THE ACTION-PROJECT DIFFERENTIATION PROCEDURE: EXTENDING THE METHOD OF A N A L Y S I S IN ACTION THEORY TO A L L O W FOR THE E X A M I N A T I O N OF SIMILARITIES A N D DIFFERENCES A M O N G SETS OF CASES by JOSE FRANCISCO DOMENE B.Sc. Trent University, 1995 M.A. Trinity Western University, 1998 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Measurement, Evaluation and Research Methodology/ THE UNIVERSITY OF BRITISH C O L U M B I A April, 2005 © Jose F. Domene, 2005 A B S T R A C T Although between-groups comparisons are a mainstay in quantitative research methods, their legitimacy within qualitative research is more tenuous. In this methodological dissertation, I explore the issue of between-groups comparative analysis within Young, Valach, and colleagues' contextual action theory research framework, in order to extend the existing action-project method of conducting social research to encompass mechanisms for examining similarities and differences between distinct groups of cases. To focus my methods-development work, I ask two specific questions, (a) "Are between-groups comparisons compatible with the assumptions that underlie the action theory paradigm?" and, (b) "What specific procedures could be used within these analyses, to generate findings that go beyond those which are attainable within the existing method?" To address these questions, I first examine the possibilities and problems with conducting between-groups comparative analyses in different forms of social constructionist research, including the action-project method. After establishing that some forms of between-groups comparison are compatible with action theory, I present a novel method for engaging in such analyses, the action-project differentiation procedure, and demonstrate its application to an existing data set. The major findings of this dissertation are: (a) between-groups comparative analyses that retain the full configuration of information from all cases are compatible with this form of qualitative research; and (b) the procedure I have developed is capable of generating useful, trustworthy findings. I conclude with some suggestions for further refinement of the action-differentiation procedure, and a discussion of the broader implications of shifting the previously existing analytical method in this direction. T A B L E OF CONTENTS Abstract 1 1 Table of Contents i i i List of Tables .. iv List of Figures iv CHAPTER I Overview of Dissertation 1 CHAPTER II Description of Action Theory 7 Introduction 7 Theoretical Assumptions of Action Theory 8 The Existing Action-Project Method 1 17 CHAPTER III Addressing the Research Questions 30 Introduction 30 Comparative Analysis in Other Types of Social Constructionist Research : 31 Comparative Analysis in Published Action-Project Research 42 Between-groups Analysis in Action Theory 52 The Action-Project Differentiation Procedure 56 Initial Conclusions 66 CHAPTER IV Demonstration of the Action-Proj ect Differentiation Procedure 69 Introduction 69 Study 1: Goal Orientation 70 Method 72 Findings 81 Discussion 89 Study 2: Gendered Dyadic Configuration 92 Method 96 Findings • 99 Discussion 103 CHAPTER V Methodological Findings of the Demonstration Studies 107 Introduction 107 Methodological Comparison of the Demonstration Studies . 107 Specific Problems and Possible Refinements I l l CHAPTER VI Discussion 120 Introduction 120 Contribution to Knowledge 120 Boundaries and Limitations of the Procedure 122 Relationship to Other Qualitative Methods 125 The Next Step 133 References • 135 Appendix I Common Findings within the Goal Orientation Groups 144 Appendix II Comparison Chart for Goal Orientation Data 146 Appendix III Preliminary draft of the Goal Orientation Findings 148 Appendix IV Common Findings within the Dyadic Configuration Groups 155 Appendix V Comparison Chart for Dyadic Configuration Data 157 Appendix VI Preliminary draft of the Dyadic Configuration Findings 159 LIST OF TABLES Page Table 1: Selected Paradigm Positions of Social Constructionism and Action 11 Theory Table 2: Aspects of Comparison in Other Social Constructionist Research 41 Methods Table 3: Summary of Existing Empirical Literature Utilizing the Action-Project 43 Method Table 4: Methods-related Differences Between Study 1 and Study 2 109 LIST OF FIGURES Page Figure 1: Part of the comparison chart for a hypothetical study on the family 64 career development projects of adolescents who have experienced an episode of early psychosis, in relation to those who have not. Figure 2: Outline of the action-project differentiation procedure, numbered to 67 facilitate tracking of the application of this method in the demonstration studies presented in chapter four. Figure 3: Approximate sequence for data collection and analysis in the Family 73 Career Development study from which the current data set was obtained. CHAPTER 1: OVERVIEW OF DISSERTATION The social sciences, including psychology, are built upon the examination of variation within, among, and between people. Individual paradigms for conducting social research each have their own methods and standards for deciding whether any given variation is sufficiently strong to warrant the conclusion that a noteworthy difference has been detected. These methods are worth studying in their own right, and can be re-examined and extended when new kinds of research questions arise within a field of study. This dissertation documents my engagement in this methods-development process, with a specific research method that is associated with an emerging framework for studying psychological phenomena: action theory. Various theories of action are present within the psychological literature. Therefore, it is important to clarify that, within the context of this dissertation, the term 'action theory' is used to refer specifically to Young, Valach, and colleagues' contextual theory of action (e.g., Collin & Young, 1992; Valach, 1988; Valach, Young & Lynam, 2002). This definition excludes other theories of action and goal-oriented behaviour (e.g., Little, 1999; Locke & Latham, 2002). I have used this narrow definition of action theory because the focus of my methods-development work is the 'action-project method' that is specifically associated with Young and colleagues' version of action theory, and is not used in other extant theories of action. The action theory to which I am referring is also distinct from action research / participatory-action approaches (e.g., Hugentobler, Israel & Schurman, 1992; Reason, 1996); although both may be qualitative in forms, the theories have different theoretical assumptions, research practices, and underlying purposes for engaging in the research process. For example, while changing the lives of participants is usually a central criterion for success in action research (Bryman, 2004), it is considered to be an acceptable side-effect in action theory. Action theory typically utilizes qualitative methods to collect data and generate findings about the actions of persons who are jointly engaged in achieving some goal or future state. As I will explain, the 'action-project' method developed by Valach, Young and colleagues currently permits the examination and comparison of variation within cases, and across all the cases within a study, but not between groups of cases. At the same time, however, an interest in examining variation linked to differences that exist between groups of people has begun to emerge in recent action theory studies. For example, a recent publication described the health-related conversations of two distinct cultural groups (Young, Lynam, et al., 2001); a current study has explored career projects in light of the presence or absence of disadvantage in the family (Arato-Bollivar et al., 2002); and a proposal to use the action-project method to examine health maintenance projects, where half the participants have cancer in the family, has been submitted to a major federal funding agency (R.A. Young, personal communication, November 21, 2003). The emergence of these kinds of research questions is somewhat problematic, however, because action theorists are seeking answers to questions that cannot be fully addressed within the existing action-project methodology. The natural solution to this problem is to extend the existing action-project method, and develop way of comparing any distinct sets of cases that may have been identified within an action-project study. There are two questions that must be addressed as a part of this methods-development process: (a) Are between-groups comparisons compatible with the assumptions that underlie the action theory paradigm? (b) What specific procedures could be used within these analyses, to generate findings that go beyond those which are attainable within the existing method? The purpose of this dissertation is to answer these two questions. I do so by engaging in a thorough examination of between-groups differentiation and comparison in action theory to establish that some forms of comparative analysis are, indeed, compatible with the paradigm; proposing a way to expand the existing action-project method to engage in between-groups comparative analysis (my 'action-project differentiation' procedure); and applying the procedure to an existing, non-simulated data set, to demonstrate its capacity to generate comparative findings. Before proceeding, however, it is necessary to more fully delineate what is meant by 'between-groups comparison' in this context because, across the spectrum of qualitative research paradigms, there is wide variation in the kinds of comparative analysis that are considered to be legitimate. Moreover, between-groups comparison in the context of action theory is different from the prediction-oriented between-groups tests of difference that are commonly found in quantitative approaches to research. In the action-project method, these comparisons can be defined as identifying and describing the similarities and differences in patterns of action around a phenomenon of interest, for distinct sets of people; that is, gaining a more elaborate understanding of the interplay between the phenomenon of interest and some identifiable construct that separates participants into distinct groups. Between-groups comparison clearly serves a different purpose in the action-project method than in the hypothesis-testing quantitative approach to research that is more commonly employed in psychology. In the latter,.tests of difference are normally used as part of the process of making inferences about the effects of one variable upon another, within the general population (Bryman, 2004). Specifically, participants are assigned to separate groups on the basis of an independent variable, then overall differences between the groups on some dependent outcome variable is measured and, finally, conclusions about those differences are used to establish a claim that some degree of variation in the outcome variable can be attributed to the grouping variable. In contrast, between-groups comparative analysis in the action-project method is designed to provide more elaborate, qualitative descriptions of phenomena; that is, the process of identifying similarities and differences forms a richer understanding of the action-related phenomenon under study and the construct by which sub-sets of cases were distinguished, without attempting to show that the grouping construct is the direct cause of any differences that are identified. To use a simplified example, a quantitative study examining the effectiveness of a novel manualized treatment for depression might include (a) assigning participants to either the psychotherapy treatment group or a wait-listed control group, (b) measuring all participants' levels of depression prior to and after the course of treatment, and (c) making decisions about the effectiveness of the treatment based on whether changes in level of depression were, on average, better in the treatment group than in the control group.1 In contrast, an action theory study of the same phenomenon might involve examining the mental health promotion actions and projects of a sample of depressed participants, some of whom received a manualized treatment for their disorder, and others of whom were attempting to overcome depression without professional assistance. If those who received professional treatment were found to form a distinct sub-group from those who did not, then it would be possible to describe the specific pattern of similarities and differences in the mental health promotion projects of these two groups. Although such a study would not make the attribution that professional treatment was responsible for any differences that were found, it would generate knowledge about the range of experiences of overcoming depression that are possible, as embedded in the distinct contexts of people's daily lives, where the experience of professional treatment is one contextual factor that was specifically attended to. However, i f the primary purpose of comparison in action-theory-oriented qualitative research.is descriptive, and descriptions of phenomena can be generated from existing action-project within-case and cross-case analysis, then a question naturally arises: why develop a new ' In reality, any true examination of the effectiveness of such a treatment would be more complex, with measures of the fidelity of the treatment administration, follow-ups of the outcome, and multiple comparison groups. However, the logic involved in making conclusions from the study would be the same as in this simplified example. between-groups analytical procedure at all, especially when between-groups comparison has a contested position within qualitative research as a whole? At a purely pragmatic level, the answer is that action theory researchers have demonstrated some inclination to make between-groups types of conclusions, even in the absence of such a procedure (see the Comparative Analysis in Published Action-Project Studies section of Chapter 3 for specific examples). These kinds of findings would be easier to justify, i f a clearly described set of procedures for conducting comparative analyses within action theory existed. A more complete answer to the question is that a between-groups analytical procedure can provide researchers with a richer understanding of phenomena than within-case and cross-case analyses alone. Systematically exploring and delineating an additional layer of contextual complexity (provided by the grouping construct) will increase the amount of information that can be obtained through the action-project method. That is, when different kinds of participants seem to be emerging within a research sample, within- and cross-case action-project analyses may not fully uncover the patterns of findings that are present in people's actions around a phenomenon. For example, gender-related differences in siblings' joint actions would be difficult to identify i f all the participant dyads are examined individually or across the entire sample, without an examination of sister-sister dyads as a group, in relation to the brother-brother dyads as a group. Therefore, addressing the problem of between-groups comparative analysis in action theory will increase the utility of the action-project method and better capture the complexities of some social phenomena than is currently possible. Furthermore, a major hesitation that many qualitative researchers have about conducting between-groups comparisons is not applicable to the present discussion. Many qualitative researchers find statistical comparisons to be objectionable on the grounds that the process of aggregating data for subsequent comparative analysis fails to attend to the unique circumstances of the specific people within the groups. Therefore, it is claimed, statistical analyses generate conclusions based only upon hypothetical 'average scores,' rather than the actual experiences of any given individual (Bryman, 2004). First, the claim that statistics cannot capture the unique circumstances of specific people fails to take into account modern statistical procedures that incorporate individual variation into their models, such as cluster analysis and hierarchical linear modelling. More to the point, however, the comparative analysis procedure that I present does not employ statistical methods of comparison. Moreover, the existing action-project method holds the unique characteristics of each individual case to be important, and as a direct extension of that method, the action-project differentiation procedure adopts the same perspective. As will be seen in Chapters Three and Four, the findings that are generated out of the between-groups analysis are then grounded back in the original cases from which the data were aggregated, precisely to prevent losing touch with the experience of individual participants during the process of making conclusions about what tends to occur for different kinds of people. To reiterate, this dissertation, completed as part of a doctoral program in Measurement, Evaluation and Research Methodology, documents my effort to examine and expand the action-project research method, in order develop a way to engage in between-groups analyses within the action theory paradigm. In so doing, I address two questions that have arisen out of research recent practices within the action theory framework: (a) are between-groups comparisons compatible with the assumptions that underlie the action theory paradigm, and (b) what specific procedures could be used to generate between-groups findings? As will be explained further in the following chapters, in this context, the notion of 'between-groups comparison' is understood to be the identification and description of similarities and differences in patterns of action around a phenomenon of interest, for distinct sets of people. CHAPTER 2: DESCRIPTION OF ACTION THEORY In order to address the research questions properly, it is first necessary to gain an understanding of Young, Valach and colleagues' action theory approach to social inquiry and the 'action-project' research method that is most closely associated with it (e.g., Collin & Young, 1992; Valach, 1988; Valach & Wald, 2002; Valach, Young et al., 2002; Young et al., 2000; Young, Valach & Collin, 1996; Young, Valach, Dillabough, Dover, & Matthes, 1994). Consequently, this chapter provides an overview of action theory, and the action-project method, as they currently exist. I begin with an introduction to its origins and the existing studies that have been conducted using this approach, then proceed with a description of its theoretical assumptions, and conclude the chapter by delineating the specific data collection and analysis procedures that are currently employed by researchers using the action-project method. Action theory is rooted in von Cranach's (von Cranach, Kalbermatten, Indermuehler & Gulger, 1982; von Cranach & Valach, 1984) understanding of human action in everyday life. It adopts a 'person-in-context' frame of understanding that is related to the ecological and contextual perspectives of John Dewey, George Mead and Lev Vygotsky (Young, et al., 1996). Action theory also makes use of the notion of 'project,' a concept that has appeared in the writings of Jones (1996), Larson (2000), Richardson (2000), Riverin-Simard (2000) and most notably, Little (1983, 1989, 1999). The research method associated with Young and colleagues' action theory has been successfully employed to examine a range of different human experiences, including health promotion activities (Young et al., 2000; Young, Lynam, et al., 2001), adolescent identity development (Domene et al., 2003; Marshall et al., 2002), suicidality (Valach, Michel, Young & Dey, 2002) and the counselling process (Michel, Dey, Stadler & Valach, in press). Action theory has also been proposed as a viable framework to study rehabilitation counselling (Valach & Wald, 2002), the effects of divorce on children (Bader & Young, 2002), the vocational rehabilitation of affectively disordered psychiatric patients (L. Valach, personal communication, January 6, 2003), career development for youth with schizophrenia (Domene, 2003), and the transition projects of First Nations peoples (S.K. Marshall, personal communication, December 11, 2003). Finally, the theory has formed the basis of a systematic program of research concerning family influences on adolescent career development (e.g., Young et al., 1999; Young, Valach, et al., 2001; Young, Ball, Valach, Turkel & Wong, 2003), which has provided the data for the applied portion of my dissertation. Theoretical Assumptions of Action Theory Action theory draws a distinction between the actions of people and 'action' as it is used to describe dynamic processes of physical phenomena. In explaining this conceptualisation of human action, Valach, Young and colleagues (2002) describe it as intentional and goal-directed, though not always rational: even i f our actions do not appear to make sense, they are performed as part of the process of achieving some desired end. This goal-directedness is absent in the actions of non-living phenomena, such as the corrosive action of acid on metal. Action is also viewed as something that is constructed, perceived, and interpreted through language and social representation. That is, human action is embedded within a social context, a context that cannot be ignored when formulating an understanding of that action. As such, action is understood to be a complex, multidimensional phenomenon. Perspectives and organisation of action: It is proposed that any particular action can be viewed from three distinct perspectives, each of which provides unique information about a phenomenon (Valach, Young, et al., 2002). Action can be understood from the perspective of 'manifest behaviour'- the readily observable sequence of behaviour involved in carrying out an activity (e.g., an adolescent talking with her psychiatrist about the possibility of altering her medication schedule). A second perspective on action is 'internal processes'- the subjective thoughts and feelings that a person experiences during an activity (e.g., the desire to be off medication; thinking that the psychiatrist does not recognise the improvements in her symptoms). Finally, it is also possible to understand action from the perspective of 'social meaning'- the explanations that people provide when describing their action to others, including the intentions and purposes they provide about the activity in question (e.g., when discussing the doctor's refusal with her parent, the daughter and mother may construct an understanding of the incident based on the doctor being a professional and knowing what is best; or, alternatively, based on the construction that the doctor is sexist, and simply unwilling to trust the opinion of a young woman). In addition to these perspectives of action, the theory proposes a three-tier hierarchical organization of action (e.g., Valach, Young et a l , 2002). At the lowest level of organisation are the 'elements' of an action: the verbal phrases, physical movements, and environmental features involved in the performance of a task. For example, the action of having a conversation may involve elements such as statements of opinion, questions, smiles, shrugging of shoulders, and sitting in an interview room. An element of an action does not occur in isolation, but in sequence with other elements. A sequence of contiguous elements that have a common function is a 'functional step,' the medial level of action organisation. Functional steps are the intentional means by which each participant moves towards their goals, and consist of sub-stages that are present in an action. For example, functional steps in a conversation may include introducing a topic for discussion, presenting an opinion, finding out information, and closing the conversation to engage in some other activity. At the highest level of action organisation are the 'goals' and 'intentional frameworks' of actions. These are the overall intentions and purposes of the people who are engaged in that action. For example, the goal for engaging in a conversation may be to discuss a person's sporting activities, to share important news with another person, or to seek advice. Normally, the goals of an action are reflected in the functional steps and elements that comprise the action. For example, the functional steps taken by a mother whose goal is to find out whether her daughter is enjoying competitive weight-lifting are very different from those involved in convincing her son to attend university. Joint projects: Another important theoretical concept that frames action over time is 'project.' Within action theory, a project is a "goal-directed mid-term process comprising individual and group actions" (Valach, Young, et al., 2002, p. 35). It is goal-directed in the sense that a project is something that an individual or group intentionally works towards; there is an identifiable end state that is being sought. A project is 'mid-term' with respect to its time frame: projects encompass more than what can be accomplished in the immediate term, and yet have some identifiable ending point (i.e., when the goals are accomplished). People engage in a variety of actions to accomplish their projects, including actions undertaken individually, and actions that multiple persons engage in together. Existing action theory research has typically examined the common projects of two or more persons, because a majority of projects that people engage in during their daily lives are social in nature and, therefore, involve joint engagement between two or more individuals (Valach, Young, et al., 2002). Paradigmatic assumptions: Action theory is a post-modern approach to social research. It falls within a "family" of methods that have been given a number of different labels by different qualitative researchers, but are most commonly known as "social constructivist," or more simply, "constructivist" (e.g., Creswell, 2003; Lincoln & Guba, 2000; Teddlie & Tashakkori, 2003). This family of techniques encompasses interpretivism, hermeneutics and constructivism (Cresswell, 2003; Schwandt, 2000), and stands in contrast to both the post-positivist and the post-structuralist paradigms for qualitative research (Lincoln & Guba, 2000). Action theory adopts a form of social constructionism as its approach to generating knowledge (Young, Valach et al, 2002). As such, it falls within the larger umbrella of constructivism. Table 1: Selected Paradigm Positions of Constructivism and Action Theory Issue Constructivism Action theory relativism: local and specifically action-referenced relativism: social constructed social realities realities are constructed, but on the Ontology b a s i s o f a c t j o n s m a t exist independently of the constructor transactional / subjectivist: findings transactional / subjectiyist; findings Epistemology ' are creations are creations (but grounded in - • - , . . external action / data) • , intertwined with validity; inquiry is a permissible, expected consequence Promotion of often incomplete without action on of the inquiry process, but not Change ^ p a r t Q ^ ^ Q participants explicitly promoted ; .'. ; ... understanding, reconstruction,. understanding, reconstruction. Aimoflnquny \ description description ' * u verbal/textual material reflecting verbal/textual material reflecting Form of Data social dialogue social dialogue, internal processes and observation of behaviour Location of ' * shared between inquirer and ' - shared between Inquirer and j*. Control participant . participant '•< Criteria for trustworthiness and authenticity trustworthiness of findings, Quality repeatability of procedures * Although the two paradigms are presented in tabular format for ease of discussion, it needs to be recognised that ways of conducting research are more analogous to a spectrum than a set of disparate categories, with varying degrees of overlap between distinct paradigms, depending on the issue at hand. Many of the underlying assumptions of action theory are similar to the positions espoused by other constructivists. However, as it is implemented through the action-project method, action theory contains a number of features that differentiate it from other forms of constructivist research. A comparison of the major philosophical and theoretical assumptions of Young and colleagues' action theory and 'typical' constructivist research is presented in Table 1. The paradigm positions that I have attributed to action theory were derived from the explicit descriptions of the paradigm provided by Young, Valach and others (e.g., Collin & Young, 1992; Valach, 1988; Valach & Wald, 2002; Valach, Young, et al., 2002; Young et al., 2000; Young, et al., 1996; Young et al., 1994), as well as the implicit assumptions of the action-project method, as it has been implemented in extant research (e.g., Domene et al., 2003; Marshall et a l , 2002; Young et a l , 1999, 2003; Young, Lynam, et al., 2001; Young, Valach, et al., 2001). The descriptions of the stances adopted by 'typical' constructivism are based upon Lincoln and Guba's (2000) latest summation of the assumptions of the major paradigms that currently exist in social science research. Many of the propositions of action theory are consistent with constructivism. For example, action theorists have stated that (a) humans, as generators of action, can only be understood using concepts different from those associated with the study of physical, inanimate objects, (b) language and linguistic encoding of thought and emotion are central to the understanding of action, and (c) the socially constructed meaning of phenomena are the highest level of action organisation. They also state "[through] conducting research within the action theoretical paradigm, we are aware of the role social representation plays in scientific conceptualization, again reminding us of the constructivist nature of the research" (Valach, Young, et al., 2002, p.28). However, action theorists do not consider social construction to be the sole basis of understanding human functioning, "social construction is a broad and complex proposition in contemporary psychology ... we consider it a part [italics added] of the conceptualization of action theory" (Valach, Young, et al., 2002, p.27). Specifically, manifest behaviour and internal processes are perceived to exist independently from any observations that are made regarding them, and can occur even when an individual is^alone. The multiple perspectives of action espoused by action theorists (particularly the fact that behavioural observations are viewed with as much legitimacy as verbalized social meaning) locate it somewhat away from the ontology of relativism found in most constructivist methodologies, and towards a perspective that approaches Cupchik's (2001) 'constructivist realism,' or Rennie's (1999, 2000) 'methodical hermeneutics.' That is, the ontology of action theory incorporates the idea that, even though social phenomena are perceived and defined inter-subjectively, they occur in the external world and exist independently of the person. The implication of this shift away from typical constructivist ontology is that some constructions are a better reflection of a phenomenon than other constructions. Furthermore, when an action is understood from the three perspectives of action that I have described, the resultant understanding is likely to be more complete than knowledge that is constructed from social meanings alone. This distinction is readily evident in the data collection process of the action-project method. As the Valach, Young, et al. (2002) describe, action theoretical research methods differ from many qualitative research methods because in action theoretical research data are sought from three perspectives rather than just one. In the case of much qualitative research, the single perspective is often that of social meaning. The video recording [of joint conversations] provides primary access to the manifest behaviour of the action. The second perspective, internal processes, is accessed through the self-confrontation procedure (p. 51). In their attempt to capture manifest behaviour and internal processes, action-project researchers strive to accurately portray what has taken place in the real world, by discovering the elements and functional steps underlying the goals and overall goal framework (Marshall et al., 2002). Furthermore, interviewers are directed to engage in self-confrontation "as close to the time of the action as possible," in order to preserve accuracy of recollection (Valach, Young, et al. 2002, p. 51). Implicit in this directive is the notion that, as time passes, recollections become less reflective of the situation that existed at the time of the interaction, with temporally distal recollections being at risk of being altered by subsequent experience. This notion presumes the existence of a time-specific external world that is independent of the person, even i f it is apprehendable only through constructivist means. For the sake of convenience, I have labelled the kind of relativism espoused by action theory.'action-referenced relativism,' to distinguish it from other, more traditional, forms of relativism. The atypical form of relativism found in action theory also manifests itself in the epistemology of the paradigm. According to Valach, Young, et al. (2002), although knowledge generation is viewed as primarily a constructive process (with findings that are inter-subjective creations rather than objective facts), it must also be grounded in collected data (i.e., the externally present actions that were present within the research). Turkel (2003) used the label 'grounded subjectivity' to describe this epistemology. To facilitate the 'grounding' of the knowledge generation process in action theory, the mechanisms underlying the action-project research method have been designed to reflect both the social, constructive nature of the knowledge generation process, and the externally present 'reality' of the data (see the following section of this chapter for a description of how this is accomplished). Differences between the action theory paradigm and other forms of constructivism are also evident in the theory's stance on interpersonal communication. In some forms of social constructionism, communication is the entire basis of people's constructions of reality (Lincoln & Guba, 2000). In contrast, the action-project method attends to observations of participants in addition to their verbal reports. Participants' descriptions alone are considered somewhat inconsistent and unreliable (Valach, Young, et al., 2002). If language-based constructions are not sufficient to gain a full understanding of what is occurring, then why are they insufficient? The answer appears to be because social phenomena occur in the external world, and have some existence/form that is independent of people's interpretations and verbal descriptions. This existence/form remains even i f it is not reflected in the language and communication of those experiencing it, and needs to be accounted for in data collection to gain a full understanding of a phenomenon being studied. In terms of promoting change in the lives of research participants, the position of action theory is located somewhere between constructivist paradigms and the stances adopted by most postpositivist approaches. The fact that research involvement induces change in the lives of the participants is not viewed as compromising the validity of a study, because people who volunteer for a specific research study are naturally interested in that topic and, therefore, are predisposed to gain new insight or take new action in that domain of functioning, irrespective of any research involvement Young et al., 2000). At the same time, however, action theory does not actively promote a 'change inducement' stance: studies are not explicitly designed to be interventions, and the empowerment of participants is not a criterion for the success of a study. Change due to participation in research is viewed as a permissible consequence of the knowledge generation process, not something that should be actively sought by researchers. In action theory, the criteria for establishing validity are generally similar to the standards adopted by other constructivist paradigms, and include the plausibility of conclusions and the credibility of the process (Domene et al., 2003; Young et al., 2001; Valach, Young, et al., 2002). Where action theory deviates from some forms of social constructionism, however, is that efforts are made to ensure that the process (though not necessarily the findings) is also repeatable. Repeatability of procedures is the idea that, although researchers are permitted the flexibility to adapt their questions and responses to the needs of the action being studied, some degree of consistency in data collection throughout a study remains necessary, so that other researchers could reconstruct and repeat the processes involved (at least in theory, i f the socio-temporal context surrounding the original study could be duplicated). Action theory is well within the range of typical constructivism for the remaining issues that define a paradigm for research: The aim of inquiry in action-project studies is to understand and provide a description of people's experiences in daily living, not to formulate general laws or predict future behaviour. The kinds of data that are collected in the action-project method are predominantly verbal and textual (although this paradigm also involves observations of manifest behaviour). Finally, efforts are made to share control of the process and products of research with the participants, which mirrors common research practice in other kinds of constructivist research. Although not a theoretical stance per se, the action-project practice of identifying and discussing findings across entire samples of participants contains certain assumptions that also require explanation. It must first be understood that the findings generated from cross-case analysis in the action-project method are not some 'mean score' of how the phenomenon manifested within the group, nor is it assumed that a particular finding is manifested in all individuals within the sample. Instead, the cross-case analysis process generates a description of the patterns of findings that are frequently present or 'common' in the group of participants being studied. Therefore, the major assumption underlying cross-case analysis in the action-project method is that it is possible and sensible to speak of findings that are common within a group of persons. If the same principle is extended to data aggregation in the context of between-groups analysis, then a further assumption must be made- that it is possible and sensible to speak of similar and contrasting aspects of findings that are common to separate groups of participants. As I explain in. chapter 3, these assumptions can also be found in other forms of constructivist research. Furthermore, they mirror the processes that people engage in during every-day living (i.e., speaking of the general tendencies and characteristics of people as a whole, and talking about how distinct groups of people tend to differ or be similar to each other without referring to average scores). However, because action theory privileges the contexts in which people's actions are embedded (Valach, Young et al., 2002), aggregation of data across the group needs to be balanced with retaining the full configuration of information from the individual cases that make up the group. In the existing action-project protocol, this retention of contextual information is accomplished by embedding the general cross-case findings within specific examples. In the procedure that I have created, this practice of grounding the findings in the actual cases is preserved: group-related findings are checked against the specific situations of the cases from which they are derived, prior to making any final conclusions (see the Description of the Differentiation Procedure section of chapter 3 for details). 77ze Existing Action-Project Method In their discussion of data management and analysis methods in qualitative research, Huberman and Miles (1994) define 'data management' as comprising the processes of data collection, storage and retrieval, within-case analysis, and cross-case analysis. Their approach serves as a useful framework for my discussion of the procedures involved in the action-project method. As a result, I will describe the action-project method's data collection techniques, issues regarding data storage and retrieval, the within-case analysis process and, finally, cross-case analysis, as reflected in published action theory studies. Although I present the method sequentially here, the actual processes of data management and collection in the action-project method occur simultaneously in practice; it is the norm for action-project researchers to be in the latter stages of analysis with some dyads of participants by the time they begin data collection with other dyads. Even within a single dyad of participants, management and analysis occur in a recursive sequence, rather than sequentially. For the purposes of this description, however, I describe the sequence of conducting action-project research as follows: an initial data collection interview, followed by preliminary analysis, then a presentation of that analysis for feedback in a second interview, next a six-month monitoring period where data is collected via telephone calls and self-report logs (and which is informed by the results of the first and second interviews) and, finally, a third data collection interview followed by an overall within-case analysis process. After the data has been examined within each case, then findings across the sample are examined all together, to generate a final set of cross-case descriptions. Data Collection. The action-project research protocol developed by Young, Valach and colleagues assesses the actions and goal-directed language of pairs of people engaged in joint projects over medial lengths of time. For example, a study may examine the action of a youth and one parent, working towards self-identified goals associated with the youth's future vocational development, over the course of approximately six months. Information is gathered using multiple techniques that allow all three perspectives on action to be reflected in the data. Specifically, the research protocol utilizes interviews, observation of conversations, a self-confrontation procedure, participant self-report logs, and periodic telephone monitoring, all of which occur in the previously described structured sequence. The first data collection interview contains several stages. The initial stage of the interview is an introduction period involving the dyad members and two interviewers. At this stage, participants are asked general questions about themselves, their perspectives and goals around the phenomenon of interest, and what is each person's role in working on their joint project. The introduction period serves to increase the participants' comfort with speaking while being recorded in the research setting, and primes them to actively think about the topic being studied. Following this initial stage, the dyad is invited to engage in a 15 - 20 minute conversation in the absence of the interviewers. Interviewers suggest that the participant discuss one or more of the topic-related issues that were raised in the preceding discussion. Ultimately, however, participants are permitted the freedom to direct the course of the conversation for themselves. The self-directed nature of the resultant conversation allows a dyad to communicate using its natural, on-going style of interaction. Upon conclusion of the joint-conversation, each participant separately engages in a 'self-confrontation procedure,' with one of the interviewers. This final portion of the first interview involves the participant viewing a videotape of the joint-conversation, pausing the tape at regular intervals, to comment on his or her internal processes. The interviewer explicitly asks the participant to describe his or her cognitions and emotions for each segment of the conversation. A l l stages of the first interview are video- and audio-taped for accurate collection of data. An initial analysis of the information from this first round of data collection is then conducted, in order to (a) generate a description of each participant's opinions and intent during the conversation and (b) tentatively identify the joint projects that the dyad is engaged in, around the phenomenon being studied. (See 'data analysis procedures' section for details.) Three narrative summaries of the resultant information and conclusions are then constructed for presentation to the dyad in the second interview: two reflecting the perspective of each individual dyad member, and one reflecting their joint engagement in ongoing projects. In the second interview, the two narrative descriptions of individual dyad members' perspectives are first presented to them alone, to elicit feedback regarding its accuracy from that participant's perspective, and to correct any errors in interpretation. Then, both dyad members and both interviewers engage in a group discussion together, to allow participants to share their narratives with each other, and negotiate which of their on-going joint projects should be the focus of their subsequent research involvement. The interviewers take an active role in this negotiation, to ensure that (a) the chosen project is one in which the dyad is presently engaged in their daily living (i.e., to guard against participants selecting something entirely new, due to their involvement in a research study) and (b) the choice is one that has some relevance to the phenomenon being studied. The nature of their chosen project is also clarified, to some degree, by embedding it in specific activities (e.g., " i f the project is working, we will be talking more openly, and will find out about disabled student support services at the local colleges"). It is recognised that projects often shift over time. Subsequently, participants complete log entries concerning their engagement in project-related activities over a period of six months. Log entries include a description of the activity, the participant's internal reactions at the time, and his or her intended goals. Each member of the dyad is asked to complete his or her own log, and to respect the privacy of the other participant's records. In addition to this self-report data, the interviewer working with each participant maintains fortnightly telephone contact to discuss progress towards the joint project, which the interviewers record in their own logbooks. Following the monitoring period, the third interview occurs. This interview is similar to the first interview, with participants engaging in another joint conversation and self-confrontation procedure, related to their project. After the self-confrontations, short, semi-structured individual interviews are conducted with each dyad member, in order to follow up on any pertinent issues that may have emerged over the course of the monitoring period. Interviewer notes and videotaped records of all stages of the final interview form the data sources in this portion of the action-project protocol. Data storage and retrieval. Existing publications do not describe the processes that are involved in storing and retrieving data in action-project studies. However, personal experience in conducting such research, combined with on-going discussion with one of the creators of the method (Richard Young), allows for a coherent picture to emerge. Although specific storage media have changed over time (ranging from audio-tapes and paper transcripts to digitized video and electronic storage of documents), most action-project studies retain common sets of information. Typically, the data record consists of (a) video or audio recordings of all sections of all three interviews; (b) written documents of the three narratives presented to participants in interview two (including any changes that were made); (c) participant and interviewer logs from the monitoring period; minutes or audio-taped records of research team meetings; (d) memo books, where research team members record all their reflections and thoughts about the study, throughout the research process; (e) written summaries of the product of the final within-case analysis for each dyad; and (f) transcripts of the joint-conversation stages of interviews one and three (transcripts are initially generated by transcribers hired for that purpose, but reviewed for accuracy by the original interviewers). Depending on storage medium, data retrieval may consist of pulling documents and r tapes out of physical files kept in filing cabinets, and/or electronic retrieval from computer hard drives. Because within-case data analysis precedes cross-case analysis, the data record is indexed primarily by case, except when doing so violates standards of research ethics (e.g., participant screening forms, which contain identifying information, must be stored separately) or makes no sense conceptually (e.g., the storage of researchers' memo books, which often contain reflections of the process in its entirety). In addition, contained within the output of the preliminary and final within-case analysis, are clear references to the raw data from which findings are derived. For example, the summary display document may contain the statement "The daughter perceived the mother to be supportive of her efforts (FASC: I23-A25)," meaning that this claim is supported by data from interviewer comment #23 to adolescent comment # 25, of the Final Adolescent Self-Confrontation interview. Within-case data analysis. Information from all three data collection times, the self-report logs, and telephone monitoring reports are examined via a systematic, action-theoretical form of qualitative content analysis (Valach, Young, et al., 2002). The content and focus of this analysis is grounded within action theory. The process for deriving constructs and themes relies upon a team of analysts, consensual decision-making, and systematic checking of conclusions against the original textual material. The concrete procedures of analysis in the action-project method are somewhat similar to Hi l l , Thompson and Williams' (1997) Consensual Qualitative Research, but the overall process can be characterized as more of a hermeneutical dialogue than Hi l l et al's purely data-driven, discovery oriented approach. That is, in the action-project method, potential interpretations and conclusions are repeatedly referenced to the theoretical understandings of action that have been adopted by action theory, as well as the nature of the data itself. In addition, the pre-existing knowledge bases and perspectives of the individual members of the research team enter into the analysis process, because understandings are constructed by multiple researchers, working in concert. As a result, coding decisions are arrived at through discussion between the entire research team, with differing possibilities (all of which reflect the collected data, the theory of action, and prior knowledge of the phenomenon being studied) being considered and discussed until a consensus interpretation is reached. However, coding decisions are also informed by the elements and functional steps that are present in the observed action of participants, because it is assumed that these lower levels of organisation reflect the nature of the higher (i.e., goals) level of action organisation. Furthermore, attending to the elements of the action serves to ground the analysis process in participants' actual experiences, thus reducing the risk of constructing interpretations that are unwarranted from the observed data. In addition, the three perspectives on action also inform the coding process, because the themes and findings that are derived must be evident across manifest behaviour, internal processes, and social meaning. The coding process, therefore, is primarily inductive (with specific codes and patterns emerging from the data rather than existing beforehand), although the propositions of action theory and researchers' pre-existing knowledge bases provide a framework for the code-generation process. During data analysis, particular attention is paid to the two joint conversations, which are analysed in the following manner: First, the elements and functional steps of each portion of the conversation are identified. Based on this, the goals of each participant for that segment are derived. Goals are assessed within socially meaningful units, and usually change over the course of the dialogue. Finally, each participant's 'intentional framework' (his or her overall purpose or goal for the conversation in its entirety) is identified. Also, information from the self-confrontation procedure is incorporated into the analysis, to obtain a greater understanding of the interna] process of the participants at the time. In the first round of analysis, information derived from this procedure is used to generate three descriptive narrative summaries of the actions and goals for each participant from the initial interview, and to tentatively identify the projects in which the dyad is engaged on a regular basis, during the course of their daily living. In the second round of analysis, this same action theoretical content analysis is applied to all the information collected from the dyad, to uncover the goals, actions taken, progress made, and themes that emerged as they worked towards their project. The product of this final within-case analysis is a written summary displaying the themes, issues and experiences that were the most salient for each dyad over time. The format of this summary has varied across studies, with some researchers employing a narrative format to report their findings (e.g., Young, Valach, et al., 2001), while others have utilized a more formal, sectional style, with divisions for each portion of the data collection process and separate sections for the various themes/topics that are of interest to the researchers (e.g., Arato-Bollivar, et al., 2002). The product of this qualitative analytical process is a series of descriptions of the projects of individual dyads around a topic of interest, the action that they have taken over time to fulfill those projects, changes that have occurred in the project and dyadic relationship over time, and how the project of interest is situated in the life of the participants (e.g., is it central or peripheral, how does it relate to other projects in which the dyad is also engaged?). These descriptions serve to enhance understanding of how a phenomenon of interest is manifested in the lives of specific pairs of individuals. Cross-case analysis. After within-case analysis has been completed for all participants in the sample, the action-project method allows for identified constructs and themes to be examined across the group in its entirety (Valach, Young, et al., 2002). This process attends to both the aspects of a phenomenon that are unique to a specific dyads in the sample, and those that are similar across the group (e.g., Young, Lynam, et al., 2001; Young, Valach, et al., 2001; Young et al., 2003). Cross-case analysis involves research team members reviewing the themes and categories present within each dyad, and reaching decisions, as a group, regarding what material is salient across the sample. Specific procedures differ across specific studies but, typically, all summary displays of the within-case analysis process are reviewed and examined together. Decisions are then made regarding which themes/constructs/codes can be judged to be commonly found among the participants, based on the consensus opinion of the research team and specified criteria for what makes a worthy of note as being present across cases. The criteria that are usually used in action-project cross-case analyses are the importance and significance of the themes, and how often they occur within the sample. This process is similar to the cross-case analysis strategy found in the Consensual Qualitative Research method (Hill et al., 1997), but tends to be more fluid, and is hermeneutic rather than purely discovery oriented in nature. Moreover, in contrast to Hi l l and colleagues' (1997) classification system of 'typical' (applying to half or more of the cases) and 'variant' (applying to less than half but more than two cases) commonalities, action-project researchers make more holistic judgements regarding cross-case similarity, to conclude that a theme is common, or that it is not. For example, in a recent study of family career development projects in Chinese Canadian families, Young et al. (2003, p. 292) described the guideline for inclusion in the report of cross-case findings as being the presence of "significant similarities across the data sets for most of the families." The final stage of the cross-case analysis entails another review of the data, at the within-case level, to confirm that the cross-case findings do, indeed, reflect the lived experience of the individual dyads of participants, and to ground those findings within specific examples. Several iterations of this back and forth process may need to occur before a satisfactory overall picture emerges. Discussion and working towards consensus among team members occurs at all stages of this analysis process. Information that is generated by cross-case analysis process, over and above within-case analysis alone, includes (a) the 'properties' (i.e., salient content and descriptive characteristics) of the joint projects that tend to occur around a phenomenon of interest, and (b) the organisation and types of relationships that may exist between the projects pertaining to the phenomenon of interest, and other life projects that dyads are pursuing at the same time. It must be noted, however, that this cross-case analysis process has not been utilized previously to examine or compare patterns of variation across distinct sub-sets of participants. Instead, its focus has been to delineate the patterns that are worthy of notice across the entire sample of participants. Criteria for rigour. The action-project method incorporates a number of safeguards to ensure that a high level of rigour is maintained throughout the analysis process. These safeguards meet existing standards for conducting qualitative research in the psychological sciences (e.g., Elliott, Fischer & Rennie, 1999; Krefting, 1991; Lincoln & Guba, 1985). One such standard is the presence of a detailed audit trail, permitting outsiders to review the research process and follow the train of logic leading to the conclusions that were made (Lincoln & Guba, 1985). In the action-project method, this trail is created by recording and retaining descriptions of all potential participants, along with extensive records of the data collection and analysis process (as I described in the "data storage and retrieval" section). To reduce the likelihood that any given researcher's biases or perspective might systematically distort the categories or conclusions drawn from the data, triangulation of researcher perspectives is required in all phases of analysis. This process is similar to the techniques commonly used to achieve inter-rater agreement in quantitative research (Young et al., 2001), although it is somewhat more complex, and involves two separate stages. In the first stage, two analysts work together to code all the raw data. Discrepancies in categorisation are discussed until a consensus opinion as to the correct interpretation is reached. The conclusions of the two analysts (for all the coding, not only the ones where a discrepancy occurred) are subsequently presented to other researchers, familiar with the method, for review. In this second stage, the peer reviewers are able to question the original analysts as to how their conclusions were made, and discuss other viable interpretations. Feedback from the review process is incorporated into the final product, producing findings that incorporate the original perspectives of all the members of the research team, and reflect a consensus decision as to what codes best fit the data. This triangulation process is performed at both the within-case and cross-case analyses, to ensure that none of the conclusions are the product of any single researcher's preconceived expectations. The action-project method also incorporates member-checking strategies. This commonly utilized method of establishing the credibility of qualitative research involves presenting research materials and findings back to the original participants, to ensure that the researchers have accurately captured their point of view (Lincoln & Guba, 1985). Presentation of the narratives and tentative joint-projects to participants after the first round of analysis reflects this member-checking process, and subsequent analyses of data are informed by the feedback that participants provide during the discussion of their narratives. Finally, the rigour of the results is enhanced through the use of multiple sources of information. Specifically, findings regarding each participant dyad are made from the interviews, self-report logs, self-confrontation procedure, and observations, to provide a broad understanding of the phenomenon in question. Convergence of information from these disparate sources provides a check against the possibility that the pattern of findings is the result of a particular data collection technique, or way of phrasing a question (Krefting, 1991). Moreover, the value of action-project findings is enhanced by including information from all three perspectives on action within the data collection and analysis process (Valach, Young, et al., 2002) That is, having descriptions that are reflective of manifest behaviour (observed from videotapes of joint conversations and interviews), internal processes (reported in the self-report logs and during the self-confrontation procedure), and social meanings (present in self-report log information, and discussions with interviewers during the three interviews) are more comprehensive than findings that could be derived from data based on any single perspective of action (e.g., only behavioural observation, or social meaning constructed from interviews). There are a number of components within this description of data collection, analysis and rigour that are essential to the action-project method. At the level of data collection, it is necessary that a phenomenon be understood from all three perspectives of action. Typically, this means using multiple sources of data (e.g., interviews, observation, physiological indicators of emotional responsiveness) in order to adequately tap manifest behaviour, internal processes, and social meaning. Another essential feature is to collect data prospectively over time (as reflected in the monitoring period and third interview).2 Specifically, it is important to collect data about people's engagement in joint action over a sufficient length of time to be capable of " I hesitate to include this as an "essential" component, because a substantial portion of action-project studies have omitted this aspect of the research protocol. Collecting data over time does, however, remain an important aspect of the method, and prospective examination of action over time seems more in keeping with the "action-referenced relativism" ontology of action theory than inferring changes from participants' self-reports alone. uncovering any changes that may occur. Another essential component of the method is that the unit of analysis in an action-project study is 'joint action'- the nature of a phenomenon as it is reflected in the activities of people engaged together to fulfill their goals and intentions. It is also essential that the analysis process be guided by the tenets of action theory, because the method was designed as a way to research the intentional action of human beings: the connection between action-project as a method and the action theory in which it is based is strong enough to render questionable any action-project analysis that is not grounded in an action theoretical understanding of the social world. The final essential component of the method is the use of consensus decision-making between multiple researchers as the basis of determining what patterns of findings are present in a data set. The involvement of a research team, working together to form common interpretations, is a required component of the analysis process, and serves to enhance the rigour of any action-project study. Overall, Young and colleagues' action theory can be understood as a distinct paradigm for conducting research in the social sciences, with its own set of theoretical assumptions. It is a specific form of constructivism that acknowledges the existence of a shared external world, against which any researcher's specific constructions can be evaluated. It also has specific data management and analysis protocols associated with it- the action-project research method. My dissertation is an examination and extension of the action-project method and, as a result, accepts action theory as its underlying paradigm. One difficulty with the action-project method, as it is currently conceptualized, is that it contains no established procedure for examining the ' actions and projects of one set of participants, in relation to the actions and projects of other, distinct, sets of participants. Other forms of constructivist qualitative research, such as grounded theory, phenomenology, and narrative research, have also wrestled with the issue of attending to variation between-groups (e.g., Glasser, 1978; Giorgi, 1985; Polkinghorne, 1995; Murray, 2003; Strauss & Corbin, 1990). Fortunately, their resolutions to the question of between-groups comparisons can serve to illuminate the territory, and provide possible avenues for me to pursue in developing such a procedure for the action-project method. This will be discussed further in the following chapter. CHAPTER 3: ADDRESSING THE RESEARCH QUESTIONS In general, the notion of comparison is much less well developed in qualitative research than in quantitative research, for a number of reasons. There is ambivalence among many qualitative researchers as to whether it is theoretically appropriate to compare findings between individuals, to say nothing of comparing sets of individuals (Bryman, 2004). Moreover, even when it has been deemed acceptable, the issue of how to perform qualitative comparisons is complicated by the plurality of analytical methods that exist (Ayers, Kavanaugh, & Knafl, 2003). In order to make this review of qualitative comparison a manageable task, I have limited myself to the research approaches that are most closely related to action theory- other forms of social constructionism. Thus, I have deliberately excluded qualitative research traditions that make the assumption that experiences of different individuals are incommensurable with each other. I have done so for several reasons: (a) action theory makes no assumption of the incommensurability of different people's experiences; (b) many existing action-project studies involve some form of clustering or grouping of participants together to examine patterns of action across the group; and (c) the identification of general findings that apply to groups of persons is also justified within the underlying paradigm of action theory. This chapter begins with a discussion of comparative analysis in some other forms of qualitative research (grounded theory, phenomenology, and narrative analysis), and how those traditions have informed the development of the action-project differentiation procedure. Next, I review all published empirical studies that have employed some form of the action-project method, focusing on issues related to comparison. Then I describe how some forms of between-groups comparative analysis are, indeed, compatible with the theoretical assumptions of action theory. Finally, I conclude this chapter by introducing the 'action-project differentiation procedure,' a formal protocol for conducting between-groups comparisons in action theory research. I have decided to use the label 'differentiation procedure,' rather than 'comparison procedure,' to highlight the distinction between the kinds of comparative analysis found in the action-project method, and the kinds of comparisons that are most commonly used in psychological research (i.e., statistical tests of difference). Comparative Analysis in Other Types of Constructivist Research Grounded theory. Although some would claim that grounded theory is more post-positivist than constructivist (Rennie, 2000), this qualitative method has been responsive to trends and changes in thinking about social research (Strauss & Corbin, 1994). Moreover, the fact that some researchers have begun to approach grounded theory from relativistic and interpretive perspectives (Charmaz, 2003; Rennie, 2000), combined with the fact that the methodology includes interesting approaches to the grouping of data and the issue of comparison, provides sufficient reason to include it in the present discussion. The grounded theory methodology of Glaser, Strauss and colleagues has been modified and elaborated upon over the past four decades (e.g., Glaser, 1978; Glaser & Strauss, 1967; Strauss & Corbin, 1990; 1998) but, throughout this time, its purpose has remained constant: to develop theory out of data that has been systematically collected and analysed. Specifically, theory is formed by allowing categories, constructs, and relationships between constructs to arise out of collected data, with ideas that are formed earlier in the data collection process being tested and refined with subsequent participants and/or follow-up interviews (Charmaz, 2003), a reversal of common practice in quantitative research, where theoretical constructs are first operationally defined, and then tested against collected data. Although its creators do not explicitly speak of 'grouping participants' within grounded theory, combining data from different persons does occur. Specifically, the theory generation process normally requires data to be collected from multiple cases (Strauss & Corbin, 1994), and benefits from having data from multiple sites or settings (Glaser, 1978). A l l resultant information pertaining to a particular construct is then considered together, regardless of which participant it came from, implying grouping of data across cases. In fact, the very notion of testing and refining emergent constructs by collecting more data from new participants, and from participants in different settings, contains an assumption that these persons can somehow be grouped together, at least on a conceptual level.3 The critical point, however, is that it is relevant information across the data set that is grouped to form a complete understanding of a construct or theoretical category, rather than the actual cases themselves (Charmaz, 2003). For example, instead of grouping participants according to structural characteristics (e.g., demographic variables) at the beginning of the analytical process, the potential influence of these characteristics is explored by reviewing data (collected from persons with and without the characteristic in question) for information pertaining to any differences that may be associated with it. As Strauss and Corbin describe, "the general procedure is to ask, What [sic] is the influence of gender (for instance), or power, or social class on the phenomena under study? -then to trace this influence as precisely as possible, as well as its influence flowing in reverse direction" (1994, p. 276). The implication is that, although cases are not grouped by gender, socio-economic status, or other differences in an a priori sense, the influence of these variables can be examined by seeking the relevant information from the data set (which includes information about the respondent's gender, socio-economic status, etc.). This grounded theory approach of grouping information rather than cases requires a slightly different conceptualization of the notion of comparison. Specifically, grounded theorists refer to 'constant comparisons,' which involves comparing, on an on-going basis 3 At the same time, however, grounded theorists make no assumption that all participants are interchangeable units. Instead, they accept individual variation in experience as adding to the fullness of understanding about a particular theoretical construct / relationship between constructs, and actively promote the collection of information from persons with differing perspectives (Strauss & Corbin, 1994). throughout the analysis process, (a) data with other data (i.e., from another participant, or from the same participant but at a different point in time), (b) data with the criteria for specific theoretical categories that are generated, and (c) categories with other categories (Glaser & Strauss, 1967). In this context, 'comparison' means examining the data and/or categories for similarities and differences, and using those similarities and differences as the basis for making decisions about the scope of each construct within the emerging theory (Charmaz, 2003). The mechanism behind this form of comparison is to (a) place the incidents to be compared next to each other (mentally or physically), (b) examine the properties and dimensions that characterize those incidents and finally, (c) formulate a judgement as to whether they are sufficiently similar to be considered the same (Strauss & Corbin, 1998). In this context, "properties are the general or specific characteristics or attributes of a category, [and] dimensions represent the location of a property along a continuum or range" (Strauss & Corbin, 1998, p. 117). The authors note, however, that the purpose of comparison is not to identify difference as an end in itself, but to use that information to gain a better understanding of the phenomenon being studied. Grounded theory introduces two ideas that are important for solving the problem of between-groups comparison in action theory. The first is the possibility that adequate conclusions regarding similarities and differences may be derived through researcher judgement (informed by an extended examination of the characteristics of the phenomenon being studied). Grounded theory methodologists also raise the possibility of exploring what could be considered group differences, by asking questions about how the differentiating characteristic influences the phenomenon of interest during analysis of the entire data set. That is, it may be possible to make conclusions about group differences by keeping those differences in mind when analysing the data, rather than requiring the different groups to be analysed separately. For example, in a study where researchers are interested in how the presence of skin cancer in the family may influence people's engagement in sun protection projects, participants who have a relative with skin cancer could be asked, "how does uncle Jim's diagnosis influence your thoughts, feelings, and actions around preventing skin cancer?" At minimum, the interviewers should keep that question in mind as they collect data from this particular dyad. Phenomenology. Based on the philosophy of Husserl, phenomenology as a research methodology seeks to understand the lived experiences of persons as they naturally occur (Giorgi, 1985); that is, "phenomenology seeks the psychological meanings that constitute the phenomenon through investigating and analysing lived examples of the phenomenon within the context of the participants' lives" (Giorgi & Giorgi, 2003, p.27). The underlying purpose of the method is to gain knowledge of the content and interconnectedness of meanings that underlie lived experiences, as opposed to measuring the frequency of their appearance or amount of similarity between them (Smith & Osborn, 2003). This knowledge is generated by capturing phenomena of interest as they are experienced in the life and context of individual participants (usually retrospectively, through open-ended interviews), deriving the meanings of that experience for each case person and, lastly, discerning the 'psychological essence' of the phenomenon across cases (Giorgi & Giorgi, 2003; Lemon & Taylor, 1997). Although this essentialist position is rather different from the social constructionist assumptions of action theory, the analytical techniques of phenomenology more closely match those found in the "constructivist" family of methods than those that are typical of post-positivism. It, therefore, remains useful to examine phenomenological research practices in generating ideas for conducting between-groups analyses in action theory, even though the underlying paradigmatic assumptions of phenomenology do not fully match those of constructivism. Mechanisms of analysis in phenomenology include (a) bracketing (becoming aware of, and attempting to limit, the influence of other sources of understanding about a phenomenon being studied), (b) close engagement with the data (reading and rereading until the researcher has entered the phenomenological world of the participant), (c) phenomenological reduction (data transformations designed to make explicit the psychological meanings underlying participants' descriptions of their experiences), and (d) intuiting / logical reflection to organize the meanings into general themes and determine the relationships between those themes (Giorgi, 1985; Giorgi & Giorgi, 2003; Lemon & Taylor, 1997; Smith & Osborn, 2003). The grouping that occurs in phenomenology is not of pre-defined structural characteristics along which the data set is theorized to systematically vary; instead, data are grouped according to the themes that emerge from the analysis process. For example, i f participants Alice, Carlos, and Dana make statements whose meanings are judged to reflect the theme of 'Youthfulness,' then those statements are grouped together and utilized to gain an understanding of Youthfulness as it applies to the phenomenon of interest. Furthermore, i f Bianca, Dana and Ephraim make other statements that are judged to reflect the theme 'Zest for Life,' then those statements are grouped together to formulate an understanding of the Zest theme, as it relates to the phenomenon. If, upon further reflection, Youthfulness and Zest for Life are judged to be two aspects of a common experience, then the researcher would attempt to summarize the essence of that single underlying theme by collapsing the various individual meaning statements from the two initial categories into a single one. If, however, it was evident that Youthfulness and Zest for Life are not a single construct, then the two sets of meaning statements would not be further grouped together. There are two aspects of phenomenological grouping and comparative analysis that are particularly relevant to the work that I am attempting to accomplish in this dissertation. First, in phenomenology, there is no a priori grouping of sub-sets of participants according to some structural characteristic (Giorgi, 1985). Instead, grouping occurs once the within-case analysis is complete, and involves seeking common meanings within the narrative descriptions of the participants. When common meanings across cases emerge in a phenomenological study, the statements reflecting that meaning are grouped together under a descriptor that reflects those meanings- a theme. Second, the primary mechanism for deciding what statements should be combined is the researcher's own judgement that the meanings derived from the analysis are sufficiently similar to warrant their being understood as parts of a single theme. There is no mathematical decision-making rule. Instead, it is assumed that the researcher will have sufficient analytical skill to be able to identify and tease apart different themes, and that themes and meanings can be readily apprehended once the phenomenological reduction process has been applied to each participant's description. Comparison in phenomenology involves identifying the similarities and differences in the meanings that underlie participants' descriptions of their experience with a phenomenon, as well as the similarities and differences in the overall themes that emerge about a phenomenon. Decisions regarding the comparability of meaning statements (or themes) are made via logical reflection and/or intuiting the structures that underlie the phenomenon, based upon (a) the data that emerges from the within-person analysis, (b) the researcher's own experience with the phenomenon of interest, and (c) an understanding of the existing literature about the phenomenon (Giorgi, 1985). It must be noted that, while external sources of understanding about a phenomenon are bracketed out through most of the phenomenological analysis process, they are allowed to enter the final integration / description phase of a study, typically the phase where themes are compared with each other (Lemon & Taylor, 1997). The process of analysis in phenomenology provides useful hints for the creation of a between-groups analysis procedure within the action-project method. As with grounded theory, in phenomenology, researcher judgement is the primary mechanism for deciding whether units of information are comparable, and identifying the specific patterns of similarity. This judgement is based upon a thorough understanding of the data (derived from successive levels of within-case coding and phenomenological reduction), and is informed by existing bodies of knowledge about a phenomenon (at least during the final phase of the research). Because it relies on the researcher's judgement, comparative analysis in phenomenology appears to be an interpretive process. The notion of comparison as an interpretive process that relies on researcher judgement and close engagement with the data is one that fits well with the existing action-project method. Although meaning is only one of the perspectives from which action is understood in action theory, the idea that distinct groupings should be formed from the findings of the within-case analysis process is one that may be valuable for an action-project differentiation method. Specifically, it may be more useful to bracket out assumptions regarding what kinds of distinct sub-sets exist, until the completion of the within-case and initial cross-case analysis. Rather than using structural characteristics (e.g., age, socio-economic status) to form the groups prior to data collection, decisions could be made on the basis of the configuration of findings from the cases themselves. For example, i f an action-project study was designed to compare action that occurs in disadvantaged versus non-disadvantaged families, it may not be the best strategy to form two distinct groups when selecting families to participate in the research. Instead, assumptions and existing knowledge about 'disadvantage' should be bracketed out of the initial data collection and analysis process because, i f disadvantage is a contextual factor that is important to families' actions, then it should emerge naturally in within-case analysis. After completion of this initial analysis, participants for whom the theme of 'disadvantage' emerged as salient could then be grouped together, and compared with participants whose action did not reveal that disadvantage was an important contextual factor in their lives. Narrative research. The concept of narrative, as defined in narrative forms of qualitative research, is a sequence of events that is emplotted (i.e., disparate actions and incidents are imputed with meaning by means of a plot), and functions as a way to organize events that occur over time, into a coherent whole (Murray, 2003; Polkinghorne, 1995). Narratives are diachronic (containing information about the temporal sequencing of events), and tend to be structured with a beginning, middle, and end (Murray, 2003; Polkinghorne, 1995). Attempting to summarize narrative analysis is a complicated endeavour, because the term encompasses a range of research practices and the analytical process is more individualized and less systematized in narrative analysis than in either grounded theory or phenomenology (Manning & Cullum-Swan, 1994; Mishler, 1986). Polkinghorne (1995) identifies two broad categories of narrative-oriented analysis that are prevalent in the social sciences: 'narrative analysis,' and 'analysis of narratives.' He states that the research task in narrative analysis is "to develop or discover a plot that displays the linkage among the data elements as parts of an unfolding temporal development culminating in the denouement," the outcome of which is a story that captures the phenomenon being studied (Polkinghorne, 1995, p. 15). This form of narrative research is of limited relevance to the present discussion because it is a within-case analytical strategy- when more than one case is being studied, the researcher generates separate stories for each case, rather than one overall summary for all the cases. In contrast, analysis of narratives (defined as the identification of common themes or structures among data that take the form of stories) provides more useful information for the development of the action-project differentiation procedure, because it more frequently involves cross-case analysis of groups of stories (Polkinghorne, 1995). In the analysis of narratives, grouping appears to occur early in the research process, at the participant selection phase. Participants from whom stories are to be elicited are chosen, in part, because they are members of a group that is of interest to the researchers (Murray, 2003). For example, i f a narrative analyst was interested in conducting research on the illness and recovery stories of women with lung cancer, all the participants would have to be women who have lung cancer; women with other illnesses or no illness, and men with lung cancer, would be excluded from the sample. Because all the participants share some degree of commonality around the topic being researched, it is assumed that cross-case analysis of their stories is possible. By definition, all the collected narratives are reflective of, for example, the range of illness and recovery experiences of women with lung cancer. The same logic underlies the selection of participants for comparative analysis of narratives between groups. For example, in Gergen and Gergen's (1987) study of the patterns that are present in life narratives, the authors selected one group of 18 to 21 year olds, and a second group of 63 to 93 year olds, for comparison. They justified their selection simply because doing so would answer their questions of interest: How do the younger and older samples compare in their self-narrative accounts? Do persons at all ages tend to view their lives as comedy-melodrama, or by extension, romantic sagas? Are the aged more likely to account for themselves as living out a happily-ever-after melodrama? Or, as often suspected by the young, is the process of growing older simply an extended regressive narrative? (p. 134-135). In terms of the analytical process, different narrative researchers attend to different facets of the narratives that they are analysing, such as (a) identifying poetic or plot structures associated with people's stories about a phenomenon, (b) summarizing recurrent content or issues that are raised, (c) examining the temporal structure of different narratives, (d) or relating individuals' narratives to different levels of social context (Murray, 2003). Although Mishler (1986) identifies certain exceptions to the rule, the mechanism that is most frequently used for generating general conclusions out of a collection of narratives is active reading and "playing" with the accounts, until appropriate meanings and conclusions emerge (Polkinghorne, 1995); 'appropriate' in the sense that they fit the theoretical framework of the researcher, as well as the material that has been recounted by the participant. Therefore, the process of comparative analysis in this form of qualitative research appears to be forming judgements based upon a close examination of the data, but also incorporating the researcher's prior knowledge in a way that is more interpretive than is typical of grounded theory or phenomenology. This strategy of incorporating the researcher's theoretical framework into the process of analysis is one that parallels the hermeneutic analysis process found in the existing action-project method and, therefore, is also relevant to the differentiation procedure that I have created. Moreover, allowing one's analysis to be guided by pre-existing knowledge and theory does not necessarily contradict the previously mentioned suggestion of bracketing and incorporating prior knowledge only at the end of the process. The two notions remain congruent i f what are bracketed out are the specific expectations about what patterns of similarity and difference should be present. In contrast, the general principles of action theory (e.g., the perspectives and organisation of action focusing on action to gain an understanding of human experience) should be incorporated into the analytical process. Another way in which the analysis of narratives has the potential to inform the development of the action-project differentiation procedure is in how comparison groups are formed. In this method there appears to be an intentional selection of participants who are members of particular groups, at the outset of the research process (e.g., Gergen & Gergen, 1987). The strategy of 'examining the characteristics of potential participants and judging them to be members of one group or the other prior to their inclusion in the study' is one that may also work in action-project studies. For example, in a comparative study of the career development projects of mentally healthy youth and youth with schizophrenia, it is necessary to collect information about mental health status during the screening of participants. An action-project research team could discuss this information and arrive at a consensus decision regarding which group the participant dyad would best fit, prior to their inclusion in the study. These discussions would be particularly useful i f the status of the youth is in doubt (e.g., a potential participant who received a diagnosis of schizophreniform disorder four months ago). Grouping through intentional selection of participants provides some assurance that the groups will be composed of distinct sets of cases, and allows participants who do not clearly fit any category or the other to be excluded at the beginning of the study, thus preserving the 'purity' of the comparison groups. In the schizophrenia example, these ambiguous participants would include persons with chronic mental illnesses that fall outside of the schizophrenia spectrum, such as bipolar disorder. Table 2: Aspects of Comparison in Other Constructivist Research Methods A l , Stance towards „ . , . A Mechanism for Method „ . Grouping of Data r , L . . . . Comparison Comparative Analysis Phenomenology necessary, to derive the data grouped according intuiting / reflection to psychological essence to the themes that they determine patterns from of phenomena are thought to reflect the phenomenological reduction process Grounded utilised in delineating ." relevant information 'constant,comparison' -Theory.'-- , theboundaries of ' ' across cases is grouped (continual examination different categories and to form-mbre-complete of data and categories ; constructs within an understandings of a for possible similarities L < IX. i 6rnergi.ng theory t construct,or category = and differences) _ Analysis of accepted in the process cases grouped active reading and Narratives of analysing sets of according to shared 'playing' with the narratives in relation to experiences or common narratives until suitable each other membership patterns emerge In summary, this review of research practices in other forms of constructivist research has yielded several useful recommendations for the development of a between-groups analysis procedure for the action-project method. As can be seen in Table 2, the dominant mechanism for making decisions about the comparability of units of information and patterns of relationship in the data set, across all three types of constructivist research, is the researcher's own judgement.4 These judgements are invariably derived from extended and close examination of the original data, and the results of prior within-case analyses. I have adopted this strategy of relying primarily on researcher judgement, informed by the specific nature of the data, for the action-project differentiation method as well. Phenomenology and narrative research have suggested two opposing strategies to the grouping of participants for subsequent comparative analysis: inductively determining which cases belong in which group from the themes that have emerged out of the within-case and cross-case analysis, or deductively selecting participants to form two conceptually distinct groups on the basis of pre-existing individual characteristics. It is possible to envision both forms of grouping working within the action-project differentiation procedure. The decision of which strategy to use may be facilitated by gaining an understanding of what kinds of comparison/differentiation questions are of greatest interest to action theory researchers. Comparative Analysis in Published Action-Project Studies To determine what kinds of comparisons are of greatest interest to researchers working within the action theory paradigm, it is necessary to review and summarize the existing body of empirical literature that has employed Young, Valach and colleagues' action-project method. Doing so will also furnish evidence for my earlier claim that no formal procedure for conducting between-groups comparative analysis currently exists within this method. Relevant studies were identified through the PsychlNFO data-base, with separate searches conducted using the key-words 'qualitative action project,' and 'action theory,' followed by a manual screening of identified abstracts, to eliminate irrelevant material (e.g., participatory action research articles, articles grounded in other theories of action). Additional research was identified by contacting 4 It may be called "intuiting," "using logical reflection," or "playing with the data," but, ultimately, the decision is made within the mind of the researcher, rather than via an appeal to some concrete, external decision-making criterion. the two principal creators of the method (Richard Young, Ladislav Valach) to request information on any other previous or current action-project research of which they were aware. This process yielded three additional relevant studies, two completed but yet unpublished articles and one master's thesis. See Table 3 for a summary of the studies that were identified. Table 3: Summary of Existing Empirical Literature Utilizing the Action-Project Method Study Participants (n) Studied Phenomenon Protocol utilized Turkel, (2003) parent-adolescent agentic language in Full action-project protocol (3 dyads) joint conversation Michel et al.. (in counsellor-clierit therapeutic 1st interview only;.supple-press)^ (18 dyads) relationship / alliance mented by skin conductance Valach et al. (2002) suicidal clients suicidality 1st interview only; supple-(40 individuals) mented by skin conductance Young et al. (1997)' parent-adolescent' emotionality in joint 1 st and 2nd"inter\ iew ! (14 dyads) conversation Young et al. (1999) adolescent peers career conversations 1st and 2nd interview (10 dyads) _ Young, Lynam, et parent-adolescent Health promotion ls'fand 2nd interview. " al. (2001) (32 dyads) ; Young, Valach et al. parent-adolescent career development Full action-project protocol (2001) (20 dyads) _ Young et al. (2003) parent-adolescent career development in Full action-project protocol (6 dyads) Chinese-Canadians ' Young, Logan, et al. parent-adolescent sun-protection Action-project analysis of a (2005) (20 dyads) grounded theory data set While the number of studies appears to be very few for a comprehensive review, it must be remembered that the action-project method was not developed until the 1990s. Moreover, the focus of the literature search was to identify and collect empirical research that has utilized the action-project method, because these are the best reflections of the method as it is actually practiced. Consequently, I excluded those articles that described the methodology or discussed how action theory and the action-project method could be applied, but did not include the actual analysis and presentation of data sets (i.e., Bader & Young, 2002; Lynam & Young, 2002; Marshall, Young & Domene, in press; Valach & Wald, 2002; Valach & Young, 2004; Valach, Young, et al., 2002; Young et a l , 1994, 2000). For her master's thesis, Turkel (2003) conducted an in-depth analysis of agentic language in the joint conversations of three parent-adolescent dyads. The three cases were purposively selected from a larger sample of 20 dyads engaged in projects around the adolescent's career development. Although the author focused her analysis primarily on the joint conversations that occurred in the first and third interviews, she drew information from all parts of the action-project protocol. She described the specific language of agency and other related themes that emerged from each of the three dyads, and employed a 'collective case study' approach to form conclusions about the nature of agency as it manifests in general parent-adolescent dialogues. Although this study generated both within- and cross-case findings, the research design did not lend itself to between-groups comparative analysis. This is one action-project study that contained little evidence of interest in questions of comparison, much less a methodology for doing so. Valach and colleagues have employed a modified version of the action-project method (consisting of the first interview only, but supplemented with a skin conductance reactivity measure of emotion) in their research program with suicidal hospital outpatients. Their program has generated two empirical articles to date. One study examined the therapeutic relationship and alliance from an action theoretical perspective, in 18 counsellor-client dyads (Michel et al., in press). It was concluded that using a client-oriented, narrative approach, gaining an empathic understanding of suicidality in the context of life career issues, and adopting a model of suicide as goal-directed action, were all beneficial for the therapeutic alliance. No research questions regarding variation due to different sub-groups of clients or types of therapeutic interventions were asked, so no between-groups comparative analysis was attempted. The second publication was a study that delineated the suicide action, suicide projects, and suicide careers of 40 clients using within-case and cross-case action-project analysis procedures (Valach, Michel, et al., 2002). This study demonstrated that suicide attempts can be understood as goal-directed, occurring within a series of suicide actions, and contextualized within suicide projects and a suicide career. Although the authors did not engage in any explicit between-groups comparative analysis, they did conclude that "patients understand their suicide-attempt actions as a part of larger systems that are goal-directed and therefore meaningful. This is even the case when the patients maintain that they cannot understand why they did it" (Valach, Michel, et al., 2002, p. 167). The researchers did not elaborate on other differences between patients who reported understanding why they attempted suicide and patients who reported that they did not. They appeared to have arrived at this conclusion by examining the general trends found in the data for all participants as a whole. Although this study hints at viable between-groups comparisons, such questions did not appear to be of interest to the authors, so no formal between-groups comparison procedure was described. In 1999, Young and colleagues examined adolescent peer interactions within the domain of careers, as manifested in the conversation of ten pairs of adolescent friends. Using a shortened form of the action-project method (omitting the monitoring period and third interview), the authors were able to identify participants' career-related goals, actions taken to reach those goals, and connections between goal-directed action and adolescent identity development projects (i.e., careers were seen as an expression of identity). The authors made one between-groups conclusion, "in contrast to parent-adolescent conversations, this study showed that conversations between peers are more egalitarian" (Young et al., 1999, p. 53). However, they appear to have based their claim on a general understanding of the nature of parent-adolescent conversations, rather than an actual comparative analysis of adolescent-adolescent versus parent-adolescent dyadic conversations. No parent-adolescent dyads were included in the sample. This study provides evidence that action theory researchers have some interest in between-groups comparisons, but not for how to conduct these kinds of comparisons within the action-project method. In a similar way, the tone of Young et al.'s (1997) article, describing the role of emotions in the career-related conversations of 14 parent-adolescent dyads, reveals a desire to make conclusions regarding group differences but is limited by the nature of the existing action-project method. This study used a shortened form of the action-project method (omitting the monitoring period and third interview) to identify and describe the function of emotion in joint conversations about career and, by extension, overall adolescent career development. Presentation of the results consisted of a descriptive summary of the ways in which emotion manifested during joint conversations, combined with detailed description of two contrasting exemplars. Some flavour of between-groups comparison was present in the discussion of the links between emotion, goal congruence and supportiveness of communication. For example, it was reported, "certainly the Smiths demonstrate that parent-adolescent conversations go more smoothly when shared goals ... are evident," and "in contrast the Joneses demonstrate how rival constructions of career and mutual feelings of tension and disappointment arise when parent and adolescent do not share similar goals" (Young et a l , 1997, p. 42). The comparison in question was generated out of the findings that emerged from the cross-case analysis process; specifically, dyads with a mutual goal versus dyads with dissimilar or conflictual goals. The apparent mechanism of comparison was to describe two representative cases in great detail and, from that, intuit conclusions regarding the differences that exist. It was presumed that these conclusions were evident in the case descriptions. 'Grouping' occurred only by embedding their conclusions in the discussion of the case examples, which were selected as the best examples of the functions of emotion in the construction of career. It is important to note that the issue of mutual versus conflictual goals formed only one, relatively minor, part of the overall discussion in the article so, naturally, the authors did not group all 14 parent-adolescent dyads according to the congruence of their goals. In a more recent study, Young, Valach and colleagues (2001) examined the joint career development projects of 20 parent-adolescent dyads. They identified and described the properties of participants' projects, and the connections between career projects and other projects in which the participants were also engaged. Although the analysis was primarily conducted at the within-case level (as reflected in the general tone of the article, and liberal use of examples from cases), the authors described a number of patterns that emerged across cases. They found variation between dyads in terms of what properties were most salient, and what kinds of 'other' projects that career-projects tended to be embedded within. The study included some degree of attending to pre-existing group differences in the way that the findings were presented, as is evident in the statement, "the cultural project was highlighted in this study because of the participation of a number of Chinese-Canadian families, although evidence of cultural projects was found in both Euro-Canadian and Chinese Canadian families" (Young, Valach et al., 2001). However, no formal comparison of these two (or any other) sub-groups of participants was conducted. Instead, where ethnicity was relevant to the discussion (e.g., in describing the relationship between career projects and cultural projects), the authors reported the ethnicity of the specific cases from which they drew their conclusions. Once again, no formal process for conducting between-groups analysis was described in the article, and differentiation between sub-sets of participants seemed to occur only at an informal level. The nature of the data generated from the Young, Valach and colleagues' (2001) study (i.e., the presence of a sizable sub-group of participants from a specific cultural background, whose projects appeared to be distinct in systematic ways from the remainder of the participants) is interesting because it left the authors with precisely the dilemma that is the focus of this dissertation: how does one draw conclusions regarding similarities and differences between sets of distinct cases within action theory research? The authors' answer was to reanalyse the data separately, for those participants whom they perceived as forming a distinct group. These results were then published in a separate article. Specifically, two of the researchers jointly reviewed and re-coded the data collected from the six Chinese-Canadian participant dyads in the original sample, and identified the "common and distinguishing themes" to describe the salient properties, connections with other projects, and cultural content of the career development projects found in Chinese-Canadian families (Young et al., 2003, p. 292). No explicit claims were made regarding the similarities and differences in patterns of results between this group and the Euro-Canadian dyads in the original sample. The strategy of separating out a particular sub-group for closer analysis and separately reporting the results of that analysis is of more limited usefulness than creating a between-groups comparison procedure for the action-project method. That is, it fails to allow for discussion of the patterns of results found in one kind of participants, in relation to those found in another kind. Therefore, researchers who are interested in exploring the interplay between action/projects and specific structural characteristics of groups of participants (or even variation within the themes that emerge from cross-case analysis) are not served by the solution of writing separate articles for each distinct grouping of participants. Separate publications for separate cultural groups was, for instance, not the approach taken by Young, Lynam, et al. (2001) in a health research study asking the question, "What are the individual and joint actions in which parents and adolescents engage in conversations about health and health promotion in Indo-Canadian and Euro-Canadian families?" (p. 40).5 This study examined the health conversations of 16 Indo-Canadian and 16 Euro-Canadian parent-3 In fact, the authors explicitly chose not to report results for the two cultural groups separately, because "[their] intention was not to create a definitive description of health practices or behaviors of Indo-Canadians," (Young, Lynam, et a l , 2001, p. 56), or presumably, all Euro-Canadians either. adolescent dyads using a shortened form of the action-project method (the monitoring period and third interview were omitted). The emphasis in this study was the within-case analysis process, from which five general categories of joint actions were then generated. The primacy of the within-case level of analysis was reflected in the fact that their method section contained a description of only the within-case analysis process; the existence of a cross-case analytical process had to be inferred from the fact that each category of action was described separately, with each description being based on the particulars of more than one specific case. In addition, the numbers of participants for whom a particular category of action was applicable were reported (e.g., "The action in 11 (4 Euro-Canadian, 7 Indo-Canadian) conversations can be characterised by the parent and adolescent sharing and exploring information, values, attitudes and beliefs about health topics," Young, Lynam, et al., 2001, p. 45). Although the authors did not conduct any explicit between-groups comparative analysis, they did make group-related claims in their discussion section. For example, they stated "adolescents in the Indo-Canadian subgroup were more likely to be concerned about differing expectations of their parents and the other communities (school and friends) of which they were a part" (Young, Lynam, et al., 2001, p. 54), and also "the relationship of the goals of these health promotion actions to the social context was also evident... this was more explicit in the conversations involving Indo-Canadian parents, single parents, and parents living within financial restraint" (Young, Lynam, et al., 2001, p. 55). Although these conclusions are evidence of a desire to engage in comparative analysis based upon distinct characteristics of sets of participants, the absence of any description of any analytical procedure beyond the within-case level makes it difficult for readers to determine what is meant by "more likely" and "more explicit," or even understand how these judgements were made. More recently, Young, Logan, Lovato, Moffat and Shoveller (2005) examined the family sun protection projects and actions of a sample of 20 parent-adolescent dyads. They described the nature of sun protection projects over time, connections to other family projects, and the implications of having projects that are focused, or diffuse. Although this study was a secondary analysis of data collected for a grounded theory study (i.e., Shoveller, Lovato, Young & Moffat, 2003), it remains important to the present discussion. The study is important because it employed the same action theoretical analysis process that is utilized in the action-project method, and explicitly presented conclusions regarding the different patterns of findings associated with different groups of participants. Specifically, Young, Logan, et al. (2005) classified participant dyads as having either 'focused' or 'diffuse' projects, and described the differences associated with this grouping construct in a comparative way. For example, they wrote, "parents who were involved in focused sun protection projects demonstrated a strong commitment to pursuing their goals, which included communication of goals and information ... Families with diffuse sun protection projects were less committed to sun protection, and there was less congruence between goals and functional steps" (p. 17-18). The authors described the general patterns of action associated with each kind of participant, and presented two concrete examples to illustrate the similarities and differences between dyads with focused projects, and dyads with diffuse projects. This manuscript goes beyond reflecting an interest in between-groups comparison and differentiation; it provides evidence that action theory researchers are actually beginning to engage in this kind of analysis. Unfortunately, the process by which Young, Logan, and colleagues (2005) arrived at their conclusions of similarity and difference is rather opaque; their methods section included descriptions of participants, data collection procedures, the general action theoretical analysis procedure, and an overview of their grouping system: Once family sun protection projects were identified and described, they were subsequently classified as either focused or diffuse. A focused sun protection project was characterized by well defined, explicit goals and functional steps shared by the parent and adolescent. A diffuse family sun protection project was characterised by few common sun protection goals or strategies ... the classification and criteria were presented to, and discussed with, two of the other researchers (Shoveller & Lovato) to seek consensual validation. (Young, Logan, et al., 2005, p. 10-11). However, the actual mechanism for deciding how this grouping construct was related to the patterns of action and connections to other family projects, and determining that differences were present between the two groups, was not explained. Instead, there appeared to be an assumption that it would be possible for readers to perceive the similarities and differences from the case examples. The absence of a description of how the authors derived their between-groups conclusions is problematic because it hampers readers' ability to evaluate the plausibility and trustworthiness of the analysis process. It also provides insufficient guidance, in terms of how to proceed with the between-groups component of the analysis process, for future researchers wishing to extend existing findings. Even though the Young et al. manuscript provides useful hints for how between-groups comparative analysis might occur within the action-project (especially in terms of how to form the comparison groups), it falls short of providing an explicit procedure for conducting these kinds of analyses. This review of existing action-project studies for material related to between-groups comparisons has yielded several important conclusions. First, while several studies contained no evidence of interest in between-groups variation in human action, others reflected an emerging interest on the part of action theory researchers to make these kinds of conclusions. Next, it is evident that no formal procedure for conducting this kind of analysis has been satisfactorily explicated in the literature, even though some degree of organising findings according to distinct sub-groups of participants is present in a number of these studies. Finally, it appears as i f researchers are interested in both variation due to pre-existing characteristics of participants (e.g., Young, Lynam, et al., 2001; Young, Valach, et al., 2001; Young et al., 2003), and also variation associated with the patterns that emerged from prior analysis (e.g., Valach, Michel, et al., 2002; Young et al., 1997; Young, Logan, et al., 2005). Given this interest in both kinds of grouping, any between-groups comparative analysis procedure should be able to accommodate groups formed on the basis of pre-existing individual differences (similar to the deductive approach found in the analysis of narratives) and groups that emerge out of the analysis process (similar to the inductive approach of phenomenology). Therefore, I incorporated both types of grouping in the action-project differentiation method that I have developed. However, simply because it is possible to form distinct groups does not mean that examining the patterns of findings of those groups in relation to each other is compatible with the theoretical assumptions of action theory. This is the issue that will be addressed next. Between-Groups Analysis in Action Theoiy The process of examining data from distinct sub-sets of cases in relation to each other contains a number of assumptions that need to be examined for their compatibility with action theory. One such assumption is that clustering cases together and make meaningful statements concerning general patterns of findings that apply to those cases together is a valid analytical process. There is nothing inherent in constructivism to preclude an examination of phenomena that are manifested across individuals, as evidenced by the use of cross-case analysis in grounded theory, phenomenology, and narrative research. Additionally, the constructivist idea that different persons living within the same socio-temporal context often share similar frames of reference when constructing their understandings of the world (Delanty & Strydon, 2003) implies that some aspects of their action are commeasurable, and can be examined across individuals. Moreover, another strong argument for the compatibility of action theory and the aggregation of information across participants is provided by the fact that many existing action-project studies utilize cross-case analysis strategies; either in a formal sense, as described in the methods and results sections (e.g., Young et al., 2003), or informally, by using examples from multiple cases to support the conclusions that the authors present (e.g., Young et al., 1997). Evidently, action theory accepts the premise that there are sufficient commonalities across the actions of individuals to permit clustering of information across a group of participants and making conclusions about the group as a whole. However, a between-groups comparative analysis procedure makes an additional assumption: that it is possible to juxtapose two or more sets of participants, and draw conclusions about the similarities and differences in the way that a phenomenon is manifested within those groups. In addition to my previous arguments demonstrating that comparing different individuals' experiences is acceptable to action theory, it is self-evident that there is variation between individuals acting within a phenomenon of interest (due to different life histories, socio-cultural contexts, and subjective interpretations of experiences). When variation in action is systematically affected by a specific historical experience, contextual factor, or subjective perspective (that is shared by some but not all participants within a research sample), then it becomes meaningful to group participants according to that construct, for subsequent exploration of the interplay between that construct and the manifestation of action. For example, when examining career development in the context of mental illness, there will be issues that are relevant to most, i f not all, participants who have experienced mental illness, but irrelevant to participants who have not shared this developmental context (e.g., deciding whether or not to disclose one's history of mental illness to potential employers). A secondary argument for the compatibility of this assumption and action theory can be derived from the theory's acceptance of the normative standard of knowledge construction found in the social sciences: findings do not exist in a vacuum, but must be understood in relation to prior knowledge (Valach, Young, et al., 2002). That is, i f an action theory researcher generates a set of findings from one study, and then conducts another study examining the same phenomenon, then he or she would be permitted (in fact, required) to discuss the findings from the latter in relation to first study. The content of this discussion usually includes a summary of the ways in which the two sets of findings support and contradict each other, which is a form of comparison. It is possible to envision situations where the participants in the first study differ L systematically from those in the second study according to some specific dimension of temporal experience (e.g., studying relationship projects of couples who grew up in the 1950s and the 1970s), social context (e.g., studying relationship projects in dating couples who have previously been divorced, and those who have never been married), or subjective perspective (e.g., studying the relationship projects of couples who are either 'starry-eyed romantics,' or 'cold, pragmatic realists'). Situating the results of the second study in light of previous research requires some discussion of the characteristics of the samples that were used in the various studies. If it is possible to compare the patterns of findings from two studies, it must also be possible to compare the findings when data for the two studies are collected and analysed simultaneously (i.e., within one larger study). Therefore, the comparison of findings associated with distinct sets of participants must be acceptable within an action theory framework. Moreover, the simultaneous collection of data from several kinds of participants (i.e., using the same research team, and within a single time frame) increases the credibility of any conclusions that are made about the similarities and differences between the two groups. The effects of the research context and person of the researcher on the construction of findings are more likely to be equivalent when they occur as part of the same study. This increased credibility means that it is not only permissible, but actually advantageous for an action-project study to conduct comparative analysis of distinct groups of participants within a single study, when systematic differences among participants are of interest to the researchers. The fact that some division of participants into groups, and analysis of data across groups, is compatible with action theory does not mean that all forms of between-groups comparison and differentiation are permissible. The theoretical tenets and paradigm positions of this theory actually impose a number of restrictions on the comparative analysis process and the kinds of conclusions that can be made. For example, the theory precludes drawing conclusions about a phenomenon based only on a single perspective of action (Valach, Young, et al., 2002). It is necessary to consider information from manifest behaviour, internal processes and social meanings, when engaged in the comparison process. In addition, because action theory accepts that action is embedded within the overall context of people's lives (Valach, Young, et al., 2Q02), it is necessary to attend to the overall configuration of a case when forming groups and evaluating patterns of action within the group. In other words, the process of between-groups analysis should involve the aggregation of cases in their entirety, rather than any form of aggregation that eliminates the specific contexts of the specific individuals that are clustered together. Removing the case context in which data are embedded, and conducting subsequent analyses of some disembodied 'mean score' of the data, would contradict the perspective of holism that is present action theory. As a result, in the differentiation procedure that I have developed, grouping constructs serve only to organise what is truly being compared (i.e., whole sets of cases) around a contextual factor that is considered to be salient; the construct must not become the focus of the comparison itself. Analyses that privilege the grouping construct to the exclusion of the life contexts of the cases in which action is embedded would violate the principles of action theory. Finally, it would be erroneous to make conclusions about directionality of influence, or to make inferences regarding the causal prominence of the grouping construct. Doing so would minimize the complexity of action as it occurs within peoples' daily lives, which is accepted as a reality within the action theory framework (Valach, Young, et al., 2002). Instead, as described in chapter one, the grouping construct is primarily a way to gain close access to some particular aspect of the context in which action occurs. It should, therefore, be viewed as having the same explanatory strength as any other contextual factors that are present in the lives of the participants; that is, it can only be understood as one aspect of the total context that influences the manifestation of action. The grouping construct may be an important factor that facilitates an improved understanding about a phenomenon, but it cannot be assumed to be the cause of variation found within that phenomenon. Therefore, the answer to the first question of interest in this dissertation is a qualified "yes." Some forms of between-groups comparative analysis are compatible with the paradigm positions of action theory. However, the theory imposes some limitations on the kinds of comparisons that can be made: the analysis must attend to all three perspectives of action that are proposed in the theory; the context of each case needs to be attended to when performing the comparison (rather than focusing exclusively on the grouping construct), and the purpose of the between-groups analysis is to provide richer descriptions of the phenomenon of interest, rather than to make causal connections between the grouping construct variable and some other 'outcome' variable. Evidently, any between-groups analysis procedure developed within the action theory paradigm will be very different from the statistical comparisons of average scores for each group that typify quantitative between-groups analysis. The Action-Project Differentiation Procedure Previously, I argued that some form of between-groups comparison is evident in various different constructivist research methods and, specifically, compatible with the action theory approach to studying human phenomena. I have also demonstrated that, despite an emerging interest in making between-groups conclusions in existing action-project publications, no specific procedure has yet been developed to conduct these kinds of analyses. Drawing upon the existing action-project method and other constructivist approaches to comparative analysis, I now attempt to correct this situation by proposing a way of performing between-groups comparisons within action theory: the 'action-project differentiation procedure.' The remainder of this chapter delineates this solution to my second research question, "what specific procedures could be for between-groups comparative analysis in action theory, to generate findings that go beyond that which is attainable within the existing method?" Rather than attempting to develop an entirely novel approach, I have extended the existing action-project method to the situation of comparatively examining distinct sets of cases. I have chosen to do so for three reasons: (a) to ensure the compatibility of the new procedure with existing action-project research practices, thus reducing the risk that my procedure contradicts the theoretical assumptions of action theory; (b) because there is sufficient relevant information within published action-project research to permit a natural extension of the analysis process to this new situation; and (c) because doing otherwise would introduce an artificial and unnecessary level of complexity for action theory researchers who are attempting to address between-group questions. The only component of the proposed differentiation procedure that is not a natural extension of the existing action-project method is the use of charts to facilitate the decision-making process. This idea was adapted from Miles and Huberman's (1994) general guidelines for data display in qualitative research, and can be found in other team-based qualitative analytical approaches (e.g., Hil l et al', (1997a, 1997b) Consensual Qualitative Research method). Despite the fact that no previous action theory research has used charting as a formal analytical tool, it does not conflict with existing action-project procedures. In any case, charts are merely a supplementary tool to facilitate the decision-making process, rather than an essential feature of the action-project differentiation procedure. Formation of groups. When clustering cases into groups for comparative analysis, the grouping construct must be a meaningful way to organize the data. This requirement ensures that there is a substantive reason for dividing participants into groups, and reduces the risk of making spurious linkages between the grouping construct and the phenomenon being studied. Although the final evaluation of the meaningfulness of a grouping construct cannot be completed until the end of data analysis (i.e., ultimately, a construct is a meaningful way of dividing participants to the extent that it produces a deeper understanding of the phenomenon of interest), it is possible to select constructs that have a higher likelihood of being a meaningful way to cluster participants at outset of the study. As I envision it, the differentiation procedure will be applied primarily in situations where various distinct sub-groups appear to be naturally from the findings of the within-case and initial cross-case analyses of an existing action-project study. That is, certain themes, characteristics or features are noted to be present in some participants but not others, and are judged to be worthy of closer scrutiny because they may be reflective of distinct kinds of participants. Any theme that emerged as meaningful in the within-case and preliminary cross-case analyses is also likely to a meaningful way to organise the between-groups analysis. Additionally, it may also be possible to plan a comparative analysis on an a priori basis, selecting meaningful comparison groups after reviewing the relevant literature to determine what contextual factors may be systematically associated with variation within the original phenomenon being studied. In this situation, the empirical knowledge and/or theory about the role of a potential grouping construct is used to evaluate whether there is sufficient justification to group participants according to it. If an initial, theory driven, choice is also supported by the pattern of findings that emerge from the within- and cross-case analyses, then an even stronger argument for the meaningfulness of that grouping construct can be made. For some phenomena, conducting a review of the literature can also guide the selection of participants for inclusion in the study. That is, once the boundaries of the groups are established by gaining an understanding of the relevant literature, it will be possible to choose only participants who clearly fall into one group or the other. In the existing action-project method, the research team reviews potential participants and discusses whether they meet the specific criteria for inclusion in the study. The action-project differentiation procedure follows the same protocol, but with the question df whether people clearly fit into one group or the other being explicitly addressed as a part of the process of selecting participants for inclusion in the sample. This issue of selecting participants who are clearly members of the groups of interest may appear to be trivial for some individual characteristics such as ethnicity or gender, but is clearly important when the grouping construct is something such as economic disadvantage, the experience of an illness, or other characteristics where the boundaries are not absolute. In fact, active selection of participants is important even for 'self-evident' characteristics, because exceptional cases may arise even then (e.g., mixed-culture families, transgendered individuals). The reason for selecting only dyads who clearly fall into one group or the other is to preserve the distinctiveness and clarity of the grouping construct. For example, a researcher who is interested in the transition-from-university projects of couples, and finds evidence in the literature that the kinds of decisions and processes they engage in differ according to whether the couple is in a casual or a committed relationship, may decide that the structural characteristic of'relationship status' is an important contextual feature to explore. During the recruitment phase of this hypothetical study, it would be important to select couples who were either clearly in the dating phase of a relationship, or clearly in the life-partner phase of their relationship. This would be accomplished by (a) collecting relevant screening information (e.g., marital status, cohabitation status, length of relationship, degree of joint economic commitment, subjective perception of commitment to the relationship), (b) discussing how well each couple fits the 'dating' or 'life-partner' group within screening meetings, and (c) selecting participants on the basis of whether they are clearly a dating couple, or clearly life-partners. Ambiguous cases (e.g., the couple that has been engaged for three years and is not co-habiting; the married couple contemplating a separation) would be excluded. For other phenomena, although it is possible to anticipate what themes may be associated with systematic variation from the literature review, it will not be possible to identify the appropriate group for specific dyads until the nature of the grouping construct has been revealed in their actions and projects over time. This is particularly true for studies where unanticipated themes, requiring comparative analysis, emerge from the data in a standard action-project study. In these situations, cases can only be clustered together after the within-case and initial cross-case analyses have been completed, rather than at the participant selection stage of the study. Once the specific manifestation of the grouping construct within each participant dyad has been delineated, the researcher can use variation in the manifestation of that theme to categorizing participants into one emergent group or another. For example, i f research has indicated that trust is an important contextual factor in the relationship success of siblings, then researchers who wanted to examine the interplay between trust and the relationship projects of siblings could only form the 'high trust' and 'low trust' groups after conducting a standard action-project study to collect data and conduct within-case and cross-case analysis. Then, i f the initial cross-case analysis indicated that trust was a salient theme in the projects, the cases whose joint actions included a high degree of trust could be grouped together, while siblings who were not trusting in their interactions would be clustered into a separate group. The action-project differentiation procedure could then be applied to explore the relationship projects of these two distinct kinds of sibling dyads. The key point, however, is that it is not possible to determine whether a specific dyad will fall into the high trust or low trust group during the screening stage. In these situations, the between-groups analysis process must have the capacity to accommodate ambiguous cases, because participants who do not fit well into either group cannot be excluded from the sample prior to data collection.6 Instead, it may be necessary to exclude those cases from the between-groups stage of the analysis process, as Young, Logan, and colleagues (2005) chose to do in their sun protection study. This is certainly a legitimate method to resolve the problem of ambiguous cases, but an additional strategy is to examine ambiguous cases for the possibility that there is sufficient commonality among them to form a coherent group of their own, and include these cases as another group to examine. For example, in studying the interplay between joint career-related actions/projects and the temporal focus of participants' projects, it might be found that a number of participants could not be classified as having either 'short-term projects' (i.e., within the next school year) or 'long-term projects' (i.e., after completing all future education). If the research team examined these ambiguous dyads and concluded that the temporal focus of all of their projects fell somewhere in between the other two groups, then it could be useful to label them as having 'medium-term projects.' This would allow the subsequent between-groups analysis to delineate the similarities and differences in career development projects in light of three different time orientations. Mechanism for determining similarity and difference. Once the groups have been formed, then the data for all members within each group can be scrutinized to determine what themes and findings tend to occur for that kind of person. This process is identical to the preliminary cross-case analysis except that, rather than the sample in its entirety, conclusions about the data are made for each distinct sub-set of participants. The existing action-project 0 In fact, even when participants are purposively selected for grouping on an a priori basis, ambiguity can creep in. To continue with the transition-from-university study, suppose a previously committed couple decided to break up over the course of negotiating their transition- into which group should they be categorised? criteria of importance and presence across cases, as perceived by the research team, are used at this stage. This 'within-group, cross-case' analytical process generates a set of findings that describe the actions, themes, and other characteristics that are commonly present for persons within each of the different groups. Next, decisions must be made regarding the comparability of the findings that emerged out of each group. As Miles and Huberman (1994) point out, in the absence of a statistical test to determine the significance of a difference, it is the practical significance of a difference that becomes the standard forjudging the noteworthiness of a difference between two units of qualitative data (and, by extension, more complex patterns of qualitative findings). Following Kirk's (1996) suggestion, 'practical significance' can be thought of as a difference of sufficient magnitude to be useful in the real world. Implicit in Miles and Huberman's suggestion is the corollary that, when the differences in collected data are not of practical significance, data can be considered to be similar. The guideline of practical significance serves as a useful standard for evaluating similarity versus difference in my differentiation procedure, although the actual mechanism for making judgements regarding the practical significance of between-groups differences still requires delineation. I propose that the same decision-making strategy used in the existing action-project method (i.e., team-based, consensual decision-making, grounded in close examination of the data and informed by the hermeneutic of action) be applied to evaluating whether the patterns of findings between the groups are sufficiently distinct to be considered practically significant. This evaluation is based on researchers' knowledge about (a) the phenomenon in question, (b) the discipline in which a particular study is manifested, and (c) the degree of difference that would be useful for the primary audience of the study. The process involves some degree of creativity and substantial negotiation between team members, as differing possibilities are considered and discussed. Consistent with the action theoretical variant of constructivist epistemology, these judgements must also be grounded in the external 'realities' of the data; that is, the conclusions must be plausible, given the nature of the collected information. This decision-making process is perhaps best illustrated by an example. In an action-project differentiation study examining the interplay between (a) whether nursing staff are working single or double shifts, and (b) surgery projects in hospitals, any observed differences in quality of care would need to be of sufficient magnitude to have practical implications for patients, before the researchers could conclude that such a difference existed between the two groups. It would also be unacceptable to draw the conclusion that operating room team performance in double shifts was of lower quality than performance in single shifts, unless the aspect of 'quality' was readily apparent in the observations and interviews, and the differences in quality were large enough to affect the patients being operated upon. To reiterate, the epistemology of action theory precludes interpretations that fail to reflect the actual nature of the observed data, no matter how strongly the research team believes in a constructed finding, nor how useful it would be to assert that conclusion. Indeed, in many constructivist approaches to research, drawing conclusions that are not apparent in the information gathered from participants raises questions about the trustworthiness of a study (Bryman, 2004; Lincoln & Guba, 1985). The analysis process is complicated by the fact that, although the data must be organised in such a way that the general patterns of findings between the groups can be examined together, the full configuration of data within each case must also be retained for consideration: ignoring the full context of each case would be a violation of one of the central beliefs of action theory. Visual presentation of the patterns of findings for both groups on a common chart is one way to facilitate the analysis process without losing important information from the cases. Specifically, a chart that lists (a) the various salient aspects of the findings, (b) the nature of their f manifestation in each group, and (c) the dyads for whom that aspect is relevant, can be generated to capture the information derived from the analysis of data across each sub-group. See Figure 1 for an example of such a chart. The research team can then examine each aspect, in turn, to make a decision as to whether the findings within that aspect of the phenomenon are similar or different between the groups. In evaluating the existence and practical significance of any difference between the groups, it would also be necessary to refer back to all cases for whom an aspect was relevant, to fully account for the original cases. After discussion has occurred and a decision has been reached, the researchers would examine the next aspect, and so on, until all pertinent aspects of the pattern of findings had been evaluated. In situations where an aspect is evident in one group but entirely absent in the other, the fact that a particular aspect of a phenomenon is not manifested at all for one kind of participant is sufficient evidence that a practical difference exists. Figure 1. Part of the comparison chart for a hypothetical study on the family career development projects of adolescents who have and have not experienced early psychosis. It is important to note that the purpose of the chart is to assist researchers in organising (Counter. SoMpbt 73, 75, 77, 78) 7 1 define "aspect" broadly, to encompass any construct, quality, or characteristic of the findings that the researchers identified during analysis, including the project properties and other relevant dimensions of the goals, action and/or projects being researched. and visualising the patterns of findings in such a way as to preserve the complexity of each participant's experiences during the between-groups analysis process. As such, it would be contrary to the spirit of this tool to use the chart in a purely reductionist way, where judgements of similarity or difference are made by attending solely to the number of cases in each cell, rather than the full configuration of information that is represented by each case. Although the frequency with which a particular action or aspect of the data set occurs is certainly important, referring back to the individual cases permits the retention of differing shades of meaning within each cell. Furthermore, attending to the unique configuration of information in the original data also permits cases to be categorised in ways that, on the surface, appear to be contradictory. To continue the example presented in Figure 1, the actions of the hypothetical participants in dyad 12 could reflect a conceptualization of career that encompasses both the notion of 'money job' and 'identity' ("I still need to earn money, so I 'll have to find a simple job, like at a fast-food place, but I don't want to be a burger-flipper for the rest of my life"). Similarly, dyad 20 may be people who perceive work primarily as a means to prevent a relapse of the mental illness and, therefore fit into neither the 'money job' nor the 'identity' category. The remainder of the research process is generally conducted according to the established practices of the existing action-project method (see chapter 2 for details). The product of an action-project differentiation study is a detailed and extensive description of the patterns of similarity and difference in distinct groups of participants' joint actions and projects, in terms of the content, properties, and connections to other projects. Together, the processes of grouping and juxtaposing findings from distinct groups yield a deeper understanding of the nature of the phenomenon of interest. Because participants are grouped via a contextual feature that is prominent within a phenomenon of interest, the resultant description provides greater understanding of the connections between the contextual factor and the phenomenon than would be possible without a between-groups comparative analysis. It must be noted that, although the group formation process and the identification of similarities and differences are presented as separate procedures in this description, the stages are not separate in the same sense that one would identify an independent variable to form groups in an A N O V A , and test to see if there are differences between the groups on some other outcome variable. Instead, in the action-project differentiation procedure, data are examined once to identify whether distinct groups exist (group formation), and examined again to generate a more detailed description of the distinctions between those groups (identification of similarities and differences). There are no separate 'independent' and 'dependent' variables. In summary, the general sequence of the action-project differentiation procedure is as follows. Cases are clustered into distinct groups on the basis of the findings that emerged from the within- and cross-case analyses. Then, the patterns of findings that emerged from the different groups can be examined in relation to each other, in order to determine the ways in which they are similar and different. At this stage, the research team uses practical significance to guide decisions about what degree of between-groups variation is sufficient to be considered a 'real' difference. Visual presentation of data, in the form of a chart, may facilitate the examination of different sub-groups together, while still retaining the full configuration of information for each case. This produces a greater depth of understanding of a phenomenon than would be possible without some form of between-groups analysis. Initial Conclusions At this point, initial answers have emerged to the questions that I posed in this dissertation. Specifically, I have shown that some forms of between-groups comparative analysis are compatible with the paradigm positions of action theory. However, the nature of the comparative analyses that are permissible within this paradigm are somewhat distinct from those found in other forms of social research. I have also proposed a way to extend to the existing action-project method, giving concrete guidance to action theory researchers who wish to conduct these kinds of between-groups comparisons. An outline of this action-project differentiation procedure is presented in Figure 2. 1 On the basis of findings that have emerged from previous analysis, and/or the extant literature about a grouping construct, researchers decide that it is worthwhile to highlight a particular contextual factor for closer analysis, within a study about any given phenomenon of interest. * 2 If no prior analysis has occurred, data should be collected following the existing action-project method, and the within- and initial cross-case analysis conducted as usual. It may be useful to select participants who maximize the distinctiveness of the grouping construct. 3 Cluster/divide the cases in the data set according to the way that the grouping construct manifested at the within- and cross-case level, with all the people who are similar to each other but different from others in terms of that construct each forming a separate group. * 4 Within each group examine the data set to identify the patterns and themes that tend to emerge for that group, using the criteria of importance and commonality of occurrence across cases. * 5 Juxtapose the findings from the different groups (possibly through charting) to examine them in relation to each other, using practical significance as a guideline. Information about the patterns of similarity and difference is used to supplement the findings from step 3, thus yielding a richer understanding of the original phenomenon of interest. * 6 Report and discuss the findings, framing the grouping construct as a contextual factor that improves understanding of the phenomenon of interest, and describing the specific ways in which the patterns of action of distinct kinds of persons are similar and different from each other. Figure 2. Outline of the action-project differentiation procedure, numbered to facilitate tracking of the application of this method in the demonstration studies presented in chapter four. However, I recognise that the Measurement, Evaluation and Research Methodology program requires an empirical, rather than purely theoretical dissertation. Therefore, rather than allowing others to use the procedure in order to determine whether it is truly a useful extension of the action-project method, I have used it myself to further analyse data from a study on the family career development projects of early adolescents. The remainder of this dissertation consists of (a) the demonstration of how this protocol was used to generate findings from a real data set (Chapter 4); (b) the results of my demonstration- a description of successes and problems encountered in implementing the procedure (Chapter 5); and (c) a more general discussion of the procedure, including suggestions for future refinement of the action-project differentiation procedure, and the broader implications of focusing on between-groups comparisons for the action theory paradigm as a whole. CHAPTER 4: DEMONSTRATION OF THE ACTION-PROJECT DIFFERENTIATION PROCEDURE Having developed a protocol for conducting between-groups comparative analysis for the action-project method, I now need to demonstrate that this differentiation procedure is capable of generating useful knowledge about psychological phenomena. In this chapter, I provide two examples of how the action-project differentiation procedure can be used to expand the kinds of questions that can be addressed in action theory research. Specifically, I use data from an existing study on family career development projects (Young, Marshall, et al., 2005), and re-analyse it in two ways, using a different grouping constructs on each occasion. The first study explores similarities and differences in the projects and patterns of action of dyads who have 'focused' or 'diffuse' goals. In the original study from which these data were obtained, we noted that participants appeared to vary on goal orientation, but lacked the capacity to systematically examine the similarities and differences that existed for people with different goal orientations (Young, Marshall, et al., 2005). As a result, the first of the two demonstration studies presented here is an example of using the procedure with a grouping construct that emerged naturally from the within-case and initial cross-case findings. The second demonstration study explores similarities and differences in the projects and patterns of action of mother-son versus mother-daughter dyads. In this study, participants were grouped on the basis of a pre-existing structural8 feature: the gender configuration of the dyad. The second study is an example of the use of the action-project differentiation procedure to examine a grouping construct that was derived from theory, rather than one that emerged from previous analysis. Although the same data set is used in both examples, each one addresses a different research question. Therefore, different sets of participants are grouped together for analysis, and different conclusions are reached in the two studies. 8 This grouping construct is "structural" in the sense that it is a static characteristic, as opposed to a process-oriented aspect of participants' action, such as their degree of participation in a project. The primary purpose of conducting these studies is to illustrate the usefulness of the action-project differentiation procedure, rather than to expand the knowledge base on family career development projects. As such, the introduction sections focus on the description of the research question and the theoretical justification for using the grouping constructs that were selected, rather than comprehensively describing career development in the context of the family. Furthermore, the method sections contain a far greater level of detail than is typically found in research articles, in order to provide a clear illustration or the implementation of the procedure. Finally, the discussion sections of these two demonstrations are somewhat abbreviated because discussion of the substantive issue of this dissertation (i.e., the viability of the action-project differentiation procedure) takes place in Chapters 5 and 6. Study 1: Goal Orientation Career development theories converge to indicate that having goals for the future facilitates career development at all ages, including adolescence (e.g., Lent, Brown, & Hackett, 1994; Peterson, Sampson, Lenz, & Reardon, 2002; Smith, 1999; Young, Valach, & Collin, 1996) . While the presence of goals is clearly important to the career development process, there is considerable variation in the nature of the goals that people form, especially when those goals are jointly constructed. This variation has been demonstrated empirically, through a series of studies about the family career development projects of adolescents and their parents, conducted by Young and colleagues (e.g., Young et al., 1994, 2001, 2003; Young, Paseluikho, & Valach, 1997) . Specifically, they found that career development projects varied considerably in terms of the kinds and importance of goals that were set by different dyads. While the specific nature of, and change in, the observed projects was, no doubt, influenced by situational or environmental factors that were unique to the specific pairs of participants, it is possible that some of the variability was also systematic. That is, perhaps different kinds of parent-adolescent dyads have distinct kinds of projects and patterns of action as they engage together in promoting the adolescent's future career. Other research within the action theory paradigm has identified one characteristic that may meaningfully organise different sub-groups of parent-adolescent dyads: the 'goal orientation' of the project- that is, whether the dyad's joint project is focused or diffuse in nature. A recent study of the sun-protection-projects of families demonstrated that participant dyads could be reliably categorized as being focused or diffuse, and that this goal orientation was a salient contextual issue for understanding the progress of parents and adolescents towards their sun-protection-projects (Young, Logan, et al., 2005). In this study, focused participants were described as having well-defined, explicit goals, shared in common by the parent and adolescent; while those who were diffuse were characterized by a lack of common understanding and disparate strategies to attain goals, or the intrusion of some other family project that took precedence over the sun protection-project. Moreover, in the two example dyads presented in Young et al.'s (1997) study on the role of emotion in the construction of career, one could be characterized as having focused goals, with a high degree of emotional support and clear, mutually agreement. In contrast, the other dyad had a career project with less congruent goals, and interactions characterized by tension and disappointment- a diffuse goal project. The notion that goal orientation is an important contextual factor in understanding the family career development projects of early adolescents is one that is congruent with the findings of a recent, non-comparative study on this phenomenon (Young, Marshall, et al., 2005). Specifically, during the analysis phase of that study, the idea emerged that that some parent-adolescent dyads had focused goals while other dyads had diffuse goals, and that the patterns of action may be different for the two groups. However, lacking a way to conduct between-groups analyses of action theory data, we were unable to explore this possibility any further. With the development of the action-project differentiation procedure, it is now possible to follow up on the idea that clustering dyads according to whether their joint projects were focused or diffuse is a meaningful way to explore variation in family career development projects. Therefore, the primary question of this study is, "what are the similarities and differences in the actions and projects of parents and adolescents, for dyads whose project-related goals are focused, in relation to dyads whose project-related goals are diffuse?" Method Data set. The data that were analysed in the current context were originally collected for a study on the career development projects of parents and adolescents from disadvantaged families (Young, Marshall, et al., 2005). The existing action-project protocol was employed in full for the original study (i.e., data collected through three interviews plus the 6-month monitoring period; findings generated through two rounds of within-case action theoretical content analysis followed by cross-case analysis of the sample as a whole. The data record consisted of (a) transcripts of all portions of the first and third interviews, (b) the narrative summaries presented to participants in the second interview, (c) researcher telephone logs and participant self-report logs for the monitoring period, and (d) summary descriptions of the results of within-case data analysis, for each participant dyad (with the exception of one dyad who failed to complete the third interview due to geographic relocation, and two other dyads who failed to return their self-report logs). I was a member of the research team for the original study, serving as a co-interviewer for six of the dyads, and participating in all team-based analysis meetings. 0 (months) 2 I L 10 12 14 20 22 1_ 24 Recruitment Cross-case Analysis I 1 Data Collection & Within-case Analysis Participant 1 Participant 19 \ Participant 2 s Participant 20 Participant 3 1st Interview 2nd Interview 3rd Interview J 1 month | 6 months I 1 month v. J -— Y ' V . v ' Transcription and Initial Analysis Monitoring Period Transcription and Final Within-case Analysis Figure 3. Approximate sequence for data collection and analysis in the Family Career Development study from which the current data set was obtained. Participants. The data were originally collected and analysed over a three year period (Autumn 2000 to Spring 2003), from 20 parent-adolescent dyads living in a small urban centre in British Columbia (see Figure 3 for details on the sequence of data collection). The adolescents were 12 to 14 years of age at the beginning of their research involvement, and all lived primarily with the parent who participated in the study. The parents were 32 to 51 years old, with 10 of them being in a sole parenting situation at the beginning of their research involvement. In addition, the original study was designed as an exploration of disadvantage (broadly defined) and career development. As such, there was an effort to select participants who described themselves as having some form of disadvantage in their lives (e.g., financial, health, family configuration, recent immigration). The participants were all volunteers, but received $100 and a certificate of participation at the completion of their involvement. The analysis process revealed that eleven of the dyads could be identified as having projects with goal orientations that were predominantly focused, eight were found to have projects with goal orientations that were predominantly diffuse, and one dyad was unclassifiable in terms of its goal orientation. Participants were recruited by having various members of the original research team give presentations about the study in grade 7 and grade 8 classes throughout the schools within the urban centre, requesting the involvement of students, and providing an explanatory letter for them to take home to their parents. The letter contained a telephone number for families to contact the research team, i f they were interested in participating. Telephone screening involved further explanation of the time commitment required of participants, and asking a number of questions to ensure that all dyads met the age and living-arrangement requirements of the study. A number of potential participants were excluded from the study at this stage, due to a lack of interest in committing so much time and energy on their part, a failure to meet the inclusion criteria of the study, or ethical limitations to their involvement (i.e., one family was excluded because the parent had a dual relationship with one of the researchers). Research team. The present study was conducted with a research team composed of three persons, two of whom were also involved with the original study. Two team members conducted an initial division of participants into groups, and an assessment of the practical significance of the various differences between the groups. The initial findings generated out of this process were then presented to the third researcher (who was familiar with the original data set, but not otherwise involved in the differentiation analysis) for further discussion, until consensus decisions about the patterns of similarity and difference were reached. Although this team is somewhat smaller than those found in the majority of existing action-project research, it conforms to the precedents set in Young et al.'s (2003) study of the family career development projects of Chinese-Canadian families, and Young, Logan, et al.'s (2005) research on sun protection-projects. Those two studies also involved novel analyses of existing action theory data sets and, in both cases, the research team consisted of two researchers who conducted the initial re-analysis and re-coding of the data, and another team member who reviewed and discussed the initial findings. It must be noted that, although the third researcher served a somewhat similar function as the 'auditor' role employed in other forms of consensus-based research (e.g., Hil l et al., 1997a), she should not be thought of as external to the research team, but was an integral member who was as familiar with the raw data as the other team-members, and whose voice carried equal weight in the analysis. The research team for both examples of the action-project differentiation procedure was composed of: (a) myself, Jose Domene, a PhD candidate in the Measurement Evaluation and Research Methodology program at the University of British Columbia (UBC), with a background in career development, child and youth mental health, and family therapy; (b) Riibab Arim, a PhD student in the Development, Learning and Culture program at UBC, with a background in adolescent social development, family relationships, and counselling psychology; and (c) Dr. Sheila Marshall, a professor in the department of Social Work and Family Studies at U B C , who was the co-investigator for the study from which the data set was originally generated. Her areas of expertise are adolescent social identity development and adolescent-parent interactions. Ms. Arim and I conducted the initial analyses, and Dr. Marshall served as the third analyst for both of the re-analyses. Analysis. Prior to examining the similarities and differences in the actions and projects of focused versus diffuse dyads, it was first necessary to establish the goal orientation of each dyad. Given the nature of the grouping construct, no a priori selection of participants was performed, because it is impossible to determine whether a dyad's project-related goals are focused or diffuse before observing their actions over time. Consequently, we proceeded directly to examining the existing data, in order to determine whether goal orientation is a salient feature in this sample, and how it was expressed in the recorded action of the participants (this corresponds to part 3 of Figure 2). The nature of dyads' project-related goals was examined by the two primary members of the research team, who engaged in close reading of the existing data set and discussion of possible interpretations. Distinct goal orientations were found to exist for different sets of families within the sample, with those different patterns being evident in the on-going action of the dyads. Specifically, 11 dyads had projects with goals that were similarly understood and shared between the parent and the adolescent; and where the specific actions that the participants engaged in were congruent with achieving those goals. For example, in one mother-daughter dyad where their project was "to continue with and expand upon their conversations about the daughter's possible future careers and other aspects of her current development," the mother conceptualized her role as "continuing to initiate conversation topics, providing suggestions and advice, and actively listening." and the daughter strived to be "engaging in conversation ... providing information and opinions, and developing strategies.to direct the topic of conversation more often." These dyads were clustered together to form the group of persons with projects that were focused in terms of their goal orientation. In eight of the dyads, the purpose of the project was either understood in different ways by different members of the dyad, and/or contained components that did not match with each other. Moreover, in these dyads, there was far less congruence between their ongoing action -and the goals that they were attempting to accomplish. For example, in one mother-son dyad, where the project was "to develop and promote a balance between the son's increasing independence and responsibility," the mother saw the project as an opportunity for the son to accept more adult roles, including taking responsibility for his own schooling, home duties and other obligations. In contrast, the adolescent's goal was to obtain greater freedoms to engage in activities that he enjoyed doing, because "I do want to become an adult, but not immediately." Lack of congruence between participants' goals and their actions is illustrated by a dyad where a goal for the project was for the daughter to become more trustworthy and deserving of freedom, but where the daughter engaged in actions such as lying to her mother about her whereabouts and companions, and subsequently getting caught by her mother. Finally, one case was unclassifiable in terms of its goal orientation, because its project exhibited some characteristics that were strongly indicative of diffusion, but also other characteristics that would be strong indicators of a focused project. Specifically, the mother and son in this dyad had different, though not mutually exclusive, goals and ideas of the purpose of project: the mother saw the project as a means to promote education, whereas the son conceptualized it as a way to "narrow down" his career options. However, despite these differing goals there was relatively little conflict or disagreement as they acted towards their project. Instead a close and emotionally supportive relationship, typical of focused families, was evident. Additionally, the classification of goals as 'focused' or 'diffuse' was facilitated by an examination of participants' commitment to their project, although the relationship between commitment and goal orientation was complex. There was substantial variability in commitment towards the project across both groups. Participants with focused goals tended to be fully committed to achieving their project and, in the few cases where there was variation in level of commitment, it was the adolescent who exhibited less commitment. In contrast, in every case where the parent, was not committed to the project, or where parent and adolescent were selectively committed to different (often opposing) aspects of their joint project, goal orientation was always diffuse in nature. For example, in one family, parent and adolescent were so ambivalent about pursuing their project that relatively few successful joint actions occurred in this domain. In this case, the end of their research involvement, the career project had been put aside in favour of focusing on a relationship project, "it didn't so much become career exploration as just communication between the two of us." Goal orientation was a salient distinguishing feature of the development projects of these twenty parent-adolescent dyads. However, we found that the nature of the goals could be better understood as occurring along a continuum, rather than as two discrete categories. For some dyads, the career development project was clearly focused or clearly diffused, while other dyads' projects reflected both kinds of goal orientation, in differing proportions. Despite the complex nature of goal orientation, it remained possible to classify 19 of the 20 dyads as having goals that were predominantly focused, or predominantly diffuse. The quality that differentiated the two kinds of goal orientations was the degree to which the goal was held in common by both dyad members (as reflected by their action) and, secondarily, the relative commitment that parents and adolescents demonstrated towards achieving the project. With virtually all of the cases successfully classified as having predominantly focused or diffuse goals, we proceeded with the next phase of the differentiation procedure (see part 4 of Figure 2), to expand understanding of what it means to be a focused dyad or a diffuse dyad, by identifying the similarities and differences in the actions and projects of people within the two groups. The comparative analysis began with an examination of the data sets of the 11 dyads with focused goals as a group, and the 8 dyads with diffuse goals as a separate group, to determine the patterns of action and kinds of project content that commonly occurred within each sub-group. The two primary members of the research team conducted these cross-group analyses using established action-project analytical methods and two specific guidelines to make their conclusions: (a) does a finding occur frequently across the group, and (b) did dyad members make any explicit claim that a particular action was important to their project? A number of findings emerged as predominant tendencies for parents and adolescents within each group. These findings covered a range of issues, including the nature of the projects that were selected, the degree of progress made towards those projects, the barriers that were identified, the kinds of project-related activities that occurred, and aspects of the parent-adolescent relationship that manifested as they engaged in their joint project. Further description of the tendencies that emerged within each goal orientation group can be found in Appendix l . 9 Next, the findings for the two groups were examined in relation to each other, to determine the ways in which they were similar and the ways in which they differed (this corresponds to part 5 of Figure 2). The two primary analysts examined the findings for the two 9 Normally, the product of this intermediary stage of analysis would not be reported. However, because this is a methodological paper, I have included them (as well as the materials presented in appendices 2 through 6) to provide a clear audit trail for readers. groups in relation to each other, scrutinising the data to determine the ways in which they were similar, and different. A chart was created to facilitate this process (see Appendix 2 for an approximation of the original, hand-written chart). This allowed us to examine the supporting and contradictory cases for each finding, across the two groups. The standard of practical significance was used to make decisions about whether the two groups were similar or different for each finding. Specifically, we used the chart to identify how frequently a theme or aspect was present across the group. Then, we referred to the individual cases within each group to determine how it was manifested in the actual data, in order to determine the magnitude/importance of that theme. This involved examining both the supporting and the contradictory cases. Then, through a process of discussion and consideration of alternative interpretations (where interpretive differences existed), we applied the consensus decision-making strategy that is typical of action-project research to the data. After making our decisions about the practical significance of the differences that emerged, we re-examined the original data sets once more, to ensure that our overall conclusions matched the actual data from the individual cases, and to ground our findings in specific examples from the data set. This process allowed us to generate a draft set of findings regarding the patterns of actions and projects of the two groups. These preliminary findings (see Appendix 3) were sent to the third analyst on the research team, for review. Feedback from the third member of the research team shed further light on the data. For the most part, the draft description of findings resonated well with her understanding of the data. However, two specific concerns were raised: (a) from her perspective, the diffuse goal families' engagement in fun activities required further clarification ("Was it that they experienced sporadic events or only remembered and reported events sporadically?"), and (b) she disagreed with the use of the term 'breakdown' in characterising the communication difficulties experienced by the diffuse families ("This term 'breakdown' is a bit problematic ... I do not think any of the dyads stopped talking ... use another term that is less like the pop culture terminology and identifies the range of difficulties that the dyads experienced or reported"). The two original analysts then met again to discuss the implications of that feedback, and to modify the written summary of the findings accordingly. This process yielded a final set of conclusions, reflecting the consensus opinion of the entire research team, about the differences and similarities in the family career development projects of dyads with a focused goal-orientation and dyads with a diffuse goal-orientation. Findings In general, we found goal orientation to be an important contextual feature in understanding the family career development projects of these adolescents and their parents., Dyads whose career projects featured goals that were focused in nature were readily distinguishable from dyads with diffuse goals. The actions and projects of the two kinds of participants differed from one another in many ways, although there were a number of similarities as well. We have organised the findings to first describe projects themselves (in terms of their nature, progress and barriers encountered), then to detail the types of project-related activities that tended to occur for persons in each group, and finally, the patterns of relationship and communication that were encountered. This framework is based on the organisation of results found in previous research on family career development projects (e.g., Young et al., 2001), but modified to describe the specific themes that emerged in this study. Kinds of projects. There were noteworthy differences in the kinds of projects that dyads within the two groups tended to engage around. For the participants with focused goals, the joint career exploration and planning for the future was very much embedded in the maintenance of parent-adolescent relationship. For example, one mother-son dyad was engaged in a project "to continue to maintain their shared relationship while exploring options for the son's future social and academic life, leading to young adulthood." For virtually all of the families with focused goals, a primary component of career development was the development and maintenance of a close, functional relationship between the adolescent and his or her parent. As one adolescent claimed, "[the project] went good because I tried to bring a strong relationship with my Mum." Although the relationship project was present in the on-going interactions of many of the diffuse goal dyads, another dominant theme in the family career development projects of this group was to work on achieving appropriate levels of independence, and negotiating (with varying levels of success) the amount of autonomy that the adolescent should be granted. As one diffuse group dyad perceived it, the main purpose of their project was to "develop and promote a balance between the son's increasing independence, and responsibility." It must be noted, however, that issues of autonomy and control were not entirely absent in the other group. Instead, it appeared as i f a majority of the parents and adolescents with focused goals were in relative agreement about what is appropriate at that particular stage in life, so that independence and autonomy were not issues that parent and adolescent needed to explicitly focus upon, within their career projects. Progress achieved. In terms of the progress made by dyads towards achieving their i • family career development projects, the data revealed substantial within-group variation for both groups. However, for focused dyads, all made at least some progress, and many made significant achievements, to the point where the project was substantially achieved. As one mother explained, "[Engaging in the project] pushed me to say this is what we gotta do. It pushed [daughter's name] into saying we want to try this.. .the reward at the end of it is being able to talk, with her having this openness that we didn't have before." The amount of progress made was somewhat different for the dyads in the diffuse group. Although some made substantial achievements, many had periods of progress alternating with periods of stalling, or even regression. The equivocal success that was experienced by many of the diffuse goal dyads is reflected in one mother's statement that "I don't know how much of the career exploration we got to because, you know, we were focusing on things that he really wanted to do, like the circus and things like that, and going to the juggling place and things like that. So he's explored some, but he hasn't explored all of them". Moreover, the diffuse group included both of the dyads where the original project was abandoned, and replaced by another project that was perceived as more important. Also, for some diffuse group dyads, their project was perceived as being successful by one of member but not the other. This difference in perspectives was most blatant in the following discussion, where a daughter was trying to convince her mother that their project was a success (but was also evident in other dyads): P35 I know how we can do it, so that's, as far as the question that you had why I didn't reach my goal? So that I'm serious why I didn't reach my goal with this bank card, and you buying your things, I didn't reach my goal A41 But you did reach your goal because you said what you were feeling. If you would have gotten angry and yelled at me and not given me a chance to speak then you wouldn't have reach your goal. But now that we've talked about it, and we've talked about it before also, I believe you reached it. P36 I don't know if that is reaching the goal or not. Because, A42 Well, our project P37 I don't know if I, I don't know if I got through. A43 Our project was to like, umm, what she said was to remain connected and learn about each other more and our expectations and me growing up as a teenager. But all that has to do with talking, and P38 Yeah we talk A44 Yeah we do but how much of that talking actually is something rather than just, P39 Yeah but A45 Than, P40 But yeah ok. So that's why I don't know if I'm answering the question there correctly or i f you're answering it correctly. A46 Well it's all your own opinion. Everyone has different opinions on how they've reached their own goal, and I think mine is P41 And my goal, my reaching the goal, I don't think I did reach it. In contrast, the dyads with focused goals tended to have very similar evaluations of how much progress had been made on their projects. For example, over the course of the monitoring period, a mother's and daughter's independent reports of how well their project was progressing matched each other well, with both perceiving the progress to be small towards the beginning of the six months, but then reporting that their project was going well several months later. Perceived Barriers. When dyads were questioned about barriers to the success of their projects, the most common response of participants within the focused group was that they lacked sufficient time to engage in activities together. For example, a focused goal mother stated, "just timing again, in you... It's, you're a family of 5 and he's not the only child, and there's other people who need the time and attention; and fitting in the time to do, to carry on the conversations." Even though dyads with focused goals had a greater frequency of engagement in joint activities, these participants also believed that they did not have enough time to spend with each other. Although a substantial number of dyads with the diffused goals also complained about not having sufficient time together, the most commonly identified barrier for that group was relational friction leading to intentional avoidance of each other. As a daughter disclosed on one occasion, "I tend to get angry right away whenever Mum calls my name," and another time, she stated, "My bad temper can still get in the way of talking ... I ' l l just go up to my room and lock myself in." Although lack of time was not unimportant to participants with diffuse goals, it the most salient barrier for dyads with diffuse projects was a relational pattern of coming into conflict or friction, followed by a period of not being able to be with each other (a pattern which was largely absent in the focused group). Activities. Regardless of the orientation of their goals, parents and adolescents working together on family career development projects engaged in range of joint activities, including casual conversations, formal discussions, use of humour and good-natured teasing, and taking the time to talk while engaged in other activities (e.g., driving to the adolescent's sporting events). Dyads reported sharing personal thoughts and emotions, as well as their concerns with each other. Adolescents reported seeking the advice and assistance of their parent, while the parents reported providing advice, support, reassurance, and practical assistance (e.g., helping to prepare a resume). There were also reports of arguments, disagreements, occasions where parent and adolescent would avoid speaking to one another for periods of time, and situations where the adolescent was selective in what he or she chose to share with the parent. A number of dyads reported engaging in civic activities together (e.g., volunteering at a soup kitchen) or working together at the mother's place of employment, and using these situations to discuss the adolescent's future; some dyads even used recreational activities (e.g., shopping together, watching movies or television, going to amusement parks) in the same manner, with the mother taking the opportunity to turn that activity into an opportunity to reflect on possible careers. Many dyads also recorded instances of simply having fun together, whether it was as simple as going on bike rides together or as momentous as taking a trip to Europe together, without the rest of the family. Interestingly, there was no practical difference in the range of activities engaged in by participants with focused goals versus participants with diffuse goals; examples of each of these activities could be found in both groups, and most were identified in about as many focused dyads as diffuse dyads. However, where goal orientation became a distinguishing feature was in the frequency of participants' joint engagement in their project-related activities. Specifically, the dyads with focused goals tended to exhibit more instances of joint action over the six-month monitoring period than dyads with diffused goals. An additional difference between the groups was noticed for recreational and social joint activities: in most of the dyads with focused goals, these kinds of activities appeared to be an integral part of participants' schedules, a naturally occurring part of the process of maintaining a good relationship. For example, one mother from a focused dyad reported engaging in recreational activities (e.g., weekends away at their cabin, playing Frisbee, visiting a corn maze) and/or attending her son's sporting activities, in every fortnightly telephone monitoring interview during the monitoring period. In contrast, diffuse group participants' descriptions of their joint activities contained relatively fewer instances of joint recreational or social activities. Furthermore, having fun together seemed to be something that was more associated with special occasions (such as going on a vacation together), rather than an integral part of these families' weekly routine. In fact, many adolescents from the diffuse group preferred to spend their free time with peers. In one diffused dyad, for example, the daughter spoke of having fun with friends in virtually all of the telephone monitoring contacts, but did not report engaging in any recreational activities with her mother. Correspondingly, the mother's logs contained only one instance of the two them participating in a joint recreational activity, over the entire six months of monitoring. Relationship and communication. By definition, the presence of conflictual goals and/or more tension in the interaction of dyads is one of the ways to distinguish diffuse goals from focused ones. The research team attended to this conflict and tension in our determination of whether a dyad fell into the focused or diffuse group. It is self-evident that between-groups differences in the way that dyad members relate to and communicate with each other exist, because this is part or the very definition of the distinct goal orientations. Specific types of communication problems that were identified in various members of the diffused goal group included: not being able to complete a conversation due to anger, avoiding talking to each other for periods of time or only responding in a minimal way, manipulative behaviour by the adolescent and/or the parent, the adolescent omitting information or engaging in outright deception, and parental ambivalence when their adolescent attempted to engage in conversation. In contrast, the dominant patterns of communication within dyads that had a focused project-related goals included a willingness by both parties to self-disclose and share personal information, adolescents who were willing to initiate conversations, attempts by the parent to listen rather than to lecture, a commitment to finding the most suitable times to for productive communication and then seeking out those opportunities on a regular basis, and frequent verbal encouragement of the adolescent by the parent. We do not intend to imply that dyads from the focused group never had problems with communication, or that families with diffused goals never experienced good communication, or even that all the specific characteristic described above were present in all members of their respective group. However, it was clear that participants from the two goal orientation groups exhibited a tendency to engage in different kinds of communication activities. Given the way that 'focus' and 'diffusion' are defined, these findings serve more to confirm the accuracy of the grouping process than to shed any new light on the projects and patterns of action that are associated with the different goal orientations. We did find one communication-related finding that was surprising. Specifically, conflict, either project-related or around other life circumstances, was evident in the joint actions of most dyads, regardless of goal orientation. In fact, a majority of the dyads in both groups reported experiencing some form of conflict between the parent and adolescent over the course of their research involvement. However, the conflicts that occurred between parents and adolescents with diffused goals were more intense and hostile than the conflicts that occurred for dyads in the focused group. One possible contributor to this difference was that a majority of the parents in the former group had a tendency to react with strong, negative emotions when having disagreements with their son or daughter. For example, one of the diffused group mothers repeatedly reported reacting with such frustration, anger and fear that she would leave the room. While this was an extreme case (the mother reported that the son's actions triggered memories of her abusive ex-husband), strong emotional reactions were frequent components of the dyadic relationship for many of the participants with diffuse goals. In contrast, very few of the focused goal parents reported reacting so strongly in their discussions, and those who did react strongly also made efforts to manage their emotions (e.g., "I was reminding myself to stay calm, to listen. Inside I was freaking out that he had to be exposed to this [violence and drug-use in the school], but relieved he would want to talk about this with us"). Another relational difference between the two groups is the differing levels of anxiety versus trust that were present in parents' relationships with their adolescents. Many parents in the diffused group appeared to be anxious and worried about the immediate and future functioning of the adolescent, typically regarding safety or academic functioning. For example, in one diffused goal dyad, the mother explained, "Like I said she's really not a, you know, a bad girl, yet. You know she is 16; she is at a dangerous age and you do hear all sorts of bad things ... we don't want her to get into trouble and be pregnant at 16 or start smoking or drinking or start failing school. We want her to have a healthy normal life that, that I don't, you know that she's not abused by anyone or anything." In contrast, most parents in the focused group believed that their previous parenting was of sufficient quality to allow their children to cope with problems in the present. In response to a question about her concerns, one of the focused project mothers stated, "I thought, oh my goodness, like what i f he decides to do something completely different, how will that affect him? ...but I don't, I know that whatever we put in for whatever reason, it is building him into a really good solid guy and somebody with a good head on his shoulders." Instead of anxiety and worry, the relationship for most of the dyads with focused goals was built on a foundation of trust. For example, "I know there's going to be some [bad decision] stuff come up, but hopefully, I'm just hoping, that we've got enough of a foundation that they're comfortable enough that - . They might not come home and the second they walk in the door tell me, but hopefully that they, that there's enough there that they can at least talk to me about it." Discussion The findings that were generated in this study demonstrate that the presence of focused versus diffuse project-related goals is an important contextual factor in understanding the family career development projects of adolescents ant their parents. Specifically, the findings indicate that participants with differing goal orientations also differed in terms of the content of their projects, the progress that was made towards those projects, the identification of relational friction as a barrier, how often they engaged in project-related activities, the intensity of their conflicts and arguments, and the level of anxiety/trust that was evident in the relationship. Ways in which the two groups were similar included the perception of lack of time as a barrier to achieving their projects, the kinds of activities that participants engaged in to achieve their project, and the presence of some degree of conflict in the parent-adolescent relationship of virtually all the dyads. Clearly, it is useful to attend to goal-orientation when attempting to understand the career development projects of adolescents and their parents. The findings also suggest that it may be important to do so as well, both in terms of future research, and career counselling practice. At the level of research, not only will examining families with focused projects separately from families with diffused projects lead to richer descriptions of their engagement action around that project, but it may uncover systematic patterns that would otherwise be missed without this separation into focused versus diffused groups. For example, the complex relationships between goal-orientation and commitment to the project required a division into separate groups, before they could be teased apart. Without attending to whether or not the projects were focused, the only conclusion would have been that in some dyads, both members are committed; in some, only one member is highly committed; and in a few, neither member appears to be highly committed towards their project. Moreover, in combination with the results of Young, Logan, et al.'s (2005) sun-protection study, these results suggest that goal orientation may be an important contextual feature of joint projects across different domains of family functioning. Future studies on family projects should attend to goal orientation, and consider conducting between-groups analyses to fully delineate the projects and patterns of action of the different kinds of participants that may be encountered. Another fruitful direction for future study is to explore the limits of goal orientation as an important feature: is it primarily within parent-adolescent projects that distinctions between people with focused versus diffuse goals arise, or do the different goal orientations also occur in the joint projects of other types of persons (e.g., counsellor-client dyads, spouses, work groups in employment settings, sports teams)? The relative merit of having focused goals in these career development projects was evident in the data. Therefore, at the level of career counselling, this study suggests that it may be fruitful to attend to goal-orientation when conducting career counselling with adolescents and their parents. If a family appears to be stalled in negotiating the career development of one of their adolescent children, there may be benefits to working with parent and adolescent to develop more common understandings of what they hope to achieve for the adolescent's future, and to focus on their mutual communication and relationship as a possible contributor their lack of progress, before proceeding with standard vocational interest/aptitude assessments, or career exploration interventions. Interestingly, eliminating conflict altogether would appear to be less important than assisting the family to manage conflict to maintain some level of communication and reduce the emotional reactivity of the family members, especially the parents. However, several steps need to occur before the findings of this study can be implemented in clinical practice. Besides the replication that would be required of any study that explores a relatively new area of inquiry, the fact remains that the action-project differentiation procedure is essentially descriptive in nature. To assume that improving communication will transform a dyad's goals from diffuse to focused, or that fostering focused goals will improve the degree of progress made by families who are 'stuck' requires causal linkages between these variables, linkages that have yet to be established. Therefore, the connections between goal orientation, communication and progress in career development projects should be explored further, using approaches that are designed to generate explanatory rather than descriptive results. Only then should the effectiveness of intervening at the level of goal orientation be examined as a viable way of improving career counselling practice. The value of this study is tempered, to some degree, by a number of limitations that existed within the design. The action-theory differentiation procedure produces descriptive, rather than explanatory results. An in-depth description of family career development projects in the context of goal orientation is certainly valuable, but further study of the phenomenon, using explanatory methods, would facilitate the application of these findings to clinical practice. A second limitation is that, despite efforts to recruit parents of both genders in the initial study, all the dyads were composed of mctf/zer-adolescent pairings. While this situation may, in itself, suggest something about fathers' engagement in projects with their children, it clearly impairs the transferability of the present findings to situations where the parent in question is male. Conducting a similar study, but with fathers as the parent in the dyad, is necessary to construct a more complete story of how focused and diffuse parent-adolescent dyads engage around the adolescent's future. Finally, during the group formation phase of the study, there were some indications that dyads' goal orientations may fall more along a spectrum than into two, truly discrete categories. Although we were still able to classify 19 of the 20 dyads as being predominantly of one goal orientation or the other, this failed to account for the different shades of meaning that may exist within goal orientation. Framing the constructs of focus and diffusion as two poles of a spectrum may lead to alternative findings. At minimum, operationalizing goal orientation in this manner will permit different kinds of questions to be asked about this important contextual feature of parent-adolescent joint projects. Overall, this study found goal-orientation (i.e., whether people's projects are focused or diffused) to be a useful way to organize parent-adolescent joint projects within the domain of career development: family career development projects could be readily categorized as being either 'focused' or 'diffuse,' and these differing goal-orientations were associated with distinct patterns of action and degrees of engagement in their projects, over time. Specifically, differences were found in the types of projects that tended to be selected, the amount of progress that was achieved, the frequency of dyads' joint engagement in project-related activities, and the kinds of relational and communications patterns that were most often evident. These findings have provided a greater knowledge of people's engagement in family career development projects than would be possible without some form of between-groups analysis. They also suggest that goal-orientation is an important dimension to attend to in research, and has the potential to be useful to consider in career counselling practice with adolescents. Study 2: Gendered Dyadic Configuration Parents and adolescents of both genders may be found in a family, which creates four possible configurations of parent-adolescent dyads who may choose to participate in an action theory study on family career development (mother-daughter, mother-son, father-daughter, and father-son). The two configurations that are relevant for this demonstration of the action-project differentiation procedure are the mother-son dyad and the mother-daughter dyad (all parents in the data set were mothers). Existing literature in the fields of both vocational psychology and family studies provide support for the contention that different mother-adolescent dyadic configurations may be an important grouping construct for examining joint career development projects. That is, there is sufficient evidence that mothers and sons interact in distinct ways from mothers and daughters around the child's career development, to warrant examining the patterns of results of these two sets of dyads in relation to each other. Research has repeatedly established that mothers are an important influence on their children's career development. Adolescents have been found to rely on parents to provide advice about major life issues (e.g., plans for education, employment options, plans for family), and those who were receptive to parental advice tend to be more positive about their future (Tucker, Barber, & Eccles, 2001). Moreover, parents are perceived to be a primary source of influence on their children's career development, from the perspective of both parents (Birk & Blimline, 1984) and the sons and daughters themselves (Otto, 2000); although it is recognised that parental influence is more easily accepted for some career decisions than for others (Bregman & Killen, 1999). Also, studies that have examined fathers and mothers separately have concluded that, while both are important, more adolescents identify mothers as a major influence, in comparison to identifying their fathers as a major influence (Kotrlik & Harrison, 1989; Otto, 2000; Schultheiss, Kress, Manzi, & Glasscock, 2001; Tucker, et al., 2001). More importantly, several studies have yielded results suggesting that the joint actions of mothers and daughters around career development may be different from the joint actions of mothers and sons. For example, adolescents' perceptions of their mothers' importance in influencing their career development differs according to gender, with daughters ranking their mothers as a significantly higher influence, than sons (Paa & McWhirter, 2000). Daughters are also more likely than sons to cite their mothers as an important influence on their career choice, at least for adolescents who intend to pursue careers outside of science/engineering (Dick & Rallis, 1991). Mannheim and Seger (1993) found that mothers and daughters are more closely matched than mothers and sons on the work values that they deem to be important. In terms of career commitment, there is a significant relationship between commitment and non-foreclosure, and between conflictual independence and close attachment for late adolescent females, but not formates (Blustein, Walbridge, Freidlander, & Palladino, 1991). Also, Grotevant and Cooper (1988) described a series of studies that they conducted in the 1980s which, in combination, indicate that communication with mothers is related to engagement in career-related exploration behaviours for adolescent females, whereas no such relationship exists for mothers and sons. Two studies have directly explored dyadic configuration differences in the career-related patterns of actions. Paa and McWhirter (2000) found a number of significant differences in the patterns of action of mother-son versus mother-daughter dyads, within the domain of career development. Specifically, mother-daughter dyadic interactions were characterized by higher levels of positive feedback, support for autonomy, and openness to discuss the youth's ideas for possible career choices. Young, Friesen and Pearson (1988) also found significant differences, but in the opposite direction: mothers were more often engaged in information provision, demonstration of interest, expression of affirmation and understanding, and use of avoidance with their sons, than their daughters. It should be noted, however, that the Paa and McWhirter (2000) study collected data from adolescents alone, while Young et al. (1988) interviewed only the parents. The somewhat contradictory nature of these two sets of findings underscores the need to examine this phenomenon from the perspective of joint action, incorporating both parent and adolescent in the data collection. Going beyond vocational psychology, literature from the study of general family relations and communication furnish additional reasons to use dyadic configuration as a grouping construct in exploring the joint projects of adolescents and their mothers. Unfortunately, a vast majority of the extant research in this area has focused on either (a) differences between mothers' and fathers' communication/relationship with their adolescent children, or (b) differences between sons' and daughters' communication/relationship with both parents (Russell & Saebel, 1997). For example, in Youniss and Smollar's (1985) extensive study of relationships in adolescence, the authors conducted comparative analyses of many aspects of mothers' versus father's interactions with children of each gender, but did not examine the possibility of statistically significant differences between adolescent girls' and adolescent boys' communication or relationship with their mothers. Unfortunately, the ways in which comparison questions are typically framed in this discipline provide little relevant information to address the specific questions that I have posed in this study. Nonetheless, the few studies that do attend to specific dyadic configuration provide some indication that mother-daughter dyads are distinct from mother-son dyads in terms of their communication. For example, it has been found that the frequency of conversation around important issues and amount of self-disclosure is higher in mother-daughter dyads than in mother-son dyads (Noller & Callan, 1990). Research on parent-adolescent conflict suggests that daughters engage in conflictual interactions with their mothers much more frequently than sons, although this varies according to the specific area of the adolescent's life (Ellis-Schwabe & Thornburg, 1986). Similarly, Poole and Gelder (1985) found that in comparison to sons, daughters made a higher number of decisions on an independent basis. However, Poole and Gelder (1985) also found that daughters reported perceiving their mothers' opinions as significantly more important to them, than did sons. It must also be noted that some studies have failed to find differences in the nature of parent-adolescent relationships (e.g., Hi l l & Holmbeck, 1987; Cooper & Grotevant, 1987). No previous research has explored family career development projects in light of the various distinct configurations of parent-adolescent dyads that exist. However, there appears to be sufficient reason to use this grouping construct as a way of framing a comparative analysis of data collected for Young, Marshall et al's (2005) study on the family career development projects of parents and adolescents. A majority of the existing findings suggest that mothers may be more important to the career development of their female adolescent children, and that the actions of mother-daughter dyads are distinct from those of mother-son dyads. It is evident from the current state of the literature that there may be similarities and differences in the actions and projects of mother-daughter dyads in relation to mother-son dyads, but that the nature of these similarities and differences has yet to be clearly delineated. The present study seeks to address this gap in the literature by answering the question, "what are the similarities and differences in the actions and projects of mother-daughter dyads and mother-son dyads engaged in family career development projects?" Parenthetically, it must be noted that there is also a large body of literature demonstrating the existence of gender differences in many other aspects of adolescent career development (see Reid & Stephens, 1985; Swanson & Gore, 2000; Whiston & Keller, 2004, for reviews). However, this literature can provide only indirect justification for the use of dyadic configuration as a grouping construct in the present context, because the present study is an exploration of the joint engagement of adolescents and their mothers around career development, rather than an examination gender differences, per se. It is impossible to ascertain which of the findings related to gender differences in career development would be relevant, because joint engagement in career project with mothers is only one of a myriad of factors that could contribute to the gendered nature of career development in adolescence. Method Data set and research team. This study utilized the same data set and research team that was used in the goal orientation study (see description of Study 1 for details). There were ten mother-son dyads, and ten mother-daughter dyads. Analysis. Because it was not the focus of the original Young, Marshall, et al. (2005) study to examine gendered dyadic configuration, participants in this study were not originally selected with this purpose in mind (as would be recommended in part 2 of Figure 2). At the same time, knowing the gender of the adolescents in the dyads ahead of time greatly simplified the process of clustering participants into groups for comparison: dyads with male adolescents were classified as having a mother-son gender configuration, while dyads with female adolescents were classified as having a mother-daughter gender configuration. However, as stated in chapter 3 (specifically, part 3 of Figure 2), a stronger argument can be made for the meaningfulness of the between-groups analysis when an a priori groupings are confirmed by the patterns that are evident in data of the individual cases. Therefore, we examined the data for each dyad, to determine whether our pre-existing categorization corresponds with the groupings that emerged out of the within-case and initial cross-case analyses. Unfortunately, dyadic configuration failed to emerge as a meaningful theme at the within-case level. With certain exceptions (e.g., one mother stated that she was more worried about, and protective of, the participating adolescent than her other children, because she was the only girl in the family), gender configuration was minimally relevant to participants' engagement in career development projects. Indeed, for many of the cases, it would not have been possible to distinguish the gender of the adolescent from the projects and patterns of action in which the dyad engaged, without referring to names, personal pronouns, or the visual record of the interview sessions. A l l the same, it must be recognised that variation may only be detectable through comparative analysis, for some phenomena, and cannot be revealed by examining individual cases alone. In these situations, comparative analysis is required to identify the differences that exist for distinct sub-sets of cases. If the relationship between dyadic configuration and family career development projects is such a phenomenon, then it may still be possible to identify differences between mother-son dyads and mother-daughter dyads, even i f dyadic configuration did not emerge as meaningful in the preliminary analyses. Therefore, it could be useful to complete the comparative analysis, to determine whether any differences that are not apparent at the within-case level emerge at the between-groups level. The potential usefulness of proceeding with a between-groups analysis is underscored by the evidence from the existing literature to suggest that different sub-groups engage in their projects in distinct ways. Given the prior empirical evidence suggesting that the two dyadic configurations of interest are an important way to distinguish people's actions and projects within the domain of career development, the team decided to proceed with the between-groups analysis. Participant dyads were classified as belonging to the 'mother-son' or 'mother-daughter' group on the basis of the adolescent's gender. Once the groups were established, the two primary members of the research team identified the patterns of action that were commonly found within each of the two dyadic configuration groups (corresponding to part 4 of Figure 2). Specifically, we attended to how often a theme manifested across the group, and its apparent importance to the participants, in order to make decisions regarding what projects and patterns of action were typical of mother-son dyads, and what projects and patterns were typical for mother-daughter dyads. A summary of this process is presented in Appendix 4. The examination of the data then proceeded in a manner similar to the process of analysis used in Study 1 (see part 5 of Figure 2). After creating a chart to aid the decision-making process (see Appendix 5), the two primary analysts scrutinized the data set and, using the standard of practical significance and the consensus approach to decision-making, generated a draft description of the patterns of similarity and difference that were present within the data set (see Appendix 6). At the conclusion of the between-groups analysis, few practical differences were found between the mother-daughter group, and the mother-son group. The preliminary findings were reviewed by the third member of the team, who agreed with the content of the analysis. She did, however, express some concern that the description left with the impression that the conclusions were based solely on numerical counts of the data. This situation was discussed by the two original analysts, who subsequently modified the wording of the findings to more accurately reflect the analysis process that had occurred. A final description of the findings, reflecting the consensus opinion of the research team, is presented next. Findings Overall, the action-project differentiation analysis of indicates that dyadic gender configuration is not an important contextual feature in understanding the family career development projects of these twenty adolescents and their mothers. Mother-son dyads and mother-daughter dyads were largely indistinguishable, in terms of their engagement in career projects. The actions and projects of the two kinds of participants were characterized primarily by similarity, rather than difference. The findings are organised in terms of the descriptions of the projects themselves (in terms of their nature, progress and barriers encountered), then the types of project-related activities that tended to occur for persons in each group, and finally, the patterns of relationship and communication that were encountered. This organisation is based loosely on the descriptions of family career development projects found in previous action theory research (e.g., Young et al., 2001), but modified according to the nature of the themes that emerged from this data set. Kinds of projects. The kinds of family career development projects in which the two sets of dyads engaged were predominantly similar, in the sense that projects across both groups tended to include a component of developing and maintaining their mutual relationship. For example, one mother-daughter dyad's joint project was to "strengthen the mother-daughter relationship, where both will feel listened to and able to fully express their ideas and needs." Similarly, in one mother-son dyad, their joint project was defined as, "for the Mother and Son to maintain their close relationship and willingness to share ideas while discovering their new roles as they adjust to the son's transition towards young adulthood." Furthermore, a large minority of dyads within both groups also had projects that included negotiation of the adolescent's independence and level of permissible autonomy. Finally, some dyads with focused goals, and dyads with diffuse goals were present in each group, indicating that mother-daughter dyads and mother-son dyads could not be distinguished in terms of the goal orientations that tended to emerge within each group. One distinction between the projects of the groups was in the degree to which explicit vocational development goals were incorporated into the joint projects. This occurred more often in the projects of mother-son dyads, (e.g., "to maintain their respectful and supportive relationship while they both engage in the process of career exploration," and "to have exploration of career development where the format of interactions will continue to be transformed into one where the son is able to openly discuss his interests"), but was present in only two of the mother-daughter dyads. This pattern of findings is an indicator that, while some mother-son dyads are already framing their career projects in terms of specific vocational choices at this age, mother-daughter career development projects tend to not to do so. Progress achieved. Division of the sample according to dyadic configuration yielded no identifiable patterns in the amount of progress achieved, with equivalent numbers of both mother-son and mother-daughter dyads achieving substantial progress, less progress, and mixture of success and failure. Although two of the mother-son dyads abandoned their initial project and none of the mother-daughter dyads did so, this was not judged to constitute sufficient evidence of an important dyadic configuration difference. Therefore, it would appear as i f dyadic gender configuration is not a salient contextual factor for the amount of progress that dyads achieve on their family career development projects Perceived Barriers. Participants from both kinds of dyads tended to identify insufficient time spent together (primarily due to one or both of the dyad members being too busy) as an important barrier to progress on their projects. One distinction between the groups was in the identification of interpersonal friction and avoidance of each other as a barrier: this was a commonly identified barrier in the mother-daughter dyads, but not the mother-son dyads. One mother provided an example of the friction, "I was thinking how frustrating and exasperating this issue is... I was feeling that i f she were younger and smaller this situation would bring me to want to spank her," and, in a different dyad, a daughter exemplified how avoidance interfered with engaging in their project, "My bad temper can still get in the way of talking ... I 'll just go up to my room and lock myself in." This difference between the two sets of dyads must be understood in light of the finding that conflict was identified as a feature of the relationships of most of the participants, regardless of the configuration of the dyad. Therefore, the fact that it only tends to be perceived as an impediment to progress in the projects for mothers and daughters raises an important question: is it the case that, contrary to the Young et al., (1988) study, mother-son dyads do not use avoidance as a strategy to cope with conflict as often as mother-daughters? Alternatively, it is possible that the conflict between mothers and sons simply does not interfere with their engagement in joint projects in the same way that it does for daughters and their mothers. Activities. The project-related activities that mother-son and mother-daughter dyads engaged in over the course of their research involvement were largely similar. Specifically, activities that were common within both groups of dyads included (a) having conversations and discussions with each other, (b) joint leisure activities and "having fun" together, (c) joint engagement in the adolescent's sporting activities (with the roles of the mother including cheering and supporting her child, and providing transport to and from practices and games), (d) working together on specific job-related activities (e.g., resume-writing, job-searching) and (e) adolescents seeking support and advice from the mother who, in turn, provided it. Joint engagement in shopping was something that many mother-daughter dyads, but very few mother-son dyads, reported as an activity. However, given the fact that this particular activity is of minimal relevance to understanding career development, it was judged to be a trivial rather than practical difference in this data set. Although many of these activities appear to be minimally related to vocational development, the participants understood them to be related to their career projects. For example, "having fun" was seen as an extension of maintaining the level of communication required for the adolescent and parent to seek and provide advice, respectively. Also, one mother-daughter dyad described their shopping trips as an opportunity to discuss possible career opportunities that they noticed in the course of shopping. Finally, a number of mothers used the time in the car (traveling to and from sporting and other activities) to have meaningful conversations about various issues that arose, including career-related ones. Relationship and communication. Few noteworthy differences were identified in the patterns of communication and relationship for the two sets of dyads: mothers and sons, and mothers and daughters both tended to have relationships that (a) were characterized as being 'close' in terms of the bond between them, and (b) featured self-disclosure and the sharing of personal information. Moreover, although approximately half of both sets of dyads experienced difficulties in communication at some point during their research involvement, there was a clear tendency for mothers to verbally encourage their children irrespective of gender. Discussion These findings appear to indicate that the family career development projects of these adolescents and their mothers are similar, regardless of the gender composition of the dyads. With a few exceptions, the themes that emerge in mother-daughter dyads also emerge in mother-son dyads, in terms of the kinds of projects that are formed, the project-related activities that participants engaged in, the amount of progress that was achieved, and the kinds of barriers that were perceived as interfering with their projects. The action-project differentiation procedure was able to identify two substantive ways in which the groups were distinct. First, we found that explicit vocational projects were a clear characteristic of the mother-son group, but were only rarely present in the mother-daughter group. Second, although conflict was present in a vast majority of the participants regardless of dyadic gender configuration, we found that interpersonal friction and avoidance was commonly perceived to be a barrier to progress in the mother-daughter dyads, but not the mother-son dyads. It is important to note that conflict was not absent in the mother-son group; it is just that, unlike the mother-daughter dyads, this conflict was not perceived as interfering with the achievement of the family career development projects of mothers and their sons. In contrast to the majority of the prior empirical evidence, these findings imply that there are many commonalities between mother-daughter dyads and mother-son dyads as they engage in joint projects around the adolescent's future development. This may, in part, be due to the fact that distinct dyadic configuration groupings failed to emerge at the within-case level of analysis. That is, because the groups were defined exclusively on an a priori basis (rather than letting the theoretical reasons for between-groups comparison be confirmed by the patterns that naturally emerge from the data), it is likely that the various dyads within each group were quite different from each other, in terms of their actions and projects. If so, then there would have been very few common tendencies that distinguished each group, which, in turn, would have left very few characteristics, features or aspects to juxtapose or examine in relation to each other. Fortunately, this apparent methodological limitation, itself, suggests some intriguing possibilities about the nature of the family career development projects of mother-daughter dyads and mother son dyads. It may be the case that individual dyads, across both kinds of gender configuration, engage in a wide range of different career-development activities and communication. That is, it may be more accurate to state that individual dyads are very different from each other, rather than concluding that the actions and projects of two groups are similar. To frame it another way, the main similarity between the groups may well be that dyads within each group tend to by highly variable and different from other dyads within the group, in terms of the actions and projects in which they engage. However, the few between-groups differences that were able to be generated have some important implications for adolescent career development, and must be highlighted. Even though the focus of the study was on career projects (rather than the myriad of other projects that parents and their children engage in), only two of the mother-daughter dyads had projects that could be characterized as explicitly vocational in nature. That is, while these mothers are actively engaged with their sons on issues around employment, school performance and future occupational choice, they are engaging with their daughters on issues that are only more generally related to future careers (e.g., development of autonomy, maintaining a good mutual relationship). Furthermore, even these general career projects are being disrupted by normative parent-adolescent conflict for the mother daughter dyads, but not the mother-son dyads. These findings are particularly troubling, in light of other research indicating that mothers are of primary importance for adolescents seeking career advice (Kotrlik & Harrison, 1989; Otto, 2000; Schultheiss, et a l , 2001; Tucker, et al., 2001). If the career projects of mothers and daughters are not explicitly vocational in nature, then who is engaging in these kinds of projects with female adolescents? At best, this study implies that parents are focusing on vocational issues earlier, and more successfully, in the lives of their sons than their daughters (the adolescent participants in this study were 12-14 years old). However, i f mother-daughter career projects do not become more explicitly vocational in focus as the daughters become older, this finding implies that adolescent girls are receiving inadequate parental support for developing the occupational/employment aspects of their future careers. Parenthetically, it must also be noted that, for mothers and their daughters at this age, and career projects are defined much more broadly than work or finding a particular job. As a result, there is a great deal of overlap between these projects and other projects. However, what distinguishes family career development projects from other projects is an underlying desire to ensure or improve the adolescent's future success (as opposed to, for example their future relationships, health, or cultural awareness). Several concrete directions for future research in the field of family career development projects in the context of dyadic gender configuration arise from the findings of this study. First, it may be fruitful to revisit the design of this study to determine whether alternative data collection procedures would increase the meaningfulness of dyadic gender configuration as a grouping construct. For example, it may be useful to state that dyadic configuration is a focus of the study during recruitment, to allow dyads for whom the gender of the adolescent is more meaningful to self-select to participate. Additionally, it may be beneficial to ask more direct questions about how the gendered nature of participants' relationship may be affecting their joint engagement (in the same way that we asked questions about barriers that they experienced, and the project-related activities that occurred). Doing so increases the likelihood that gender configuration will emerge at the within-case level of analysis, and may generate a very different set of findings. However, i f the same pattern of 'more variation among individual dyads that between groups of dyads' emerges even after these additional measures are put in place, then a stronger case could be made that the career-related actions of mothers and sons are not very distinct from the career-related actions of mothers and daughters, at this stage of development. Additionally, the pairings that were examined in the present study are only two of the four possible dyadic gender configurations that are possible. It may be interesting to research all four kinds of dyads in relation to each other, to determine what overall patterns of similarity occur in the actions and projects of mother-son, mother-daughter, father-son, and father-daughter dyads. Doing so will provide a broader and possibly richer understanding of how parents and adolescents of different genders are jointly engaged in career development projects. Finally, another direction for future research is to extend this study to examine the actions and projects of older adolescents and their mothers. It could well be the case that between-groups differences become more important at later stages of development, a possibility that is highlighted by the fact that the adolescents in the Paa and McWhirter (2000), and Young et al (1988) studies were older than those who participated in the present study. Conducting a new study with dyads where the adolescent is closer to the end of High School will help to determine whether the same patterns of (a) mother-son dyads having projects that are more vocational in nature, and (b) conflict presenting a barrier to mother-daughter but not mother-son engagement in their projects, continue into the developmental period where adolescents need to be making concrete decisions about their future occupations. If this is the case, then the nature of mothers' engagement in career projects with their daughters versus their sons may be an important contributing factor to the gender differences in career outcomes that continue to persist in society, despite massive efforts to ameliorate these disparities. It is, therefore, important to pursue this line of research further. Achieving a more thorough understand the nature of dyadic gender configuration as a contextual feature of career development, across a greater range of adolescent ages and parental genders, is a necessary precursor to future research in this field. Once the nature of career-related action in the context of dyadic configuration is more fully explored, then other methods of social inquiry can be used to establish the nature and direction of influence between parent-adolescent career-related joint action, and the various gender-related differences in occupational outcomes that continue to exist in society. In this way, the action-project differentiation procedure will complement other qualitative and quantitative approaches to building knowledge in this field. CHAPTER 5: METHODOLOGICAL FINDINGS OF THE DEMONSTRATION STUDIES The demonstrations of the action-project differentiation procedure that were presented in Chapter 4 reveal several important qualities of the procedure, and lead directly to a number of specific suggestions for improvement. In this chapter, I address the method-related findings of the two demonstration studies. First, I discuss the overall differences between the design of the two demonstrations and the findings that were generated, and interpret the meaning of these differences, in terms of the development of the method. Then, I discuss specific problems that were encountered at different stages of the analytical process, and propose refinements and possible solutions to overcome those difficulties in future action-project studies. Methodological Comparison of the Demonstration Studies Both demonstrations of the action-project differentiation procedure succeeded in generating findings that went beyond those presented in the original Young, Marshall, et al. (2005) study, and extended current knowledge about family career development projects in the context of goal orientation, and dyadic configuration, respectively. As such, the studies provide evidence that the procedure is both useful (i.e., it is capable of answering between-groups research questions within the action theory framework) and, to a large extent, viable (i.e., we were successful in implementing the required procedures, albeit with some minor modifications). The experience of implementing the procedure with real data also revealed the strengths and limitations of various parts of the procedures, and several possible refinements that should be considered for adoption in future comparative action theoretical research. However, there was also an important difference in the kinds of grouping constructs that were used in the two demonstration studies, resulting in several differences in the implementation of the method, and leading to different levels of success in identifying similarities and differences between the groups. Specifically, in Study 1, groups of cases were formed on the basis of the patterns of actions that were noticed at the within-case and initial cross-case level. In contrast, the grouping construct of interest in Study 2 was not identified as salient in the within-case and initial cross-case analyses. This had several consequences for the analysis process (see Table 4 for details), and resulted in far more differences being identified in Study 1 than in Study 2. Table 4: Methods-related Differences Between Study 1 and Study 2 Study 1 (goal orientation) Grouping construct is an intrinsic part of the action process Grouping construct identified as a potentially meaningful way to group participants after it emerged from initial analyses of the data Most dyads were found to fit into one group or the other, but with some variation in goodness of fit Analysis across cases, for each sub-group yielded arange o f common/dominant patterns of action Examining findings of groups in relation to each other yielded many similarities and differences Study 2 (dyadic gender configuration) Grouping construct is a structural feature of the people engaged in action Grouping construct identified.as a potentially meaningful way^to group participants from prior theory/research , A l l dyads were classifiable on the basis of the adolescent's gender, but the grouping construct was not evident in the within-case patterns of action Analysis across cases for each sub-group yielded revealed few common - distinguishing features for either group ' \ Examining findings of groups in relation to each other yielded few similarities or differences The reasons behind the differences in implementing the procedure across the two studies have important implications for the development of the action-project differentiation method and, therefore, must be examined more closely. For the first study, the grouping construct was chosen because it emerged from the collected data as a salient way of organising that data. This created a situation where there was some predisposition for finding distinct tendencies that were common within each group (if the researchers had not observed these commonalities, then goal orientation would not have suggested itself as a potential way of grouping participants for further analysis). No equivalent predisposition was present in the second study. In fact, it would have been possible to form the two dyadic configuration groups prior to any data analysis. The a priori nature of the grouping construct in Study 2 appears to have created a situation where the dyads were much more variable and different from each other within each of the two groups. Consequently, there were few aspects/ dimensions/ characteristics of the dyads' actions that could accurately be labelled as being 'common' across participants within each group. Successful implementation of the remainder of the between-groups analysis hinges on the presence of these common tendencies. This explains the paucity of the subsequent findings: i f only a few common aspects of action 'exist' for any given group, then it is only possible to compare the groups on those few aspects. The implication is that, in the second study, there was such a large range of people within each group that they cannot truly be characterized as a single 'kind.' The action-project differentiation was not designed for use in situations where participants are not similar to each other and, evidently, very few meaningful results can be generated when this is the case. The demonstration studies do not provide evidence that the differentiation procedure is inherently unsuitable for use with grouping constructs that are chosen primarily from theory, or based on structural characteristics of persons. It may still be possible to generate sets of cases that are similar to each other, yet distinct from others, using these kinds of constructs. What has been revealed, however, is that a priori groupings lack the predisposition for similarities among the cases within each group that are present in naturally emergent groupings. A potential remedy for this situation would be to find some alternative mechanism to decrease the variation among cases within each group, when an a priori structural feature is of interest to action theory researchers. Specific possibilities for increasing the amount of 'commonness' among cases within a group are discussed next, in the "specific Problems and Possible Solutions" section. A l l the same, i f a grouping construct is truly unimportant to peoples' joint action within a domain, then no amount of remediation will allow those sets of cases to become similar within each group, yet distinct from the other groups in a sample. Specific Problems and Possible Refinements Group Formation. In the goal-orientation example, the categorisation of dyads into the focused and diffuse groupings was accomplished by examining the emergent patterns within the data collected for each dyad. In the dyadic gender configuration study, the decision of whether a dyad should be placed in the mother-son or mother-daughter group simply involved identifying the gender of the adolescent, rather than attending to the full configuration of within-and cross-case findings. Consequently, the experience of forming the comparison groups differed in the two demonstrations. In the goal orientation study, the research team found that the ease with which dyads could be placed into one group or another varied according to the specific configuration of data for that dyad. Some projects were clearly focused or diffuse, while others contained elements of both, requiring more effort and discussion to classify. In one case, it was simply not possible to assign the dyad to either group, because the project equally exhibited characteristics of both a focused and a diffuse project. Despite these difficulties, the findings of the goal orientation study demonstrates that, when the grouping construct is process-oriented, and the categorisation of specific cases emerges out of the within-case and initial cross-case analyses, many patterns of similarity and difference can be generated and described. The dyadic gender configuration study also yielded some interesting conclusions regarding the viability of the action-project differentiation procedure, when groups are formed a priori, on the basis of structural features alone. We discovered that it is possible to identify some distinctions at the between-groups analysis phase, even when a grouping construct fails to emerge as important at the within- and initial cross-case case levels of analysis. However, certain anticipated differences (e.g., degree of communication, Grotevant & Cooper, 1988; frequency of joint activities, Noller & Callan, 1990) were not evident in the findings. As discussed in the "Methodological Comparison of the Demonstration Studies" section, this inability to find differences was probably due to a lack of commonalities within the mother-son and mother-daughter groups, and appears to be an inherent risk when researchers wish to form groups on the basis of a priori, structural features. In future, it is recommended that, even when groups are defined beforehand and cases are categorized according to some structural aspect of the participants' backgrounds, efforts be made to maximize the similarities among the members of each group, thereby increasing the likelihood that distinct tendencies and common aspects will emerge. These efforts could include purposive selection10 of participants for whom the grouping construct appears to be a salient issue, and/or explicitly asking questions related to the grouping construct during the collection and preliminary analysis of the data. For example, i f the dyadic gender configuration study was to be conducted again, the it may be useful for the interviewers to ask, "how, i f at all, did the fact that the two of you are the same gender influence your project," or, for the daughter specifically, "does your mother treat you differently that she treats your brothers, when it comes to preparing for your future?" In implementing these modifications, researchers must bear in mind two caveats. First, i f distinct groups fail to emerge out of the preliminary analyses even after such measures are put into place, then it is likely that the grouping construct in question is truly unimportant in understanding the action of those people, and continuing the between-groups analytical process is unlikely to generate many similarities or differences. Second, when asking questions related to the grouping construct, researchers must be willing to pursue and explore responses that suggest it is not the grouping construct that matters. To continue the 1 0 That is, choosing participants for inclusion in the sample because it is believed those individuals are cases who will best enhance understanding about a phenomenon of interest, rather than through random or representative sampling procedures (Bryman, 2004). earlier example, i f the daughter's response was, "well, my brothers are quite a bit older than me and getting jobs is more on their minds right now, so my mom has to talk with them more," then the interviewer should probe for the possibility that age differences, rather than dyadic gender configuration, is the critical issue in this family. Sequence of Analysis. In the first demonstration study, the sequence of (a) identifying commonly occurring themes and characteristics within each sub-group first, and then (b) examining those findings in relation to each other, was found to be an effective way of identifying similarities and differences between the groups. Fewer patterns of similarity and difference were generated in the second study, but this was due primarily to difficulties that were encountered with the group formation process, rather than the sequence of analysis itself. One problem that arose in both studies was that, although the procedures were sequential 'on paper,' the research team found it somewhat difficult to maintain a clear separation between the two stages in practice: we developed a tendency to anticipate the between-groups stage of the analytical process, when identifying what was common across the second of our two sub-groups. Anticipating the comparative analysis at that stage is problematic because it could result in a limited understanding of the data. Specifically, i f a research team were to use the themes and characteristics that were evident in the first sub-group as their exclusive guide for identifying patterns in the second sub-group, then characteristics of the data set that are prevalent in the second sub-group, but absent in the first sub-group, could be missed. To illustrate this problem in concrete terms, the projects of the focused-goal participants (the first group to be analysed in Study 1) tended to centre on their mutual relationship. Therefore, no "independence/autonomy project" category existed when we summarised the dominant patterns of the focused goal dyads. Had we utilized only the categories from this group when exploring the projects of the diffuse goal group, we would have missed the fact that the projects of many of the latter group of participants centered on issues of autonomy and independence. That is, the dyads in question would simply have been coded as not having relationship projects as a focus of their joint engagement. While true, this was clearly only part of the story. The solution that we employed was to re-examine the data for new categories of findings, after evaluating the data within the second group. If any new findings emerged in the second group, we reanalysed the data for the first group of dyads, to determine i f and how the new finding was present in that group. To continue the earlier example, after it became evident that independence/autonomy projects were dominant in the diffuse goal group, we had to re-examine the focused-goal dyads to determine which cases could be classified as having a significant independence/autonomy development component in their joint projects. The strategies of using multiple iterations of analysis, and evaluating existing categories in light of newly emergent findings are found in various forms of constructivist research (e.g., Giorgi, 1985; Glaser & Strauss, 1967), and are compatible with the idea of discussing and evaluating multiple interpretations of data. Given the tendency to leap prematurely to comparative analysis that we experienced, I recommend that that process of 're-examining the patterns of findings from earlier groups in light of information that emerges from later groups' be explicitly incorporated into future applications of the action-project differentiation procedure. Consensus. The process of using team discussion and consensus decision making to reach conclusions about patterns of similarity and difference between the groups was similar to my experience with these analytical strategies in other action-project studies: some patterns were self-evident and required little discussion to reach consensus, while others were open to multiple interpretations and required more intense discussion and scrutiny of the data to reach a decision. Furthermore, the team occasionally came across situations where various aspects of certain cases remained indeterminate, with no achievable resolution. For example, in three of the dyads, it was not possible to determine whether parental encouragement was a feature of their mutual relationship, due to inconsistent and/or conflicting pieces of information that were present in different portions of their data sets. In situations like that, the affected cases must remain unclassified for that particular dimension of their projects. Because the use of consensus to generate findings at the between-groups level was very similar to its use in generating within-case and cross-case findings, there is no need to discuss it further, other than to state that this aspect of the analysis is readily transferable from the existing action-project method to the new action-project differentiation procedure. Practical significance. I had anticipated that the standard of 'practical significance' would provide a clear, relatively simple heuristic for making judgements of similarity or difference. Although the standard is clear, its implementation was not always simple. That is, the research team encountered numerous situations where there was ambiguity as to whether the degree of difference between the groups was of sufficient magnitude to be considered practical. Some decisions about whether a difference was present or absent were obvious (e.g., some level of conflict was noted in a vast majority of dyads, across all the configurations of sub-groups). Other decisions were complicated by the presence of discrepant facets of the available information (e.g., abandonment of the original project in favour of another was present in only two of twenty dyads; however, the fact that abandonment of a project is even possible, and occurred only in the diffuse group, could be important in understanding the interplay between goal orientation and how projects progress over time). Finally, the relative magnitude of the differences were sometimes inherently ambiguous (e.g., is important that four mother-son dyads had projects directly focused on vocational issues, while only two mother-daughter dyads had such projects, in a total sample of twenty dyads?). As a result, it was my experience that applying the standard of practical significance requires substantial researcher judgement. Fortunately, the consensus decision-making process allows for different possible interpretations to be discussed and evaluated, and the ambiguity of what 'practical significance' means did not hinder the generation of results. Therefore, I recommend continued use of this standard, evaluated through both importance and the degree to which a characteristic or theme was present across the group, as a major component of making decisions about the ways in which the actions and projects of distinct groups are similar and different from each other. Researchers who are less familiar with qualitative analysis may find the amount of 'subjective' judgement required to implement this standard to be somewhat discomforting at first, but as they learn to trust in the consensus decision-making process, the use of practical significance will permit the generation of substantive findings. Charting. For the most part, the charts that were created during the analysis of the two studies (see Appendices 2 and 5) proved to be useful in facilitating the comparative analysis process. That is, using the chart as a reference point to locate information in the data set allowed for rapid, convenient juxtaposition of the common themes that emerged for the two groups in each example. The charts also allowed the research team to easily identify which cases reflected a particular dimension of commonality across the group, as well as any cases that were counter-examples of that dimension. This, in turn, facilitated the process of referring back to the original cases to ensure that the conclusions did, indeed, reflect the lived experience of the cases, and to ground the findings in concrete examples and participant statements. Most aspects of the data set were relatively easy to capture and summarise in a tabular format. For certain dimensions, however, it was necessary to deviate from the practice of listing examples and counter examples. Specifically, for "degree of progress made" and "perceived barriers," we found it more informative to list all the different possible options under one heading, rather than attempting to develop separate headings for each possible option. For example, for the issue of how the projects progressed over time, rather than using the dichotomous coding scheme of "made progress" versus "Did not make progress," we found that the range of possible progression was better captured by using "substantial progress," "less progress, but still increasing over time," "mixed progress (periods of progression alternating with periods of regression)," and "original project completely abandoned in favour of an alternative project." Moreover, one dimension of the data set was not amenable to being charted at all. Many activities that participants reported in the monitoring period were highly individualised, and difficult to aggregate with the activities of other dyads (e.g., one mother organised a "destination imagination" creative play club for her son and his friends; another mother wrote about racing on all-terrain vehicles with her son). This resulted in a situation where attempting to summarise the findings on a chart needlessly complicated the analysis process. Therefore, for this dimension of the data, the charting was abandoned in favour of directly examining the monitoring logs to summarize the general kinds of activities in which parents and adolescents jointly engaged, and how often they occurred. This was, however, the only aspect of the data set where charting impeded, rather than facilitated, the process of analysis. Although the charts were a useful tool for conducting the differentiation analysis, they were also a source of some complications. Specifically, the presence of the chart increased the temptation to use simple counting rather than the standard of practical significance as the guideline for evaluating similarity/difference. This problem was most noticeable for the team-member with little prior experience in qualitative analysis, who often found herself referring exclusively to the number of cases, during discussions of possible interpretations of the data. However, it also influenced the thinking and writing of the more qualitatively experienced initial analyst as well, to the point that the third team-member found it necessary to question whether practical significance had been attended to, in the second demonstration presented in Chapter 4. The relative number of cases in which a particular characteristic or theme is present within the group is an important part of evaluating the practical significance of a difference, but it is not the sole criterion. Therefore, as decisions were made about whether the two groups were similar or different in any particular way, it was necessary to review our initial conclusions to ensure that decisions were not made solely on the basis of how many dyads were involved; that is-, we needed to explicitly remind ourselves that it is the content of the cases within each of the cells on the chart that is important for data analysis, rather than the numbers of cases alone. This complication causes me to feel somewhat ambivalent about retaining the charting tool within the action-project differentiation procedure. Charts certainly facilitate the process of juxtaposing and examining together the pattern of findings that are evident in different groups of participant. At the same time, I am concerned that the chart will become the data, rather than being used as intended; that is, as a tool to facilitate access to the actual data (i.e., the information contained in the data records of each dyad). One potential solution is to emphasise strongly that the charts are merely a reference tool, and that consensus decision-making and practical significance are the criteria for evaluating similarity and differences. In combination with examples illustrating the proper (and improper) use of charting, this solution may be sufficient to ensure that the process of analysis will occur as intended, at least for researchers who are open to constructivist and action theoretical approaches to dealing with qualitative data. In summary, I have demonstrated that the procedures developed as part of this dissertation are largely successful at achieving their purposes of (a) forming distinct sub-sets of participants, and (b) identifying and describing the patterns of similarity and. difference in the actions and projects of those distinct groups. The demonstration studies also revealed several unanticipated problems, which required minor modifications of some aspects of the procedure. Specifically, I believe that the existing procedure will be strengthened by (a) ensuring that even structural grouping constructs will emerge in the preliminary within- and cross-case data analyses; (b) explicit acknowledgement that several iterations of re-examining data within each group, in light of the findings that were derived from subsequent groups, may be necessary to obtain an adequate understanding of the patterns in the data; and (c) emphasising that charts function primarily as a reference tool (rather than being the data themselves). Fortunately, these are minor modifications that do not violate the underlying ontological and epistemological assumptions of action theory, and can be readily incorporated into the method that I have developed. CHAPTER 6: DISCUSSION The previous two chapters extend my initial answer to the second research question of this dissertation, "What specific procedures can be used to generate findings about between-groups comparisons within an action theory research framework?" The demonstration studies confirmed that my extension for the existing action-project method is capable of generating findings that cannot be obtained without some form of between-groups analysis and, does so in a manner that is consistent with the underlying tenets of Young and colleagues' action theory. Although the previously described refinements to the action-project differentiation procedure need to be further examined through their use in future research, the action-project differentiation procedure appears to provide a workable solution to the problem of between-groups comparisons in action theory. In this final chapter, I provide a summary of what I have accomplished, describe the boundaries and limitations of the procedure, and conclude with a discussion of how the action-project differentiation procedure relates to other methods of conducting qualitative research. Contribution to Knowledge In this dissertation, I addressed the problem of generating comparative findings about the actions and projects of participants, within an action-project study, when there is some specific contextual factor that distinguishes different kinds of participants from each other. Specifically, I have answered the questions of whether this kind of between-groups comparison is compatible with the assumptions that underlie the action theory paradigm, and what specific procedures can be used to generate such findings. In doing so, I first described the compatibility of conducting between-groups comparative analyses with a number of epistemological, ontological and other paradigmatic assumptions of Young and colleagues' (e.g., Collin & Young, 1992; Valach, 1988; Valach et al., 2002) action theory. After establishing that some forms of between-groups analysis (specifically, comparisons that (a) attend to all the perspectives of action that are proposed in the theory; (b) include the specific situation of each case, rather than focusing on the grouping construct alone; and (c) are descriptive of specific samples situated in their contexts rather than being inferential in nature) are, indeed, permissible within the action theory framework, I created a set of specific, concrete procedures for conducting this kind of analysis: the action project differentiation procedure. The two empirical demonstrations of the differentiation procedure demonstrated that the procedure is capable of generating useful findings and addressing questions that the existing action-project method is unsuited to answering. However, the demonstration studies also suggested that it may be more difficult to apply the procedure to situations where the groups are formed via a priori structural characteristics, when these characteristics do not also emerge from the within-case findings. Finally, the process of conducting the demonstration studies yielded a number of adjustments and refinements to the procedure, which should be incorporated into future applications of this comparative method of analysis for action theory. By engaging in this process, I have made two main contributions to knowledge in the field of 'measurement, evaluation and research methodology' or, more specifically, qualitative research methodology. First, I demonstrated that some types of between-groups analyses are compatible with the theoretical tenets and philosophical assumptions of action theory. Prior to this point, action theory researchers had begun to become interested in between-groups questions due to the patterns of data that were emerging in their studies, but lacked clarity as to whether engaging in such analyses would be consistent with the qualitative, constructivist nature of action theory. Second, I have developed a set of concrete procedures to allow action theory researchers to engage in systematic comparative analysis of the patterns of action that are present in distinct sets of persons, and demonstrated that my action-project differentiation method is capable of generating useful findings. Although it requires refinement through further usage in actual research practice, my method is definite extension of the existing action-project method and, for the first time, provides action theory researchers with a concrete way to address between-groups questions in a clear and transparent manner. Boundaries and Limitations of the Procedure One of the boundaries on the utility of the action project differentiation procedure was introduced in Chapter 5, during the comparison of the two demonstration studies. For this kind of between-groups analysis to function, it is necessary to have groups that are composed of cases that truly are similar to each other, in terms of their patterns of action. If the phenomenon of interest is one where each dyad's experiences are unique, or where there are not common tendencies in actions across each group, then it will not be possible to generate findings that characterise that group. Consequently, the process of juxtaposing the findings of each group and examining them in relation to each other will fail. Although discovering that people's actions within a domain of interest tends to be unique for each case is, itself, an important finding, the fact remains that the differentiation procedure will not function in this situation; between-groups analysis requires the existence of 'true' groups within a sample. A second limitation to the procedure is imposed by the function the comparative analysis within action theory. The purpose of the action-project differentiation procedure is to explore and develop a richer understanding of human action in context, highlighting a particular aspect of the context (what I have labelled the 'grouping construct') for closer scrutiny during the exploration. The procedure cannot be used to draw the conclusion that the grouping construct is directly responsible for the differences and similarities that emerge. To use a statistical metaphor, this between-groups analysis procedure is more akin to a correlation than a regression. Therefore, before any of the findings from an action-project differentiation study can be translated into concrete recommendations for changes to practice (e.g., that counsellor would be able to improve parent-adolescent communication within the vocational domain by somehow promoting the development of a more focused project), it would first be necessary to confirm that the direction of influence between the grouping construct and the between-groups differences that were found flows in a particular direction. An inferential statistical framework, where the purpose is to isolate sources of variation and determine the contribution of certain variables to the systematic variation found in other variables, would be better suited to generating those kinds of conclusions. In this situation, the descriptions generated in an action-project study serve primarily to build the theory that will eventually guide the development of specific research questions, which could then be tested using quantitative methods. By itself, the differentiation procedure cannot provide all the answers that are necessary to develop practice strategies in career counselling, or any other field of professional practice. Additionally, the quality of the findings that are generated in any given action-project differentiation study are limited by the quality of the team of persons doing the analysis. Making decisions about what dimensions and aspects are common within a particular group, or how much of a distinction between groups constitutes a practically significance difference requires team members who can propose and actively discuss alternative interpretations. As Hil l et al (1997a, 1997b) recognised in their description of their consensual qualitative method, when the analysis relies upon team-based discussion, the integrity of the decision-making process can be negatively impacted by (a) a high level of conflict between team members, or (b) a high level of acquiescence and deference by some team members towards other team members. Therefore, the utility of the action-project differentiation procedure is bounded by the ability of the researchers to engage in a truly team-based process of analysis. The composition of the research team also imposes another limit on the usefulness of this method. The power of the analysis is, in a sense, dependent on the ability of the team members to comprehend the data, notice meaningful patterns, and relate the findings form the different groups to each other. Each of these tasks requires a certain level of perception, background knowledge, and willingness to thoughtfully consider alternative interpretations of the data. In short, the kinds of skills that are primarily required to create a good discussion section in a quantitative study, are necessary for both the discussion of findings and the actual generation of those findings in an action-project differentiation study. If one or more team members are deficient in these skills, then this will negatively affect the team's ability to identify and describe the similarities and differences in people's actions. The issues raised in the preceding two paragraphs reflect the strong reliance of the procedure on researcher judgement, and the potential risks of doing so. However, this risk is mitigated by the consensual discussion and analysis process, and a willingness to consider multiple possible understandings. Additionally, Hoskins (2001) has made several suggestions for promoting the fidelity of qualitative analysis, suggestions that can readily be adopted by researchers engaged in action-project differentiation research. These suggestions include: (a) being intentional in examining the congruence between research practice, assumptions and phenomenon of interest; (b) attending to, rather than filter out, the various simultaneous perceptions generated by their various senses and intuition; and (c) becoming comfortable with groundlessness and letting go of the need to exert tight control over the research process. In is evident that the differentiation procedure will never lead to objective findings of the kind that are assumed to exist in quantitative research. Fortunately, there is no need to strive towards objectivity, because action theory accepts that multiple legitimate interpretations can be constructed from any given data set. Instead, the ideal is to achieve authenticity and fidelity to the data as one engaged in the process of making decisions in research. Finally, it needs to be recognised that, to date, the action-project differentiation method has only been applied to research situations where the cases in question are comprised of pairs of people. As noted in Chapter 2, any set of individuals working together towards a common purpose qualifies as a viable case for analysis in action theory. However, the data collection and analysis procedures presented in this dissertation assume that cases are composed of dyads of persons. Although the principles should remain the same, action theory researchers who wish to study the actions of distinct groups of triads (or larger units) engaged in common projects may have difficulty adapting the concrete procedures presented here for their studies. It may, for example, be more difficult to identify common tendencies across groups of these kinds of cases, because the larger number of individuals within each case may result in greater variation in action between the cases in any given group. Until the procedures have been adapted and tested for use in these kinds of situations, it is prudent to limit claims about the usefulness of the action project differentiation procedure to studies examining the joint actions of pairs of people. Relationship to Other Qualitative Methods Constructivist methods: The procedures that I have developed are related to, but somewhat distinct from, solutions to the problem of comparison found in other forms of social constructionism. One of the primary ways in which the action-project differentiation procedure differs from grounded theory is in the underlying purpose of the comparative analysis: we are attempting to identify and describe patterns in people's joint actions and projects within a domain, rather than to develop a complete, data-guided theory about that domain of functioning. Due to the nature of the research questions asked by grounded theory researchers, data is typically examined across the entire set of participants in such a study (after all, the theory should be able to encompass all of the cases). In contrast, a researcher who adopts the action-project differentiation procedure does not search for the common core that represents the experience of all cases, but instead, proposes that the action of different kinds of persons may be distinct in important ways, necessitating a comparison of distinct groups. Despite these fundamental differences, there are a number of commonalities in the way that these two methods arrive at their findings and conclusions. Grounded theory and the action-project differentiation procedure both assume that it is best to allow concepts and definitions of constructs to emerge from analysis of the within-case data, rather than through prior definition. Also, my recommendation to engage in multiple iterations of examining the data at the 'within-group' phase of analysis, and modifying the descriptions of the themes and characteristics that emerged in previous groups in light of the findings obtained in subsequent groups, is reminiscent of the 'constant comparison' analytical method that is a hallmark of grounded theory (see Chapter 3 for an overview of constant comparison). Finally, the underlying principle of relying on researcher judgement to make decisions and generate findings from qualitative, textual data is another way in which these two methods are closely related. The relationship between the action-project differentiation procedure and phenomenology is similar to the relationship between the differentiation procedure and grounded theory, in the sense that the fundamental purpose of the two methods are very different, yet share some common elements within their analytical strategies. Phenomenology is designed to understand the meanings associated with various experiences, across different individuals (Giorgi & Giorgi, 2003), while the action-project differentiation procedure seeks to describe similarities and differences in the action of different individual. Therefore, meaning is the sole focus of the analysis in phenomenology, while action (encompassing manifest behaviour, internal processes, and socially constructed meaning) is focus of analysis in the differentiation procedure. More importantly, in phenomenology, data are examined across cases to develop an understanding of the common core of the phenomenon in question. In other words, a phenomenologist would be interested primarily in the similarities that are present in the findings of distinct kinds of persons. In contrast, the action-project differentiation procedure was designed to describe both the similarities and the differences that are present in the findings of distinct kinds of persons. Despite these substantial contrasts, certain elements of the analytical mechanisms found in phenomenology and the action-project differentiation procedure are quite closely related to each other. Specifically, the two methods (a) share a preference for allowing variables, constructs and themes to emerge naturally from the data, rather than imposing an a priori definition, and (b) acknowledge the subjectivity of the analysis process, relying primarily on researcher judgement to draw conclusions from textual data. It is somewhat more difficult to relate the action-project differentiation procedure to narrative approaches to comparative analysis, because there are a range of theoretical positions and methodological practices within the scope of what is called narrative research. For example, some narrative researchers would eschew the notion of grouping different people's stories for subsequent comparison, assuming that individuals' stories are unique and incommensurable. However, other narrative researchers accept the notion of grouping the stories of different individuals together, and conducting between-groups comparisons of the stories of different kinds of persons (e.g., Gergen & Gergen, 1987). What distinguishes the action-project differentiation method from even these kinds of narrative studies is the fact that this procedure is designed to describe joint actions around a phenomenon, rather than the way in which people describe their experiences or history. To fulfill this purpose, data collection in an action-project differentiation study involves attending to manifest behaviour, internal processes and social meaning, rather than social meaning alone (as reflected in the stories that are constructed by narrative research participants). In summary, it would appear that a complex pattern of relationship exists between the action-project differentiation method and grounded theory, phenomenology and narrative research. As described, in terms of the general purpose and the product of analysis, the action-project differentiation method is quite distinct. In contrast, there are a number of commonalities that exist between the methods, in terms of the specific mechanisms of analysis. This pattern of relationship should not, however, be terribly surprising. The action-project differentiation procedure reflects the action theory paradigm of research, which differs somewhat from other forms of constructivism in terms of the phenomenon of inquiry and the assumptions that are held (see Table 1 in Chapter 2 for a summary of these differences). At the same time, the solutions that grounded theory, phenomenology, and narrative research have developed to deal with the problem of comparison were examined as part of the process of building my method, so it is natural that the mechanisms for analysis employed in these methods are related to the ones found in the action-project differentiation procedure. Broader paradigm issues: Adoption of an explicit set of procedures for conducting between-groups analysis also has broader implications for the relationship between action theory and other paradigms within the spectrum of qualitative research. As a paradigm for qualitative research, action theory generally falls within social constructionism. However, it falls at the more highly data-oriented end of the range of constructivist methods, and even contains some qualities that are more closely aligned with post-positivism. As such, it is different from emancipatory and post-structural approaches to qualitative research. Although action theory research was moving in the direction of between-groups analysis prior to the development of the action-project differentiation method (e.g., Arato-Bollivar et al., 2002; Young et al., 2003; Young, Logan, et al., 2005), the adoption of my procedure by action theorists will emphasise the differences between action theory and the emancipatory / post-structural paradigms. For example, researchers using the action-project differentiation procedure are much less inclined to take an active stance regarding the phenomenon being studied, prior to data collection. Although the framing of research questions is influenced by the existing literature, analysis should be driven primarily by the data rather than any specific position, such as the need to fight against historical gender oppression that is present in many feminist research paradigms. Also, the procedure is designed to describe phenomena as they are observed, rather than to disrupt dominant social discourses or promote positive change within the participants. Although most action theorists would be pleased i f these outcomes were to occur, they are not measures by which the success of an action theory study is based. That is, promoting a particular perspective and inciting participants or the broader society to change are of secondary importance; what is of primary importance is describing and generating findings about the phenomenon of interest. Action theory's less advocacy-oriented approach to inquiry would probably be viewed by many emancipatory and post-structural qualitative researchers as somewhat regressive, to the point of being indistinguishable from the position of traditional, quantitative research. Parenthetically, it is perhaps necessary to explain that action-theory is not anti-transformation, in the sense of many quantitative approaches that seek to minimize the influence of the research procedures on the outcomes being observed. In fact, the act of participating in an action-project differentiation study can induce important changes in the lives of participants, particularly during the self-confrontation interviews, and self-monitoring period. However, unlike emancipatory and post-structural approaches, transforming the lives of the participants is not the underlying purpose for conducting research within an action-theory paradigm, nor is it a criterion by which to gauge the success of an action-project study. Similarly, although we describe ourselves as adopting a collaborative stance in action theory research (citing such steps as the self-confrontation procedure and member-checking during the second interview), there is a substantial amount of analysis and interpretation that occurs without participants' input in the action-project differentiation procedure. In fact, once the analysis progresses past the within-case level, the interpretation of data and generation of findings is completed exclusively by the researchers. This echoes existing practices found in the action-project method, but is contrary to the way that inquiry is conducted by many emancipatory and post-structuralist qualitative researchers (Lincoln & Guba, 2000). Finally, as I described in Chapter 3, the very notion of between-groups comparison assumes a high level of commensurability of experiences across individuals. This reflects an ontology that is, once again, closer to the position of post-positivism than some other forms of qualitative research. Although I have designed the action-project differentiation procedure to retain the individual contexts of the cases and included safeguards to ensure that the analysis is not reduced to pure counting, the mere act of asking a between-groups research question is foreign to many post-structuralists and, to a lesser degree, emancipatory researchers. On an even broader level, in discussing the present and future state of qualitative inquiry in the Handbook of Qualitative Research, Lincoln and Denzin (2000), and Gergen and Gergen (2000) identified a number of issues that they perceived as important ones in the current, "seventh moment" of qualitative research. These issues, which they perceive as needing to be addressed in the immediate future, include (a) formulating and refining new, relativist perspectives of science to replace the foundationalism that had previously been dominant in social research; (b) negotiating questions of voice, authority, and the struggle with how researchers can legitimately represent others; (c) grappling with moral dimensions and inclusion of the sacred, into the processes and discourses of social inquiry; (d) coping with the impact of the "technorevolution" on how research is defined and conceptualized; (e) establishing (or at least discussing) the appropriate place of the political within the endeavour of research; and (f) re-visioning and retraining the nature of validity in light of non-postpositivist approaches to knowledge generation. However, even in the few years that have passed since Lincoln and Denzin, and Gergen and Gergen wrote their summary articles, circumstances within and outside of the field have shifted the place of qualitative research, at least in relation to the specific discipline of psychology. In the past few years, there has been an increased recognition of the legitimacy and value of qualitative methods as ways of generating knowledge in psychology. For example, close to twenty years after their special issue on quantitative research methods, the American Psychological Association's Journal of Counseling Psychology is finally publishing a special issue on qualitative methods in April 2005 (Beth Haverkamp, personal communication, December 20, 2004). Also, in a special issue on qualitative research in Canadian psychology, O'Neill uses the metaphor of earthquake and "tectonic change" to explore the idea that qualitative research is on the way to becoming " i f not the dominant paradigm in psychology, at least closer to sharing dominance with hypothesis-testing quantitative research" (2002, p. 190)." Finally, it has been my personal observation that an increasing number of psychology-related, undergraduate research methods/design textbooks are incorporating information on qualitative research into their chapters (e.g., Coolican, 2004; Davis & Smith, 2005; Langdridge, 2004; Salkind, 2003), albeit with varying degrees of integration. A second circumstance that has shifted the place of qualitative research is the emergence of mixed methods approaches the social research. Mixed methods research has been described as a "third research movement" that allows social researchers to move beyond qualitative and quantitative disputes, using a pragmatic philosophical stance to combine techniques from both 1 1 It must be acknowledged, however, that many of the other articles within that special issue describe continued resistance and subtle prejudice towards qualitative research in Canadian psychological academia. (Johnson & Onwuegbuzie, 2004). The emergence of mixed methods research has rekindled the examination of qualitative and quantitative research together, despite the fact that this family of research methods remains in its formative stages, with lively debate currently occurring among mixed methods researchers, about a variety of basic definitional issues12 (Teddlie & Tashakkori, 2003). However, unlike earlier manifestations of the 'quantitative-qualitative debate,' the purpose of current discussions appears to finding commonalities and ways that one can strengthen the other, rather than demarcating the boundaries of what is qualitative and quantitative, or highlighting the inadequacies of one kind of research or the other. Indeed, Creswell (2003) has gone so far as to claim that the designs of current social research studies can no longer be classified simply as 'quantitative' or 'qualitative,' but must be examined in terms of how quantitative or qualitative they are; with an implication that there is often some element of both in any given design. In this climate of increased acceptance of qualitative research within psychology and decreased separation between qualitative and quantitative research designs, the development of the action-project differentiation procedure raises a new question to add to Lincoln and Denzin's (2000), and Gergen and Gergen's (2000) lists of what must be addressed in the seventh moment of qualitative research: "what is the full scope of research questions that can be answered by each qualitative method?" Specifically, i f between-groups comparisons can be successfully conducted within the action theory paradigm, then it may also be possible to adapt other qualitative methods to answer between-groups questions, while remaining consistent with the underlying paradigmatic assumptions of the method in question. The notion that between-groups comparative analysis is solely the provenance of quantitative research is now open to 1 2 Teddlie and Tashakkori (2003, p. 4) list six, "(1) the nomenclature and basic definitions used in mixed methods research; (2) the utility of mixed methods research (why we do it); (3) the paradigmatic foundations for mixed methods research; (4) design issues in mixed methods research; (5) issues in drawing inferences in mixed methods research; and (6) the logistics of conducting mixed methods research."-question, and qualitative research methodologists may do well to explore how such analyses could be conducted within other constructivist methods, or possibly even post-structuralist/ emancipatory approaches to knowledge generation. The Next Step Ultimately, the final standard for the successful development of any method is its acceptance within the scientific community. Waiting for the action-project differentiation procedure to be published in a major journal, or for other action theorists to complete studies using this method is, unfortunately, beyond the scope of this dissertation. However, some initial indicators that this method will be accepted within the scientific community have already occurred. A number of action theory researchers have expressed an interest in using the method in their own future studies. Furthermore, the content of the demonstration studies and descriptions of the differentiation procedure, have been presented at a number of regional, national and international conferences (i.e., the 2004 conferences of the Canadian Psychological Association, the European Association for Research on Adolescence, and the International Network on Personal Meaning; also the 2003 and 2004 Conry Conferences on Measurement, Evaluation, and Research Methodology), where the material has generally been well received. For example, when the goal orientation study was presented at the 2004 conference of the Canadian Psychological Association, it won the 'best poster' prize for the counselling psychology section. I acknowledge, however, that the development of research methods, especially qualitative methods, is a recursive process. Even when I applied the procedure in Chapter 4,1 found it necessary to implement number of modifications to the original processes proposed in Chapter 3. I fully anticipate that, as it is applied to new research questions and different data sets, the action-project differentiation procedure will continue to be altered and refined. Moreover, the adaptation of the procedures described here, to deal with cases that are composed of three or more persons working towards a common goal, has not even begun to occur. Nevertheless, sufficient work has been completed to allow other researchers to address questions of between-groups similarity and difference in their own action-project studies, in a manner that is consistent with the principles of action theory. As such, this dissertation represents a significant advancement of the action theory framework for social research. Future development of the action-project differentiation procedure now requires that the method be presented to the scientific community, so that its utility and viability can be independently assessed. Furthermore, the refinements and modifications suggested in Chapter 5 will benefit from further examination as the procedure is applied in future studies. These are the tasks that I will seek to accomplish in the immediate future. The creation of an explicit set of procedures to conduct between-groups comparative analyses in action theory also provides action theory researchers with the opportunity to ask new kinds of questions as they design their future studies (e.g., the joint career development projects of focused and diffuse romantic partners), and provides a mechanism to revisit existing data sets to more systematically explore actions and projects in the context of specific grouping variables (e.g., re-examining existing action-project data sets in light of the goal orientation of the participants). Formulating and empirically examining these new questions is another important task for the future, one that may eventually promote a more wide-spread adoption of Young and colleagues' contextual action theory within the psychological sciences. REFERENCES Arato-Bollivar, J., Young, R.A., Marshall, S.K., Hayoun, R., Marshall, E.G., Turkel, Ff., & Domene, J.F. (2002, May). Conversations of "disadvantaged"parents and adolescents as Family Career Development Projects. Paper presented at the annual meeting of the Canadian Psychological Association, Vancouver, BC. Ayers, L. , Kavanaugh, K. , & Knafl, K . A . (2003). Within-case and across-case approaches to qualitative data analysis. Qualitative Health Research, 13, 871-883. Bader, D., & Young, R.A. (2002). The reasoning of parents and children in conversations about marital transitions. In L. Valach, R.A. Young & M.J. Lynam (Eds.), Action theory: A primer for applied research in the social sciences (pp. 139-152). Westport, CT: Praeger. Bank, B.J. (1995). Gendered accounts: Undergraduates explain why they seek their bachelor's degree. Sex Roles, 32, 527-544. Birk, J.M., & Blimline, C.A. (1984). Parents as career development facilitators: An untapped resource for the counselor. The School Counselor, 37,310-317. Blustein, D.L, Walbridge, M . M . , Friedlander, M.L . , & Palladino, D.E. (1991). Contributions of psychological separation and parental attachment to the career development process. Journal of Counseling Psychology, 38, 39-50. Bregman, G., & Killen, M . (1999). Adolescents' and young adults' reasoning about career choice and the role of parental influence. Journal of Research on Adolescence, 9, 253-275. Bryman, A. (2004). Social research methods (2nd ed.). New York: Oxford University Press. Bumpus, M.F., Crouter, A.C. , & McHale, S.M. (2001). Parental autonomy granting during adolescence: exploring gender differences in context. Developmental Psychology, 37, 163-173. Charmaz, K. (2003). Grounded theory. In J.A. Smith (Ed.), Qualitative psychology: A practical guide to research methods (pp.81-110). Thousand Oaks, CA: Sage. Collin, A . & Young, R.A. (1992). Constructing career through narrative and context: An interpretive perspective. In R.A. Young and A Collin (Eds.), Interpreting career: hermeneutical studies of lives in context (pp. 1-12). Westport, CT: Praeger. Coolican, H. (2004). Research methods and statistics in psychology (4th ed.). London: Hodder & Stoughton. Cooper, C.R., & Grotevant, H.D. (1987). Gender issues in the interface of family experience and adolescents' friendship and dating identity. Journal of Youth and Adolescence, 16, 247-264. Creswell, J.W. (2003). Research design: Qualitative, quantitative, and mixed-methods (2nd ed.). Thousand Oaks, C A : Sage. Cupchik, G. (2001). Constructivist realism: An ontology that encompasses positivist and constuctivist approaches to the social sciences. Retrieved August 8, 2003 from Forum Qualitative Sozialforschung /Forum: Qualitative Social Research, 2 (1): www.qualitative-research.net/fqs-texte/1-01/1-01 cupchik-e.htm. Davis, S.F., & Smith, R.A. (2005). An introduction to statistics and research methods: Becoming a psychological detective. Upper Saddle River, NJ: Pearson Delanty, G., & Strydon, P., (2003). Part 6: New Directions and challenges. In G. Delanty & P. Strydon (Eds.), Philosophies of social science: The classic and contemporary readings (pp. 365-383). Philadelphia, PA: Open University Press. Denzin, N .K. (1989). Interpretive interactionism. Newbury Park, CA: Sage. Dick, T.S., & Rallis, S.F. (1991). Factors and influences on high school students' career choices. Journal for Research in Mathematics Education, 22, 281-292. Domene, J.F., Young, R.A., Marshall, S. K. , Arato-Bollivar, J., Hayoun, R., & Marshall, E.G. (2003, May). Aspects of possible selves revealed in "Family Career Development Projects. " Poster session presented at the annual conference of the Western Psychological Association, Vancouver, BC. Eccles, J.S. (1994). Understanding women's educational and occupational choices: Applying the Eccles et al. model of achievement-related choices. Psychology of Women Quarterly, 18, 585-609. Elliott, R., Fischer, C.T., & Rennie, D.L. (1999). Evolving guidelines for publication of qualitative research studies in psychology and related fields. British Journal of Clinical Psychology, 35,215-229. Ellis-Schwabe, M . , & Thornburg, H.D. (1986). Conflict areas between parents and their adolescents. Journal of Psychology, 120, 59-68. Farmer, H.S., Wardrop, J.L., Anderson, M.Z. , & Risinger, R. (1995). Women's career choices: Focus on science, math, and technology careers. Journal of Counseling Psychology, 42, 155-170. Gergen, K.J . , & Gergen, M . M . (1987). Narratives of relationship. In R. Burnett, & P. McGhee (Eds.),. Accounting for relationships: Explanation, representation and knowledge, (pp. 267-288).. New York: Methuen Gergen, M . M . , & Gergen, K.J . (2000). Qualitative inquiry: Tensions and transformations. In K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 1025-1046). Thousand Oaks, C A : Sage. Giorgi, A . (1985). Phenomenological and psychological research. Pittsburgh, PA: Duquesne University Press. Giorgi, A . , & Giorgi, B. (2003). Phenomenology. In J.A. Smith (Ed.), Qualitative psychology: A practical guide to research methods (pp.25-50). Thousand Oaks, C A : Sage. Glaser, B.G. (1978). Theoretical sensitivity. M i l l Valley, CA: Sociology Press. Glaser, B.G. , & Strauss, A . L . (1967). The discovery of grounded theory. Chicago: Aldine. Gottfredson, L.S. (2002). Gottfredson's theory of circumscription, compromise, and self-creation. In D. Brown (Ed.), Career choice and development (4th ed., pp. 85-148). San Francisco: Wiley. Grotevant, H. D., & Cooper, C. R. (1985). Patterns of interaction in family relationships and the development of identity exploration in adolescence. Child Development, 56, 415-428. Grotevant, H. D., & Cooper, C. R. (1988). The role of family experience in career exploration: A life-span perspective. In P.B. Baltes, D.L. Featherman, & R . M . Lemer (Eds.), Lifespan development and behaviour (pp. 231-258). London: Earlbaum. Harris, I.D., & Howard, K.I. (1984). Parental criticism and the adolescent experience. Journal of Youth & Adolescence, 13, 113-121. Hi l l , C.E., Thomson, B.J., & Williams, E.N. (1997a). A guide to conducting consensual qualitative research. The Counseling Psychologist, 25, 517-572. Hi l l , C.E., Williams, E.N., & Thomson, B.J. (1997b). A rejoinder to Stile's, Hoshmand's, and Tinsley's comments about " A guide to conducting consensual qualitative research." Counseling Psychologist, 25, 606-614. Hi l l , J.P., & Holmbeck, G.N. (1987). Disagreements about rules in families with seventh-grade girls and boys. Journal of Youth and Adolescence, 16, 221-246. Hoskins, M . L . (2001). True grit and the new frontier: Cultivating new ground for psychological research. Qualitative Inquiry, 7, 659-675. Huberman, A . M . , & Miles, M.B. (1994). Data management and analysis methods. In N . K. Denzin & Y . S. Lincoln (Eds.), Handbook of qualitative research (1st ed., pp. 428-444). Thousand Oaks, C A : Sage. Hugentobler, M.K. , Israel, B.A. , & Schurman, S.J. (1992). An action research approach to workplace health: Integrating methods. Health Education Quarterly, 19, 55-76. Johnson, R.B., & Onwuebuzie, A.J . (2004). Mixed methods research: A research paradigm whose time has come. Educational Researcher, 33 (7), 14-26. Jones, C. (1996). Careers in project networks: The case of the film industry. In B . M . Arthur & D . M . Rousseau (Eds.), The boundaryless career: A new employment principle for a new organisational era (pp. 58-75). New York: Oxford University Press. Kirk, R.E. (1996). Practical significance: A concept whose time has come. Educational and Psychological Measurement, 56, 746-759. Kotrlik, J.W., & Harrison, B.C. (1989). Career decision patterns of high school seniors in Louisiana. Journal1 of Vocational Education, 14, 167-178. Krefting, L. (1991). Rigor in qualitative research: The assessment of trustworthiness. American Journal of Occupational Therapy, 45, 214-222. Kvale, S. (1996). Interviews: An introduction to qualitative research interviewing. Thousand Oaks, C A : Sage. Langdridge, D. (2004). Introduction to research methods and data analysis in psychology. Harlow, Essex, England: Pearson. Larson, R.W. (2000). Toward a positive psychology of youth development. American Psychologist, 55, 170-183. Lemon, N . , & Taylor, H. (1997). Caring in casualty: The phenomenology of nursing care. In N . Hayes (Ed.), Doing qualitative analysis in psychology (pp. 227-243). Hove, East Sussex, U K : Psychology Press. Lent, R.W., Brown, S.D., & Hackett, G. (1994). Towards a unifying social cognitive theory of career and academic interest, choice, and performance. Journal of Vocational Behavior, 45, 79-122. Lincoln, Y.S. , & Denzin, N .K. (2000). The seventh moment: Out of the past. In K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 1047-1065). Thousand Oaks, CA: Sage. Lincoln, Y.S. , & Guba, E.A. (2000). Paradigmatic controversies, contradictions, and emerging influences. In K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 163-189). Thousand Oaks, C A : Sage. Lincoln, Y.S. , & Guba, E.A. (1985). Naturalistic inquiry. Beverly Hills, CA: Sage. Little, B. (1983). Personal projects: A rationale and methods for investigation. Environment and Behavior, 15, 273-309. Little, B. (1989). Personal project analysis: Trivial pursuits, magnificent obsessions, and the search for coherence. In D. M . Buss & N . Cantor (Eds.), Personality psychology: Recent trends and emerging directions (pp. 15-31). New York: Springer. Little, B. (1999). Personal projects and social ecology: Themes and variations across the life span. In J. Brandtstadter & R. M . Lerner (Eds.), Action and self-development: Theory and research through the life span (pp. 197-221). Thousand Oaks, CA: Sage. Lynam, M.J, & Young, R.A. (2002). Empowerment, praxis and action theory in health-promotion research. In L. Valach, R.A. Young & M.J. Lynam (Eds.), Action theory: A primer for applied research in the social sciences (pp. 123-138). Westport, CT: Praeger. Mannheim, B., & Seger, T. (1993). Mothers' occupational characteristics, family position, and sex role orientation as related to adolescents' work values. Youth & Society, 24, 276-298. Manning, P.K. & Cullum-Swan, B. (1994). Narrative, content, and semiotic analysis. In N . K. Denzin & Y . S. Lincoln (Eds.), Handbook of qualitative research (1st ed., pp. 463-477). Thousand Oaks, CA: Sage. Marjoribanks, K. (1986). A longitudinal study of adolescents' aspirations as assessed by Senger's model. Merrill-Palmer Quarterly, 32, 211-230. Marshall, S.K., Young, R.A., Turkel, H. , Arato-Bollivar, J., Hayoun, R., Marshall, E.G., & 'Domene, J.F. (2002, May). Adolescent possible selves as family projects. Paper presented at the annual meeting of the Canadian Psychological Association, Vancouver, BC. Marshall, S.K., Young, R.A., & Domene, J.F. (in press). Possible selves as joint projects. In J. Kerpelman & K. Dunkel (Eds.), Possible selves: Theory, research, and application. New York: Nova Science Publishers. Mau, W.C., & Bikos, L .H . (2000). Educational and vocational aspirations of minority and female students: A longitudinal study. Journal of Counseling & Development, 78, 186-194 McWhirter, E.H. (1997). Perceived barriers to education and career: Ethnic and gender differences. Journal of Vocational Behavior, 50, 124-140. Michel, K., Dey, P., Stadler, K. , & Valach, L. (in press). Therapist sensitivity towards emotional life-career issues and the working alliance with suicide attempters. Archives of Suicide Research. Miles, M.B. , & Huberman, A . M . (1994). Qualitative data analysis: An expanded sourcebook (2nd ed.). Thousand Oaks, CA: Sage. Murray, M . (2003). Narrative psychology. In J.A. Smith (Ed.), Qualitative psychology: A practical guide to research methods (pp. 111-131). Thousand Oaks, C A : Sage. Noller, P., & Bagi, S. (1985). Parent-adolescent communication. Journal of Adolescence, 8, 125-144. Noller, P., & Callan, V.J. (1990). Adolescents' perceptions of the nature of their communication with parents. Journal of Youth and Adolescence, 19, 349-362. O'Neill, P. (2002). Tectonic change: the qualitative paradigm in psychology. Canadian Psychology, 43, 190-194. Otto, L .B. (2000). Youth perspectives on parental career influence. Journal of Career Development, 27, 111-118. Paa, H.K., & McWhirter, E.H. (2000). Perceived influences on high school students' current career expectations. Career Development Quarterly, 49, 29-44. Papini, D.R., Frank F. Farmer, F.F., Clark, S.M., Micka, J.C., Barnett, J.K. (1990). Early adolescent age and gender differences in patterns of emotional self-disclosure to parents and friends. Adolescence, 25, 959-977. Peterson, G.W., Sampson, Jr., J.P., Lenz, J.G., & Reardon, R.C. (2002). A cognitive information processing approach to career problem solving and decision making. In D. Brown (Ed.), Career choice and development (4th ed., pp. 312-372). San Francisco: Wiley. Polkinghorne, D.E. (1995). Narrative configuration in qualitative analysis. Qualitative Studies in Education, 8, 5-23. Poole, M.E. , & Gelder, A.J . (1985). Family cohesiveness and adolescent autonomy in decision-making. Australian Journal of Sex, Marriage, and Family, 5, 65-75. Reid, P.T., & Stephens, D.S. (1985). The roots of future occupations in childhood: A review of the literature on girls and careers. Youth & Society, 16, 267-288. Rennie, D.L. (1999). Qualitative research: A matter of hermeneutics and the sociology of knowledge. In M . Kopala & L.A. Suzuki (Eds.), Using qualitative methods in psychology (pp. 3-13). Thousand Oaks, C A : Sage. Rennie, D.L. (2000). Grounded theory methodology as methodological hermeneutics: Reconciling realism and relativism. Theory and Psychology, 70,481-502. Richardson, M.S. (2000). A new perspective for counsellors: From career ideologies to empowerment through work and relationship practices. In A. Collin & R.A. Young (Eds.), The future of career (pp. 197-211). Cambridge, U K : Cambridge University Press. Riverin-Simard, D. (2000). Career development in the changing context of the second part of working life. In A. Collin & R.A. Young (Eds.), The future of career (pp. 115-129). Cambridge, U K : Cambridge University Press. Russell, A . & Saebel, J. (1997). Mother-son, mother-daughter, father-son, and father-daughter: Are they distinct relationships? Developmental Review, 17, 111-147. Salkind, N.J. (2003). Exploring research (4th ed.). Saddle River, NJ: Prentice Hall. Schultheiss, D.E.P., Kress, H.M. , Manzi, A.J. , & Glasscock, J.M.J. (2001). Relational influences in career development: A qualitative inquiry. The Counseling Psychologist, 29, 214-239. Schwandt, T.A. (2000). Three epistemological stances for qualitative inquiry: Interpfetivism, Hermeneutics, and Social Constructionism. In K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 189-213). Thousand Oaks, CA: Sage. Shoveller, J.A., Lovato, C.Y., Young, R.A., & Moffat, B. (2003). Exploring the development of sun tanning behavior: A grounded theory study of adolescent and parental experiences with getting a tan. InternationalJournal of Behavioral Medicine, 10, 299-314. Smith, J. (1999). Life planning: Anticipating future life goals and managing personal development. In J. Brandtstader & R . M . Lerner (Eds.), Action and self development: Theory and research through the life span (pp. 223-255). Thousand Oaks, Ca: Sage. Smith, J.A., & Osborn, M . (2003). Interpretive phenomenological analysis. In J.A. Smith (Ed.), Qualitative psychology: A practical guide to research methods (pp.51-80). Thousand Oaks, C A : Sage. Strauss, A . & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques (1st ed.). Thousand Oaks, C A : Sage. Strauss, A. & Corbin, J. (1994). Grounded theory methodology: An overview. In N . K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (1st ed., pp. 273-285). Thousand Oaks, CA: Sage. Strauss, A. & Corbin, J. (1998). Basics of qualitative research: Techniques and procedures for developing grounded theory (2nd ed.). Thousand Oaks, CA: Sage. Swanson, J.L., & Gore, P.A. (2000). Advances in vocational.psychology theory and research. In S.D. Brown & R.W. Lent (Eds.), Handbook of counseling psychology (3rd ed, pp. 233-269). New York: Wiley. Teddlie, C , & Tashakkori, A . (2003). Major issues and controversies in the use of mixed methods in the social and behavioral sciences. In C. Teddlie & A . Tashakkori (Eds.), Handbook of mixed methods in social and behavioral research (pp.3-50). Thousand Oaks, CA: Sage. Tucker, C.J., Barber, B.L. , & Eccles, J.S. (2001). Advice about life plans from mothers, fathers and siblings in always-married and divorced families during late adolescence. Journal of Youth and Adolescence, 30, 729-747. Turkel, H. (2003). The language of agency in parent-adolescent conversations about career. Unpublished Master's Thesis, University of British Columbia, Vancouver, British Columbia, Canada. Valach, L. (1988). A theory of goal-directed action in career analysis. In R.A. Young & W.A. Borgen (Eds.), Methodological approaches to the study of career (pp. 107-126). Westport, CT: Praeger. Valach, L., & Wald, J. (2002). Action theoretical perspective in rehabilitation. In L. Valach, R.A. Young & M.J. Lynam (Eds.), Action theory: A primer for applied research in the social sciences (pp 173-198). Westport, CT: Praeger. Valach, L., Michel, K., Young, R.A., & Dey, P. (2002). Stories of attempted suicide: Suicide career, suicide project, and suicide action. In L. Valach, R.A. Young & M.J. Lynam (Eds.), Action theory: A primer for applied research in the social sciences (pp 173-198). Westport, CT: Praeger. Valach, L. , & Young, R.A. (2004). Some cornerstones in the development of a contextual action theory of career and counselling. International Journal for Educational and Vocational Guidance, 4, 61-81. Valach, L., Young, R.A., & Lynam, M.J. (2002). Action theoiy: A primer for applied research in the social sciences. Westport, CT: Praeger. Von Cranach, M . , & Valach, L. (1984). The social dimension of goal directed action. In H. Tajfel (Ed.), The social dimension of social psychology (pp. 285-299). Cambridge: Cambridge University Press. Werrbach, G.B., Grotevant, H. D., & Cooper, C. R. (1992). Patterns of family interaction and adolescent sex role concepts. Journal of Youth & Adolescence, 21, 609-623. Whiston, S.C., & Keller, B .K. (2004). The influences of the family of origin on career development: A review and analysis. The Counseling Psychologist, 32, 493-568. Young, R.A., Antal, S., Bassette, M.E. , Post, A. , DeVries, N . , & Valach, L. (1999). The joint actions of adolescents in peer conversations about career. Journal of Adolescence, 22, 527-538. Young, R.A., Ball, J., Valach, L. , Turkel, H. , & Wong, Y.S. (2003). The family career development project in Chinese Canadian families. Journal of Vocational Behavior, 62, 287-304. Young, R.A., Friesen, J.D., & Pearson, H . M . (1988). Activities and interpersonal relations as dimensions of parental behavior in the career development of adolescents. Youth & Society, 20, 29-45. Young, R.A., Logan, C , Lovato, C.Y., Moffat, B., & Shoveller, J.A. (2005). Sun protection as a family health project in families with adolescents. Journal of Health Psychology, 10, 333-344. Young, R.A., Lynam, M.J., Valach, L., Novak, H. , Brierton, I., & Christopher, A . (2000). Parent-adolescent health conversations as action: Theoretical and methodological issues. Psychology and Health, 15, 853-868. Young, R.A., Lynam, M.J., Valach, L., Novak, H., Brierton, I., & Christopher, A . (2001). Joint actions of parents and adolescents in health conversations. Qualitative Health Research, 11, 40-57. Young, R.A., Marshall, S.K., Domene, J.F., Arato-Bolivar, J., Hayoun, R., Marshall, E.G., & Zaidnam-Zait, A . (2005). Relationships, Communication, and Career in the Parent-adolescent Projects of Families with and without Challenges. Manuscript in preparation. Young, R.A., Paseluikho, M.A. , & Valach, L., (1997). The role of emotion in the construction of career in parent-adolescent conversations. Journal of Counseling and Development, 76, 36-44. Young, R.A., Valach, L., Ball, J., Paseluikho, M.A. , Wong, Y.S. , DeVries, R.J., et al. (2001). Career development in adolescence as a family project. Journal of Counseling Psychology, 48, 190-202. Young, R.A., Valach, L., & Collin, A . (1996). A contextual explanation of career. In D. Brown & L. Brooks (Eds.), Career choice and development (3rd ed., pp. 477-512). San Francisco, C A : Jossey-Bass). Young, R. A. , Valach, L., Dillabough, J., Dover, C , & Matthes, G. (1994). Career research from an action perspective: The self-confrontation procedure. Career Development Quarterly, 43, 185-196. Youniss, J. & Smollar, J. (1985). Adolescent relationships with mothers, fathers, and friends. Chicago: University of Chicago Press. APPENDIX 1: C O M M O N FINDINGS WITHIN THE FOCUSED A N D DIFFUSE G O A L ORIENTATION GROUPS Focused Group Findings The following patterns of findings emerged as common within the 11 dyads whose project-related goals were focused in nature. Career projects were embedded in, and often secondary to, relationship projects. There was variation in the amount of progress made towards the project goals over time, with most of the families making great progress but substantial numbers making only a medial amount of progress. Although different families identified various specific barriers to their progress, the one barrier that consistently occurred was lack of time. Most focused families reported engaging in a wide range of joint activities, including having conversations, discussions and disagreements, providing/seeking parental support and advice, providing/seeking concrete assistance, disclosure of private information and emotions, and spending leisure time/having fun together. Parents from focused dyads were highly involved and engaged in the project, while adolescents varied considerably in terms of their involvement and engagement. The parent-child relationships of the focused group tended to be close (in terms of emotional connection and awareness of the details in each other's lives) to begin with, and became even closer over time. There also appeared to be a good agreement between parent and adolescent in terms of where the balance between granting/taking autonomy and independence on the one hand, and setting/accepting limits and boundaries on the other should lie. At the same time, conflicts and disagreements were present in virtually every single focused dyad. Another characteristic that was commonly in families with focused goals was the presence of substantial parental encouragement or their participating adolescent. Diffuse Group Findings For the eight dyads whose project-related goals were diffuse in nature, although relationship issues were a component of many projects, their projects tended to be more explicitly oriented towards issues of independence and autonomy. There was also considerable disagreement between parents and adolescents over what degree of independence or control is appropriate, both at the beginning and also over the course of their research involvement. Not surprisingly, then, the diffuse group exhibited considerable variation in the amount of progress made, ranging from a few dyads with considerable progress all the way to two dyads who abandoned their original project in favour of an alternate one. There were also a number of cases where different dyad members had conflicting perceptions about how their project had progressed. Two barriers were frequently identified by participants in this group: time and friction within the joint relationship. Most diffuse families reported engaging in a wide range of activities, although individual activities and time/actions with peers rather than families were surprisingly frequent. In terms of the parent-child relationship, some dyads were close, but more were variable or even clearly not close to each other. Moreover, there was frequently a lack of trust in the relationship. Also, most of the diffuse parents were very worried about their adolescents' safety and future. Conflict, including some cases of very intense conflict, was a common feature across the diffuse group. APPENDIX 2: COMPARISON C H A R T FOR G O A L ORIENTATION D A T A N B : a) this chart is a reproduction of the original, hand-written one; b) one case (2004) could not be classified as predominantly focused of diffuse and is, therefore, omitted from these charts. Focused Project Group Diffuse Project Group Relationship project was a focus: 2001, 2006, Relationship project was a focus: 2005, 2009, 2008, 2010, 2011, 2016, 2017, 2018, 2020, 2014,2021 2022 Not: 2002 Not: 2003, 2015 (but became so over time): 2013, 2019 Independence/control project was a focus: Independence/control project was a focus: 2006, 2018 2003, 2009, 2013, 2014, 2015, 2021 Not: 2001, 2002, 2008, 2010, 2011, 2016, Not: 2005, 2020 2017,2020,2022 Unclassifiable: 2019 Degree of progress made: Degree of progress made: Lots: 2001, 2002, 2016, 2018, 2022 Lots: Less: 2010, 2011,2017 Less: 2003, 2009, 2015,2021 Mixed: 2006, 2008 Mixed: 2005,2014 Abandoned: - Abandoned: 2013, 2019 Unclassifiable: 2020 Conflict: Conflict: 2001, 2002, 2006, 2008, 2010, 2011, 2016, 2003,2005, 2009, 2013, 2014, 2015,2019, 2018, 2020 2021 Not: 2017,2022 Not: -Communication problems: Communication problems: 2010, 2018 2003, 2005, 2009, 2013, 2014, 2015, 2019, 2021 Not: 2002, 2006, 2008, 2011, 2016, 2017, Not: -2022 Unclassifiable: 2001, 2020 Parental encouragement: Parental encouragement: 2001, 2006, 2008, 2011, 2016, 2017, 2018, 2004,2009,2013,2015 2020, 2022 Not: - Not: 2014, 2019, 2021 147 Unclassifiable: 2010 Unclassifiable: 2003, 2005 Level of closeness: Level of closeness: Very: 2001, 2002, 2006, 2008, 2010, 2011, Very: 2013 2016, 2017, 2018, 2020, 2022 Less: Less: 2003, 2005, 2009, 2014, 2015, 2019, 2021 Perceived barriers: Perceived barriers: Time: 2001, 2002, 2006, 2008, 2010, 2011, Time: 2003,2005,2009, 2013 2017, 2022 Friction/Avoidance: 2008, 2011 Friction/Avoidance: 2009, 2013, 2014, 2015, 2019, 2021 Finances: 2002, 2008 2020 Finances: 2003 Family problems: 2001, 2008, 2010, 2017 Family problems: 2003 Mother's health: 2022 Mother's health: -"None": 2016, 2018 "None": -APPENDIX 3: PRELIMINARY DRAFT OF THE GOAL-ORIENTATION FINDINGS Kinds of projects. There were noteworthy differences in the kinds of projects that dyads within the two groups tended to engage around. For focused group,-career exploration and working towards goals for the future were very much embedded in the maintenance of parent-adolescent relationship. For example, the project that one mother-son dyad engaged in over the course of their research involvement was described as "to continue to maintain their shared relationship while exploring options for the son's future social and academic life, leading to young adulthood." It appeared that, for virtually all of the families with focused goals, a principal component of career development was the development and maintenance of a close, functional relationship between the adolescent and his or her parent. As one adolescent claimed, "[the project] went good because I tried to bring a strong relationship with my Mum". In contrast, although the relationship project was present in the on-going interactions of many of the diffuse dyads, the dominant theme in the family career development projects of this group was to work on achieving appropriate levels of independence, and negotiating (with varying levels of success) the amount of autonomy that the adolescent should be granted. As one diffuse group dyad perceived it, the main purpose of their project was to "develop and promote a balance between the son's increasing independence, and responsibility." It must be noted, however, that issues of autonomy and control were not entirely absent in the other group. Instead, it appeared as i f a majority of the parents and adolescents with focused goals were in relative agreement about what is appropriate at that particular stage in life, so that independence and autonomy were not issues that parent and adolescent needed to explicitly focus upon as a career project. Progress achieved. In terms of the progress made by dyads towards achieving their family career development projects, the data revealed substantial variation among different dyads within each of the two groups. However, for dyads with focused goals, all made at least some progress, and many made significant achievements, to the point where the project was substantially achieved. As one mother explained, "[Engaging in the project] pushed me to say this is what we gotta do. It pushed [daughter] into saying we want to try this.. .the reward at the end of it is being able to talk, with her having this openness that we didn't have before." The range in progress was somewhat different for the dyads in the diffuse group. Although some made substantial achievements, many had periods of progress alternating with periods of stalling, or even regression. The equivocal success that was experienced by many of the diffuse goal dyads is reflected in one mother's statement that "I don't know how much of the career exploration we got to because, you know, we were focusing on things that he really wanted to do, like the circus and things like that, and going to the juggling place and things like that. So he's explored some, but he hasn't explored all of them". Moreover, the diffuse goal group included all of the dyads where an original project was abandoned, and replaced by another project that was perceived as more important. Also, for some of the diffuse group dyads, their project was perceived as being successful by one of member of the dyad but not the other. This difference in perspectives was most blatant in the following discussion where a daughter was trying to convince her mother that their project was a success, but was evident in several other dyads, as well: P35 I know how we can do it, so that's, as far as the question that you had why I didn't reach my goal? So that I'm serious why I didn't reach my goal with this bank card, and you buying your things, I didn't reach my goal A41 But you did reach your goal because you said what you were feeling. If you would have gotten angry and yelled at me and not given me a chance to speak then you wouldn't have reach your goal. But now that we've talked about it, and we've talked about it before also, I believe you reached it. P36 I don't know if that is reaching the goal or not. Because, A42 Well, our project P37 I don't know if I, I don't know if I got through. A43 Our project was to like, umm, what she said was to remain connected and learn about each other more and our expectations and me growing up as a teenager. But all that has to do with talking, and P38 Yeah we talk A44 Yeah we do but how much of that talking actually is something rather than just, P39 Yeah but A45 Than, P40 But yeah ok. So that's why I don't know if I'm answering the question there correctly or if you're answering it correctly. A46 Well it's all your own opinion. Everyone has different opinions on how they've reached then-own goal, and I think mine is P41 And my goal, my reaching the goal, I don't think I did reach it. In contrast, the dyads with focused goals tended to have very similar evaluations of how much progress had been made on their projects. Activities. Parents and adolescents working together on family career development projects engaged in range of joint activities, including casual conversations, formal discussions, use of humour and good-natured teasing, and taking the time to talk while engaged in other activities (e.g., driving to the adolescent's sporting event). Dyads reported sharing personal thoughts and emotions, and their concerns with each other. Adolescents reported seeking the advice and assistance of their parent, while the parents reported providing advice, support, reassurance, and practical assistance (e.g., helping to prepare a resume). There were also reports of arguments, disagreements, occasions where parent and adolescent would avoid speaking to one another for periods of time, and situations where the adolescent was selective in what her or she chose to share with the parent. A number of dyads reported engaging in civic activities together (e.g., volunteering at a soup kitchen) or working together at the mother's place of employment, and using these situations to discuss the adolescent's future; some dyads even used recreational activities (e.g., shopping together, watching movies or television, going to amusement parks) in the same manner, with the mother taking the opportunity to turn that activity into an opportunity to reflect on possible careers. Many dyads also recorded instances of simply having fun together, whether it was as simple as going on bike rides together or as momentous as taking a trip to Europe together, without the rest of the family. Interestingly, there was no practical difference in the range of activities engaged in by participants with focused goals versus participants with diffuse goals; examples of each these activities could be found in both groups, and most were identified in about as many focused goal dyads as diffuse goal dyads. However, where goal orientation became a salient distinguishing feature was in the frequency of participants' joint engagement in their project-related activities. Specifically, the dyads where their project was focused tended to log and report, over the telephone, more instances of joint action over the six-month monitoring period than dyads with diffuse goals. An additional difference between the groups was noticed for recreational and social joint activities: In most of the dyads with focused goals, these kinds of activities appeared to be an integral part of participants' schedules, a naturally occurring part of the process of maintaining a good relationship. For example, one mother with a focused goal reported engaging in recreational activities (e.g., weekends away at their cabin, playing Frisbee, visiting a Corn Maze) and/or attending her son's sporting activities in every fortnightly telephone monitoring interview during the six-month monitoring period. In contrast, many of the dyads with diffuse goals appeared to have fun together only on a sporadic basis, or on special occasions such when going on a vacation together, with many of these adolescents preferring to spend time with peers. In one diffuse goal dyad, for example, the daughter spoke of having fun with friends in virtually all of the telephone monitoring contacts, but did not report engaging in any recreational activities with her mother. Correspondingly, the mother reported only one instance of the two them participating in a joint recreational activity, over the entire six months of monitoring. Relationship and communication. Conflict, either project-related or to do with other issues in their lives was evident in the joint actions of dyads, regardless of goal orientation. In fact, a majority of the dyads in both group reported experiencing some form conflict between the parent and adolescent over the course of their research involvement. However, the conflicts that occurred between parents and adolescents with diffuse goals often seemed noticeably more intense and hostile than the conflicts that occurred for dyads in the focused goal group. One possible contributor to this difference was the fact that a majority of the parents in the former group had a tendency to react with strong, negative emotions when having disagreements with their son or daughter. For example, one of the diffuse group mothers repeatedly reported reacting with such frustration, anger and fear toward her son that she would to leave the room. While this was an extreme case (the mother reported that the son's actions triggered memories of her abusive ex-husband), strong emotional reactions were frequent components of the dyadic relationship for many of the participants with diffuse goals. In contrast, very few of the focused goal parents reported reacting so strongly in their discussions, and those that did reported making efforts to manage their emotions (e.g., "I was reminding myself to stay calm, to listen. Inside I was freaking out that he had to be exposed to this [violence and drug-use in the school], but relieved he would want to talk about this with us"). Another relational difference between the two groups is the differing levels of anxiety versus trust that was present in parents' relationships with their adolescents. Many parents in the diffuse goal group appeared to be anxious and worried about the immediate and future functioning of the adolescent, typically regarding safety or academic functioning. For example, in one diffuse goal mother-daughter dyad, the mother explained, "Like I said she's really not a, you know, a bad girl, yet. You know she is sixteen; she is at a dangerous age and you do hear all sorts of bad things ... we don't want her to get into trouble and be pregnant at sixteen or start smoking or drinking or start failing school. We want her to have a healthy normal life that, that I don't, you know that she's not abused by anyone or anything." In contrast, most parents in focused group appeared to believe that they had done a sufficiently good job of parenting in the past, that their children would now be capable of coping with problems. In response to a question about her concerns, one of the focused goal mothers stated, "I thought, oh my goodness, like what i f he decides to do something completely different, how will that affect him? ...but I don't, I know that whatever we put in for whatever reason, it is building him into a really good solid guy and somebody with a good head on his shoulders." Instead of anxiety and worry, the parent-adolescent relationship for most of the dyads with focused goals seemed to be built on a foundation of trust. For example, "I know there's going to be some [bad decisions] stuff come up, but hopefully, I'm just hoping, that we've got enough of a foundation that they're comfortable enough that - . They might not come home and the second they walk in the door tell me, but hopefully that they; that there's enough there that they can at least talk to me about it". The more strained relationships and increased levels of parental worries that were present in the diffuse goal group naturally led to the existence of breakdowns in communication for those dyads. Including breakdowns associated with anger, avoidance, manipulativeness, adolescents lying to their parents, and parental ambivalence about wanting to engage in conversation. In contrast, communication between dyads that had a focused goal were more open and consistent within their relationship. I do not intend to imply that dyads from the focused goal group never had problem with communication, or that the communication breakdowns that were evident in the diffuse goal group were constant, or that all the specific characteristic described above were present in all members of their respective group; however, the dominant patterns of communication that were most commonly found in dyads with differing goal orientations were clearly distinct. Perceived Barriers. When dyads were questioned about barriers to the success of their projects, the most common response of participants within the focused group was to identify a lack of time to engage in project-related activities as a barrier. For example, a focused group mother stated, "just timing again, in you... It's, you're a family of 5 and he's not the only child, and there's other people who need the time and attention; and fitting in the time to do, to carry on the conversations." Even though dyads with focused goals had a greater frequency of engagement in joint activities, these participants also believed that they did not have enough time to spend with each other. Although a substantial number of dyads with the diffuse goals also complained about not having sufficient time together, the most commonly identified barrier for that group was relational friction leading to intentional avoidance of each other. As a daughter disclosed on one occasion, "I tend to get angry right away whenever Mum calls my name," and another time, she stated, "My bad temper can still get in the way of talking ... I ' l l just go up to my room and lock myself in." Although lack of time was not unimportant to participants with diffuse goals, it was evident the most salient barrier for participants with diffuse goals was a relational pattern of coming into conflict or friction, followed by a period of not being able to be with each other (a pattern which was largely absent in the focused goal group). APPENDIX 4: C O M M O N FINDINGS WITHIN THE MOTHER-DAUGHTER A N D MOTHER-SON D Y A D I C CONFIGURATION GROUPS Mother-son Group The following patterns of findings emerged from our analysis of the themes and patterns that characterised the mother-son group. In terms of the nature of the family career development project, a majority of mother-son projects focused on maintaining or improving their mutual relationship, with issues of independence, and future vocation being the next most common. No dominant pattern emerged regarding the degree of progress that was achieved: some dyads made substantial progress towards their projects, others made minimal progress, and many fell somewhere in between. Mother-son dyads experienced a wide range of barriers to achieving their project, but the most prominent barrier was being too busy or lacking the time to devote to the project. The most common joint activities within this group were engaging in conversations, spending time together in fun or leisure activities, joint engagement in competitive sports, (with sons being involved on teams and mothers taking an active role in supporting, cheering and/or transporting them), the son seeking and the mother providing concrete assistance with job-related activities (e.g., resume writing, searching for available positions) and the seeking/provision of parental support and advice. In terms the parent-adolescent relationship, most dyads within the mother-son group reported having a close relationship, and had interactions featuring high levels of disclosure and parental encouragement. At the same time, however, the group also tended to experience some degree of conflict or disagreement with each other during their research involvement. In terms of the quality of joint communications, no clear pattern emerged, with approximately half of the group maintaining good communications, while the other half experienced problems with communication. Although different mother-son dyads perceived a range of barriers to achieving their joint project, the only barrier that was identified by a majority of the group was a lack of time. Mother-daughter Group For the mother-daughter group, the dominant family career development projects were ones that involved maintaining or improving their mutual relationship, or promoting the adolescent's independence/autonomy. There was substantial variation within the group in terms of the degree of progress that was achieved on their joint projects. The kinds of joint activities that this group tended to engage in included having conversations with each other, shopping together, spending time together in fun or leisure activities, the mother supporting the daughter's involvement in competitive sports, joint engagement around job-related activities (e.g., resume writing, searching for available positions) and the seeking/provision of parental support and advice. Most mother-daughter dyads had a close relationship, although all but one also reported experiencing some degree of conflict or disagreement within that relationship. High levels of disclosure and parental encouragement were also common in this group. Examination of the way that mothers and daughters communicated yielded no dominant patterns. Lack of time was the most frequently cited barrier to achieving their project goals, although the experience of friction and choosing to avoid each when in conflict was another common barrier. APPENDIX 5: COMPARISON CHART FOR D Y A D I C CONFIGURATION D A T A N B : this chart is a reproduction of the original, hand-written one. Mother-Son Group Mother-Daughter Group Relationship project was a focus: Relationship project was a focus: 2004, 2005, 2006, 2011, 2017, 2918, 2001, 2008, 2009, 2010, 2014, 2016, 2020 2021,2022 Not: 2003 (but became so over time): 2013, Not: 2015,2002 2019 Independence/control project was a focus: Independence/control project was a focus: 2003,2006, 2013,2018 2009, 2014, 2015,2021 Not: 2004, 2005, 2011, 2017, 2020 Not: 2001, 2002, 2008, 2010, 2016, 2022 Unclassifiable: 2019 Explicit vocation project was a focus: Explicit vocation project was a focus: 2004,2005,2019,2020 2016, 2022 Not: 2003, 2006, 2011, 2013, 2017, 2018 Not: 2001, 2002, 2008, 2009, 2010, 2014, 2015, 2021 Degree of progress made: Degree of progress made: Lots: 2004, 2018 Lots: 2001,2002, 2016, 2022 Less: 2003,2011,2017 Less: 2009, 2010, 2015,2021 Mixed: 2005, 2006 Mixed: 2008,2014 Abandoned: 2013, 2019 Abandoned: -Unclassifiable: 2020 Conflict: Conflict: 2003, 2005, 2006, 2011, 2013, 2018, 2001, 2002, 2008, 2009 2010, 2014, 2015, 2019,2020 2016, 2021 Not: 2004,2017 Not: 2022 Communication problems: Communication problems: 2003,2005,2013,2018, 2019 2009, 2010, 2014, 2015,2021 Not: 2004, 2006, 2011,2017 Not: 2002, 2008, 2016, 2022 Unclassifiable: 2020 Unclassifiable: 2001 Interactions featured self-disclosure: Interactions featured self-disclosure: 2004, 2006, 2011, 2017, 2018, 2019, 2020 2002, 2010, 2014, 2015, 2016, 2022 Not: 2005 Not: 2009, 2021 Unclassifiable: 2003,2013 Unclassifiable: 2001,2008 Level of closeness: Level of closeness: Very: 2004, 2006, 2011, 2013, 2017, Very: 2001, 2002, 2008, 2010, 2016, 2022 2018,2020 Less: 2003, 2005, 2019 Less: 2009, 2014, 2015, 2021 Parental encouragement: Parental encouragement: 2004, 2006, 2011, 2013, 2017, 2918, 2001, 2002, 2008, 2014, 2015, 2016, 2022 2020 Not: 2019 Not: 2009, 2021 Unclassifiable: 2003, 2005 Unclassifiable: 2010 Perceived barriers: Perceived barriers: Time: 2003, 2005, 2006, 2011, 2013, 2017 Time: 2001, 2002, 2008, 2009, 2010, 2022 Friction/Avoidance: 2011, 2013, 2019 Friction/Avoidance: 2008, 2009, 2014, 2015, 2021 Finances: 2003, 2004, 2020 Finances: 2002, 2008 Family problems: 2003, 2017 Family problems: 2001, 2008, 2010 Mother's health: - Mother's health: 2022 "None": 2018 "None": 2016 APPENDIX 6: PRELIMINARY DRAFT OF THE D Y A D I C CONFIGURATION FINDINGS Kinds of projects. The kinds of family career development projects in which the two sets of dyads engaged were predominantly similar, in the sense that projects across both the groups tended to include a component of developing and maintaining their mutual relationship. For example, one mother-daughter dyad's joint project was to "strengthen the mother-daughter relationship, where both will feel listened to and able to fully express their ideas and needs." Similarly, in one mother-son dyad, their joint project was defined as, "for the Mother and Son to maintain their close relationship and willingness to share ideas while discovering their new roles as they adjust to the son's transition towards young adulthood." Furthermore, a large minority of dyads within both groups also had projects that included negotiation of the adolescent's independence and level of permissible autonomy. Finally, no differences were discernable in terms of the goal orientations of mother-daughter dyads and mother-son dyads, with focused goals and diffuse goals being evident in both groups. The one distinction between the groups for this aspect of the data was in the incorporation of explicit vocational development goals within joint projects. This was relatively common in the projects of mother-son dyads, (e.g., "to maintain their respectful and supportive relationship while they both engage in the process of career exploration," and "to have exploration of career development where the format of interactions will continue to be transformed into one where the son is able to openly discuss his interests"), but was identified in only two of the mother-daughter dyads did so. This difference is an indicator that, while some mother-son dyads are already framing their career projects in terms of specific vocational choices at this age, mother-daughter career development projects tend to not to do so. However, it is important to recognise that relational and autonomy-related projects were more important for parents and adolescents, regardless of the gender configuration of the dyad. Activities. The project-related activities that mother-son and mother-daughter dyads engaged in over the course of their research involvement were largely similar. Specifically, activities that were common within both groups of dyads included (a) having conversations, and discussions with each other, (b) joint leisure activities and "having fun" together, (c) joint engagement in the adolescent's sporting activities (with the roles of the mother including cheering and supporting her child, and providing transport to and from practices and games), (d) working together on specific job-related activities (e.g., resume-writing, job-searching) and (e) adolescents seeking support and advice from the mother who, in turn, provided it. Joint engagement in shopping was something that many mother-daughter dyads, but very few mother-son dyads, reported as an activity. However, given the fact that this particular activity is of minimal relevance to understanding career development, it was judged to be a trivial rather than practical difference in this data set. Progress achieved. Division of the sample according to dyadic configuration yielded no identifiable patterns in the amount of progress achieved, with equivalent numbers of both mother-son and mother daughter dyads achieving substantial progress, less progress, and mixture of success and failure. Although two of the mother-son dyads abandoned their initial project and none of the mother-daughter dyads did so, this was not judged to constitute sufficient evidence of an important dyadic configuration difference. Therefore, it would appear as i f dyadic gender configuration is not a salient contextual factor for the amount of progress that dyads achieve on their family career development projects Perceived Barriers. Participants from both the mother-daughter and mother-son dyads most commonly identified an insufficient amount of time spent together (primarily due to one or both of the dyad members being too busy) as a barrier to progress on their projects. One distinction between the groups was in the identification of interpersonal friction and avoidance of each other as a barrier: this was a commonly identified barrier in the mother-daughter dyads, but not the mother son dyads. One mother's provided ah example of the friction, "I was thinking how frustrating and exasperating this issue is... I was feeling that i f she were younger and smaller this situation would bring me to want to spank her," and, in a different dyad, a daughter exemplified how avoidance interfered with engaging in their project, "My bad temper can still get in the way of talking ... I ' l l just go up to my room and lock myself in." This difference between the two sets of dyads must be understood in light of the finding that conflict was identified as a feature of the relationships of most of the participants, regardless of the configuration of the dyad. Therefore, the fact that it only tends to be perceived as an impediment to progress in the projects for mothers and daughters raises an important question: is it the case that, contrary to the Young et al., (1988) study, mother-son dyads do not use avoidance as a strategy to cope with conflict as often as mother-sons? Alternatively, it is possible that the conflict between mothers and sons simply does not interfere with their engagement in joint projects in the same way that it does for mother-daughter dyads. Relationship and communication. Few important differences were identified in the patterns of communication and relationship for the two sets of dyads: mothers and sons, and mothers and daughters both tended to have relationships that were characterized as being "close" in terms of the bond between them, featuring self-disclosure and the sharing of personal information. Moreover, although approximately half of both sets of dyads experienced difficulties in communication at some point during their research involvement, there was a clear tendency for mothers verbally encourage their children irrespective of the adolescent's gender.