PROCESS-OUTCOME RESEARCH AND THE FAMILY-BASED MODEL: REFINING AND OPERATIONALIZING KEY THEORETICAL CONCEPTS The document below is a reproduction of Chapter 12 from the book Treating The Tough Adolescent: A Family-Based, Step-by-Step Guide, pages 259-309, by Dr. Scott P. Sells. Click Here For More Information      The focus of this chapter is to show how I used the process-outcome research method of "task analysis" (Rice & Greenberg, 1984) to create a 15-step family-based model for difficult adolescents. I will highlight how I used videotaped counseling sessions and focus group interviews with both counselors and clients to shape and refine the model itself. Research findings served as feedback to clarify and strengthen the theoretical concepts contained within the model. For example, results from focus group interviews revealed that parents reported that disrespect was an "ace" that neutralized their effectiveness. This information led to a change in the original model and the addition of disrespect as one of the "Five Aces." This chapter will also show how concepts drawn directly from this model can be operationalized and tested using outcome measures to determine their effectiveness. Future implications of this type of process-outcome research for family therapy and other mental health fields will be discussed and highlighted. Current Challenges and Controversies      Ready or not, our field is caught up in a health care revolution that demands accountability. Health care insuers require that our services demonstrate effectiveness with a particular problem and treatment population. Many counselors write books and articles that claim effectiveness but fail to demonstrate the process and empirical outcomes that back up such claims. This lack of accountability gives little comfort or credibility to third-party payers, legislators, students or fellow professionals. This lack of accountability can be traced to three main causes: (1) Treatment models with procedures that are abstract, generalized and difficult to implement; (2) Outcome studies that answer the question "Does it work?" before answering the question "How does it work?" and (3) A failure to combine both process and outcome research to create, refine or operationalize treatment models. 1I would like to give special recognition to Neil Schiff and Jay Haley for their review and help with this chapter. Without their help and guidance, this chapter would not have been possible. They are among the first to open up their videotape library for the scrutiny and analysis of an entire case study. Because of their foresight and vision, this research was possible. Treatment Models That Lack Specificity      Most treatment models either lack specificity or contain procedures that are abstract, generalized, and difficult to implement. A particular model is often employed because it is popular at the time, because it fits with the particular counselor's treatment philosophy or because it mirrors that of the school where the counselor received his training. In an article on the effectiveness of marital and family therapy, Pinsof and Wynne (1995) conclude that "in almost all of the MFT research, it is impossible to know what actually occurred in counseling" [p. 606]. This should certainly concern counselors as we enter the 21st century. Without the specification of key concepts, it is difficult if not impossible to assess what takes place in any particular counseling session so that accountability can take place. Counseling then becomes a mystical process behind closed doors rather than a systematic one that is well articulated. Under these conditions, it is not surprising that healthcare providers are leery about funding undefined and untested treatments. "Who Won" Studies      The majority of current research uses comparative-competitive or "who won" studies that pit one treatment approach against another (e.g. cognitive therapy outperformed structural family therapy with depressed women aged 19 to 35) but fail to specify what factors within the model were associated with improvement and deterioration. Without this information, the study may have little relevance for the individual counselor. The counselor reads that one treatment approach is better than another but has no idea what particular techniques might have caused this change. Outcome research without process research is therefore minimally informative. In a review summarizing trends in theory and research from 1980 to 1987, Bednar, Burlingame, and Masters (1988) state that 140 family counseling studies revealed a virtual absence of treatment variables drawn from systems theory literature. The reviewers concluded by saying that rigorous experimental outcome research was premature for a field that had yet to operationalize its essential theoretical concepts. Wynne (1988) reached a similar conclusion stating: The term 'research' is often understood by psychotherapists as referring to confirmatory studies, such as comparative studies of the outcome of the two methods of counseling. In sharp contrast to this usual view, at the present stage of development of the family therapy field, a strong emphasis should be given to exploratory, discovery-oriented and hypothesis-generating research, rather than primarily or exclusively to confirmatory research. [p. 251]      Before a particular counseling model can be applied, the concepts must be operationally defined. Outcome studies that answer the question "Does it work?" before answering the question "How does it work?" are suspect and premature. A Failure to Combine Process and Outcome Research      Once you have an idea of how a model performs through process research, you must conduct outcome studies to determine if the model does work. Often, one step is conducted without the other. If the outcome studies fail to show effectiveness, this is invaluable feedback. It informs the researcher that parts of the model are not working or that this particular model does not work with a particular problem (i.e. alcohol or drug use, depression, psychosis) or treatment population (i.e. adult, individual, child). The researcher is then forced to reevaluate the model to strengthen or revise it in specific ways. Goals and Objectives of the Project      To address these current challenges, I focused on the following two research objectives: (1) Development of a Treatment Manual (2) Combining Process and Outcome Research. How these areas addressed each research gap will be highlighted. Development of a Treatment Manual      Before embarking on this project four years ago, I realized that I had to address each of the above research gaps. First, I had to provide the counselor with a road map of step-by-step procedures, techniques, themes, and therapeutic maneuvers. This was needed because current books and articles on treatment with difficult adolescents often lacked specificity. The current treatment models (i.e. multitarget ecological treatment (MET), functional family counseling, social learning counseling, strategic counseling and structural counseling) articulated key theoretical concepts (e.g. hierarchy, boundaries, power, ecosystems, coercive interaction patterns) but failed to provide the reader with a step-by-step process of how and when these concepts should be implemented. In addition, there was an abundance of "who won" studies with difficult adolescents (cf., Chamberlain & Rosicky, 1995) but these studies lacked relevance for the individual counselor because they failed to provide information on specific concepts that effected change.      With such a challenging population, I wanted to know what to do and when to do it if a, b, or c should occur. For example, what should I do the next time the parents refused to take charge? Could I choose from a menu of creative and innovative techniques? While I realized there is no magic formula, I needed a more explicit road map to increase my chances of success and keep one step ahead of the cunning adolescent and resistant parent.      Second, the treatment manual also had to be flexible enough to be customized to meet the needs of individual clients without stifling the flexibility, innovation and creativity of the counselor. My goal was not to produce a rigid application of treatment but give guidelines that were systemic yet adaptable enough to encompass novel situations and circumstances. For example, what does one do with a single parent who cannot take charge because she has no support systems? Or with an adolescent who is protected by a highly dysfunctional set of peers? Many manuals feature a simplistic one-size-fits-all approach that is simply unrealistic with difficult adolescents. One only has to be around them for a short time to realize that you must be very creative and think on your feet quickly.      Finally, I felt that the manual itself had to emerge directly from an intensive case by case study of counseling sessions and focus group interviews with both clients and counselors rather than strictly from a literature review in the library. As stated earlier, this was important because many family counseling models have yet to operationalize essential theoretical concepts or map out these concepts in clinical practice rather than in a laboratory setting. As a result, theory had to be linked directly with clinical practice. Combining Process and Outcome Research      To accomplish this goal, I used a "task analysis" method (cf., Rice & Greenberg, 1984) to conduct the process research portion of the study and then used outcome measures to test key theoretical concepts that emerged. A task analysis methodology would help me discover key moments of change within counseling sessions from an intensive analysis of videotaped interviews and self-reports from clients and counselors. These moments of change could then be written as hypotheses and tested through outcome measures. Results from these outcome measures would then be used to clarify or strengthen these moments of change. For example, it was found that pretest measures supported the hypothesis that difficult adolescents and their parents enter treatment with severe conflict and a lack of nurturance. This outcome data strengthened the need for the procedural step of "restoring nurturance" within the 15-step model. I wrote about the benefits and procedures of blending qualitative process research and quantitative outcome research within the same study and how these two methods can reciprocally help clarify, strengthen or refine key theoretical concepts (see Sells, Smith,& Sprenkle, 1995 for a detailed discussion on this topic). A Task Analysis Approach      In this section, I will illustrate through sample flow charts and coding manuals how the 15-Step Family-Based Model was created. The entire process can be referred to as discovery-oriented because key concepts are generated not from a review of literature but directly from an intensive study of clinical practice cases. The following five phases were employed within the task analysis process research study: (1) Creating Idealized Performance Models; (2) Creating Revised Performance Models; (3) Broadening the Range of Application; (4) Theoretical Yield and (5) Combining Process and Outcome Research. Phase I: Creating Idealized Performance Models      During the first stage of the project, I went to the library and located books and articles that outlined theoretical concepts and treatment procedures for difficult adolescents (i.e. teenagers between the ages of 12 and 18 who meet the DSM IV diagnostic criteria of either oppositional defiant or conduct disorder). I then extracted the major concepts and placed them into a spreadsheet format. I combined these spreadsheets into three idealized performance models of what should theoretically happen throughout the treatment process. These concepts were then placed in a step-by-step treatment manual format. It is important to note that the procedural steps were not clearly specified or laid out within the literature. I consolidated the many different concepts on a spreadsheet and placed them into a hypothesized series of optimal procedural steps. The process was like taking hundreds of tiny puzzle pieces and trying to place them together in the proper order with little to go on except similar shapes and colors. The three idealized theoretical models were differentiated into three separate stages based on "markers," clinically significant events that appeared to change the course and direction of the treatment process. Each of three idealized models is reproduced below to illustrate the step-by-step process of how each model was created. 2The books and articles under study came from structural, strategic, solution-focused, and multidimensional treatment models (cf., Fishman, 1988; Haley, 1976, 1980; Keim, 1996; Liddle, 1995; Madanes, 1991; Minuchin, 1974; Minuchin, Montalvo, Guerney, & Rosman, 1967; Price, 1997; Selekman, 1993). Other treatment models were not selected because they were not theoretically congruent with the 15 Step Family Based Model or a systems theory framework. All of the models selected for review were theoretically congruent with a family systems perspective. Model Two/Stage Two: Troubleshooting Crises and Relapses      Within the literature, treatment seemed to begin with the counselor calling the parents to set up the first appointment and end with the adolescent functioning without behavior problems. There also appeared to be a proverbial "fork in the road" at which point the parents would choose to accept or not accept a position of changing their teenager's problem behavior. Which road the parents chose often seemed to depend on the quality of the parents' rapport with the counselor and whether or not the counselor was seen as a credible expert. Once the decision was made, a series of steps would follow. If the parents refused to take charge, the teenager or other outsiders would take charge and treatment would end unsuccessfully. If the parent did take charge, the counselor would navigate the parents through a series of steps to keep the parents in this position of authority, troubleshoot potential problems and stop the extreme behavior problems.      Each step within the three idealized models was operationalized into observable behaviors. Behaviors were then defined within a coding manual. For example, the concept of "Pre-Session Preparation" in Step 1 of the first model was defined in the following manner: Step 1: Pre-Session Preparation: Before the first session, the counselor personally contacts the parents and explicitly asks them to come in with their son or daughter to help him or her with the identified problem by providing information and guidance that only they can provide. They should not be asked to come in to have "therapy" because few people want "therapy." Following this same rationale, members of the extended family including other siblings, grandparents, etc. are also asked to attend. This observational code is defined with statements made by the counselor to the parents asking them to come in with their teenager "to help the teenager with his problems by providing valuable information to the counselor." Everyone in the family including other siblings are asked to attend. Model One/Stage One is illustrated below with the procedural steps hypothesized to be optimal for change with difficult adolescents and their families. Each of these concepts emerged directly from the literature and was pieced together in a treatment model format. Model Two/Stage Two: Troubleshooting Crises and Relapses      From the literature, Stage Two begins with a relapse of the difficult teenager and ends with the parent weathering the storm by devising a plan to prevent further relapses. There are basically two reasons for this relapse.      Once the teenager functions without problems, the parents are then lulled into a false sense of security. Parents think that these changes are permanent. However, the teenager is not likely to hand over their power and authority without a fight and at least one major relapse to test the waters. The teenager wants things to return to status quo. A teenager functioning without behavior problems has not yet had enough time to realize that they can get most of their needs met by being good. For many teenagers being "good" is a change in identity and feels awkward and different. As a result, the risk and temptation is high for at least on major relapse.      In addition, the teenager' problem may be a conscious or unconscious attempt to shift the focus off more threatening issues in the family such as marital conflict, depression, or alcohol and drug abuse. If the parent or other family members remain focused on the teenager's problem, other issues are not addressed. Consequently, every time the teenager begins to function normally and without problems, the family becomes unstable and other problems surface. The adolescent must again function incompetently and relapse to shift the focus off these other problems so the family can restablize. This cycle will repeat itself again and again until the underlying family issues are solved or resolved.      Both of these reasons may be occuring separately or together. The teenager will most likely always want to test the waters regardless of whether or not there is an interconnection with other family issues. The counselor can spot if the interconnection piece is also an issue if other family problems surface immediately or soon after the behavior problems are solved.      In either case, the parents reaction to the relapse is usually negative. They feel personally betrayed and possess a "here we go again" attitude. When this happens, the parent's inclination is either to remove the teenager from the family and place them into an institution or to feel apathetic and give up. At this point, the counselor must take charge and somehow convince the parent to stand firm and not give up. Instead, they must devise a plan of action to address the present relapse and prevent further relapses from occurring in the future. The counselor must try to prevent institutionalization or risk starting from scratch when the teenager finally returns home. Model Two/Stage Two is illustrated below with the procedural steps hypothesized as to be optimal for change with difficult adolescents and their families. These concepts also emerged directly from the literature. Model Three/Stage Three: Peace Sets In & Teenager Moves Into Adulthood      From the literature, Stage Three begins when the adolescent continues to the test limits but does so in a way that is not extreme (violence, running away, truancy, etc.). The stage ends when relapse ends on a permanent basis. The adolescent is then able to freely move through the development stage of individuation by leaving home and becoming an adult.      In essence, at the end of Stage One, the parent survives the initial onslaught of the hurricane and briefly experiences calmness in the eye of the storm but does not buckle or fold when the hurricane resumes its gale force winds in Stage Two. In Stage Three, the parent survives the hurricane and can finally enjoy the fruits of their labor as the hierarchy is permanently reversed and they maintain their authority even when it is tested. As a result, calmness and peace set in within the household on a consistent basis. The teenager is now free to move their time and energy from trying to maintain their power and authority to pursuing employment, dating, sports, college, and eventually leaving the home to become an adult. It is important to note that this disengagement from the family is also contingent upon the resolution of underlying family issues. Otherwise, the teenager will be unable to disengage from the family and display self-destructive behavior that debilitate or incapacitate the adolescent from leaving home to become self-supporting.      Model Three/Stage Three is illustrated below with the procedural steps hypothesized as to be optimal for change with difficult adolescents and their families. These concepts also emerged directly from the literature. Phase II: Creating Revised Performance Models      When Phase I was completed, Phase II began by following Rice and Greenberg's (1984) recommendation to acquire and analyze videotapes from "expert clinicians regarded by colleagues, trainees and clients as being instrumental in facilitating substantial amounts of positive client change" [p. 291]. I analyzed each session and made a coding manual and a performance model diagram to accompany each videotaped counseling session. Each procedural step was a "marker," or series of interventions hypothesized to be optimal for promoting change.      After each tape was completed, I compared the diagram and coding manual with the three idealized theoretical models to locate similarities and differences. If the marker in the actual counseling session was similar, the matching idealized concept was strengthened. If there were discrepancies or new discoveries, a revised performance model was produced. For example, after the first videotaped session was analyzed, one marker closely matched the idealized model concept of engagement. This idealized theoretical concept was therefore supported and strengthened. In contrast, a new concept also emerged from this session that involved the parents' redefining the son's problem behavior. This led to a revised performance model that contained the new procedural step of "defining and redefining the problem."      This process of shifting back and forth between analyzing actual videotapes and revising the idealized models continued after each new tape. A final revised performance model emerged at the conclusion of Phase II containing both original theoretical ideas as well as new and exciting discoveries from the videotapes of Jay Haley and Neil Schiff. Below is a brief description of the steps in Phase II with illustrations of several revised performance models to demonstrate the research process. The final revised performance model is also reproduced to highlight the developmental steps in creating the 15-Step Family-Based Treatment Model. Step 1: Using Expert Clinicians to Construct Revised Performance Models      As stated earlier, an intensive study of videotaped interviews from expert clinicians is an ideal first step. This is because most process researchers select the work of student counselors or counselors who are not regarded as experts in the model under investigation (Mahrer, 1988, Rice, 1984). A closely related issue is a lack of "treatment integrity" or the failure of counselors to adhere to the guidelines specified within the treatment model (Pinsof & Wynne, 1995). When this happens, it becomes increasingly difficult to locate key moments of change when one is uncertain if change is even occurring or if the treatment model guidelines are being followed.      To address these problems, I asked Jay Haley, the founder of strategic family therapy, if I could analyze videotaped counseling sessions that he felt were instrumental in facilitating significant amounts of positive change. After hearing about the project, he consented and suggested an intensive, beginning-to-end analysis of a 28-session case involving a 18-year-old male who exhibited extreme behavior problems (i.e. threats and acts of violence). Haley stated that this case contained all the essential theoretical concepts and procedures to promote change with difficult adolescents and was therefore representative of how to work with this population. In addition, Haley stated that the counseling was successful as indicated by an annual follow-up for ten years that showed no relapses or return to previous problem behaviors. The adolescent had since graduated from college and was successfully functioning as a high school art teacher.      Dr. Neil Schiff was the counselor while Jay Haley supervised each session from behind a one-way mirror. Both Neil Schiff and Jay Haley are considered by colleagues, trainees and clients to be the leading experts in treating difficult adolescents and their families. Step 2: Constructing the Performance Model      Briefly, the initial performance model was built and constructed from an intensive analysis of all 28 sessions. Each videotaped session was transcribed, the markers were operationalized within a coding manual and the procedural steps were illustrated on a schematic diagram. Each code was accompanied by actual transcript from the session to support the inclusion of that particular code. For example, the marker of "task check" emerged from the third session and was operationally defined in the following manner. Step 3: Task Check      The counselor asks the parents and the adolescent if the teenager and/or the parents completed the tasks assigned at the end of the last session. After giving a task, the counselor should always ask for a report at some time in the next interview. In this way, the teenager and parents are accountable for completing each task.      This observational code is defined as one or more statements by the counselor asking the parents and the teenager if the tasks assigned at the end of last session were completed. Counselor Comments      The counselor asks the parents and problem teenager if the tasks assigned at the end of the last session were completed by the teenager and/or parents. Sample Dialogue from Session Time: 3 min and 30 seconds into video session 67 Ther: Well I'm delighted that you had some more normal 68 moods, but I'm disturbed about you crashing, and I wish, I 69 hope that there's something we can set up that will alleviate 70 the pain associated with that. Anyway, let me go on a bit and 71 then come back to this. Did you register for a course?      Below is the schematic diagram that emerged from the third counseling session. Notice how the parents vacillate between taking charge and refusing to take charge in Steps 8, 21, and 22. In response, the counselor asks the parents in Step 8A to take charge and set tasks with specific timeframes for completion. Notice how the parents struggle with defining their son as frail, incapable and not responsible for his extreme behavior under Step 8. The counselor counters by redefining the adolescent as normal, capable and responsible for his actions in Steps 13, 17 and 20. The counselor also emphasizes the normality of the son in Steps 1 and 2 by convincing the parents to decrease the son's medication. As stated earlier, the use of medication can result in the adolescent being labeled as a chemically-imbalanced mental patient rather than a misbehaving teenager responsible for his behavior. This theme of emphasizing normality, capability and responsibility also emerges in later s essions and was influential in the construction of Parental Empowerment in Step 3 and Defining and Redefining the Problems in Step 2 of the Family-Based Treatment Model. Step 3: Informant Verification      After each session was analyzed, the coding manual and diagram were sent to both Neil Schiff and Jay Haley. I interviewed each by phone and asked if they agreed or disagreed with my conceptualization of each procedural step contained in the coding manual. If there were discrepancies, the concept was modified accordingly. For example, Schiff and Haley both stated that a third relapse was avoided due to Schiff's use of troubleshooting while I described the same event as problem solving. The concept was then modified to fit the definition of troubleshooting and operationalized based on Schiff and Haley's description. This process of "informant verification" refers to the extent to which a set of meanings held by multiple observers are sufficiently congruent so that they describe the phenomenon in the same way and arrive at the same conclusions (LeCompte & Goetz, 1982). Informant verification was used to assess the reliability of the codes by checking to see if Schiff, Haley and I independently described the codes in the same way and arrived at the same conclusions.      In addition to checking reliability, these interviews helped shape the performance model by expanding my conceptual definitions and highlighting key interventions that promoted optimal change. These brainstorming sessions by telephone generated rich clinical data. For example, we discovered that the use of "dry runs" was essential to prepare the parents for future confrontations with their teenager. A turning point in one case came when the father asked his son to return the house keys and move out because of his extreme acts of violence. Schiff prepared the father for this critical confrontation by playing the part of the son while the father practiced his delivery of what he would say. We all felt that this preparation was key to the father's ability to take charge. These valuable discussions led to the creation of "dry runs" as a mini-step of the larger step of troubleshooting. Step 4: Comparing Performance with Idealized Models      After the performance model was further refined based on the Haley and Schiff interviews, I compared it with the three idealized models. If there were similarities, then the concepts within the idealized model were validated. On the other hand, if there were discrepancies, a new revised performance model would emerge. On the other hand, if there were discrepancies, I made revisions accordingly. I demonstrate this process by providing and discussing three diagrams: one for the idealized stage model, or Model 1 (Figure 5A); one for the videotape-based performance model for Session 1 of counseling (Figure 5B); and one for the revised performance model for Session 1 that resulted from my integration of these two models (Figure 5C). Comparison of Session One and the Idealized Model A schematic diagram from session one is reproduced and then compared with the Model One/Stage One Idealized Model. After the comparison is completed, a revised performance model emerges that integrates the theoretical concepts from both the first idealized model and the first session performance model. Notice how the first three steps within the session one performance model are identical to those of the idealized model. From this point on, however, there are discrepancies and similarities. Step 4 [family's explanation of the problem] and Step five [calling on family member's expertise] of the performance model did not occur anywhere within the idealized model. Thus, these concepts were organized under Step 4 of the revised performance model under the broader category of "Parental Empowerment." This was done because calling on the family's expertise and asking them to define the problem are both strategies designed to empower the family to solve their own problems. The two models are similar with regards to the parents' decision to take charge or not take charge. As a result, these concepts were substantiated in the idealized model. An exciting breakthrough came with the discovery that two key procedural steps reoccurred throughout the counseling session. The performance model from session one indicated that parents vacillated between assuming and refusing authority throughout the counseling session. In addition, engagement was not a one-time step and continued to surface throughout the first counseling session. So within the idealized model, it appeared that certain steps occurred only once throughout the counseling process. However, the performance models indicated the opposite. Counseling with difficult adolescents was discovered to be a very fluid and circular process as key steps continued to resurface again and again. These patterns made it easier to identify steps that were optimal for change. Other discrepancies between the models centered around the use of solution-focused hypothetical and exception questions as well as the concept of revisiting central themes. As more performance models emerged, these steps were also rearranged under broader categories. For example, the use of solution-focused questions was later classified under "Clear Rules and Consequences Outlined" when analysis showed that the counselors use these types of questions as way to clearly define rules and consequences. The Session One Revised Performance Model retained the idealized concepts from Model One but added the following new components: (1) Empowerment; (2) Recurrence of Taking Change; (3) Recurrence of Engagement; (4) Solution-Focused Hypothetical and Exception Questions; and (5) Revisiting Central Themes. Steps within the idealized model that were not observed or that were still unaccounted for after analyzing session one were listed next to the revised performance model in the event that they arose in future videotaped sessions. In sum, each cycle of looping back and forth between the idealized model and a performance model both clarified and further operationalized existing theoretical concepts. The process also uncovered new and exciting discoveries. The process could be comparable to an anthropologist who has books and ancient descriptions of a particular population but goes to the ruins of this civilization to discover new artifacts that both confirm and deny these original writings. These artifacts not only lead to an expansion and clearer definition of these original writing but also generate new and exciting discoveries. Step 5: Revised Performance Model      After analyzing all 28 taped sessions and comparing the results with the three idealized models, one final revised model was constructed by integrating each of the revised performance models and noting similarities and discrepancies. Steps that were similar in concept were placed under one main category. For example, several revised performance diagrams contained steps that included seeking concreteness, outlining specific tasks, and setting specific dates. Each of these described the same basic process of how to outline clear rules, consequences and task procedures. Consequently, these steps were integrated under one main category labeled "Step 5: Clear Rules, Tasks and Consequences Outlined."      Another interpretation of this process is that "mini-steps" united into one main step. For example, executing "dry runs" and "what if scenarios" were the mini-steps the counselor took in conducting the larger step of "Troubleshooting." These mini-steps or "small-o" outcomes were particularly helpful because they operationalized the step-by-step procedures used to achieve the end result. This process of integrating smaller, related steps under one main category continued until one final revised performance model remained. Below is a reproduction of this completed revised performance model.      The question mark next to Step 8 indicates that although the concept of soothing sequences was contained within the idealized models, it was not identified in any of the 28 sessions. This step was implied during several of Schiff's interventions but without the clarity needed to make these observable actions into a distinct step. As a result, further investigation was needed. Phase III: Model Building: Broadening the Range of Application      After Phase II was completed, the next step was to take the final revised performance model and test its procedural steps in the field with a variety of different counselors, clients and conduct and oppositional behavior problems. The goal was to fine tune the model by pursuing any anomalies or new ideas resulting from a broader range of its application. If anomalies were discovered, the revised model was modified accordingly. I continued this process until analyses of the focus group interviews and the videotaped sessions failed to provide any new information.      It is important to note that this does not mean additional discoveries cannot still be found. If a variety of counselors in different parts of the country use this model with a larger sampling of difficult adolescents, additional concepts could potentially emerge. The idea that any treatment model can be theoretically saturated and produce no new concepts is naive and misleading to the reader. One of the goals of this chapter is that other counselors will utilize this model and provide feedback on whether or not new discoveries were made requiring further refinement. Thus, this model should be seen as a "work in progress" rather than "the decisive treatment model for difficult adolescents."      Model-building really begins during Phase III as the process of enrichment and elaboration of the treatment steps through a broader application unfolds. In developing the 15-Step Family-Based Model, I started with three idealized models and moved in a progressive fashion toward more intricate models that more closely reflected the complexity of working with difficult adolescents. The basic steps in creating the performance models of Phase III are identical to those outlined for Phase II. The differences between this phase and Phase II included (1) the analysis of sessions by four counselors involving a variety of cases rather than the analysis of one case by two expert clinicians and (2) qualitative focus group interviews with both clients and counselors to gain access into their thoughts about the use of this treatment model. Description of Counselors, Clients and Target Problems      A team of four counselors was selected to implement field-testing of the final revised performance model from Phase II. The counselors included three recent graduates of a Master's of Social Work program with three to seven years of experience and a supervisor with a Ph.D. and more than ten years of experience. Over a two-year period, 83 difficult adolescents and their families were seen for a minimum of five sessions. The average length of counseling was ten sessions.      Most of the adolescents treated were males (78.9%) with an average age of 15.3 years (sd=1.5). Over half of the adolescents (52.2%) had a history of fighting or assault; 56% had a history of stealing or shoplifting; 43.5% were truant from school; 39.9% had drug or alcohol problems; 39.1% had problems with running away; and 8.7% had charges of vandalism, weapons or sexual abuse. These percentages reflect that the majority of the adolescents had multiple problems. The demographic data showed that 54% of all families served had an income of $10,000 to $34,000 dollars. A majority of the adolescents had numerous stays in detention, prison or residential treatment (m=2.1). These adolescents were also multiple offenders, having an average of 3.3 arrests each.      In sum, the three counselors selected were recent graduates with only limited experience in treating difficult adolescents. The adolescent population was primarily characterized by severe conduct problems and low-income households. This use of inexperienced counselors and a population of difficult adolescents was intentionally selected for two reasons. First, the applicability of the model and the ease of its implementation might be better understood. If inexperienced counselors demonstrated success through pre-post outcome measures, a case could be made that the treatment manual was prescriptive and highly applicable. In addition, if the model was effective in producing change with extremely difficult problems, the model may generate even greater success with less severe cases and families with greater economic and social resources.      Secondly, anomalies or unsuccessful change episodes would be more revealing. Since the counselors were rather inexperienced, they might make mistakes on the most basic of steps. In turn, these mistakes would force me to make each treatment step clearer, more concrete and more user-friendly. In addition, unsuccessful change episodes would be much more common and challenging with extremely difficult adolescents from families with limited economic and social resources. This would force me to become more innovative with the treatment steps. An example is the development of "Neutralizing the Five Aces" in which a creative menu of strategies was outlined to help struggling counselors stop extreme behavior problems. Treatment Integrity      As stated earlier, a major problem in research is whether or not the counselor treating the case is actually following the steps of the treatment model (Wynne, 1988). To address this concern, two areas were addressed. First, each counselor was given the performance model to read and memorize like a play book before entering the first counseling session. Each was then asked to demonstrate each step through role-plays. If there were problems during implementation, the supervisor would stop the role-play and model the correct procedures. In addition, the supervisor observed each counselor through a one-way mirror during the first three sessions and once a week afterwards to ensure that the treatment steps were being followed fairly closely. Videotapes of the sessions were also analyzed.      It is important to note that the training and role-playing was not a rigid application of each treatment step but a demonstration of general guidelines. This forced the counselors to hone their skills of creativity and intuition. In addition, the model was still a work in progress and many of the steps had not yet been developed (i.e. restoring nurturance, neutralizing the "five aces") or were not yet concretely defined (i.e. troubleshooting, working with outsiders). These gaps revealed the weaknesses and the strengths of the model and where it needed to be revised or more clearly defined. Focus Group Interviews      As stated earlier, clients' thoughts and feelings towards counseling are as important as observable behaviors since a comprehensive process analysis requires both (Pinsof, 1988). Comparisons of how different family members view treatment provide valuable information about practice effectiveness as a prelude to clinical-trial outcome research (Gurman, Kniskern, & Pinsof, 1986).      To address this issue, the procedures from an earlier study by Sells, Smith, & Moon (1996) served as a template for conducing client interviews. In this study, the researchers used ethnographic interviews that immediately followed counseling sessions to elicit the clients' thoughts and feelings about their sessions. Since one goal of this project was to tap into these same areas, procedures from this earlier study were replicated during this phase of the project. The counselors asked each client a series of questions at the end of every third session. This timeframe would give interventions a chance to prove successful or unsuccessful. The answers revealed a wealth of information on what clients perceived as effective and ineffective interventions, important counselor qualities and recommendations for future counseling sessions. The following eight questions were asked: Could you tell me in detail all the things that have been most helpful so far? What are the most helpful things I have done or said as your counselor so far? Could you tell me in detail all the things that have been least helpful so far? What are the least helpful things that I have done or said as your counselor so far? What needs to be done in the future to make your sessions more useful or helpful? How would you describe to a friend what we do here or the approach that I am using? What are all the things you like about it? What are all the things you dislike about it?      Each interview was either audiotaped or videotaped and then transcribed. Major themes were uncovered from these interviews and coded in the same manner as the videotaped interviews. These codes were then compared with the idealized and performance model for discrepancies and similarities.      Several very interesting findings emerged that help refine and shape the 15-step Model. For example, a series of interviews revealed that 37 teenagers felt the opportunity to regain trust was one of the most important things needed to improve future counseling sessions. Teenagers reported that when they lost the opportunity to rebuild trust they lost hope and resentment set in.      Before these interviews, the area of trust was not looked at as an intervention. After the interviews, it was placed as one of the seven micro-strategies for restoring softness and nurturance between parent and teenager. I then field- tested the concept and closely analyzed the videotapes when this marker was being used by the counselor. We also conducted more focus interviews with both teenagers and parents to locate any further inconsistencies. Each new set of information led to additional refinements of this micro-step. For example, we found that parents must give trust in increments proportional to the level of supervision (mandatory, structured, or limited) at which the teenager is currently functioning. The level chosen would guide the parents on how much trust to give and how much to hold back. Explanation of Anomalies      The most important action of this step was the intense scrutiny of instances in which the model did not appear to work for the client. When this happened, I looked for potential explanations of the anomaly by asking the following four questions: Was the counselor marker or the concept poorly defined? What factors could account for the anomaly? Was it something specific the counselor did or said that accounted for the intervention not being effective? Were there particular characteristics of the client that seemed to make the counselor's intervention particularl difficult or impossible?      One of the anomalies that emerged from this analysis is presented below with a diagram illustrating the micro-steps uncovered. Discovering the Micro-Step of Outlining Rules and Consequences During the analyses of two different videotaped sessions, I found that the procedural step of setting rules and consequences did not appear to be working. This was indicated by the fact that the parents failed to follow through with rules and consequences that were outlined with the counselors the week prior. After considering the four questions above, I began to find reasons for this anomaly. First, both counselors outlined the rules but only in very vague terms. The rule of "showing respect," for example, was not operationally defined by listing concrete behaviors considered disrespectful such as swearing and refusing parental requests. In addition, consequences were not clearly defined. One consequence might be grounding but there was no discussion of when or how long the grounding would be or who would enforce and monitor it. Second, the counselor markers of "Clear Rules and Consequences Outlined" were poorly defined. The two counselors had to be shown the micro-steps involved in helping the parents operationalize specific rules and consequences. The following seven strategies were developed from observed mistakes made by the counselors and focus group interviews with clients and counselors. After the counselors were trained to successfully implement these seven strategies, analyses of future interviews with the same families revealed that the rules and consequences were enforced. In addition, parents reported that "a clear road map of rules and consequences" enabled them to be more effective as parents.      In sum, an intensive analysis of video performances and focus group interviews enabled me to fine tune and create the 15 Step Family Based Model. The idealized and the revised performance models revealed the macro steps but its broader application revealed the micro steps needed to achieve each major procedural step. At the end of this phase, the 15 Step Family-Based Model emerged. It is illustrated here with a summary of both the 15 procedural steps and the mini-steps or specific micro-theories that can be used to generate change within each macro step. Phase IV: Theoretical Yield      In this section, I will summarize some of the major ideas that emerged to expand my thinking about treatment with difficult adolescents. I will highlight key moments of discovery throughout the three previous phases that shaped the 15-Step Family-Based Treatment Model. Of particular interest is the discovery of what I call "micro-steps," the strategies a counselor must engage in to accomplish the larger goal. For example, to restore nurturance and tenderness (Step 11), it was discovered that the counselor may have to employ as many as seven different strategies.      It is important to note that the concepts outlined are not altogether new and many of them originated from the three idealized models. Concepts such as engagement, parental empowerment and relapse have been in the field for a long time. What is new is the way these concepts and their surrounding steps are mapped out. The task analysis methodology allowed me to gain an in-depth clinical understanding of the "what, when, why and how" of interactions between difficult adolescents and their families. Patterns and themes were identified providing a clearer road map and reducing the complexity of communications between counselor, parent, teenager, and outside systems. This bridged the gap between research, theory and direct practice.      The key discoveries that re-shaped my thought process during this study centered around four areas: (1) relapse, (2) the "Five Aces," (3) the hard and soft sides of hierarchy and (4) rules, consequences and troubleshooting. There were other discoveries but these are the ones that stood out as facilitating substantial amounts of positive client change. Relapse      From the focus group interviews and videotaped session analyses, exciting discoveries were made in the area of relapse. Each time the adolescent appeared to be doing better and the parents were hopeful, the adolescent would relapse the next week or soon thereafter. When this happened, the parents felt betrayed, angry and even more hopeless than before. This often caused the teenager to shut down and become recalcitrant. At this point, the family would sometimes leave treatment or commit a series of no-shows.      The treatment model was not working at this point so I went back to the drawing board to determine what factors accounted for this anomaly. The second idealized model in stage two revealed that the teenager relapsed following a period of normal functioning. The solution was to return to the techniques that worked earlier and reformulate clear rules and consequences to prevent a second relapse. However, this strategy often proved unsuccessful with the difficult families we worked with. They were either too burned out, hopeless or bitter to entertain the idea of refining their rules and consequences. New solutions had to be found.      An intensive analysis of the videotaped sessions in Phase III provided the clues needed to find these solutions. The examinations revealed a common pattern of intervention among the three counselors. In each case, the counselor failed to predict the relapse or prepare the parents for the possibility by dicussing what they should do if it occurred. Every time an adolescent relapsed, the counselor appeared just as surprised and frustrated as the family.      These common patterns led to the creation of Step 13, "Relapse Predicted," and the development its micro-steps of normalizing, prediction and troubleshooting. The counselors were trained to normalize the behavior by explaining to the parents that relapse is common and often expected. Most parents understood this rationale, helping to decrease the bitterness and hopelessness they felt when a future relapse occurred. Additionally, counselors were trained to predict relapse each time the adolescent began to function normally to actually prevent the relapse from happening. When a relapse is predicated, the motivation of the teenager and the family to rally together to prove the counselor wrong frequently increases. Finally, counselors were taught to use "what if" scenarios to design troubleshooting strategies in the event of a relapse.      After the counselors learned these new strategies, more sessions were videotaped and more focus group interviews were conducted. Analysis revealed that the clients' reactions to future relapses were much less negative and destructive. The parents seemed to take the relapses in stride and worked with the counselor to get back on track as quickly as possible. In addition, the parents reported in the focus group interviews that troubleshooting allowed them to feel prepared so that they were not caught off guard when relapse occurred.      Future research in this area would need to test these micro-steps with a broader population before these findings could be generalized. An outcome measure would have to be developed or located that was sensitive enough to pick up changes in the results of these relapse prevention steps. One specific hypothesis to be tested would be to determine if relapse prediction immediately following normal adolescent functioning prevented future relapse and allowed parents to feel less hopeless and more willing to prevent future relapses. This is a good example of how specific micro-theories of change can be generated from process research and tested using outcome research. The "Five Aces"      This was one of the most exciting theoretical yields from the study. As the study broadened to include other counselors and clients, I began to notice a very interesting pattern. Whenever the parents tried to restored their authority, the difficult adolescent would use an extreme behavior to induce the return of their authority from the parents. The use of extreme behaviors initially defeated the counselors in this study due to the difficulty of coming up with effective consequences. The entire process reminds one of a savvy poker player who always has a hidden ace up his sleeve to defeat his opponent at the precise moment the opponent thinks he has won. In the same way, adolescents seem to use their own "aces" to defeat parents and counselors anytime they seem to be winning.      A developmental timeline illustrates how the concept of the "five aces" evolved and how both observables and self-report methods led to this theory. The videotaped sessions revealed a pattern of extreme behaviors neutralizing the parents' and counselors' effectiveness. The following four extreme behaviors seemed to produce the neutralizing effect on the parents' and counselors' authority: (1) running away, (2) threatening suicide, (3) truancy and poor school performance, and (4) threats or acts of violence. A closer analysis of videotaped interviews demonstrated that the adolescents initiated one or more of these extreme behaviors following the parents' attempts to implement predetermined consequences or change their confrontational style. When the rules and consequences were effective and the teenager was unsuccessful in controlling the mood and direction of arguments, the authority shifted to the parents. As the parents became stronger, the teenagers' typical method of regaining their authority by yelling, refusing to comply, nagging, inducing guilt, etc. was no longer effective. The teenagers would then "pull out one of their aces" to counter earlier defeats and regain authority. Parents and counselors appeared not to know what to do to stop these behaviors. If the adolescent was successful, he kept using the ace until the parent and the counselor gave up and things went back to status quo. The parents would then hand their authority over to an outside source (institution, police, extended family) and treatment often ended unsuccessfully. The family-based model was revised to include Step Seven and arrows to illustrate what happened if the parent and counselor failed to stop the teenager's ace(s). The word "aces" is coined to describe these extreme behaviors because they appeared to parallel a savvy poker player who always seemed to have a hidden ace up their sleeve to defeat the parents at the precise moment they appeared to be winning. Focus group interviews with parents revealed a fifth ace called "disrespect" Some parents report that disrespect pushed their "buttons" often to a greater degree that all the other aces combined. In turn, this caused the parents to lose control of their temper and lose their rational thought process. Parents then reacted out of emotion and were unable to maintain consistency or follow through on predetermined rules and consequences. The family-based model is revised again to include this fifth ace of disrespect. The literature is revisited to discover methods and techniques that were effective in neutralizing extreme behaviors. The writings by Haley (1980), Schiff & Benson (1987), Price (1997), and Keim (1997) provided helpful suggestions. A menu of strategies or consequences was developed to effectively neutralize each ace. The team of counselors are given this menu and trained in the use of these strategies. Counselors then implemented these strategies with parents and teenagers during Step Eight. Videotapes and focus group interviews from these interventions were then analyzed and compared back to the revised performance model. Anomalies are discovered or times when these interventions failed or were not effective.      In sum, this timeline demonstrates the iterative process involved in operationalizing theory within actual practice. As new discoveries were made, the researcher had to take the patience and time necessary to explain anomalies at any point during the treatment process. The answers to these questions led to more questions. For example, a discovery of how the parents were defeated through extreme behaviors led to questions on how to stop these behaviors. In turn, these questions led the researcher back to the literature in search of answers. These answers led to the implementation of mini-steps custom built to neutralize a specific extreme behavior. A videotape analysis of these mini-steps led to more questions and changes. In addition, collaboration with clients through self-report interviews led to new information and the creation of a fifth ace, disrespect. Each of these discoverie led to a further refinement of essential concepts and gave the counselor a better road ma p to follow. Hard versus Soft Side Hierarchy      An intensive analysis of Haley and Schiff's work during Phase II, did not result in the discovery of the soft side of hierarchy. Jim Keim's (1996) writings appeared to be the only place in the literature where the concept was described. These writings, however, lacked specification in their implementation. For example, Jim Keim (1996) wrote about concepts such as initiating "soothing sequences" of communication, physical touch, and special outings. However, it remained unclear as to how and when to implement these steps in the overall treatment process. Even though these concepts lacked specification, it appeared that these principles were essential contributors to the probability of treatment success. Constant negative communication patterns between parent and difficult teenager created a dearth of softness and this was needed for permanent change to take place. As a result, I decided tha t it was important to try to implement these procedures somewhere in the overall treatment process. For this to occur, three main questions had to be answered: Where and when should the concept of "restoring nurturance" be implemented within the overall treatment process? In other words, what is the optimal timing of this intervention and what are the parental characteristics that influence this timing? What are all the "mini-steps" or strategies that can be used to successfully implement the treatment intervention of restoring nurturance? Once these steps are discovered how can they be operationalized in such a way that they can implemented in a step-by-step fashion?      Below is a summary of the developmental stages that took place to answer these questions. Points within this developmental timeline where these questions were answered will be highlighted. The timeline contained the following 9 developmental steps: After reading Keim's (1996) writings, the decision is made to implement this concept during Phase III of the project. The team of counselors are trained on how to restore tenderness and nurturance. The supervisor discusses the mini-steps of physical touch, initiating special outings between parent and teenager, and creating soothing sequences or nurturing communication patterns between parent and child. However, since these steps are still in the development stage, the supervisor asks the staff to use their creativity and intuitiveness in implementing each step and the timing as to when they should be introduced. The team is told to videotape every session that this intervention is used. Videotape and focus group interviews are analyzed and compared back to the idealized and revised performance models. Anomalies or failed sessions are carefully scrutinized for patterns and themes. Results from the analysis revealed the answer to the first question: When and where should nurturance be implemented and what client characteristics influence this decision? It is discovered that If the adolescent exhibits one or more extreme behaviors (one of the five aces) or the problem is chronic, the counselor must stop these behaviors first before trying to introduce the concept of nurturance. Videotapes revealed parental hostilities whenever nurturance was introduced before the extreme behavioral problems were solved. In addition, parent self-reports supported this finding when parents (n=9) stated that "one must stop the bleeding" by stopping the problem behavior before they had the time or energy to consider the issue of softness. Parents (n=22) also reported that they had to first establish trust and confidence in their counselor's ability to help them. This was in part established when the counselor could design a consequence to stop an extreme behavior when others had tried and failed. Armed with this confidence and trust, the parent was more willing to take the risk of opening their heart up again and being soft. Based on this information, the step of "Restoring Nurturance and Tenderness" is placed as Step 11 or after the five aces have been neutralized and the teenager is functioning without behavior problems. Focus group interviews partially answered the second question: What "mini-steps" or strategies can be used to successfully implement nurturance. Many of the teenagers interviewed (n=37) reported that they needed an opportunity to regain lost trust. Without trust, resentment and bitterness set in and there was little hope for nurturance. In addition, some of these teenagers (n=17) reported that the potential for nurturance was blocked when parents constantly criticized or failed to understand how they felt. Based on this information, the following new "mini-steps" were adopted and implemented: (1) Opportunities to build trust, (2) A new approach to criticism and praise, and (3) Acceptance of underlying feelings. An analysis of videotapes employing these interventions, helped to refine and operationalize these steps. These results helped answered the third question: Once these steps are discovered how can they be operationalized in such a way that they can implemented in a step-by-step fashion? The strategies of special outings and physical touch were implemented and the videotapes of these interventions closely scrutinized. Sessions that did not work provided the necessary clues needed to successfully implement these strategies. For example, videotape analysis revealed that special outing sessions were unsuccessful when the counselor failed to get the parent and teenager to set a specific date and time. In these session, the counselor also failed to troubleshoot all the things that could go wrong. After these problem variables are identified, counselors were shown how to incorporate specifics and troubleshooting to increase the probability of success. An analysis of videotapes revealed that soothing sequences of communication were conversations where the parent not the child controlled the mood, topic, and direction of the discussion. These discussions did not contain elements of criticism or attacks on teenager's character but rather elements of praise, acceptance, positive rewards, special outings, acceptance of feelings, opportunities to build trust, or signs of good physical touch. Failed sessions revealed that to be successful the counselor must first engage in careful preparation and then practice their delivery. The counselor must show the parent how to properly deliver soothing communication sequences through the use of dry runs or role-plays. These discoveries were invaluable.      In sum, this process is a good example of how a relatively new theoretical concept can be field tested and operationalized using process research. Using task analysis methodology, counseling sessions were intensively studied to reveal clues on how and when to implement this particular strategy. In addition, the interventions were operationally defined. This example also demonstrated how clients could collaborate with researchers to direct them to new areas of investigation and strategies that custom fit their needs. Rules, Consequences, and Troubleshooting      Another important discovery was the correlation between the successful implementation of rules and consequences and the use or nonuse of troubleshooting. In cases where troubleshooting was employed, the parents had a greater degree of success in implementing rules and consequences. In the cases where troubleshooting was not used, the adolescent did something unexpected that often rendered the rule or consequence ineffective. After these patterns were observed, I looked for potential variables within the videotapes to explain these occurances. Throughout this analysis, I asked myself the following question: "What factors could account for this pattern?".      The answer to this question was revealed when tapes that employed troubleshooting were compared with tapes without troubleshooting. The following developmental timeline describes how the answers to this question were revealed: An comparative analysis between videotapes that employed troubleshooting and those that did not revealed the following important discrepancies. The tapes without troubleshooting showed the implementation of rules and consequences without the preparation of "dry runs" or "what if " scenarios. When this happened, rules and consequences were often ineffective as the teenager was able to skillfully push their parent's "buttons" to make them lose control of their emotions or outmaneuver them by thinking two steps ahead. For example, if the consequence was grounding on the weekend for missing school, the teenager would get up early and be out the door before the parent woke up to enforce this consequence. Behaviors like this were unexpected and preplanned by the teenager to throw the parent off track and render the consequence ineffective. Tapes using the strategy of troubleshooting were analyzed that employed the techniques of "dry runs" and "what if" scenarios. The tapes showed that parents who used these strategies were able to deliver rules and consequences more effectively than those who had not used these interventions. Parents who had practiced their delivery through role-plays did not allow their teenager control the mood the discussion or "throw them off track." In addition, the teenager would still try to outmaneuver the parent as before but this time there was a Plan B in place if a particular rule is violated.      As stated earlier, difficult adolescents have both enhanced social perception abilities and the ability to push their parent's buttons. As a result, a comparative videotape analysis revealed the importance of troubleshooting in countering these special skills. The impact of troubleshooting and the timing of its use would have gone undetected without an intensive analysis and comparison of both successful and unsuccessful change episodes. In addition, an analysis of successful change episodes led to the operationalization of the mini-steps of "dry runs" and "what if" scenarios. Phase V: Combining Process with Outcome Research      Once it was determined how the model works, the final step was to use outcome research to determine if the model does work. In this section, I will report the results of a two year pretest/postest outcome study with 82 difficult adolescents and their families. How these results supported or disconfirmed the theoretical concepts that emerged from the process study will be highlighted. Specifically, the following five research questions were tested using standardized outcome measures: At the end of treatment, did the parents show a significant change in negative attitudes with their difficult teenager? At the end of treatment, did the results show a significant change in the parent's role and their ability to be in charge and maintain control over their teenager's problem behavior? At the end of treatment, did both parent and teenager show a significant change in the areas of affective responsiveness or nurturance and tenderness? At the end of treatment, did both parent and teenager show a significant change in negative communication patterns? At the end of treatment, did both parent and teenager indicate client satisfaction with the overall treatment process, even those clients that were involuntarily committed to treatment?      In sum, the standardized measures used were sensitive enough to test the effectiveness of four theoretical constructs from the 15-step Model: (1) Parents ability to take charge; (2) Changing the timing and process of confrontation patterns; (3) Parents ability to neutralize behavior problems (Five Aces); and (4) Restoration of Nurturance and Tenderness. Other concepts such as changes in troubleshooting, relapse, and rules and consequences still need to be tested. These were not tested here because I was unable to locate standardized measures that were theoretically congruent and sensitive enough to measure changes in these areas. It is important to note that most "measures are chosen because they are widely used and have become standard instruments, not because they provide the best test of the impact of a particular family treatment" (Anderson, 1988, p. 83). As a result, the outcome results in this study could only go as far as the sensitivity of the standardized measure used and the extent to which it is theoretically congruent with the concepts being evaluated. Target Population Characteristics      Of the eighty-two families who participated in this study, 68.4% were Caucasian, 10.5% were African-American, 7.9% were Asian-American, and 5.3% were Hispanic. Almost half the parents participating in the study were married (42.1%), with 28.9% divorced, 18.4% single, and the remainder 5.2% were either separated or divorced.      The families averaged 2.9 children (sd=1.3). Over 2/3 of the parents were employed (69.4%), with 19.4% being unemployed, and 11.1% homemakers. Almost 1/4 (20.8%) of the families made less than $10,000; almost 1/3 (29.2%) earned between $10,000 and $20,000; 1/4 (25%) earned between $20,000 and $35,000; and nearly 1/10 (8.3%) earned between $35,000 and $50,000. The remainder of the families (8.4%) earned more than $50,000.      As stated earlier, most of the adolescents treated were males (78.9%) with an average age of 15.3 years (sd=1.5). Over half of the adolescents (52.2%) had a history of fighting or assault; 56% had a history of stealing or shoplifting; 43.5% were truant from school; 39.9% had drug or alcohol problems; 39.1% had problems with running away; and 8.7% had weapons charges, vandalism, or sexual abuse charges. These percentages reflect that a majority of the adolescents had multiple problems. Design and Measures      A non-experimental pretest/posttest design was implemented. Parents and teenagers completed the Family Assessment Device (FAD) and the Client Satisfaction Scale (CSI) separately and independently from one another before treatment began and again after counseling ended. Some families completed treatment in five sessions and some completed treatment within 10 sessions with an average session length of 6.7 sessions. Only the parents completed the Index of Parental Attitudes (IPA). Responses from the pre and post program measures were evaluated to determine the level of change that occurred after the 15-Step Family-Based Treatment Model was implemented. A summary of the scales used are presented below.      Family Assessment Device (FAD) (Epstein, Baldwin, & Bishop, 1983) The Family Assessment Device (FAD) is a 60-item questionnaire designed to evaluate overall health and pathology of a family as well as changes in the families organizational properties and communication patterns that have been found to distinguish between healthy and unhealthy families. Families are given a series of statements ("We are too self-centered"; "Anything goes in our family") and asked whether or not they Strongly Agree, Agree, Disagree, or Strongly Disagree with that particular statement. The FAD identifies and distinguishes between seven key areas of family functioning: (1) Problem Solving, (2) Communication, (3) Roles, (4) Affective Responsiveness, (5) Affective Involvement, (6) Behavior Control, (7) General Functioning.      Index of Parental Attitudes (IPR) (Hudson, 1992) The Index of Parental Attitudes (IPA) is a 25-item questionnaire designed to measure the extent, severity, or magnitude of the parent's overall positive or negative attitude toward their teenager. If the overall pretest or average mean score is above 30, it suggests a clinically significant problem and indicates that the parent has an extremely negative attitude towards their teenager. In addition, there is an increased risk that the parent is experiencing extreme stress with a clear possibility that some type of violence would be considered or used by the parent to deal with the problem. Using this scale, parents are given statements about their child or teenager ("I really enjoy my child" or "I resent my child") and asked to respond with None of the Time, Very Rarely, A Little of the Time, Some of the time, A Good Part of the Time, Most of the Time or All of the Time.      Client Satisfaction Inventory (CSI) (McMurty, 1994) The Client Satisfaction Inventory (CSI) is a 25-item questionnaire designed to measure the overall satisfaction a client has with treatment and their individual perception of how good or bad the services were in general. If the average mean score is below 30, it suggests a clinically significant problem in the clients perception in the quality of treatment. It suggests that the client is extremely unhappy with treatment or the counselor's style and "bedside manner." Scores above 30 indicate the complete opposite. Using this scale, clients are given a statement about counseling or the counselor ("I feel much better now than when I first came here" or "People here are only concerned about getting paid") and are asked to respond with None of the Time, Very Rarely, A Little of the Time, Some of the time, A Good Part of the Time, Most of the Time or All of the Time. Research Questions and Subsequent Results 1)     At the end of treatment, did the parents show a significant change in negative attitudes with their difficult teenager?      Looking at Table 1, parents reported a statistically significant change (m=23.17, p<.05) in negative attitudes towards their difficult teenager. The posttest mean score also dropped below a cut off mean score of 30 and suggests a decreased risk that the parent is experiencing extreme stress or that violence would be used to deal with the teenager's problem behavior.      The pretest mean score of 33.01 indicated that before treatment the parent had an extremely negative attitude towards their teenager. Following treatment, however, this score significantly dropped. This indicates that the family-based model was effective in changing two key areas that affect overall parental attitudes: (1) The timing and process of confrontations, and (2) The restoration of nurturance and tenderness between parent and teenager. During the process study, it was discovered that parental attitudes were affected by these two areas. If the communication was mostly negative, the attitude of the parent would also be negative. In turn, these negative attitudes would severely limit the possibility of injecting nurturance back into the relationship. Parents often reported in focus groups that they "loved their son or daughter" but they did "not like them anymore." In sum, the si gnificant changes in postest parental attitudes supports the hypothesis that changes in confrontational patterns and nurturance can have a positive change on negative relationships between parent and teenager. 2)     At the end of treatment, did the results show a significant change in the parent's role and their ability to be in charge and maintain control over their teenager's problem behavior? Looking at the bar graphs and table 2 and table 3, both parents and teenagers reported significant changes in the parent's role through their ability to be in charge and maintain control over the problem behavior. This is indicated on changes within the FAD subscales of roles, behavior control, and general functioning.      The FAD subscale of "roles" focuses on whether or not the family has a clear set of rules and consequences and if parents clearly assign roles and tasks to the children. Examples of questions include: "We discuss who is to do household chores" or "We make sure members meet their family responsibilities." The posttest mean scores of 1.96 for parents and 1.91 for teenagers indicated that the family-based model was effective in clarifying roles and hierarchy between parent and teenager. This supports the notion that the parent was able to maintain and accept a position of authority following treatment. In turn, this supports the hypothesis that the family based model was effective in putting the parent in charge and maintaining this position of authority.      The FAD subscale of behavior control measures how effective the parent is in controlling problem behaviors and setting up rules and consequences. Examples of questions include: "We have rules about hitting people" or "We have parents who control behavior problems." The posttest mean scores of 1.56 for parents and 1.65 for teenagers indicated that the parent was able to stop or control the problem behavior. In turn, this finding supported the hypothesis that the family-based model was effective in helping the parent to neutralize the teenager's five aces by controlling the particular behavior problem. This supports the inclusion of Step 8: Neutralizing the Adolescent's Five Aces.      The FAD subscale of general functioning measures the overall health or pathology of the family. Examples of questions include: "We don't get along well together" or "We cannot talk to each other about the sadness we feel." The posttest mean scores of 1.85 for parents and 1.81 for teenagers indicated that the overall health and general functioning improved. A basic premise of the family based model is that if the hierarchy is congruent and the parents are back in charge, the overall health of the family will improve. In turn, these results support the inclusion of the family-based model principle of putting the parent in charge of the adolescent's behavior. 3)     At the end of treatment, did both parent and teenager show a significant change in the areas of affective responsiveness and affective involvement or nurturance? Looking at table 4 and table 5, both parents and teenagers reported significant changes in the areas of affective responsiveness or nurturance and tenderness. This is indicated from changes within the FAD subscales of affective responsiveness and affective involvement.      The FAD subscales of affective responsiveness and affective involvement measure whether or not family members show tenderness, concern, and affection for one another. Examples of questions include: "We express tenderness" or "We cry openly." The posttest mean scores for parents and teenagers on both subscales indicate that both parents and teenagers were able to show tenderness, nurturance, and more concerned for one another's welfare. In turn, this finding supported the hypothesis that the family-based model was effective in injecting nurturance back into the relationship thereby supporting the inclusion of Step 11: Restoring Nurturance and Tenderness. 4)     At the end of treatment, did both parent and teenager report a significant change in negative communication patterns? Looking at table 6 and table 7, both parents and teenagers reported a significant change in the area in negative communication patterns. This is indicated from changes within the FAD subscale of communication.      The FAD subscale of communication defines the quality of the exchange of information among family members. The focus is on whether or not verbal messages are clear in content and direct in the sense that the person spoken to is the person for whom the message is intended. Examples of questions include: "We are frank with each other", and "People come right out and say things instead of hinting at them." The posttest mean scores for both parents and teenagers indicate that after treatment there was improved communication and a decrease in bitter and negative confrontations. This finding supported the hypothesis that the family-based model was effective in changing communication patterns thereby supporting the inclusion of Step 6: Changing the Timing and Process of Confrontations. 5)     At the end of treatment, did both parent and teenager indicate client satisfaction with the overall treatment process, even those clients that were involuntarily committed to treatment? Within a ten session framework, 88% of the parents and 83% of the teenagers reported that they were satisfied with treatment on the Client Satisfaction Inventory (CSI). What makes these percentages particularly significant is the fact that a majority of clients were involuntary and came to treatment only because they were required to do so by their probation officer. This was a particularly surprising result and one that indicates that this model has a great deal of promise for the future. The counselor often asks the parent to make a great deal of sacrifices to regain control of the household and take charge of the problem behavior. Therefore, unless the client is satisfied with treatment, there is no reason why they will make any sacrifices or listen to their counselor. In addition, if the client is involuntary, they will often come into the first session with a negative attitude and resist any suggestions for change. This is especially true with difficult teenagers.      Some of the focus group interviews provided clues as to why client satisfaction was so high. Several parents (n=7) reported that this treatment was different and better because it "got down to the business of giving them specific tools to solve the problems they came into counseling for." Parents (n=5) also said that "this kind of counseling was not just office work." The counselor was available 24 hours a day and came to where the problem was whether it be home visits, school visits, or church visits. Teenagers (n=14) reported that the counseling showed their parents how to talk to them differently and without yelling. They also liked the fact that they knew what to expect as it pertained to rules and that there were both negative and positive consequences. Finally, teenagers (n=37) reported that they liked the fact that they had chances to earn back trust and that their opinions and ideas were heard and even integrated into the contract.      In sum, these preliminary findings indicated that parents were satisfied with treatment because the model was clear cut, provided specific tools and strategies, 24 hour a day on call service, and was not limited to one hour a week in office sessions. Teenagers reported that they were satisfied because the model provided better methods of communication, clear rules and consequences, a voice, and opportunities to earn back trust. Further research is needed to tap into additional reasons why clients were satisfied with treatment. Answers to this question will provide valuable insight into the models strengths and weaknesses. Beyond the 21st Century: Conclusions and Future Implications/font>      This chapter demonstrated how process and outcome research can be used together within a single study to operationalize key theoretical concepts and build a treatment model directly from clinical practice. Rather than beginning with hypotheses testing, this study moved toward hypothesis testing as the final step of a rigorous program of discovery process research. In this study, outcome measures supported the effectiveness of four theoretical constructs from the 15-step Model: (1) Parents ability to take charge; (2) Changing the timing and process of confrontation patterns; (3) Parents ability to neutralize behavior problems (Five Aces); and (4) Restoration of Nurturance and Tenderness. In turn, these findings can lead to future task analysis investigations to find out why clients are so satisfied with this treatment model and the specific parts within these four theoretical construct s that produced positive change. In this way, theory building process and verification outcome research are interdependent and complementary because process theory development directed outcome empirical research and now these results are directing the researcher toward further process or task analysis probing. The use of these two methods can lead to better research questions and a better and more refined definition of theoretical concepts within the 15-Step Family-Based Model. Future Implications      As stated earlier, it is my hope that readers will see this model as a work in progress and replicate the research process within this chapter to refine and develop the model further. The next logical step is to custom fit the model to an even large set of variables and possible scenarios. For example, suppose you are presented with these variables and this particular scenario in your practice: A 15 year male comes from a single parent home. He belongs to a gang, uses drugs, and relapses before Step Thirteen in the model and before other family issues surface in Step Twelve. Given this scenario and these variables, what are your best treatment steps and options?      These type of exceptions to the overall model are common. Counselors will want a custom road map for their particular case and problem. One day in the future, the clinician might be able to sit down in front of a computer and type in a custom set of variables. After these variables are analyzed, an interactive computer disk would present the counselor with a custom set of procedures that fit the particular scenario and the set of variables presented. This kind of fine tuning is needed in the future and expected in a society that wants positive change quickly and a health care system that demands it.      As we enter the 21st century, it seems timely for the field to reconsider and reassess its conceptual base by defining its treatment models through process and outcome research. We must struggle to conduct research that moves us closer to answering Frank's (1991) central question in his classic work Persuasion and Healing: The question is not whether psychotherapy works; that goes without saying. Rather, the central question is, what are the central ingredients within a particular treatment method that account for its effectiveness with a particular population and clinical problem. [p.6]      To accomplish this goal, we must return to our roots of discovery-oriented research and a collaboration between the counselor and researcher. A Return to Our Roots      In the 1950s, family therapy was born from discovery-oriented observations behind a one-way mirror of family members sitting around a circle in the same room. Initially, the therapeutic goals and procedures, if any, were only vaguely specified. These observations, however, yielded rich theoretical concepts (i.e. metacommunication, family homeostasis, double bind) and generated new research hypotheses and clinical enthusiasm. In the 1960s, these concepts were incorporated into a diversity of family counseling models (i.e. Structural, Bowenian, MRI Brief Therapy). Family therapy teaching and theorizing flourished and was both conceptually interesting and provocative. Since the 1970s to the present, however, family therapy has become disconnected from its discovery-oriented research base and lost its original zest and focus. The field now resembles the "flavor of the month club" by moving from one fad to the next or "who won" outcome studies that fail to move the field to the next step: Finding out how a particular treatment works and why with a particular population and presenting problem? Direct practice counselors and students are hungry for answers to this question and want mini-steps to find their way within the complexity and multiple layers of a problem.      In addition, a split occurred during the 1970's between those who did research and those that did clinical work. This split is described by Haley (1978): In the 1950's it was taken for granted that a counselor and researcher were of the same species (although the counselor had a more second class status)...Today it seems more apparent that the research stance and the counselor stance are quite different. The researcher must explore and explain all the complex variables of every issue since he is an explorer of truth. The counselor stance is much different. He must use simple ideas that will accomplish his goals and not be distracted by the explorations into interesting aspects of life and the human mind. It seems evident that the creation of the researcher and the creation of the counselor are different enterprises. (pp. 73-74)      This split continues today and must end before the gap between research, theory, and practice can be bridged.      In sum, one must ask themselves this central question: "Is what we are currently doing as a field in model building working?" If the answer is no, then we must look for exceptions or times in the past when it was working and do more of it. I believe that time was in the 50's and 60's with discovery-oriented research that employed a here and now process/outcome template. This chapter and book hopefully represents a beginning step toward this future. The Future of Counseling      At the Evolution of Psychotherapy conference in 1995, Salvador Minuchin and Donald Meichenbaum tried to answer the question: "What is the future of counseling as we approach the 21st century?" Both gave very different answers to this question, but these answers symbolize both the fears and hopes of many counselors as well as presenting a snapshot of the future. I will present these two different viewpoints and show how this model and this research process may offer one possible solution to the dilemmas presented by each speaker.      Meichenbaum was the first to speak. He stated that field of counseling would need to move to manualized treatment models that could evaluate its effectiveness with specific populations (children, adolescents, adults) and treatment issues (anxiety disorders, depression, conduct disorders). These manualized treatments would have to be able to show strengths as well as limitations in order to be reimbursed by third-party payers. Meichenbaum envisioned a time in the 21st century where there would be computerized manuals or interactive disks whereby the counselor or actors would perform and demonstrate each essential strategy and technique. The family member would then take these disks home between sessions and practice each strategy between sessions. He called these disks "catalytic supplements." In addition, there may be a time when a person can go on the internet or tune to a cable station and receive these supplements through a push of the button.      Minuchin then spoke and presented a different viewpoint. He stated that these manualized treatments would be unable to mirror or reflect all that transpires within a particular session. Manualization would also limit the intimacy of treatment and the therapeutic relationship between counselor and client. Finally, this would lead the counselor towards a ridge application of treatment without responding to the individual needs of the client or adequately capturing novel situations and circumstances.      Both viewpoints were different but posed this central dilemma: How can you produce manualized treatments that do not sacrifice the intimacy of the therapeutic relationship but are flexible enough to respond to the client's custom needs without stifling the innovation and creativity of the counselor? The goal of this model was to accomplish this task in the following manner. First, the model offers "treatment guidelines" or a generalized template rather than a ridge application of treatment steps. Even though the steps are numbered, they are done so only to give the counselor a sense of direction. Each chapter gave numerous case examples and "what if" scenarios to give the counselor as many situations as possible within a particular procedural step. For example, in the restoring nurturance chapter, I provided the reader with three possible choices or scenarios based on the particular client characteristics. These different scenarios emerged from the research study and were outlined to give the counselor flexibility.      Second, the intimacy and importance of the therapeutic relationship is never undermined but expanded and written about in almost every step. It is described as a seperate step "engagement" but it is also talked about within many other steps. For example, within "Chapter 8: Working with Outsiders" I specifically talk about the importance of rapport and trust between counselor and client or counselor and outside system. In this way, engagement is not presented as a one time static step but as reoccurring throughout the treatment process.      Third, the model was not produced in a laboratory setting but from actual practice sessions and in collaboration with expert clinicians (Jay Haley and Neil Schiff), counselors, parents, and teenagers. Each time a new discovery was made the concept was field tested using a variety of counselors and clients. These sessions were then analyzed for anomalies. In addition, clients were asked about their perceptions and feelings about a particular intervention or series of interventions. In turn, this feedback was used to refine the model further. In this way, the model was grounded in direct practice and the principles reflected all the complexity and "curve balls" a difficult family could present. This gives the counselor information on how to respond to the client's custom needs without stifling their innovation and creativity.      In sum, this model represents an attempt to address the central dilemmas posed by Minuchin and the needs of the 21st century posed by Meichenbaum. I have already begun the process of experimenting with "catalytic supplements." For example, parents now take home and read the "Neutralizing the Five Aces" chapter or view videotapes of actors demonstrating ways to skillfully change the timing and direction of the confrontation. Preliminary focus group interviews with parents indicate that this has been very helpful addition to treatment by providing clarity to specific strategies. It will always remain a challenge to juxtapose manualization and the complexity of the human relationship. However, as we enter the 21st century, we cannot afford to try. Theory construction can no longer remain as a back-room activity. It must be moved front and center to improve methods for constructing testable theories. References Anderson, C. M. (1988). The selection of measures in family therapy research. In L. C. Wynne (Ed.), The state of the art in family therapy research: Controversies and recommendations. New York: Family Process Press. Bednar, R. L., Burlingame, G. M. , & Masters, K. S. (1988). System of family treatment: Substance or semantics? Annual Review of Psychology, 39, 401-434. Chamberlain, P., & Rosicky, J.G. (1995). The effectiveness of family therapy in the treatment of adolescents with conduct disorders and deliquency. Journal of marital and family therapy, 21(4), 441-459. Epstein, N.B., Baldwin, LM., & Bishop, D.S. (1983). The McMaster family assessment device. Journal of marital and family therapy, 9(2), 171-180. Forehand, R., & Long, N. (1988). Outpatient treatment of the acting out child: Procedures, long-term follow-up data, and clinical problems. Advances in behavior research and therapy, 10, 129-177. Frank, J.D. (1991). Persuasion and healing: A comparative study of psychotherapy. Baltimore: Johns Hopkins University Press. Goetz, J.P., & LeCompte, M.D. (1984). Ethnography and qualitative designs in educational research. New York: Academic Press, Inc. Gurman, A. S., Kniskern, D. P., & Pinsof, W. M. (1986). Research on the process and outcome of marital and family therapy. In S. Garfield & A. Bergin (Eds.), Handbook of psychotherapy and behavior change (3rd ed.). New York: Family Process Press. Haley, J. (1978). Ideas which handicap therapists. In M.M. Berger (Ed.), Beyond the double bind: Communication and family systems, theories, and techniques with schizophrenics. New York: Brunner/Mazel. Haley, J (1980). Leaving home. New York, N.Y: McGraw Hill. Haley, J. (1976). Problem solving therapy Hudson, W.W. (1992). The WALMYR Assessment Scales Scoring Manual. Tempe, AZ: WALMYR Publishing Company. Keim, J.P. (1996, October). Strategic coaching of parents of oppositional kids. Presentation at the American Association of Marital and Family Therapy, Toronto, Canada. Mahrer, A. (1988). Discovery-oriented psychotherapy research: Rationale, aims, and methods. American psychologist, 43 (9), 694-702. McMurty, S. (1994). The WALMYR Assessment Scales Scoring Manual. Tempe, AZ: WALMYR Publishing Company. Meichenbaum, D & Minuchin, S (1995, October). The future of therapy. Presentation at the Evolution of Psychotherapy Conference, Las Vegas, NV. Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press. Minuchin, S. & Fishman, H.C. (1981). Family therapy techinques. Cambridge, MA: Harvard University Press. Minuchin, S., Montalvo, B., Guerney, B.G., Rosman, B.L., and Schumer, F. (1967). Families of the Slums: An exploration of their structure and treatment. New York: Basic Books. Pinsof, W. M. (1988). Strategies for the study of family therapy process. In L. C. Wynne (Eds.). The state of the art in family therapy research: Controversies and recommendations. New York: Family Process Press. Pinsof, W.M. & Wynne, L.C. (1995). The efficacy of marital and family therapy: An empirical overview, conclusions and recommendations. Journal of Marital and Family Therapy, 21(4), 585-610. Price, J.A. (1996). Power and Compassion: Working with difficult adolescents. Guildford Publications: New York,NY. Rice, L N. & Greenberg, L. S. (Eds.) (1984). Patterns of change: Intensive analysis of psychotherapy process. New York: Guilford Press. Rice, L. N., & Saperia, E. P. (1984). Task analysis of the resolution of problematic reactions. In L. Rice & L.S. Greenberg (Eds.), Patterns of change: Intensive analysis of psychotherapy process. New York: Guilford Press. Schiff, N.P. & Belson, R. (1988). The gandhi technique: A new procedure for intractable problems. Journal of marital and family therapy, 14(3), 261-266 Sells, S. P., Smith, T.E., & Moon, S. (1996) An ethnographic study of client and therapist perceptions of therapy effectiveness in a university-based training clinic. Journal of Marital and Family Therapy, 22(3), 321-343 Sells, S. P., Smith, T. E., & Sprenkle, D. (1995). Integrating quantitative and qualitative meth