Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Making strengths-based practice work in child protection : frontline perspectives Oliver, Carolyn 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_spring_oliver_carolyn.pdf [ 2.24MB ]
Metadata
JSON: 24-1.0165904.json
JSON-LD: 24-1.0165904-ld.json
RDF/XML (Pretty): 24-1.0165904-rdf.xml
RDF/JSON: 24-1.0165904-rdf.json
Turtle: 24-1.0165904-turtle.txt
N-Triples: 24-1.0165904-rdf-ntriples.txt
Original Record: 24-1.0165904-source.json
Full Text
24-1.0165904-fulltext.txt
Citation
24-1.0165904.ris

Full Text

    MAKING STRENGTHS-BASED PRACTICE WORK IN CHILD PROTECTION:  FRONTLINE PERSPECTIVES by Carolyn Oliver B.A., The University of Oxford, 1992 M.S.W., The University of London, 1995     A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Social Work) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) March 2014   ? Carolyn Oliver, 2014  ii  Abstract Strengths-based practice has been widely promoted as a preferred approach for statutory child protection work, but its complexity and inconsistent implementation suggest that it may be hard to do. This pragmatic mixed methods study asked frontline workers whether and how they applied strengths-based and solution-focussed ideas in their daily child protection practice and what supported and impeded their efforts. Via 26 semi-structured interviews and an online survey, 225 child protection workers from the Ministry of Family Development in British Columbia, Canada, gave their views.  The study found nearly all workers to be positive about strengths-based practice but 70% had implicit rules about the clients and situations for which strengths-based solution-focussed work was not appropriate or possible. Five definitions of strengths-based practice were identified from their descriptions of the approach, each linked to particular challenges and supports. The most applicable definition, 'Enacting firm, fair and friendly practice', illustrates a way for practitioners to navigate the strengths-based child protection relationship while managing the risk and authority inherent to their role. It involves maintaining a shifting balance between inviting maximum client collaboration and using authority purposefully and suggests the importance of judging impartially, being transparent, attending to the worker-client interaction and seeing clients as fellow human beings. It may support workers to navigate a developmental progression from occasionally doing strengths-based practice to identifying as strengths-based practitioners.  The study highlights the importance of developing organisational capacity to provide the time, resources and culture necessary for strengths-based practice. It suggests that 'fearlessness',   iii  emotional self-regulation and comfort with mandated authority support strengths-based practice and recommends greater emphasis be placed on these qualities in child welfare education. Other recommendations include adopting a system-wide child protection-applicable definition of strengths-based practice like 'Enacting firm, fair and friendly practice', implementing 'strengths-based management' and providing ongoing opportunities for practitioner reflection and education.      iv  Preface This dissertation is an original intellectual product of the author, C. Oliver. Ethical approval for this research study was obtained from the University of British Columbia?s Behavioural Research Ethics Board. The certificate number for the study, entitled 'Strengths-based practice in child protection', is H11-03423.     v  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents .......................................................................................................................... v List of Tables ................................................................................................................................ xi List of Figures .............................................................................................................................. xii Acknowledgements .................................................................................................................... xiii CHAPTER 1 : INTRODUCTION ............................................................................................... 1 Background to the Study .............................................................................................................. 1 Strengths-Based Practice in Child Protection Work ................................................................. 1 Research Rationale ....................................................................................................................... 3 Research Questions ....................................................................................................................... 4 Theoretical Framework ................................................................................................................ 5 Overview of the Dissertation ........................................................................................................ 5 CHAPTER 2 : LITERATURE REVIEW .................................................................................. 8 Strengths-Based Solution-Focussed Child Protection Practice ................................................ 8 The Roots of Strengths-Based Practice ....................................................................................... 8 The Roots of Solution-Focused Practice ................................................................................... 11 The Blurring of Theoretical Boundaries ................................................................................... 14 The Strengths-Based Solution-Focussed Child Protection Approach ....................................... 15 The Strengths-Based Solution-Focused Child Protection Relationship .................................... 20 The Implementation of Strengths-Based Solution-Focused Child Protection .......................... 23 Historical Changes in the Child Protection Worker-Client Relationship ............................. 27 1870 - 1930................................................................................................................................ 29 1930 - 1960................................................................................................................................ 38 1960 - 1990................................................................................................................................ 49 1990 - Present ............................................................................................................................ 56   vi  Current Contextual Challenges ................................................................................................. 59 1) Systemic Pressures ................................................................................................................ 59 2) Worker Capacity ................................................................................................................... 60 3) Violence ................................................................................................................................ 61 4) Mandated Authority .............................................................................................................. 62 5) Differing Perspectives on Good Practice .............................................................................. 63 Relational Social Work .......................................................................................................... 64 Mandated Relationship Approaches ...................................................................................... 65 Summary ...................................................................................................................................... 69 CHAPTER 3 : STUDY DESIGN AND METHODOLOGY ................................................... 71 Epistemology ............................................................................................................................... 71 Pragmatism ................................................................................................................................ 71 Methodology ................................................................................................................................ 73 Mixed Methods ......................................................................................................................... 73 Quantitizing (Quantifying) Qualitative Data ......................................................................... 76 Interpretive Description ......................................................................................................... 77 Methods ........................................................................................................................................ 81 Recruitment ............................................................................................................................... 81 Survey .................................................................................................................................... 81 Interviews .............................................................................................................................. 83 Delegated Aboriginal Agencies ................................................................................................ 84 Consent ...................................................................................................................................... 84 Confidentiality and Data Management ..................................................................................... 85 Data Collection ............................................................................................................................ 87 Online Survey ............................................................................................................................ 87 Survey Design........................................................................................................................ 87 Attitude Questions ................................................................................................................. 89 Think Aloud Interviews ......................................................................................................... 91 Interviews .................................................................................................................................. 93 Data Analysis ............................................................................................................................... 96 Plan for Statistical Analysis of Survey Data ............................................................................. 96 Qualitative Analysis of Interview and Survey Data ................................................................ 101 Quantitizing (Quantifying) the Qualitative Data ..................................................................... 105 Quality ........................................................................................................................................ 107 Summary .................................................................................................................................... 111    vii  CHAPTER 4 : DESCRIPTIVE INFORMATION................................................................. 112 British Columbia Ministry for Children and Family Development (MCFD) ..................... 112 Definition of MCFD Terms ...................................................................................................... 115 Intake (ITK) Teams ................................................................................................................. 115 Family Development Response (FDR) Teams........................................................................ 116 Family Service (FS) Teams ..................................................................................................... 116 Integrated (INT) Teams ........................................................................................................... 116 Integrated Case Management (ICM) ....................................................................................... 117 Description of Survey Participants.......................................................................................... 117 Survey Sample Representativeness ......................................................................................... 119 Description of Interview Participants ..................................................................................... 120 Researcher Perspective ............................................................................................................. 121 CHAPTER 5 : SURVEY FINDINGS ...................................................................................... 123 Knowledge and Use of SBP ...................................................................................................... 123 Support for SBP ........................................................................................................................ 124 Applicability of SBP .................................................................................................................. 126 'Sometimes' and 'Always' Applicable ...................................................................................... 129 The Use of SBP Techniques ..................................................................................................... 130 When SBP is Inappropriate ..................................................................................................... 133 Challenges .................................................................................................................................. 135 Organisational Factors............................................................................................................. 136 Client Factors .......................................................................................................................... 140 Practice Factors ....................................................................................................................... 142 Comparing Challenges Identified by the 'Sometimes' and 'Always' Groups....................... 143 Supports Needed ..................................................................................................................... 145 Comparing Supports Identified by the 'Sometimes' and 'Always' Groups .......................... 149 CHAPTER 6 : INTERVIEW FINDINGS .............................................................................. 150 Different Definitions of SBP ..................................................................................................... 150 Group 1. Relating Therapeutically (N=3) ............................................................................... 153 Challenges ............................................................................................................................... 155 1) Shifting Roles and Severing Trust .................................................................................. 156 2) Managing Other Parties to the Relationship ................................................................... 158 3) Glossing Over the Risks .................................................................................................. 160   viii  Supports ................................................................................................................................... 161 1) Applying SBP Strategically ............................................................................................ 162 2) Believing SBP Empowers the Disempowered ................................................................ 163 Group 2. Supporting Client Self-Determination (N=7) ......................................................... 164 Challenges ............................................................................................................................... 166 1) Balancing Support and Addressing Risk ......................................................................... 166 2) Being Left Out on a Limb ............................................................................................... 168 3) Needing Client Buy-In .................................................................................................... 169 4)  Paying a Personal Price .................................................................................................. 170 Supports ................................................................................................................................... 171 1) Experiencing Success ...................................................................................................... 172 2) Being Transparent ........................................................................................................... 173 3) Holding Convergent Values ............................................................................................ 174 4) Having Team Support ..................................................................................................... 174 Group 3: Connecting to Internal and External Resources (N=6) ........................................ 175 Challenges ............................................................................................................................... 179 1) Finding the Time ............................................................................................................. 180 2) Managing Emotions ........................................................................................................ 180 3) Needing Client Buy-In .................................................................................................... 183 4) Lacking Support From Above ......................................................................................... 183 5) Lacking Resources .......................................................................................................... 184 Supports ................................................................................................................................... 185 1) Caring for Clients ............................................................................................................ 185 2) Learning Through Training ............................................................................................. 186 3) Setting Clear Boundaries ................................................................................................. 186 4) Having Team Leader Support ......................................................................................... 187 Group 4: Pursuing a Balanced Understanding (N=4) ........................................................... 187 Challenges ............................................................................................................................... 193 1) Balancing Support and Addressing Risk ......................................................................... 193 2) Lacking Clinical Support ................................................................................................ 194 3) Lacking the Time ............................................................................................................ 195 Supports ................................................................................................................................... 196 1) Receiving Skilled Feedback ............................................................................................ 196 2) Being Open to Learning .................................................................................................. 198 3) Having a Clear Sense of Purpose .................................................................................... 199 4) Believing That All People Have Strengths ..................................................................... 201 5) Being Transparent About Potential Conflict ................................................................... 202 Group 5: Enacting Firm, Fair and Friendly Practice (N=4) ................................................. 203 Being Transparent ................................................................................................................ 204 Judging Impartially .............................................................................................................. 209 Using Authority Purposefully .............................................................................................. 211 Inviting Maximum Collaboration in the Process ................................................................. 213 Attending to the Interaction ................................................................................................. 215   ix  Seeing Clients as Human ..................................................................................................... 217 Using Strengths.................................................................................................................... 220 Challenges ............................................................................................................................... 222 1) Surviving an Unsupportive Culture ................................................................................. 222 2) Sustaining the Effort........................................................................................................ 224 Supports ................................................................................................................................... 225 1) Using Self-regulation Skills and Strategies ..................................................................... 226 2) Being Comfortable with Mandated Role and Authority ................................................. 228 3) Having Self-Confidence .................................................................................................. 229 4) Having Convergent Values ............................................................................................. 230 5) Being Humble ................................................................................................................. 231 Characteristics of Interviewees Holding Different Definitions of SBP ................................ 233 Changing Definitions of SBP.................................................................................................... 235 The Alternative to SBP ............................................................................................................. 239 General Implementation Factors............................................................................................. 241 Management Support .............................................................................................................. 241 Fear and Fearlessness .............................................................................................................. 245 a) Engaged ........................................................................................................................... 248 b) Firm ................................................................................................................................. 249 c) Open ................................................................................................................................ 250 What Makes Workers Fearless? .............................................................................................. 251 Summary .................................................................................................................................... 252 CHAPTER 7 : DISCUSSION, RECOMMENDATIONS AND CONCLUSION................ 254 Introduction ............................................................................................................................... 254 Acceptance of SBP ideas......................................................................................................... 254 Applicability of SBP ideas ...................................................................................................... 258 Different Definitions of SBP ................................................................................................... 264 The Three Least Applicable Definitions of SBP ................................................................. 265 Pursuing a Balanced Understanding .................................................................................... 271 Enacting Firm, Fair and Friendly Practice ........................................................................... 273 Being Transparent ................................................................................................................ 276 Judging Impartially .............................................................................................................. 277 Seeing Clients as Human ..................................................................................................... 279 Authority: The Use of Professional Power ............................................................................. 281 Becoming a Strengths-Based Child Protection Practitioner ................................................... 285     x  General Implementation Factors: ........................................................................................... 293 Organisational Supports .......................................................................................................... 293 Time and Resources ............................................................................................................. 295 Firm, Fair and Friendly Management .................................................................................. 296 Fear and Fearlessness .............................................................................................................. 302 Design Limitations, Strengths and Areas for Further Research .......................................... 307 Recommendations ..................................................................................................................... 310 Conclusion ................................................................................................................................. 313 References .................................................................................................................................. 317 Appendix A: Recruitment Email ............................................................................................. 333 Appendix B: Interview Consent Form .................................................................................... 334 Appendix C: Survey Questions ................................................................................................ 337 Appendix D: Initial Interview Guide ...................................................................................... 341     xi  List of Tables  Table 1 Attitude Questions ........................................................................................................... 90 Table 2 Questions for Statistical Analysis .................................................................................... 97 Table 3 Self-Assessed Knowledge and Use of SBP (N = 224)................................................... 123 Table 4 Support for SBP (N = 224) ............................................................................................ 124 Table 5 Correlations Among Worker Characteristics and Attitudes of Support for SBP .......... 125 Table 6 Attitudes Regarding the Applicability of SBP (N = 224) .............................................. 127 Table 7 Correlations Among Worker Characteristics and Attitudes about SBP Applicability .. 128 Table 8 Use of SBP Techniques in Previous 10 Adult Client Conversations ............................ 131 Table 9 Correlations Between Worker Characteristics and Use of SBP Techniques ................. 132 Table 10 When SBP is Not Appropriate (n = 102) ..................................................................... 134 Table 11 Factors Making SBP Challenging (n = 200) ............................................................... 136 Table 12 Challenges Identified by the 'Sometimes' and 'Always' Groups .................................. 144 Table 13 The Locus of Challenges ............................................................................................. 145 Table 14 Supports Needed to do SBP (n = 204) ......................................................................... 145 Table 15 Needed Supports Identified by the 'Sometimes' and 'Always' Groups ........................ 149 Table 16 Characteristics of Interviewees by Definitional Group ............................................... 234 Table 17 Team and Qualification of Interviewees by Definitional Group ................................. 234 Table 18 Reasons for Fearlessness ............................................................................................. 251 Table 19 The Values and Strategies of Strengths-Based Management ...................................... 297    xii  List of Figures  Figure 1 Quality Criteria for Mixed Methods Studies ................................................................ 108 Figure 2 Five Definitions of SBP................................................................................................ 152 Figure 3 Challenges to 'Relating Therapeutically' SBP .............................................................. 155 Figure 4 Supports for 'Relating Therapeutically' SBP ................................................................ 162 Figure 5 Challenges to 'Supporting Client Self-Determination' SBP ......................................... 166 Figure 6 Supports for 'Supporting Client Self-Determination' SBP ........................................... 172 Figure 7 Challenges to 'Connecting with Internal and External Resources' SBP ....................... 179 Figure 8 Supports for 'Connecting to Internal and External Resources' SBP ............................. 185 Figure 9  Challenges to 'Pursuing a Balanced Understanding' SBP ........................................... 193 Figure 10 Supports for 'Pursuing a Balanced Understanding' SBP ............................................ 196 Figure 11 'Enacting Firm Fair and Friendly Practice' SBP ......................................................... 204 Figure 12 Challenges to 'Enacting Firm, Fair and Friendly Practice' SBP ................................. 222 Figure 13 Supports for 'Enacting Firm, Fair and Friendly Practice' SBP ................................... 226 Figure 14 Becoming a Strengths-Based Practitioner .................................................................. 286      xiii  Acknowledgements  To the British Columbia Ministry of Children and Family Development child protection workers who participated in this study. For sharing your time and wisdom and reminding me why we all do this work - thank you.  To Doug Hughes and Jayn Tyson. Without your commitment this study would not have happened - thank you.  To Dr. Grant Charles, Dr. Sheila Marshall, Dr. Richard Sullivan and Dr. Shafik Dharamsi. For your  knowledge, encouragement and insight - thank you. And in particular to Grant, for your constant support and astute advice - thank you.  To Matthew. To Georgia. For everything - thank you.  1  CHAPTER 1 : INTRODUCTION Background to the Study I have worked in the statutory child protection field for most of my professional career. My child protection training and much of my practice was oriented to carrying out the statutory duty to investigate reports of abuse and neglect and assess the risk of future harm. My challenge was always how to meet statutory obligations while working with the child?s parents in ways that were supportive, compassionate and effective.   At the turn of this century the ideas of strengths-based solution-focussed practice began to enter the discourse, education and practice of child protection work. Its emphasis on the quality of the worker-client relationship reawakened my interest in how to most effectively bring about change through casework relationships combining care and control. My efforts to integrate strengths-based ideas into my practice and teaching of new child protection workers raised questions as to the compatibility of the approach with child protection work and the conditions necessary for its enactment. The worker-client relationship described as necessary to make strengths-based practice work in child protection appeared to be so complex that I wondered whether it demanded relational skills and supports beyond the reach of many workers. The outcome is this study, in which it is those tasked with making strengths-based solution-focussed practice work who provide their answers.     Strengths-Based Practice in Child Protection Work Contemporary strengths-based solution-focussed child protection practice originated in the middle of last century in two very distinct areas of mental health work. The first was strengths-  2  based case management (Marty, Rapp, & Carlson, 2001; C. Rapp, 1993; C. Rapp & Wintersteen, 1989; R. Rapp & Lane, 2012; Saleebey, 2012). This was developed by social workers for the voluntary clients of mental health services. The second was the brief therapeutic intervention called solution-focussed therapy (De Shazer, 1982; C. Franklin, Trepper, McCollum, & Gingerich, 2012; Shazer et al., 1986). Solution-focussed work has been called a strengths-based approach (Bond, Woods, Humphrey, Symes, & Green, 2013; C. Rapp, Saleebey, & Sullivan, 2006) and over the last decade ideas from these two traditions have increasingly merged under the broad category of strengths-based practice (Gray, 2011; R. Jack, 2005; Lietz, 2011; Skrypek, Idzelis, & Pecora, 2012; Skrypek, Otteson, & Owen, 2010). In this dissertation strengths-based practice, abbreviated as SBP, refers to all ideas within this broad category, drawing both from the Kansas strengths-based case management and solution-focussed therapy traditions.        SBP has needed adaptation for child protection work and a number of models have now been developed. The original adaptations were Insoo Kim Berg and Susan Kelly's solution-focussed work (2000) and Andrew Turnell and Steve Edwards' Signs of Safety? model (1999). These described a similar approach, which remains the foundation of contemporary strengths-based child protection practice. The core of strengths-based solution-focussed work remains intact, requiring practitioners to elicit and reinforce the client?s strengths and goals. However workers must simultaneously assess and prioritize the child?s need for safety. Child protection workers carry a legal mandate to respond to reports that a child may be at risk of abuse or neglect. They must assess the child?s situation and intervene if necessary to ensure the child?s safety and well-being. They work primarily with the child?s adult caregivers, who are frequently involuntary parties to the relationship. The strengths-based solution-focused child protection literature suggests that to make this work, practitioners must engage with their mandated adult   3  clients in complex relationships balancing therapeutic support and the use of their mandated authority.   Research Rationale SBP has been heralded for more than a decade as a progressive new approach to statutory child protection work and introduced to statutory child protection agencies in the U.S.A., Canada, Western Europe, Australia, New Zealand and Japan (Antle, Christensen, van Zyl, & Barbee, 2012; Bunn, 2013; Idzelis Rothe, Nelson-Dusek, & Skrypek, 2013; Turnell, 2012). Current policy within the Ministry of Children and Family Development (MCFD), the provincial child protection agency in British Columbia, positions SBP as the primary approach (Ministry for Children and Family Development, 2012). Some child protection workers appear to be enacting strengths-based solution-focused ideas and there is early evidence of positive results (Antle et al., 2012; Bond et al., 2013; Bunn, 2013; Department for Child Protection, 2010; Idzelis Rothe et al., 2013; Keddell, 2012; Skrypek et al., 2012; Skrypek et al., 2010). However, we do not know how typical these workers are or what it was that makes it possible for them to work in a strengths-based way. Some workers clearly find it hard to do SBP with child protection clients (Antle et al., 2012; Lietz, 2011; Roose, Roets, & Schiettecat, 2012; Skrypek et al., 2010; Smith & Donovan, 2003) and broad concerns have been expressed that much social work done in the name of SBP is not strengths-based at all (Blundo, 2001, 2012; Grant & Cadell, 2009; C. Rapp et al., 2006).   In this study I set out to explore what could be learnt from the different ways in which workers applied the ideas of SBP in their child protection work. There has been a tendency in the SBP literature to frame the approach's inconsistent implementation as indicative of individual   4  and organisational failings. Informed by pragmatism, my approach was to view the differing application of strengths-based ideas as providing valuable information about the usefulness of the approach in the child protection context. I sought to identify the child protection workers, clients, situations and contexts for which strengths-based solutions-focused practice worked, and those for which did it not. I wanted to hear what attributes, experiences, beliefs, practices and situations enabled some to do strengths-based solution-focused practice more readily than others who faced similar resource constraints and statutory demands. This I hoped might support the continued adaptation of the approach to the particularities of child protection practice and the provision of better supports for its enactment.  It was important to me that this study offered the opportunity for practitioners to reflect on their experiences, share their hard-earned wisdom and contribute to the development of a key practice approach. As described by Lipsky (2010), frontline workers interpret policy and enact practice in very contextually specific ways. I hoped that by listening to the many ways in which this happened I might develop a more nuanced understanding of the fit between SBP and child protection work. The goal was that this would inform action to increase the accessibility and usefulness of the approach to those on the frontline.  Research Questions The research questions guiding this study were: 1. Do child protection workers apply the ideas of strengths-based solution-focused practice, and if so, how?  2. What do they perceive as helping and hindering them in this process?     5  Theoretical Framework The overarching theoretical perspective for this study was pragmatism (Dewey, 1920/2004; Haack & Lane, 2006; Ketner, 1992). This guided all decisions regarding methodology and research design. Pragmatism proposes that human experience is best understood as a process of continuous interaction between meaning-making and action. The test of ideas is in their application; it is whether and how we use ideas that determines their value. Pragmatism highlights the contribution of both individual and structural factors to human experience which is envisaged as the product of a continuous reciprocal process of interaction between people and their environment.    One implication of a pragmatist perspective is that research design decisions and methods are dictated by an assessment of what will best answer the research question rather than being made to align with a detailed and pre-constructed conceptual framework. It supports researchers to understand a phenomenon from multiple angles and is highly congruent with the mixed methods and interpretive description approaches employed in this study. It orientates the researcher to focus on the goal of useful knowledge that might inform the target audience, in this case personnel within the MCFD child protection system, to take more productive action regarding a practice problem.     Overview of the Dissertation This first chapter provides a brief introduction to the study. In the second chapter I lay out the dominant themes in the related literature in order to identify how this study is informed by and contributes to child protection knowledge. I firstly discuss the theoretical roots of SBP, examine how it has been adapted for child protection work and detail the relationship between worker and   6  client envisaged in literature relating to these child protection adaptations. I highlight the importance of the broader political, social and economic context by describing key changes in the dominant conceptualisation of the worker-client relationship since the inception of the modern child protection enterprise. I then consider some of the common challenges of statutory child protection work that might impact the ways in which practitioners enact SBP.   In Chapter Three I explain my epistemological and methodological choices. I describe how I conducted the study, with particular attention to data collection and analysis processes and the criteria for quality. In Chapter Four I give an overview of the British Columbia Ministry for Children and Family Development, the government child welfare agency partnering with me in this research. I define key terms used within MCFD and this study. I then describe key characteristics of the frontline child protection workers employed by MCFD who participated in the study. I finish this chapter by describing the assumptions, experiences and theoretical lenses that I brought to the study.     In Chapter Five I present findings from the online survey, starting with survey responses regarding participant knowledge and use of SBP and support for the approach. I then describe results regarding perceptions of the applicability of SBP and the barriers and supports for its enactment.  In the next chapter I describe findings from the interview data. I outline the five different definitions of SBP identified from interviewee accounts, with accompanying supports and barriers to their implementation. I detail two factors which appeared important to the implementation of all five definitions of SBP: management support and a quality of 'fearlessness'.   In the final chapter I discuss the meaning and implications of the findings. I link the findings both to the literature outlined in Chapter Two and to theory related to learning and    7  developmental processes. I consider the strengths and limitations of the study and suggest areas for future research before outlining recommendations for MCFD that might support SBP to be implemented more broadly and experienced as more applicable to frontline child protection practice.      8  CHAPTER 2 : LITERATURE REVIEW  Strengths-Based Solution-Focussed Child Protection Practice The last decade has seen the development of strengths-based solution-focused approaches for use by frontline child protection workers with mandated clients in statutory child welfare agencies. These are centred on the principles and techniques of solution-focused therapy, although are often referenced generically as 'strengths-based' and situated within the tradition of strengths-based practice (SBP). In the next section I describe the distinct theoretical roots of strengths-based practice and solution-focussed work and how these have found expression in strengths-based child protection practice.   The Roots of Strengths-Based Practice Strengths-based practice developed in the United States through work with people with serious and chronic mental illness. It was prompted by the financial demands of managed care (Berg & Kelly, 2000) and the failings of the community treatment model implemented from the 1950?s onwards as a response to the mass deinstitutionalization of psychiatric patients. By the mid 1980?s mental health services tended to be delivered through day treatment programs, highly paid professionals offering individual therapy, and residential treatment in the event of crisis (Kisthardt, 1997; Macias, William Farley, Jackson, & Kinney, 1997; C. Rapp & Chamberlain, 1985). This left clients with little help to navigate challenges of daily living like securing and maintaining housing, income, education and community supports. What help there was tended to be offered on an ad hoc crisis-driven basis by staff in the day treatment or residential programs (Macias et al., 1997), or fell under the remit of therapists who had little interest in this work (C. Rapp & Chamberlain, 1985). In some areas case managers operating on a 'service broker' model   9  assessed and linked clients to community resources. However the lack of such community resources, and of a meaningful relationship with clients in a model that was often telephone-based, problem-focused and separated functional and therapeutic tasks between case manager and therapist meant many clients fell through the cracks (Kisthardt, 1997; C. Rapp & Chamberlain, 1985).  Led by Rapp, a group of academics at the University of Kansas School of Social Welfare piloted the strengths case management approach (Kisthardt, 1997; Modrcin, Rapp, & Poertner, 1988; C. Rapp, 1993; C. Rapp & Chamberlain, 1985; Weick, Rapp, Sullivan, & Kisthardt, 1989). It was a model of 'aggressive outreach' requiring regular contact with the client outside of the office (Modrcin et al., 1988). The case manager accompanied the client, psychologically and physically, on their journey to achieve their goals in community living. This working relationship was seen as central to the achievement of client goals and was increasingly recognized as a therapeutic intervention in itself and as a key factor in the reduction of hospital admissions and psychiatric symptomology. The focus was on client strengths rather than pathology and clients were seen as the directors of their own case planning. Case managers completed a strengths assessment with the client and the work was oriented to the specific achievable goals outlined by clients using a personal planning tool.   While there was little interrogation of its theoretical roots in the early writing of the Kansas group, the strengths approach was later described as developing from constructivism and research in the areas of health and wellness, resilience and human development that countered ideas of a fixed developmental path and the inevitable route from adversity to pathology (Saleebey, 1996, 2012). It was founded on a critique of the ?empirical?technological paradigm? (Weick, 1987, p. 224) and the belief that social work's emphasis on problems merely gave them a   10  new vitality (Weick et al., 1989). Key principles of the approach were that all clients had the capacity for continual growth and that the community, rather than being a barrier to wellness, was rich in health-promoting resources (Modrcin et al., 1988). The view that human behaviour is largely a function of available resources (C. Rapp & Chamberlain, 1985) also suggested the influence of ecological and system theories (Germain, 1978; Pincus & Minahan, 1973).    It is notable that a key feature of the Kansas strengths model was that case managers were ?preprofessional personnel? (C. Rapp & Chamberlain, 1985, p. 419). In the first studies they were social work students (Kisthardt, 1997; C. Rapp & Chamberlain, 1985) or workers with no human service experience or advanced degrees (Modrcin et al., 1988). This was deemed significant because it meant case managers really did feel they had something to learn from clients (Kisthardt, 1997). Client self-determination was the core principle of this approach which held that it is impossible for even the best trained professional to judge how another person should best live his or her life. The nonjudgmental attitude in social work dictates that not only should social workers not judge but that social workers cannot judge (Weick et al., 1989, p. 353).  It appears that case managers either did not exercise statutory powers over these clients or in the occasional instance where clients were identified as mandated by legal order to attend outpatient services (Kisthardt, 1997, p. 103) this was not deemed to present problems for the strengths approach.  Saleebey, also from the University of Kansas, broadened the appeal of the strengths approach by describing it as a general social work perspective applicable to clients beyond the realm of adult mental health (Saleebey, 1992, 2006). However by the turn of the new century   11  little had been written about translating this approach directly to work with families (Early & GlenMaye, 2000). Noble, Perkins and Fatout (2000) cited Rooney's (1992) advice for work with involuntary clients when they called for child protection workers to both be clear about their statutory role and to apply strengths-based ideas in the Kansas tradition. More recent papers, however, have highlighted the theoretical incompatibility of this tradition of SBP and statutory work (Mirick, 2013; Murphy, Duggan, & Joseph, 2013). Few authors have attempted to describe SBP in child protection without the addition of solution-focussed ideas.  The Roots of Solution-Focused Practice Current strengths-based child protection models draw most heavily on the ideas of solution-focused therapy, which developed very separately from the Kansas strengths approach. Solution-focused therapy emerged in the United States in the 1980?s from therapeutic work with adults and families (De Shazer, 1982; Shazer et al., 1986). It was the culmination of attempts over the previous two decades to develop a model of brief therapeutic treatment that was more effective than shortened forms of conventional treatment (Weakland, Fisch, Watzlawick, & Bodin, 1974). Therapists at the Mental Research Institute in Pal Alto, California, had developed the Focused Problem Resolution model (Weakland et al., 1974) which offered clients no more than 10 sessions and focused on changing present observable behavioral interactions rather than mining the past for causative explanations. Drawing on general systems theory (Bateson, 1979) and Milton Erickson?s brief hypnotherapy approach (Haley, 1993), it framed problems as interactional and situational and saw the therapist?s role as being to interrupt the habitual patterns in thinking and behavior that maintained the problem. A small change at any point in the system, a new behavior or even just the relabeling of a behavior, had the power to change the whole   12  system. Steve de Shazer and his colleagues at the Brief Family Therapy Centre in Wisconsin articulated these ideas as the ?ecosystemic approach? (De Shazer, 1982) which underpinned their new approach of solution-focused therapy.   The most radical element of solution-focused therapy was to move away from the interest in problematic behaviors which were the subject of other brief therapy models to a focus on solutions. All clients were believed to want change and to be already doing things that would help solve their presenting problem. The therapist explored all attempts to cooperate and all exceptions to the problem situation in order to harness the client?s inherent motivation and to match the intervention to the client?s ways of thinking and behaving to increase the chance that it might be well received. The therapist worked with the client?s view of reality while introducing sufficient ?news of a difference?  (De Shazer, 1982, p. 8) to expand the client?s behavioral or cognitive responses to a problem and prompt system change.  The solution-focused therapist?s main interest after the first session was not in revisiting the problem situation but in exploring what the client did that was different. The therapist used techniques like miracle, exception and coping questions to elicit detailed descriptions of client goals and successes (De Jong & Miller, 1995). He challenged the client?s construction of problems as continuous and reinterpreted taken-for-granted behavior and assumptions as a matter of client choice. By focusing on narratives of strengths and success he reinforced the client?s motivation to choose what worked. This built on constructivist and narrative ideas about narrating new interpretations opening up new possibilities for action: ?The solution-focused language game is designed to persuade clients that change is not only possible, but that it is already happening. It is, in other words, a rhetorical process designed to talk clients into solutions to their problems? (Miller & de Shazer, 1998, p. 7).   13   The brief therapy tradition framed the therapist?s authority in a very different way from the Kansas strengths approach. The therapist was an expert who delivered directives, tasks and reinforcing messages for strategic ends: Since we as therapists are by definition experts, giving authoritative instructions on both thinking and acting, another pervasive element of paradox is created by the fact that ordinarily we do so only tentatively, by suggestions or questions rather than direct orders, and often adopt a "one-down" position of apparent ignorance or confusion. We find that patients, like other people, accept and follow advice more readily when we avoid "coming on strong?. (Weakland et al., 1974, p. 9) In solution-focused work the therapist first delivered ?compliments? from the therapy team to help the client to feel sufficiently understood to accept the therapist?s authority, before offering ?clues? or strategic suggestions for behavior change. Taking a ?one-down? position in which the client was promoted as the expert on their own situation circumvented client ?resistance?. The approach, from Erickson onwards, was underpinned by a belief in the untapped creative healing capacities of all clients. The discourse of partnership was more complicated here than with the Kansas strengths approach however, as it was both a core value and a strategic maneuver: Simply, the start of the therapeutic message is designed to let clients know that the therapist sees things their way and agrees with them. This, of course, allows the clients to agree easily with the therapist. Once this agreement is established, then the clients are in a proper frame of mind to accept clues about solutions, namely, something new and different. (Shazer et al., 1986, p. 8)     14   Work with involuntary clients was important to the development of the solution focused approach (De Jong & Hopwood, 1996). Many of the clients of the Brief Family Therapy Center were referred or mandated to attend by public agencies like courts and child welfare agencies. The approach was described as working particularly well for mandated clients as it did not seek to directly challenge resistance (De Jong & Berg, 2001). It borrowed from Erickson?s ideas about accepting and working with whatever the client offered and about the value of implicit, indirect and paradoxical interventions to induce small cognitive or behavioral changes without directly challenging the client?s frame of reference (Haley, 1993): When we work with mandated clients we do not attempt to transform potentially resistant clients into cooperative ones by influencing techniques . . . As they reflect, puzzle, and struggle to answer solution-focused questions, new possibilities for doing something different often emerge and cooperation naturally happens. (De Jong & Berg, 2001, p. 372)  It was not uncommon in solution-focused work, for the therapist to differentiate himself from the mandating agency and to ally with the client on the question of ?how can I help you to convince them you are doing well?? (De Jong & Berg, 2001). The client?s wish to escape the clutches of mandated services could be a key motivator for, and goal of, change (Berg & Miller, 1992).  The Blurring of Theoretical Boundaries While they have very distinct roots, the last decade has seen a blurring of the theoretical distinctions between strengths-based and solution-focused practice. Indeed Rapp, Saleebey and Sullivan have defined solution-focused practice as one of four strengths-based approaches (C. Rapp et al., 2006). Academic writing in child welfare is beginning to talk about a strengths and   15  solution-focused approach (Christensen & Antle, n.d.; Skrypek et al., 2012; Skrypek et al., 2010) and to draw on references from both the Kansas and solution-focused therapy literature (R. Jack, 2005; Lietz, 2011). A common viewpoint is that the strengths approach is best described as an attitude or standpoint (Saleebey, 1997; Staudt, Howard, & Drake, 2001), while solution-focused therapy is a model operationalizing this standpoint (Gray, 2011).   The different ways in which the two approaches conceptualize practitioner expertise and authority make this convergence somewhat problematic. The strengths-based approach arguably has a less complicated relationship with worker authority, framing it as the opposite of collaboration: ?The strengths approach to social work practice values empowerment of individuals seeking services and advocates a relationship of collaboration as opposed to one of authority? (Grant & Cadell, 2009). Solution-focused practice however depends on the exercise of therapeutic expertise and authority within collaborative worker-client relationships. It provides a firmer theoretical base than the Kansas strengths approach for the surfacing and management of conflict.   The Strengths-Based Solution-Focussed Child Protection Approach The key theorist in the development of strengths-based solution-focussed child protection practice was Insoo Kim Berg, a therapist at the Brief Therapy Center, the wife of Steve de Shazer and the mentee of John Weakland of the Mental Research Institute. Through her work with statutory child welfare agencies in Michigan State she laid the theoretical groundwork for a solution-focused approach to child welfare (Berg, 1994; Berg & Kelly, 2000). She located the approach within a strengths-based paradigm but made little reference otherwise to the Kansas group.    16   In 1991 Berg and de Shazer travelled to Perth, Australia where they worked with Andrew Turnell, a brief family therapist and social worker, and Stephen Edwards, a child protection worker (Turnell & Edwards, 1999). Turnell and Edwards went on to develop the ?Signs of Safety' approach in collaboration with frontline child protection workers (Turnell, 2012; Turnell & Edwards, 1999; Turnell, Lohrbach, & Curran, 2008). Beyond recommending the literature of the Kansas group ?if the reader wishes to consider a strengths-based perspective? (Turnell & Edwards, 1999, p. 66) their approach cited the solution-focused rather than the strengths tradition. 'Signs of Safety' is now the most prevalent model of strengths-based solution-focused child protection practice, having been implemented in 50 to 100 jurisdictions across Australasia, North America and Europe (Bunn, 2013; Turnell, 2012). In addition to providing principles and tools for the assessment and discussion of risk, family strengths and what is needed to ensure child safety, it has evolved to include tools, like the "Three Houses" and "Words and Pictures" (Turnell, 2012), to engage children and family members in child-focussed discussions.   Recent peer-reviewed literature links two other solution-focussed approaches to child protection work. The first is the Strengths and Skills approach (Corcoran, 2005). This combines ideas from solution focused, cognitive behavioral and motivational interviewing work and claims to be applicable to child welfare practice (Corcoran, Jones, & Ankerstjerne, 2005; Hohman, Kleinpeter, & Loughran, 2005). However I have been unable to find evidence of this approach being implemented in child protection work. The second is Solution-Based Casework (Antle, Barbee, Christensen, & Martin, 2008; Antle et al., 2012; Barbee, Christensen, Antle, Wandersman, & Cahn, 2011; Pipkin, Sterrett, Antle, & Christensen, 2013). Also known as 'Family Solutions', it is based on solution-focused therapy, family life cycle theory and relapse prevention theory. It has been implemented in some statutory child welfare agencies in Kentucky, Tennessee, New Hampshire, Washington State and Florida. While the creators of   17  Solution-Based Casework have written about its implementation and outcomes, they provide little detail about what the model means for the relationship between worker and client. This question is most comprehensively answered in works authored or co-authored by Berg or Turnell.   A common theme of strengths-based solution-focussed child protection approaches is that the ideas of partnership and client self-determination inherent to SBP are limited by the need for the worker to assess risk, make judgments and exercise the authority of the mandated role (Turnell, 2004). The worker?s role is to help the client expand his perception of available options by systematically assessing and reinforcing his strengths and supporting the client to utilize internal and external resources to work towards self-identified goals (Berg & Kelly, 2000). At the same time, the worker conducts ?forensic, rigorous professional inquiry? (Turnell, 2010, p. 21) into the situation of children perceived to be at risk of abuse or neglect. Child safety is always an explicit goal of the interaction with parents. The worker seeks to support the right and capacity of parents to make decisions that promote child safety, while simultaneously carrying the statutory power to limit those parents? rights within systems typically characterized as authoritarian (K. Healy & Darlington, 2009). As Turnell comments,  The Signs of Safety approach does not set problems in opposition to a strengths and solution-focus, nor does it frame forensic, rigorous professional inquiry as something that diminishes or erases the possibility of collaborative practice. Quite simply, the best child protection practice is always both forensic and collaborative. (2010, p. 21)   Common principles of strengths-based solution-focussed models developed explicitly for child welfare work are:    18  1) The client is a respected partner in creating safety for the child The basis of the approach is a collaborative worker-client relationship in which the client is respected as a fully participating partner at all stages of the case. The client is seen as an expert on his own life, who, if given the opportunity, will lead the worker towards solutions. The worker?s job is to elicit, understand and validate the client?s perspective. The worker should accept that this perspective is likely to be different from her own and be open to ideas that are not her own. The client should always have choices. Key to this approach is the worker suspending their own assumptions and adopting a ?not-knowing? stance.  2) Every family has strengths on which the solutions to child welfare problem can be built It is assumed that parents wish to parent well, and can do so with sufficient resources and support. The focus of the work is on identifying and developing the strengths and solutions that lie within every client. This means considerable time is spent exploring the details of exceptions to the problematic behavior.  3) The goal of the work is the future safety of the child and this requires assessment of risk The work is present and future-oriented. The guiding question is ?what can be done now to increase child safety in the future?? The details of past abuse and neglect are important but only in so far as they link to the child?s future safety. This means that workers do not need a complete chronology of past problems and do not need the parents to admit culpability for past abuse or neglect.   19  4) The work is driven by the motivation of clients to achieve their own goals It is assumed that the client will be motivated to work towards his own goals. The worker finds out what the client wants and supports him to work towards these goals. The client?s goals most often directly increase the child?s safety. Sometimes the worker may link support for client goals with the goal of safety. It is assumed that if nothing else, the client?s wish to be free of statutory protection services provides the common goal on which a working partnership can be built. Goals are small, achievable and related to specific new behaviors described in detail. Small changes in thinking or behavior can trigger significant personal and systemic change.  5) The worker and client together co-construct solutions and motivation through  relationship and language Motivation is not an individual trait, but a co-constructed product of the worker-client interaction. The way a problem is discussed opens or closes the door to different solutions. Workers can create an expectation of change by framing problems as temporary, solutions as accessible and describing clients in terms of their capacities. The worker uses the client?s language and every interaction with the client is a potentially powerful intervention.   6) Coercion and partnership are not mutually exclusive  The worker possesses significant power and must exercise his statutory authority if the child?s safety cannot be secured without it. However the use of coercive force can co-exist with successful and ongoing efforts to construct a genuine partnership informed by both the client?s and worker?s perspectives.     20  7) The worker strategically uses the tools of solution-focused therapy in the interaction with the client More than a perspective, standpoint or set of principles, the approach requires the use of specific therapeutic tools to inform the casework process. These can include compliments, exception questions, scaling questions, coping questions and the miracle question.  The Strengths-Based Solution-Focused Child Protection Relationship Strengths-based solution-focused child protection approaches envisage the worker-client relationship to be ?the principle vehicle for change? (Turnell & Edwards, 1999, p. 47). It is through the relationship that the client develops a sense of validation, personal competence and trust in the helping process and the worker elicits the best information. It is through the relationship that the client and worker co-construct solutions and the motivation to achieve them. This puts a premium on preserving a positive cooperative alliance.   The worker is responsible for establishing a positive relationship based on the principle that clients are ?people worth doing business with? (Turnell & Edwards, 1999, p. 30). By adopting a ?non-knowing posture? the worker steps out of her expert role and communicates curiosity and ?genuine awe and respect? (Berg & Kelly, 2000, p. 98) for the client?s perspective.  The client is perceived to have considerable powers of self-determination: ?the strengths approach is built upon a simple belief ? the power is in the hands of the clients? (Calder, 2008b, p. 138). The worker is required to respect the client?s right to feel the consequences of rational decisions and freely-exercised will: In interviews we discipline ourselves to ask questions with no investment in client outcomes. This posture does not represent an uncaring attitude toward clients or the   21  welfare of the community but an acceptance of the reality that we are working with human beings who make choices. (De Jong & Berg, 2001, p. 372)   This is a purposeful goal-centred relationship for which the worker hypothesizes future goals before he and the client meet, and the client?s goals are used as ?leverage? for cooperation (Turnell & Edwards, 1999). The worker wields a great deal of power, guiding client interviews with structured and persistent communication and therapeutic techniques (Turnell & Edwards, 1999). She responds to client anger by reframing it and uses cognitive-behavioural strategies of rehearsal and reinforcement. She acknowledges the client?s anxiety and fears. As Berg and Kelly (2000) explain, "What we can do is to influence clients in such a way that they believe it is in their desire and in their best interest to change. This is all done with talking? (p. 80). Although the client is encouraged to share his perspective, pragmatism dictates that wholehearted acceptance of the client?s perspective be limited by its usefulness to the goal of behavioral change:  Ask the client to explain the events as she sees them, without correcting her or arguing with her, however outrageous her story seems. Then ? much later ? you follow her logic and push it to an extreme, it will come to sound pretty incredible to her also. (Berg & Kelly, 2000, p. 88) The worker is honest and open in the discussion of goals, although not about her theoretical positioning and use of clinical techniques.  Workers gather information, assess risk and exercise their authority judiciously. They are advised that ?making your requests in a calm, quiet firm voice makes them difficult for clients to resist? (Berg & Kelly, 2000). Attentive listening and encouragement for client self-determination   22  is balanced by clarity about the child protection concerns and honesty about mandated authority (Turnell & Edwards, 1999). The worker is explicit about what is and is not negotiable and straightforward about consequences and expectations. To maximize the therapeutic potential to work through the client?s interpretations, the worker holds her own in abeyance for as long as possible. Workers have to ?hold at least five different stories in their head at one time?  (Turnell & Essex, 2006, p. 38) and carry judgments ?lightly? on the constructivist assumption that all knowledge is incomplete. This position demands continual movement between expressing empathy for the client and establishing expectations of the client (Turnell et al., 2008) as the worker navigates a relationship of mutuality in which sufficient separation is maintained to enable judgments to be made in the best interests of the child.   The relationship relies, however, as much on the worker?s warmth, spontaneity and openness as her skills in eliciting and working with the client?s position. When the worker demonstrates genuine empathy, caring and curiosity, the social distance between worker and client is reduced and the client develops the sense of reciprocity and trust necessary for engagement  (Forrester, Kershaw, Moss, & Hughes, 2008). The client needs significant and continual encouragement to avoid discouragement (Antle et al., 2012). In order to meet this need, workers are advised to ?maintain humour, hope and gratitude. Do not take yourself too seriously? (Berg & Kelly, 2000, p.73). Turnell et al (2008) tell the story of a social worker confronted with a screaming swearing one-eared father during an investigation and, after meeting his demand that she ?f*** off? with the request to know when she can ?f*** back? proceeds to say "I couldn`t help notice the fact that you hardly have a left ear, and if you don`t tell me how that happened I don`t think I?ll be able to concentrate on what we have to talk about"   23  (p. 105). Even while strategically guiding the relationship for therapeutic ends, the whole worker, not just the professional persona, is required to show up.  The Implementation of Strengths-Based Solution-Focused Child Protection  There has been a general movement in social work over the last two decades to adopt strengths-based approaches. There have been, however, few attempts to interrogate theoretical adherence to strengths-based models and what practitioners are actually doing in their name (Lietz, 2011; C. Rapp et al., 2006; Staudt et al., 2001). There is a concern that the principles of strengths-based practice are not being adopted by social workers (Blundo, 2001, 2012; Grant & Cadell, 2009; Roose et al., 2012). Two studies into the experiences of families in the child welfare system found that workers were at best inconsistent in their application of a strengths approach (Lietz, 2011; Roose et al., 2012).   There is some evidence of the implementation of the Signs of Safety approach. Turnell (2012) reported that,  there are currently nearly 100 jurisdictions in 12 countries undertaking some type of substantive implementation of the Signs of Safety . . . Beyond Western Australia the most substantial system-wide implementations are occurring (or have occurred) in?Minnesota counties; Gateshead Children?s Services Authority, England; Bureau Jeugdzorg in Drenthe, The Netherlands; Open Home Foundation, New Zealand; all Copenhagen boroughs in Denmark; Ktunaxa Kinbasket Child and Family Services, British Columbia, Metis Child Family and Community Services, Manitoba, Canada, Saitama City, Japan. (p. 6)    24  Bunn more conservatively estimated that in 2012 Signs of Safety was being used in at least 50 jurisdictions (Bunn, 2013). That same year 11 states in the United States used the Signs of Safety approach, three as their sole model for practice and eight in conjunction with the Structured Decision-Making risk assessment model  (Harbert & Tucker-Tatlow, 2012). Some jurisdictions had developed hybrid practice models that drew on the ideas of Signs of Safety but no longer went primarily by the Signs of Safety name. California's 'Safety-Organized Practice', for instance, re-articulated the Signs of Safety in combination with Structured Decision-Making and strategies like motivational interviewing, family meetings and group supervision (Harbert & Tucker-Tatlow, 2012).   In Minnesota the Signs of Safety approach was first introduced in Olmstead County in 1999 and became part of statewide training in 2009 (Idzelis Rothe et al., 2013; Skrypek et al., 2010). An independent statewide implementation study found inconsistent implementation of the model, with workers deterred by a lack of trust in their organisation's commitment to the model, the efficacy of the approach and the capacity of families (Skrypek et al., 2010). It found that  those who were earliest along in Signs of Safety implementation were more likely to rate themselves as further along in their understanding and integration of the model than those counties who had more experience and exposure to Signs of Safety. This may be attributed to the fact that while the Signs of Safety tools are relatively simple and straight-forward, it is using them in practice that results in the real learning and understanding of the model. Individuals who have been practicing Signs of Safety for a longer period of time are more likely to recognize the complexity of the approach and the challenges of fully integrating it into all aspects of their practice. (p. 2)     25   A later study involved interviews with 24 parents of children whose child welfare cases had recently been closed (Skrypek et al., 2012). It concluded that there was evidence that social workers were using the model and that the majority of parents were positive about the approach. However only 42 of the 100 parents initially approached to participate in the study consented to be contacted by researchers and this first approach was often made by the parent?s worker. It is likely that those most positive about their experience participated in the study and so while it appears that some workers were using the Signs of Safety approach to good effect, it is not clear how typical they were. A third evaluation report a year later showed that since the introduction of the Signs of Safety to Olmstead and Carver Counties, there had been improvements in such outcomes as the rate of entry to care and the recidivism rate (Idzelis Rothe et al., 2013). The implementation of the approach shortly after radical changes in service delivery like the introduction of differential response and Structured Decision Making made it hard, however, to isolate the cause of such improvements.    A study of the implementation of Signs of Safety in statutory child welfare services was completed in Andrew Turnell?s home ground of Western Australia in 2010 (Department for Child Protection, 2010). Eighty-eight percent of respondents found the approach useful or very useful, and the majority reported that positive effects on clients included enabling greater participation, voice and collaboration, clearer goals and a better understanding of the impact of harm. Two thirds of workers reported that it had increased their job satisfaction.  However these results were based on a 17% response rate, with only 251 responses to invitations sent to 1460 staff, and again it is not clear how representative they are.  On his website Andrew Turnell reports that 380 workers were trained in the Signs of Safety approach between 2005 and 2008 in the Borough of Copenhagen Child and Family   26  Services. He reports that an independent evaluation of the project (Sorenson, 2009) was conducted in which 171 child welfare practitioners were interviewed. It found that over 70% of interviewees said Signs of Safety had changed the way they worked with families and increased their focus on family resources. Sixty nine percent reported that they used the approach with families. Just over half said it had increased their inclusion of family strategies and solutions and just under half said they now gave families more responsibility. The study is only in Danish and I have unable to independently assess its merits.    Much of the remaining literature on the implementation of the Signs of Safety approach is small-scale and descriptive (Bunn, 2013) and while offering case examples of workers using the approach, offers little insight into how typical they are. The majority were written by people intimately involved as workplace leaders or paid 'Signs of Safety consultants' in the introduction of the approach (Hogg & Wheeler, 2004; Lohrbach et al., 2005; Shennan, 2006; Turnell et al., 2008; Wheeler & Hogg, 2012). There may be some reason to be skeptical of claims by those  who stand to gain commercially from the success of the Signs of Safety approach. An added challenge when evaluating the evidence base for this approach is that Turnell advocates an implementation and evaluation approach based on the principles of Appreciative Inquiry (Turnell, 2012). While there may be an acknowledgement of challenges, studies informed by Appreciative Inquiry have focussed on what works and cases that practitioners are proud of (Keddell, 2012; Shennan, 2006). This perspective may obscure the voices of those for whom the approach does not work and limit the discussion of problems encountered with the approach.   Finally, the Solution-Based Casework model was developed with frontline child protection workers in Kentucky and has since been implemented in that state, Washington State and some parts of Florida, New Hampshire and Tennessee (Barbee et al., 2011). Early research   27  found that workers who had been trained in the Solution-Based Casework principles and skills demonstrated them inconsistently (Martin, Barbee, Antle, & Sar, 2002). A study asking whether child welfare workers could implement the model concluded that when compared to a team in which the supervisor received one day of training in the model, a team in which all members received 5 days of training and 24 monthly consultations not only used the model significantly more, they also achieved greater cooperation from families, achieved more goals and saw fewer child removals and referrals to court (Antle et al., 2008). Later interviews with 12 public child welfare workers ?identified challenges of the shift from a pathology-orientation to a solution-focused and strengths-based perspective, the importance of supervisory support, and the struggle to understand complex elements of the model? (Antle et al., 2012, p. 344). It concluded that in a large sample of cases workers were implementing the model with a sufficiently high degree of fidelity to result in lower recidivism rates and to produce better outcomes in child safety, permanency and well-being when compared to cases in which there was lower adherence to the model (Antle et al., 2012). However this research program did not interrogate whether such different outcomes were due to the use of the model itself or other factors.    Historical Changes in the Child Protection Worker-Client Relationship  One question motivating this study was the extent to which it was reasonable to expect frontline workers to engage in such apparently complex client relationships under the challenging conditions of contemporary child protection practice. Context matters: since the inception of the modern child protection movement the worker-client relationship has been shaped by broader social, economic and political forces. Changes in the conceptualisation of this relationship have been closely linked to changes in available resources and the struggle to define the social work   28  profession and the place of child protection within it. The following section traces these changes over the course of the development of contemporary statutory child protection systems in the United States, Canada, the United Kingdom and Australia.   It is important to note that there has been no linear temporal progression from one conceptualization of the worker-client relationship to the next. Relational models have coexisted and informed each other. The worker-client interaction is influenced by the personal characteristics and immediate environment of worker and client as well as broader contextual factors (Ross, Polaschek, & Ward, 2008). It should also be noted that there have been differences in the culture and forms of service delivery, and in the timing, nature and impact of child welfare changes across the countries discussed in this section. The national child welfare systems in these countries are, however, very similar (Gilbert, 1997; R. Hetherington, 2002) and it is possible to trace shared historical trends in the dominant understanding of the relationship between worker and adult client.   There is plentiful evidence in the social work literature from the United States and United Kingdom that those countries experienced the broad historical trends outlined in this review. Australia (Fogarty, 2008) and Canada (Jennissen & Lundy, 2011) have been described as following a similar historical path. Each has been informed by a liberal political philosophy (R. Hetherington, 2002) predicated on a reluctance to intervene into the private realm of the family unless absolutely necessary (Dingwall, Eekelaar, & Murray, 1995). In each of these countries strengths-based solution-focused practice has gained ground since the turn of this century. In each this has been presented as a response to the problems of a deficit-based investigative approach and the ineffectiveness of the child welfare system. The majority of the literature on these subjects has been written in and about the United States, Canada, Australia and United   29  Kingdom. It has focussed on the mainstream child welfare systems in these countries, often excluding the voices of non-dominant cultural groups (Jimenez, 2006; Trocm?, Knoke, & Blackstock, 2004) and fathers (Strega et al., 2008). Unless otherwise stated the claims outlined in this section do not extend beyond these countries.   1870 - 1930 The roots of the modern child welfare system are often traced to responses to the industrialization which swept across North America, Australia and many parts of Europe in the 19th century. By 1900 mass immigration to industrial centres had overwhelmed urban infrastructure and prompted social dislocation (Myers, 2004). About one quarter of the British population lived in poverty and in the poorest areas the same proportion of babies died before 12 months of age (Rowntree, 1901). Overcrowded unventilated slums bred filth and disease. Women and children worked long hours to supplement low family wages and ?baby-farms? offering cheap childcare flourished (Zelizer, 1988). A patchwork of private, voluntary and religious organizations offered services to support families, but the primary response to the problem of abandoned and abused children was to consign them to the poor house, orphanage, district schools or indentured servitude (Hendrick, 1994; Katz, 1986).   Until the 1870?s no agency was responsible for enforcing the laws against child cruelty and for finding, investigating and rescuing children from situations of abuse (Costin, 1991). However this changed after the 1874 American case of Mary Ellen Wilson crystallized concern about a system of child-serving agencies which, despite considerable dependence on government subsidy, was largely unaccountable to the tax-payer (Costin, 1991). Acting at the request of a missionary worker, the director of the New York Society for the Prevention of Cruelty to   30  Animals turned to the courts to protect nine year old Mary Ellen from the violent abuse of her stepmother. The case caught the attention of the newspapers and Mary Ellen became the poster child for ?child rescue? campaigners. Supported by a judiciary eager to expand into new territory, the influential men of the animal protection movement financed and led the first Society for the Prevention of Cruelty to Children (SPCC) in New York. That 'child-saving' should become attached to the cause of animal protection made sense in light of the influence of Darwinism and progress in the science of anaesthesia (Costin, 1991). These contributed to a cultural rejection of suffering and redefinition of children as vulnerable, in need of protection and only one small step above animals on the evolutionary ladder (Costin, 1991; Katz, 1986).  By 1900 there were 157 private and voluntary societies in the US with responsibility for the rescue of abused and neglected children (Antler & Antler, 1979). The UK saw its first SPCC in 1884 and by the end of the century handled over 28,000 cases a year (Ferguson, 1996). In Australia the Victoria SPCC opened in 1896 (Fogarty, 2008). In Canada the first Society opened in Toronto in 1881 and the following year the Nova Scotia Society for the Prevention of Cruelty to Animals assumed guardianship responsibilities for neglected children (Jennissen & Lundy, 2011).    The early SPCCs operated as, the arm of the law (which) seizes the child when it is in an atmosphere of impurity, or in the care of those who are not fit to be entrusted with it, wrenches the child out of those surroundings, brings it to the court and submits it to the decision of the court ? unless, on the other hand, its reaches out that arm of the law to the cruellist. (McCrea, 1910, pp. 138-139)   31   In the UK concerns for their public legitimacy made SPCC?s focus more on educating parents in their duties rather than removal and prosecution (Hendrick, 2005). Common across SPCC?s, however, was the belief that their role properly lay with the enforcement of laws and parental responsibilities rather than the provision of charity.   The NSPCC  ?inspectors? of the early twentieth century were men drawn from the ranks of police and army, ex-truant officers and poor law officials (Carstens, 1921). Their main tool was their authority, reinforced by close relationships with the police. They carried warrant cards, often wore uniforms and could be contacted through local police stations (Clapton, 2009). Their primary role was evidence gathering for criminal court and their involvement ended once a child was removed from the home and the case transferred to a child-serving charity (Antler & Antler, 1979). They had little training in, or wish to undertake, preventative work.  These early protective agencies were ?dominated by discipline rather than by discretion on the part of the workers, by the following of definite legal rule or procedure rather than by the looser adaptation of means to ends in individual cases? (McRae, 1910, p. 139). They did give some material assistance in the form of food and clothing, performed wider duties like returning runaways to their parents and securing work permits for children (Myers, 2004) and, at least in Scotland, they often engaged in kind acts which belied their stern public reputation as the ?Cruelty Men? (Clapton, 2009). However, they worked more through advice, threats and referral to the courts than supportive empathetic relationships: The work has to be done in cold blood, so to speak, with a deaf ear to the pleadings and entreaties of mothers and fathers whose love for their children is never so strong as when they think their children will be taken from them. The best interests of the children, and not the feelings of those closely connected to them, have to be   32  considered, and a trained charity worker is seldom qualified by his or her training to absolutely banish the spirit of ?love for fellow-man? from his or her work. The SPCC worker, on the other hand, looks at the matter from a colder (a legal) viewpoint, and therefore the only training he should receive is such that will enable him to detect cruelty and apply the remedy, regardless of how the application affects those who are responsible for the cruelty. (McCrea, 1910, p. 135) By the first decade of twentieth century the Cruelty Men?s powers of surveillance and child removal were so ?wholesomely feared by the evil-doer? (Carstens, 1921, p. 136) that many parents avoided contact with them altogether (Ferguson, 1996).   There were two alternative visions for child welfare work in 1900, and together they provided the foundations on which the new profession of social work was built. They were driven by women and informed by social reform movements for women?s suffrage, temperance, social purity and action to mitigate the worst excesses of industrialization (Costin, 1991). The first vision was promoted by the Settlement House movement, which started in 1884 in the United Kingdom and was embodied by the American work of Jane Addams (1860 -1935). She became the first female president of the National Conference of Social Work in 1909. She believed that protecting children entailed advocating for legislative reforms against child labour and prostitution, building community networks of social and material support for families (D. L. Franklin, 1986) and lobbying government and charities for ?support for deserted women, insurance for bewildered widows, damages for injured operators, furniture from the clutches of the installment store? (Addams, 1960, p. 8). Partly due to her political partisanship, anti-professionalisation stance and pacificism during World   33  War 1, however, her approach to child welfare failed to garner widespread political acceptance (D. L. Franklin, 1986).   The Settlement House workers were encouraged to make relationships of equality, mutuality and camaraderie with their adult clients (Addams, 1960). Addams sought to reduce the social distance between herself and the community by refusing to accept a salary and fighting attempts to professionalise social work. She lived in a house in the Chicago slums, installed three baths so her neighbours could wash there, and invited elderly women to stay with her for holidays from the Poorhouse. Community members joined her as residents and hosted community reading, music and debating groups and sports, dancing and cultural events. She called her clients ?neighbours? and would only give public lectures in their presence to prevent any sense that she was claiming more expertise than they (Addams, 1960). She believed that living side by side with those in need cultivated the intimate community knowledge and mutual compassion and respect needed for collective action.   Addams believed that the middle class women who learnt social work at the Settlement Houses gained as much as their clients from the cross-class relationships which allowed them to escape their ?undernourished, over sensitive lives? (Addams, 1960, p. 12). All benefited from the exercise of the natural instinct for relatedness and mutual support. The women?s movement had generated some sense of gender solidarity and Addams understood democratic action as the means to perform her Social Christian duty to relieve human suffering (Phillips, 1996). However these relationships were supposed to be driven by a sense of authentic affection and admiration for clients rather than by obligation.   An alternative conceptualisation of child welfare work was offered by the "friendly visitors" of the Charitable Organisation Societies (COS). In the United States this work was   34  exemplified by Mary Richmond (1861-1928), who went on to found the first social work School of Civics and Philanthropy in 1907 (D. L. Franklin, 1986). Like Addams, she believed children were best protected through work with the whole family, but her means was not political activism but what came to be known as casework. Friendly visitors were to employ a combination of ?education, advice, persuasion and assistance? (Carstens, 1930, p. 1). They taught parenting and home-management skills, advocated with schools and community agencies for appropriate supports for those deemed to be deserving and gave the kind of direct help reflected in case notes like ??Visited family, left 5?Took 1 to employment agency. Got job?. ?Took Mary to the hospital. Had tooth pulled. Sent Christmas basket?? (Robinson, 1930, p. 110). Influenced by the scientific philanthropy movement (Bremner, 1956), this kind of casework was increasingly dependent on ?social diagnosis? (Richmond, 1917).  This was the detailed and systematic assessment of the family situation intended to individualise intervention and deliver charity more rationally and efficiently.   Mary Richmond related to her clients more like a strict devoted aunt than Addam?s comrade-in-arms. In her early writing Richmond described workers developing relationships with clients over the course of five or six years, through which they gained ?intimate and continuous knowledge of and sympathy with a poor family?s joys, sorrows, opinions, feelings and entire outlook on life? (Richmond, 1899, p. 180). Friendly visiting was ?intensely personal work? (Richmond, 1899, p. 193), demanding great commitment and a dogged perseverance. The friendly visitor might take the children to her home in the country, or invite a mother to her kitchens for cooking lessons, or spend evenings teaching parlour games to the entire family. She was expected to bring humour and energy to the task, as well as ?gifts of cooking, growing plants, pictures and simple decorations? (Richmond, 1899, p. 139).   35   While the friendly visitor had the power to arrange a child?s removal, her prime tool was intimacy, by which she might translate her moral and social superiority into the influence to lift a family out of dependence. ?They must get it just as our friends get an influence over us, by long, patient contact and by the slow, natural growth of friendship? (Richmond, 1899, p. 183). Relationships were characterised by a 'tough love' transparency. Once intimacy had been established the visitor told clients why relief was being withheld from them or that they had a dirty home, in order to raise their expectations and prompt their striving for a physically and morally cleaner life. She did so with energy and humour as ?poor people are no fonder of dismal folk than the rest of us? (Richmond, 1899, p. 129).   Richmond described her clients with great warmth. She implored friendly visitors to abandon their prejudices and to see the poor as inherently worthy of compassion and respect. However, her protestant and liberal individualist values led her to emphasise the moral causes of dependence (D. L. Franklin, 1986). An important task for the case worker was to root out the 'undeserving' poor. She argued that careless charity undermined client self-determination and resourcefulness: "It helps a man to know that someone cares and will help him to find work; but it cripples him to let him feel that he can sit idle and let his friend do all the searching and worrying" (Richmond, 1899, p. 41). To avoid the moral pitfalls of worker over-indulgence and client dependence a  peculiar kind of reciprocity was required: ?Would it not be well if, instead of always giving sympathy, we sometimes asked for it?Such mutual relationship broadens their meagre lives, and makes our contact with them more human? (Richmond, 1899, p. 185).   During World War 1 the casework model of child welfare won out over the alternative police and community development models (Ferguson, 1996; Jennissen & Lundy, 2011; Shoshani, 1984). SPCCs slowly expanded to offer a range of protective services beyond simple   36  court intervention. Their broadening remit included such concerns as ?beating or other physical cruelty?children begging and accompanying  an  organ  grinder?children  sent  out  by  parent  or  guardian to beg, attempted assault?abandonment  of  child, need  of  medical care, child found intoxicated, child living in immoral resort? (Carstens, 1921, p. 138). In 1909 the White House Conference on Children concluded that children should not be removed from their homes for reasons of  poverty (Myers, 2004), signalling a shift in the official goal of child welfare work from removal to rehabilitation (Hendrick, 2005; Myers, 2004). By the 1920s protective services began to be situated within preventative child and family agencies and protection work was viewed as part and parcel of general child welfare practice (Antler & Antler, 1979; Hendrick, 2005; Jennissen & Lundy, 2011).    Child welfare was increasingly seen as a matter for the professional expertise, assessment and intervention embodied by Richmond?s social diagnosis approach. This was partly fuelled by a positivist faith in the potential of ?a treatment of behavior so scientific that results instead of being accidental will be subject to intention and prediction? (Taft, 1922, p. 3). It was partly political, as the COS gained from their active support for the war effort and collaboration with the Red Cross in meeting the needs of returning solders and their families. Richmond had close links with the medical profession; her address to the National Conference of Charities and Corrections (NCCC) in 1912 was entitled ?Medical and Social Cooperation? and many medical students volunteered in the COS (D. L. Franklin, 1986). In the UK too doctors and NSPCC workers often conducted home visits together and the case files of both contained increasingly sophisticated medical tools like percentile charts (Ferguson, 1996). The social diagnosis approach used medical language and both it and the female-dominated discipline of social work   37  gained professional legitimacy from collaboration with an influential male-dominated medical profession which was rapidly expanding its reach.  As demand rose for assessments of eligibility and social situation, taking years to establish intimate knowledge of a family became impractical. In any case, the COS policy of ?not alms but a friend? (Richmond, 1899) meant relationships with clients were often far from friendly. Systematic inquiry replaced long relationships as the basis for individualized service. The worker gathered facts, digging beneath the client?s story by interviewing family, employers, neighbours and analyzing documents to construct ?as exact a definition as possible of the social situation and personality of a given client? (Richmond, 1917, p. 51). The worker-client relationship became more goal-directed and the worker claimed a professional expertise which increased her emotional and social distance from the client.  This was a time when ?workers were universally thought to be wiser and better informed than clients, advice was one of the visitor?s stocks in trade? (Woods & Hollis, 1964/1990, p. 116). While it remained important to be ?sincere and direct and open-minded? (Richmond, 1917, p. 200),  friendship was no longer the goal except as a means to secure useful information and exert influence. The tension between care and control, support and authority, was never far from the surface in this casework relationship and the pull of child welfare?s policing function remained strong: Such words as ?rescue?, ?prosecute? ?investigate? ?evidence? ?compel? are found in abundance, while ?assist? ?persuade?, ?development of resources?, ?treatment? are almost wholly absent?Historically child protection is a police, law-enforcing movement. Today we practice some, and talk more, of case work methods. But the heritage of the past is a heavy burden and we have not freed ourselves from it. (Falconer, 1931, p. 2)   38  1930 - 1960  The Wall Street Crash of 1929 marked the beginning of a decade of depression for all western industrialized nations. Many privately funded agencies disappeared and caseloads in those that remained increased rapidly (Hendrick, 2005; Myers, 2004). Roosevelt?s New Deal (1931) and Social Security Act (1935) in the United Sates, and creation of children?s departments following the Beveridge Report (1942) in the United Kingdom, signalled a trend towards the state taking responsibility for filling gaps in family services. However many places, particularly in rural areas, remained without child protection services until the 1960?s, leaving police and probation officers to operate a residual service along more traditional lines (Meyers, 2008).   Social workers found themselves with few resources with which to tackle the intractable problems of poverty and mass unemployment. They turned instead to the task of supporting individual adjustment through the resolution of intra-psychic difficulties. Clients of protective agencies were increasingly seen as needing skilled therapeutic casework. Since the end of the 19th century psychiatry had extended its professional reach beyond institutional control of the 'insane' to 'mental hygiene' (Alexander, 1972; Shoshani, 1984) and social problems like delinquency had been reframed as psychological problems requiring research and treatment rather than punishment or moral advice (W. Healy, 1915). These ideas had been passed on to social workers through close collaboration with psychiatrists during the war and in the child guidance clinics of the 1920s (Shoshani, 1984). By the time the crises of depression and the second world war had given way to relative prosperity in the 1950?s, the social work profession was heavily committed to psychological theories that assumed ?the environment was a given ?reality?, ?normality? or whatever, and that the client was by definition ?maladjusted?, or ?abnormal? in some way, and required ?treatment?" (Jordan, 1987, p. 25). Workers were keen to   39  take advantage of the new clinical opportunities offered by the creation of the welfare state and greater acceptance of public intervention into private life. The rise of fascism and advent of the Cold War also generated a climate of conservatism which supported the focus on individual pathology.    From around 1930 the therapeutic nature of the worker-client relationship became increasingly important. The psychodynamic theories of Freud, and the later ego theories of his daughter Anna and wayward prodigy Rank, attributed social dysfunction to undeveloped id and ego functions which could be modified through the therapeutic relationship. This therapeutic relationship became seen as the primary means by which workers understood the client?s internal and external worlds and clients named and changed intrapsychic patterns (Biestek, 1957; Hamilton, 1951). It became a common perception that through the therapeutic relationship the social worker ?opens doors and windows to let in air, light and sunshine, so that the client can breathe more easily and see more clearly? (Biestek, 1957, p. 106).  This period saw the emergence of two distinct approaches to generalist social work. The Diagnostic school inherited Richmond?s focus on the interconnection of person and situation and the need for clear assessment. Treatment included environmental modification, psychological support, clarification and insight development (Woods & Hollis, 1964/1990). Clients of diagnostic social workers were still asked about their social histories, and the detail of early childhood, feeding and weaning experiences became more important (Hamilton, 1949). However, the client?s ?own statement of his own situation, his spontaneous expression of difficulties and wants ?replaced? the tendency to try to extract from the patient by various devices the facts that we needed to have? (Hamilton, 1933, p. 517). This approach was informed by an understanding of the dynamic multi-causal nature of experience and interaction of subject   40  and object (Hamilton, 1941). The transference relationship was now recognized as a source of rich information and ?while one can say that caseworkers do not focus on intrapsychic unconscious conflict as such, they must understand and be able to handle it as it shows itself in attitudes, behavior, adaptive and defensive structures? (Hamilton, 1949, p. 214).  Functionalist social workers saw their role as being to present what the agency could offer and allow the client to take full responsibility for his response. The social worker?s job was to create a boundaried therapeutic space, in which clients could decide ?what use will he make of the worker?? (Robinson, 1930, p. 136). Past history mattered only if the client brought it into the present relationship. It was through the relationship that the client discovered his will, both in opposition to the worker and through ?the sense of security and protection which he derives from the relationship and?his tendency to identify with the worker and take over attitudes and interests which the workers suggests? (Robinson, 1930, p. 132). The helping relationship was the space in which the client struggled with conflicting lifelong desires to individuate and merge. Functionalist social workers saw this struggle as the means for the client to realize his innate capacity for self-determination and self-creation (Robinson, 1930; Smalley, 1967).   The psychotherapist Carl Rogers (1902-1987) developed these ideas into the ?client-centred? model for helping relationships which was to become a staple of social work education. He shared the functionalists? faith that ?if I can provide a certain type of relationship, the other person will discover within himself the capacity to use that relationship for growth, and change and personal development will occur? (Rogers, 1951, p. 33). What kind of relationship best fostered psychic transformation and compensated for deficits of empathy and attunement in past relationships was the subject of an explosion of writing in the 1940s and 1950s by objects relations theorists (Klein, 1959; Sullivan, 1953/2003). Later the reparative therapeutic   41  relationship was to be described as a ?holding environment? (Winnicott, 1965) in which the client received from the therapist the nurturing care, empathy and attunement which compensated for deficits in previous relationships and enabled the development of a healthy self (Mandin, 2007).    Common to diagnostic and functionalist approaches was the worker?s wholehearted ?acceptance of the other person as he is ? in whatever situation, no matter how unpleasant, or uncongenial, with whatever behavior, aggressiveness, hostility, dependency, or lack of frankness he may manifest? (Hamilton, 1951, p. 52). Rogers described this as ?unconditional positive regard?, and ?caring for the client as a separate person, with permission to have his own feelings, his own experiences? (Rogers, 1957, p. 98). This acceptance enabled the worker to see the client as an individual. It created the safe space in which the client could relax psychological defences so that ?as the caseworker accepts him, the client begins to accept himself? (Biestek, 1957, p. 76). Acceptance was achieved through ?a constant search for deeper meanings which the client may be struggling to express, rather than a passive toleration of the attitudes he may assert of the surface? (Robinson, 1930, p. 124). It demanded a non-judgmental attitude, which implied in social work not the absence of judgment altogether, but rather the refusal to allocate blame or disapprove of social conduct (Biestek, 1957).    The worker offered ?warm human connection? (Robinson, 1930, p. 100), intimacy and a caring that some even described as love (Biestek, 1957; Robinson, 1930). Most importantly he offered a deep empathy for the client?s experience: The social worker must be a person of genuine warmth with a gift for intimacy. He must be willing to enter into the feeling experience of another, willing to listen to the   42  other?s view of his problem and willing to go patiently along with him in his struggles for a solution. (Hamilton, 1951, p. 28) This required the worker to be ?freely and deeply himself? (Rogers, 1957, p. 95) in order to truly listen and go beyond mere understanding to feeling with the client. The emotional identification however, was limited by the need to ensure ?the worker would not be personally lost in that experience, be overwhelmed, depressed or elated by it? (Robinson, 1930, p. 170). The worker had to maintain the self-awareness to ?sense the client?s private world as if it were your own, but without ever losing the ?as if? quality? (Rogers, 1957, p. 99).   The worker?s emotional connection with the client was controlled by the therapeutic purpose of the relationship: ?Professional relationships are not just friendly associations. Contact is not for the sake of contact?The professional self is controlled towards the end one is serving ? namely to understand and meet the psychosocial needs of client? (Hamilton, 1951, p. 29). The worker was neither a friend nor an equal and there was little room for spontaneity or reciprocity in these carefully crafted relationships where the worker?s intervention ?must be a selective response, which is guided by purpose, including the over-all purpose of this particular interview, and the immediate purpose of the response? (Biestek, 1957, p. 59). Freudian and ego psychology supported workers to set boundaries and to maintain a sense of detachment to allow full expression to the client and enable the identification and management of transference.  Diagnostic social workers retained a clear authority to diagnose and treat, albeit with client input (Hamilton, 1951). Functional workers used time limits, and agency procedures to set the boundaries which motivated change and protected against excessive dependence (Smalley, 1967). Their role was to clearly present what the agency could offer, and allow the client to take full responsibility for his response. All social workers needed ?fairness and scrupulous honesty   43  with himself, as well as with the client. The worker should not connive with the delinquent, no more than he should be pulled into the client?s neurosis? (Hamilton, 1951, p. 289).  Self awareness and self-management were essential.    The transition to these more therapeutic relationships happened as ?a painful, slow, creaky process of wresting workers away from longstanding habits of practice, rather than as a sudden ?deluge?? (M. H. Field, 1980, p. 399). Even within preventative family agencies evidence of a more therapeutic approach varied from worker to worker well into the 1940s (Woods & Hollis, 1964/1990). Intensive therapeutic work was often referred out, leaving protection workers with more of a monitoring and advocacy role (Skehill, 2010). It may be true that ?the  average social worker, whatever her setting, usually confronted a caseload too large to permit the differential social diagnosis envisioned by Mary Richmond, let alone the intensive examination and therapy symbolized by the child guidance movement? (Lubove & America, 1965, p. 104).   By the 1950s social work journals were beginning to discuss a new casework approach for involuntary clients. The courts were perceived to be the appropriate venue to manage the ?small proportion of persons? so resistive to guidance, where neglect and pathology are gross? (Hamilton, 1951, p. 290), but until this time workers had had little guidance about how to combine court intervention with casework relationships founded on acceptance and empathy. The conceptualization of the therapeutic 'holding relationship' did not extend to holding clients against their will. The functionalists demanded practice that was ?individual, non-moral, non-scientific, non-intellectual, which can take place only when divorced from all hint of control? (Taft, 1933, p. v). There was general agreement that casework based on client self-determination was incompatible with the use of authority exerted either through covert ?persuasion? or legal sanction (Biestek, 1957; Rogers, 1951). There was nothing in the diagnostic approach to prevent   44  the combined use of care and control; Richmond had done this from the beginning of the century. However until the 1950s there was little recognition that these involuntary relationships might need a different approach.   This posed problems for child protection work. Workers carrying protection responsibilities were accused of being embarrassed by the overt use of authority, and practicing as if they were in child guidance clinics (Beck, 1955). The shift to more therapeutic work led to the exclusion of  ?the extremely disturbed and disorganized persons who had received attention from the protective agency? (Beck, 1955, p. 2) but were now viewed as poor candidates for successful treatment. Social workers became so ?concerned with the neurotic? that treatment became so pre-occupied with the inner life as almost to lost touch with outer reality? (Hamilton, 1958, p. 22). Protective workers were accused of ineffectiveness, manipulation and failing those who needed them most (Keith-Lucas, 1953).  New "aggressive" (Overton, 1953) casework techniques began to be developed through work with juvenile delinquents and their families. These included home visits, meeting client attempts to withdraw from service with increased contact, and holding on to the belief that contact was beneficial, whether it was wanted or not. Ego-psychology justified the clinical use of ?restraint of impulses? (Hamilton, 1951, p. 46) to meet the client?s child-like need to develop ego-strength through a period of enforced dependency (Studt, 1954). Workers were advised to be clear at the beginning of service about abuse concerns, the restrictions on themselves and their clients and the areas for negotiation (Beck, 1955). They needed to be comfortable using their social authority to motivate, re-educate and treat clients (Studt, 1954). They were advised to deal immediately with negative transference and accept the client?s increased need for dependency due to their reduced social status (Studt, 1954). Workers had first to translate their social   45  authority to psychological authority before they could motivate clients to become self-determining. However, as evidenced by contradictions in statements about the use of authority, considerable ambivalence remained:  Social workers are discovering how a positive use of authority and a certain tenacity can help clients to induce inward change. Social workers are interested in no practice which invades the civil rights of individuals, which usurps the proper power of the courts, or which puts them in the position of arbiters of social morality. (Beck, 1955, p. 19)   By the beginning of the 1960s increasing interest was being paid to ?the suddenly visible poor, to endemic civil and social wrongs suffered by minorities, to rotting slums and rising crime? (Perlman, 1979, p. 14). Social workers started to turn their attention outwards again and the frontline worker added locator of service, interpreter of need, mediator and advocate to her job description (Woods & Hollis, 1964/1990). The functions of the casework relationship expanded to include ?manipulation of the environment, reassurance, persuasion, direct advice and guidance, suggestion, logical discussion, exercise of professional authority and immediate influence? (Woods & Hollis, 1964/1990). While the two dominant generic practice models of psychosocial (Hollis and Woods, 1964) and problem-solving social work (Perlman, 1957) were very different, they shared a common view of the worker-client relationship. It continued to be a therapeutic space in which to challenge dysfunctional relational patterns and build ego-strength.  However the relationship once again became more of a means to the end of intervention into the wider systems in which it was embedded.    46   Central to the social work relationship was the client?s faith in the expertise and professional authority of the worker. The worker ?is accorded a leadership or pattern-setting role in some area of living. In such a relationship the individual either identifies with and imitates the worker, or subscribes to the worker?s values, or accepts his or her assessment, suggestions and advice (Woods & Hollis, 1964/1990, p. 94). This was a one way relationship ?from helper to help seeker? (Perlman, 1979, p. 67), in which there was considerable room for education and frank advice.   The worker?s role was to lead and push for change: it was this ?expectation of becoming? that led Perlman to reject Roger?s definitions of unconditional positive regard and nonjudgmentality. It was balanced, however, by the functionalists' faith in partnership with the client as self-determining expert on his own situation and an active participant in all stages in the work. The relationship was ?an undertaking in which two people work together on a problem. They have mutual respect and mutual interest in improving the client?s well-being. Frankness and openness contribute to a feeling of mutuality? (Woods & Hollis, 1964/1990, pp. 110-111). The client had the power to reject any plan. Precise and clear communication averted misunderstandings and hidden agendas. Honest, direct and rational negotiation of the work culminated in explicit contracts. Lengthy relationship-building was not a prelude to assessment; the worker headed straight into eliciting the client?s view of their problems with courage, curiosity and frequent checks on her own and client understanding.  Empathy, warmth and acceptance remained necessary, although insufficient, conditions for effective and meaningful work. The client needed ?compassionate connectedness along with the rent money,? (Perlman, 1957, p. 53) because the very act of requesting help raised feelings of vulnerability and resistance which would undermine participation unless met with genuine   47  concern.  ?The creative but controlled ?use of self? requires that the worker keep a constant and conscious balance between head and heart, distance and closeness? (Woods & Hollis, 1964/1990, p. 206). This balance was achieved by a clear orientation to professional role and therapeutic purpose. Reassurance, acceptance and compassion were exercised as ?sustaining techniques? (Woods & Hollis, 1964/1990) which were to be calibrated in accordance with the client?s needs. There was ?continuous movement between momentary merging with his client/patient and regaining his objective stance as a professional responsible assessor and actor on the client?s behalf? (Perlman, 1957, p. 59). The worker used time and self-reflection to manage her transference to keep a clear and objective focus on the client?s needs. In conflicts between professional and personal responses the professional response was required by social workers who were  supposed to be warm, empathic, caring about each and all of the people who need their help. And then they encounter . . . the surly man who has beaten his baby into unconsciousness; the bedraggled woman who locked in her three hungry children while she sat at a neighbourhood bar. Can one genuinely care for such persons? (Perlman, 1979, p. 99). Perlman answered that if she did not really care, she needed to act as if she did, as all contact needed to serve a therapeutic purpose.  The psychosocial and problem-solving models both claimed to address the needs of involuntary clients, although they proposed different techniques to do so. Perlman advised the worker to see client resistance as the first problem to be addressed. Client and worker examined it together using transparency and curiosity before moving on to other problems once common interests had been established. Woods and Hollis claimed the combination of ?sustaining   48  techniques? and satisfaction of material needs to be particularly effective: ?With the ?hard to reach? much generous ?doing for? may be necessary to overcome the stereotype of the caseworker as an interfering hostile do-gooder? (1964/1990, p. 40). Common to both approaches was the idea that resistance was a time-limited issue which, once addressed, was unlikely to reappear. The overt use of social authority remained a last resort and a failure of the therapeutic relationship. If it was necessary it should be executed with sufficient confidence to override the client?s temptation to resist.   This type of relationship demanded a high degree of clinical skill. It was developed at a time when ?weekly conferences of several hours were held regularly for discussion of complicated or baffling cases? (Woods & Hollis, 1964/1990, p. 16). There were frequent recommendations for social workers to seek personal therapy and skilled supervision (Smalley, 1967; Woods & Hollis, 1964/1990). They needed emotional maturity and "a sense of wholeness, of being put together, of knowing who and what one is, what one?s guiding values are, and as a result, of being on fairly good terms with oneself" (Perlman, 1979, p. 60). As Perlman added, ?That?s a tall order? (p. 60). It is questionable how much workers in mainstream child welfare agencies were able to build these relationships involving ?the demanding self-discipline, the self-controlled practice, the putting themselves out, the steadfast giving, the tasting and bearing of the pain of others or the slings and arrows of some of the outrageous person and problems they encounter? (Perlman, 1979, p. 203). They did not have much time practice before the ?discovery? of physical abuse made this complex kind of relationship practically unattainable.    49  1960 - 1990 While there was a recognition that the interests of the adult client and her children may conflict (Perlman, 1957), a popular view mid-century was that deliberate or wilful maltreatment was extremely rare (Lynch, 1985). This began to change after 1962 when an American pediatric radiologist co-wrote a paper entitled ?The Battered-Child Syndrome? (Kempe, Silverman, Steele, Droegemueller, & Silver, 1962). It was quickly followed by papers on the same subject by doctors in the United Kingdom (Griffiths & Moynihan, 1963) and Australia (Wurfel & Maxwell, 1965).  Doctors had questioned unexplained fractures in children in the past (Caffey, 1946; Lynch, 1985), but these papers caught the attention of media and led to a redefinition of child protection work. Its prime concern became physical injury, not neglect and delinquency. This new definition of child abuse was heavily promoted by a powerful medical profession, and in particular by pediatricians, diagnostic radiologists and forensic pathologists who were keen to use increasingly sophisticated technology to expand into new professional territory (Wattam, Parton, & Thorpe, 1997). In the UK the NSPCC seized the opportunity to carve out a new role to compensate for their declining power as public authorities took over front line work. It established a research unit dedicated to publicizing this new understanding of child welfare (Parton, 1979).  In this newly medicalised discourse, abuse was perpetrated by the ?character-disordered family? (Anderson & Shafer, 1979). Such families needed treatment (Howe, 1992) and this required social workers to resume a role similar to that of the ?socially diagnosing? workers of the 1920s. As Kempe said,  ?Early on you need a very careful diagnosis of family abnormality. About 80% of our cases are dependant, inadequate, yearning people?another 10% are, frankly,   50  mentally ill. They are paranoid schizophrenics, psychopathic personalities, aggressive psychopaths? (1973, p. 805). While social work as a whole had begun to embrace structural explanations for social problems, with the help of ecological (Germain, 1978) and systems theories (Pincus & Minahan, 1973), child protection practice moved back to ideas of individual pathology. There was some resistance. For instance, a proposal was made to the 1973 United States Senate hearings on the Child Abuse Prevention Act that abuse be defined as ?inflicted deficits, or gaps, between the specified rights and the actual circumstances of children, irrespective of the sources or agents of the deficits: Every child?should be entitled to equal social, economic, civil, and political rights? (Gil, 1975, p. 112).  However at a time of great optimism in medical science, individualized treatment and re-education for physically abusing parents promised cheaper rewards than the long-term structural interventions necessitated by seeing abuse as a problem of neglect and neglect as a problem of poverty.     In America in 1963 responsibility for child protection investigation was still spread across police, social work and public assistance agencies and many cases fell through the gaps (Besharov, 1985). The new focus on physical abuse led to a decade of theoretical and administrative confusion as governments scrambled to reorganize their statutory systems, and referrals rose with the implementation of mandatory reporting and an expanding definition of abuse to include ?practically every physical and emotional risk to children? (Newberger & Bourne, 1978). Resource constraints increased the need for child protection to adopt simple, short-term intervention models. Indeed science itself seemed to support this shift as researchers suggested that it was the first few sessions of the therapeutic encounter that produced change (Reid & Shyne, 1969) and that traditional psychotherapy was ineffective (Eysenck, 1952). There was increasing interest in behaviourism (Thomas, 1970) and the emerging casework approaches   51  of crisis intervention (Rapoport, 1970) and task-centred work (Reid & Epstein, 1972). Neither crisis intervention nor task centred methods were intended for involuntary clients or those with the chronic difficulties experienced by those child welfare clients for whom ?being in a state of crisis is a life style? (Rapoport, 1970, p. 304). However all three approaches became integrated into child welfare social work education.  From the mid-1960s court referrals for child welfare cases increased rapidly (Newberger & Bourne, 1978), with influential figures arguing that every case reported by a doctor should be brought before a judge (Kempe, 1973). In England the trend was accelerated by the first public inquiry into the death of a child involved with child welfare services, after seven year old Maria Colwell was killed by her stepfather (Parton, 1979). Amidst a moral panic fuelled by the media, concerns about children falling through the cracks of the newly organized public system and the New Right reaction against an over-interventionist welfare state (Howe, 1998), the British government seized the chance to take a clear stance on a popular issue (Parton, 1979). This and subsequent inquiries established the judiciary as the final arbiter of child protection decisions and indeed of the health of the entire protection enterprise. They implied that ?not only is social work secondary to the law and can only be understood within its legal framework, in effect social work activity is the functioning of one area of the law in practice? (Parton, 1986, p. 514). In the United Kingdom, police quickly became integral members of the Area Review Committees established in 1974 to oversee protection cases (Parton, 1979) and the focus shifted from welfare to justice (Howe, 1992), from helping to the gathering of evidence.  With the remit of child welfare redefined as deliberate abuse, it was easy to make the case for specialist services, distinct from mainstream supportive social work. Legislation like the 1974 Child Abuse Act in the United States established this specialist child protection system. As   52  the rights of parents and their children became defined as dichotomous within an adversarial legal-based system, the goal of intervention shifted to protecting children from their parents rather than returning the whole family to healthy functioning (Howe, 1992). The rising tide of neo-liberal residualism, financial cuts and contracting out of supportive services during the 1980s (Scarfe & Sullivan, 2007) accelerated the trend for workers to be seen as information-seekers rather than helpers. Increasing attention to the substantiation of child protection reports left statutory workers little time for therapeutic work. One study found child protection meetings focused so heavily on establishing the details of the child abuse or neglect that an average of only nine minutes was left for discussion of an intervention plan (Farmer & Owen, 1995). In America the average family under supervision received five visits in six months, ?after which the case is closed or forgotten in the press of other business? (Besharov, 1985).   The casework methods that emerged in the 1970s required workers to take a more directive role with their clients. As a ?behavioural engineer? (Sheldon, 1982) the worker taught clients to record their behavior and rehearse new skills, and sometimes directly delivered positive or negative reinforcement. In task-centred work he used ?systemic communication? to facilitate the negotiation and completion of contracted tasks (Reid & Epstein, 1972). In crisis intervention he met the client?s need for leadership until homeostasis was regained (Rapoport, 1970). Drawing on social learning theory (Bandura, 1977) the worker offered advice, encouragement and education. This approach required greater transparency than past ?pseudo-therapeutic and indirect methods of influence? (Reid & Epstein, 1972, p. 112). The worker was to openly explain her specialized techniques and theories, share her hypotheses to reduce client anxiety, and justify her choices ?since the client?s co-operation is an obvious requirement for successful casework, and since he can cooperate better if he knows what the caseworker and he are supposed to do?   53  (Reid & Epstein, 1972). The relationship was now more openly acknowledged as a partnership in which the client rationally gave power to the worker who used her professional expertise to lead him through a process of recovery.  In this conceptualization of the worker-client relationship the worker?s leadership was no longer balanced, as it had been in the previous conceptualisation, by a nurturing therapeutic role. Psychoanalytic theory was going out of fashion, taking with it the ideas that intimacy and an acute sensitivity to the transference relationship were necessary to understand the client. Empathy had been recognised as an important variable in the psychotherapy literature (Truax & Carkhuff, 1967), but in social work now tended to be conceptualised more as a cognitive skill to ensure responsiveness to the client?s meaning than an emotional experience for either client or worker. This change should not be exaggerated. A positive, trusting and warm and mutually respectful relationship remained the foundation for client engagement. The client still needed to feel that the worker had her best interests at heart and ?the worker must maintain conventional, culturally patterned civility . . . is courteous, friendly, and open with clients and others in the modification effort? (Thomas, 1970, p. 197). However ideas of care and nurturing were less prominent in cognitive approaches and indeed were to be seen as increasingly problematic in light of the social worker?s growing use of her mandated authority: The effect of the present public relations over-kill is to mystify and deceive both client and worker as to the potential outcome of their transactions. Both parties are placed in continual double binds with regard to honesty, confidentiality, ethics and morality. In the old days an ?agent? of the Society went out to ?investigate? a ?report?. Today that isn?t done. The community sends out a ?social worker? to extend   54  ?casework services? in response to a ?request? from one of its members (Fortin, 1975, p. 87).   During the 1960s and 1970s parents involved in child welfare services were increasingly defined as different and dangerous: ?the fact is that some parents don?t like their children. They hate them? (Kempe, 1973, p. 804). In child welfare literature of the period parents were described as untrustworthy and unable to engage in a helping relationship: ?The lack of normal conscience and behavioral control, the dangerous and repetitive acting out of these adults, the unreliability of their promises all make authoritative intervention imperative for the protection of children? (Anderson & Shafer, 1979). For them, the coercive use of mandated authority was deemed acceptable and ?may actually be therapeutic; that is, they need the pressure of protective services or the court to enable them to focus on the problems? (Faller, 1985).   The relationship between child protection workers and clients was heavily influenced by the new emphasis on investigation (G. Jack, 1997). Howe described the worker's increasingly authoritative stance: "Clients are expected to comply and conform; they are not diagnosed, treated or cured. If they know the rules, it?s up to them to decide whether or not to abide by them" (1996, p. 88). Workers were criticized for naivety and excessive optimism (Blom-Cooper, Beal, Brown, Marshall, & Mason, 1985) and were encouraged to develop a more forensic, policing approach (Besharov, 1985). This required professional objectivity and emotional detachment, leading to accusations that ?the protective worker will become the IBM of the profession: proficient, efficient, straight, but not terribly human? (Fortin, 1975, p. 92).  The rising tide of neo-liberal residualism, financial cuts and contracting out of supportive services during the 1980s (Parton, 1994, 1998; Scarth & Sullivan, 2007) accelerated the trend for   55  workers to be seen as information-seekers rather than helpers. The intrinsic therapeutic value of the worker-client relationship was largely disregarded. The child protection literature of this period emphasises the science of interviewing, intervention techniques and risk assessment over relational skills. Removals increased, the media ravaged the public reputation of child protection workers (Ayre, 2001) and parents retreated from voluntary involvement with statutory authorities (Meddin & Hansen, 1985). When it took several visits to even contact increasingly alienated clients (Faller, 1985) it was harder to build meaningful relationships. Even investigations came to rely more on information from other professionals than from clients themselves.   Arguably the increasing use of facilitated interprofessional conferences and alternative dispute resolution processes increased the physical and emotional distance between client and worker. Ecological systems theory (Germain, 1978) had contributed to the perception that abuse was a multi-faceted social problem too complex to be managed by any one profession. From child protection conferences to family group decision-making conferences and mediations, social workers slowly handed much of the skilled work of negotiating plans over to 'neutral' third parties. Time for direct work with parents was increasingly taken up with passing information from investigation and surveillance around the interprofessional network (Parton, 2008) and managing inter-professional relationships (Reder & Duncan, 2003). Child protection workers were described as information managers in bureaucratic systems which depersonalized clients by categorizing them according to their assessed risk status in a futile attempt to render them predictable (Blaug, 1995; Howe, 1992).  This conceptualisation of the worker-client relationship did not exclude empathy and compassion and some child protection clients reported feeling valued and cared for by their workers (Winefield & Helen, 1995). Davis (1995) found that meeting child protection clients?   56  needs for nurturing was an important strategy to gain compliance. However immediate material help was emphasized over unconditional acceptance as the basis on which a trusting relationship might be built (Faller, 1985). The investigative and helping roles, the goals of client self-determination and mandated behavior change, were widely seen as incompatible (Parton, 1995; Rothman, 1989). Role conflict was blamed for burnout (Drews, 1980) and protective social workers were painted as depressed, detached from their clients and overwhelmed by professional demands (Fryer Jr, Poland, Bross, & Krugman, 1988).   1990 - Present  By the 1990s, a series of scandals involving children removed from their families after flawed sexual abuse allegations was widely seen as evidence that the over-interventionist child protection system had lost its way. In the United States child protection reports had tripled since 1980 (Berg & Kelly, 2000) and the alarming recidivism rate raised concerns as to the ineffectiveness of the response (Faller, 1985; Inkelas & Halfon, 1997). Similar trends were seen in Canada, Australia and the United Kingdom (Lonne, Parton, Thomson, & Harries, 2008). Most reports were of neglect and did not fit the narrative of dangerous physically and sexually abusing parents which underpinned the investigative approach. Most cases were closed without services being provided (Parton, 1997).  Calls for greater client participation in services since the 1970s (Fortin, 1975; Mayer & Timms, 1970) became more urgent in light of the devastating impact of investigations on families (Cleaver & Freeman, 1995). Supporters borrowed the participatory discourses of feminism, community development and the disability rights movement (K. Healy, 1998). They were joined by neoliberals and consumerists on the political right who looked to market   57  principles and user involvement to correct inefficient and unaccountable public services (Croft & Beresford, 1994). Psychotherapy researchers had convincingly demonstrated the importance of a positive worker-client relationship in effecting change (Horvath & Symonds, 1991; Orlinsky, Ronnestad, & Willutzki, 2004). A discourse of partnership with self-determining clients connected to extensive informal resources (Coady, 1993) offered statutory agencies the opportunity to target protection services and divest responsibilities back to families (Parton, 1997) at the same time as engaging clients more effectively (Dumbrill, 2006). This led in the United Kingdom to the 1989 Children Act which can be seen as the first legislative attempt to make child protection work more about partnering with parents than policing them (Lonne et al., 2008).  Many jurisdictions implemented differential response systems (T. Hetherington, 1999; Marshall, Charles, Kendrick, & Pakalniskiene, 2010; Merkel-Holgu?n, Kaplan, & Kwak, 2006; Trocm?, Knott, & Knoke, 2003). These rested on the assumption that different types of case warranted different responses and that the prevalent practice of leaving families until their situations had deteriorated sufficiently to be deemed high risk was reactive, inefficient and dangerous for children (Waldfogel, 1998).  High risk cases, typically those involving imminent harm, serious physical injury or sexual abuse, were still investigated (Kaplan & Merkel-Holguin, 2008; Merkel-Holgu?n et al., 2006). However even investigating workers were expected to pay more attention to parental rights to participate in decision-making processes (Bell, 1999). In child protection cases deemed to involve lower risk an alternative to investigation was offered.    Most child welfare systems today offer some kind of differential response. The format differs widely across jurisdictions, and is delivered both by statutory child protection agencies and community agencies on their behalf  (Conley, 2007). However core components of   58  differential response approaches are a family assessment, collaborative approach to meeting a child?s needs and emphasis on services rather than surveillance (Kaplan & Merkel-Holguin, 2008). There is no formal determination or substantiation of child abuse or neglect and, although parents are often advised that service refusal may lead to mandatory child protection intervention, their involvement in differential services is at least nominally on a voluntary basis.   By pursuing alternative responses to investigation it is intended that ?the family voice and commitment for child and family safety and well-being is leveraged, underscoring the notion of child protection as a shared concern and responsibility? (Christenson, Curran, DeCook, Maloney, & Merkel-Holguin, 2008). This is the same idea that underpinned the introduction of a raft of new family involvement processes, from family case planning conferences to family group decision making conferences and mediation. Indeed in many jurisdictions differential response and family involvement strategies have been promoted not as discrete strategies but intertwined in a new ?partnership? approach (Christenson et al., 2008; Comer & Vassar, 2008).  One example of this has been the British government's attempt to move beyond the provision of differentiated child protection and family services to create an integrated service oriented to the needs of families and 'safeguarding' children?s wellbeing rather than simply protecting them from neglect and abuse (Parton, 2006). Strengths-based solution-focused practice is widely seen as another means to enact the principles of parental partnership and participation and to provide a differential response that takes into account the unique features of every family.   The new emphasis on parental participation saw the emergence by the mid 1990's of a more complex model of the worker-client relationship. However the shift to a more participatory relational model was difficult as it was ?welded onto a system for classifying risk which, arguably, was not constructed properly to support it? (Bell, 1999). Parents were often 'groomed'   59  for participatory events (Bell, 1999) and their success relied on their willingness to accept and demonstrate contrition for the social worker?s version of concerns (Holland, 2000). Workers were often committed to the idea of partnership but ?despite the rhetoric of participation, professionals did not encourage parents to share their views and understandings . . . and did not welcome information that undermined or obscured the clarity of their point of view (Corby, Millar, & Young, 1996). Indeed partnership was seen as a troublesome concept when conflict was so central to the work (Corby et al., 1996). Arguably it was only the introduction of solution-focused and strengths-based ideas to child protection practice that provided the theoretical foundations for a change in the working relationship.   Current Contextual Challenges A number of features of contemporary child protection systems appear problematic for the implementation of strengths-based ideas. These are discussed in the next section.  1) Systemic Pressures Today's frontline child welfare workers frequently carry high caseloads and operate with insufficient resources (Bennett et al., 2009; Herbert, 2003). The assumption of strengths-based solution-focussed approaches that ?most people can change their behavior when provided with support and adequate resources? (Berg & Kelly, 2000, p. 63) may not translate well to contexts in which support and resources are less than adequate. Clients often experience multiple chronic problems which are not easily resolved even in the context of a positive working relationship. Change often relies on scarce services with long waitlists and on people outside the worker-client dyad.   60   There are strong environmental pressures to retreat to relationships organized around the risk assessments still needed by courts and child protection systems, and risk management has become a central organizing feature of child welfare work (Houston, 2000; Parton, 1998, 2011; Smith & Donovan, 2003; Spratt, 2001). Procedural and information management responsibilities have been identified as undermining the ability of child welfare workers to exercise clinical judgment and to build relationships with clients (Parton, 2008). Workers operate within systems designed more for managing client volume than for clinical effectiveness and must work within organizational policies based on priorities that they may not share (Hasenfeld, 1987; Lipsky, 2010). In order to cope child welfare workers have been found to reinterpret strengths-based policies in ways that allow them to routinize their case management, de-prioritize anything not seen as a core function, and withdraw from client contact (Smith & Donovan, 2003). This means that social control is arguably still the dominant feature of the child welfare role (Scourfield & Welsh, 2003).   2) Worker Capacity Strengths-based solution-focused work requires the worker to hold on to a genuine faith in the client?s potential to change and to continually instill this hope and the optimism in the face of disappointments. We do not know what happens to the worker-client relationship when the workers stop feeling the optimism on which it relies. The question is important as high levels of stress and the combination of emotional exhaustion, depersonalisation and low self-efficacy that constitutes burnout are typical in child protection work (Boyas, Wind, & Kang, 2012). Worker turnover is high (Boyas et al., 2012; Cyphers & Association, 2001; K. Healy, Meagher, & Cullin, 2009). This means that child protection agencies have increasingly recruited graduating social   61  work students (Cyphers & Association, 2001) and the frontlines are manned disproportionately by inexperienced workers at the beginning of their careers (K. Healy et al., 2009).  Child protection workers rarely receive the reflective case consultations and relationally-oriented supervision that has been linked to the sense of containment necessary for relational practice (Ruch, 2007; Trevithick, 2003). While solution-focused practice is a therapeutic model child protection workers no longer see themselves as therapists and typically lack training in therapeutic work (Yatchmenoff, 2005) and competence in counselling skills (Forrester et al., 2008). As Healy et al. (2009) comment, ?Many workers are simply not in the role long enough to develop the strong body of context-based knowledge and skills required for expert child protection practice? (p. 301). The opposite problem has also been identified in that social worker education can reinforce the orientation to helping and sense of professional expertise which can stand in the way of the genuine appreciation of client capacity required for strengths-based solution-focused work (Blundo, 2001; Grant & Cadell, 2009).   3) Violence   We know little about the way in which social workers reconcile the new emphasis on openness, honesty and use of self with the anxiety (Ruch, 2007) and ?pervasive sense of powerlessness and fear? (Barter, 2008, p. 94) associated with child protection work. Child protection workers are at significant risk of violence and abuse (Ferguson, 2005; B. Harris & Leather, 2012; Littlechild, 2008). One study of Canadian workers found more than half reported receiving threats of violence and almost a quarter reported being assaulted by a client (Regehr, Hemsworth, Leslie, Howe, & Chau, 2004). Their declining public reputation and increasing use of mandated authority over the last 40 years means they are often perceived as unwelcome intruders into   62  clients' lives. Paradoxically, the recent shift to emphasize family support over investigation may increase the risk of violence, as it leaves both worker and client with more ambiguous roles and increased uncertainty about power (Littlechild, 2008). It is hard to reconcile the picture of the worker who feels relief when the door is not answered (Ferguson, 2005) with Turnell et al?s (2008) description of the strength-based worker who jokes in the face of an angry, screaming client in his isolated apartment late on a Friday afternoon. In situations of threat workers may engage in protective strategies focused more on pleasing or avoiding their clients than on establishing cooperative partnerships (Calder, 2008a; Ferguson, 2009; Goddard & Tucci, 1991; Stanley, Goddard, & Sanders, 2002).  4) Mandated Authority In countries like Canada, United Kingdom and United States, contemporary child protection operates in the grey area of ?semi-compulsion? (R. Hetherington, 1998). The threat of court intervention is ever present but normally so distant that it does not dominate contacts with clients. This creates the potential for what Mary Richmond identified in 1917 as the ?temptation to indirectness, subterfuge, concealment or ambiguity, in which you might drift in spite of yourself while striving to help? (p. 109). Even when workers are clear about their mandated authority, the emphasis on validating the client?s perspective, starting every session ?what is better?? (Berg & Kelly, 2000) and building cooperation by amplifying client successes may obscure the worker?s power. Some clients already assume that workers act for them and are unaware of alternative loyalties to the child, the courts and the employing agency (Regehr & Antle, 1997) and social worker discomfort with power means it tends to be downplayed or   63  denied (Bar-On, 2002; Bundy-Fazioli, Quijano, & Bubar, 2013; R. Hetherington, 1998; van Nijnatten, Hoogsteder, & Suurmond, 2001).  It may be difficult to resolve contradictions between the client?s narrative truth, on which the strengths-based solution-focused working relationship is built, and the historical truth demanded by courts and the quasi-legal processes on which child protection depends. An extreme example of this is the ?similar but different? technique that developed through strengths-based solution-focussed work with parents who deny child abuse concerns (Turnell & Essex, 2006). The worker leads the parents through role plays featuring hypothetical abusive clients to enable them to demonstrate knowledge of children?s safety needs while maintaining their position of denial. While the worker is warned not to trick clients into confession or analyze the results for historical truth, there is little guidance on how to use sophisticated therapeutic manoeuvres like these without including the unintended results in formal assessments. Clients can be tricked by empathy into disclosures that may later by used against them (Strasburger, Gutheil, & Brodsky, 1997) and the potential for the worker-client relationship to end in accusations of betrayal after professionals testify has led to clear recommendations from some mental health practitioners to separate therapeutic and forensic roles (Greenberg & Shuman, 1997, 2007; Wright & Odiah, 2000).   5) Differing Perspectives on Good Practice As has been the case since the birth of the modern child welfare movement, frontline workers navigate conflicting messages as to what constitutes good practice. Over the last decade the increasing prominence of relational social work and approaches specific to mandated clients have brought to light some of these tensions in theoretical perspectives.    64   Relational Social Work The development of strengths-based solution-focused child protection approaches can be seen as part of a broader trend in generalist social work over the last decade towards relational (or relationship-based) work. This has been informed by attachment theory (Howe, 1998), feminist relational theory (Freedberg, 2009) and the relational psychodynamic paradigm (Borden, 2000; Ornstein & Ganzer, 2005). Relational social work is based on the ideas that humans develop in relationship to others and to their environment and the worker-client interview is a therapeutic space in which the client learns experientially through an accepting, nurturing and empathic relationship. The creation of a healing relationship is both an end in itself, and the means to better assessments and more effective client connections with the broader social environment (Trevithick, 2003).  In this approach both client and worker are active participants in the relationship, influencing and influenced by each other (Freedberg, 2009). Each have a stake in defining the nature and intimacy of the bi-directional relationship (Alexander & Charles, 2009; O'Leary, Tsui, & Ruch, 2013). Rather than containing the emotional reactions of self and client by setting rigid interpersonal boundaries, the worker engages with, interrogates and learns from them (Freedberg, 2009; Ornstein & Ganzer, 2005). Indeed she may intentionally seek intense emotional attunement to make herself vulnerable and open to the client?s experience (Freedberg, 2009). The separation of the personal and professional selves is neither possible nor desirable (Mandell, 2008). The worker can feel, disclose and therapeutically use deep emotional empathy, compassion and even love for the client (Maidment, 2006). She constantly shifts between assessing her own and her client?s feelings (Freedberg, 2009) and engages in ongoing reflexivity   65  to ensure use of self is in the service of the client (Mandell, 2008; Ruch, 2005). According to Weick (2000), this means that the caring and caretaking at the heart of social work is reclaimed. It is notable that little of the relational social work literature engages with the question of power and how practitioners might concurrently negotiate intimacy and authority.  Client research supports the call to move further towards relationships characterized by caring, empathy and reciprocity (Beresford, Croft, & Adshead, 2008; Coady, 1993; Ornstein & Ganzer, 2005; Schreiber, Fuller, & Paceley, 2013). Clients want workers to put aside their professional authority and prejudgments (Drake, 1994) and demonstrate ?compassionate ordinariness? (Huxley, Evans, Beresford, Davidson, & King, 2009). Although worker expertise is important, more so are the natural ?human qualities of kindness, warmth, compassion, caring, sensitivity, empathy and thoughtfulness? (Beresford et al., 2008). Clients want social workers who show their humanity by being honest, down-to-earth, sharing their emotions and disclosing information about their lives (De Boer & Coady, 2007; Schreiber et al., 2013). They want relationships which are comfortable and friendly and that make them feel listened to and cared about (Ghaffar, Manby, & Race, 2012; Schreiber et al., 2013). Perhaps most surprisingly some value positive relationships with social workers as real friendships, characterized by mutual affection, informality, reciprocity and strong emotional bonds (Beresford et al., 2008; De Boer & Coady, 2007).  Mandated Relationship Approaches Solution-focused child protection theorists join the small number of contemporary social work authors (Barber, 1991; Ivanoff, Blythe, & Tripodi, 1994; Rooney, 2009; Trotter, 1997, 2006) who suggest that mandated work needs to proceed on a different basis from work with the   66  ?vulnerable voluntary? (Rooney, 2008, p. 116). As Ferguson (2005) writes, ?Empathy, sensitivity, warmth and Rogerian ?unconditional positive regard? are still consistently identified as what social work should be about. But these are deeply problematic in work with involuntary clients? (p. 793). For many parents of children deemed to be at risk of abuse or neglect, engagement with child welfare services is unwelcome. This suggests the value of models of engagement from social work and other fields of practice like corrections and compulsory mental health treatment that specifically address the needs of mandated clients.   Much writing about mandated clients draws on social exchange theory (Cook & Rice, 2006) and cognitive behavioural theory (Sheldon, 2011). It suggests the client is a rational decision maker who makes change if it is in his interests to do so. To make rational decisions the client needs reliable information. This makes transparency about authority, goals and process one of the most important requirements of the casework relationship. Explicit discussion of the worker?s role and authority is the foundation of the working alliance. It teaches the client the limitations of the relationship, what is and is not negotiable and enables the client to perform her role (Barber, 1991; Ivanoff et al., 1994; Rooney, 2009; Trotter, 2006). It maximizes the congruence between worker and client expectations which in turn increases the chance that clients will experience the relationship as supportive (Svensson, 2003). Trotter found that child protection clients tend to be confused about the role of their social worker, perceiving them as friend and helper or as supervisor and investigator, but rarely as both (Trotter, 2004). Those who understood clearly that their workers held a dual surveillance and helping role tended to have better outcomes.   This open discussion about the nature and limits of the casework relationship is more than an initial engagement strategy, but revisited regularly through the life of a case. The worker   67  strives for honesty and consistency in the management of rules and decisions rather than positive feelings and mutuality (Calder, 2008). Trotter takes this furthest by asserting that ?whether clients are voluntary or involuntary they have a right to know about a process whereby their attitudes and actions may be influenced or changed? (Trotter & Ward, 2013, p. 10). He advocates explaining to clients how therapeutic techniques and positive and negative reinforcement will be used to shape their behavior in order that they might make fully informed choices. This level of transparency may be difficult to achieve when workers pursue conflicting goals of efficiency and care. As Rooney comments,  ?practitioners and agencies often do have hidden agendas . . . while greater candor on all sides might make for a better world, it would be more realistic at this point to accept lack of complete candor as expectable in involuntary transactions? (Rooney, 2009, p. 198). The issue is complicated by a lack of practical guidance on how to perform these repeated and explicit discussions of power.   A second element common to approaches with mandated clients is a focus on the negotiation of interests, goals and tasks. Negotiation and conflict management skills are the primary requirement and a trusting relationship is not a precondition for assessing or making demands of the client (Ferguson, 2005). Barber (1991) advises worker and client to begin with no expectations of common ground and negotiate an intervention plan which maximizes opportunities for client self-determination. The worker seeks out client problem definitions, interests and goals and endeavors to include these as far as possible in a process of negotiation based on a clear understanding about what changes are desired and what are required (Rooney, 2009; Trotter, 2006).    A third common element of approaches for the mandated client is the creation of a microclimate of sanctions and rewards for the client?s goal-oriented efforts (P. Harris, 2008).   68  Pro-social modeling and reinforcement for pro-social behaviour appear to influence outcomes more than empathy (Trotter, 2006). Work with offenders indicates that this is particularly true with clients with psychopathy (Ross et al., 2008), depression, avoidant behavior, multiple interpersonal problems and difficulties making relationships (Orsi, Lafortune, & Brochu, 2010). These clients are common in child protection work.  Offering praise and material rewards for wanted behaviors is not inconsistent with strength-based and solution-focused theory. However the advice (Rooney, 1992; Trotter & Ward, 2013) to ignore or actively confront rationalizations and anti-social comments instead of accepting them as a valid reflection of the client?s experience, is. Several theorists suggest that the overt use of power may be more effective than the relational bond in motivating some involuntary clients who will only in the later stages of the change process become motivated by more intrinsic desires (De Leon, 1988; P. Harris, 2008). Using strategies of control may allow time for the worker to discuss concerns in such a way that the client achieves the insight necessary to engage in voluntary treatment (S. Morgan & Hemming, 1999).   Finally, while approaches for mandated clients tend to play down the emotional aspects of the relational bond it is generally accepted that the working relationship should be characterised by warmth, openness, interest and respect for the client. The punitive use of authority has poor outcomes (Skeem, Louden, Polaschek, & Camp, 2007). Empathy, conveyed thorough attentive listening, encourages client self-expression (Orsi et al., 2010; Trotter, 1997). A caring relationship characterised by acceptance, support, trust and openness helps mandated clients to feel that rules are firm but fair, and it is this perception of fairness which correlates with compliance (Skeem et al., 2007). However, there is evidence that empathy can overwhelm and alienate some mandated clients (Ivanoff et al., 1994), and at the very least prove ineffective   69  if not accompanied by pro-social modeling (Trotter, 2006). Different clients are likely to need a different balance of empathy, support and authority but with mandated clients a positive casework relationship may not in itself be a sufficient basis for change.   Summary  In strengths-based solution-focused child protection work the relationship between worker and adult client is the key mechanism of change and the point at which tensions between competing visions of social work play out. Strengths-based practice (SBP) in child protection straddles the somewhat incongruent strengths-based and solution-focussed traditions, requiring workers to perform a complex relational dance to assert both their own authority and support client self-determination. While SBP has been compared to Settlement House work (Kisthardt, 1997) and identified as a continuation of the functionalist tradition (McMillen, Morris, & Sherraden, 2005; Weick et al., 1989), it appears to require the child protection worker to operate more like the skilled therapist of Perlman and Hollis? days. The emphasis on transparency and goal-setting echoes key ideas from work with mandated clients that identifies conflict as the relationship's defining feature (Barber, 1991). Yet links can also be drawn to the very different approach of relational social work which describes relationships founded on an assumption not of conflict but of cooperation.    We know little about how frontline workers navigate these tensions to make sense of and enact SBP with their adult clients. A review of the changing conceptualisation of the worker-client relationship since the beginnings of modern child protection systems demonstrates that the decision as to how to proceed does not rest solely with individual workers, nor even within the   70  worker-client dyad. The protection agency, courts and general public are parties to the worker-client alliance (Leung, 2002), which is shaped by broader organisational, social, economic and political forces. This raises the question of whether workers see SBP as a realistic and effective approach for contemporary child protection settings. If they do, how do they apply SBP ideas and what helps and hinders them in the attempt? It is these questions this study was designed to explore.     71  CHAPTER 3 : STUDY DESIGN AND METHODOLOGY Epistemology Pragmatism The epistemology informing this study is pragmatism. This is a philosophical tradition founded in the United States in the 1870's by Charles Sanders Peirce (Ketner, 1992) and initially developed in the ideas of John Dewey (1920/2004), William James (1907) and George Herbert Mead (1934). It has many different strands; even in 1908 it was claimed that there were 13 different pragmatisms, all of which were logically independent of each other (Biesta & Burbules, 2003). It is the ideas common to the ?neo-classical? variations which stay close to Peirce?s original philosophy (Haack & Lane, 2006; Misak, 2007) on which this research draws.   The core of pragmatism is the idea that objects or ideas can only be understood contextually by reference to their real-world application. When we understand all the possible consequences of the application of a concept, we understand its full meaning. We cannot come to know the world through the application of any idea of abstract or universal truth. This is because all knowledge is refracted through language, context and individual meaning-making (Ketner, 1992). It is both a product and reflection of the continual process of transaction of human organism and environment (Dewey, 1920/2004). The way in which we understand the world depends on the way we engage with it. This means that an abstract understanding of strengths-based solution-focused ideas is of far less value than an understanding of how such ideas are interpreted in practice.  Pragmatist researchers gain knowledge in a cyclical process which starts with inferences from observations and experiences. These inferences are informed by previous knowledge. This knowledge is made explicit in order that others might interrogate the ways in which it helped or   72  hindered the researcher's understanding of the subject matter and contributed to the inferences made. This is important because pragmatists judge the value of any theory partly by the extent to which it is supported by existing evidence and fits with supporting beliefs that are themselves well-founded and comprehensive (Haack, 2000). However, in pragmatism the key question is 'do the consequences of our ideas turn out as we say?' As Dewey said, ?The test of ideas, of thinking generally, is found in the consequences of the acts to which the ideas lead, that is in the new arrangement of things which are brought into existence" (1929, p. 136). It is the correspondence of suggested meaning and realized meaning that makes knowledge sufficiently useful to allow the prediction of future lines of action (Biesta & Burbules, 2003).  What all this means for pragmatist studies is that researchers make design decisions with reference to what will most usefully answer their practical research question. As James said "The pragmatic method is primarily a method of settling metaphysical disputes that otherwise might be interminable...The pragmatic method in such cases is to try to interpret each notion by tracing its respective practical consequences" (James, 1907/1991, p. 23). The research project is understood in terms of its goals and purposes and researchers may adopt whatever perspective and tools appear most useful for answering the research question. Both qualitative and quantitative methods are acceptable if they help address the problem at hand and the researcher understands the practical application and limitations of any chosen strategy (Onwuegbuzie & Leech, 2005; Tashakkori & Teddlie, 2003).   There are two potential ethical problems with pragmatist research. The first is that there may be no resolution of the question "For whom is a pragmatic solution useful?" (Johnson & Onwuegbuzie, 2004, p. 19). In this study my explicit orientation is to solutions that further the goal of ensuring the child?s safety and are workable for both practitioner and adult client. The   73  second more general criticism of pragmatism is its potential to disintegrate into a crude attachment to the idea that whatever is useful is true. This is prevented by a commitment to fallibilism. In pragmatism all truths are partial and fallible because they are mediated through individual and contextual perceptual frames and because what worked in the past can never be guaranteed to work in the future in a complex dynamically shifting world. This also means that researchers do not have to be immobilized by the search for certainty before they take action. They may make reasonable preliminary assertions based on available evidence as part of the ongoing cycle of action and inquiry that characterizes a pragmatic approach.   Methodology Mixed Methods This is a mixed methods study. Mixed methods is ?research where the researcher mixes or combines quantitative and qualitative research techniques, methods, approaches, concepts or language into a single study? (Johnson & Onwuegbuzie, 2004). It emerged in the mid 1990s as a distinct research approach, producing its own methodologists, community of practice and, in 2007, a peer-reviewed journal (Denscombe, 2008; Plowright, 2011). Its development was partly a response to the polarization of the debates regarding the relative value of quantitative and qualitative methodologies and the epistemologies in which they were embedded. Mixed methods researchers challenged the contention that quantitative and qualitative methods were incompatible (Johnson & Onwuegbezie, 2004) or even distinct (Onwuegbuzie & Leech, 2005). Those advocating pragmatism as a way to move beyond claims regarding the incommensurability of epistemological paradigms have tended to promote mixed methods (Johnson & Onwuegbuzie, 2004; D. Morgan, 2007), and pragmatism remains the most common   74  philosophical foundation for mixed methods research (Feilzer, 2010; Tashakkori & Teddlie, 2009).  A common argument for mixed methods studies is that complex problems are best addressed using different strategies and types of data. In combination each method has the potential to compensate for the limitations of the other (Creswell, Plano Clark, Gutmann, & Hanson, 2003).  Each can illuminate a different facet, or provide a different perspective on, the research problem. Greene, Caracelli and Graham (1989) proposed five reasons for combining quantitative and qualitative methods. These were:  1) Triangulation: results from one method corroborate results from another  2) Complementarity: results from one method add to or illustrate results from another  3) Initiation: contradictions between results from different methods generate new lines of  inquiry  4) Development: results from one method inform how the other method proceeds  5) Expansion: different methods addressing different elements of the research problem  lead to an expanded understanding of the problem   Mixing methods was consistent with my pragmatist perspective that a complex issue like the meaning and enactment of SBP was best tackled from a variety of angles. I added a quantitative element to what had been originally intended as a qualitative study as a result of discussions with senior staff at the Ministry for Children and Family Development (MCFD) in which I was granted access to all MCFD child protection workers. This gave me the opportunity to explore SBP not only at a micro level, through detailed description of individual practitioner   75  experience, but also at a macro level through analysis of broad trends (Onwuegbuzie & Leech, 2005).    Many typologies for mixed methods designs have been proposed (Creswell et al., 2003). Most typologies reference the chronological ordering and the relative importance of the quantitative and qualitative components of the study. Some seek to avoid the terms quantitative and qualitative altogether (Plowright, 2011), or to distinguish between methods according to their function as exploratory or confirmatory (Onwuegbuzie & Teddlie, 2003). Quantitative and qualitative components of a study may be conducted independently of each other throughout the data collection and analysis stages (Creswell et al., 2003) or may mix and inform each other from the design of the research question onwards (Teddlie & Tashakkori, 2006).   In this study quantitative and qualitative components were mixed at the design, data collection, analysis and reporting stages. In order to adequately answer the research question 'do social workers use strengths-based practice?' I needed to assess patterns of SBP use in the child protection worker population, and for this I required quantitative data. To understand the complexities of how workers made sense of SBP, I needed qualitative data. I used two methods of data collection. The chronological first was an online survey emailed to the entire population of 824 fully delegated child protection workers employed by MCFD and completed by 224 workers. This survey contained both quantitative questions, whose responses were statistically analysed, and open-ended questions whose responses were subject to both qualitative and quantitative analysis. The second part of the study comprised of interviews with 24 participants.  The data from these interviews was treated as qualitative.   Statistical analysis of the survey data informed the development of the interview questions. Analysis of the interviews and qualitative survey data led me to construct new   76  quantitative variables which I used in a second round of statistical analysis of the survey data. I also converted qualitative data from the open-ended survey questions into quantitative data by counting the frequency of codes. In my findings I attempted to integrate the conclusions drawn from the quantitative and qualitative procedures.  Quantitizing (Quantifying) Qualitative Data A common technique in mixed methods research is quantitizing, or the numerical translation of qualitative data (Sandelowski, Voils, & Knafl, 2009). It involves coding qualitative data with a consistent set of codes, which may be either inductively or deductively derived. The frequency of each code is then counted. One purpose is, to form qualitative data in ways that will allow analysts to discern and to show regularities or peculiarities in qualitative data they might not otherwise see or be able simply to communicate, or to determine that a pattern or idiosyncrasy they thought was there is not. (Sandelowski et al., 2009) Quantitizing the data provides a way to summarise patterns within it. The converted data is also frequently used in statistical analysis.   Mixed methodologists have described this quantitizing process as 'classical content analysis' (Leech & Onwuegbuzie, 2008; Onwuegbuzie & Combs, 2010), and as 'quantitative content analysis' (Bergman, Tashakkori, & Teddlie, 2010; D. Morgan, 1993). Its roots are often traced to Berelson's (1952) summary of techniques to study the content of mass communication, in which Berelson described content analysis  as "a research technique for the systematic, objective, and quantitative description of the manifest content of communication" (1952, p. 18).While many researchers would now question the claim to objectivity, this kind of content   77  analysis "seeks to answer questions about what and how many"  (D. Morgan, 1993, p. 116) in as systematic a way as possible. It is less good at supporting complex descriptions of phenomena or answering questions of causation as it requires the simplification and decontextualisation of data and much meaning can be lost in the process (Sandelowski et al., 2009). When coding is performed by one person only its reliability can be problematic (Rourke & Anderson, 2004), but this can be tested by matching codes generated by a second coder for a random subset of responses.  Interpretive Description The methodology guiding the qualitative portion of this study is interpretive description (Thorne, 2008; Thorne, Kirkham, & MacDonald Emes, 1997). Interpretive description is highly congruent with, and indeed rooted in, a pragmatist perspective. Its theoretical heritage has previously been identified in symbolic interactionism (Oliver, 2012), which is an approach to the conceptualisation and study of social relations that evolved from pragmatism. Blumer, who first named and wrote about symbolic interactionism, described it as an interpretation of the pragmatism of his mentor Mead (Blumer, 1977, 1986). It has commonly been described as pragmatist philosophy reformulated into the language of sociology and social research (Lewis & Smith, 1980; Shalin, 1991). The pragmatist philosophy underpinning symbolic interactionist inquiry in general and interpretive description research in particular can be seen in the focus on the problems of everyday practice, support for multiple strategies to find a workable resolution to those problems, naturalistic inductive approach, and attention to the ways in which action and meaning-making interact for all those involved in the study.    78   Interpretive description developed within the field of nursing. It was intended to address the problem that applied researchers often had to abandon methodological orthodoxy to follow more flexible research procedures if they were to produce the timely knowledge, sitting ?somewhere between fact and conjecture? (Thorne, 2008, p. 15) needed by the applied disciplines to address practice issues. Interpretive description legitimized drawing from a variety of research traditions and techniques by providing a ?framework within which the design decisions that work for your particular questions can be effectively set forth? (Thorne, 2008, p. 103). It requires a logical, systematic and defensible research strategy to address practice issues in a way that makes sense to the discipline and can reasonably be expected to advance disciplinary knowledge and inform practice change.   There are some basic guidelines for interpretive description inquiry. To maximise the chances of producing knowledge that is useful to practitioners, interpretive description researchers begin with a critical analysis of existing theoretical and clinical knowledge within the discipline. This provides the starting conceptual frame which evolves over the course of the study through an iterative process of data collection and analysis. I conducted an extensive literature review to locate my study in existing knowledge in the field of child protection. This is an important starting point for the ongoing examination of how the researcher's perspective interacts with that of the participants.   In interpretive description studies researchers are expected to be aware of, transparent about and understand the effects of, the perspectives with which they approach their subject. This requires ongoing reflexivity. I wrote regularly in a research journal to record and interrogate my assumptions and developing analysis. I also took steps like avoiding value-laden prompts in interviews, discussing different perspectives with my committee and very consciously adopting a   79  position of curiosity to maximise the extent to which findings were based on the experiences and meaning-making of participants rather than my own preconceptions.  The researcher using interpretive description employs purposeful and theoretical sampling to focus inquiry on the most potentially valuable data sources and to explore and expand emergent ideas that seem most relevant to answering the research question. In an analytical process in which it is important to avoid making a premature commitment to a particular conceptual framework, Thorne suggests that coding and constant comparative analysis are useful tools. In the latter method borrowed from grounded theory (Glaser & Strauss, 1967), units of data are compared to each other and to their context to better describe their properties and explore relationships between them. Thorne describes this as replicating clinical reasoning as the researcher "alternates between asking 'what is going on?' and 'how does this relate to what else is known?'" (Oliver, 2012, p. 412).  Interpretive description does not employ a traditional approach to member checking in which participants are asked to review their own transcript or the analysis for accuracy. This is because the researcher is not seen as the conduit for participant thoughts but is an interpretive actor whose role is to produce conclusions which synthesise the input of all participants rather than mirror the thoughts of individuals. As Thorne (2008) says, simply asking participants whether they said what the researcher believes them to have said "can lead to false confidence if they confirm what you thought, and potentially derail you from good analytic interpretations if they do not" (p. 159). She does however suggest asking participants to comment on emergent synthesised interpretations in order to expand the analysis and assess the extent to which it resonates with participants.    80   In place of detailed procedure interpretive description provides general criteria by which the researcher can examine his or her design decisions and account for adaptations of traditional methods. These must be based on a clear understanding of the relationships between the research question, a chosen method and the limitations of the knowledge it affords. For instance, I combined two different approaches to analysing the interview data. When I identified a relatively common concept like 'transparency' or 'supports' that might be seen as part of a shared disciplinary language I created a broad code which I later broke down into smaller categories according to the object and consequences of worker transparency or the type of support. At other times my analysis moved in the opposite direction, beginning with detailed codes which were grouped into progressively larger categories.  The first strategy did not enable me to interrogate in detail the different meanings of words like 'transparency' and 'supports' and it is possible that workers attributed somewhat different meanings to these words. However, I used this strategy as my research questions were focussed on the actions the words described rather than the minutiae of language use. The second strategy resulted in an overwhelming number of decontextualised codes. However, I used it to limit my natural inclination to move too quickly to broad conceptualisation and to help me attend in detail to the words of participants. I found it helpful throughout the analysis to shift repeatedly between the two analytical positions. Coding pieces of text in multiple ways enabled me to  consider the data from a number of different angles, to keep in mind both the context and the detail of participant contributions and to more easily change and abandon codes as my analysis evolved. In interpretive description credibility rests on the researcher?s ability to be thoughtful about these kinds of decisions and to analyze and justify them in a logical way.   81       The end point of interpretive description studies is a "thematic summary or a conceptual description" (Thorne, 2008, p. 164). It should capture the main elements of the phenomenon and provide practitioners with a new way of illuminating the practice problem, enabling them to approach the problem in a more productive way. As Thorne (2008) suggests, The ideal might be described as a research report that makes visible and accessible the clinical wisdom of a passionate and thoughtful expert practitioner for whom a similar understanding had been acquired through extensive pattern recognition and reflective practice observations (p. 169). As such, the goal is to construct not a generalisable theory, but a heuristic that is sufficiently memorable and relevant to practitioners that it might lead to different actions in regard to the practice problem.    Methods Recruitment Survey The population of interest in this study was current frontline MCFD child protection workers holding C6 delegation. C6 delegation is the legal authority enabling Ministry workers to undertake the full range of child protection duties.  I was provided by the Ministry with a list of all staff in the province who had C6 delegation. This list included frontline workers, team leaders, managers and those such as consultants who worked in positions of support to the frontline. Following a meeting with the Provincial Director of Child Welfare, an email was sent from the Director to all names on this list of C6 delegated workers. The email expressed the Ministry's support for the upcoming study   82  and gave basic information as its purpose and nature. Dillman (2007) recommends this type of pre-contact letter as a recruitment tool. In order to increase awareness of the study I also attended two conference calls to brief senior MCFD managers and worked with communications personnel to have information about the study posted on the Ministry's internal website.    I identified from the C6 delegation list 850 workers whose job or position title was 'child protection worker'. I sent a recruitment letter by email to each of these workers (see Appendix A). I received responses from 26 people, most often via an automated email, indicating clearly that they were not within the sampling frame as they were either did not hold C6 delegation or were not currently practising as a child protection worker for MCFD (being on leave for a period of at least six months including the period of the study, practising in another position or having left the Ministry). I removed the names of these 26 workers from the sampling frame. This left 824 workers as the final population for the study.  The recruitment letter I had emailed to all 850 workers contained the live link to the online survey. It also invited any currently practising child protection worker who was interested in being interviewed for the study to contact me. The last question of the online survey also invited workers interested in being interviewed to indicate their interest and to provide their contact information.  The online survey was open for one month. After two weeks I sent a reminder email to the study population of 824 workers in which I also thanked those who had already participated. In the final week of the survey I sent another reminder email only to those who had not yet participated. In each email I made it clear that their responses were valued and would be used. The intent was to maximise response rates by making multiple contacts with potential   83  respondents but to stop short of a level of contact that might become annoying (Dillman, 2007; Nulty, 2008; Shannon, Johnson, Searcy, & Lott, 2002).  Interviews Sixty nine workers who had completed the survey indicated via the final survey question their interest in being interviewed. In addition, one person who had not completed the survey emailed me to express interest in participating in an interview. I contacted 64 workers to thank them for their interest, and to send them further information regarding consent (See Appendix B). I was not able to make contact with the remaining six workers. Sixty percent (n = 42) of the workers who had expressed interest indicated their consent to being interviewed.  From this convenience sample I used purposive sampling, beginning by ensuring that I interviewed workers representing the range of teams and range of responses to the survey question 'Strengths-Based Practice is hard to do in child protection work'. As most people willing to be interviewed appeared to be generally positive about SBP, I made sure to interview the one person who had expressed very negative views, in order to understand as wide a range of experiences with SBP as possible. When analysis of survey data indicated that years of SBP experience and age were important, I included these factors in my sampling criteria, choosing to prioritise contact with younger workers with less experience until there was more variation within the sample. I attempted to interview 36 workers. Of this group 24 completed interviews and 12 chose not to participate further in the study.   84   Delegated Aboriginal Agencies In British Columbia there are currently nine Aboriginal agencies providing full child protection services to the Aboriginal communities they serve. Known as Delegated Aboriginal Agencies, they have assumed child welfare responsibilities through delegation agreements with the Provincial Director of Child Protection, to whom they remain accountable.   MCFD staff provided me with a list of all C6 delegated workers in the Delegated Aboriginal Agencies and hoped that I would include these agencies in the study. I redesigned the online survey to include a question asking workers to identify whether they worked for a Delegated Aboriginal Agency. It was suggested to me that as these agencies fell under the remit of MCFD, support for their inclusion from the Provincial Director of Child Welfare and Aboriginal Services Branch meant that separate permission from these agencies may not be necessary. However, UBC ethics approval did not extend to these agencies and I did not wish to proceed without full consultation with these agencies and the communities they served. The timelines for this study were too short to allow for this consultation. For this reason, workers from the Delegated Aboriginal Agencies were not included in the study.  Consent  Full information about consent and the rights of research participants was contained in each of the three emails inviting workers to participate in the study. This included the information that consent to participate in the survey was indicated by completing the survey questions.  When workers contacted me regarding their interest in being interviewed I explained the consent process and emailed them the consent letter. Initially I advised that I would contact them   85  again after 48 hours. However, the large number of workers interested in being interviewed, and the difficulty of making contact, meant that I changed to asking workers to contact me once they had read the consent letter if they still wished to be interviewed. Some workers signed, scanned and returned the consent letter but due to difficulties experienced by some workers in doing this, I advised that an email from them indicating that they had read the consent letter and wished to proceed would suffice. At the beginning of all but the first interview I reminded participants of consent issues and of their ability to end the interview at any time.  I interviewed two participants twice. This was due to one highly experienced worker giving very detailed and comprehensive feedback that could only be fully explored with the help of a second interview, to which he was happy to consent. I contacted the second worker again after I discovered that it was not possible to retrieve the audio-recording of her interview. She had expressed a high level of commitment to participating in the research and I felt it only respectful to advise her of the situation and ask her how she wished it to be managed. She immediately chose to be interviewed again. I was later able to retrieve the data from her first interview. I asked these two participants for their consent when we initially discussed the idea of a second interview and again at the beginning of the second interview.  Confidentiality and Data Management The online survey was hosted by an independent survey company approved for UBC research. It used encryption software and stored all data in Canada in accordance with British Columbia's Freedom of Information and Protection of Privacy Act. No person at MCFD could access the raw survey data, which was encrypted and password protected. Notwithstanding this, the fact that participants were completing the survey using MCFD computers and in response to an email sent   86  to their work addresses made it essential that they trusted their information was confidential and that they could choose to remain anonymous. For this reason they were not required to give identifying information in order to complete the survey. Only the names of those people who identified themselves as interested in being interviewed were recorded in the downloaded data. I did not ask for demographic characteristics like gender, location and cultural identity. Although of potential interest, I assessed that these questions might be deemed too identifying and might discourage participation.  Interviewees determined the means of contact and the interview location. I only used the MCFD email system to arrange contact if participants had given permission for me to do so. When I emailed interviewees to invite their comments on my emergent findings, I advised them of the options of communicating by personal email or telephone if they were concerned about MCFD access to their email.   I downloaded the survey data into a document from which I removed all identifying information. I password protected this document before storing it on my personal computer. I assigned each participant an identification number and used this number in this and all other records. The audio recordings were transcribed by a professional transcriptionist who signed a confidentiality agreement. She removed all identifying information from the interview transcripts and deleted the recordings after transcription. I stored both the audio recordings and transcripts in a password protected file on my personal computer and deleted the recordings from the recorders.   I made one master copy of the data which I kept on an external hard drive in a locked filing cabinet. A master sheet linking the names of participants to their identification number was password protected and stored separately from all other records in a locked filing cabinet.   87    For the purposes of analysis I uploaded the anonymised interview data into the Atlast-ti (v.7) computer program (Atlas-ti, 2012) and the quantitative survey data into the PASW Statistics (v.18) program (SPSS, 2009).   Data Collection Online Survey As stated previously, the first method for data collection was an online survey. This method provided an efficient way to offer participation to the entire population of workers, all of whom had computer access and had to be computer-literate in order to carry out their job functions. Studies also suggest that participants tend to answer more honestly in online surveys than in researcher-administered telephone or interview surveys (Brace, 2008; Dillman, 2007). One disadvantage is that participants in online surveys have no opportunity to clarify misunderstandings and great care is needed to ensure instructions and questions are clear.  The link to the survey was emailed to all workers in the sampling frame. Three hundred and forty two workers, or 42% of the population of 824, clicked on the survey link. This brought them to a front page describing the purpose of the survey and asking only frontline workers to continue. Two hundred and twenty four workers, or 65% of those who had clicked the link, completed the survey.      Survey Design The survey consisted of 15 questions (see Appendix C). It was designed to take no more than 10 minutes to complete. The main research questions were addressed directly with survey questions asking workers to rate the frequency and duration of their SBP use and the rate of use of specific   88  SBP questions (miracle, scaling, exceptions and coping questions) commonly advocated as strengths-based child protection techniques. They were also addressed with open response questions asking workers to define SBP, describe their use of it in practice and describe supports for and barriers to its implementation. At no point were respondents given a definition of SBP and I left questions regarding specific techniques associated with SBP to the end so as to avoid leading participants towards a particular definition of SBP.  Further survey questions were intended to enable exploration of the extent to which the application of SBP ideas was related to variables theoretically linked to SBP implementation. The literature supporting these theoretical links is described in Chapter Two. The view of some SBP researchers that SBP implementation required developing worker capacity and training (Antle et al., 2008; Idzelis Rothe et al., 2013; Lietz, 2011; Skrypek et al., 2010) suggested that the level of educational qualification and SBP knowledge might be important factors in the application of SBP ideas. The link between worker maturity and their capacity to undertake complex relational work (Perlman, 1979) suggested that age might also be relevant. A historical understanding of the changing conceptualisation of child protection work suggested that years of child protection experience might be important to the exposure and response of workers to different models of child protection practice. Evidence of the importance of the immediate organisational context for SBP (Barbee et al., 2011; Idzelis Rothe et al., 2013; Pipkin et al., 2013; Turnell, 2012) prompted a question asking workers to identify their team type. Finally, the suggestion that implementation of SBP requires workers to make a significant attitudinal commitment to the approach (Blundo, 2001, 2012; C. Rapp et al., 2006) led to the eight Likert-type response questions, discussed in the following section, regarding worker attitudes.    89   To increase the likelihood that workers would complete the survey, I did not ask for information related to demographic characteristics like gender, location and ethnicity. This was because they may have deterred participation by being perceived as identifying, especially when analysed in conjunction with characteristics like age and team type for which there was stronger theoretical support. I also followed Brace's recommendations (2008) to make survey completion as easy as possible. I included several skips to avoid asking participants questions that were not relevant. The questions were laid out over several pages to avoid scrolling, and included a progress bar which in short surveys tends to motivate participants to continue.  I started with general questions that were relatively easy to answer. Questions that were potentially more threatening as they required a greater degree of self-disclosure or discussion of practice challenges came later, and were interspersed with easier questions in order to maintain participant interest.   Attitude Questions I designed a series of simple statements about SBP which were intended to elicit the extent to which participants perceived SBP to be appropriate, possible, effective and a good approach for child protection work (see Table 1). These statements were informed by the literature reviewed in Chapter Two and the pragmatist perspective that SBP ideas would be applied by frontline workers to the extent to which they were perceived as useful and workable. The fact that SBP originated in a different field of practice and has needed adaptation for child protection work led to the statements regarding its appropriateness. Evidence of inconsistent implementation and of the importance of training and support suggested the value of asking child protection workers whether they saw the approach as within their capacity to perform. Both the historical overview   90  of the child protection worker-client relationship and comparisons between contemporary approaches like relational social work and approaches specific to mandated clients illustrated competing perspectives on best practices with adult clients. This suggested that the question of SBP's effectiveness was unlikely to be a settled issue for frontline workers. Finally, claims that SBP required a radical shift in individual thinking and practice made it important to establish the extent of worker support for the approach and analyse the relative importance of this support.  Table 1 Attitude Questions Is SBP... Statement Appropriate? 1. Strengths-Based Practice is appropriate in every situation  2. With some clients Strengths-Based Practice is not the right approach Possible? 3. It is always possible to do Strengths-Based Practice 4. Strengths-Based Practice is hard to do in child protection work Effective? 5. Children are left at greater risk when protection workers do Strengths-Based Practice 6. Doing Strengths-Based Practice increases the chances I will be successful with my clients A Good  Idea? 7. Strengths-Based Practice is a good approach for child protection work 8. I would like to do Strengths-Based Practice more with my clients   Participants were asked to indicate on a Likert-type scale whether they strongly disagreed, disagreed, neither agreed nor disagreed, agreed or strongly agreed with each statement. As participants tend to answer to the left of scales and to agree with statements when presented with Likert-type responses (Brace, 2008), I placed the 'disagree' option to the left of the scale to maximise the extent to which these tendencies might cancel each other out. I also included positive and negative statements of most attitudes and randomised the order of the statements to mitigate the effect of patterned responses (Brace, 2008; Dillman, 2007).   91   A mid-point Likert-type option like "neither agree nor disagree" can attract responses from those who are ambivalent, who are indifferent, who wish to minimise effort and who wish to avoid expressing an undesirable view (Johns, 2005). The presence of the latter two groups in particular raises questions regarding the extent to which it can be seen as the midpoint measure of a continuous underlying negative?to?positive dimension. However, in previous research (Oliver, 2010) child protection workers had told me of their frustration about being forced to adopt a false positive or negative stance due to the absence of a middle option. Johns (2005) found that when the majority of potential mid-point responders are motivated by ambivalence or indifference, removing the option of a midpoint increases both frustration and measurement error. I had taken steps to minimise both the effort required to complete the survey and the pressure on respondents to answer in socially desirable ways. When the response 'neither agree nor disagree' did not appear problematic during piloting of the survey I decided to include it to represent a middle position of having no strong view one way or another on a scale that went from a strong negative view to a strong positive view.   Think Aloud Interviews I initially intended to pilot the survey with at least 10 MCFD child protection workers. However, senior management at MCFD preferred that they review the survey and pass back their feedback to me in a conference call. Their feedback was generally positive and initiated no specific changes to the survey. The survey was piloted by my four committee members. I also elicited feedback regarding the survey design by completing Think Aloud interviews (Ericsson & Simon, 1993) with two people who had considerable recent experience of frontline child protection work   92  with MCFD. They were recruited through personal contacts and an email to UBC social work students with recent MCFD experience.  The Think Aloud process was initially developed to explore cognitive operations like information recall and problem solving (Ericsson & Simon, 1993). It has become a common method in product usability testing (McDonald, Edwards, & Zhao, 2012) and instrument development (Collins, 2003). It involves participants verbalising their thinking as they work their way through the survey or test. The researcher uses prompts like 'keep talking' and 'what are you thinking now?' to encourage participants to become aware of their thoughts about the survey or test and to immediately express them.   Some participants find it difficult to concurrently problem solve and verbalise their cognitive processes and this approach works less well for complex problems and with people who have cognitive limitations (Johnstone, Bottsford-Miller, & Thompson, 2006). However, it is generally agreed to be an efficient, effective and user-friendly means of detecting design problems in tests (Collins, 2003; McDonald et al., 2012). It requires few participants as it tends to produce more detailed information than standard test pilots (Johnstone et al., 2006).   I audio recorded the Think Aloud pilots. After each participant had worked their way through the online survey while expressing their thoughts, I asked them to reflect on the experience. I asked them to talk about the information they felt the survey was intended to elicit and any problems they had encountered. As a result of these interviews I made changes to the survey, for instance breaking down the question 'What do you do with clients that could be described as SBP?' into the two separate questions of  'How do you define SBP?' and 'What would someone watching you see you do with clients that you would describe as SBP?'     93  Interviews The second method of data collection was interviews with practising child protection workers. Interviewing is an easily understood and accessible strategy and  ?interviews are particularly well-suited for studying people?s understanding of the meanings in their lived world, describing their experiences and self-understanding, and clarifying and elaborating their own perspective on their lived world? (Kvale & Brinkmann, 2008, p. 116).  Interviews elicit the interviewee's subjective experiences and these are influenced by the relationship with the interviewer and the context of, and expectations regarding, the interview. There is always the risk that interviews say more about culturally accepted discourses than about the interviewee's perspective (Thorne, 2008). With this in mind I saw my role as being to create an atmosphere in which interviewees might feel as free as possible to fully describe their understandings and experiences.  I began by offering interviewees as much control as possible over the timing and form of their interview. Most chose to be interviewed during the working day although some chose evenings and weekends. I offered interviewees the choice of talking via telephone, Skype, or, for those who were in the Lower Mainland, a face-to-face meeting.  Four people chose to meet, two at their office, one at a local coffee shop and one at my home. The remainder chose telephone interviews. I was surprised at the number of people who chose to be interviewed at work, and checked with them before beginning the interview whether they felt they had sufficient privacy to speak openly with me. Three of the interviews were briefly interrupted by the need for workers to attend to colleagues or clients.   The mean time for the interviews was 74 minutes, with the shortest interview lasting 35 minutes and the longest 127 minutes. The 35 minute interview was considerably shorter, and provided less comprehensive data, than the other interviews. However this was all the time the   94  participant was able to offer, and I felt it important that I respect the contribution of this participant, and the busy nature of child protection work, by including this data. As stated previously, I interviewed two people twice.   I began the interviews by thanking workers for their participation, reviewing consent and asking if they had any questions. I explained the purpose of the interviews as being to explore whether and how frontline workers understood and used SBP in their child protection work. I stated my belief in the importance of the perspectives of frontline workers, and that I was interested not in how others told them they should work or in ideal practice but in the reality of whether and how SBP worked for them in their daily work. I told them that I understood the importance of anonymity and described how this would be preserved. In taking this approach I hoped as much as possible to create an environment in which participants felt free to speak openly.    I used an interview guide (see Appendix D) to help me focus the interviews on the research questions. I hoped that by opening with a question about the practice of other workers I would elicit descriptions of SBP being enacted while enabling the participant to 'settle' into the interview before asking what might be perceived as more intrusive questions about personal practice. Informed by pragmatism, I tried to approach each research question in multiple ways. To elicit interviewee understandings of SBP I asked about MCFD's motivation for the approach and sought practice examples from the interviewee's own and others' work. To explore perceived supports and barriers I asked about them directly, but also sought specific practice examples in which SBP was easy or inapplicable and asked about personal qualities and experiences that led interviewees to relate to clients in the ways that they did. Motivated by the idea, outlined in Chapter Two, that the strengths-based worker-client relationship in child protection is complex   95  and the site of competing perspectives about the proper role of the worker, I included several questions designed to help interviewees reflect on this relationship.    I tried to ask all questions on this guide for the first interviews to guard against closing down avenues of inquiry by prematurely judging what constituted important information. About halfway through the series of interviews I adapted the questions in order to explore emergent concepts, for instance asking about high risk situations and about whether participants had felt anxious or fearful with clients.   I attempted to adopt the role of "an encouraging and judgmentally neutral facilitator so that an individual can explain him or herself as fully as possible" (Thorne, 2008, p. 129). I often prompted participants to go into more detail or to give examples. This was both to gain the best understanding I could of the interviewee's experience and to try to move beyond common discourses about strengths-based and child protection practice. Even so, there were times when I realised after the interview that I had assumed a shared understanding of common child protection words or issues that might not exist. I found it challenging to suspend my judgements about good practice to move into the role of inquirer. These issues are discussed as typical of interviewers who have previously shared the practice role of their interviewees (Hunt, 2011; Thorne, 2008). I found it helpful to reflect before each of the interviews on the importance of challenging my own preconceptions, to frame interviewees as experts of their own experience and to give myself visual prompts to remain curious during the interviews.  In pragmatist research there is always a different perspective to be taken on the data and the contribution of each new participant has the potential to change, challenge or expand the conceptualisation. However, I was aware of my limited resources and my commitment to provide research findings to MCFD in a timely manner. There came a point at which I felt that I   96  had sufficient data on which to base an analysis that was both well grounded in participant's experiences and threw new light on the challenges associated with SBP.  This is seen as the point in an interpretive description study at which data collection can reasonably end (Thorne, 2008).  Data Analysis Plan for Statistical Analysis of Survey Data I began the statistical analysis of the survey questions outlined in Table 2 shortly after the survey closed and continued throughout the months in which I was interviewing workers. I started by conducting descriptive statistics for each variable. Team membership and qualification were treated as categorical variables, while all other variables, including the Likert-type response items, were treated as continuous. I generated a frequency table for each variable and a histogram, boxplot, mean, median, mode, standard deviation, variance and range for the continuous variables. These summaries of the data enabled me to describe the demographic characteristics of survey participants and to begin to answer the questions of whether and how child protection workers applied SBP ideas.   As a preliminary to conducting inferential analysis, I visually assessed the normality of each distribution of the continuous variables using P-P plots and histograms, and I calculated the z score for skewness and kurtosis. I also conducted Kolomogorov-Smirnov and Shapiro-Wilks tests to compare the distribution of scores in the sample with those in a normal distribution with the same mean and standard deviation. However, as the large sample size meant that small deviations from normality produced significant findings on these last two tests, I relied more on the visual assessment and absolute value of the z scores for skewness in deciding whether the assumption of normality for parametric tests was sufficiently met (A. Field, 2009).      97  Table 2 Questions for Statistical Analysis Questions  Response Options What is your age?  Years For how long have you done SBP?  Years What is the total number of years you have worked as a child protection worker with full authority to investigate and remove children?   Years What best describes your current team?   Intake, Family Service, Family Development Response, Integrated, Other What qualification do you hold?  Bachelor of Social Work, Master of Social Work, Bachelor of Child and Youth Care, Master of Child and Youth Care, Other How much do you know about Strengths-Based Practice?  Nothing, A little, A fair amount, A good amount, A lot How often do you do Strengths-Based Practice in your child protection work?   Never, Occasionally, Some of the time, Most of the time, Always  To what extent do you agree or disagree with the following statements: ? Strengths-Based Practice is appropriate in every situation ? With some clients Strengths-Based Practice is not the right approach ? It is always possible to do Strengths-Based Practice ? Strengths-Based Practice is hard to do in child protection work. ? Children are left at greater risk when protection workers do Strengths-Based Practice ? Doing Strengths-Based Practice increases the chances I will be successful with my clients ? Strengths-Based Practice is a good approach for child protection work ? I would like to do Strengths-Based Practice more with my clients   For each statement: Strongly disagree, Disagree, Neither agree nor disagree, Agree, Strongly agree  Thinking of the last 10 times you talked with a parent on your child protection caseload, how many times did you use the following: ? Coping question ? Scaling question ? Exceptions question ? Miracle question    For each question: Not used, Used once or twice, Used three or four times, Used five or more times, I don't know what this is   98   I firstly examined whether there were significant differences in age or child protection experience across workers from different teams or holding different qualifications. While this analysis was not directly related to the research questions, it enabled me to assess whether any later findings of relationships between the application of SBP ideas and team type or qualification were likely to be confounded by these variables. In this and all subsequent analyses including the team variable, the team category 'Other' was coded as missing due to its small size. I chose one-way Analysis of Variance (ANOVA) tests when, as for worker age, the z score for skewness was not greater than the absolute value of three and Levene's test showed that there was not significant variance across groups. This enabled posthoc tests to be conducted and interpreted. When these conditions were not met, as with years of child protection experience, I chose ANOVA's non-parametric equivalent, the Kruskal-Wallis test.  The goal of the remaining statistical analysis described in this section was to explore for any patterns in the distribution of the variables pertaining to the understanding and application of SBP ideas, namely self-reported frequency and duration of SBP use, use of the questioning techniques commonly associated with SBP in child protection, and self-reported SBP knowledge and attitudes towards SBP. I was specifically interested in whether any of these variables were related to each other or to the worker characteristics of age, length of child protection experience, team membership and level of qualification. I hoped that this exploration would help me answer the main research questions of whether and how workers applied SBP ideas.  I conducted correlations to explore associations between pairings of the continuous variables. Carifio and Perla (2007) maintain that Likert-type responses can be analysed using parametric tests like Pearson's correlation and ANOVA, as these tests are robust to violations of the interval data assumption and to moderate skew. For the questions with Likert-type responses   99  asking participants to rate how often they used the coping, miracle, scaling and exceptions questions, I combined the 'I do not know what this is' response and the 'I do not use this technique' response to create a continuous scale of self-reported use of these techniques, ranging from 'no use' to 'used more than five times'. Two-tailed Pearson's Product Moment correlations were conducted when the z score for skew and kurtosis for both variables was less than the absolute value of three. In all other cases Spearman's correlation was used as it assesses the relationship between ranks of scores and does not require that variables be normally distributed.    One-way ANOVAs were conducted to explore differences across teams and levels of qualification with respect to knowledge of SBP and frequency of use, the use of the scaling and exceptions questions and all attitudes except the attitudes that SBP is a good approach for child protection and SBP leaves children at greater risk. When a significant difference across groups was indicated I conducted Hochberg's GT2 post hoc test to locate the difference. This post hoc procedure is recommended when Levene's test indicates that unequal groups have homogenous variance (A. Field, 2009).   The non-parametric equivalent to ANOVA, the Kruskal-Wallis test, was conducted for the remaining five variables related to SBP, either because the z score for skewness was greater than the absolute value of three (for SBP experience, the attitude that SBP is a good approach for child protection, the attitude that SBP leaves children at greater risk and use of coping questions), or because the Levene's test showed unequal variance across groups (use of the miracle question). When the Kruskal-Wallis test indicated significant differences across the teams or qualification levels, post hoc Mann Whitney U tests were selected to examine where these differences lay. A Bonferroni correction, dividing the critical value of .05 between each test, controlled the Type 1 error rate.   100    The distribution and content of responses to the survey questions regarding the applicability of SBP suggested that the distinction between those who saw SBP as an approach that could be used all the time and those who saw it as having more limited applicability was an important one. This finding prompted me to code survey respondents into two groups based on this distinction in order to further explore patterns in the perception of SBP's applicability. The 'Sometimes' group (n = 128) consisted of participants who indicated disagreement to the question 'It is always possible to do SBP' or 'SBP is appropriate in every situation' or agreement to the question 'With some clients SBP is not the right approach' or who clearly indicated in their responses to open-ended survey questions that there were situations in which SBP was not appropriate or possible. The 'Always' group (n = 54) answered in the opposite direction on all three questions above and in the open-ended survey questions did not identify any situations in which SBP was not possible, not appropriate or particularly challenging. Those who appeared to fall into the 'Always' group but who responded 'Neither agree nor disagree' to any of the three questions or who, in their responses to other survey questions, named groups or situations for which SBP was particularly challenging, were coded as missing (n = 42). The latter were coded as missing because in the interviews and responses to open-ended survey questions some participants conflated the idea that SBP was challenging for particular clients with the idea that SBP could not be done with those clients.   The end result was that the 'Always' group consisted only of people who had clearly indicated that SBP was always applicable and had given no information that might contradict this position. The recoding addressed the problem that responses to the three relevant survey questions under-represented the perceived limits on SBP as some people who responded that   101  SBP was always possible or appropriate went on later in the survey to identify situations in which it was clearly not.  I conducted t-tests to explore whether workers in the 'Sometimes' group differed from those in the 'Always' group in their mean age, self-assessed knowledge and use of SBP or desire to do SBP more with clients. T-tests were chosen for these analyses because the z score for skew in each group was under the absolute value of three and Levene's tests indicated homogeneity of variance across the two groups. For the variables where the z score for skew was greater than the absolute value of three (years of SBP experience, child protection experience and the attitudes that SBP is hard, is a good approach for child protection, increases the chances of success and puts children at greater risk) I conducted Mann Whitney U tests. These are the non-parametric equivalent to t-tests, requiring that the distribution in the two groups be the same shape, but not that it be normal. I conducted chi-square tests for differences in these groups across team and qualification categories.   Qualitative Analysis of Interview and Survey Data There were two sources of qualitative data: the interviews and the written responses to the following open-ended survey questions: 1) How do you define SBP? 2) What would someone watching you see you do with clients that you would describe as SBP ?  3) What do you find most challenging about SBP? 4) What supports do you need to do SBP?  5) In which situations, or for which clients, is SBP not appropriate?   102  I began my analysis of both data sources after completing my preliminary statistical analysis of the survey data and shortly after the first interview. Initially, the process for analysing the qualitative data from the survey and the interviews was very similar. I loaded into Atlas-ti the responses relating to each of the open-ended survey questions, and the transcripts of each interview. Taking each open-ended survey question in turn, I read the compiled responses several times. I also read each interview transcript several times and listened again to the audio recording. This helped me get a sense of each interview and the array of responses to each question as a whole, and was something I did several times during the analysis process to remind myself of how the codes in the data related to the larger context.    I started by coding each "meaning unit" (Thorne, 2008, p. 145), or what I perceived to be each distinct idea, which were often marked by breaks or punctuation in the text. These codes described what I understood to be the overt meaning of that data unit and I tried to keep close to participant's words, frequently using them as codes. For instance, when participants talked of 'going onto autopilot' or 'having a hard shell' these phrases became codes. This first sweep of coding resulted in a variety of seemingly unrelated codes like 'clients with mental illness', 'it's not personal' and 'we work in isolation'.   After this initial coding I re-read the transcripts and survey responses, coding specifically for workers' descriptions of their actions, thoughts and feelings and information that related directly to the questions of what actions and meanings workers attached to SBP and what were the supports and barriers to its use. This resulted in codes like 'SBP is asking questions',  'thinking outside the box', 'SBP feels good', 'learn from training' and 'clients do not trust it'. I also did what Maxwell (2013) calls contiguous coding to capture relationships in time and space between ideas. For instance, I noticed that workers often talked about 'strengths' and 'risks' in the   103  same sentence, with some describing these concepts as opposites and others as complementary. I created codes like 'as well as' and 'instead of' and 'mitigates' to capture the different relationships between the two concepts.  In addition to these detailed codes I used some very broad codes to group together larger pieces of text. These codes are described by Maxwell (2013) as organizational in that they "function primarily as bins for sorting the data into further analysis" (p. 107).  I used them to code common ideas like 'transparency' that were used by workers as if they had a shared and unproblematic meaning and to code broad topics that were directly related to the research questions like 'supports' and 'challenges'. I also used the codes of 'Sometimes' and 'Always', derived from the statistical analysis described above, to code the interviewees into two groups according to whether they described  SBP was a practice that was always or only sometimes applicable.   Following the initial coding I went through an iterative process of comparing and contrasting coded data pieces to each other and to the entire data set to analyse for similarities and differences. This process of constant comparative analysis is drawn from the grounded theory tradition (Corbin & Strauss, 2008) and is an important tool in interpretive description studies. It enables researchers to develop descriptions of each code and to analyse relationships between and within data pieces. It provides a way to move from simple descriptions of content to more conceptual analysis, or "from data to pattern and from pattern to relationship" (Thorne, 2008, p. 158). It enabled me to group similar codes into progressively larger and more conceptual categories, for instance linking codes like 'observing others' and 'religious values inform SBP' into the larger category of 'learn from' which ultimately fell under the broad theme of 'supports'. I also worked in the other direction, breaking broad categories down into smaller   104  ones with distinct properties. For instance, within the broad category of 'transparency', I noticed a set of comments about the object of transparency and a set about the consequences of transparency. This I broke down into a series of distinct but related codes like 'transparent about expectations', 'transparent about process' and 'transparent creates trust'.    Finally, I also began to visually map the codes. This provided me with a different perspective on the data and helped me to explore relationships between the different levels of codes and categories. I found that this kind of visual representation helped move my conceptualisation forward because, as Corbin and Strauss (2008) describe, "most of all, doing diagrams force a researcher to think about the data in "lean ways"; that is in a manner that reduces the data to their essence" (p. 125).  As the analysis progressed I began to construct themes, including the five definitions of SBP outlined in Chapter Six. To further develop my analysis I asked interviewees to reflect on these five definitions and to identify the definition that best fit the way they practised SBP. Their responses offered some confirmation that my tentative conceptualisation was perceived to be relevant and provided important information to develop both my description of the five definitions and to expand my thinking about the possible developmental relationships between them. These five definitions of SBP then became an organising framework within which to analyse supports and barriers to the implementation of these particular definitions of SBP. A second set of themes, described in Chapter Six, was related to two additional implementation issues (management support and fear) which were pertinent to all the definitional groups.         105  Quantitizing (Quantifying) the Qualitative Data I quantitized (Sandelowski et al., 2009) some of the qualitative data from the surveys and interviews when it enabled me to concisely summarise responses without losing in the process what I perceived to be important information. For instance, interviewees offered a number reasons for their fearlessness, but either did not elaborate on these during the interview or provided contextual information that was too identifying for me to present in narrative form. I therefore coded the reasons and counted their frequency to summarise participant ideas on this issue. I used this method again to summarise responses to the open-ended questions about supports, barriers and situations in which SBP was not appropriate. The responses to the open-ended survey questions varied widely in length, ranging from one-word answers to paragraphs of text. They were often in the form of lists of actions or ideas. Quantitizing these responses seemed to be the most effective way of summarising information which was important, but lacked the context and detail necessary for deeper qualitative analysis.   Through the interpretive description coding process previously described I constructed a set of descriptive codes relating to each of the open-ended survey questions. For instance, for the question regarding what participants found most challenging I coded the response "Not having the time or the resources to properly do the work" with the two codes 'lack of resources' and 'lack of time'. I compiled a list of the codes and outlined the properties of each code to create a coding scheme. I then went through the set of responses for each question several times in order to code consistently according to this scheme before counting the frequency of each code. This analysis resulted in a relatively simple set of findings, presented in Chapter Five, which outlined what workers perceived to be the situations in which SBP was inappropriate, the barriers to be overcome and the supports needed if SBP was to be implemented.   106   My previous analysis of the qualitative interview and survey data and of the Likert-type response questions about SBP's applicability, suggested that workers who thought that SBP was always applicable might define SBP differently from those who thought it only sometimes applicable. To explore this distinction further, I visually compared the frequency of codes for respondents in the 'Always' group with those in the 'Sometimes' group. I also conducted a chi-square analysis to assess whether there was a statistically significant difference in the challenges identified across these two groups, as the coding for the question about challenges had created sufficiently large groups to support this type of analysis.  Finally, I attempted to categorise all survey respondents into one of the five definitional groups described above, based on their responses to the open-ended survey questions 'How do you define SBP?' and 'What would someone watching you see you do with clients that you would describe as SBP?'. My hope was that I might establish the frequency of each definition of SBP and then use tests like ANOVA to examine differences across these definitional groups. However, I concluded that the data from these two questions was not sufficiently thick to support this categorisation. The survey did not include any questions specifically designed to distinguish between these definitional groups as I had constructed the five definitions of SBP from my qualitative analysis after the survey was finished. The five definitions were characterised by different degrees of complexity, with the more complex definitions integrating ideas from simpler definitions. While survey responses to questions as to how participants defined SBP varied considerably in length, there was no way to assess whether this was related to how workers perceived SBP or to other factors. Data for coding the more complex definitions was particularly thin. I assessed that this made the gathering of evidence for construct validity too problematic to proceed.   107  Quality There is as not yet consensus as to what constitutes appropriate criteria by which to assess a good mixed methods study (O'Cathain, 2010; Onwuegbuzie & Johnson, 2006). Several mixed methods theorists have suggested that distinct quality criteria are needed to assess unique features of mixed methods studies like how methods are combined and analysis integrated (O'Cathain, 2010; Onwuegbuzie & Johnson, 2006; Tashakkori & Teddlie, 2008). Criteria proposed specifically for mixed methods incorporate criteria from quantitative and qualitative traditions and address the extent to which studies are conducted in ways that enable the researcher to answer the research question in a logical, transparent and useful manner, that respect the strengths and limitations of each method and that take into account the nature of the knowledge each method provides. O'Cathain (2010) collated these proposed criteria, which are summarised in Figure 1.   I endeavoured throughout the study to meet the quality criteria in Figure 1. My attempts were motivated to a large extent by my pragmatic goal that my findings be useful and relevant for child protection workers and their clients. This goal led me many times to reconsider my method, data and analysis in an attempt to be as rigorous and open to the experience of participants as possible. It encouraged me to make decisions about how to proceed through the study in a way that was logically linked to answering the research questions and respected the complexities of child protection practice. It also kept me alive to the possibility of contradictory or surprising information. While an important quality criteria was that inferences be consistent with some existing knowledge or theory, the question of which existing knowledge or theory could not be predetermined if I was to remain open to inferences that contradicted accepted ways of thinking about practice.   108   Design All aspects of the design are appropriate for answering the research questions and consistent with the epistemological standpoint There is clear and detailed information about all elements of the study The choice of mixed methods is justified The strengths and weaknesses of methods are considered to avoid overlapping weaknesses  The study is situated within and informed by a comprehensible critical review of the literature Implementation  Methods are implemented with rigor Methods are implemented in a way that remains true to the study design Sampling technique and size for each method are adequate Analytic techniques, including those that integrate qualitative and quantitative data, are appropriate for the research question and performed properly Interpretation  It is clear which methods produced which findings Inferences are consistent with the findings on which they are based  Inferences are consistent with current knowledge or theory Others are likely to reach the same conclusion based on the findings and methods presented Inferences are more credible than other conclusions considered Meta-inferences adequately incorporate inferences from the qualitative and quantitative findings Inconsistencies between inferences are explained Inferences correspond to the research questions and purpose Translation The findings can be transferred to some other context, group or time  The findings are usable Reporting of findings is accessible and clear  Figure 1 Quality Criteria for Mixed Methods Studies (adapted from O'Cathain, 2010)     I was also influenced by Thorne's (2008) criteria for a good interpretive description study. This is very similar to the criteria outlined above. She adds the requirement that the study contribute to disciplinary knowledge, be morally defensible and take into account the possibility that findings will be put into practice. I considered that any findings that might inform practice needed to be logically justifiable, rooted in participants' experiences rather than my own preconceptions and provide a perspective that might help address the challenges of SBP in child protection work.   I have endeavoured to make all sections of this dissertation sufficiently comprehensive and detailed to make accessible the reasons for my methodological choices and inferences. This   109  enables readers to assess the extent to which they might reach similar conclusions and these conclusions are more credible than alternatives. I have reported discrepant data and inconsistent or non-significant findings and attempted in my sampling for interviews to include a wide variety of viewpoints.   Throughout the study I sought to engage in a reflexive process of journaling to examine my own biases and assumptions. I wrote regular memos to capture my starting assumptions, evolving perspectives and questions to consider in future interviews and analysis. For instance, after each interview I wrote a memo documenting a synopsis of the interview and my impressions and initial thoughts about both its content and process. I found that this helped me to identify what had stood out for me about the interview and made me more able to move beyond these first impressions.  The memos became part of the "audit trail" (Thorne, 2008) in which I tried to lay out the theoretical assumptions and personal perspectives that informed my inquiry.   I endeavoured to implement both the quantitative and qualitative components of this study as rigorously as possible. I attempted to address threats to credibility, although was unable to eliminate these completely. Three examples are outlined below:  1) In the explanatory emails for the survey I was clear that the study was intended for all frontline child protection workers, not only those who had done SBP or who had particularly strong views on the practice. However, it is possible that the findings were affected by unknown sources of sampling bias and in particular that respondents were disproportionately people who were particularly invested in SBP.   2) I attempted to create for interviewees an environment in which they would feel able to talk openly about their experiences. It is likely however that their responses were constrained by such contextual elements as their assessment of the expectations held by MCFD and myself. This   110  is particularly a concern because many chose to be interviewed in their place of work and during work hours.  3) I endeavoured to follow principles of good survey design and the pilots I conducted suggested that the survey had face validity. However in the absence of a more extensive validation process, the extent to which respondents understood questions in the way I had intended and the impact of any measurement error are hard to assess. It appears that some elements, like the definition of an integrated team, were not clear.   In light of the fact that it was not possible for me to account for ways in which these factors might have influenced the findings, my credibility depends on understanding the limits of the knowledge gained by each method and ensuring my claims do not extend beyond these limits. The survey question asking participants how often they did SBP for instance, produced data not about the actual occurrence of SBP but about the perceived occurrence of what participants defined as SBP. Particularly in light of the findings regarding the varied definitions of SBP, this is an important distinction. Even the clearest findings from this study should be viewed as contextually limited and fallible. While they may lead to suggestions for action, these should be seen in the light of the pragmatist commitment to ongoing assessment and critique.   Finally, in order to assess the extent to which my findings were perceived to represent a perspective that might be helpful, I shared with all interviewees a synopsis of my analysis and asked them to identify which, if any, of the definitions represented their understanding of SBP. I also discussed the findings at a forum of child protection consultants convened by MCFD, and asked participants to identify the definition of SBP that was most appropriate for their work. All participants who responded to these requests appeared able to locate their standpoint within the analysis and many described the analysis as helpful or resonant. This suggests that my emergent   111  conceptualisation might have value, although the true test of this will only become clear if the findings inform practice.  Summary This was a mixed methods study informed by a pragmatist epistemology. The population of interest consisted of 824 currently practising and fully delegated frontline child protection workers currently employed by MCFD. The sample of 225 workers was recruited through emails sent to the entire population. I collected data from 224 participants via an online survey which was designed for this study and tested using Think Aloud interviews. I also interviewed 24 participants. The survey respondents constituted a convenience sample and the interviewees were recruited through a combination of convenience and purposeful sampling. The quantitative data was analysed using descriptive and inferential statistics. The qualitative data was analysed using inductive coding and constant comparative techniques within an interpretive description approach. I also quantitized some qualitative survey data, before conducting descriptive and inferential statistical analysis. Components of the resulting analysis were presented to interviewees and to key MCFD staff for feedback regarding its resonance and relevance.      112  CHAPTER 4 : DESCRIPTIVE INFORMATION This study took place within the Ministry for Children and Family Development (MCFD), the main statutory child welfare agency in British Columbia, Canada. MCFD was a relatively early adopter of SBP, having first introduced the approach in 2003. This is two years earlier than the Minnesota jurisdiction of Carver County which, with Olmstead County, are often identified as two of the longest-term implementers of the Signs of Safety framework anywhere in the world. This long history with the model provides an opportunity to trace the dynamics involved in implementing a Child Protection intervention strategy such as Signs of Safety, and identify lessons learned for other jurisdictions interested in doing so. (Idzelis Rothe et al., 2013, p. 6)     MCFD has invested in province-wide training to support the approach and since 2008 has expected all protection workers to implement SBP at all stages of case management. The organisation has had time to embed SBP into its policies and expose workers to its ideas. This made it less likely that a study into whether and how child protection practitioners applied SBP ideas in their frontline work would pull up short with the simple conclusion that the approach had not yet been sufficiently propagated, and increased the chances that it might say something about the fit of the approach with the conditions and goals of contemporary child protection practice.    British Columbia Ministry for Children and Family Development (MCFD) The Ministry for Children and Family Development is a branch of the provincial government with 429 offices located across British Columbia (Charles, Oliver, Lach, Torrans, & Dudding,   113  2012). Its staff provide full child protection services, including assessment and investigation, services to support families and to mitigate and monitor child protection concerns, and guardianship and adoption services to children in its legal care. These services are provided to all children in the province, with the exception of those Aboriginal and Metis children who are served by Delegated Agencies. These are agencies who have signed Delegation Agreements with the Provincial Director of Child Protection which enables them to deliver services to a specified Aboriginal or Metis community. Currently nine Delegated Agencies are delegated to hold full child protection responsibilities while another 14 have partial delegation to provide some child welfare services.   MCFD introduced SBP in 2003 (Ministry for Children and Family Development, 2004a). when the first Family Development Response (FDR) teams were created to offer a differential response to child protection concerns (Marshall et al., 2010). This was intended to be a more collaborative response than traditional child protection investigations, focussed on negotiated goals and using the principles of SBP. FDR teams received training in strengths-based questioning and in the Signs of Safety model (Ministry for Children and Family Development, 2004c). MCFD practice standards required those delivering FDR to conduct a "strengths-based assessment of the family's capacity" (Ministry for Children and Family Development, 2004b). Written material supporting the new policy of FDR referenced the strengths-based work of Turnell, Edwards and Saleeby (Ministry for Children and Family Development, 2004c).    The following decade saw SBP promoted as the preferred response to child protection concerns on Intake, Family Services and Integrated teams in addition to the FDR teams. By 2008 it was expected that workers on all child protection teams would implement the approach (J. Wale, personal communication, July 31, 2013). There were several initiatives to train workers in   114  SBP. The following course titles and descriptions were taken from the Learning Management System, the central database of  training for MCFD employees. This is not a complete record of all training offered and it is likely that more training was offered over this period (J. Morais, personal communication, August 29, 2013).   In 2006 50 team leaders received training in "Collaborative, Strength Based Practice". In 2008 SBP training was included in the core training for newly hired child protection staff. A course description stated that during the two day course "participants will identify the characteristics of a strengths-based perspective when working with families and children and how to identify family strengths?. Some regions offered classes in skill-building to support the focus on strengths-based, collaborative practice.  A new initiative to train workers in SBP was launched in 2011, although there was variation in the type and duration of training that was offered across the administrative districts. One region offered 10 sessions of ?Collaborative Response Model?, which was described as focussing on strengths-based practice and skills development. A second region conducted nine two-day workshops that "focussed on SBP and skills development where participants are given  an overview of how strengths-based practice supports current practice shifts in terms of expanding Family Development Response". A third region offered two sessions of a course titled ?FDR Strengths Based Practice? which addressed how "we implement strengths-based practice in our work and the tools that encourage strengths-based practice, including scaling questions, searching for exceptions, discovering family strengths and the miracle question". In 2011 nine groups of new child protection workers received a two day course in SBP as part of the core training they received when first hired. A course description stated that "the course emphasizes interventions that tap into client resources, talents, knowledge, motivation, and environmental   115  assets. The strengths-based approach to working with families aims to support and strengthen family functioning by identifying and building upon a family?s existing strengths". Seven more sessions of this training ran in 2012 and 2013.  There have been a number of updates to policies in this area since 2003. The current Child Protection Response Policy (Ministry for Children and Family Development, 2012) describes expectations that workers implement a strengths-based approach at all stages in the child protection process. They are, for instance, expected to "use strengths-based, solution-focused communication techniques when discussing the family's situation and exploring possible solutions" (p. 3-53), to "take an approach to planning that recognizes that families are experts regarding what interventions or services will be most supportive to them" (p. 3-95) and to "recognize that family members have strengths upon which they can draw as they work towards the kind of positive change that will improve the safety of the child/youth and the family's overall well-being" (p. 3-95).   Definition of MCFD Terms The terms listed below are common terms within MCFD and are used in this study:  Intake (ITK) Teams: The primary function of these teams of child protection workers is to assess child protection reports, make community referrals and work for up to 30 days to support families and mitigate child protection concerns. They are typically the first point of contact for families regarding child protection concerns. Families needing longer-term work are referred to a Family Development Response, Family Service or Integrated teams.    116  Family Development Response (FDR) Teams: These teams were first established in 2003. Their primary function is to provide intensive time-limited support to families to address child safety concerns. They use a strengths-based approach modelled on the Signs of Safety (Turnell & Edwards, 1999). They accept referrals from Intake teams when the children are deemed to be at lower risk, and the parents are deemed to be willing to cooperate with an assessment and intervention. A variety of factors, including the age and vulnerability of the child, the history of child protection concerns and the capacity of the family to care for the child contribute to the assessment as to whether a FDR team is appropriate (Ministry for Children and Family Development, 2004a).   Family Service (FS) Teams: These teams provide ongoing assessment and services to families who have been referred from Intake and who are deemed ineligible for a Family Development Response. They most commonly work with children who are perceived to be at higher risk, or for whom resolution of the child protection concerns is expected to take more than three months. They service families who are involved in child protection court proceedings.     Integrated (INT) Teams: The definition of these teams varies. It can refer to teams in which child protection workers carry out more than one of the following child protection functions: intake, family development response, family service or guardianship and adoption planning for children who have become wards of the state. It can also refer to teams combining child protection workers and colleagues from another field of practice like child and youth mental health services, youth probation services or addictions services. In this study the term covers   117  both definitions in that it is used to describe teams which include frontline child protection workers and which carry out more than one of the above functions.   Integrated Case Management (ICM): This is the name of the computerized information system shared by MCFD, the Ministry of Social Development and Social Innovation (SDSI) and the Ministry of Technology, Innovation and Citizens? Services (MTICS). The system was first introduced in 2010, although it did not become the primary information recording and retrieval system for child protection workers until 2012. The intent of ICM was to upgrade what was deemed to be obsolete information technology and to facilitate better information sharing and coordination relating to key government social programs. However, difficulties experienced by child protection workers in using the system to access and record information has meant that remedial action has been required to adapt the system to their needs (Shera & Litton, 2013).     Description of Survey Participants  The survey was completed by 224 participants. This represented 27% of the total population of 824 fully delegated frontline child protection workers in MCFD workers at the time of the survey. The following description of survey participants is based on the analysis, described in Chapter Three, of their responses to the demographic questions in the survey.   Twenty seven percent (n = 60) of survey participants said they belonged to a Family Service team, 21% (n = 48) to an Intake team and 5% (n = 10) to a Family Development Response Team. Initially 37% (n = 83) reported belonging to an Integrated team and 12%  (n = 27) to the category of 'Other'. However some of the descriptions of 'Other' were of teams with multiple functions, for instance one person identified as belonging to a team that combined   118  intake, investigation, guardianship, adoption, resources, probation, child and youth mental health and addictions services. I recoded as 'Integrated' the team membership of all participants in the 'Other' category who identified as belonging to teams that had more than one function. This left 45%  (n = 101) on Integrated teams and 2% (n = 5) on single function teams not previously mentioned, like Youth Services.   It should be noted that the demographic characteristics of workers did not differ significantly across these team types. A one-way Analysis of Variance (ANOVA) test showed no significant difference in mean age of workers across teams and Kruskal-Wallis tests, chosen due to the non-normal distribution of years of SBP experience and child protection experience, showed no significant differences in these variables across teams.    Sixty eight percent (n = 153) of survey participants held a Bachelor of Social Work (BSW) qualification, 18% (n = 41) held a Bachelor of Child and Youth Care and 8% (n = 18) held a Master of Social Work (MSW). Five percent (n = 12) identified as holding a qualification in the category of 'Other'. These included degrees in the arts, psychology, education and nursing.  Participant years of experience as a fully delegated child protection worker ranged from  0 to 32, with an average of 8.41 (SD = 6.08), median of 7 and a mode of 5 years. Their years of experience doing SBP ranged from 0 to 40, with an average of 8.28 (SD = 6.81), median of 7 and a mode of 10 years. Participant ages ranged from 23 to 65 years, with the average age being 41.9 years (SD = 10.46), the median being 41 and the mode being 35 years. Ten participants declined to give their age.    119   Survey Sample Representativeness  One question related to the description of survey participants is the extent to which they can be seen to be representative of the population of fully delegated MCFD child protection workers (N = 824). The survey link was clicked by 342 workers, or 42% of the population. The survey was completed by 224 workers, representing a response rate of 27%. The response rate suggests that in the absence of non-sampling error, results could be said to represent the population with a margin of error of 5.59%, 95% of the time.  In August 2013 MCFD provided me with data regarding the age of child protection workers. The data excluded workers in supervisory or management positions but, unlike the study sample, included workers at all levels of delegation. It was, however, the nearest proxy available for the age of frontline delegated child protection workers. The mean age of these workers was 42.58 (SD = 10.9), median age was 41.26 and mode was 34.2. I conducted a one-sample t-test to ascertain whether, in terms of age, the study population might be seen as representative of frontline child protection workers in MCFD. The test showed no significant difference between the mean age of the survey sample (M = 41.9 years, SD = 10.46) and the mean age of the current population of frontline child protection workers, t(213) = -.948, p = .34. In terms of their age the sample can be seen as representative of the broader MCFD child protection population.   In a sample of this nature, however, it is likely that there is non-sampling error. The sample is comprised of workers willing and able to respond. Those who responded may well have different characteristics from those who did not respond, and it is impossible to estimate the extent and effect of such differences (Fricker, 2008). This problem is often unavoidable in   120  exploratory research (Palys & Atchison, 2003). However it should be remembered that the inferential statistics used in this study do not take into account non-sampling error. This means that caution is needed in generalising from the sample to the population. Findings are best interpreted as providing theoretical support for possible patterns in the population, rather than an accurate description of the population.  Description of Interview Participants  I interviewed 24 frontline MCFD child protection workers. Their ages ranged from 28 to 64 with a mean age of 38.75 (SD = 9.74). Their years of child protection experience ranged from 0 to 23, with a mean of 7.17 (SD = 5.80). They claimed between 1 and 40 years of experience doing SBP, with a mean of 9.79 (SD = 9.35).  Fourteen interviewees worked on Integrated teams, six on Family Service teams, two on Family Development Response teams, one on an Intake team and one on a Youth team. Seven were male and seventeen female. Sixteen workers had a BSW, two had an MSW and six had a qualification they designated as 'Other.' These included qualifications in the psychology, counselling, medical and criminal justice fields.   One of my sampling criteria was that interviewees have a range of views as to whether SBP was hard to do in child protection work. The final sample comprised of 10 people who disagreed to some extent that SBP was hard, eight who agreed, three who neither agreed nor disagreed and one who did not complete this survey question.     121  Researcher Perspective Interpretive description requires the researcher to 'forestructure' what she brings to the study (Thorne, 2008). This means laying out the assumptions, personal agendas, and professional experiences which will potentially keep her from grasping the worlds of those she seeks to understand. I did not see this reflection as a one-off task, but rather as an ongoing process throughout the study. However in this section I have endeavoured to describe some of the perspectives I held at the outset of the research.   When I started data collection it had been 17 years since I completed a Masters in Social Work and embarked on a career in the statutory child protection field. My social work degree and introduction to child protection practice in England during the mid 1990s was heavily influenced by a forensic investigative practice model. My focus was on establishing the facts of  alleged abuse or neglect and assessing and mitigating the risk of future harm. When I began working as a statutory child protection worker in British Columbia in 1998, the legislation, policy and culture within MCFD all supported this focus. As I became more experienced, my frontline practice and work training and mentoring child protection workers was grounded in this 'risk-focused' approach.  It was from this perspective that I witnessed the introduction of SBP to MCFD. I was asked to train workers in both the traditional investigative approach and in SBP, and I became increasingly intrigued with the question of whether and how the two approaches were compatible. My experience told me that statutory child protection settings were characterized by scarce resources, demoralized workers, involuntary clients and frequent divergence between the needs of children and their parents. I was skeptical as to whether SBP was possible in these contexts. It certainly seemed to me that it required highly sophisticated relational skills and a   122  great deal of flexibility if it was to work. I found SBP appealing, having often found myself unable to make effective relationships with clients when using a more traditional deficit-focused approach. However, I wondered if SBP was more suited to voluntary mental health practice than to child protection work. I saw SBP as a new approach for child protection work, albeit one that echoed past child protection practices, and I suspected that there were limits on its usefulness.   I came to the study believing that most child protection workers wish to do the best they can for their clients in a job that can be rewarding but also very hard. I believed that frontline workers were well-placed to say what approaches worked best and that they had to see new practices as both helpful and realistic if they were to implement them. I also believed that if the needs and practice wisdom of these workers were not taken into account, the needs of clients were unlikely to be met.   Finally, my views on child protection had been informed by the history described in the literature review. During my development as a social worker I had been heavily influenced by systems and ecological theories and I saw the person-in-environment perspective as one of the core disciplinary tenets of social work. I saw the child protection field as being rooted in the traditions and theory of social work and the person-in-environment perspective as an important element of the child protection knowledge base. It led me to see all phenomena as the product of the complex interaction between individuals and their environment and to believe that useful explanations of human behavior tended to incorporate both individual and contextual processes.      123  CHAPTER 5 : SURVEY FINDINGS Knowledge and Use of SBP In this study I set out to understand whether and how child protection workers applied the ideas of SBP. The first two survey questions were designed to explore this by asking participants "How much do you know about Strengths-Based Practice?" and "How often do you do SBP in your child protection work?". Participants were asked to choose a response from a Likert-type scale. The possible responses ranged from zero, representing knowing nothing about SBP and never doing SBP respectively, through to four, representing knowing a lot about SBP and always doing SBP (see Table 3).   Table 3 Self-Assessed Knowledge and Use of SBP (N = 224) I know ... Frequency Percentage of N  I do SBP... Frequency Percentage of N Nothing 0 0  Never 0 0 A little 6 3  Occasionally 5 2 A fair amount 46 21  Some of the time 34 15 A good amount 121 54  Most of the time 130 58 A lot 51 23  Always 55 25   All participants reported that they knew at least a little about SBP (M = 2.97, SD = .74) and did SBP in their child protection work (M = 3.05, SD = .70). A two-tailed Pearson's Product Moment correlation was conducted to explore whether there was a relationship between self-reported knowledge and use of SBP. This showed a statistically significant positive relationship (r(222) =.57, p =.000), accounting for 32% of the variance in the association.   Spearman's correlations and one-way ANOVAs were conducted to explore whether self-reported knowledge or use of SBP varied with age, years of SBP or child protection experience, level of qualification or team membership. Both were positively correlated to SBP experience,   124  with SBP experience accounting for 7.6% of the variance in ranks for self-reported knowledge of SBP (r(222) =.28, p =.000), and 3.5% of the variance in ranks for frequency of SBP use (r(222) =.19, p =.005). SBP experience was the only variable that was found to be related to self-reported use and frequency of SBP.     Support for SBP  A pragmatist perspective suggested that whether workers implemented SBP would be connected to the extent to which they felt positively about its application to child protection work. Four questions, rated on a continuous scale ranging from 0 (strongly disagree) to 4 (strongly agree) were designed to assess the level of support for the approach. Responses (see Table 4) showed a high degree of support for SBP, with 89% (n = 200) in agreement that SBP was a good approach for child protection work and 85%  (n = 191) in agreement that SBP increased the chances of success with clients. Only 9% (n = 21) felt that using SBP increased the risk to children.   Table 4 Support for SBP (N = 224)   Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree M SD  SBP is a good approach for child protection work  Frequency 3 3 18 122 78 3.20 .75 Percentage 1 1 8 55 35   Doing SBP increases the chances I will be successful with my clients Frequency 2 7 24 112 79 3.16 .80 Percentage 1 3 11 50 35   I would like to do SBP more with my clients Frequency 2 6 59 101 56 2.91 .83 Percentage 1 3 26 45 25   Children are left at greater risk when protection workers do SBP Frequency 63 104 36 17 4 1.08 .95 Percentage 28 46 16 8 2      125   Two-tailed Spearman's correlations were conducted to explore whether relationships existed between positive worker attitudes towards SBP and self-reported knowledge of SBP, frequency of SBP use and duration of SBP experience. They showed that self-reported SBP knowledge, frequency of use and length of SBP experience were indeed all positively correlated with the attitudes 'SBP is a good approach for child protection work' and 'SBP increases the chances I will be successful with my clients' and that the first two variables were negatively correlated with the attitude that SBP increases risk for children (See Table 5).  Two-tailed Spearman's correlations were also used to ascertain whether support for SBP varied with age and length of child protection experience. These showed no relationship between attitudes of support for SBP and length of experience as a child protection worker. However there were significant positive correlations between worker age and the attitudes that SBP is a good approach for child protection work, that it increases the chances of success and that  workers would like to do SBP more.  Table 5 Correlations Among Worker Characteristics and Attitudes of Support for SBP   I would like to do SBP more with my clients SBP is a good approach for child protection work Children are left at greater risk when protection workers do SBP Doing SBP increases the chances I will be successful with my clients Age  (n = 214) .14* .19** -.04 .14* SBP Experience (N =224) .09 .23*** -.09 .22** Knowledge of SBP  (N = 224) .09 .28***    -.26*** .30*** Frequency of SBP  (N = 224) .18** .33***    -.31*** .41*** * p < .05, ** p < .01, *** p < .001 Shaded correlations are Pearson's r. All other correlations are Spearman's rho   One-way ANOVAs and their non-parametric equivalent, the Kruskal-Wallis test, were conducted to explore whether worker support for SBP varied with their qualification or team   126  type. While these showed no association with qualification, there was a significant relationship between team membership and the belief that SBP increased risk to children, H(3) = 9.13, p = .028. The mean scores on this attitude were .50 for FDR workers (SD =.71), 1.04 for Integrated workers (SD =.95), 1.08 for Intake workers (SD =.92) and 1.33 for Family Service workers (SD =.97). This suggested that any significant differences were likely to be between the FDR and the Family Service workers. To explore further, I conducted two post-hoc Mann Whitney U tests to assess whether differences between the FDR and FS workers, and between the FDR and Intake workers, were significant. The FDR workers (Mdn = 0) agreed significantly less than workers on Family Service teams (Mdn = 1) with the statement "Children are left at greater risk when protection workers do SBP",  U  = 150.50, z = -2.68, p = .007, r = -.32. Team membership accounted for 9% of the variance on this attitude between the two groups. There was no significant difference in this attitude between the FDR workers and the Intake workers (Mdn = 1), U  = 153.00, z = -1.90, p = .058, and therefore it could also be concluded that there was no difference between members of the FDR and Integrated teams.    Applicability of SBP The close relationship described in Chapter Two between changes in frontline practices and in the perception of the role of child protection workers had led me to hypothesise that the application of SBP ideas would depend somewhat on the extent to which workers saw them as possible within and relevant to their role. Four Likert-type response survey questions were intended to elicit worker attitudes about these aspects of SBP's applicability to child protection.   Of all the attitude questions, these four elicited the most even spread of participant responses (see Table 6). Nearly the same number of participants agreed as disagreed with the   127  statements regarding the appropriateness of SBP. Forty-five percent (n = 101) thought SBP was appropriate in every situation and 40% (n = 90) that it was not. Forty-four percent (n = 99) thought it was right for all clients while 43% (n = 97) thought that it was not. Regarding the two questions pertaining to how difficult it was to do SBP, approximately twice as many workers thought that SBP was not hard than thought it was hard and approximately twice as many thought it was always possible than thought it was not always possible.   Table 6 Attitudes Regarding the Applicability of SBP (N = 224)   Strongly Disagree Disagree Neither Agree nor Disagree Agree Strongly Agree M SD SBP is appropriate in every situation Frequency 20 70 23 66 35 2.12 1.26 Percentage 9 31 15 30 16   With some clients SBP is not the right approach Frequency 29 70 28 78 19 1.95 1.23 Percentage 13 31 13 35 9   SBP is hard to do in child protection work  Frequency 30 99 29 58 8 1.62 1.11 Percentage 13 44 13 26 4   It is always possible to do SBP Frequency 9 54 29 92 40 2.45 1.16 Percentage 4 24 13 41 18       I conducted correlations to ascertain whether there was any relationship between workers' self-reported knowledge and use of SBP and their attitudes regarding SBP's applicability. These showed that as self-reported knowledge of SBP, frequency of SBP use and length of time using SBP increased so did the extent to which SBP was perceived to be possible and appropriate (see Table 7). The largest effect, of medium size, was for frequency of use, which accounted for 10 -   128  16% of variance in these attitudes as to SBP's applicability. Length of SBP experience accounted for 4 - 6% of variance in these attitudes and self-reported SBP knowledge accounted for 3 - 7%.  Table 7 Correlations Among Worker Characteristics and Attitudes about SBP Applicability   SBP is hard to do in child protection work It is always possible to do SBP SBP is appropriate in every situation With some clients SBP is not the right approach Age  (N=214) -.23** .14* .11 -.11 Years of SBP Experience  (N=224) -.20** .25*** .24*** -.24*** Knowledge of SBP  (N=224) -.22** .24*** .17* -.27*** Frequency of SBP  (N=224) -.34*** .40*** .32*** -.33*** * p < .05, ** p < .01, *** p < .001 Shaded correlations are Pearson's r. All other correlations are Spearman's rho     I also sought to establish whether differences in attitudes regarding the applicability of SBP were related to such worker characteristics as age, length of child protection experience, qualification or team membership. Age was significantly correlated to workers seeing SBP as less hard and more possible, accounting for 2 - 5% of variance in these attitudes. It was not, however, correlated to attitudes regarding the appropriateness of SBP. Neither years of child protection experience nor level of qualification were related to any of the four attitudes outline in Table 7.   One way ANOVAs did, however, indicate a significant difference across the four team types (Intake, FDR, Family Service and Integrated) for the attitude 'SBP is appropriate in every situation', F(3,215) = 2.71, p = .046,   = .15. A Hochberg's GT2 post hoc test, chosen because the groups were different sizes but had homogeneous variance, was used to make pairwise comparisons between the team types. It showed a significant difference between the Intake and   129  Integrated workers on the attitude 'SBP is appropriate in every situation' (p =.032, r =.22). Members of Integrated teams (M = 2.26, SD = 1.29) were more likely than members of Intake teams (M = 1.65, SD = 1.16) to feel that SBP was always appropriate. The effect size for this comparison was small, with team membership accounting for only 5% of the variance between the two team types.   A one-way ANOVA also showed a difference according to team membership that was on the boundary of statistical significance for the attitude 'With some clients SBP is not the right approach', F(3,215) = 2.64, p = .05,   = .15. A Hochberg's GT2 post hoc test showed a difference between the Intake and FDR workers that was also on the boundary of statistical significance (p =.053, r =.33). The FDR team members (M = 1, SD = 1.49) were less likely to agree with the attitude than the Intake team members (M = 2.1, SD = 1.53). There was a medium effect size for this difference, accounting for 11% of the variance between workers from the two team types.  'Sometimes' and 'Always' Applicable  To further explore emergent differences in the attitudes regarding SBP's applicability, I coded participants into those who thought SBP was always applicable and those who thought it was not. The criteria for this coding are detailed in the Plan for Statistical Analysis of Survey Data in Chapter Three. Forty two participants were coded as missing, leaving 70% of the remaining sample of 182 in the 'Sometimes' group (n = 128) and 30% in the 'Always' group (n = 54). A number of statistical tests were then conducted to ascertain whether membership of these two groups was related to worker demographic characteristics or self-reported SBP knowledge or use.     130   Chi-square tests were used to explore whether those who clearly saw SBP as always applicable differed in their team membership or level of qualification from those who clearly saw SBP as only sometimes applicable. They did not. A t-test also indicated that the groups did not differ in mean age, and a Mann Whitney U test showed no difference in child protection experience.   There was, however, a significant difference between the groups in their length of SBP experience, knowledge of SBP and use of SBP. Mann Whitney U tests showed that the median years of SBP experience for the 'Always' group (Mdn = 10) was significantly higher than that for the 'Sometimes' group (Mdn = 5), U = 2395.50, z = -3.28, p = .001, r = -.24. A t-test showed that the mean SBP knowledge score of the Always group (M = 3.28, SD = .60) was also significantly higher than for the 'Sometimes' group (M = 2.85, SD = .733); (t(180) = -3.778, p = .000, r = .27). These differences across the groups constituted a small effect. Self-reported use of SBP was also related to whether participants felt SBP was sometimes or always applicable. A t-test showed that those in the 'Always' group (M = 3.43, SD =.54) did SBP significantly more frequently than those in the 'Sometimes' group (M = 2.88, SD = .70); (t(180) = -5.18, p = .000, r = .36).     The Use of SBP Techniques Scaling, miracle, coping and exceptions questions are, at least in theory, important techniques of strengths-based solution-focussed practice in child protection work. Survey respondents answered a series of questions asking them to rate how often they had used these techniques in their last 10 conversations with parents of children on their child protection caseload. These questions elicited the rate of self-perceived use of these techniques and offered some insight into the extent to which participants saw themselves applying the SBP model described in much of   131  the child protection literature. It should be noted that the responses were not expected to provide an accurate accounting of the use of these techniques, as participants tend to perceive past events as happening more recently than is true and to generally have difficulty recalling this kind of  detailed information (Brace, 2008).  The self-perceived use of each of the four SBP questioning techniques, is detailed in Table 8.   Table 8 Use of SBP Techniques in Previous 10 Adult Client Conversations   Not used Used once or twice Used three or four times Used five or more times I do not  know this technique n M SD  Miracle Question Frequency 53 58 41 61 4 217 2.47 1.19 Percentage 24 27 19 28 2 Exceptions Question Frequency 28 52 53 63 21 217 2.50 1.30 Percentage 13 24 24 29 10 Scaling Question Frequency 36 57 50 67 7 217 2.62 1.18 Percentage 17 26 23 31 3 Coping Question Frequency 15 27 60 102 11 215 3.06 1.16 Percentage 7 13 28 47 5   To explore these groups further, I added the responses of those workers who reported not knowing the techniques to the responses of those workers who reported not using the techniques. This created a continuous scale for each technique of self reported use, coded from zero (no use of the technique) through to three (technique used five or more times in the last 10 client conversations). For the miracle question the mean was 1.49  (SD = 1.16), for the exceptions question it was 1.60 (SD = 1.13), for the scaling question it was 1.65 (SD = 1.17) and for the coping question it was 2.11 (SD = 1.04).    132   As all four variables regarding use of these specialised SBP techniques had skew or kurtosis ratios greater than three, I conducted Spearman`s correlations to explore relationships between self-reported use of the techniques and worker age, SBP experience, child protection experience, use and knowledge of SBP and attitudes towards SBP (see Table 9).   Table 9 Correlations Between Worker Characteristics and Use of SBP Techniques   Miracle Question n = 217 Exceptions Question n = 217 Scaling Question n = 217 Coping Question n = 215 Knowledge of SBP .168* .237** .091 .141* Frequency of SBP  .190** .228** .030 .132 SBP Experience -.023 .037 .000 -.014 Age -.063 -.149* -.065 .037 Child Protection Experience -.106 -.197** -.093 -.032 Belief that using SBP leaves children at greater risk -.072 -.198** .117 -.029 * p < .05, ** p < .01   Spearman's correlations showed no relationship between reported use of any of the four techniques and years of SBP experience. There was also no relationship between use of the scaling question and any of the variables. Those who reported knowing more about SBP reported using the miracle, exceptions and coping questions significantly more. There was also a statistically significant relationship between reported use of the SBP approach in general and reported use of the miracle and exceptions questions, although no correlation with use of the coping and scaling questions.   Age and child protection experience were significantly related to use of the exceptions question only. As each increased, use of this question decreased. The only attitude which was   133  significantly correlated to use of any of these techniques was 'Children are left at greater risk when protection workers do SBP'. People who agreed with this statement were less likely to use exceptions questions. It should be noted that effect sizes for all these significant correlations were small, ranging from .14 to. 24.    One way ANOVA's showed there to be no significant differences in the mean scores for the miracle, exceptions or scaling questions across levels of qualification. The non-parametric equivalent of the Kruskal-Wallis test was conducted for the coping question, as scores for this variable were not normally distributed within the qualification groups. This showed there to be no differences in the mean reported use of the coping question across qualifications.   One way ANOVA's showed the mean score of the miracle question, and this question only, to be significantly related to team membership, F(3,209) = 2.78, p = .042,   = .49.  Hochberg's GT2 post hoc tests indicated that Intake workers (M = 1.80, SD = 1.17) were more likely than Family Service workers (M = 1.19, SD = 1.14) to report use of the miracle question, (p =.039, r = 0.26) although the effect size for this comparison was small.  When SBP is Inappropriate I offered respondents who had agreed with the statement 'SBP is not right for some clients' or disagreed with the statement 'SBP is appropriate in every situation' the opportunity to answer an open-ended question asking in which situations, or for which clients, SBP was not appropriate. Using the interpretive description coding and quantification processes described in Chapter Three, I inductively coded the content of their responses to this question. I collapsed these codes into larger categories which summarised the clients and situations for which SBP was deemed   134  inappropriate. I then recoded all the responses to this question using these categories and counted the frequency of responses in each category.  Respondents across all team types stated that there were situations for which, and clients for whom, SBP was inappropriate. One hundred and two of these respondents offered one or more examples of such situations or clients (see Table 10). It should be noted that while this question specifically asked about the limits of the appropriateness of SBP, a small number of people had clearly conflated the concepts of 'appropriate', 'possible' and 'effective' and described situations in which SBP was particularly challenging or ineffective. Their responses were, however, included in this analysis.  Table 10 When SBP is Not Appropriate (n = 102) When SBP Is Not Appropriate Percentage of n With clients who are not willing to work collaboratively 30 When the worker must act to secure the child's immediate safety 26 With (some) clients with substance use issues 25 With (some) clients with mental illness 21 With clients who are hostile or aggressive 18 With clients who deny concerns 14 With (some) clients who sexually abuse 12 When the abuse was particularly severe 10 With (some) clients with a long history of child welfare involvement 7 With clients with limited cognitive capacity 5 In some situations of family violence 4 With clients who are sociopathic or psychopathic 4 With clients who deliberately harm their child 2 With clients who are involved in criminal activity 2 Unique responses: when the worker had already made a relationship with the client; investigation cases; child protection cases; when there was court involvement; when the worker was on the witness stand; with clients of differing cultural backgrounds; in cases involving homelessness or prostitution; and with parents who did not care for the welfare of their children   135   Some respondents suggested that SBP was not appropriate for a broad cross-section of typical child protection clients, for instance one worker cited "child protection cases" (260, FS1) and another, "mentally ill clients, clients where severe physical or sexual abuse has occurred, domestic violence. Severe substance using clients" (90, INT). In most cases workers specified that SBP was inappropriate for entire client populations, for instance people with mental illness or addictions. Sometimes they qualified their responses with words like 'some' and 'sometimes' to suggest that SBP was inappropriate only for certain people within these categories.   Challenges  One goal of this study was to understand the factors perceived by workers to be barriers and supports to SBP in child protection practice. To this end, I included in the survey an open-ended question asking workers what they found to be most challenging about SBP. Their responses are summarised in Table 11.   Two hundred respondents answered this question by identifying challenges. Often they identified more than one challenge. I coded the responses using the inductive coding techniques described in Chapter Three. This coding process led to the development of a series of codes describing practitioner challenges, grouped into the larger categories of 'organisational factors', 'client factors' and 'practice factors' to reflect where respondents located the source of their challenges. I then recoded all survey responses to this question using these codes and categories and counted the frequency of responses within each.                                                   1In reporting these findings I have used quotations to illustrate categories and I have identified these quotations with the respondents' survey identification number and team type, abbreviating ITK for Intake, INT for Integrated, FS for Family Service, FDR for Family Development Response and OTH for Other.   136   Table 11 Factors Making SBP Challenging (n = 200) Factor Percentage of n Organisational Factors 46 Lack of time 25 Unsupportive organisational culture 12 Lack of resources 8 Legislation, tools and policies 8 Unsupportive supervisor 4 Unsupportive management 3 Other agencies uncooperative 2 Lack of effective training 1 Client Factors 31 Clients who are unwilling to engage/acknowledge concerns 12 Hostile clients 6 Clients with whom it is hard to find strengths 4 Clients with substance use issues 3 Clients with mental health issues 3 Clients with a long history of child welfare involvement 4 Clients who lack capacity 3 Clients do not trust SBP 4 Practice Factors 31 Balancing supportive and directive roles in client relationship 14 Identifying strengths 6 Keeping the right attitude 6 Being strengths-based in high risk situations 4 SBP requires a high degree of skill 2 Not being phony 1 Managing conflict 1 Not ignoring structural barriers 1 Doing SBP with non-English speakers 1   Organisational Factors  Ninety-two of the 200 respondents to this question identified challenges relating to organisational supports for SBP. The most common challenge, identified by 49 workers, was the lack of time to do SBP. Respondents talked of lacking the necessary time to construct strengths-based  relationships with clients, to explore strengths and to develop plans. This was an approach   137  that was seen as requiring considerably more investment from workers than 'business as usual'. As one worker said, "It is a lot easier to tell a client what they need to do as opposed to taking the time to involve the family and extended family. Strength based work can be very time consuming" (232, FS). Another wrote that,  Strengths-based practice takes a great deal of time, effort and patience. You must build a strong relationship with the client in order to weather the storm of real change. This can be difficult in a child protection setting but it is possible! I have been amazed time and time again at what families and youth are able to do when they feel empowered to make change. Although it takes tremendous work up front, the shift away from adversarial relationships and top down power dynamics free both client and worker to focus on the real needs and issues. (212, OTH)   While some talked of the benefits of being able to take the time with clients they felt was necessary, others identified that attempting SBP without sufficient time had significant costs for the client: I have found that an asset search can be difficult and painful for a client if only a short time is available as clients can be left with only a handful of strengths and feel terrible.  When I've had time to do really thorough asset searches, clients have been surprised by how many positives they have. However, I've found that clients can freeze up when asked about good things in their lives and this needs to be drawn out over several meetings. (168, INT)   138  Workers frequently explicitly linked the lack of time to excessive caseload size. Administrative demands like form-filling and data entry on the ICM computer system were also seen to take social workers away from strengths-based work with families:  All the paperwork (particularly with the onslaught of the hideous ICM) that forces social workers to do "virtual" social work rather than "real" social work. Mark my words, the whole system is going to come crashing down like a deck of cards if the current situation is not rectified . . . in a hurry!  Social workers are all sagging under a crushing load of paperwork! (55, INT)    Twenty-three workers identified the organisational culture of MCFD as a major barrier to SBP. It was characterised as deficit-based, constantly changing and overly concerned with allocating blame: I do not believe that we work from a strength based practice approach. I think that we give lip service to the practice, but when it comes down to the work and the tools that we have, it is still from a place of what is lacking. (297, FS) There was concern that an organisational focus on expedience undermined SBP, which had been reframed as "the best way to save a buck" (281, INT). Colleagues were seen as unsupportive of SBP and inclined to be risk-averse and punitive. There was a lack of understanding as to what SBP in child protection work entailed:  I find that how people interpret and teach strength based practice to be inconsistent and often inaccurate. I have been taught that it is a practice that does not say the bad stuff and only speaks in positive terms. This is not strength based practice. But there are many workers who are confused. I have found sometimes MCFD likes to grab an   139  idea without doing all the research to know exactly what they are talking about and then teach things out of their intended context, which brings confusion. (254, INT)   Fifteen workers identified MCFD legislation, policies and tools as creating challenges for SBP. For some, the child protection mandate itself was a barrier to doing SBP. As one worker said, "it is difficult at times to work strength-based in child protection when child protection is inherently an intrusive and mandatory process" (54, FS). For others, 'risk-focussed' assessment tools and policies were problematic. One worker, for instance, wrote about the focus on eliminating, rather than managing, risk and gave the example of a strengths-based plan that had been undermined by the decision that someone with an unrelated criminal conviction could not provide care to a child. The policy that workers consult clients before interviewing children was also identified as problematic at times.  Sixteen workers identified lack of resources as problematic for SBP. Strengths-based plans needed to be supported by flexible and adequate services and these were seen as generally lacking. As one worker said, "it is incredibly time consuming and resource ($) heavy and we are not provided with either the time or resources to practice effectively in this manner. The workload is unethical" (209, FS). This created the sense that SBP was a "false front" (266, FS), or a means to divest responsibility to families without providing them with the means to properly assume it: In this province strengths-based practice is an ideological investment in the form of cost-saving that allows influential state agents to offload care responsibilities - which the state or the community should shoulder - to family members - usually women -   140  who do not receive the training or economic compensation to sustain themselves or their dependents. (281, INT)    Team leaders and managers were explicitly identified by twelve workers as barriers to SBP. They were described as lacking knowledge of SBP, being scared to use it or actively disagreeing with the approach. They were also described as insufficiently creative and strengths-oriented for strengths-based planning to work.  Client Factors Sixty-two of the two hundred respondents identified the challenges of SBP as lying with particular groups of clients. The group most commonly identified as problematic for SBP  consisted of clients who were unwilling to acknowledge concerns or to engage in a collaborative process. This included clients who absented themselves from the relationship physically, those who did not acknowledge child protection concerns, those who wanted simply to be told what they should do, those who were not prepared to look at the situation in a strengths-based way, those who were unwilling to share information, those who lied to workers and those who did not follow through on plans. As one worker wrote of SBP,  I find that it assumes that people are in a place where they recognize that something needs to change in their family/life/situation, however in my experience most people that we work with in child protection are unwilling to acknowledge this. (240, INT)   141     Eleven people described the challenge of doing SBP with clients who were hostile or aggressive. As one said "It is extremely challenging to be faced with a disrespectful client and find it within yourself to remain SB" (216, INT). Another worker commented that, I don't always know how to use this practice when clients are non-responsive or very aggressive. I believe relationship building is the key but you can't fake that as sometimes the connection just isn't working. When clients are aggressive I don't always think of this practice as I am in "protective mode" for myself. (268, FS)   With some clients it was simply difficult to find strengths. Some clients with mental health or substance use problems were also identified as presenting particular problems for this approach, in addition to clients who had had extensive involvement with MCFD and either did not trust a new approach or were locked into negative patterns of interacting with MCFD. These were described by one worker as "the chronic clients that are 'lifers' and have a hard time accepting that things are going to be done differently this time around" (388, INT).  SBP was also challenging with people who did not have the capacity to come up with their own solutions or run their own family plans, either because they had limited cognitive abilities or simply because the task was too hard. One worker wrote that the problem lay with,  the belief that the client truly is the 'expert' of their situation. Some clients do not possess the insight, intelligence to even see a way out of their situation or if they do it is entirely impractical. While I do see the benefits in C.P. work, this approach strikes me as coming out of university based therapists who have no concept of what some of   142  our clients are truly like. 'They' base the practice on looking at people thru 'rose coloured glasses'. (336, ITK)  Practice Factors The third category of challenges, described by sixty-two respondents, was located within the strengths-based interaction and pertained to the enactment of SBP. The most common challenge in this category, identified by 28 workers, was finding the right balance in the worker-client relationship between being supportive and being directive, intrusive or firm about child protection concerns.  It was hard for some workers to maintain a relationship while ensuring that the focus on strengths did not obscure the risks and undermine their ability to address them:  I think it is a fine balance between strengths-based practice and needing to be clear about addressing the child protection concerns and mitigating risk to the child. At the Intake level we are always assessing risk while also trying to support the family unit. MCFD investigations and FDR assessments can often be very intrusive processes. I believe the challenge is working through the steps of the investigation while continuing to be respectful, professional, and strengths-based. (299, ITK) Another described this challenge as being about, balancing the positive messages and the negative messages. Making sure the issues of concern are being heard and are being given the appropriate weight by everyone. Making sure that recognition of the positives does not "drown out" the concerns, thus leading to confusion on why MCFD is involved if "everything is going so well". Skilful application of strengths-based work - some workers may feel that recognition of strengths is strengths-based, which is not accurate as there needs to be   143  acknowledgement of what is not going well and an appropriate plan created to address the concerns. (116, INT)   A second challenge was identifying strengths in families. Situations in which this was particularly difficult included when children appeared to be at high risk, when families experienced multiple problems and when the social worker simply was not in the right frame of mind. As one worker said "there are times and situations where it is difficult to look at the good things within the families. There are times when the situation can seem hopeless and it makes it difficult to move past the negative and into the positive" (208, FS). Twelve workers identified their challenge as keeping the right mental perspective for an approach that was seen as requiring a great deal of effort and patience and was considerably harder to maintain than simply telling the client what to do. One worker wrote "I find that strengths-based practice requires constant effort . . . a constant "presence of mind", in order to be genuine and effective. I believe that clients are able to discern between genuineness in "collaborative spirit", and simple "lip-service" (133, INT). Eight workers identified that SBP was hard to practice in situations of immediate or high risk.  Comparing Challenges Identified by the 'Sometimes' and 'Always' Groups  To explore whether participants who thought SBP had limited applicability perceived similar or different challenges from those who thought it was applicable at all times, I categorised all responses to the question 'What do you find most challenging about SBP?' according to membership of the 'Sometimes' and 'Always' groups. I then counted the frequency of the identified challenges in each group (see Table 12).     144  Table 12 Challenges Identified by the 'Sometimes' and 'Always' Groups  Always (n = 54) Sometimes (n = 128)  Percentage of n Percentage of n Organisational Factors 54 34 Lack of time 24 22 Lack of resources 11 6 Legislation, tools and policies 9 6 Unsupportive organisational culture 19 7 Unsupportive supervisor 4 3 Unsupportive management 4 2 Other agencies uncooperative 2 2 Lack of effective training 2 1 Client Factors 9 21 Clients who are unwilling to engage/acknowledge concerns 0 12 Hostile clients 0 5 Clients with whom it is hard to find strengths 2 2 Clients with substance use issues 0 3 Clients with mental health issues 0 3 Clients with a long history of child welfare involvement 4 2 Clients who lack capacity 0 4 Clients do not trust SBP 6 1 Practice Factors 20 29 Balancing supportive and directive roles in client relationship 13 14 Identifying strengths 2 8 Keeping the right attitude 0 6 Being strengths-based in high risk situations 0 5 SBP requires a high degree of skill 2 2 Not being phony 0 2 Managing conflict 2 0 Not ignoring structural barriers 2 0 Doing SBP with non-English speakers 0 1    To look at this another way, I coded each member of the 'Always' and 'Sometimes' groups according to whether they identified the locus of challenges as being with the organisation, with clients or with the practice of SBP. Twelve workers identified challenges as emanating from more than one of these locations and they were allocated to a fourth group labelled 'combined'.   145  The locus of challenges identified by the 'Sometimes' and 'Always' groups are summarised in Table 13. A chi square test showed that differences between these four groups were not statistically significant.  Table 13 The Locus of Challenges   Organisation Client Practice Combined Always n = 54 Frequency 29 5 11 2 Percentage 54 9 20 4 Sometimes n = 128 Frequency 45 27 33 9 Percentage 35 21 26 7   Supports Needed To elicit what helped workers to do SBP, survey respondents were asked the open-ended question 'What supports do you need in order to do SBP?' Two hundred and four respondents answered the question. I coded and counted responses to this question using the coding and quantitizing techniques described in Chapter Three. Participant suggestions are summarised in Table 14.  Table 14 Supports Needed to do SBP (n = 204) Supports Needed Percentage of n More time 36 More resources for families 27 More support from supervisor 20 More training 20 More support from management 16 Changes to tools and policies 12 More supportive organisational culture 6 More collaboration with community partners 5 More support from team 4 Unique responses: support from own family members; support from the client; a high degree of maturity   146   The most common support needed to do SBP was more time. This was identified by 73 workers, 44 of whom said that more time could be created through smaller caseloads and more staff: More staff... and this is a political issue! I've heard that other provinces in Canada have addressed caseload issues in their statutes, i.e. that if caseloads go above 25, government is required to create more delegated positions.  I think this is a HUGE issue for the client group I work with... The (named) team... has caseloads in the 60's - how can they POSSIBLY do preventative work with clients? With caseloads that high, ALL they can do is respond to immediate risk and emergencies. (189, INT) Other suggestions included reducing paperwork and ICM requirements and ensuring timely backfill for vacant positions and paying a salary comparable to other provinces. The time was needed to support families because, as one worker put it, "relationships and trust aren't built overnight" (132, FS).   Fifty-six workers identified that to do SBP more resources were needed. They called for more flexible and responsive services to support strengths-based plans for families. The most common specific services identified were family support outreach workers, mental health  services and support in the form of gas vouchers, bus passes or money for client transportation. Other specific services identified were addiction services, family education programs, parent-teen mediation, resources for men who had been violent, homemakers and doulas. More support from public health and preschool programs and access to legal aid and increased CPP compensation for dependents of deceased and disabled contributors were also discussed. Other suggestions included strengthening supports for out-of-care placements, creating a web of respite services, and providing more long-term support to birth families. Some workers called for access   147  to budgets to support parent-child activities, to provide honoraria to elders, provide practical supports like food vouchers and generally to finance creative plans that met the unique needs of a family.   Forty-one workers identified the need for more support from their direct supervisor. The most common support identified was for more regular supervision, and that this be "CLINICAL supervision: meaning supervision that is not simply focused on the day to day tasks and documentation but is about doing the work and its impact on both worker and client" (390, INT). Several people emphasised the need for supervisors to be both clinically skilled and committed to the values of SBP. Six workers asked specifically that supervisors better understand SBP and what it requires of the workers. As one said, clinical skills and leadership were needed "to remind me, mentor me and inspire me to stay on the positive thought process and not get bogged down in negativity" (138, FS). Other workers identified that supervisors needed to be creative, flexible and level-headed. They also needed to have child protection experience "as there are massive gaps in understanding when they come from other fields" (327, INT).  Separate from support from their direct supervisor, 33 workers identified that they needed more support from management. About half of this group asked simply that managers show greater understanding of, and commitment to, SBP. Others asked for more practical supports from management. Their suggestions included managers being more open to creative planning, supporting workers to take on risk,