in a
positive evaluation of the present situation as “better.”
Still another therapist intervened first at the most general level (evaluation). The client
initially reported that things were “not better” and moved downward from this evaluation to the
indicator that she was feeling insecure about her ex-husband, and episodes such as a tense
argument with her ex-husband. This therapist began deconstruction at the most general level by
challenging her initial report, until the client started questioning this evaluation herself, denying
the idea that “everything is awful”. This seemed to have exerted a downward contextual force
on the indicators, which in the elaboration phase changed gradually from “feeling insecure” to
“being clearer” about her desires. Thereafter, the episodes acquired a different meaning for her,
for example, from “a tense argument” to “an opportunity to realize” that she did not love him.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 15
In short, a hypothetical change mechanism for deconstructing initially negative reports
could be the progressive construction of topics relevant to the client, through the upward and
downward forces that episodes, indicators and evaluations exert.
Implications for practice
At the beginning of this project, starting from a list of deconstruction questions and a
general definition of deconstruction (de Shazer, 1988), it seemed that it might be possible to
uncover a relatively simple pattern of therapeutic interviewing, which would be useful for
achieving the deconstruction of initially negative reports. However, there was no single way that
the therapists in our sample accomplished (or did not accomplish) deconstruction. Specific
conclusions about possible differences between excerpts that ended with improvement and those
that did not are limited. There were several different ways to proceed. For example, therapists
intervened in two different ways (the direct and the indirect approach), but both approaches
could be found for either outcome. Instead of a fixed formula, several discoveries emerged:
First, this study suggests that deconstruction of clients’ initial no-improvement reports is
best seen as a complex process that involves more than simply using deconstruction questions
and then moving on to elaborate on exceptions. In fact, deconstruction per se seems to be only
one phase in a more extended process that can include up to five different phases. These phases
do not necessarily follow a fixed order, as therapists and clients may move back and forth from
one phase to another, depending on each other’s responses. Seen from this perspective,
deconstruction of no-improvement is related to the broader theme of the construction of change,
a process that might extend over a whole session.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 16
Second, although solution-focused therapists might be inclined to focus on solutions from
the beginning of every session, these data suggest that often it might be necessary to initially join
the client who is stating that things have not improved. Connection with client’s problem-
centered discourse may be an important part of the deconstruction process, both as the starting
point of the indirect approach and as a possible resource to come back to if difficulties appear in
the deconstruction phase or the elaboration phase.
A third discovery was the value of alternative topics. Our assumption before carrying out
this study was that asking clients about future steps and hypothetical solutions was the natural
option only after actual improvements had been described and expanded. In contrast, the
findings suggest that another good option may be to discuss alternative topics that might generate
talk about improvement. Questions such as “What will be the first small sign that you will see
happening once things get better?” (future improvement) or “Well, suppose that you had handled
his tantrum differently last week, how would he have noticed?” (hypothetical solution) may be
useful options in the preparation phase. For some clients, it might be easier to move from
negative to alternative topics and only then to positive ones, than to move directly from negative
to positive in the deconstruction phase. The focus on future scenarios or hypothetical situations
might also help clients to generate realistic expectations (Lloyd & Dallos, 2006).
Fourth, the data suggest that understanding therapist-client exchanges from the perspective
of ordered generality/specificity levels might be useful in SFBT. Coordinated Management of
Meaning (Pearce & Cronen, 1980) offers an interesting theoretical framework for the
understanding of in-session change. At the practice level, it is possible that moving from
episodes to indicators and evaluations and vice-versa is an invitation to include narrative
questions (White & Epston, 1989) in SFBT sessions, especially in the elaboration phase.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 17
Finally, relevance seems to be a central element of the overall deconstruction process.
Seen from this perspective, the job of the therapist is not only to locate specific improvements or
exceptions (i.e. the deconstruction phase) but to keep a global perspective on the therapeutic
process, constantly checking with their clients the implications and meaning that these
improvements have in relation to their broader therapeutic goals (i.e. the consolidation phase).
For example, if the therapist insists prematurely on evaluating those “small things done in spite
of everything” as an important advance, the client could reply something like “Can’t you see that
I’m not doing anything special for me, just trying to survive day to day? I’ve said that I feel
Limitations and future possibilities
There are necessarily limitations to any qualitative study (Elliott & Williams, 2001). In
spite of all precautions, there is always the risk of “confirmatory skew,” in which the researchers
tend to find what they expect (Glaser & Holton, 2004). The fact that there were many
unexpected insights suggests that this was not a major problem in this research project.
The small sample was carefully selected and matched, but it was limited to SFBT sessions
conducted at one clinical service. The decision to include sessions of both trainees and trainers
is both a weakness and a strength of the study. It could be argued that trainees are not as
competent as trainers in handling a complex process like deconstruction, but analyzing the
sessions of therapists with varying degrees of expertise brings this study closer to real-world
Future research to clarify further the conditions that make deconstruction helpful could
make use of triangulation (Mertens, 1998), by including both therapists’ and clients’ perspectives
about what makes the difference in those sessions that end with improvement. It would also be
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 18
useful to take into account the possible moderating effect of the therapeutic relationship: It might
be that a good therapeutic alliance allows therapists to use the direct approach in the
deconstruction process, whereas a less solid alliance might make an indirect approach more
advisable. Finally, it would be interesting to conduct quantitative sequential analyses (Bakeman
& Gottman, 1986) of how therapeutic interaction unfolds moment-by-moment, in order to test
specific hypotheses on the deconstruction process.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 19
Anderson, H., & Goolishian, H. A. (1988). Human systems as linguistic systems: Preliminary
and evolving ideas about the implications for clinical theory. Family Process, 27, 371-393.
Bakeman, R., & Gottman, J. M. (1986). Observing interaction: An introduction to sequential
analysis. New York: Cambridge University Press.
Bavelas, J. B., McGee, D., Phillips, B., & Routledge, R. (2000). Microanalysis of
communication in psychotherapy. Human Systems: The Journal of Systemic Consultation
& Management, 11, 47–66.
Berg, I. K., & Miller, S. D. (1992). Working with the problem drinker. A solution-focused
approach. New York: Norton.
Beyebach, M. (2006). 24 ideas para una psicoterapia breve [24 ideas for a brief psychotherapy].
Beyebach, M., & Carranza, V. (1997). Therapeutic interaction and dropout: Measuring relational
communication in solution-focused therapy. Journal of Family Therapy, 19, 173-212.
De Shazer, S. (1988). Clues: Investigating solutions in brief therapy. New York: Norton.
De Shazer, S. (1994). Words were originally magic. New York: Norton.
De Shazer, S., & Berg, I. K. (1992). Doing therapy: a post-structural re-vision. Journal of
Marital and Family Therapy, 18, 71-81.
Elliott, M. S., & Williams, D. I. (2001). Paradoxes of qualitative research. Counseling and
Psychotherapy Research, 1, 181-183.
Franklin, C. (1996). Solution-focused therapy: A marital case study using recursive dialectic
analysis. Journal of Family Psychotherapy, 7, 31-51.
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 20
Gale, J., & Newfield, N. (1992). A conversation analysis of a solution-focused marital therapy
session. Journal of Marital and Family Therapy, 18, 153–165.
Glaser, B. G., & Holton, J. (2004). Remodeling grounded theory. Forum Qualitative
Sozialforschung, 5(2), Paper 4. Retrieved on June, 12th, 2006 from http://www.qualitative-
Greenberg, L. S. (1984). Task Analysis: The general approach. In L. N. Rice & L. S. Greenberg
(Eds.), Patterns of Change (pp. 124-148). New York: Guilford.
Herrero de Vega, M., & Beyebach, M. (2004). Between-session change in solution-focused
therapy: A replication. Journal of Systemic Therapies, 23, 18-26
Lloyd, H., & Dallos, R. (2006). Solution-focused brief therapy with families who have a child
with intellectual disabilities: A description of the content of initial sessions and the
processes. Clinical Child Psychology and Psychiatry, 11, 367-386.
Mertens, D. M. (1998). Research methods in education and psychology. Thousand Oaks: Sage.
Mühr, T. (1996). Atlas.ti short user manual. London: Scolari.
Nau, D. S., & Shilts, L. (2000). When to use the miracle question: Clues from a qualitative study
of four SFBT practitioners. Journal of Systemic Therapies, 19, 129-135.
O’Hanlon, W. H., & Weiner-Davis, M. (1989). In search of solutions. New York: Norton.
Pearce, W. B., & Cronen, V. E. (1980). Communication, action and meaning: The creation of
social realities. New York: Praeger.
Reuterlov, H., Lofgren, T., Nordstrom, K., Ternstrom, A., & Miller, S. D. (2000). What is better?
A preliminary investigation of between-session change. Journal of Systemic Therapies, 19,
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 21
Rhodes, R., & Greenberg, L. S. (1994). Investigating the process of change: Clinical applications
of process research. In P. F. Talley, H. Strupp & S. Butler (Eds.), Psychotherapy Research
Practice: Bridging the gap (pp. 227-245). New York: Basic Books.
Sluzki, C. (1992). Transformations: A blueprint for narrative changes in therapy. Family
Process, 31, 217-230.
Tomm, K. (1987a). Interventive interviewing: Part I. Strategizing as a fourth guideline for the
therapist. Family Process, 26, 3-13.
Tomm, K. (1987b). Interventive interviewing: Part II. Reflexive questioning as a means to
enable self-healing. Family Process, 26, 167-183.
Trepper, T. S., Dolan, Y., McCollum, E. E., & Nelson, T. (2006). Steve de Shazer and the future
of solution-focused therapy. Journal of Marital and Family Therapy, 32, 133-139.
White, M., & Epston, D. (1989). Narrative means to therapeutic ends. Adelaide: Dulwich Centre
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 22
Figure 1. Basic model of in-session change. Deconstruction sequences developed in two
different ways, represented by continuous arrows. The direct approach started with
deconstruction and moved on to elaboration and consolidation. The indirect approach began by
adding two steps prior to the deconstruction phase: connection and preparation. Broken arrows
indicate possible sidesteps along the main sequences (described in the text).
“NOTHING IS BETTER”
“NOTHING IS BETTER”
Running head: SFBT RESPONSES TO “NO IMPROVEMENT” 23
Figure 2. How contextual and implicative forces could be hypothetical change mechanisms for
the client: As described in the text, the sequence began on the far left, which illustrates a downward
contextual force: The client’s initial evaluation (“not better”) included intermediate indicators
such as “loneliness” and specific episodes (e.g., staying at home crying). The numbers represent
the in-session change sequence. At 1, the therapist began by changing the focus of the conversation
at the indicators level, from negative aspects into client’s opportunities to see herself “overcoming
difficulties”. Once indicators shifted to more positive aspects, a new contextual force (at 2)
allowed the client to realize that some exceptional episodes had already happened (e.g., “forcing”
herself to cook for a visit). These examples of little things that she did in spite of all difficulties
exerted an upward implicative force (at 3), which eventually led to an overall positive evaluation
(valuing small things that she had given up in the past).
Staying at home crying,
“without doing anything for me”.
Buying some clothing for herself.
Not knowing what to do
with her kids (self-blaming).
Buying Christmas presents for her kids.
Preparing Christmas decoration together.
Tension and isolation
with a group of friends.
“Forcing” herself to cook for a visit.
Going out and having a good experience
Valuing small things that she had given
up in the past and now represent a
”difference that makes a difference”.
Doing something for herself
in spite of it.
Changes in client’s discourse over the course of the session