Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influence of client social class on therapists’ perceptions and evaluations Hutchings, Anita Clare 1998

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

Item Metadata

Download

Media
831-ubc_1998-0478.pdf [ 9.17MB ]
Metadata
JSON: 831-1.0053932.json
JSON-LD: 831-1.0053932-ld.json
RDF/XML (Pretty): 831-1.0053932-rdf.xml
RDF/JSON: 831-1.0053932-rdf.json
Turtle: 831-1.0053932-turtle.txt
N-Triples: 831-1.0053932-rdf-ntriples.txt
Original Record: 831-1.0053932-source.json
Full Text
831-1.0053932-fulltext.txt
Citation
831-1.0053932.ris

Full Text

THE INFLUENCE OF CLIENT SOCIAL CLASS ON THERAPISTS' PERCEPTIONS AND EVALUATIONS By Anita Clare Hutchings B.A., Bishops University, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Counselling Psychology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 1998 ©Anita Clare Hutchings, 1998 In p resent ing this thesis in partial fu l f i lment o f t h e requ i rements f o r an advanced degree at the Univers i ty o f Brit ish C o l u m b i a , I agree that the Library shall make it f reely available f o r re ference and s tudy. I fu r ther agree that permiss ion fo r extensive c o p y i n g o f this thesis f o r scholar ly pu rposes may b e g ran ted by t h e head of m y d e p a r t m e n t or by his o r her representat ives. It is u n d e r s t o o d that c o p y i n g or pub l i ca t ion of this thesis fo r f inancial gain shall n o t be a l l owed w i t h o u t my w r i t t e n permiss ion . The Univers i ty o f Brit ish C o l u m b i a Vancouver , Canada D e p a r t m e n t DE-6 (2/88) 11 ABSTRACT This study was designed to explore whether initial information on client social class would impact negatively on the therapist's perception, evaluation and proposed treatment of the client. Respondents were randomly selected (N= 300) from the membership lists of the British Columbia Psychological Association and the Canadian Guidance and Counselling Association (British Columbian members). The return rate was 37% (N= 111). Of these, 49 were returned from the British Columbia Psychological Association and 62 from the Canadian Guidance and Counselling Association. Each respondent read the client information forms. They then filled out a questionnaire containing a seven-point Likert scale, a qualitative response section, a semantic differential of opposing adjectives, a respondent relationships form, and a demographics section. Half of the questionnaires contained the description of a lower social class client, and half contained the description of a middle class client. All remaining information was identical. Examination by total sample population, association and gender all produced significant results linked to client social class. The lower class client was evaluated as less intelligent, less knowledgeable, less capable, less clean, less hostile, louder, and more emotional than the middle class client. The CGCA respondents produced lower frequencies of presenting problems for the lower class client. Doctoral degree respondents produced less significant bias than respondents with master's degrees. An unusual pattern of evaluation linked to respondents' years of experience was apparent, however results by client social class were not significant. Caution is advised when translating results to real-life situations since counsellors may not necessarily respond in the same way in a real counselling setting as they would to a hypothetical client. iii TABLE OF CONTENTS Abstract ii Table of Contents iii List of Tables vi List of Figures viii Preface ix Acknowledgement x INTRODUCTION 1 LITERATURE REVIEW 5 Portrait of Life in the Working Class 5 Life Choices and Experiences 6 Education 6 Work 10 Leisure 11 Stressful Life Events 12 Health 13 Self Esteem and Self-Worth 15 Values and Beliefs 18 Social and Political Ideology 18 Parental Values 21 Education 23 Sex Roles 24 Emotional Expression 24 The Schematic Model 26 Schemas 26 Categorization 27 Stereotypes 28 Lower Class Stereotypes 32 The Schematic Model Applied to Counselling 34 Class and Psychotherapy 38 Acceptance for Therapy 39 Continuation in Therapy 40 Perceptions and Expectations 43 Resistance 44 iv TABLE OF CONTENTS Testing Bias 45 Adaptive Functioning, Presenting Problem, and Outcome 47 QUESTIONS/HYPOTHESES 53 METHOD 55 Design 55 Independent variable = Client Socio-economic Status 55 Pilot Study 58 Participants 59 Procedure 60 Characteristics of Respondents 61 Stimulus Material 68 Dependent Variable 69 Measures 69 Data Analysis 71 Summary 72 RESULTS 73 Respondents' Client Ratings and Evaluations 73 Problem Evaluation Results 73 Choice of Therapy, Presenting Problems, and Length of Therapy 75 Choice of Therapy 75 Explanations for Choice of Therapy 79 Presenting Problems 80 Length of Therapy 8 5 Summary 86 Client Characteristics 87 Respondents' Personal Network and Relationship To Client Evaluations 93 Respondents' Social Class Background 96 Respondents' Levels of Education 98 Respondents' Years of Experience 102 Summary 108 DISCUSSION 109 Client Characteristics Results 109 Findings by Total Sample Population 110 Findings by Association 111 Findings by Respondent Level of Education 112 Findings by Gender 113 V TABLE OF CONTENTS Findings by Respondents' Years of Experience 114 Explanations for Choices of Therapy 115 Issues/Presenting Problems 116 Findings by Association 117 Findings by Gender 117 Suggested Length of Stay in Therapy 120 Findings by Association 120 Problem Evaluation Results 122 Choice of Therapy 123 Summary 124 Implications 125 Strengths and Limitations 127 Suggestions for Future Research 129 Conclusion 131 BIBLIOGRAPHY 133 APPENDICES 139 Appendix A: Therapist Decision-Making 139 Appendix B: Independent T-Tests Results of Respondents' Years of Counselling Experience 148 Appendix C: Independent T-Tests of Respondents' Evaluations of the Problem 149 Appendix D: Chi-Square Analyses of Important Issues 164 Appendix E: Client Characteristics - T-Tests for Independent Samples 234 Appendix F: Client Characteristics - Analyses of Variance 244 Appendix G: Problem Evaluation - Analyses of Variance 254 VI LIST OF TABLES Table 1 Hollingshead's Occupational and Educational Characteristics of Five Socio-economic Levels 56 Table 2 Student Vignette Ratings of Hypothetical Client's Social Class 58 Table 3 Demographic Characteristics of Respondents 62 Table 4 Total Sample Scores on Respondent Ratings of the Class III and Class V Client 74 Table 5 Respondents' Choice of Therapy by Association and by Total Sample Population 76 Table 6 Choices of Cognitive and Behavioural Therapies by Total Sample Population and by Respondent Association 77 Table 7 Choices of Cognitive and Behavioural Therapies by Total Sample Population and by Respondent Association 78 Table 8 Respondents' Choices of Most Important Client Issues Needing Address 82 Table 9 Respondents' Average Recommended Length of Stay in Therapy Based on Client Social Class (N = 87) 86 Table 10 Effect of Client Social Class on Respondents' Ratings 89 Table 11 Significant Results of T-Tests of Client Ratings by Association 90 Table 12 Significant Results of T-Tests on Evaluations of Client Characteristics by Respondents' Gender 92 Table 13 Respondents' Personal Network 94 Table 14 Client Characteristics Showing Significance By Client Social Class and Respondent Degree 100 Table 15 Rating Means For Motivation For Change and Prognosis Based On Respondents Years of Experience 104 vii LIST OF TABLES Table 16 Mean Client Characteristic Scores by Respondents' Years of Experience 105 Vlll LIST OF FIGURES Figure 1 Respondents' Social Class Background 97 Figure 2 Respondents' Years of Experience 103 Preface Counsellors and psychologists enter their professions out of a sincere desire and commitment to provide caring, respectful help to their clients. As a result of their training and personal values, they are probably better equipped than most members of the general population to see clients as individuals and to view their problems in a non-judgemental, unbiased way. At the same time, however, these professionals live and work in our larger society, and it would be truly surprising if the stereotypes and prejudices of that society had no effect on these individuals and their work. The research that is reported in this document examines psychologist and counsellor bias in the area if client social class. It documents examples of bias, and, as such, must be a cause for concern to any therapist who honours the values of respect and client welfare. At the same time, there are several areas where the levels of bias could be interpreted as being reassuringly low. While longitudinal interpretations are beyond the scope of the current study, it is possible that this snapshot has captured some positive change; however, that is a question that remains for future research. My personal experiences from having grown up in a poor family, and my deep concern for disadvantaged individuals, is reflected in the choice of this research topic and, perhaps, in the perspective I bring to discussing my findings. While I acknowledge the possibility that my own perceptions may reflect some bias, I believe that our open discussion of these issues will only further a common desire to be of assistance to our clients. X Acknowledgements First and foremost, I want to thank Dr. Beth Haverkamp for her patience, assistance, and support, and for her interest in the subject of social class in the counselling setting. Having her at the university provided me with the safety and support I needed to continue to hold on to my own beliefs and values. Secondly, I would like to thank my children, Roxanne, Tchad, Leon and Jonathan, and my brother Rick Hutchings for their own stories of courage, and for their belief in my ability to succeed. Next, I want to thank Jenifer Jones for her dedication of energy and hours in proof-reading the final copy and assisting with the appendices. I also want to thank her for her emotional support and taking on more of the housework during the last few months of putting this research project together. I also want to thank Fiona Graham and Janet Bianic for their support and encouragement over the past several years. Finally, to the Linedrivers who, without knowing it, helped keep me safe and sane on numerous occasions over the years, I send my love and thanks. 1 Introduction A counsellor's primary task is to understand a client and his/her problems in order to encourage and/or facilitate change. While therapists are aware of the importance of and strive towards unprejudiced and honest dealings with every client, these interactions occur in an interpersonal context and, as such, are susceptible to errors and biases in perception. There is a considerable body of research that seems to indicate that people, including counsellors, are poorer judges of their interactions than they perceive themselves to be (Arkes, 1981). While we can often recognize perceptual errors in others, many of us remain oblivious to our own errors in perception. Therapists are no exception to this occurrence. For the purposes of this study, I chose to investigate the subject of counsellor biases that may occur before or during the counselling process, and are related to differences in client social class. The literature, while relatively sparse, clearly establishes that social class has been a powerful factor in how persons are perceived and judged by others (Garfield, 1986; Herr, 1989; Lorion, 1986). While the majority of counsellors come from middle class backgrounds, the majority of clients who arrive for treatment are from lower class backgrounds (McCarthy et al., 1991). The environment from which the client and counsellor come, including cultural values, may be very different. In addition, we live within a society that upholds middle class values and attitudes as being the norm to strive towards. The values that we hold help create our personal attitudes towards people, places and situations (Herr, 1989; Lerner, 1973). These personal attitudes affect our beliefs about and our feelings towards 2 others. This being the case, the extent to which a majority of counsellors understand and relate to lower social class individuals is questionable. The dominant culture's beliefs, stereotypes, prejudices and judgements of a group will often impact in negative ways on members of the non-dominant culture (Herr, 1989; Morris & Williamson, 1989). Research in the area of schema and, more importantly, stereotype formation and activation, suggests that membership within a particular target group works to prime or activate the group stereotype within the perceiver's memory. Once an individual is categorized as belonging to a particular group or category, individual characteristics and behaviour are evaluated in ways that support the existing belief or stereotype. Characteristics and behaviour that do not support the stereotype may be overlooked while behaviour that supports the stereotype receives special attention. When the socially defined stereotype consists of predominantly negative traits, the impact within a counselling perspective can be particularly significant. Since clinicians are not exempt from the cultural attitudes, beliefs and stereotypes that make up their own societal background, and initial evaluation is of particular importance within the counselling domain, detection and awareness (or lack of) of bias and stereotyping is of particular relevance. As stated by Herr (1989), "Poor people are not rich people without money. Their values, risk-taking, sense of self, sense of power, and focus on the past, present, and future are all conditioned by the environment in which they find themselves." (p. 152) In addition, by lacking a familiarity with and understanding of the client's value system, the counsellor's own values and intervention strategies may conflict with the client's. Therapists may label a client as resistant or unresponsive when, in fact, the 3 counsellor's own lack of societal and cultural awareness may contribute to create a situation where the client is inappropriately labelled and judged. Within the field of psychology, the focus has traditionally been on individual and intra-psychic influences on mental health. In recent years, the focus has expanded to include the impact of social and environmental factors on psychological well-being. While cross-cultural counselling, in particular, has taken on prominence within the psychological literature, social class appears to have a place within the research and counselling domains only in so much as ethnic and racial minorities are disproportionately represented in the lower classes (Pederson, 1981). This could lead readers to conclude that while social class is an important factor for cross-cultural counselling, it is not necessarily a factor to be considered when counselling non-ethnic minority members. Much of the existing research on social class has been carried out from a sociological perspective. These studies have revealed that social class and social class background impact on the individual in a variety of ways including their life experiences and choices, self-esteem and self-worth, and on the values they hold (Coclough & Beck; Rosenberg & Pearlin, 1978; Rubin, 1976). The issue of social class appears to have received only limited recognition and study within the psychological literature. Psychological research has revealed a class bias in diagnosis and treatment, with lower and working class clients being more likely to receive more severe diagnoses, and less likely to receive talk-oriented treatment (Garfield, 1986). Epidemiological studies have revealed significantly higher rates of health-related problems and psychological disorder among the lower social classes (Frank & Mustard, 1994; Wilkins, 1988). In addition, a national study (Wilkins, 1988) indicates that factors directly related to social class such as education and income are strongly related to rates of disorder. Most of the psychological research pertaining to social class was carried out in the United States in the 1960s and 1970s. There has been little psychological research performed within Canada focused on social class. Attention to class has been especially sparse within the past fifteen years. Considering the economic decline within the past 10-15 years, this appears to be a serious gap within the Canadian research area. In addition, a trend in recent sociological research has been to move away from using the term lower class, and incorporate a generalised term such as the disadvantaged, or subgroup-specific terms such as transient workers, and the homeless. While the term lower class has a negative connotation, and could be interpreted as perpetuating a stereotype, North America society has tended to view itself as relatively classless. We are rarely encouraged to define or articulate differences in terms of class, or in terms of the inherent power differences. For these reasons, I have chosen to continue to use the terms lower social class and lower socio-economic status throughout this research study. For this study I was particularly interested in examining how counsellor beliefs, attitudes and decisions influence a lower socio-economic status client's acceptance for and duration of therapy, and whether the counsellor's awareness of a client's socio-economic status impacts on how the client is perceived, evaluated, and treated. I investigated this question within an experimental design, using a random sample of psychologists and counsellors from across British Columbia. 5 Literature Review The task of this literature review, is to examine what research pertaining to social class has been generated to date. Three broad areas appear to be of particular relevance. First, in order to develop a more realistic and less stereotypic view of the lower socio-economic status client, I will attempt to shed light on the possible effects upon the individual of growing up and living within the lower social classes. Second, since counsellors perceive and understand their clients using the same cognitive processes that other individuals use in perceiving people, it is important to present a conceptual overview of the social cognitive processes involved with person perception and evaluation, with special focus on the development and maintenance of biases in the form of stereotypes. Thirdly, I will explore the relationship of psychology and social class as a significant area for the activation and maintenance of negative stereotyping. In summary, I will attempt to include studies which provide an overview of life in the working class, rudimentary information on the cognitive schematic model and stereotyping, as well as studies in the field of psychology that examine the client's acceptance into therapy, socio-economically affected dropout rates, therapist-perceived client attributes and therapist expectations of the lower class client, and outcome variables of therapy. Portrait of Life In The Working Class To facilitate change, the counsellor must be capable of understanding and accepting the client. Failure to adequately recognize and respect social class differences may be of particular relevance within the field of counselling psychology. In order to be of greatest service to the client, the counsellor must be capable of putting aside socially prescribed definitions and stereotypes of the working class and the poor in order to develop a clear understanding of the particular client and the contexts of his or her life. Various studies, predominantly from within the field of sociology, have revealed that class status and class background have an impact on individuals in a variety of ways, including their life experiences and choices, self-esteem and self-worth, and values and beliefs. In both the United States and Canada, higher incomes have been associated with higher status, power, and unlimited access to a variety of resources, while lower incomes are associated with lower status, less power, and restricted access to resources (Bullock, 1995). In this section I will attempt to elaborate on several life variables that may be effected by class membership. Life Experiences and Choices Education. Rubin (1976), in her in-depth study of life in the white working class in America, provides a powerful insight into the effects of class on a person's opportunities, choices, and educational experiences. Rubin interviewed 50 white working class couples, and another 25 white middle class couples as a comparison group. She found that the working class participants had gotten married and had children at an earlier age than their middle class counterparts, requiring them to get jobs to support their families. While one might assume that the working class participants had the same life and educational opportunities as the middle class, and simply made different choices, Rubin (1976) stresses that the struggles and crises experienced in the working class families limit the alternatives they perceive and experience, and therefore the choices they make. While theoretically both the American and Canadian school systems are open and accessible to all, middle class children in fact have a head start throughout childhood (Lareau, 1987; 7 Rubin, 1976). Middle class children attend school having had adequate food, properly attired for the season and educational activities, having less daily disruptions (e.g. frequent moving, less health problems), more parental support, and more exposure to, information about, and opportunities for higher education (Lareau, 1987; Rubin, 1976). Laureau (1987), in her qualitative study of family-school relationships in white working-class and middle-class communities, found that while schools have standardized views of the acceptable and proper roles of parents in schooling, lower social class parents have less access to the educational, capital and cultural resources required to comply with teachers' requests for parental participation (e.g. money, time, parents' level of education, materials, personal connections, etc.). Guppy, Mikicich and Pendakur (1984) carried out research in order to systematically examine historical changes in Canadian educational inequalities based on the social class background of the students. Guppy et al. examined groups of birth cohorts spanning a forty year period from 1913 to 1952, using a total sample of population of 44,868 respondents. Social origin was operationalized using the parents' levels of education and father's occupation (which was collapsed into four main categories: 1) professional/managerial; 2) white collar; 3) blue collar; and 4) farmer. Parental education was based on three categories: on whether one parent, neither parent or both parents had completed high school. Results showed that the disparity in high school completion rates based on social origin has diminished over time, although significant disparities were still evident. Guppy et al. (1984) also found that, while the average years of education has increased for all social classes, the actual reduction in educational disparities across social 8 classes appeared significant only when the 1938-1942 birth cohort was considered. They suggest that the reduction of disparities evidenced at that time may be directly related to the educational reforms that occurred in Canada during the 1950s and 1960s. Because of their findings, they questioned whether the economic decline of the 1980s (and presumably the 1990s as well) might not lead back to a further retrenchment of educational disparities based on social class background, as was evident earlier in the century. Guppy et al. (1984) noted that significantly greater disparities based on social class origin existed at the university level, and showed only a modest trend towards reduction over time. However, in noting the greater levels of social class disparity at the university level, they stated that examining why the disparities tend to heighten at higher levels of education did not fall within the scope of their study. While their study did suggest that inequalities have moderately declined over time in the area of general level of education, they also stress that their study did not show a diminishment of the impact of social class on the acquisition of specific types of education that might yield more value to the individual. For instance, if college and undergraduate degrees become increasingly ineffectual for occupational and financial attainment, then professional and graduate degrees may become increasingly important as a means of intergenerational mobility for the lower social classes. Coclough and Beck (1986), carrying out a national longitudinal research study involving a sample of 5,671 American male high school students, suggest that the American educational structure reproduces social class and contributes to the maintenance of class-based inequalities inherent in its society. They make three major assertions. First, they assert that private schools, primarily accessible to the middle and upper classes, result 9 in the reproduction of advantaged class positions. Second, recent trends show that the poor, Black and other ethnic minorities, and less educated individuals, remain in inner city areas, while the more affluent tend to move to the suburbs. The differentiation of financial resources available between inner city and suburban schools means that more affluent schools acquire and maintain more resources than the inner city schools. Even though some schools designated as 'Inner City' may receive extra funding, schools in suburban areas remain ahead in fund-raising capacity, volunteer professional resources, space, and equipment (Colclough and Beck, 1986). Third, although curriculum tracking is considered to be ability based, standard intelligence tests, developed by a middle class education system, still may tend to reflect middle class white male perspectives, and therefore again provide white middle class children with a head start. In addition, Coclough and Beck (1986) note that past studies (Rosenbaum, 1976, cited in Coclough & Beck, 1986) have demonstrated that while tracking fosters the illusion of meritocracy, factors such as individual levels of motivation and ability have had only a limited effect on actual placement within a particular track. Coclough and Beck's (1986) study found that the majority of children placed in the 'manual' rather than the 'mental' track were from the working class, while the opposite was true for the middle class. A seven-year follow-up revealed that being placed in the college or vocational track had a significant influence on the likelihood of class mobility and class reproduction (Colclough and Beck, 1986). A significant limitation of this study, however, is its lack of generalizability to the female student population, as the study involved only male students. Rubin (1976) and Laureau (1987) noted that most of the working class families in their studies did not expect their children to attend college or university, and in 10 Rubin's (1976) study, this expectation was lower for daughters than for sons. While statistics show that Canadian students are becoming better educated, by 1991 students from the lowest social class in Canada were dropping out of high school at a rate 2.5 times that of non-poor families (Canada Council on Social Development, 1996). Work. By the age of 25 years, more than half of the working class men in Rubin's (1976) study had held an average of 6-10 jobs. Generally, these jobs were in blue collar positions such as construction labourer, assembly-line worker, or gas station attendant. Some left jobs because of low pay and poor working conditions. For many, the job changes were involuntary due to shutdowns, lay-offs, and industrial accidents (Rubin, 1976). Most blue collar jobs allow for less independent judgement, freedom, and autonomy than white collar jobs and offer few intrinsic rewards. Frequently, they entail a truncated career ladder and incur little social status and prestige. While some of the men at the higher end of the blue collar ladder felt a certain amount of pride and satisfaction with their jobs, they remained frequently dependent upon unions and striking for pay increases. The fear of unemployment and one's family doing without necessities, was a everyday fear among the working class men (Rubin, 1976). Because, traditionally, career has been the significant marker of a successful man, Rubin (1976) noted that many of the blue collar workers expressed a sense of failure and discontent with their life situations. Those who managed to secure more stable long-term positions expressed a sense of resignation to a situation consisting of stable yet stagnant long-term work. This resignation is very poignantly noted by Rubin: "Imagine the consequences to the shape and form of that human life. Imagine, too, an environment in which the same paucity of choices is the reality of most lives - no friends or relatives 11 around who see a future with plenty of possibilities stretching before them; no one expects very much because experience has taught them that such expectations end painfully" (Rubin, 1976, p. 163). By 1994, over 17% of Canadians (20% of all children) were living in poverty (Statistics Canada, 1995). Close to half of this number includes families and individuals who are earning wages that keep them below the poverty line (Statistics Canada, 1995). While in times of economic distress the middle-class will also suffer more instability and unemployment, the levels of poverty and distress are far more significant for those already living near the edge even in the periods of comparative economic stability. Leisure. Rubin (1976) found that relationships with family members were at the heart of working class social life, while activities such as dinner parties (almost non-existent among the working class) were the central focus of middle class leisure life. While one of the leisure activities of the working class men in Rubin's study included going out to "have-a-drink-with-the-boys," fitting neatly into one of the classic stereotypes of the blue collar worker in our society, Canadian statistics (Wilkins, 1988) report that frequent drinking is more common among the middle and upper income groups than among the poor. In addition, infrequent drinking (less than one drink a week) or abstention is most common among the very poor, those with little education, and persons whose first language is neither English nor French (Wilkins, 1988). Rubin found that the lack of financial resources meant that leisure activities within the working class families included the children. Individual sporting activities, such as skiing and tennis, were much less popular among the working class, probably because of the associated financial cost, and low-cost sporting activities such as bowling were far 12 more popular (Rubin, 1976). Stressful Life Events. Several researchers have focused on the differences between middle class and lower class individuals' experiences of, and reactions to, stressful life events and living conditions, as well as the frequency of such events. Belle (1990) proposes that low income and minority women experience more frequent, threatening, and uncontrollable life events such as crime and violence, illness and death of children, imprisonment of husbands, lack of adequate housing, food, clothing, and childcare, living in dangerous neighbourhoods, and everyday discrimination based on gender, class status, and race or ethnicity. Their everyday living involves enduring significantly higher stressful living conditions than their middle class counterparts. Dohrenwend (1973) surveyed 124 heads of households to examine whether lower and working class families experienced more stressful life events than middle and upper class families. The sample included men and women from a variety of religious, racial and ethnic backgrounds. Participants were divided into two categories of social class: those who had not graduated from high school, and those who had graduated high school or more. Using a checklist of events, Dohrenwend (1973) found that lower social class members and women were exposed to significantly higher numbers of stressful life events that produced high rates of change and instability in their lives. Her findings also indicated a positive correlation between stressful life events and levels of psychological distress (Dohrenwend, 1973). Rubin (1976) found that, while the middle class participants in her study would generally fantasize about being a child again, the working class participants would not. In fact, the dominant memories of the working class individuals tended to be ones of both 13 emotional and material pain and deprivation. Even in the families where the recollections were of a loving and stable home, the working class adults recalled at least periods of unemployment, poverty, and deprivation. Health. According to Frank and Mustard (1994), in their review of research focusing on determinants of health, virtually all research investigating health and its relationship to social class has found that health status is directly related to social class. Health status is generally best among the higher income groups and worse among the lowest income groups. Moreover, they found that a social class gradient exists for almost every major disease studied world-wide. As one moves up or down the ladder of social class, both health and length of life increases with upward movement and decreases with downward movement. In a longitudinal study spanning almost twenty years, the Whitehall Study (Frank & Mustard, 1995) followed the health of over 10,000 British civil servants from four different occupational groups. These were: administrative, professional/executive, clerical, and labourers. Frank and Mustard (1995) note that what is significant in this study is that the obvious social gradient related to health was evidenced in a working population not faced with hazardous working environments, poverty, unemployment, or extreme affluence. Those lower down in the socio-economic hierarchy had a higher rate of death from coronary heart disease, strokes, gastro-intestinal diseases, smoking and non-smoking related cancers, accidents and suicides. The risk of heart attack was 2.5 times greater for labourers than for administrators. Age standardized mortality rates were 3 times higher for those at the bottom than at the top of the socio-economic ladder. While the study revealed that lifestyle issues such as smoking were strongly influenced by where 14 the individual was in the social hierarchy, and that those on the lower end of the scale were more likely to smoke, Marmot (as cited in Frank & Mustard, 1995) revealed that adjusting for factors such as cholesterol, smoking, and blood pressure reduced the social class gradient for coronary heart disease by only 25%. Moreover, the Whitehall study concluded that differences in medical care did not account for the startling difference in mortality linked to position within the work hierarchy. Through their review of the research, Frank and Mustard (1995) suggest that 1) the quality of support and interaction an individual has bears heavily on one's health and sense of well-being, and 2) that one of the dominant factors influencing health and well-being is an individual's sense of achievement, self-esteem, and control over one's work and life. This belief is also strongly supported by extensive reviews of health determinants world-wide carried out by the Canadian Public Health Association (1997). Moreover, this review (Canadian Public Health Association, 1997) and Mustard and Frank (1995) found that adverse physical (nutrition and housing) and psycho-social (social ties and social support) environments in early childhood have a significant negative effect on health, behaviour, competence and coping skills in adult life. Lower social class families are more frequently faced with a lack of adequate social support systems. While universal access to health care has existed in Canada since the early 1970s, the Canadian Public Health Association (1996) points out that it has had little impact on reducing health inequalities linked to socio-economic status. The Canada Council on Social Development (1996) points out that, among poor families in Canada, the incidence of low birth weight and premature birth is significantly 15 higher. They state that poor children are also at a greater risk of injuries, and are more likely to smoke and develop drug problems, than children from higher social class families. Results of a 1990 province-wide government study in Ontario (Canada Council on Social Development, 1996) found that teens in lower income families (less than $30,000 annually) were 1.8 times more likely to smoke than teens in higher income families (over $50,000 annually). Children from families in the lowest income levels are 1.7 times more likely than children in other families to have a psychiatric disorder, and die at a rate twice as high as children from families at the highest income level (Canada Council On Social Development, 1996). Lower income families are far more likely to live in over-crowded housing, experience more frequent and extended periods of unemployment, stressful family relationships and parental mental health issues. The inability to adequately provide for or improve future circumstances within the lower class families, in itself, produces increased levels of stress within the family. The ability to fulfil one's personal needs can create a sense of efficacy and well-being frequently lacking for many lower social class families. Self-Esteem and Self-Worth Several researchers (Belle, 1990; Rosenberg and Pearlin, 1978; Rubin, 1976) have proposed that many working class individuals feel little control over their lives, leading to anxiety, depression and a negative sense of self. As previously mentioned, while some members of the working class may hold jobs that pay well, most of these jobs tend to have little job security, may be dangerous, and frequently lead to health problems. Consequently, many working class individuals feel constantly on the edge, and do not feel as though they have succeeded, even when they are earning a higher wage (Rubin, 16 1976). In addition, most of the semi and unskilled blue collar jobs are significantly lacking in occupational autonomy, self-direction and job complexity, areas consistently found in research (Gecas & Seff, 1989; Sennett & Cobb, 1972) to be positively associated with self- esteem. Those individuals without jobs are even more profoundly affected by a feeling of inadequacy, given that they live within a society that sees success as an individual matter with all people having a roughly equal chance at the rewards (Bullock, 1995; Morris & Williamson, 1982; Rubin, 1976). Rosenberg and Pearlin (1978) analysed data collected from a sample of 1,988 school students in grades three through twelve in the Baltimore area and 2,300 Chicago area adult residents aged eighteen to sixty-five. Schools were randomly selected based on median income of census tract and proportion of non-white students. Students were then randomly selected from within the schools. Children's self-esteem was measured using a six-item Guttman scale, while adults' self-esteem was measured using the 10-point Rosenberg scale. Social class was measured using the Hollingshead Index of Social Position (Hollingshead & Redlich, 1958) which includes levels of education, occupation and income to determine social class. Correlations were then computed between social class and each of the items encompassed in the two scales for measuring self-esteem. Results showed no difference in levels of self-esteem among children linked to social class. However, the older the individual, the lower the self-esteem and this correlated positively to social class, with self-esteem being lowest among the lower social class adults. Although results were significant, correlations were low. Problems with this study include the fact that both scales used to measure self-esteem had low numbers of items. In addition, it is possible that the instruments 17 themselves may not be an accurate measure for assessing the construct of self-esteem. However, the study did provide interesting information about the possible long-term effects of social class standing. Gecas and Seff (1989), in attempting to further delineate the relationship between social class and self-esteem, carried out a phone study involving parents of high school students from five school districts within Eastern Washington. The sample population was a subset of a larger study that had administered questionnaires to high school students within the five districts. For the phone study, 173 (67%) fathers in intact families with adolescent children agreed to participate. The mean age of the fathers was 46, and approximately half held blue collar and half held white collar jobs. Subjects were asked to rate themselves on a 12-item semantic differential, which was comprised of adjectives descriptive of high to low levels of self efficacy and self worth. Results of the semantic differential were then correlated with social class, education, and responses to questions relating to work conditions such as job complexity, job control, job routinization, job supervision, and job prestige. Results showed a direct significant effect of level of education on self-evaluations and self-esteem. Results also showed significant indirect effects of social class on self evaluations through the occupational conditions of control over job, routinization, and complexity of work. They found that, overall, the effect was stronger for self-efficacy than for self-worth. They concluded that this was not surprising when one considers that the workplace is a major context for access to and expression of efficacy. In jobs involving little control over one's work, low levels of complexity, and routinization, it is predictable that self-esteem would be negatively affected. Values and Beliefs 18 The values one develops are affected by social class membership. Class differences have been noted in social and political beliefs and voting patterns, in the characteristics parents value in their children, in education, in sex roles, and in emotional expression (Johnston and Ornstein, 1985; Rubin, 1976). Social and political ideology. Several research studies have found significant differences in socio-political beliefs linked to social class. Johnston and Ornstein (1985) make note of various empirical studies which concluded that individuals from lower ranking occupations, less education and less income are more likely to support union movements, government economic intervention, increased spending on social welfare measures, and a general reduction in social inequality. They note that the research concludes that individuals from more privileged backgrounds, higher incomes, and higher occupational status are more likely to participate in politics and to support existing inequalities. These studies also found that the disadvantaged, on the other hand, were more likely to support social change, yet did not possess the political resources to bring about the equality. Johnston and Ornstein (1985), using data accumulated through a national Canadian sample survey involving 3300 personal interviews, attempted to systematically examine possible political ideological class differences. In their study, the social classes of each of the respondents was classified according to three differing although overlapping Marxist typologies and also by their occupation, occupational status, level of education, and income. The subjects were all asked to respond to questions dealing with workers' rights, social welfare issues, redistribution of income, distribution of power, involvement and 19 satisfaction with government, and the legitimacy of protest activities. Significant class differences were found towards trade unions, provision of social welfare benefits, and the redistribution of income on all three of the class measures in attitudes. Differences along class lines appeared stronger than using any other measure of stratification (e.g. income alone). The results also clearly showed that even though none of the class measures may be exact in themselves, and while the magnitude of the difference between the working class and the bourgeoisie (or middle class) may change depending on the issue, the working class (and lower class) remained consistently politically on the left in comparison with higher levels of class (Johnston & Ornstein, 1985). Interestingly, their research also found that additional schooling appeared to create results further to the right as level of education increased, although they had no explanation for this finding. While one might assume that higher levels of education tends to produce more liberal views of society and more support for societal change, Canadian studies indicate that increased levels of education appears to have a conforming effect in support for the dominant ideology and dominant institutions (Baer & Lambert, 1982; Johnston & Ornstein, 1985). Baer and Lambert (1982), using data collected from a Canadian sample of 3,288 persons all above the age of eighteen, found that the higher the levels of education: 1) the less the respondents supported spending on social services; 2) the less they felt there was too much of a difference between the rich and the poor; 3) the less they felt that the rich should pay a greater proportion of the tax bill; and 4) the less they felt that government has an obligation to provide jobs for the unemployed who want to work. The lack of positive support for changing the existing social inequalities gradually increased among 20 those individuals who had completed high school up to those holding doctorate and professional degrees. All results were highly significant at the .0001 level. Conversely, those individuals with less than high school completion (a variable of class) supported more funding for social services, more government sponsored jobs, higher taxes for the rich, and felt that the gap between the rich and poor was too great. Baer and Lambert (1982) also investigated the effects of parental occupation and education of the respondents and found no significant differences. The respondents' own level of education remained significant in every case. In observing possible differences based on language, Baer and Lambert (1982) also found that French Canadians showed more of a tendency to support egalitarian social changes, except for those individuals at the highest educational levels. Griffin and Okeneba-Sakyi (1993), in investigating causal attributions of poverty, found differences based on class background in the beliefs of the causes of poverty. Using a sample of 207 undergraduate students in Sociology at an American university, their analysis showed that, with all other independent variables held constant, those students coming from working class families were 40% less likely to attribute poverty to individuals' actions than those students coming from upper class families. Griffin and Okeneba-Sakyi (1993) suggest that further studies should investigate whether differences in attributional styles reflect socialization processes whereby attributions to individuals or groups are reinforced by learned schematic stereotyping. A limitation of their study is that working- class background university students are not necessarily representative of the working or lower classes. However, their research would appear to support previous studies that suggest more egalitarian and social effect leanings on the part of individuals 21 from the lower social classes. Parental Values. Rubin found that working class participants in her study tended to place greater value on those characteristics that reflect conformity, such as neatness and obedience to authority (Rubin, 1976). Conversely, middle class participants placed greater value on characteristics that reflect self-direction, such as curiosity and self-control. Thus, the higher the social class, the less value was placed on conformity and obedience, and the more was placed on independent thought and direction (Rubin, 1976). Kohn and Schoenbach (1993) analysed data from studies carried out in the United States, Poland, Ireland, Germany, Italy, Thailand, and Japan to explore whether certain parental values were limited to national or political boundaries. Earlier work by Kohn (1969), consisting of a national study in the United States involving a sample size of over 3,000, had revealed class differences in characteristics fathers valued in their children. The working and lower class participants had expressed a greater value for characteristics that reflected conformity, neatness and obedience. Middle class participants had placed greater value on characteristics reflecting self-direction and independence. Results of the international data analysis noted an underlying dimension of self-direction versus conformity in parental values. Further investigation of the studies in Poland, the United States and Japan noted a correlation of social class and parental values of self-direction versus conformity among both fathers and mothers for both boys and girls in the United States and Poland. Parents from the lower social class displayed a preference for conformity and obedience in their children, while the higher social classes preferred independence and self-direction in their children. In Japan, however, while the class-based correlations for boys resembled that of the United Stated and Poland, conformity was 22 considered a valued characteristic for girls in both social classes (Kohn & Schoenbach, 1993). A Canadian study by Pineo and Looker (1983), using a sample of 400 families from the Hamilton Regional Assessment lists, found similar patterns for higher values of self-direction among the white collar parents and lower values of conformity than among the blue collar parents. Pineo and Looker (1983) also noted that, on the average, Canadian parents tend to place a higher value on self-direction and a lower value on obedience to parents than do their American counterparts. In addition, Canadian parents tended to place a greater value on being a "good student." Pineo and Looker speculate that Canadians are more willing to cede authority over their children to outside agencies such as schools, and therefore see the responsibility for control of the child less exclusively that of the parents. While there are evident differences within the two countries on specific values, the overall tendency remains the same. The Canadian blue collar parents, although not as dramatically as the American blue collar parents, endorsed higher levels of conformity than the white collar parents. As Rubin (1976) points out, working class individuals usually work in jobs where the values of conformity to the rules (being paid to work, not think) far outweighs the benefits of stepping out of line. In addition, jobs for the working class tend to involve little freedom for creativity and individuality. Sennett and Cobb (1972) point out that the limited freedom on the job can contribute to a sense of inadequacy and low self-esteem. Middle class careers, on the other hand tend to place greater value on individual thought and expression, which can enhance one's sense of self-value and dignity. 23 Education. Values regarding education also reveal some significant class differences. Rubin (1976) and Laureau (1987) found that working class parents felt that school should be strict, while middle class parents believed that it should be enjoyable and loose. Rubin (1976) suggests two alternatives for the class differences. On the one hand, the working class parents' own feelings of failure or inadequacy may stimulate fears of failure for their children if the children do not learn important academic skills. On the other hand, the working class parents may also have high fears that encouraging creativity and independence in their children will not provide them with the necessary skills to make it in the working class world, where conformity and following orders are the keys to any degree of success. Neither Laureau (1987) nor Rubin (1976) found any class difference in parental desire for their children to succeed. Both working-class and middle-class parents equally valued educational success, and wanted their children to do well in school. Laureau (1987) also found that both working-class and middle-class parents all saw themselves as supportive and helping their children to succeed educationally. However, the ways in which they promoted educational success differed. Working-class parents were much more likely to turn over the responsibility for their children's education to the school, while middle-class parents tended to see their children's educational as more of a collaborative and interactive effort. Laureau (1987) states that the lower-class parents' lower levels of education, lower occupational prestige with the teachers, and limited time and financial resources produces a situation within which lower-class parents are more apt to feel that the educational success of their children is better left in the hands of the professionals. 24 Sex Roles. Depictions of working class families in the media have typically portrayed working class men as macho and demanding, with relatively subservient wives (e.g. Archie Bunker; Homer Simpson). Rubin (1976), in her exploration of class differences in sex roles, found that middle class couples, while asserting an ideology of equality between the partners, were no more egalitarian in their behaviour than the working class couples in her study. In working class couples, the women openly used language acknowledging the power and authority of the husband. In contrast, the middle class women used language suggesting equality, yet the power differential within the relationships remained evident in behaviour. Rubin (1976) hypothesized that since middle class men have more power and authority in the work place than do working class men, their tendency may be to exhibit a less overtly authoritarian role within the family. Conversely, since working class men have relatively little power on the job, they may tend to fulfil their traditional male role expectations by asserting more overt power within the family. Emotional Expression. Several researchers have proposed that the value and emphasis on emotional expression varies based on class membership. A qualitative study carried out among a group of working class women from a Toronto Rape Crisis Centre revealed that emotional expressiveness was taught and valued in their families (Ignagni et al., 1988). By comparison, middle class individuals are taught to operate from the head, and to talk about feelings rather than expressing them. The working class women also reported that their expression of strong personal emotions such as anger or frustration made many middle class women uncomfortable (Ignagni et al., 1988). Both Cardea (1985) and Hughes (1987), in their personal narratives of their 25 experiences as white working class women, describe their emotional expression of anger as being socially unacceptable to middle class individuals and institutions. Both felt that, while the middle class may emphasize emotional awareness, the working class value emotional expression (Cardea, 1985; Hughes, 1987). However, Hertzberg and Eschbach (1982) and Hertzberg (1989) propose that working class individuals are less likely to express strong emotions. They suggest that working class individuals survive the experiences of trauma, exploitation and dehumanization almost inherent to living in poverty by remaining frequently disconnected from their emotions. They maintain that being "in touch" with one's feelings presupposes a level of self-acceptance frequently not available for individuals who are societally labelled as inferior or unacceptable, and whose life experiences differ from traditional values and ideals. While being out of touch with one's emotions would frequently be labelled as dysfunctional, Hertzberg and Eschbach (1982) contend that not being in touch with feelings may be a functional survival skill for individuals recurrently experiencing oppression. As the above review indicates, class status and background can influence not only one's experiences, but also ones choices, values and one's beliefs about self and others. In the next section, I will attempt to explore conceptual factors and their possible impact on how lower socio-economic status clients are perceived, interpreted and evaluated. 26 The Schematic Model Schemas In examining possible counsellor bias based on differences in social class, it is important to consider the cognitive processes that may be involved. Our understanding of people, situations and events does not begin anew with each new instance we might encounter. Each time we encounter new people and situations, we bring with us "our past experiences, knowledge, beliefs, and feelings about similar situations and people" (Crocker et al., 1984, p. 197). In order to understand both oneself and others, information must be acquired, interpreted, and accessible (Leyens, Yzerbyt, & Schadron, 1994). One representation of how individuals acquire, store and use information is in the form of cognitive structures called schemas. A schema is an abstract cognitive knowledge structure, stored in memory, that represents an individual's general knowledge about a particular 'domain stimulus,' including the defining features, relevant attributes, and interrelations among the attributes (Crocker et al., 1988; Fiske & Taylor, 1984). Schemas help us to interpret, organize, and integrate new information. Through the process of categorization, schemas simplify the encoding, storage and retrieval of relevant information, and shorten the time it takes to both process information, and to solve problems (Crocker et al., 1984). Schemas also function as an information filter in that information considered as irrelevant may be ignored or forgotten, while information interpreted as relevant to a particular schema may be integrated. When a perceiver is faced with information incongruent to his or her individual 27 schemata, the response may be one of assimilating the new information to fit within the schema, overlooking the new information, or accommodating the new information through "modifying or altering the schema in response to the demands of the environment" (Crocker et al., 1984, p. 198). Individuals may possess well developed schemas for particular situations, individuals or groups, or no schema at all, depending on the situation and the past knowledge and experience of the particular perceiver. Maturing schemas evolve and change with increasing experiences to which the perceiver may be exposed. The more information is acquired, the more developed the individual's schematic structure may become (Crocker, Fiske, & Taylor, 1984). While schemas may change through exposure to information that does not fit the particular schema that the perceiver or group possesses, a remarkable number of studies have pointed to the fact that although schemas do change, they can also be highly resistant to change (Crocker et al., 1988). In fact, Crocker et al. (1988) have noted that schemas can frequently "bias and distort the encoding, retention, and retrieval of schema-relevant information..." (Crocker et al., 1988, p. 198). Categorization. Integral to the understanding of the schematic model is the concept of categorization. Different pieces of incoming information may be integrated to form a category; subsequent category cues then activate an existing schema about a specific kind of person (e.g. politician), group of people (e.g. women) or situation (e.g. birthday party). Categorization simplifies the cognitive process by summarizing all of the features contained in one category under a particular title or name (Leyens et al., 1994), shortening the amount of time needed to understand or interpret the object of perception, 28 as well as providing meaning and familiarity to it. For instance, if we hear that a particular person is a long-distance runner, we have a certain idea of aspects or attributes entailed within the category and we have certain expectations of that person. We may assume the person has developed stamina, exercises regularly, and probably eats well. This constitutes our schema category of "long-distance runner." Categorizing provides us with a means of managing a myriad of information. It helps to make the environment more understandable and predictable by allowing the integration of new information with older information. In addition, when gaps occur within our schematic structure, inferences drawn from the schema are used to fill in the missing information. Contradictory or ambiguous information may be interpreted in ways that allow for its integration within the existing schema (Leyens et al., 1994). Stereotypes Stereotypes may be seen as a particular type of social role schema that organizes one's knowledge and expectations about other people who fall within certain socially defined categories. Stereotypes involve socially shared beliefs about personality traits and behaviours of specific groups of people (Fiske & Taylor, 1984; Leyens, Yzerbyt, & Schadron, 1994). As with schemas, stereotypes help simplify and organize information, and assist in the sorting and remembering of details (Leyens, Yzerbyt, & Schadron, 1994). While cognitive explanations of stereotyping stress the important function of stereotyping in thinking and communication, stereotyping holds a negative connotation (Leyens, Yzerbyt, & Schadron, 1994). Stereotyping groups and individuals can lead to perceptual distortions of them because there is no assurance that the content of the schema is accurate. Implicit in the concept of stereotype is the fact that certain groups of 29 individuals are negatively interpreted and evaluated. By stereotyping individuals, we perceive and judge an individual as interchangeable with other members of a particular category or stereotype. The individual is then expected to have certain attributes, behaviours, ,and standards based on their membership within the category, and individual differences can be overlooked or ignored (Leyens, Yzerbyt, & Schadron, 1994). Stereotypes frequently favour one group over another (Fiske & Taylor, 1984; Jones, 1988). In North American and European society, men are generally favoured over women, Whites over Aboriginals, Blacks, Hispanics, and other racial and ethnic groups, and the middle and upper social classes over the lower social classes. Expectations that members of certain groups will be more successful due to internally based attributes are significantly higher for the members of the favoured groups (Jones, 1988; Morris & Williamson, 1988). For instance, studies have shown that when a man and a woman perform exactly the same behaviour, a man's performance may be evaluated more favourably than a woman's (Foshi, Lai & Sigerson, 1994). The same is true of evaluations of individual capacity for change and levels of mental health based on membership in a particular socio-economic group (Frank, 1994; Sutton & Kessler, 1986). Stereotyping also impacts on the value assigned to specific tasks. Tasks viewed as appropriate for men are more favourably valued and perceived as more difficult than tasks usually designated as appropriate for women. This same upgrading and downgrading occurs for class related job occupations as well (Morris & Williamson, 1988; Rubin, 1976). Causal attributions for performance are also important and may be associated with a stereotype. The expectations and perceptions of favoured or "in-groups" are more 3 0 positive than those of unfavoured or "out-groups." When individuals behave according to expectations, their behaviour is attributed to stable or internal causes; when individuals behave inconsistently with prior expectations, their behaviour is attributed to unstable or external causes (Fiske and Taylor, 1984; Morrow & Deidan, 1992). Consequently, when favoured or in-group members succeed at tasks, their behaviour will be attributed to internal, stable causes, while when out-group members succeed, the behaviour will more likely be attributed to external causes, or the individual's behaviour may be perceived as exceptional and not representative of the group's usual behaviour (Fiske & Taylor, 1884; Morris & Williamson, 1988). Causal attributions can reinforce negative stereotypes by offering explanations of both stereotypic and counter-stereotypic behaviours that match the original stereotypic expectation. Behaviours which fall outside of that which is considered usual does not necessarily change the stereotype, but is simply perceived as an unstable or external event (Fiske & Taylor, 1984). Studies have shown that perceptions of out-groups show much less variability and complexity than perceptions of in-groups. One's own multiplicity of experiences and interactions with members of one's own groups produces a more varied view and understanding of those groups. Conversely, fewer interactions with members of out-groups allow for perception of less variability and complexity. Because out-groups are perceived as less variable and complex, individuals tend to make inferences about out-group members based on little information. When a person is perceived as belonging to an out-group schema, the fit of the individual within the stereotype of the out-group is seen as being particularly tight (Fiske & Taylor, 1984). 31 In addition to variability and complexity, categorizing someone as an instance of a particular schema slants perception of what the person does. An analogue study by Sutton and Kessler (1986) found that the behaviour of lower SES (socio-economic status) clients was evaluated as more dysfunctional than the same behaviour of middle SES clients (Sutton, 1986). Perception of out-group members' behaviour is not only less variable and less complex, but more negatively perceived than the same behaviour by in-group members. (Fiske, 1984; Sutton, 1986). Stereotyping also tends to shape the perceivers' recollection of the individual's past behaviour and background. Fiske and Taylor (1984) describe a study in which a later revelation that a particular person was gay caused the perceivers to remodel and reconstruct memory so that factors were interpreted as leading up to the person's sexual orientation, and new factors were perceived that had not been obvious before. The retrieved information was moulded to fit the socially prescribed stereotype of the out-group member. Research (Devine, 1989) has shown that schema-consistent information is remembered and favoured over schema-inconsistent information. Devine (1989), in a series of three studies, examined the effects of automatic and controlled components of cultural stereotyping. Using a commonly known stereotype in the United States, that Blacks are aggressive, Devine found that stereotypes are automatically activated in the presence of members of the specific target group. Even those White subjects who had been classified as 'non-prejudiced,' through their responses to the Modern Racism Scale, evaluated behaviours of others in stereotype-congruent ways when they were not capable of consciously monitoring the activated stereotype (Devine, 1989). Devine (1989) 32 suggests that, since individuals are frequently exposed to negative stereotyping, they may also be involuntarily exposed to automatically activated stereotypes, without conscious awareness, and that conscious personal monitoring may be a necessary component of counteracting negative stereotyping. Lower Class Stereotypes Leahy (1981), in a study of children ranging in age from 5-18 found that with increasing age there was a significant tendency to view the rich and poor as differing both in physical characteristics as well as in terms of being different kinds of people. As age increased, the descriptions increasingly involved perceiving the poor as lacking in ability and lacking in effort to change their circumstances. Leahy (1981) also noted that upper-middle-class children were more likely to describe the poor as possessing negative personality traits, while poor children were more likely to describe poor people as worrying about having very little money. One of the earliest studies (Gough & Heilbrun, 1980) using the Adjective Check List to investigate the subjects of both class and race, asked black and white American college students to select, from a list of 85 adjectives, the ones they believed most descriptive of upper-class white Americans, lower-class white Americans, upper-class black Americans and lower-class black Americans. The most notable finding of the study was that the social class of the group influenced the choice of descriptors much more so than did race. The most frequently chosen adjectives for both upper-class white and black Americans, chosen by both white and black students were: intelligent, ambitious, industrious, neat and progressive. Both the white and black lower-class groups were 33 described as lazy, loud, ignorant and dirty. No specific traits emerged to describe the groups as a whole based on race (Bayton, McAlister, & Hamer, 1956, cited in Morris and Williamson, 1982; Gough & Heilbrun, 1980). Morris and Williamson (1982) provide a review of contemporary images of the lower class and poor that continue to reflect negative stereotypes based on social class, occupation, and education. The poor are frequently portrayed as being destitute because they share and pass on to their children a set of defective behaviours, values and personality traits. These supposed traits include laziness, inability to delay gratification, lack of respect for and interest in education, unwillingness to work, dishonesty, and sexual promiscuity combined with a disinterest in or ignorance about birth control (Ehrenreich, 1987, 1990; Sullivan, 1989). Bullock (1995), in her review of the research on social class, found that middle-class perceptions of welfare recipients included that they are typically dishonest, dependent, lazy, disinterested in education, and promiscuous. Single-parent women on welfare were typically stereotyped as devaluing the two-parent family, as having larger families in order to collect more welfare, and as being sexually promiscuous. The stereotype of the working-class man, typified in television series such as Archie Bunker, and more recently, Homer Simpson, portrays them as patriarchal, aggressive, racist, dishonest and slightly stupid. Common themes in movies tend to emphasize story lines suggesting love conquering all in poor girl/rich boy stories such as Pretty Woman, or rags to riches story lines such as in the television series The Fresh Prince of Bel-Air. A plethora of movies and television series suggest that many individuals succeed through intelligence, persistence and hard work, suggesting that the 34 lack of these qualities causes the resultant poverty among the lower class (Bullock, 1995). Few movies or sitcoms ever tend to portray the external causes of poverty and class systems such as lack of opportunity, an unequal education system, lack of access to material, financial and human resources, lack of job opportunities, lack of political power, lack of understanding of social class, and various forms of bias. North America has tended to view itself, until recently, as a relatively classless society. We are rarely encouraged to define or articulate differences in terms of class. The opportunity for less than a conscious awareness of class and the implicit stereotyping that goes on can allow for members of the lower class to simply be seen in derogatory terms such as ignorant, lazy, impulsive or loud. Over the past two decades, much effort has gone into the recognition of biases based on a person's gender, race or ethnicity. In a society that frequently denies the existence of a class structure, class bias may be a much more insidious problem. Differences from the middle class in-group are frequently viewed as individual deviances from the norm or individual failure. While the stereotype may automatically arise in the presence of a lower socio-economic status individual, the perceiver may not necessarily be aware that he or she is perceiving, interpreting and evaluating the individual through the biased stereotype of lower class. This may be equally true of counsellors, who are predominantly members of the middle and upper classes (McCarthy et al., 1991). The Schematic Model Applied to Counselling In applying the schematic model to a counselling setting involving lower SES clients, it seems apparent that the possibility of errors of perception, interpretation and evaluation are high. Most counselling therapists do not come from a lower class 35 background. Schematic stereotypes appear to be automatically activated in the presence of out-group or target members (Devine, 1989). Stereotypes tend to contain simplified and at times erroneous perceptions and interpretations of out-group members. The fit between the individual members and the stereotype of the group is frequently seen as being particularly tight, and schemas are resistant to change (Crocker et al., 1988). If, in addition to this, the counsellor has not developed an awareness of actual class differences and the life experiences of lower socio-economic status clients, the lower socio-economic status client may be entering a situation in which he or she will continue to be negatively perceived and evaluated (Hertzberg, 1990; Hertzberg & Eschbach, 1982). As previously noted, the typical stereotype of the poor in North America includes attributes such as irresponsible, lazy, uncontrolled, erratic, and an inability to delay gratification, etc. (Morris & Williamson, 1988; Muller, 1993). Newspapers and the media frequently portray the poor as welfare scam artists, or people who feel that society owes them a living (Morris & Williamson, 1988). It would seem illogical to believe that counsellors are immune to the societal stereotyping involving the lower social classes. The North American Work Ethic holds each individual responsible for his or her success or failure. This societal ideology implies that those who work hard will be rewarded with success, and that failure to reach one's goals and aspirations is frequently a result of personal deficiencies (Morris & Williamson, 1988). Traditional psychotherapy has focused more on individual responsibility for one's life, rather than including the importance of social context. This bias not only includes initial stereotypic reactions to clients, but can also influence the counsellor's choice of explanatory hypotheses as to the presenting problem and issue at hand, the direction of 36 therapy, the questions asked and unasked, and the counsellor's evaluation as to the appropriateness and success of the client's treatment (Smith, 1988; Turk & Salovey, 1988). Clinicians employ the same method of cognitive processing as other individuals. Initial intake contacts with a particular client, including presenting problem and physical appearance, may trigger schemas categorizing the client as a member of a particular out-group (Smith, 1988). Studies show that, at intake, lower SES clients have been frequently deemed less acceptable for therapy and are more likely referred for psychotropic medication than middle and upper class clients (Garfield, 1986). In assigning responsibility for events, counsellors are more likely to attribute an individual's problems and levels of distress to internally based personality traits, while the client may attribute his/her actions and distress to external factors. As pointed out by Morrow and Deidan (1992), "This bias may be especially potent for female and minority clients who often have legitimate claims of discrimination and may often face significant external stressors." The disagreement in origin of the problem may, whether verbalized or not, produce a situation where the client is negatively evaluated. In addition, a lower class client, feeling that the professional is always right, may begin to believe that he or she is responsible for all of the problems that may beset him or her, producing increased levels of guilt, stress and low self-esteem (Hertzberg, 1990; Morrow and Deidan, 1992). Stereotypes of the lower social classes may produce biased counsellor perceptions of a client, while excluding individuating information. As pointed out by Fiske and Taylor (1984), once an individual is categorized into a particular category, features attributed to the stereotype may be attributed to the client, new client information and actual client 37 attributes may be disregarded or overlooked, and category-consistent information may become salient. Research (Fiske & Taylor, 1984) shows that perceivers reconstruct reality in ways that support current beliefs about others. Since lower SES has been linked to higher levels of psychiatric disorder, it may be that attributes expressed by the client that fit certain disorders may also receive greater attention based on the client's SES. Studies have indicated that lower SES clients are categorized as having higher levels of dysfunction than middle class clients even when the categorization is based on the exact same behaviours (Frank, 1992; Sutton, 1986). Lower social class clients may appear distinctive in manner of dress, language, differences in levels of and extent of emotional expression, or choice of problem focus. Viewing the lower class client as distinct (i.e., different) may produce a greater focus of attention to expressions of attributes considered part of a lower social class client stereotype, and a lack of focus on expressions of attributes that do not fit the stereotype. Research investigating hypothesis testing has found that once a counsellor has adopted a particular hypothesis about a client, there is a significant tendency to perceive confirmatory information, while disconfirmatory information is overlooked or ignored. While previous research outcomes suggested that counsellors are generally unbiased in their questioning strategies to test client hypotheses, recent studies have found a strong confirmatory bias when investigating counsellor-generated hypotheses (Haverkamp, 1993; Strohmer et al., 1990), and no support for the previous finding that counsellors tended more towards the employment of a neutral hypothesis testing strategy when testing client-generated hypotheses (Haverkamp, 1993). As pointed out by Haverkamp (1993), previous testing for confirmatory bias had presented subjects with limited optional-choices 38 for hypotheses. Results of a strong confirmatory bias when self-generated hypotheses are available, may be reflecting an influence of pre-existing beliefs on counsellor choices and questioning patterns. Strohmer et al. (1990) noted a strong retrospective confirmatory bias in hypothesis testing. Even when client reports contained more disconfirmatory than confirmatory information, experienced counsellors remembered more of the hypothesis confirmatory information. Initial examination of research on schematic functioning, stereotyping and counselling suggests that there is ample room for bias to enter into the therapeutic interaction and process. Therefore in the final section of this literature review, I will focus attention on research that has specifically examined possible counsellor bias linked to client socio-economic status. Class and Psychotherapy Numerous studies carried out in both the United States and Canada have demonstrated higher rates of psychological disorders among the lower social classes (Dorenwend, 1973; Lorion & Felner, 1985; Wilkins, 1988). Reviews of the research also indicate a significant class bias in psychotherapeutic diagnosis and treatment (Garfield, 1986). Many of these studies were carried out in the 1960s and 1970s, when some attention was directed towards socio-economic status as an influencing factor in therapy. Little if any research has shown progress in eliminating class bias in therapy and, since the mid-seventies, research and education focused on socio-economic status has been minimal. Garfield (1986) and Lorion and Felner (1986) have suggested several areas where counsellor bias related to client socio-economic status might influence the therapeutic 39 process. These areas are: client acceptance for therapy, length of or continuation in therapy, counsellor perceptions and expectations of clients, and counsellor interpretation and evaluation of client levels of adaptive functioning, presenting problem and therapeutic outcome. Acceptance for Therapy Garfield (1986) reports that the overwhelming majority of the studies that focus on who gets accepted for therapy show, to varying degrees, a social class bias against acceptance of lower class clients. Garfield found that the more expert the therapeutic staff, the more selective the procedure for acceptance tended to be. Generally, this meant that clients who were more highly educated, perceived as more intelligent, more verbal and more motivated, were more frequently accepted for therapy. Kadushin (1969), in a study of the different kinds of clinics in New York City, concluded that social class was the most important feature distinguishing the applicants to the various clinics, and the more traditionally minded (orthodox psychoanalytic movement) a clinic was, the higher the social class of its accepted applicants. Hollingshead and Redlich (1958), in their classic New Haven study of social class and psychotherapy, found a significant difference in the services offered to clients based on social class. Long-term psychoanalytic treatment was offered mainly to middle and upper class clients, while more inpatient and drug therapy was offered to lower class clients. This held true even in state-supported mental health clinics offering reduced rates for treatment. While Garfield (1986) noted a tendency for therapists to perceive lower socio-economic status clients as more apt to refuse therapy when offered, Garfield found only one study supporting such a belief, and concludes that, because of the lack of research in 40 this area, results are inconclusive. Continuation in Therapy Garfield (1986) reports that while many studies report a positive relationship between client education and length of stay in therapy, several reports have shown no difference at all. Garfield found some evidence that suggests that the lack of agreement between studies may be due partially to study factors such as pre-screening and acceptance procedures and the administration of drugs along with treatment. Garfield points out that, 1) where screening procedures are most stringent, dropout rates tend to be lower, but the samples tend also to be biased in favour of better educated clients; and 2) when administration of drugs is concurrent with therapy, dropout rates for all socio-economic classes are lower. Several literature reviews have noted that dropout rates are significantly higher for lower class clients than for middle class clients. Berrigan and Garfield (1981) found no correlation based on age or sex of the clients, but a significant correlation between socio-economic status (using the Hollingshead Index of Social Position) and premature termination. They found increasing proportions of dropouts as social class decreased, with a range of zero in Class I (highest social class) to 50% dropout rates in Class 5 (lowest social class). Furthermore, Class V patients tended to miss more scheduled appointments than any other socio-economic group. Edward Beckham (1992), in an attempt to identify characteristics of dropouts, examined four possible predictor variables. These were 1) patient attitudes about the nature of psychological disorders and psychotherapy; 2) practical difficulties for patients attending therapy; 3) patient's views of the therapist's ability to relate to the patient, and 4) 41 social class. The study involved 93 applicants accepted for treatment at a mental health clinic. Applicants filled out intake questionnaires, Cause of Illness Inventories, and the Barrett-Lennard Relationship Inventory (which measures the patient's initial impression of the therapeutic relationship). The only significant differences found were that family income was higher for the non-dropouts (perhaps indicating less practical problems usually associated with higher levels of dropouts), and a tendency for dropouts to have initially seen their therapists as less warm, empathic and genuine. This would appear to show that the patient's first impression of the therapist is very important. While Beckham makes no link between lower family income levels and initial impression of therapist's ability to relate, it seems possible that these may be connected even though general scores on socio-economic status and early termination were not. Lower class individuals experience significantly higher numbers of stressors in their lives than the middle class, and frequently on a daily basis (Herr, 1989). The prevalence of psychological disorders is highest in the most socio-economically disadvantaged segments of our society (Lorion & Felner, 1985). Yet, research shows that the average number of sessions for both male and female working class clients is much lower than that of middle class clients (Hollingshead and Redlich, 1958; Garfield, 1986). In fact, McCarthy et al. (1991) state that the modal number of visits for working class individuals is one. Another study (Pettit, Pettit, & Welkowitz, 1974) investigating the relationship between values, social class and duration of therapy, found a relationship between the interaction of social class and the discrepancy between patient and therapist values and continuation. Pettit et al. (1974), using a sample of 249 clients at a moderate cost outpatient clinic, found a significant relationship between social class and all six factors of 42 personal values produced through the results of four separate values tests given to each client. On one particular factor (Authoritarian-Submissive versus Independent), high scores on 'respectful of and submissive to authority' on the part of the lower SES clients in combination with low scores on 'respect for independence' on the part of the therapist correlated with continuation in therapy; the greater the discrepancy, the longer the duration of therapy. No such discrepancy existed for middle and upper SES clients in relationship to continuation. Pettit et al. (1974) found that when the interaction of this particular factor was taken into account, there was a social class effect on continuation, with lower SES clients remaining in therapy for shorter durations. In another study (Brill & Storrow, 1960), where acceptance for psychotherapy related positively to social class status, clients were also evaluated as to their 'Psychological Mindedness' (understanding of and readiness for therapy). Low social class was found to be significantly related to low estimated intelligence, a tendency to view the problem as physical rather than emotional, a desire for symptomatic relief, lack of understanding of the psychotherapeutic process, and a lack of desire for psychotherapy. It is important to remember that these evaluations were not client self-evaluations but intake interviewer evaluations. In addition, the intake interviewer had less positive feelings for lower-class patients and saw them as less treatable by means of psychotherapy. Brill and Strorrow conclude by stating that these findings seem to suggest some form of interaction between the attributes and expectations of lower class clients, and the attitudes and interpretations of middle-class therapists. While they do not go so far as to strongly suggest therapist bias, they indicated that it is a significant possibility. Perceptions and Expectations 43 It has been shown that lower class clients are more likely to be perceived as emotionally disturbed than are middle class clients (Hollingshead & Redlich, 1958; Lerner, 1973; Sutton & Kessler, 1986). Hollingshead and Redlich's study (1958) evaluated lower socio-economic status patients to be more unreliable, impulsive and having difficulty delaying gratification. Therapists also sensed a lack of rapport with low-income patients, and perceived them to be hostile, suspicious and uncertain as to the appropriateness or efficacy of therapy. Lorion and Felner (1985) also mention that more recent studies continue to make these same attributions to lower SES clients. As pointed out by Lorion (1974) and Lorion and Felner (1985), such noted attributes are not based on direct assessment, but on subjective therapist assessments from inferred client behaviour. Lorion (1974) carried out an investigation of treatment attitudes and expectations based on client social class. His sample included 90 applicants for outpatient treatment divided into six subgroups of white Class III, IV and V males and females matched on age, marital status and religion. Results on the Mental Health Attitudes Survey (used to measure attitudes towards seeking help), and the Treatment Expectations Survey, showed no significant effects based on socio-economic status or sex. Lower SES applicants showed no more negative pre-treatment attitudes or expectations than their middle SES counterparts. This being the case, it would appear that direct responses from applicants and clients show no more obvious negative attitudes than higher SES applicants. Frank, Eisenthal and Lazare (1978), in their study and review, were critical of the view that lower class individuals have treatment expectations that are inappropriate and different from those of other patients. Using an 84-item questionnaire, the researchers 44 obtained responses from 278 walk-in-clinic patients and found no marked social class differences in the kinds of requests for help desired. Class I through IV showed no differences at all while Class V showed only nonsignificant differences on such items as requests for social interventions, administrative help and psychological expertise. They concluded that social class differences in treatment disposition and outcome might be reflecting the attitudes and beliefs of middle-class therapists rather than the attributes or behaviour of lower class patients. Resistance. Presumed unreliability (Hollingshead & Redlich, 1958) or, as defined by the more psychodynamically minded, 'resistance to therapy' is supposedly witnessed in lower SES clients in higher frequencies of missed sessions, arriving frequently late, or bringing young children into the sessions. However, as explained by several researchers (Garfield, 1986, Lorion & Felner, 1985; McCarthy et. al, 1991) lower socio-economic status individuals are more likely to arrive late or miss sessions due to lack of transportation, unreliable transportation, or no money for transportation. They may bring their children into sessions because they cannot afford day care or because the care they can afford is unreliable. Women coming to counselling may miss sessions or drop out for fear of being beaten for sharing information that a partner feels is threatening to him. Families receiving counselling through government run programs may fear revealing things for fear of losing custody of their children. When factors such as these combine, a therapist may assume a client is resistant or unreliable when, in fact, the social context of his/her life puts constraints on his/her ability to be available and open up to the same extent as more fortunate individuals. Both Brill and Strorrow (1960) and Lorion and Felner (1985) found that lower 45 SES clients desired somewhat more direct intervention on the part of the therapist, and therapy aimed at symptom relief. Hollingshead and Redlich (1958) also felt that lower SES clients were disappointed in not receiving direct advice and guidance with their problems. As pointed out by Lorion and Parron (1985 in Lorion and Felner, 1986), this is not surprising considering the contexts of the lives of many lower SES clients: Poverty, if nothing else, is characterized by an endless series of unpredictable crises which demand constantly reallocating meagre resources from one necessity to another. Everything one owns is already at risk for needing repair, credit, when available, is expensive and short-term, and savings must always be diverted from what is desired to what is needed. The therapist who understands the daily decision-making necessary to survive economic hardship is likely to appreciate both the client's resourcefulness and insistence that treatment be quick and symptoms be addressed immediately, (p. 83-84) What may be viewed as impulsivity, or difficulty in delaying gratification, may be the reflection of a realistic need to get things done quickly in order to deal with the daily life crises frequently the norm for individuals living in the lower socio-economic classes. In fact, Lorion and Felner (1985) suggest that the practical problems inherent to living in poverty may outweigh all other factors in accounting for high dropout rates for lower SES clients. Testing Bias. There is an abundance of research that shows that growing up in the lower classes affects how people perform on tests of abstract thinking, tests of intelligence, and tests of academic achievement (Frank, 1994). The norms established to evaluate people's performance on psychological tests are frequently developed through 46 pretesting on sample middle class or unidentified mixed classes of individuals. Ehrenreich (1990), in his research on the possible effects of social class on normative responses to TAT cards, points out that while the research hints to possible differences in the ways middle class and lower class populations may respond, individuals are evaluated against the standard norms which are frequently middle-class based. Ehrenreich's subject population included 70 second-year community college students, 35 of who were middle class and 35 of who were lower class. Class was determined using both the subject's own class standing and the family of origin standing. Five TAT cards were shown on screen for 20 seconds each. The subjects were then given four minutes to write stories on each card. The stories were scored for numbers and intensity of drive expressions, levels of defences used, patterns of dependency, and locus of control. Significant differences were found in two areas, those of locus of control and dependency. The working class subjects more frequently portrayed the principal character as having an external locus of control, while the middle class subjects more frequently described the principal character as having an internal locus of control. The working class subjects were also more likely to provide stories reflecting themes of dependency. Ehrenreich (1990) notes that, by aggregating all of the stories written, the working class subjects showed a slightly higher tendency to portray bad objects rather than good objects in their stories, and portray stories with more themes of danger than those of safety. While this would produce different normative responses than middle class populations, it would appear quite normal to expect these results as lower class individuals do experience more stress and crises in their daily lives. 47 Ehrenreich (1990) points out that his research is limited in that his subjects are all female college students. Lower class college students are not necessarily representative of the lower class population, and males may not necessarily react in the same fashion as females to the TAT cards. Frank (1994) has demonstrated differences in people's responsiveness to the Rorschach (a test to assist in determining psychopathology) based on socio-economic status. However, research has also demonstrated that the same Rorschach protocol has been evaluated as reflecting greater levels of psychopathology when the clinician believes the respondent is an individual from a lower socio-economic status than when the clinician believes the individual is from a higher socio-economic status. Since this again appears to be clinician bias, it would be important to investigate what are acceptable healthy responses to a variety of psychological tests in order to determine whether responses are evaluated based on class held beliefs about what is health and what is illness, or whether the tests actually measure objective criteria for healthy functioning. Frank (1994) suggests further investigation of the Rorschach based on socio-economic status. This suggestion would do well to be applied to much of the personality and psychological testing presently available. Adaptive Functioning, Presenting Problem, and Outcome Eric Hillerbrand (1988), in a study including 163 randomly selected counselling files of university student clients which were divided into three class groupings (upper, middle, and lower class), and 21 counsellors who were either counselling psychologists, interns, or counselling psychology practicum students, found no significant differences among the clients in presence of clinical syndromes or client symptoms, level of 48 psychosocial stress, number of sessions attended, cancellations and no-shows, and premature termination. Some of these results may be partially due to the fact that student counselling centres tend to focus on short-term goal focused therapy rather than long-term psychoanalysis. In addition, because only three class groupings were used, working class background students were categorized into either the middle class or the lower class. A more adequate grouping would have been to provide clearer class distinctions allowing for a working class grouping. Furthermore lower socio-economic status university students tend to be a select group and are not necessarily a representative sample from the lower class. Results did show, however, that lower SES student-clients were rated on intake as more dysfunctional than middle and upper SES student-clients. A significantly higher proportion of the middle and upper SES clients were also rated by their counsellors as having reached their counselling goals at termination in comparison with the lower SES clients. In addition, concordance rates between the client and intake interviewer for identification of presenting problems were much lower for lower SES clients than for the middle and upper SES clients, with higher proportions of the lower class students perceiving their problems to be more vocational than personal. These three ratings (adaptive functioning, presenting problem, and outcome) were all subjectively based intake interviewer assessments. The only significant differences in this particular study, then, lay in areas involving subjective judgement and interpretation on the part of the intake interviewer. This would seem to support much of the research investigating errors of perception, interpretation and evaluation carried out by social cognition researchers (Devine, 1989; Crocker et al., 1988; Haverkamp, 1993; Strohmer et 49 al., 1990). A review carried out by Beutler (1981) found a consistent and strong relationship between similarity of clients' and therapists' beliefs and values and subsequent improvement in therapy. Therapists appeared to be more comfortable with clients who spoke the same language and were more similar to them. This supports social cognition research that suggests more positive evaluations and higher levels of acceptance of in-group members (Fiske & Taylor, 1984; Jones, 1988). Evans, Acosta, Yamamoto and Skilbeck (1984) evaluated 29 therapists on their efficacy with lower-income and minority patients before and after attendance in an orientation program to increase therapists' knowledge of and sensitivity to lower income and minority patients and their concerns. Patients of oriented therapists reported higher levels of satisfaction with therapy, felt they were better able to handle their problems, and were more likely to return for therapy in the event of future problems. Lerner (1972) and Lerner and Fiske (1973) carried out two studies to examine possible client and therapist characteristics contributing to positive outcome in psychotherapy. For both studies, 30 consecutively accepted outpatients at a large state-supported outpatient clinic received therapy from 14 therapists of varying theoretical orientations. Of the 30 outpatients, 10 were men, 20 were women, 17 were black and 13 were white. Ages ranged from 16 - 57, with a median range of 32. In the first study (Lerner, 1972), 11 client factors including social class, severity of psychological impairment (assessed by both the therapist and by an independent diagnostic test), present social unproductivity, past psychiatric hospitalization, age, race, and factors from two scaling tests (Klopfer Prognostic Scale and the California Scale) were examined. 50 Results showed no significant relationship between these client factors and therapeutic outcome. Outcome was found to be related to therapist factors, with the most significant factors for positive outcome being a democratic attitude on the therapist's part, and a preference for working with severely impaired and lower class clients. In the second study, Lerner and Fiske (1973) chose to examine client personality traits that previous researchers had chosen as being indicators of good prognosis for therapeutic outcome. These traits were: low preoccupation with somatic complaints, low tendency to externalize blame, low suspicion and guardedness, low tendency to act out, high levels of introspection, high general ego strength, high sophistication regarding the therapeutic process, high verbal fluency, high motivation for change, high estimated intelligence, and moderate degree of experienced anxiety. Measures of outcome were gathered through 1) clients' ratings; 2) therapists' ratings, and 3) the Rorschach Psychological Functioning Scale (using blind raters) pre-therapy to post-therapy change scores. Lerner and Fiske's (1973) results showed no significant relationship between high scores on the good prognosis personality traits and positive therapeutic outcome on any of the three measures. Clients with many favourable attributes did no better than clients with very few favourable attributes. Furthermore, sums of the good client attributes for therapy were related to both high social class and to low levels of psychopathology. As Lerner and Fiske (1973) point out, "lower-class and severely impaired clients do tend to be significantly more guarded and suspicious, less sophisticated about the therapeutic process than their more socio-economically and psychologically fortunate brethren. In addition, they seem slightly more likely to emphasize somatic complaints, to externalize 51 blame, and to strike their therapists, on the basis of initial contact at least, as less intellectually impressive." (p. 275) However, there was no indication that lower class clients tended to act out impulsively or lack introspective ability. The authors point out that while some therapists state that their client preferences have nothing to do with social class, many of the qualities many therapists look for in evaluating a client's prospect for therapy are very much linked to social class. Lerner and Fiske's (1973) study, however, did contain several limitations. There was no analysis made to determine whether race, sex or age had any significant effects on the study results. The therapists' use of a variety of different theoretical orientations with the clients may also be viewed as another confound. In addition, the original sample number had been 45, but 15 had either dropped out or had not finished therapy at the time of data collection. It would have been interesting to analyse client characteristics for those who had dropped out, including socio-economic data, as their leaving may have produced skewed results for the remaining 30 participants. In 1986, Sutton and Kessler carried out a study involving a random selection of clinical psychologists chosen from the membership register of the 1980 American Psychological Association. The psychologists were asked to evaluate three hypothetical clients on seven scales. Three social class variations (Class III, IV and V) of the same case history were prepared and reliability established. The seven ratings included: 1) the client's prognosis; 2) the client's motivation to change; 3) the client's self-concept; 4) the severity of the client's disorder; 5) the therapist's own personal interest in treating the client; 6) the likelihood of using psychotherapy as the main modality of treatment, and 7) the likelihood of referring for medication. 52 On five of the seven variables, the Class V, or lower class client received the poorest ratings. The only two scores that the Class V client was rated higher on were severity of illness and likelihood of referral for medication. With the same case history, Class V clients were given a poorer prognosis, their self-concepts were viewed as lower, there was significantly less interest in treating them, they were less likely to receive psychodynamic therapy, and were more likely to be referred for medication. While Sutton and Kessler mention that, since lower class patients tend to perform poorly on clinical measures, it is possible that psychologists are using this, along with other social cues to effectively judge lower class patients. However, they argue that this does not explain the lowered personal interest in treating lower class patients or the tendency to refer them for medication. Sutton and Kessler concluded that poorer scores on the parts of lower class clients in comparison to middle and upper class clients' scores may be as much determined by therapist attitudes as they are by lower class clients' attitudes. Their study clearly showed some bias on the part of the therapists in their subjective interpretations and evaluations of lower SES clients. 53 Questions/Hypotheses Research carried out to date indicates that lower social class individuals have at times been deemed less acceptable for treatment, had fewer outpatient visits, were treated more frequently with medications rather than psychotherapy, and were more likely to be negatively perceived and evaluated (Garfield, 1986; Herr, 1989; Lorion & Felner, 1986). Some of the research also appears to indicate that a significant portion of the bias is therapist related (Lerner, 1973; Lerner & Fiske, 1974; Lorion, 1974; Sutton & Kessler, 1986). Lorion and Parron (1985) suggest that therapists have various stereotypes about the lower class client. These stereotypes include descriptions of lower class clients as unreliable, irresponsible, impulsive, and hostile. Lorion and Parron (1985) also note that therapists described lower class clients negatively in terms of styles of dress, language patterns, attitudes towards work, general lifestyles, and dependence on the welfare system. However, much of the psychological research is outdated and most of it is within the American context. I am interested in examining, within a Canadian context, whether current research will continue to indicate a negative influence of client social class on therapist perceptions, evaluations and decision-making. More specifically, would there be evidence that client socio-economic status impacts upon therapists' acceptance of clients for treatment, on diagnosis, on prognosis, on the probability of referring for psychiatric assessment, on perceived client attributes, on choice of therapy offered and presenting problems, and on recommended length of stay in therapy. 54 My hypotheses are: 1) Therapists will evaluate a lower class client more negatively than they will a middle class client. 2) Therapists will be less likely to accept a lower class client for therapy 3) Therapists will offer shorter durations of therapy for a lower class client. 4) Since higher levels of education appear to have a conservative effect on individuals, there will be more negative and stereotyped descriptions and evaluations of the lower class client by respondents with higher levels of education. 5) Respondents who have higher levels of interaction with lower social class individuals will rate them less negatively than respondents who have little interaction with lower social class individuals. 6) Respondents' years of counselling experience will not significantly impact on ratings. 7) Due to a heightened awareness of gender biases and the need for inclusion of social context within feminist counselling, female respondents will produce less negative evaluations of the lower class client than male respondents. 55 Method Design To examine and test my hypotheses, I chose to carry out an experimental field study based on the design of Sutton and Kessler (1983). Two short written vignettes representing intake interviews were prepared to present the case histories of two hypothetical clients. The two clients (and vignettes) were identical on all variables except for the variable of social class. In addition, two fictitious client intake information sheets were prepared. These contained identical information except for levels of education and occupations of both the client and his parents. Identical scores on results of psychological testing were also provided. Independent Variable: Client Social Class Social class (or socio-economic status) was operationalized through the development of the two client vignettes and accompanying intake information forms illustrating a middle and a lower socio-economic status client. The independent variable consists of the two hypothetical clients, identical on all variables except for socio-economic status and background, which was identified through a combination of level of education, occupation and parents' levels of education and occupations. The Hollingshead Two-Factor Index of Social Position (ISP) was the primary source used to develop the social class differences for the independent variable because it is the most commonly used instrument for measuring socio-economic status in research in psychotherapy. In Hollingshead's model, socio-economic status (See Table 1) is divided into five classes based on a combination of levels of education and occupation 56 (Hollingshead & Redlich, 1958). Briefly, Classes I and II represent the upper and upper-middle classes. Class III represents the middle class. Class IV represents the working class, and Class V represents the lower-working class or poor. Individuals are classified according to their years of education and their (or their father's) occupational status. Developed during the fifties, Hollingshead's ISP was based on one family income and Table 1 Hollingshead's Occupational and Educational Characteristics of Five Socio-economic Levels Class Occupational Level Educational Level I Salaried positions in policy making Professional degrees; A.B. level executive level; private practice and beyond professionals II Salaried positions in business and professions; minor professionals included A.B. level or partial college III "Middle class" administrative, clerical, sales, technical, and semi-professional positions High school diploma IV "Working class" skilled and semi- High school or technical school skilled manual occupations in diploma with some below tenth unionized trades and industries grade V "Poor" semi-skilled and unskilled High school diploma infrequent manual occupations nonunionized with many not completing eighth with irregular employment grade Note. The socio-economic status levels are referenced in Hollingshead and Redlich (1958). Table from Sutton (1983). 57 occupation. Levels of education have also changed over the past 40 years in that, average levels of education now tend to be higher by several years (Guppy et al.; 1984). Therefore, education levels for both the middle and the lower class clients were adjusted upward by several years. Using the modified Hollingshead Two-Factor Index of Social Position (1957) and incorporating the ideas of Wright et al. (1982) and Cardachi (1975) on degrees of control and decision making, half of the vignettes described a Class III, or middle-class client from a middle-class background, with a Bachelor of Applied Science, who worked as a buildings inspector with a relatively stable work history. The other half described a Class V, or lower class client, from a lower class background, with a grade nine education, who worked as an unskilled labourer (blue collar) with an unstable work history. An abundance of sociological literature points out that education and occupation alone do not fully address the issue of the degree of decision-making and control one possesses over one's work - factors closely linked to one's standing within the social class hierarchy. Some individuals within the same occupation may have little control over their job situation, while others have high levels of autonomy and control. For instance, the occupation of carpenter, usually classified as working class, could be a long-term unemployed individual (lower class), a worker in a company (working class), a semi-autonomous employee (any class), a manager (middle class), or a business owner (middle to upper class). Using a Marxist analysis of class, the works of Eric Olin Wright (Wright, Costello, Hachen & Sprague, 1982) and G. Cardachi (1975) on class structure have incorporated the concepts of control over investments, decision making, other people's work and one's 58 own work in their analyses of and research into class structure in America. In short, those with the least control comprise the workers (the lower and working classes), those with some control and decision making capability in their work situations would be the managerial group and semi-autonomous workers (roughly the equivalent of the middle and upper-middle classes), while those with high levels of autonomy and control over others and investments would comprise the upper classes or the bourgeoisie. The hypothetical clients were developed with this in mind. Pilot Study Initial pretesting of the client social class variable and the questionnaire were carried out through the participation of volunteer students attending three Summer graduate classes in Counselling Psychology at the University of British Columbia (See j Table 2). Students were asked to rate the client's social class as either upper, middle, Table 2 Student Vignette Ratings of Hypothetical Client's Social Class Student Rating Hypothetical Client upper middle working lower unsure middle class client 1 10 2 1 lower class client 14 1 2 working or lower class. Of a total of 31 students who participated, fourteen rated the middle class vignette and seventeen rated the lower class vignette. Of the fourteen students who rated the middle class client, 10 rated him as middle class (77%), 2 rated him as working class, one as upper class, and one was unsure. Of the 17 who rated the lower 59 class client, 14 rated him as working class, 1 as lower class, and 2 were unsure. Based on their judgements of the client's social class, I proceeded to modify the social class aspect of the client vignettes to further clarify the lower class client's status and background. The final client vignettes were reviewed by my thesis committee supervisor and three independent participants with 100% inter-rater reliability, classifying the vignettes as predicted. In response to participants' comments on the questionnaires, adjustments were also made to the demographic information and the semantic differential sections to further clarify responding procedures. Participants Participants were members of the British Columbia Psychological Association (BCPA) and the Canadian Guidance and Counselling Association (CGCA) within British Columbia. Project approval and assistance were requested and received from both associations. To ensure adequate representation, random sampling was carried out by the associations to choose 300 practising therapists from within British Columbia, selected from the two associations' membership mailout lists. Each randomly selected member from each of the two associations received a questionnaire package through the mail. Participation was completely voluntary. To encourage participation, each respondent was offered the opportunity to enter their names in a drawing for a $25.00 gift certificate to the restaurant of their choice. To ensure anonymity and confidentiality, small separate envelopes for the respondent's name and address were included with the questionnaire responses to be detached from responses and put in a draw box. Respondents were also provided with a request form to be mailed in separately, in order to receive results of the 60 study after completion. A successful return was judged to require approximately 100 responses. Procedure Three hundred psychologists and counsellors, representing the 150 randomly selected members from each of the two participating associations, were randomly distributed either a middle or a lower class client case history, with 50% each of the middle and lower class client case histories allotted to each association. The case history included a completed intake information form, results of psychological testing, and notes from an initial intake session. Each subject received a package containing only one of the two case histories, a questionnaire including Likert rating scales and short answer questions, a client characteristics semantic differential scale using adjectives noted in the both sociological and psychological literature as frequently used descriptors for members of the lower social class, a section linking respondent's personal relationships according to occupation, a demographic questionnaire, and preliminary information with instructions for responding (See Appendix A). Two to three weeks after the initial mailout, a follow-up letter was sent out, encouraging those who had not already responded to respond. Responding procedure involved reading the case history information and answering the attached questionnaire. Subjects were instructed not to turn back to read the case history again when answering the questionnaire. By manipulating the case history on one variable, that of client social class, one can then observe if there is a relationship between the client's social class and respondents' ratings, responses, and evaluations. After responding to the client information questions, the respondents were to fill in the Personal Relationships form, and lastly to fill in the demographic information section. 61 Of the 300 questionnaire packages mailed, 8 were returned stamped that the individual had moved with no forwarding address, 3 disqualified themselves because they were not practising therapists, one returned the questionnaire stating that he chose not to participate, and one had received the questionnaire from both associations. In all, 111 (37%) responded. Of the returned questionnaires, 56 contained responses to information on the lower class client and 55 contained responses to information on the middle class client. Characteristics of Respondents Table 3 presents descriptions of the sample population. Respondents from both of the participating associations were predominantly Caucasian and from middle to upper class backgrounds. All in all, roughly half of the respondents were female (N= 58) and half were male (N= 52). However, there were differences in the numbers of males and females who responded from within the two associations. From the British Columbia Psychological Association approximately 41% of the respondents were female, while 59% were male. From the Canadian Guidance and Counselling Association approximately 61% of the respondents were female and 39% were male. These numbers are roughly the inverse of each other. The average level of education among CGCA respondents was a Master's degree, and among BCPA respondents it was a doctoral degree. Most of the respondents from both associations had attained degrees from either Counselling or Clinical Psychology programs. More than 75% of the respondents were over the age of forty. Of the 111 respondents, 105 were white. No significant differences between respondent groups (based on the client's social class) were noted on respondents' age, gender, race, or degree obtained. 62 Table 3 Demographic Characteristics of Respondents CGCA (#=62) BCPA (N=49) Ethnicity / % f % White 56 90.3 49 100. Aboriginal 1 1.6 Asian 2 3.2 East Indian Black Hispanic Other 3 4.8 Total 62 100. 49 100. Gender / % / % Female 38 61.3 20 40.8 Male 24 38.7 29 59.2 Total 62 100. 49 100. (table continues) 63 Table 3 (continued) CGCA (N = 62) BCPA (N = 49) Highest Degree Obtained / % / % Bachelor's 6 9.7 1 2.0 Master's 36 58.1 18 36.7 Doctorate 12 19.4 27 55.1 Other 1 1.6 1 2.0 Unanswered 7 11.3 2 4.1 Total: 62 100. 49 100. Program of Highest Degree / % / % Counselling Psychology 27 43.6 18 36.8 Clinical Psychology 11 17.7 17 34.8 Educational Psychology 3 4.8 5 10.2 Counsellor Education 6 9.7 2 4.1 Social Work 1 1.6 Other 12 19.4 6 12.2 No graduate degree 1 1.6 Unanswered 1 1.6 1 2.0 Total: 62 100% 49 100% (table continues) 64 Table 3 (continued) CGCA BCPA (N = 62) (N = 49) Respondent Age / % / % 30-35 3 4.8 3 6.1 36-41 5 8.1 6 12.3 42-48 21 33.9 13 26.5 49-54 18 29.0 14 28.6 55-60 7 11.3 6 12.3 61-66 7 11.3 1 2.0 67 or older 1 1.6 6 12.3 (table continues) 65 Table 3 (continued) CGCA BCPA (N = 62) (N = 49) Mothers' Education Level / % f % No formal education Some grade school 6 9.7 Completed grade school 5 8.1 7 14.3 Some high school 11 17.7 9 18.4 Completed high school 19 30.6 17 34.7 Some college 4 6.5 5 10.2 Completed college 5 8.1 4 8.2 Some university 3 4.8 Undergraduate degree 6 9.7 1 2.0 Some graduate work 1 2.0 Graduate degree 2 3.2 5 10.2 missing (01) (1.6) (table continues) 66 Table 3 (continued) CGCA (N = 62) BCPA (N = 49) Father's Education / % / % No formal education Some grade school 6 9.7 5 10.2 Completed grade school 7 11.3 2 4:1 Some high school 11 17.7 7 14.3 Completed high school 14 22.6 9 18.4 Some college 5 8.1 6 12.2 Completed college 3 4.8 2 4.1 Some university 2 3.2 1 2.0 Undergraduate degree 3 4.8 7 14.3 Some graduate work Graduate degree 10 16.1 9 18.4 missing (01) (1.6) (01) (2.0) (table continues) 67 Table 3 (continued) CGCA BCPA Total Years experience / % / % / % 0- 4 10 16.4 6 12.5 16 14.4 5- 9 18 29.5 9 18.8 27 24.3 10-14 12 19.7 11 22.9 23 20.7 15 - 19 12 19.7 8 16.7 20 18.0 20 + 9 14.8 14 29.2 23 20.7 61 100.0 48 100.0 109 100.0 CGCA BCPA Total Father's Occupation / % / % / % owner 10 16.13 7 14.29 17 15.32 manager/supervisor 16 25.81 18 36.74 34 45.95 skilled professional 12 19.36 9 18.37 21 64.87 semi/unskilled white collar 1 1.61 1 2.04 2 66.67 skilled blue collar 13 20.97 7 14.29 20 84.69 semi/unskilled blue collar 3 04.84 1 2.04 4 88.29 farmer 7 11.29 4 8.16 11 98.20 father deceased/missing 2 4.08 100.00 68 No significant difference was noted on respondents' numbers of years experience and client social class. Respondents to the questionnaires containing the lower class client had an average of 11.72 years experience, while respondents to the middle class client had an average of 13.50 years experience (p = .23). There was, however, a significant difference in the numbers of years experience based on association membership, with the CGCA respondents having, on the average, 10.5 years experience, and the BCPA respondents having an average of 15.5 years experience (See Mests in Appendix B). Both fathers' and mothers' average levels of education for respondents from both associations was high school completion. While this was higher than the average for individuals attending high school in the 1930s and 1940s (Guppy et al., 1984), it is not surprising when one considers the education levels of respondent population. Stimulus Material The case history describes a client with symptoms of generalized anxiety disorder (See Appendix A for vignettes descriptions), as described in the Diagnostic and Statistical Manual of Mental Disorders (1995). These symptoms include, restlessness, feeling of being on edge, trouble falling and staying asleep, irritability, and worry and anxiety over at least two life circumstances. The client is described as a 34-year-old white male who presented himself at a counselling centre for help with marital problems. Personal intake information lists him as being the third of five children, whose father had left when the client was quite young, leaving the mother to care for and support the children. During the intake interview the client's behaviour is described as agitated and anxious. The client is presently unemployed and stated that he had carelessly spent most of his income for that week. He stated that he 69 and his wife had argued over money and he had become angry and broken several objects before hitting his wife. He also stated having high fears of losing his children, little hope that counselling will help him, and mentioned that he had been drinking excessively lately. Standard psychological testing showed elevations at the 80th percentile for anxiety, the 82nd percentile for depression, and the 71st percentile for paranoia. Dependent Variable. The dependent variables for this study are the subjects' written perceptions and evaluations of the client, as noted through responses in the form of Likert rating scales, short answer questions, and a semantic differential scale. Measures. The questionnaire consists of items designed to note therapists' perceptions of, attitudes towards, beliefs about, and evaluations of lower SES clients (See Appendix A for copy of questionnaire). In Part I of the questionnaire, respondents were asked to rate the client along six likert scales with seven point ranges. The client was rated on: 1) the severity of the presenting problem: from very mild to very severe; 2) client motivation to change: from very strong to very weak; 3) client self-concept: from excellent to poor; 4) client prognosis: from favourable to bleak; 5) the probability of referring the client for psychiatric assessment: from very high to very low; and 6) counsellor's level of personal interest in treating the client: from very strong to very weak. To prevent response bias, the polarities of the scales were randomized. The counsellors were also asked to select what type of individual treatment they might use with this client, what issues and problems they felt were most important to address first, and suggested length of stay in therapy. The second part of the questionnaire was a semantic differential scale. The semantic differential scale was originally developed by C. E. Osgood (1969) in order to 70 examine cross-cultural polarization of attitudes. For my study, the respondents were asked to rate the hypothetical client on a series of opposing pairs of adjectives that appear on a seven-point scale ranging from "very closely describes client" at points seven and one, to "unrelated" at the midpoint of four. By choosing to use a semantic differential scale, it was possible to select adjectives that were more directly relevant to the study of social class bias. The adjectives chosen for inclusion were selected from opposing pairs of adjective lists used by Osgood (1960) and adjectives that had been previously noted in the literature to have been stereotypical descriptors of lower socio-economic status individuals and/or groups (Hollingshead & Redlich, 1958; Jones, 1982; Morris & Williamson, 1982). The semantic differential allows for an evaluative judgement of a given object by the respondent. The object in this case is the client from either a middle or a lower social class. As in the case of the six initial rating questions, the polarity of the scales were randomized. Osgood et al. (1970; 1984) extensively studied the reliability for lists of opposing adjective pairs, and found test-retest studies for reliability to produce coefficients ranging from .85 (Osgood, 1984) to .93 (Osgood et al, 1970). While also considered by Osgood and others (1970; 1984) to possess a high level of face validity in comparison with other attitudinal instruments of measure, some researchers have suggested its use only in conjunction with other measures. My choice for inclusion of a semantic differential scale focuses on the importance of using more than one measure to test for possible bias. While Sutton and Kessler were able to note bias in their original work carried out in 1981,1 felt that close to fifteen years later, taking into account a greater awareness of studies noting bias in a variety of forms, it was important to use several measures of difference. The 71 semantic differential scale, while showing some overlap with the initial six likert rating scales, allows for investigation of whether the subjects' perceptions and evaluations of the client are influenced by specific stereotypical social class descriptors. Part III of the questionnaire asked subjects to check off types of relationships they have with individuals from a variety of occupations, including individuals receiving long-term social assistance. The types of relationships listed were acquaintance, close friend, relative, colleague and client. This section is a modification of a question utilized by Tindall (research in process) in his investigation of characteristics of individuals participating in the environmental movement on Vancouver Island. My intention in incorporating this section was to investigate whether various forms of contact with lower socio-economic status individuals might have a positive impact on the subjects' perceptions and evaluations of the client. The final section, Part IV, was a demographic questionnaire that included information on the participant's gender, race, age, marital status, social class background, including information on whether their family had had to rely on public assistance at any time, both parents' and respondent's educational background, counselling experience, and present work environment (e.g. private practice, non-profit agency, university setting, etc.). Data Analysis. Since it is essentially the same case history, a comparison of the rating scales results according to client SES, made it possible to determine whether there was a relationship between a client's social class and counsellors' perceptions and evaluations. Statistical analyses in the forms of f-tests and ANOVAs were used to examine both the Likert rating scales and semantic differential evaluation results. MANOVAS 72 were used to examine results on both scales based on the combination of therapist's years of experience, education, and therapist's degree of interaction with lower SES individuals and client's social class. For ease of analysis and reliability, a rater was recruited for Part I, to examine and categorize client presenting problems as noted by the respondents. Using chi-square analysis, differences in respondent choice of therapy and frequencies of client issues linked to client social class was examined. Comparison of results based on association membership and gender were also investigated for possible differences. For statistical significance, I chose to use a cutoff level using alpha = .05, however, results up to alpha = .10 are discussed as possible trends. All statistical analyses were conducted using the SPSS-X statistical program. Summary The premise of the current study is that, if the lower socio-economic status client received less positive ratings, was found less acceptable for treatment, or was descriptively evaluated less positively, based solely on the difference in social class, it would provide support for the hypothesis that therapists' perceptions are related to social class bias, lack of social class awareness and/or negative social class stereotyping. The next chapter will examine the results of this researcher's findings to determine whether there are any differences in therapist ratings, perceptions, decisions and evaluations linked to client social class. 73 Results This study was developed to investigate possible therapist bias linked to client socio-economic status (SES). Previous reviews of the research have found a pattern of negative perceptions, evaluations, and treatment of lower social class clients. Questionnaire results were analysed by total sample population, by professional association, and by gender. To examine whether information on client SES had an impact on the therapists' judgements and evaluations of the client, analyses of variance and f-tests were performed on the six continuous variables evaluating the therapist's conception of the client's problem (hereafter referred to as problem evaluation) and on the impression of the client measured with 19 pairs of opposing adjectives that comprised the semantic differential scale (hereafter referred to as the client characteristics). In addition, chi square analyses were performed to examine possible relationships between client social class and type and frequency of participant responses to open-ended questions on aspects of counselling. Possible interrelationships between a variety of respondent demographic information and responses linked to client social class were also investigated. Each section will be followed by a short summary of results as well as a final summary at the end of this chapter. Respondents' Client Ratings and Evaluations Problem Evaluation The six variables comprising the problem evaluation were severity of presenting problem; perceived client motivation for change; perceived client concept; prognosis; likelihood of referring client for psychiatric assessment, and personal interest in treating 74 the client. These variables were measured on a seven point Likert-type scale (See Appendix A). For the purposes of comparison, all seven variable ratings in Table 4 were arranged with 1 equalling the most positive rating, and 7 equalling the poorest rating of the client. Overall, respondents rated client self-concept as relatively poor, with means of 5.85 and 5.87, respectively, for the middle class and lower class client. Client means for severity of presenting problem also received rather high ratings at 4.79 and 4.57 for the middle and lower class client. Means for interest in treating the client fell below mid-range at 2.91 for the lower class client and 3.02 for the middle class client, indicating a higher than average interest in treating (where 4 represents the midpoint). Motivation for change, prognosis, and likelihood to refer the client for psychiatric assessment fell approximately mid-way for both hypothetical clients. Table 4 Total Sample Scores on Respondent Ratings of the Class III and Class V Client Class III Class V Univariate F-tests (#=1,111) Client ratings N M SD N M SD F. P Severity 55 4.79 .81 56 4 57 .83 1.832 .18 Motivation 55 4.35 1.22 56 4 21 1.32 .296 .59 Client Self-Concept 55 5.85 .93 54 5 87 .80 .009 .92 Prognosis 53 3.42 1.31 56 3 70 1.14 1.453 .23 Psychiatric Assess 54 3.54 1.67 56 3 18 1.93 1.09 .30 Personal Interest 54 2.91 1.44 56 3 02 1.60 .144 .71 75 In contrast to Sutton and Kessler's (1981) research, multiple analyses of variance (F = 1.343,/? = .246) produced no significant differences overall in participant responses linked to client social class for any of the problem evaluation variables examined by total sample population, by association, or by respondent gender (See Table 4 for individual analyses of variance by total sample population). Because of the exploratory nature of the study, independent /-tests were also performed (See Appendix C) comparing the responses by total sample population, by association and by gender. Results produced significance on only one variable. On the variable prognosis for change, female respondents of the Class V client predicted a less favourable prognosis than the female respondents of the Class III client (Ms = 4.04, 3.28, SDs = 1.14, 1.31 respectively) t = 2.34, df= 54, p = .02. Male respondents produced no significant difference on the variable of prognosis linked to client social class. Choice of Therapy. Presenting Problems, and Length of Therapy The second section of Part I included 3 open-ended questions (See Appendix A) that asked the respondents 1) to describe the theoretical and/or treatment approach they would use with the client and reasons for their choices; 2) to indicate which problems the respondent viewed as most important to address first; and 3) estimated length of stay in therapy. Choice of Therapy. The most frequent choices of therapy among all respondents and for both hypothetical clients were cognitive (n = 59), behavioural (n = 53) and affective (Ar= 18). Table 5 provides a breakdown of choices of therapy by association and by total sample population. In all, 61% of the BCPA respondents chose to include the use 76 of a cognitive approach and 50% included a behaviourally based approach (n - 30 and 25 respectively). Of the CGCA respondents, 47% chose to include a cognitive approach, and 47% included a behaviourally based approach (n = 29 and 28 respectively). Of the 23 BCPA respondents who returned questionnaires based on responses to the middle class client, 16 or 70% stated that they would include a cognitive approach, and 12 or 52% stated that they would include a behavioural approach. Of the remaining 26 participants from the BCPA who returned questionnaires containing ratings and evaluations of the lower class client, 14 or 54% stated that they would include a cognitive and 13 or 50% would include a behavioural approach in dealing with this client. Of the 32 CGCA subjects who returned middle class client questionnaires, 15 or 47% included the use of a cognitive approach, while 13 or 41% stated that they would include a behavioural approach. Of the remaining 30 who returned lower class client Table 5 Respondents' Choice of Therapy by Association and by Total Sample Population Choice of Therapy CGCA BCPA Total Cognitive 11 8 19 Cognitive-Behavioural 10 19 29 Cognitive-Affective 1 2 3 Cognitive-Behavioural-Affective 7 1 8 Behavioural 10 5 15 Behavioural-Affective 1 1 Affective 5 1 6 77 Table 6 Choices of Cognitive and Behavioural Therapies by Total Sample Population and by Respondent Association CGCA BCPA Total Class III Class V Class III Class V Class III Class V (w = 32) (» = 30) (w = 23) (w = 26) (« = 55) (« = 56) Cognitive 14(47%) 14(47%) 16(70%) 14(54%) 31(56%) 28(50%) Behavioural 13(41%) 15(50%) 12(52%) 13(50%) 25(46%) 28(50%) questionnaire responses, 14 or 47% suggested some form of cognitive therapy, while 15 or 50% suggested some form of behavioural approach. In examining choice of therapy by gender, of the 25 male respondents of the Class III client, 52% (n = 13) chose to include a cognitive approach and 52% (n = 13) chose to include a behavioural approach. Of the 28 male respondents of the Class V client, 53.6% (n = 15) chose to include a cognitive approach and 42.9% (n = 12) chose to include a behavioural approach. Among the 31 female respondents who chose to include a cognitive and/or behavioural approach for the Class III client, 61.3% (w = 19) chose a cognitive approach and 42% (n = 13) chose a behavioural approach, while for the Class V client (N= 27), 48.1% (« = 13) chose cognitive approach and 60% (n = 16) chose a behavioural approach. As can be seen in Table 7, the male respondents chose a cognitive approach equally for the middle and lower class client (based on percentage of total numbers). However, they chose behavioural therapy slightly more frequently for the middle class 78 client than for the lower class client. Female respondents chose cognitive therapy more frequently for the middle class client, and behavioural therapy more frequently for the lower class client. Table 7 Choices of Cognitive and Behavioural Therapies by Total Sample Population and by Respondent Association Male Female Class V Class III in = 25) Class V (n = 28) Class III (w = 31) (n = 27) Cognitive 13 15 19 13 Behavioural 13 12 13 16 No other respondent differences linked to the client's social class were noted in the types of therapy chosen. Respondents tended towards a directive and structured approach with both the Class III and Class V clients. Of the 68 respondents from the total sample population (N= 111) who stated that their approach would be either structured or open-ended, 51 stated they would use a more structured approach (BCPA = 26, CGCA = 25), 16 stated they would chose an open-ended approach (BCPA = 3, CGCA = 13), and one stated the use of both a structured and an open-ended approach. Of the 64 respondents who noted either a directive, nondirective, or combination approach with their client, 49 chose a directive approach (BCPA = 26, CGCA = 21), 11 chose a nondirective approach (BCPA = 2, CGCA =11), and 4 chose the use of both directive and nondirective approaches (BCPA = 1, CGCA = 3). Several respondents from both client groups noted 79 that they felt this particular client would do better with a more concrete, task-oriented form of therapy. Explanations For Choice of Therapy. In examining respondents' explanations for choosing their approach with their particular client, a striking pattern emerged. While this data is qualitative in nature, it seems important to note. From among the BCPA questionnaire responses to the lower class client, comments implying judgement of the client's ability, competence, or intelligence included the following: "Possible structured, cognith^ehavioural ....[reason] He's probably not too bright." "His personal information profile does not suggest that one would get very far with subtle." "Structured and directive at onset, cognitive-behavioural methods. Why: I.Q. likely average to low average; likely not therapy minded; concrete thinker; low education; ...suspicious, paranoid tendencies likely to be exacerbated by ..." "Considering education level, sophistication of thinking ..." "I would approach Robert in a practical manner' what can be done to help you' manner. I would want an assessment of learning abilities and vocational counselling/training." "He would have more immediate benefit from doing tasks that effect change as opposed to verbal dialogue which would likely increase his anxiety and inferior position with the therapist." No such statements relating to low aptitude, low level learning capabilities or general intelligence appeared in the comments referring to the middle class client. Although the clients were the same in all aspects except for those defining social class, comments about the middle class client included less negatively stereotyped descriptive statements such as: "His 'style' and orientation seem more consistent with a structured ..." "Cognitive, as he would be more uncomfortable with a more 'feeling' or open ended 80 approach." Or comments such as: "Robert appears to be a bright, reasonably focused fellow", or "Client could readily understand the concepts, tasks, and exercises involved and would benefit readily, quickly." From the CGCA respondents, comments for the lower class client again included statements suggesting lower intelligence, or the need for intelligence testing: "Vocational aptitudes and interests - check learning disabilities training." "Structured, directive and behavioural -1 suspect his cognitive abilities are limited." For the middle class client, comments included: "Client would need to appreciate method in a practical cognitive sense..." "Because client is a Engineer by training and profession, I would use a quite structured, problem-solving, solution-focused approach." None of the respondents who returned the middle class client questionnaires with the same results on psychological testing, and the same presenting symptoms and problems as the lower class client, suggested intelligence, aptitude, or learning disability testing. Presenting Problems. The second question asked respondents to identify the problems and issues the respondent felt would be the most important to address first with this client. Table 8 lists the issues generated by association and by total sample population. Respondents were not provided with lists of issues nor were they given a limit on the specific numbers of issues they might choose to note. Issue groupings were then developed and refined, through the assistance of an independent volunteer rater, based on issues listed by the respondents. Independent categorization of a random sample of one-third of the questionnaires, for this question only, was performed by the volunteer rater. The individual recruited to 81 assist in the scoring is a 48-year-old female who presently works with individuals and groups as a career consultant. She is also an action-oriented trainer and writer. The rater was provided with a brief set of written instructions and score sheets (See Appendix D), and was asked to carry out two tasks: 1) to list the issues noted by the respondents into prearranged categories on the score sheets, and 2) note any way that the categories might be rearranged to be more accurately descriptive of the issues being noted. The results produced a rate of agreement between this researcher and the volunteer assistant of 92% on first scoring and 100% on the second scoring. The bulk of the difference on first scoring occurred between two categories, which were fear of losing his children, and fearfulness/paranoia. After a brief discussion, it was agreed to collapse the two categories into one on second scoring. Other very minor differences included missed issues and psychological wording that needed clarification for the assistant. As is evident is Table 8, respondents tended to provide slightly higher frequencies of issues for the middle class client than for the lower class client. For the middle class client (N= 55), there were 216 notations of presenting problems to be addressed, in comparison to 190 for the lower class client (JV= 56). The difference in numbers is predominantly related to the CGCA respondents, who generated 110 and 134 frequencies of issues for the lower and middle class client respectively. Among BCPA responses, there were 80 frequencies of issues for the lower class client and 82 for the middle class client. Where expected frequencies for specific issues were large enough, chi square analyses were performed by total sample population, by association, and by gender (See Appendix D). By total sample population, frequencies of five of the seventeen issues 82 Table 8 Respondents' Choices of Most Important Client Issues Needing Address BCPA CGCA TOTALS Issue: Class V (n = 26) Class III (« = 23) Class V (n = 30) Class III (/i = 32) Class V (w = 56) Class II] (n = 55) alcohol abuse 11 8 11 15 22 23 anger management 8 7 9 8 17 15 anxiety 3 9 4 7 7 16 childhood issues 2 2 7 6 9 8 depression 7 6 4 8 11 14 family/relationship issues 5 5 8 7 13 12 family violence 7 10 9 21 16 31 finances/money spending 1 3 5 3 6 6 general health issues 3 4 3 3 6 7 mental health issues 6 5 4 2 10 7 feelings/mood 1 8 1 8 2 fearfulness/paranoia 4 5 5 15 9 20 self-directed 1 1 3 5 4 6 self-esteem issues 4 5 8 10 12 15 suicide 3 3 1 4 4 7 unemployment 6 3 7 9 13 12 vocational assistance 2 1 7 9 1 other 7 4 8 9 15 13 Total 80 82 110 134 190 216 83 produced differences shown to be significant. For the middle class client, three issues were generated significantly more frequently. These were, with df=\, anxiety .Y2 (n = 23) = 4.24, p = .04, family violence X2 (n = 47) = 7.84, p = .01, and paranoiaX2 (n = 29) = 5.38, p = .02. For the lower class client, two were generated significantly more frequently. These were coping with feelings/mood X2 (n = 10) = 4.08, p = .04, and vocation X2(n= 10) = 4.08, p = .04. The issues of both coping with feelings/mood and vocation had expected frequencies in one cell each of less than five, however, with Yates continuity correction, there was still a probability of .09 that results were significant. A point of interest emerges when one separates out the frequencies of fear of losing his children (n = 13) from the issue of fearfulness/paranoia (n = 29). Of those who had noted fear of losing his children, 10 listings were for the middle class client (5 from BCPA respondents and 5 from CGCA respondents) and 3 were for the lower class client (all 3 from CGCA respondents). Among CGCA responses, with df=\, family violence X2 (n = 30) = 6.57, p = .01, and fearfulness/paranoiaX2 (n = 20) = 5.6, p = .02, produced significant differences, with significantly higher frequencies for the middle class client than for the lower class client. Frequencies of coping with feelings/mood X2 (n = 9) = 7.50, p = .01, and vocation X2 (n = 9) = 4.81, p = .03, were significantly higher for the lower class client than for the middle class client. Although the last two issues had expected frequencies of less than five, the Yates continuity correction was .02 for mood and .07 for vocation. Among BCPA responses, with df= 1, the frequencies of one issue produced significance: anxiety X2 (n = 12) = 5.02, p = .03. BCPA respondents noted anxiety significantly more frequently for the middle class client than for the lower class client. 84 Differences in response patterns linked to client social class were also evident when examined by respondent gender. Female participants produced significant differences in the frequencies of family violence X2 (1, n = 24) = 4.97, p = .03, and unemployment X2 (1, n = 16) = 4.38, p = .04. Female respondents noted the issue of family violence significantly more frequently for the middle class client, while unemployment was noted significantly more frequently for the lower class client. They also exhibited a tendency to generate the issue of fearfulness/paranoia X3(l,n = 20) = 3.6, p = .07 more frequently for the middle class client. Male participant responses produced no significant frequency differences on any of the issues, however, a tendencies towards difference were noted on the issues of anxiety X2(\, n = 9) = 4.08,/? = .04, and family violenceX2 (1, n = 23) = 3.06, p = .08. Note that on the issue of anxiety the alpha level is significant, however total frequency was only nine. For both of the issues mentioned, frequencies were higher for the middle class client than for the lower class client. The female respondents generated higher overall frequencies of presenting problems than the male respondents (TVs = 232 and 174 respectively). In examining results of the male respondents, 15 of the 18 issues had cell frequencies of less than 5, in comparison with 7 of the 18 issues for the female respondents. Because of this, any determination of gender differences based on these results would be speculative. Length of Therapy. The third open-ended question of Part I asked respondents to estimate the recommended number of sessions for this particular client. For recommended length of stay in therapy, many respondents produced ranges such as 10-15 sessions. Others omitted answering this question , or wrote that recommended length of 85 stay was, as yet, unknown (24 of 111 did not specify a recommended length of stay). Therefore, results for this question are based on 87 responses from the total sample population (See Table 9). For statistical purposes, range results were calculated with the midpoint treated as an average. Even numbers were rounded to the higher number in all cases (e.g. 10-15 sessions was calculated as 13 sessions). Table 9 Respondents' Average Recommended Length of Stay in Therapy Based on Client Social Class (N= 87) n M SD SE of Mean C G C A middle class 25 17.68 12.62 2.52 lower class 21 11.71 6.05 1.42 BCPA middle class 22 12.45 7.16 1.53 lower class 19 14.79 9.34 2.14 Total middle class 47 15.23 10.65 1.55 lower class 40 13.17 8.02 1.29 approximately 15 sessions suggested for the middle class client and 13 suggested for the As shown in Table 9, examination of responses based on total sample population did not produce a significant difference in recommended length of stay in therapy with lower class client. However, a significant difference in estimated length of stay in therapy 86 linked to client social class was produced from among the CGCA responses. The average recommended length of stay from the CGCA responses is 17.68 sessions for the middle class client and 11.71 sessions for the lower class client. Independent -^testing shows this difference to be significant at the .05 level (t = 2.06, df = 37.13). BCPA respondents did not produce a significant difference in the suggested number of sessions between the middle and lower class client. The suggested average number of sessions was 12.45 for the middle class client and 14.79 for the lower class client (r = -.91, df= 39, p = .37). When results on suggested length of stay in therapy were examined by respondent gender, there was a notable difference in the male respondents' results for length of stay in therapy linked to client social class. Male respondents suggested 17.52 sessions for the middle class client and 11.95 sessions for the lower class client (SDs = 12.76, 7.61 respectively, t = 1.71, p = .096). While the difference appears apparent, standard deviations are large, suggesting that the male respondents were not a homogeneous group in their responses to this question. Female respondents suggested approximately the same length of stay for the middle class client and the lower class client. With middle class client means and standard deviations listed first, the female respondents produced the following results: Ms =13.04, 14.40, SDs = 7.80, 7.43, t = -.55, p = .58 (See /-tests in Appendix DI). Summary. Differences related to client social class were noted in respondents' explanations for choice of therapy. Among CGCA responses, the overall frequencies of issues generated were lower in those questionnaires returned by respondents who had received the lower class client questionnaires. Certain presenting problems (most notably 87 anxiety and family violence) were less frequently generated by respondents of the lower class client by total sample population and by association. In addition, from among the CGCA sample population, the issues of coping with feelings/mood as well as vocational assistance were more frequently generated by respondents to the lower class client, while the issue of fearfulness/paranoia was more frequently noted by the respondents to the middle class client. The lower class client also received recommendations for a significantly shorter suggested length of stay in therapy than the middle class client from the CGCA respondent population. Since respondents were not choosing from prescribed lists in answering these three open-ended questions, caution must be exercised in examining the findings. Broad ranges of responses were evident both in choices and descriptions of treatment offered, in respondents' choices of presenting problems and issues to be addressed, and in length of stay in therapy. Client Characteristics Part II of the questionnaire contained a semantic differential scale using client characteristics as descriptors, in an attempt to note possible respondent stereotyping linked to client social class. The scale prepared for this research contained 19 pairs of opposing adjectives scored on a seven point continuous scale. The client characteristics variables included (See Table 10), were chosen because they had appeared in past psychological and sociological research as social class descriptors (Hollingshead & Redlich, 1958; Morris & Williamson, 1982). Common descriptors of lower social class individuals and groups noted in 88 psychological and sociological research and literature include adjectives such as ignorant, hostile, dirty, unknowledgeable, emotional, unfriendly and lazy (Morris & Williamson, 1982). Other descriptors were chosen because they have been used to describe other minority groups and individuals (e.g. based on gender, ethnicity, race, sexual orientation, etc.) where negative stereotyping has been detected and researched (Devine, 1989; Miller, 1988). Multivariate analyses of variance of differences by client social class were performed and provided the following results: for the total sample population, F (19, 91) = 3.71, p = .00, for CGCA respondents, F (19, 42) = 3.18, p = .00, for BCPA respondents, F (19, 29) = 2.89,/? = .01, for female respondents, F (19, 38) = 2.08,/? = .03, and for male respondents, F (19, 33) = 2.26, p = .02. All results were highly significant. Table 10 provides the results of univariate analyses of variance based on client social class for the total respondent population. Differences can be noted on the following pairs: unintelligent/ intelligent (p = .00), ignorant/knowledgeable (p = .00), friendly/hostile (p = .03), loud/quiet (p = .02), emotional/controlled (p = .00) and incapable/capable (p = .01). In addition, results on the variable: clean/dirty, produced a tendency towards difference (p = .07). The middle class client was evaluated as less friendly than the lower class client. All of the remaining significant differences described the lower class client as less intelligent, less knowledgeable, louder, more emotional and less capable (and possibly less clean) than his middle class counterpart. Independent /-tests proved significant (including the descriptor of clean/dirty) on all seven of the same opposing pairs (See Appendix E). 89 Table 10 Effect of Client Social Class on Respondents' Ratings Class III Class V Univariate F-tests (TV =55) (7V= 56) with(#= 1, 97) Descriptor: M SD M SD F P expressive/inexpressive 4.43 1.87 4.38 2.17 .16 .69 pessimistic/optimistic 2.04 1.32 2.29 1.62 .03 .87 trustworthy/untrustworthy 3.93 1.67 3.80 1.53 .12 .73 competent/incompetent 3.89 1.44 4.23 1.64 .15 .70 unintelligent/intelligent 5.25 1.44 3.45 1.53 35.15 .00 involved/withdrawn 5.04 1.66 4.67 1.88 1.24 .27 industrious/lazy 3.58 1.41 3.82 1.18 .06 .80 impulsive/reliable 2.24 1.26 2.49 1.55 .01 .94 responsible/irresponsible 4.49 1.81 4.38 1.75 .07 .79 ignorant/knowledgeable 4.29 1.56 3.39 1.40 14.11 .00 friendly/hostile 4.76 1.20 4.21 1.49 4.96 .03 loud/quiet 4.65 1.50 3.96 1.63 5.97 .02 clean/dirty 2.98 1.37 3.71 1.30 3.28 .07 emotional/controlled 3.89 2.10 2.95 2.00 12.69 .00 moral/immoral 3.33 1.43 3.34 1.40 .01 .94 incapable/capable 4.20 1.66 3.55 1.56 6.98 .01 defensive/ receptive 2.47 1.32 2.79 1.62 .11 .74 aggressive/peaceful 2.56 1.23 2.57 1.19 .01 .92 dishonest/honest 4.23 1.52 3.84 1.56 1.04 .31 90 In examining the two associations separately, some differences in client evaluation were noted. Independent /-tests found significant differences on five of the nineteen variables within the BCPA responses (See Table 11). These were: unintelligent/intelligent (p = .00), ignorant/knowledgeable (p = .02), loud/quiet (p = .05), clean/dirty (p = .02) and incapable/capable (p = .05). The lower class client was evaluated as less intelligent, less knowledgeable, louder, dirtier, and less capable than the middle class client by the BCPA respondents. Table 11 Significant Results of T-Tests of Client Ratings by Association BCPA CGCA ( J V = 4 9 ) ( J V = 6 2 ) M SD /-value p M SD /-value p unintelligent/intelligent middle class 5.54 1.02 5.01 .00 5.03 1.68 4.28 .00 lower class 3.68 1.52 3.27 1.53 ignorant/knowledgeable middle class 4.42 1.25 2.43 .02 4.19 1.78 2.18 .03 lower class 3.52 1.33 3.29 1.47 loud/quiet middle class 4.75 1.26 2.03 .05 4.82 1.42 1.39 .17 lower class 3.96 1.46 4.35 1.23 clean/dirty middle class 3.33 1.01 -2.32 .02 2.71 1.55 -2.01 .05 lower class 4.00 1.00 3.48 1.48 incapable/ capable middle class 4.29 1.55 2.03 .05 4.39 1.71 .92 .36 lower class 3.36 1.66 4.03 1.32 Examination of CGCA responses linked to client social class produced significant results on the following three variables: unintelligent/intelligent (p - .00), ignorant/knowledgeable (p = .03) and clean/dirty (p = .05). The lower class client was evaluated as less intelligent, less knowledgeable, and less clean. In addition, two other descriptors produced a tendency towards difference from within the CGCA responses. These were (with middle class means and standard deviations listed first): friendly/hostile (Ms = 4.68, 4.03, SDs = 1.35, 1.60) t = l.71,p = .09, and emotional/controlled (Ms = 3.97, 3.00, SDs = 2.29, 2.04) t = 1.15,p = .08. In these two cases the lower class client was rated as friendlier and more emotional than the middle class client. Investigation of differences linked to client social class by respondent gender also produced significant results (See Table 12). Evaluations by the female respondents produced significant differences on three of the sets of opposing pairs. These were: unintelligent/intelligent (p = .00), ignorant/knowledgeable (p = .03), and clean/dirty (p = .05). The lower socio-economic status client was evaluated by female respondents as less intelligent, less knowledgeable, and less clean than his middle class counterpart. Significant differences were noted in the male respondents' evaluations on five of the pairs of opposing adjectives. These were unintelligent/intelligent (p = .00), ignorant/knowledgeable (p = .02), loud/quiet (p = .05), clean/dirty (p = .04), and emotional/controlled (p = .05). The descriptor incapable/capable produced a tendency towards difference with an alpha level of .06. Male respondents evaluated the lower class client as less intelligent, less knowledgeable, louder, less clean, less controlled and possibly less capable than his middle class counterpart. 92 Table 12 Significant Results of T-Tests on Evaluations of Client Characteristics by Respondents' Gender (N= III) Female Respondents Male Respondents (n = 58) (n = 53) Descriptor: M SD /-value P M SD /-value P unintelligent/intelligent middle class 4.97 1.67 3.86 .00 5.60 1.04 5.62 .00 lower class 3.30 1.59 3.61 1.47 ignorant/knowledgeable middle class 4.13 1.78 2.15 .03 4.48 1.26 2.53 .02 lower class 3.21 1.45 3.57 1.35 clean/dirty middle class 2.60 1.45 -1.97 .05 3.44 1.12 -2.15 .04 lower class 3.36 1.47 4.07 1.02 loud/quiet middle class 4.55 1.65 1.69 .10 4.84 1.31 2.04 .05 lower class 3.78 1.83 4.07 1.41 emotional/controlled middle class 3.97 2.29 1.77 .08 3.88 1.87 1.98 .05 lower class 2.96 1.99 2.82 2.00 Overall, significant differences linked to client social class were evident when results were examined by total sample population, by association and by gender. BCPA respondents produced significant differences on 5 and CGCA respondents on 3 of the descriptor pairs, and the variables. Male respondents produced significant differences on 5 and female respondents on 3 of the descriptor pairs. 93 Respondents' Personal Network and Relationship to Client Evaluations Part IV of the questionnaire is a personal relationship grid (See Appendix A). Respondents' were asked to report whether they have relationships with individuals within particular occupations based on the following relationship types: friend, acquaintance, relative, colleague, and client. My aim in including the grid was to investigate whether respondents' contacts with individuals of the lower social classes (as defined simply by occupation for this section) would impact on respondents' evaluations and judgements of the lower social class client. The two categories of close contact and no contact with individuals from the lower social classes in the form of close friend, relative or client, were used as the independent variables. Responses to the problem evaluation and client characteristics scales were the dependent variables. Results showed no significant differences in evaluations of the clients based on contact with individuals of the lower social classes as listed in the personal network grid. However, the form was not designed to note how many contacts within a particular category any one respondent had. For example, while one respondent may have twenty friends on welfare and another may have one, both could check off the category of having a friend on welfare. The form used did not include any statistical data on numbers within any category, or qualitative information on closeness of the relationships. Another problem that arose in results within this section were the low numbers within certain categories. For the category welfare recipients, while 74 of the 111 respondents indicated having at least one client who was a welfare recipient, only 9 of the 111 indicated having a welfare recipient as a friend, and only 4 out of the 111 indicated having a relative who was a welfare recipient. Several other occupations showed 94 Table 13 Respondents' Personal Network Occupations friend relative colleague client business owners/managers 65 64 52 63 professors 55 37 62 29 teachers 60 54 35 61 bankers 15 15 4 31 accountants 23 22 7 40 engineers 32 31 5 33 bus drivers 9 6 2 28 seasonal farm workers/fishers 13 8 3 38 auto mechanics 19 8 3 38 plumbers/ electricians 10 12 2 36 waiters/waitresses 11 18 3 58 police officers 18 13 8 37 loggers/mill workers 10 13 3 48 firefighters 3 8 2 25 federal/provincial politicians 10 3 2 10 computer specialists 24 28 20 46 biologists/botanists/chemists 18 16 11 19 nurses/physiotherapists 51 36 40 58 doctors 50 31 43 37 architects 12 5 3 20 lawyers 38 24 20 37 unskilled construction workers 7 9 60 janitors/cleaning personnel 5 2 4 50 gas station attendants 3 2 2 28 unskilled factory workers 2 3 38 welfare recipients 7 8 74 truck drivers (non owners) 9 4 41 social workers 50 19 68 29 95 similar patterns within the categories of friends and relatives. Few respondents listed contact in the form of friendships or family relations with individuals from the lower and working class occupations listed in the grid. Within the category of friends, the eight occupations that were the least checked off were all lower to working class occupations. Those occupations were: bus drivers, fire fighters, unskilled construction workers, janitorial/cleaning personnel, gas station attendants, unskilled factory workers, and non-owner truck drivers. The eight most frequently checked off occupations within the category of friends were business owners/managers, professors, teachers, nurses/physiotherapists, doctors, lawyers, and social workers. In the category of relatives the pattern, although not identical, is similar. The lowest checked off occupations for relatives of respondents were: bus drivers, seasonal farm workers/fishers, auto mechanics, fire fighters, janitors/cleaning personnel, gas station attendants, unskilled factory workers, welfare recipients, truck drivers, architects, and federal/provincial politicians. The most frequently checked off in the category of relatives were: business owners/managers, professors, teachers, engineers, computer specialists, nurses/physiotherapists, doctors and lawyers. It is evident, noting the categories checked off in the grid, that respondents have noticeably higher levels of contact through friendships and family relatives with nearly all of the occupational categories used to represent middle and upper class groupings. While the list of occupations is limited, one can assume, based on responses to the occupations used, that the respondents predominantly have much less contact through friendships or through blood relations with lower and working class individuals. In contrast, within the category of clients the most frequently checked off group was welfare recipient. 96 In examining for differences in problem evaluation and on the client characteristics, responses were grouped based on contact versus no contact with social class occupational groupings. Responses were compared in the categories of friendship, relative, and client. Those occupations chosen as representative of the lower class were the following: unskilled construction workers, janitors/cleaning personnel, unskilled factory workers, welfare recipients, gas station attendants, and non-owner truck drivers. (Further analyses also included loggers/mill workers and waiters/waitresses.) Using independent /-testing, a difference was found in the response pattern on only one of the problem evaluation scales. Respondents (n = 16) who had clients in all five of the lower class categories mentioned above, rated the lower class client significantly higher (t = 3, df- 20, 54, p - .01) on motivation for change than those respondents (n = 95) who had no clients in any of the five lower class occupations. Caution must be used however, since the group of respondents having clients in all five of the lower class occupations was dramatically lower than respondents who did not have clients in all five of these occupations. No differences were found on responses to the client characteristics linked to the contact and no contact groups. Attempts at further analysis were unproductive because of the low numbers of respondents who had contact with individuals within lower class occupations. Respondents' Social Class Background Through the demographic section of the questionnaire, information was gathered relating to the respondents' mothers' and fathers' levels of education and work experience while the respondents were growing up (See Appendix A) in order to estimate the respondents' social class backgrounds. More than 85% of the respondent population were over the age of forty. This means that many of them were children at a time when less 97 women were regularly participating in the paid workforce. Close to half of the respondents reported their mothers as not having worked outside of the home, making analysis of mothers' occupations less statistically useful. Therefore, for the purposes of analysis, respondent's social class background was based on the father's occupation(s) and level of education. In two cases where the father was listed as missing or deceased, social class background was based on the mother's occupation and education. Not all social class groups were equally represented from within the sample population (See Figure 1). Using the father's occupations and level of education, seventy-three (65.77%) of the 111 respondents were from middle to upper class backgrounds. Of the remaining 38 respondents, 20 (18.02%) stated that their fathers were skilled blue Figure 1. Respondents' Social Class Backgrounds Figure 1. Respondent's social class background based on father's occupations and level of education (N = 111). 98 collar workers, 11 (9.91%) farmers, 5 (4.51%) semi - unskilled blue collar workers, and 2 (1.80%) semi - unskilled white collar workers (See Figure 1). As reported in the Methods section, more than half of the respondents' parents had completed high school. Considering the highly educated sample population used for this study, it is not surprising that the parents had levels of education and positions within the occupational structure that were above average for their time. As pointed out by Guppy (1984), there is a strong correlation between parents' standing within the socio- economic hierarchy and children's levels of education. Using multiple analysis of variance, no significant differences in response patterns linked to respondent social class background were noted for the problem evaluation or for the client characteristics results. However, analysis of responses linked to client and respondent social class was impaired by the significant absence of respondents whose backgrounds were from either the semi/unskilled blue collar, or the semi/unskilled white collar groups. Respondents' Level of Education To examine possible effects of the combination of client social class and respondent level of education on client evaluations, I chose to use only those respondents holding either a master's or a doctoral degree. Categories excluded were undergraduate degrees (n = 7), 'other' category (» = 2) and missing data (n = 9). In all, 45 of the 49 BCPA members (18 masters degrees, 27 doctorate degrees) and 48 of the 62 CGCA members (36 masters degrees, 12 doctorate degrees) were included in the analysis. No significant interactions by client social class and respondent degree were found on the problem evaluation (See Appendix F). In examining results on the client 99 characteristics by client social class and respondent degree, five of the nineteen client characteristics produced significant results (See Table 14). Analyses of variance found significant interactions between client social class and respondent's degree on the following client characteristics: involved/withdrawn, dishonest/honest, incapable/capable, emotional/controlled, and loud/quiet. Friendly/hostile also showed a slight tendency towards an interaction at an alpha level of .085. (See analyses of variance for all client characteristics in Appendix F). To assist me in understanding and interpreting the data, I performed Mests on each of the six descriptor pairs separately, by client social class and by respondent degree. For the following results, middle class means and standard deviations are listed first. InvolvedAVithdrawn. For the descriptor pair involved/withdrawn, master's respondents rated the middle class client as significantly more withdrawn than the lower class client (Ms = 5.60, 4.14, SDs = 1.12, 2.17) f = 3.03,/? = .004. Doctoral respondents exhibited a tendency to rate the middle class client as less withdrawn than the lower class client, however, the results did not reach statistical significance (Ms = 4.41, 5.41, SDs = 1.89, 1.28) t = -1.88,/? = :069. Dishonest/Honest. On the descriptor pair dishonest/honest, master's respondents rated the middle class client significantly more honest than the lower class client (Ms = 4.60, 3.59, SDs = 1.04, 1.90) t = 2.38,p = .02. Doctoral respondents did not produce a statistically significant difference in ratings between the middle and lower class client (Ms = 3.68, 4.24, SDs = 1.67, 1.09) t = -1.18,/? = .25. Incapable/Capable. For the descriptor pair incapable/capable, master's respondents rated the middle class client significantly more capable than the lower class 100 Table 14 Client Characteristics Showing Significance By Client Social Class and Respondent Degree (N= 93. df= 1) Client Characteristic SS MS F P involved/withdrawn by class 1.18 1.18 .40 .53 by degree .04 .04 .01 .91 class by degree 33.98 33.98 11.51 .00 dishonest/honest by class 1.19 1.19 .51 .48 by degree .41 .41 .18 .68 class by degree 13.74 13.74 5.97 .02 incapable/capable by class 6.94 6.94 2.76 .10 by degree 1.08 1.08 .43 .51 class by degree 10.92 10.92 4.34 .04 emotional/controlled by class 19.64 19.64 5.30 .02 by degree .00 .00 .00 1.00 class by degree 38.16 38.16 10.30 .02 loud/quiet by class 17.01 17.01 6.77 .01 by degree .10 .10 .04 .84 class by degree 13.93 13.9 5.54 .02 client (Ms = 4.60, 3.35, SDs = 1.32, 1.61) t = 3.10,/? = .003. Doctoral respondents produced no significant difference in their ratings of the middle and lower class client on this variable (Ms = 3.68, 3.82, SDs = 1.94, 1.38) t = -.26,p = .80. Emotional/Controlled. On the variable emotional/controlled, master's respondents rated the lower class client as significantly more emotional than the middle class client (Ms = 4.76, 2.52, SDs = 1.90, 1.86) t = 4.37, p = .000. Doctoral respondents produced no 101 significant difference in the rating of the middle and lower class client on this variable (Ms = 3.46, 3.82, SDs = 2.04, 1.91) t = -.58,/? = .57. Loud/Quiet. On the descriptor pair of loud/quiet, master's respondents rated the middle class client quieter than the lower class client. The difference was significant (Ms = 5.04, 3.38, SDs = 1.10, 1.92) t = 3.82,/? = .000. Doctoral respondents again produced no difference in their ratings on the loud/quiet scale (Ms = 4.32, 4.24, SDs = 1.96, .83, t = .16, p = .87). Friendly/Hostile. On the descriptor pair friendly/hostile, master's respondents rated the middle class client more hostile than the lower class client (Ms = 4.76, 3.83, SDs = .88, 1.79) t = 2.36,/? = .02. Doctoral respondents produced no significant difference linked to client social class on this variable (Ms = 4.59, 4.65, SDs = 1.37, .86) t = -.15,/? = .88. Ignorant/Knowledgeable. As with the total sample population, both associations and both genders, on the variable of ignorant/knowledgeable, masters degree respondents rated the lower class client as significantly less knowledgeable than the middle class client (Ms = 4.44, 3.00, SDs = 1.19, 1.49) t= 3.88,/? = .000. Doctoral respondents however, did not rate the lower class client as significantly less knowledgeable than the middle class client (Ms = 3.96, 3.53, SDs = 1.96, 1.23) f = .78,/? = .44. Clean/Dirty. By respondent degree, master's respondents produced no significant difference on the variable of clean/dirty (Ms = 3.08, 3.55, SDs = 1.28, 1.74) t = -1.11,/? = .27. Doctoral respondents however, rated the middle class client significantly cleaner than the lower class client (Ms = 3.05, 3.94, SDs = 1.46, .43) t = -2.44,/? = 02. Unintelligent/Intelligent. Both groups rated the lower class client as significantly 102 less intelligent than the middle class client. For the master's respondents results were Ms = 5.52, 3.38, SDs = 1.09, 1.74, t = 5.32,/; = .00. For the doctoral respondents, results were Ms = 4.86, 3.53, SDs = 1.86, 1.28, t = 2.53,/? = .02. Doctoral degree respondents also produced a tendency towards difference linked to client social class on the descriptor pair: impulsive/reliable (Ms = 2.23, 3.05, SDs = 1.19, 1.64, t = -1.84, p = .07). The lower class client was rated as more reliable than the middle class client. In all, master's degree respondents produced significant differences on eight of the opposing pairs of descriptors, and doctoral degree respondents produced significant differences on two (and a tendency towards difference on two other) of the opposing pairs of descriptors. Respondents' Years of Experience The impact of respondents' years of counselling experience on client evaluations was investigated using factorial analyses of variance (See Appendix G). Analyses were performed using the following five levels of counselling experience: 0-4 years (n = 16), 5 - 9 years (» = 27), 10 -14 years (n = 23), 15 -19 years (n = 20), and 20 or more years experience (n = 23). Within the problem evaluation scales, main effects linked to years of experience by client social class were noted for ratings of client motivation for change (F = 4, df= 4, p = .01) and for prognosis (F= 3.21, df= 4,p = .02). On motivation for change, respondents with 0-4 years experience produced a mean score of 5.50 for the middle class client and 3.50 for the lower class client. This suggests that the difference produced on this variable is due predominantly to the spread of scores within this 0-4 years experience group 103 Figure 2. (See Table 15). Both the 0-4 and the 15 - 19 years experiences groups rated the lower class client as having a stronger motivation for change than the middle class client. All three other groupings based on years of experience rated the middle class client as having a slightly higher motivation for change than the lower class client. On the variable of prognosis, as with motivation for change, scores do not appear to follow any linear pattern, making it difficult to draw any clear conclusions. In examining years of experience and evaluations of client characteristics as listed on the client characteristics scale, no significant differences linked to an interaction between respondents' years of experience and client social class were evident on any of the nineteen descriptor pairs. Although analyses of variance did not produce significant interactions, there were noteworthy differences evident in Table 16. However, because no consistent/linear pattern by experience is apparent, results are difficult to interpret. 104 Table 15 Rating Means For Motivation For Change and Prognosis Based On Respondents Years of Experience Number of Years 1 -4 5 -9 10- 14 15 - 19 20+ Experience: (n=\6) (n = 27) (n = 23) (n = 20) (n = 23) M M M M M Motivation for change Class III 5.50 4.15 3.92 4.91 4.21 Class V 3.50 5.00 4.18 3.78 4.44 F=4.23 /? = .003 Prognosis Class III 3.00 3.46 2.67 4.36 3.33 Class V 3.50 4.14 4.00 3.22 3.44 F=3.18 p = .0\l 105 Table 16 Mean Client Characteristic Scores by Respondents' Years of Experience Number of Years 1 -4 5 -9 10- 14 15 - 19 20+ Experience: (ft =16) (ft = 27) (ft = 23) (ft = 20) (ft = 23) M M M M M expressive/inexpressive Class III 3.75 5.39 4.67 4.46 3.64 Class V 4.33 5.29 3.64 4.33 3.78 F=.40 /> = .81 pessimistic/optimistic Class III 1.50 2.31 2.17 1.73 2.14 Class V 2.58 3.00 1.46 1.67 2.56 F=1.09 p = 37 trustworthy/untrustworthy Class III 3.50 3.85 3.25 5.09 3.93 Class V 3.82 4.07 3.55 3.78 3.78 F=.93 p = A5 competent/incompetent Class III 3.00 4.31 3.58 4.55 3.54 Class V 4.42 5.00 3.36 4.33 3.67 F=.85 p=.50 unintelligent/intelligent Class III 4.00 5.77 4.67 5.46 5.42 Class V 3.83 3.36 3.20 4.00 2.78 F=1.80 p = A3 involved/withdrawn Class III 6.00 5.54 4.33 5.64 4.50 Class V 4.25 5.29 4.64 5.33 3.78 F=.77 p = .55 (table continues) 106 Table 16 (continued) Number of Years 1 - 4 5 - 9 1 0 - 14 15 - 19 20+ Experience: ( w = 1 6 ) (n = 27) (n = 23) (n = 20) (n = 23) M M M M M industrious/lazy Class III 4.00 3.15 3.83 4.27 3.21 Class V 3.75 4.36 3.73 3.78 3.22 F = 1 . 6 0 p = AS impulsive/reliable Class III 4.00 3.15 3.83 4.27 3.21 Class V 3.75 4.36 3.73 3.78 3.22 F = . 5 6 p = .70 responsible/irresponsible Class III 3.25 4.77 4.50 5.27 4.07 Class V 4.58 4.50 3.70 5.11 3.89 F=.7\ p=.59 ignorant/knowledgeable Class III 3.50 4.62 3.42 4.73 4.50 Class V 3.67 3.43 3.18 3.89 2.78 F = 1 . 1 9 p = .32 friendly/hostile Class III 5.00 4.92 4.50 4.73 4.79 Class V 4.67 4.79 3.55 4.22 3.44 F = . 7 3 p = .57 loud/quiet Class III 4.25 4.46 4.33 5.09 4.79 Class V 4.33 4.21 3.36 4.22 3.56 F = . 5 2 p=J2 (table continues) 107 Table 16 (continued) Number of Years 1 -4 5 -9 10- 14 15 - 19 20+ Experience: (n = 16) (n = 27) (n = 23) (n = 20) (n = 23) M M M M M clean/dirty Class III Class V emotional/controlled Class III Class V moral/immoral Class III Class V incapable/capable Class III Class V defensive/receptive Class III Class V aggressive/peacefiil Class III Class V 2.75 3.15 3.00 2.64 3.21 3.75 4.14 3.18 3.78 3.55 F=.54 p = .l\ 4.00 4.39 2.92 4.73 3.64 3.17 3.57 2.64 3.11 2.20 F=.31 p = .S3 4.50 3.46 2.33 3.91 3.36 3.67 3.64 3.27 3.56 2.67 F=1.49 p = .2l 5.25 3.92 3.75 3.91 4.64 3.67 4.00 3.27 4.00 2.79 F=\.61 p = .16 2.00 2.62 2.17 2.64 2.71 2.92 2.79 3.18 2.89 2.33 F=J5 p=.56 2.00 3.15 2.33 2.18 2.71 2.42 2.86 1.91 2.89 2.67 F=.81 p = .52 (table continues) 108 Table 16 (continued) Number of Years 1 -4 5 -9 10- 14 15 - 19 20+ Experience: (n = 16) (n = 27) (n = 23) (n = 20) (n = 23) M A/ M M M dishonest/honest Class III Class V 3.75 4.54 3.75 3.73 4.64 4.75 4.00 3.27 3.67 3.22 F=\36 p = .25 Summary While few differences were found in the problem evaluation scales used to partially replicate Sutton and Kessler's (1981) original research carried out in the U. S., significant differences arose in responses to the open-ended questions, and in the client characteristics ratings. Differences linked to client social class also occurred when results were examined by professional association and by gender. Significant interactions were also evident between client social class and respondent level of education. It was impossible to carry out response analyses based on respondents' race or ethnicity as the sample population was overwhelmingly Caucasian (95%). In addition, response differences linked to respondent social class, while investigated, were of little value because there were so few respondents from lower (semi - unskilled blue and white collar) class backgrounds (n = 7) using father's occupations and parents' levels of education as social class indicators. In the following chapter, the results will be discussed with consideration given to the relevant literature, and the implications for clinical practice and training. Discussion 109 This study was designed to explore whether initial information about a client's social class would impact negatively on the therapist's perception, evaluation, and proposed treatment for the client. A review and discussion of the predominant findings, and implications of those results, appears below. Overall, this study lends support to previous psychological and sociological research into the effects of client socio-economic status (SES) on therapist perception, judgement, evaluation, and proposed treatment for the lower social class client. Since both hypothetical clients were identical except for information identifying social class and class background, it would appear that information identifying the one client as lower social class (education, occupation, family social class background, and manner of dress) had a negative impact upon respondents' perceptions, evaluations, and proposed treatment for that client. The implications of these findings are serious in that most of the respondents reported having clients within all levels of social class occupations; and, in fact, the client population category with the highest levels of frequency (i.e., the respondent has clients who fall within that category) was welfare recipient (See Respondent's Personal Network, p. 94). I have chosen to review the findings beginning with areas that I believe provide the most persuasive evidence of respondent bias linked to client social class. Client Characteristics Results on the client characteristics were significant when examined by total sample population, by association, by respondents' level of education, and by gender. 110 Findings bv Total Sample Population. Findings on the results of the client characteristics revealed that respondents receiving a description of a lower social class client provided significantly less positive evaluations of their client than did respondents receiving a description of a middle class client on 6 of the 19 pairs of opposing adjectives (p = .05). These were unintelligent/intelligent, ignorant/knowledgeable, loud/quiet, emotional/controlled, incapable/capable and clean/dirty. While it may be argued that the lower level of education attributed to the lower class client suggests a somewhat lower level of intelligence, and less knowledgeability and capability, it is imperative for counsellors to begin the counselling process with as little bias as possible. A variety of research studies from Canada and the United States have found that social class has a significant negative impact on the treatment of, and on the education received, by lower social class students (Baer & Lambert, 1982; Colclough & Beck, 1986; Gamoran, 1987; Guppy, Mikicich & Pendakur, 1984). This, in combination with the added stressors of childhood poverty, can produce higher dropout levels for lower social class students. To present for therapy, and to be initially perceived, evaluated and/or treated as less intelligent, capable or knowledgeable by a helping professional can in no way benefit a client from any social class background, let alone lower class individuals who interact in a society that already labels them as "losers". While some may make the claim that level of education is one valid reflection of intelligence, knowledgeability, or capability, the lower SES client was also evaluated as louder, more emotional, and less clean than the middle class client. There would appear to be no reason, based on the client information presented to the respondents, for these results except that these descriptors fit the commonly held stereotype of a lower class Ill member, as described in both the sociological and psychological literature (Hollingshead & Redlich, 1958; Jones, 1982; Morris & Williamson, 1982). Interestingly, the middle class client was given a more negative rating on the one variable of friendly/hostile. Lower class individuals have historically been stereotyped in the psychological literature as hostile (Hollingshead & Redlich, 1958; Morris & Williamson, 1982). Therefore, this one result falls outside of the stereotypic expectation. Perhaps, although both clients' behaviour and communication was the same, respondents may have assumed that the middle class client would have a more difficult time opening up to the therapeutic experience due to the embarrassment of presenting oneself for therapy. Another possible explanation may the existence of a Type I error. Findings bv Association. The findings also indicated that, while the lower class client received more negative ratings than the middle class client from both the Canadian Guidance and Counselling Association and the British Columbian Psychological Association respondents, BCPA respondents (N = 49) rated the lower class client more negatively on five of the variables in comparison with three of the variables for the CGCA respondents (N= 62). BCPA respondents rated the lower class client as less intelligent, less knowledgeable, louder, dirtier, and less capable than the middle class client. CGCA respondents rated the lower class client as less intelligent, less knowledgeable, and less clean. While at first glance, it would appear that the BCPA respondents exhibited more differential perception, two other variables within the CGCA respondent ratings also produced a tendency towards difference. These were hostile/friendly (p = .09) and more emotional/controlled (p - .08). The lower class client was rated as friendlier and more emotional than the middle class client. From these results, it is not possible to determine 112 that the BCPA responses produced more bias than the CGCA responses. However it is possible to state that both associations produced significant results that portray the lower class client in a more negative light, and that these differences fall within the societal stereotypes of lower social class members. Findings bv Respondent Level of Education. Some researchers (Baer & Lambert, 1982; Johnston & Ornstein, 1985) have suggested that higher levels of education are associated with more conservative socio-political viewpoints, although this may be somewhat mitigated by field of study. Longitudinal research (Guimond & Palmer, 1990) investigating the causal attributions of poverty found that social science students were more likely to endorse structural explanations for poverty over commerce or engineering students, who were more likely to endorse individualistic explanations. The respondent population's average level of education was a doctoral degree for the BCPA sample, in comparison with a master's degree among respondents from the CGCA. However, analysis of findings based on respondents' highest degree of education rather than association membership found that respondents with master's degrees produced more highly significant results on the client characteristics than did respondents with doctoral degrees. Respondents with doctoral degrees produced significant results on two of the descriptor pairs. These were: clean/dirty and unintelligent/intelligent. The lower class client was perceived to be dirtier and less intelligent than the middle class client. Master's degree respondents produced significant results on eight of the descriptor pairs. These were: ignorant/knowledgeable, friendly/hostile, loud/quiet, emotional/controlled, incapable/capable, dishonest/honest, unintelligent/intelligent, and 113 involved/withdrawn. The lower class client was perceived to be more ignorant, friendlier, louder, more emotional, less capable, less honest, less intelligent but more involved than his middle class counterpart. While respondents of both levels of education produced significant results that involved a predominantly negative perception of the lower class client, the group of respondents holding masters degrees produced bias on far more of the variables than the group of respondents holding doctoral degrees. It would appear that, contrary to one of my original hypotheses, higher levels of education within the general field of psychology had a mitigating effect on levels of biased perception. This would suggest that education positively impacted on the degree of negative stereotyping. However, since many masters level individuals will not continue on towards a doctoral degree, and doctoral respondents also produced significant bias in their responses, it would seem highly probable that providing training and education specifically on class and stereotyping would have a positive impact on both master's and doctoral students' ability to offer unbiased services to all social classes. Findings by Gender. The lower class client received more negative evaluations on the client characteristics than the middle class client when results were examined by respondent gender. Male respondents produced more negative evaluations of the lower class client on five of the descriptor pairs (unintelligent/intelligent, ignorant/ knowledgeable, loud/quiet, clean/dirty, emotional/controlled), and a strong tendency towards difference on one other (incapable/capable). Female respondents produced significant differences on three descriptors. These were unintelligent/ intelligent, ignorant/knowledgeable, and clean/dirty. While it is not possible to infer with certainty 114 that because the male respondents produced significant results on two more of the descriptors that they are, in fact, exhibiting more bias, it seems reasonable to assume that feminist attitudes in counselling have affected women more than men, and the feminist emphasis on taking into account the social contexts of a person's life may have had a modifying impact on female respondent ratings on the client characteristics. Findings bv Respondents' Years of Experience. In examining the data informally, an unusual pattern of response linked to respondents' level of experience was apparent. No consistent upward or downward pattern of evaluation linked to client social class and respondents' years of experience was evident; however, ratings varied dramatically at times, from one experience group to another. On those adjectives that had produced significance by social class for the total sample population (unintelligent/intelligent, ignorant/knowledgeable, friendly/hostile, loud/quiet, emotional/controlled, and incapable/capable), the 20+ years experience group produced the greatest spreads of differences linked to client social class, with the lower social class client receiving the poorer evaluations. Replicating the results found by Sutton (1983), therapists with the most years of counselling experience appear to have shown more negative clinical evaluations towards the lower social class client. It may be reasonable to assume that many of these respondents received their training at a time when it was believed that lower social class clients were not good candidates for psychotherapy, due to a variety of personal deficiencies such as lower levels of cognitive and verbal abilities and an inability to delay gratification (Hollingshead & Redlich, 1958). In addition, gender and cultural differences and other social context issues were not considered topics for therapeutic discussion, making recognition of biased perception more difficult. This may 115 have had an influence on ratings provided by the respondent population with 20+ years' experience. However, lack of statistically significant findings requires caution in advancing this hypothesis. More research may be required into the specific effects of years of counselling experience, theoretical orientation, and number of years since attending training or graduate school. In all, when results on the client characteristics were examined by total sample population, by association, by educational degree, and by gender, all produced significant differences linked to client social class. Most of these differences produce a negative perception of the lower class client unrelated to the information provided to the respondents in the client information material. Explanations for Choices of Therapy Explanations for choices of therapy contained more negatively descriptive and evaluative statements of the lower class client (e.g. "I suspect his cognitive abilities are limited."). The most obvious explanation for this would appear to be the belief that lower class clients (regardless of whether they present in the same manner as middle class clients) are less able than their middle class counterparts to think in the abstract, less apt to progress with insight-oriented therapy, more apt to require a vocational therapeutic focus, and more apt to have learning disabilities. Statements made were reinforced when examining evaluations produced on the client characteristics. There appeared to be a tendency to view the lower class client as less intelligent, less capable, and less apt to progress if offered insight-oriented therapy. The impact of such beliefs on the lower class client could no doubt effect both the progress made and their choice of remaining in therapy. To attempt to secure assistance 116 for a marital problem only to have vocational, aptitude or intelligence testing suggested would appear to be disrespectful of the client's own choices for coming to therapy, let alone ring of biased judgement. Previous literature (Spengler et al., 1990) has suggested a vocational overshadowing1 when both vocational and personal issues are brought out in the counselling setting. In this case, although both clients were unemployed, and neither verbalised vocational issues, only in the case of the lower class client was a vocational focus taken by some of the respondents. Rather than a vocational overshadowing, a personal overshadowing in favour of the vocational emerged for the lower class client. This vocational focus, in itself, does not necessarily prove respondent bias. It can be argued that because of the greater risks involved with poverty, respondents who chose to include a vocational focus were demonstrating an awareness of and concern for the socio-economic situation faced by the lower class client. However, the language used to explain the respondents' choice of therapy for the lower class client (e.g. "He's probably not too bright." "His personal information profile does not suggest that one would get very far with subtle.") leave one feeling that one is observing decisions influenced by biased thinking. In either case, it seems apparent that the lower class client would not be offered the equivalent therapy available to the middle class client. Issues/Presenting Problems Respondent differences linked to client social class also occurred in the area of presenting problems. The total number of problems generated by the respondents for the lower class client was slightly less than those generated for the middle class client. In 1 'Vocational overshadowing' occurs when a client presents with both personal and vocational issues, and a preference is shown on the part of the therapist to focus more on the personal issues. 117 addition, the issues of anxiety, family violence, and paranoia were generated significantly more frequently for the middle class client, while the issues of feelings/mood and vocation were generated significantly more frequently for the lower class client. A greater portion of these differences were due to responses from the CGCA sample population. Findings bv Association. CGCA respondents produced significant frequency differences on four issues. Family violence and paranoia were generated significantly more frequently for the middle class client while coping with feelings/mood and vocation were generated significantly more frequently for the lower class client. In addition, the CGCA respondents produced an overall frequency for presenting problems of 134 for the middle class client and 110 for the lower class client. BCPA respondents generated the one issue, anxiety, significantly more frequently for the middle class client than for the lower class client. Overall frequencies for the middle and lower class client were virtually identical (f= 80 and 82 respectively). It appears that, on this question, BCPA respondents exhibited less difference in perception of the problem between the middle and lower class client than did respondents from the CGCA sample population. The average levels of education were a master's degree for the CGCA sample population, and a doctoral degree for the BCPA sample population. Significant differences on the Client Characteristics were linked to levels of education. This may have been the determining factor in significant differences in frequencies of presenting problems; however, statistical investigation of presenting problems was not carried out by respondent level of education. Findings by Gender. Female respondents generated a significantly higher frequency of the issue of family violence for the middle class client. The issue of paranoia 118 produced a tendency towards difference (p = 06), with a higher frequency generated for the middle class client. Female respondents generated the issue of unemployment significantly more frequently for the lower class client. Male respondents produced no significant frequency differences on any of the presenting problems. While one of the frequencies reached significance (anxiety,/? = .04), total frequencies were too low for chi square analysis. Initially, one might assume that the male respondents are exhibiting no difference between the middle and lower class client in issues generated, however, results are inconclusive because, on 15 out of 18 issues generated, cell frequencies were too low to carry out reliable analyses. Overall, the male respondents generated lower frequencies (f= 174) than the female respondents (f= 232) for all of the presenting problems. The most frequently generated issue by total sample population, by gender and by CGCA respondents was family violence. Family violence was generated significantly more frequently for the middle class client than for the lower class client. Recent studies (Davidovich, 1990; Peterson, 1980; Smith, 1990) have revealed an increase in family violence moving from the middle to lower social class. These studies also argue that the lack of finances and certain forms of social supports increase the possibility of family violence. Economic factors tend to be the strongest reasons why women remain in violent situations (Ignagni et al., 1988). Why then, was family violence as an important issue, noted significantly less frequently for the lower class client? One possibility may be that counsellors/therapists were less likely to note family violence as an important issue for the lower class client because violence is more frequently expected (less shocking) when occurring within the context of poverty or 119 limited financial resources (Morris & Williamson, 1982). Counsellors may downplay family violence related to the lower class client because they implicitly recognize the violence as a result of the more difficult socio-economic situation the lower class family faces. While plausible, this explanation still appears weak in that recognition of social context should not blind one to the issue. An alternative explanation for the apparent lowered emphasis on the issue of family violence may be linked to schema theory and stereotyping. When faced with an individual from a social group other than one's own, one is likely to be less perceptive of a variety of particular aspects and characteristics of that individual than for individuals who fall within or close to one's own social group (Fiske & Taylor, 1984; Jones, 1988; Leyens et al, 1994; Morris & Williamson, 1988). The less one identifies with an individual, and the more one is affected by the stereotype, the less apt one is to take note of all of the specific issues or characteristics that make up that particular individual. Studies have shown that perceptions of out-group members show much less variability and complexity than perceptions of in-group members (Fiske & Taylor, 1984). While therapists are trained, more than most professionals, in acceptance of individuals different from themselves, research has shown that they are not immune to stereotypes or their immediate activation when facing out-group members (Smith, 1988). This being the case, therapists who interact with a lower social class client are faced with a stereotype; and, because of this, may be less apt to notice as many of the predominant issues. The lowered frequencies of issues generated for the lower social class client in this particular section of my research parallels results of previous out-group schema research (Fiske & Taylor, 1984). 120 Suggested Length of Stay in Therapy No significant difference was measured in suggested length of stay in therapy when results were examined by total sample population, or by respondent gender. By total sample population, the lower class client was offered approximately 13 sessions and the middle class client was offered approximately 15 sessions. By gender, female respondents suggested approximately 13 sessions of therapy for the middle class client and 14 sessions of therapy for the lower class client. Although the male respondents produced noticeably different results in suggested length of stay in therapy linked to client social class (18 sessions for the middle class client and 12 sessions of therapy for the lower class client), results were not significant. This is, most probably, due to the wide range in responses for this question, producing large standard deviations, and indicating that, on this particular question, the male respondents were not a homogeneous group. Findings by Association. Differences were significant when results were examined by professional association. One might intuit that the lower class client would be determined to need approximately the same number of therapy sessions (if not more) as the middle class client. BCPA respondents suggested a nonsignificantly longer length of stay in therapy for the lower class client than for the middle class client (Ms = 14.79, 12.45 sessions respectively, p = .37). However, CGCA respondents suggested a statistically significant shorter length of stay in therapy for the lower class client than for the middle class client (Ms = 11.71, 17.68 sessions respectively, p = .05). One explanation could be that some respondents perceived the lower class client as a less appropriate candidate for long-term therapy. The lower class client was perceived by some respondents as less psychologically-minded, and therefore, less apt to respond 121 well to insight-oriented therapy. However, as noted in the Literature Review, when lower class clients have been offered long-term therapy in an accepting environment, dropout rates have not been significantly higher than those of middle class clients (Beutler, 1981; Lerner, 1972). Nearly all of the respondents fall within and originate from higher social class backgrounds. As pointed out by schema research (Fiske & Taylor, 1984; Leyens et al, 1994), seeing an individual as different from ourselves produces less of an attraction to that individual. This possible perceived difference between respondent and client could have produced lower motivation on the part of the respondents to provide long term therapy for the lower class client. However, this does not explain the differing results between the CGCA and BCPA respondents. Another possible explanation for the significant difference on the part of CGCA respondents could be that the client's ability to pay was taken into account in respondents' estimates of suggested length of stay in therapy. It may be argued that more BCPA members are able to charge fees for services directly through the provincial medical plan -an option not available to most CGCA members. However, since only one CGCA respondent mentioned ability to pay, it is difficult to ascertain whether this had any significant impact in their decisions. Previous studies dating back to the 1950s (Hollingshead & Redlich, 1958) showed that, even with payment guaranteed, therapists displayed a lowered interest in treating lower social class clients. Both previous (Dohrenwend, 1973; Kessler & Geary, 1980) and more recent (Harder et al, 1990) longitudinal studies on levels of psychological distress and prognosis have revealed that lower class individuals develop greater levels of psychopathology and 122 take longer to recover. Virtually all health studies (Canadian Public Health Association, 1997) reveal that health is directly related to socio-economic status, and individuals from the lower socio-economic levels are more apt to fall ill and take longer to recover from a wide range of both physical and mental illnesses. This being the case, it would seem more logical and effective to offer the greater length of stay in therapy to the lower class client. Problem Evaluation Results In reviewing results of the problem evaluation, no differences linked to social class were found to be statistically significant when examined by total sample population or by association. Female respondents to the lower class client produced a significantly lower prognosis when compared with female respondents to the middle class client. While it is not possible to determine with confidence why female respondents rated the lower class client less positively on prognosis, there are several possible explanations. One explanation may be that female respondents were demonstrating more bias than male respondents. This seems less likely however, when one considers the rating results on the client characteristics. Female respondents' evaluations produced less significant differences linked to client social class on more of the variables than did male respondents' evaluations. A more plausible explanation for this difference may be that female respondents were somewhat more apt to take into account social contexts and therefore, provide a more realistic prognosis for the lower class client considering the added stressors that accompany poverty, social bias, and lower access to and availability of a variety of social supports. Since this particular evaluation was the only one to reach significance, one very 123 possible explanation may be the occurrence of a Type I error. Choice of Therapy No significant differences arose in respondent choice of therapy offered and client social class. BCPA respondents displayed a slightly higher tendency to offer cognitive therapy to the middle class client (middle class client = 70%, lower class client = 54%). Whether this tendency is linked to a belief that lower class clients would benefit less from an insight-oriented form of therapy is difficult to ascertain. Research carried out in the 1950s - 1970s suggested that lower class clients benefited less from psychodynamic psychotherapy (Hollingshead & Redlich, 1958; Lorion & Felner, 1986). However, Brill and Strorrow (1960) found that long-term psychoanalytic treatment was offered much more frequently to middle and upper class clients, while drug therapy and inpatient services were offered more frequently to lower class clients regardless of whether the presenting problems were alike. Beutler's (1981) study found no outcome differences linked to client social class, using cognitive-based therapy. While CGCA respondents offered cognitive therapy equally to both middle and lower class clients (middle class client = 50%, lower class client = 50%), behavioural therapy was suggested slightly more frequently for the lower class client (middle class client = 41%, lower class client = 50%). Behaviour therapies focus on producing a behavioural change. With the added stressors impacting on almost every aspect of life in the lower social class, a focus on effecting behaviour change may have been viewed by some as more of an imperative for the lower class client. Brill & Strorrow (1961) and Lorion & Felner (1986) found that lower class clients desired somewhat more direct intervention on the part of the therapist, and therapy aimed at symptom relief. Further 124 research in this area could help clarify whether, because of the more difficult social situation for many lower class clients, a behavioural approach might provide more options for short-term change. Summary It is evident in these findings that respondents demonstrated a differential perception, evaluation and proposed treatment of the lower social class client. Comparison by professional association produced inconclusive results. While BCPA respondents produced more negative evaluations of the lower class client on the client characteristics, CGCA respondents produced more significant differences in the frequencies and choices of presenting problems and in the suggested length of stay in therapy also linked to client social class. Findings on the client characteristics linked to gender do appear to lend limited support to my hypothesis that, due to a heightened awareness of gender biases and the inclusion of social context within feminist counselling, female respondents produced less negative evaluations of the lower class client than male respondents. However, results on the problem evaluation, presenting problems and suggested length of stay in therapy linked to respondent gender and client social class remain inconclusive. In comparing results by respondent levels of education, bias appeared more strongly among master's respondents than among doctoral respondents on the client characteristics. This finding is in contradiction to my original hypothesis that higher levels of education would produce more bias. 125 Implications Clinicians employ the same method of cognitive processing as other individuals. Initial intake contacts with a particular client, including client demographic information, presenting problem and physical appearance, may trigger schemas categorising the client as a member of a particular out-group (Smith, 1988). The findings suggest that therapists may use indicators of socio-economic status to evaluate a client's suitability for therapy as well as to evaluate a client's characteristics in terms that are linked to social class stereotypes. This can lead to the perpetuation of social class stereotyping and class bias in the clinical setting (Hertzberg & Eschbach, 1982). Under these circumstances, it is realistic to infer that the lower social class client is not offered as thorough or fair a therapeutic assessment or environment as is offered to the middle class client. Until recently, social context was given short shrift within the counselling setting. Individualistic explanations for a client's problems and life situation dominated the field. This atmosphere tends to be one that places responsibility as well as blame for most life situations on the individual. This neglect of the social context allowed for little understanding or empathy for the realities for lower class individuals within a society where there are, in fact, considerably more obstacles and less opportunities in nearly every aspect of their lives. Some may argue that lower and middle class clients, due to a variety of external and internal factors, are different and, therefore, therapists are evaluating the lower class client based on a correct judgement or evaluation. One might realistically expect certain client differences linked to client social class. However, research has been inconclusive (and/or disregarded) on exactly what many of these differences may be and how they 126 might be best handled in a counselling setting. If, as research suggests, some real differences linked to client social class do exist, then both research into the exact differences and the provision of appropriate academic training for counselling lower social class clients is significantly lacking. In view of the fact that, within the category of clients on the Personal Network Grid, the most frequently checked off category was welfare recipient, more effective and thorough treatment programs need to be designed and integrated into already existing programs. On the other hand, if, as suggested by Sutton and Kessler's (1986) study and in the results of this work, therapists are exhibiting negative evaluation linked to social class stereotypes, then academic as well as other training programs must actively participate in moderating and eliminating social class bias among its trainees and trainers. It is probable that both negative stereotyping and actual differences are in effect when counselling lower social class clients. The implication of both of these issues is that lower social class clients are evaluated more negatively, and may not be offered as positive or thorough a therapeutic environment as the middle class client. It is important for graduate programs to develop and offer courses and written material that educate graduates about social class and class stereotyping. Further, closer ties with departments that have a tradition of carrying out research on class issues and have already proven to be more responsive to social class need to be encouraged and cultivated. Non-university training programs, frequently attended by practicing professionals, also need to incorporate social class into their existing agendas. Cultural and gender training and the inclusion of culturally and gender relevant materials within university counselling programs serves to increase counsellors' awareness 127 of and sensitivity to the importance of culture and gender on one's life experiences. The lack of focus on social class as a predominant factor can serve to perpetuate class stereotypes, and maintain a lack of awareness of the effects of social class on psychological health and wellbeing. When one considers therapists' years of training, clinical supervision, and the constant need to demonstrate at least a moderate degree of compassion and acceptance of people dissimilar to themselves, the finding that ignorance or bias continues to persist within this group, suggests that social beliefs and stereotypes may be deeply ingrained. Research in schema theory and stereotyping has revealed that stereotypes are long lasting and are resistant, although not impervious, to change (Fiske & Taylor, 1984). For individuals needing and expecting unbiased assistance in a therapeutic environment, the likelihood that they will be faced with persistent social biases, as demonstrated in previous research and in this study, constitutes a severe indictment of the profession. Strengths and Limitations This study has several strengths. By providing all respondents with the same case history, with the only differences being in aspects of social class and class background, it was possible to clearly isolate the effects of the social class information for analysis. In addition, by provided a Caucasian male as the client, it was possible to avoid the confounds of gender and race effects. To my knowledge, there has never been a study of therapist and/or counsellor judgement differences based on client social class in Canada. Using members of both the CGCA and the BCPA in British Columbia, allowed for a representative sample of counsellors and therapists across the province, and cross-comparison based on a variety of respondent demographics (e.g. degree, level of 128 education, gender, association membership, etc.). Therefore, this study, while partially replicating a similar American study carried out in 1981, may be more representative of Canadian therapist and counsellor responses. Another strength is the return rate on the questionnaire mailout. More than one-third (N = 111) of the random sample (N = 300) returned completed questionnaires, providing a large sample size from which to carry out analyses and comparisons. Analogue studies also contain certain built-in weaknesses. As with all analogue studies, a major limitation of the current research lies in its generalizability to real life counselling practice. It is not always possible to infer that responses to a written case study are analogous to actual clinical evaluation. It is difficult to determine whether negative clinical perception, treatment and evaluation would be lessened or increased within an actual clinical setting with real clients. However, some studies suggest that individuals who display significant racial bias in responses to written questionnaires demonstrate higher levels of bias when tested in real life situations (Crosby et al., 1980). Many counsellors and therapists do rely on written summaries in client evaluation and therapist decision-making. By presenting the respondents with as much information as possible, in the form of a personal information sheet, results of psychological testing and the initial interview, I hoped to replicate a clinical evaluation as closely as possible. Another limitation is that this research does not extend into an understanding of potentially real differences between middle and lower class clients. The lines between actual differences and bias are not clearly deligneated. This is an important point because, if differences are real, therapists need to learn more about these differences in order to increase an understanding of and comfort for the client. 129 Other limitations include the fact that the hypothetical client was male and findings may not necessarily apply to lower class female clients. In addition, the results represent responses within the province of British Columbia only and may not be generalizable to other areas of the country. One further limitation was related to respondents' social class and class background. Studies have found that client/therapist similarity of background and values has produced positive treatment and outcome ratings (Sladen, 1982). However, the significantly low numbers of lower social class background respondents made it impossible to determine whether this would have had a positive impact on clinical evaluation. If the sample population is representative of the British Columbian general population of therapists and counsellors, it is reasonable to assume that very few come from lower class backgrounds. One might assume that negative stereotyping might be lower among individuals whose class background was similar to those of lower class clients. On the other hand, receiving education in a predominantly middle class environment and value system, with material focused and based predominantly on middle class populations may impact on the perceptions of all students. Suggestions For Future Research Recent published research in the area of therapist perceptions, evaluations and treatment of lower socio-economic status clients and their effects on outcome in therapy is significantly lacking both within Canada and the United States. A Canadian-wide study linked to client social class could help to shed light on whether differences noted in this study are particular to British Columbia or reflections of a nation-wide difference in therapists' perceptions, evaluations and treatment linked to 130 client social class. Since stereotypes tend to be enduring and difficult to change, it is also important to design research that examines ways to initiate stereotype change and incorporate this into training designed to educate students about social class. In conjunction with this is the need for studies of possible actual differences in clients from differing social class backgrounds, in what ways these differences affect the counselling experience, and how best to work with lower social class clients. Studies of the impact of differing counsellor and client value systems, and in what ways these impact on the counsellor/client relationship and counselling experience, might also provide insightful and useful information both for training and clinical purposes. A review of course materials and training programs to examine possible social class bias (either explicitly or by way of exclusion of class context) in both research and education could also be undertaken and prioritised. Few individuals from lower socio-economic levels reach graduate levels of education. A study to explore lower and working class students' experiences of attending institutions that reflect and uphold middle class culture, values and ideology might be enlightening. Since few lower and working class students continue on into graduate programs, investigation of experiences of students, from a social class perspective, could shed light on the obstacles or hindrances they might face, and deterrences from entrance to and continuation in graduate school. Because social class is more invisible than culture and gender, and because North American society encourages a belief that a class system does not exist here, lower class students remain, to some extent, invisible. This can serve to deny or invalidate their experiences and values systems, and allows the assumption that 131 all students share similar histories, values and beliefs. Another possible area of research is a qualitative analysis and comparison of both lower, working and middle class individuals' experiences in counselling. Psychological research linked to the study of the lower social classes has been minimal. As described by Reid (1992), personal affiliation leads individuals to study populations more like themselves. For example, until women began participating in producing research, most studies involved male populations. In a similar manner, most research, including the testing of theories and standardized tests has been carried out on white, middle class populations. Lower class populations are then compared to this standard. Because few individuals from the lower social classes reach higher levels of academia (or have done so despite barriers associated with social class), little enthusiasm has been shown for extensive studies of lower social class populations. This in itself, has meant an inadequate focus on a significantly large portion of our entire population. Conclusion Over the past twenty years, a variety of cross-cultural and feminist counselling models have been developed, as well as guidelines for counselling women and ethnic groups. While some of these address the issue of counselling the poor, frequently this is done only in so much as the poor are members of an ethnic minority or single-parent women. The lack of training or focus on the specific needs and life-experiences of all individuals within the lower social classes would appear to perpetuate their invisibility and the notion that they are more responsible for their negative life situations. If therapists perceive lower class clients as less suitable for therapy, and as exhibiting less positive outcomes, do we then believe that the therapists' expectations and 132 beliefs have no effect on the course and outcome of therapy? If an individual is already uncertain about involvement in mental health treatment, the lower socio-economic status client's awareness of therapist messages may constitute a hindrance to the client's therapeutic process and outcome. Lower socio-economic status clients have limited access to mental health and other counselling services. In addition, the services available are frequently less culturally and socially relevant than for more privileged clients. The emphasis within our society and within counselling to focus on individual action as the prime determinant of success or failure adds an implicit negative judgement on those less economically successful. To see a person as separate from the contexts of his or her life devalues the experience of many individuals, handicaps the therapist's capacity to provide help and facilitate change, and in addition does not allow for examination and questioning of one's own class based beliefs and value structures. Clinicians employ the same method of cognitive processing as other individuals. Initial intake contacts with a particular client, including presenting problem and physical appearance, appear to trigger schemas categorising the client as a member of a particular out-group (Smith, 1988). If therapists' perceptions, evaluations and treatment choices affect the counselling experience in ways that may be either less helpful or detrimental to a particular group of clients, then research must shed light on these issues in order to develop ways of educating therapists, and in so doing create the possibility of providing a safer and healthier helping environment for all clients regardless of race, culture, age, gender, and the neglected and often hidden variable of social class. 133 Bibliography American Psychiatric Association. (1995). Diagnostic and statistical manual of mental disorders (4th ed., rev.). Washington, DC: American Psychiatric Association. Arkes, H. R. (1981). Impediments to accurate clinical judgement and possible ways to minimize their impact. Journal of Consulting and Clinical Psychology, 49(3), 323-330. Baer D. E. & Lambert R. D. (1982). Education and support for dominant ideology. Canadian Review of Sociology and Anthropology, 19(2), 173-195. Beckham, E. E. (1992). Predicting patient dropout in psychotherapy. Psychotherapy, 29(2), 177-182. Belle, D. (1990). Poverty and women's mental health. American Psychologist, 45(3), 385-389. Berrigan, L. P. & Garfield, S. L. (1981). Relationship of missed psychotherapy appointments to premature termination and social class. The British Journal of Clinical Psychology, 20, 239-242. Beutler, L. E. (1981). Convergence in counseling and psychotherapy: A current look. Clinical Psychology Review, 1, 79-101. Brill, N . Q. & Storrow, H. A. (1961). Social class and psychiatric treatment. Archives of General Psychiatry, 3, 340-344. Bullock, H. E. (1995). Class acts: Middle-class responses to the poor. In B. Lott and D. Maluso (eds.), The social psychology of interpersonal discrimination, (pp. 118-159), New York: Guilford Press. Canada Council on Social Development (1996). An Introduction to Poverty in Canada. The Canadian Fact Book on Poverty. Sept, 1996. Canada Council on Social Development, Ottawa. Canadian Public Health Association (1997). Health impacts of social and economic conditions: Implications for public policy. Board of Directors Discussion Paper. March, 1997. Health and Welfare Canada, Ottawa. Cardea, C. (1985). The lesbian revolution and the 50 minute hour: A working-class look at therapy and the movement. Lesbian Ethics, 7(3), 47-68. Colclough, G. & Beck, E. M . (1986). The American educational structure and the reproduction of social class. Social Inquiry, 36, 456-476. 134 Crocker, J., Fiske, S. T. & Taylor, S. E. (1984). Schematic bases of belief change. In J. R. Eiser (ed.), Attitudinaljudgement (pp. 197-226). New York: Springer-Verlag Inc. Davidovich, J. R. (1990). Men who abuse their spouses: Social and psychological supports. In Clinical treatment of the criminal offender. (PP. 27 - 43). Haworth Press Inc. Crosby, F., Bromley, S. & Saxe, L. (1980). Recent unobtrusive studies of black and white discrimination and prejudice: A literature review. Psychological Bulletin, 87(3), 546-563. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56(1), 5-18. Dohrenwend, B. S. (1973). Social status and stressful life events. Journal of Personality and Social Psychology, 28(2), 225-235. Ehrenreich, J. H. (1990). Effect of social class of subjects on normative responses to TAT cards. Journal of Clinical Psychology, 46(4), 467-471. Evans, L. A., Acosta, F. X., Yamamoto, J., & Skilbeck, W. (1984). Orienting psychotherapists to better serve low income and minority patients. Journal of Clinical Psychology, 40, 90-96. Fiske, S. T. & Taylor, S. E. (1984). Social cognition. New York: Random House. Foschi, M. , Lai, L. & Sigerson, K. (1994). Gender and double standards in the assessment of job applicants. Social Psychology Quarterly, 57(4). Frank, A., Eisenthal, S., & Lazare, A. (1978). Are there social class differences in patients'treatment conceptions? Archives of General Psychiatry, 35, 61-69. Frank, G. (1994). Socioeconomic status and the Rorschach. Psychological Reports, 74, 95-98. Frank, J. W. & Mustard J.F. (1995). The determinants of health from a historical perspective. Daedalus, Journal of the American Academy of Arts and Sciences, 123(4), 1-17. Garfield, S. L. (1986). Research on client variables in psychotherapy. In S.L. Garfield and A. E. Bergin (eds.), Psychotherapy and behaviour change (pp. 213-256). New York: Wiley. Gecas, V. & Seff, M.A. (1989). Social class, occupational conditions, and self-esteem. Sociological Perspectives, 33(3), 353-364. 135 Gough, H. G. & Heilbrun, A. B., Jr. (1980). The Adjective Check List Manual (Rev.ed.). Palo Alto, Ca: Consulting Psychologist Press. Griffin, W. E. & Oheneba-Sakyi, Y. (1993). Sociodemographic and political correlates of university students' causal attributions for poverty. Psychological Reports, 73, 795-800. Guppy, N., Mikicich, P. D. & Pendakur, R. (1984). Changing patterns in educational inequality in Canada. Haverkamp, B. E. (1993). Confirmatory bias in hypothesis testing for client-identified and counselor self-generated hypotheses. Journal of Counseling Psychology, 40(3), 303-315. Harder, R. C , Strauss, J. S., Greenwald, D. F., Kokes, R. F., Ritzier, B. A , & Gift, T.E. (1990). Predictors of outcome among adult psychiatric first-admissions. Journal of Clinical Psychology, 46(2), pp. 119-128. Herr, E. L. (1989). Counseling in a dynamic society. Alexandria, VA: American Association for Counseling and Development. Herr, E. L. & Niles, S. (1988). The values of counseling: Three domains. Counseling and Values, 33, 4-17. Hertzberg, J. (1990). Feminist psychotherapy and diversity: Treatment considerations from a self-psychology perspective. Women and Therapy, 9(3), 215-291. Hertzberg, J. & Eschbach, A. (1982). Social class and therapy: The dynamics of exclusion. Issues in Radical Therapy, 10(2), 6-11 & 50-53. Hillerbrand, E. (1988). The relationship between socioeconomic status and counseling variables at a university counseling center. Journal of College Student Development, 29, 250-254. Hollingshead, A. B., & Redlich, F. C. (1958). Social class and mental illness. New York: Wiley. Hughes, N. (1987). Why can't I write about class. Fireweed: A Feminist Quarterly, 25, 21-24. Ignagni, E., Parent, D., Perreault, Y. & Willats, A. (1988). Around the kitchen table: Toronto rape crisis centre working class caucus. Fireweed: A Feminist Quarterly, 26, 69-8L Jones, R. A. (1988). Perceiving other people: Stereotyping as a process of social cognition. In A. G. Miller (Ed.), In the eye of the beholder (pp. 41-92). New York: Praeger Publishers. 136 Kadushin, C. (1969). Why people go to psychiatrists. New York: Atherton. Kohn, M . L. (1969). Class and Conformity: A Study in Values. Chicago: University of Chicago Press,. Kohn, M . L. & Schoenbach, C. (1993). Social stratification, parents' values and children's values. In D. Krebs and P. Schmidt (eds.), New directions in attitude measurement (pp. 119-151). New York: Walter de Gruyer. Lareau, A. (1987). Social class differences in family-school relationships: The importance of cultural capital. Sociology of Education, 60(2), 73-85. Lerner, B. (1972). Therapy in the ghetto. Baltimore: John Hopkins University Press. Lerner, B., & Fiske, D. W. (1973). Client attributes and the eye of the beholder. Journal of Consulting and Clinical Psychology, 40, 272-277'. Leyens, J. P , Yzerbyt, V. & Schadron, G. (1994). Stereotypes and social cognition. Thousand Oaks, Ca: Sage Publications Inc. Lockhart, L. L. (1987). A reexamination of the effects of race and social class on the incidence of marital violence: A search for reliable differences. Journal of Marriage and the Family ^ (August 1987), 603-610. Lorion, R. P. (1974). Patient and therapist variables in the treatment of low-income patients. Psychological Bulletin, 81(6), 344-354. Lorion, R. P. & Felner, R. D. (1986). Research on mental health interventions with the disadvantaged. In S.L. Garfield and A. E. Bergin (eds), Psychotherapy and behaviour change (pp. 739-775). New York: Wiley. Lorion, R. P. & Parron, D. L. (1985). Countering the countertransference: A strategy for treating the untreatable. In P. Pererson (Ed), Handbook on cross-cultural counseling and therapy (pp. 79-86). New York: Greenwood Press. McCarthy, P. R , Reese, R. G., Schueneman, J. M. , & Reese, J. A. (1991). Counselling working class women. Canadian Journal of Counselling, 25(4), 581-593. Martin, D. (1977). Battered Wives. New York: Pocket Books. Millar, J. (1993). Determinants of health. Unpublished manuscript. Provincial Health Officer. Victoria, B.C. 137 Morris, M., & Williamson, J. B. (1982). Stereotypes and social class: A focus on poverty. In A. G. Miller (Ed.), In the eye of the beholder (pp. 411-465). New York: Praeger Publishers. Morrow, K. A , & Deidan, C. T. (1992). Bias in the counseling process: How to recognize it and avoid it. Journal of Counselling and Development, 70, 571-577. Miiller, W. (1993). Social structure, perception and evaluation of social inequality and party preferences. In D. Krebs & P. Schmidt (Eds), New directions in attitude measurement (pp. 94-117). New York: Walter de Gruyter. Osgood, C. E., May, W. H. & Miron, S. M. (1970). Cross cultural universals of affective meaning. Chicago: University of Illinois Press. Osgood, C. E. (1984). The nature and measurement of meaning. In C. E. Osgood, & O. C. S. Tzeng (Eds.), Language, meaning, and culture: The selected papers of C. E. Osgood (pp. 159-202) New York: Praeger. Pederson, P. (1981). Cross Cultural Training For Counselors and Therapists. In P. Pererson (Ed.), Handbook on cross-cultural counseling and therapy (pp. 79-86). New York: Greenwood Press. Pettit, I., Pettit, T., & Welkowitz, J. (1974). Relationship between values, social class, and duration of psychotherapy. Journal of Consulting and Clinical Psychology, 42, 482-490. Peterson, R. (1980). Social class, social learning, and wife abuse. Social Service Review, September, 390-406. Pineo, P. C. & Looker, E. D. (1983). Class and conformity in the Canadian setting. Canadian Journal of Sociology 8(3), 293-317. Powell, S. (1992). Social class: Therapists'awareness and training. Unpublised doctoral dissertation. Southern Illinois University at Carbondale. Reid, P.T. (1993). Poor women in psychological research: Shut up and shut out. Psychology of Women Quarterly, 17, 133-150. Rosenbaum, J. E. (1976). Making inequality: The hidden curriculum of high school tracking. New York: John Wiley and Sons, (in Colclough, G. & Beck, E. M. (1986). The American educational structure and the reproduction of social class. Social Inquiry, 36, 456-476.) Rosenberg, M. & Pearlin, L. I. (1978). Social class and self-esteem among children and adults. American Journal of Sociology, 84(1), 53-77. 138 Rubin, L. B. (1976). Worlds ofpain: Life in the working-class family. New York: Basic Books Inc. Sennett, R, & Cobb, J. (1972). The hidden injuries of class. New York: Knopf Inc. Smith, A. F. (1988). Perceiving the client. In D. C. Turk & P. Salovey (Eds.), Reasoning, inference, andjudgment in clinical psychology (pp. 73-89). New York: Free Press Inc. Smith, D. M . (1990). Sociodemographic risk factors in wife abuse: Results from a survey of Toronto women. Canadian Journal of Sociology, 15(1), 39-58. Spengler, P.M., Blustein, D.L. & Strohmer, D.C. (1990). Diagnostic and treatment overshadowing of vocational problems by personal problems. Journal of Counseling Psychology, 37(4), 372-381. Statistics Canada (1995). Income Distributions by Size in Canada 1993. Ottawa, Statistics Canada, 1994. Cat. 13-207. Strohmer, D. C , Shivy, V. A. & Chiodo, A. L. (1990). Information processing strategies in counsellor hypothesis testing: The role of selective memory and expectancy. Journal of Counseling Psychology, 37(4), 465-472. Sutton, R. J. (1983). A national study of clients' SES on clinical psychologists' professional judgments. Unpublished doctorate thesis. University of Vermont. Sutton, R. G. & Kessler, M . (1986). National study of the effects of clients' socioeconomic status on clinical psychologists' professional judgments. Journal of Consulting and Clinical Psychology, 54(2), 275-276. Tindall, D. (1995). The Vancouver Island Wilderness Preservation Movement Study. (research in process). Wilkins, R. (1988). Special report on the socially and economically disadvantaged. Health and Welfare Canada. Wright, E. O., Costello, C , Hachen, D. & Sprague, J. (1982). The American class structure. American Sociological Review, 47, 709-726. 140 Appendix A THERAPIST DECISION-MAKING Directions; 1. Remove, read, and retain the letter of consent. 2. Read the client Personal Information Form, Standardized Testing Results and Intake Interview on pages one and two of this questionnaire booklet. 3. After reading all of the client information, complete the questionnaire in order. Please do not look ahead or go back to change your answers once you have completed them. 4. If you wish to enter the prize drawing, fill in your name and address on the attached form and seal it in the small envelope. 5. Return the questionnaire and sealed draw envelope to me in the large stamped and addressed envelope. 6. If you would like to receive a synopsis of the results, mail in the Request For Synopsis of Results form in a separate envelope. Do not include this in the draw envelope or with your questionnaire. This questionnaire will take approximately one-half hour of your time and contains a number of questions that ask for your impressions of a hypothetical client. We recognize that no client can be described in absolute terms with such limited information. However, for the purposes of this research, your best estimate of characteristics and probable results is requested. Thank You. 141 P E R S O N A L INFORMATION F O R M This form is to help your counsellor become better acquainted with you Please fill in as much as you can Omit any item you do not want to answer. Name / & ^ a J c " : Date / / (Last) (First) (Initial) (d) (m) (y) A q e ^ V Date of birth / / Sex(M /F)_£2_ S3 . (d) (m) (y) Ethniclty/Raclal origin C^^i^LLa^ Religious Preference O i c r v u i ^ Education: Number of years 3 Degrees _ Occupation r^^iL.T^y^ L L*JLj>J Presently. Employed Unemployed < / _ Marital Status: Single Manned \ * / Separated Divorced Widowed How many times have you been married 2, Number of children: 3 How many of them live with you? »3_ Family Background While you were growing up, were your parents: MarriecWiving together Divorced ^ Separated Father remarried Mother remarried Deceased: Mother Father Father's occupation ^o^JL Mother's occupation -^j^r^L^^Se^fu Father's education ^sr^i^JU^ 6> Mother's education <^/ui^*kes Number of brothers and sisters # How many were younger/older than you? Younger — Older 2 - _ Physical Conditions and Characteristics Heioht£>gl. /jLu. . Weight ZOQ^JL. Physical disabilities or health problems? (Describe) : Any lengthy hospitalizations or severe injuries? (Describe) U Q * Presenting Problem^) Why you have chosen to come for counselling? Underline any of the following that apply toyou : F requen t h e a d a c h e s , b a c k a c h e s , d r ink ing or drug p rob lems, d izz iness , overeat ing, poor appet i te , vomi t i ng , d iges t ive p rob lems , s l e e p l e s s n e s s , fa int ing spe l l s ,pens ion , con t inued t i redness , tea r fu lness , v i sua l or hea r ing p r o b l e m s . 142 Name: Results of standard psychological testing showed elevations on the following scales: Anxiety: 80th percentile Depression: 82nd percentile Paranoia. 71st percentile Intake Interview: Robert -is a talZ, slightly stocky man, who arrived {or this {-Out session dressed in a T-shirt and jeans. Robert sat somewhat sti{{ly throughout the {irst meeting, and made. onZy intermittent eye.-contact. H-ii predominant mood appeared to be somewhat agitated and anxious. Robert stated that he, tost kis job -in the, past year and -is currently supporting hti {amiZy on monthly wel{are cheque*, Robert haA concerns about being able, to care {orhts {amity. Hehashadmany bout* o{ unemployment over the years. Robert's {other lc{t the family when Robert was quite young, leaving hi* mother to ratse the £ive children on welfare. Robert shared that one week, previously, hi* wi{e had &sked him {or tome money. Robert admitted to her that he had blown mo-it o{ h-is welfare cheque that week, which he stated he has problems stopping himsel{ {rom doing lately, H-is wi{e had become angry at thts, and Robert lost his temper and broken several objects before slapping hi* wile to stop her {rom "nagging " him. Robert also mentioned that one day this past week he had kept bis children home {rom school, as he {eared he would never see them again. H-U wi{e was not happy with his behavior and told him he needed to "get his act together ". He described his thoughts recently as racing all the time lately, including at night, preventing him {rom sleeping. He also stated that he {orgets things, and {eeli cu>i{heis going crazy. He holds tittle hope that counselling will be help{ul. Robert alio said that he has been drinking more than usual recently. Now that you have read the client information, please answer the questions in each of the sections that follow. You will be asked questions about the client, his concerns, and possible choices for treatment 143 PART I: Case History Rating Scales: Please circle one number on each of the bar graphs that best describes your impression of the client. 1. Severity of presenting problem: Very Severe 1 2 Very Mild 2. Perceived client motivation for change: Very Strong 1 2 3 _ Very 7 Weak 3. Perceived client self-concept: Excellent 1 Poor 4. Prognosis: Favourable Bleak 5. Likelihood of referring this client for psychiatric assessment: Very High 1 2 3 4 5 Very 7 Low 6. Your personal interest in treating this client: Very Strong 1 2 3 _ Very 7 Weak Case History Questions: 1. Many counsellors modify their approach from one client to another. Please describe the approach (e.g. structured vs open-ended, directive vs nondirective, cognitive vs behavioral vs affective, etc.) you would use with this client and why. 144 2. What problems and issues do you feel it would be important to address first with this client? 3. Recommended number of sessions in counselling/therapy: PART H: Client Characteristics: Based on your brief first impression, please check to what extent each of the following characteristics might best describe the hypothetical client. Your first response is best. Please do not refer back to the case history as you answer. expressive pessimistic trustworthy competent unintelligent involve industrious impulsive responsible ignorant friendly loud clean emotional moral incapable defensive aggressive dishonest very closely only not only closely very closely describes slightly related slightly describes closely describes client describes describes client describes client client client client inexpressive optimistic untrustworthy incompetent intelligent withdrawn lazy reliable irresponsible knowledgeable hostile quiet dirty controlled immoral capable receptive peaceful honest 145 PART HI: Personal Network: For this section, please indicate whether you know anyone in each of the occupations listed. Do not put more than one tick for any one person. Do not put more than one tick in any one box. Example of how to answer this section: TYPE OF JOB: TYPE OF RELATIONSHIP: Do you know anyone who works in the following industries? Acquaintance Close Friend Relative Colleague Client Example 1: secretary X x Example 2: counsellor X X x Example 1, would indicate that you know an acquaintance and a client who are both secretaries. Example 2, would indicate that you know a close friend, a relative and several colleagues, all three of whom are counsellors. TYPE OF JOB: TYPE OF RE LATIONSHIP: Do you know anyone who works in the following industries? Acquaintance Close Friend Relative Colleague Client business owners or managers university or college professors primary or secondary school teachers bankers accountants engineers bus drivers seasonal farm workers or fishers auto mechanics plumbers/electricians waiters/waitresses (nonsupervisory) police officers loggers/mill workers firefighters (non volunteer) provincial/federal politicians computer/electronic specialists biologists/botanists/chemists nurses/physiotherapists doctors architects lawyers unskilled construction workers janitorial/cleaning personnel gas station attendants unskilled factory workers longterm welfare recipients truck drivers (non owners) social workers 146 PART IV: Demographic Information: The following questions pertain to important background information to help ensure that a range of professionals have responded. Please answer as honestly as you can. Results are completely anonymous. Thank you for responding. 3. Your ethnic/racial identification (Please circle only one.): 1 White 2 Aboriginal 3 Asian 4 East Indian 5 Black 6 Hispanic 7 Other - Please specify 4. Highest level of education your parents or significant adults you lived with completed while you were growing up: (If there are more than two important adults, please answer for the ones you lived with the longest.) Mother/Significant Adult Father/Significant Adult 1. Your age: 2. Your gender: D Female • Male 1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11 No formal education Some grade school Completed grade school Some high school Completed high school Some college Completed college Some university Completed undergraduate degree Some graduate work Completed graduate degree 5. Your father's/significant adult's occupation(s) while you were growing up: (If more than one occupation, list the two longest held starting with the more recent) 6. Your mother's/significant adult's occupation(s) while you were growing up: (If more than one occupation, list the two longest held starting with the more recent) Please continue on to the last page 147 7. Did your father/mother/significant adult receive public assistance (e.g. welfare, unemployment insurance, disability, etc.) while you were growing up? • Yes • No If Yes, approximately how frequently? Often Sometimes Rarely 8. Number of years of professional therapeutic practice after completing your training: 9. Your highest degree and the year it was completed: D Bachelor's D Master's D Doctorate 10. Program from which you received your highest degree: 1 Counselling Psychology 2 Clinical Psychology 3 Educational Psychology 4 Counsellor Education 5 Social Work 6 Other (Please Specify): 11. Where did you receive your highest degree? 1 within British Columbia 2 outside of British Columbia, within Canada 3 outside of Canada Please indicate which country: 12. Your usual work setting: (If more than one applies, please circle the two most frequent.) 1 University or college counselling centre 2 Nonprofit community agency 3 Government agency 4 Private profit agency 5 Private practice 13. Residential setting of the majority of your clients: 1 Rural 2 Suburban 3 Urban 4 Mixed THANK YOU AGAIN FOR PARTICIPATING!! 148 Appendix B Independent T-Test Results of Respondents' Years of Counselling Experience by Client Social Class Number Variable of Cases Mean SD SE of mean Years Middle Class 56 13.50 7.72 1.03 Exp. Lower 54 11.72 7.72 1.05 Class Mean Difference = 1.78 Levene's test for Equality of Variances: F=091 P = .764 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal 1.208 108 .230 1.78 1.47 (-1.14, 4.96) Unequal 1.208 107.855 .230 1.78 1.47 (-1.14, 4.96) Independent T-Test Results of Respondents' Years of Counselling Experience by Professional Association Number Variable of Cases Mean SD SE of mean Years CGCA 62 10.37 7.00 .89 Exp. BCPA 48 15.54 7.72 1.11 Mean Difference = 5.17 Levene's test for Equality of Variances: F = .799 P = .337 T-Test for Equality of Means Variances t-value df 2-Tail Sign Mean Diff SE of Diff 95% CI for Diff (-7.96, -2.38) Equal -3.672 108 .000 -5.17 1.41 Unequal -3.626 95.895 .000 -5.17 1.43 (-8.00, -2.34) 149 Appendix C independent /'-Tests on Respondents' Evaluations of the Problem Total Sample Population Presenting problem Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 2 .2182 2 . 4286 . 809 . 828 . 109 .111 Mean Difference = -.2104 Levene's Test for Equality of Variances: F= .030 P= .863 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of D i f f 95% CI for D i f f Equal Unequal -1.35 -1.35 109 109.00 . 179 . 179 155 155 (-.519, .098) (-.519, .098) Client self-concept Variable Number of Cases Mean SD SE of Mean middle class lower class 55 54 5.8545 5.8704 .931 . 802 . 126 . 109 Mean Difference = -.0158 Levene's Test for Equality of Variances: F= .201 P= .655 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of D i f f 95% CI for D i f f Equal Unequal .09 .10 107 105.22 . 925 .924 . 167 .166 (-.346, .315) (-.34.6, .314) Appendix C Independent T-Tests on Respondents' Evaluations of the Problem Total Sample Population Motivation for change Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 4.3455 4 .2143 1.220 1.317 . 165 . 176 Mean Difference = .1312 Levene's Test for Equality of Variances: F= .249 P= .619 t-t e s t for Equality of Means Variances t-value df 2-Tail Sig SE of Diff 95% CI for D i f f Equal Unequal . 54 . 54 109 108.63 . 588 . 587 .241 .241 (-.347, .609) (-.347, .609) Personal interest in treating the client Variable Number of Cases Mean SD SE of Mean middle class lower class 54 56 2.9074 3.0179 1. 444 1 . 601 . 197 .214 Mean Difference = -.1104 Levene's Test for E q u a l i t y of Variances: F= .309 P= .579 t- t e s t for Equality of Means Variances t-value df 2 - T a i l Sig SE of D i f f 95% CI for D i f f Equal Unequal -.38 108 .38 107.53 .705 . 705 .291 .291 (-.688, .467) (-.686, .466) 151 Appendix C Independent /'-Tests on Respondents'.Evaluations of the Problem Total Sample Population Prognosis Variable Number of Cases Mean SD SE of Mean middle class 53 3.4151 1.307 .180 lower class 56 3.6964 1.127 .151 Mean Difference = -.2813 Levene's Test for Equality of Variances: F= 1.641 P= .203 t-test for Equality of Means 95% Variances t-value df 2-Tail S i g SE of D i f f CI for D i f f Equal -1.21 107 .231 .233 (-.744, .181) Unequal -1.20 102.79 .233 .234 (-.746, .184) Likelihood of referring for psychiatric assessment Number Variable of Cases Mean SD SE of Mean middle class 54 3.4630 1.668 .227 lower class 56 3.8214 1.927 .257 Mean Difference = -.3585 Levene's Test for E q u a l i t y of Variances: F= 1.304 P= .256 t-t e s t for Equality of Means 95% Variances t-value df 2 - T a i l S i g SE of D i f f CI for D i f f Equal -1.04 108 .300 .344 (-1.041, .324) Unequal -1.04 106.78 .299 .343 (-1.039, .322) 152 Appendix C Independent T-Tests on Respondents' Evaluations of the Problem (BCPA only) Presenting problem Number Variable of Cases Mean SD SE of Mean middle class 23 2.4783 .846 .176 lower class 26 2.5385 .948 .186 Mean Difference = -.0602 Levene's Test for Equality of Variances: F= .047 P= .830 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -.23 47 .817 .258 (-.579, .459) Unequal -.23 46.99 .815 .256 (-.576, .455) Client self-concept Number Variable of Cases Mean SD SE of Mean middle class 23 5.8696 .757 .158 lower class 24 6.0417 .690 .141 Mean Difference = -.1721 Levene's Test for Equality of Variances: F= .867 P= .3 57 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -.82 45 .419 .211 (-.598, .253) Unequal -.81 44.19 .420 .212 (-.599, .254) 153 Appendix C independent 7-Tests on Respondents' Evaluations of the Problem (BCPA only) Motivation for change Number Variable of Cases Mean SD SE of Mean middle class 23 4.2174 1.166 .243 lower class 26 4.2308 1.177 .231 Mean Difference = -.0134 Levene's Test for Equality of Variances: F= .008 P= .928 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of Diff CI for D i f f Equal -.04 47 .968 .335 (-.688, .662) Unequal -.04 46.37 .968 .335 (-.688, .662) Personal interest in treating the client Number Variable of Cases Mean SD SE of Mean middle class 22 3.2727 1.518 .324 lower class 26 3.8077 1.833 .360 Mean Difference = -.5350 Levene's Test for Equality of Variances: F= 2.209 P= .144 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of Diff CI for D i f f Equal -1.09 46 .282 .491 (-1.524, .455) Unequal -1.11 45.98 .274 .484 (-1.509, .439) 154 Appendix C Independent /-Tests on Respondents' Evaluations of the Problem (BCPA only) Prognosis Number Variable of Cases Mean SD SE of Mean middle class 23 3.3478 i.265 .264 lower class 26 3.8846 1.177 .22 1 Hean Difference = -.5368 Levene's Test for Equality of Variances: F= .204 P= .653 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for Diff Equal -1.54 47 .131 .349 (-1.239, .166) Unequal -1.53 45.25 .133 .351 (-1.243, .169) Likelihood of referring for psychiatric assessment Number Variable of Cases Mean SD SE of Mean middle class 23 3.6087 1.559 .325 lower class 26 3.8846 1.904 .373 Mean Difference = -.2759 Levene's Test for Equality of Variances: F= .411 P= .524 t-test for Equality of Means 95% Variances t-value df 2- T a i l Sig SE of D i f f CI for D i f f Equal -.55 47 .585 .501 (-1.285, .733) Unequal -.56 46.74 .580 .495 (-1.272, .720) 155 Appendix C Independent T-Tests on Respondents' Evaluations of the Problem (CGCA only) Presenting problem Variable Number of Cases Mean SD SE of Mean middle class lower class 32 30 2.0313 2 .3333 740 711 131 130 Mean Difference = -.3021 Levene's Test for Equality of Variances: F= .083 P= .775 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of Diff 95% CI for D i f f Equal Unequal -1. 64 -1. 64 60 59.96 . 107 . 106 .185 .184 (-.671, .067) (-.671, .067) Client self-concept Variable Number of Cases Mean SD SE of Mean middle class lower class 32 30 5.8438 5.7333 1. 051 . 868 . 186 .159 Mean Difference = .1104 Levene's Test for Equality of Variances: F= .004 P= .952 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of D i f f 95% CI for D i f f Equal .45 60 .655 .246 (-.381, .602) Unequal .45 59.10 .653 .244 (-.378, .599) Appendix C •Independent /'-Tests on Respondents' Evaluations of the Problem (CGCA only Motivation for change Number Variable of Cases Mean SD SE of Mean middle class 32 4.4375 1.268 .224 lower class 30 4.2000 1.448 264 Mean Difference = .2375 Levene's Test for Equality of Variances: F= .506 t-test for Equality of Means . 480 95% Variances t-value df 2-T il Sig SE of Diff CI for Diff Equal .69 60 .494 .345 (-.453. .928) Unequal .69 57.77 .496 .347 (-.457, .932) 156 Personal interest in treating the client Number Variable of Cases Mean SD SE of Mean middle class 32 2.6563 1.359 .240 lower class 30 2.3333 .959 .175 Mean Difference = .3229 Levene's Test for Equality of Variances: F= 2.469 P= .121 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal 1.07 60 .287 .300 (-.278, .924) Unequal 1.09 55.84 .282 .297 (-.273, .918) 157 Appendix C Independent T-Tests on Respondents' Evaluations of the Problem (CGCA only) Prognosis Number Variable of Cases Mean SD SE of Mean middle class 30 3.4667 1.358 .248 lower class 30 3.5333 1.074 .196 Mean Difference = -.0667 Levene's Test for Equality of Variances: F= 1.695 P= .198 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -.21 58 .834 .316 (-.700, .566) Unequal -.21 55.08 .834 .316 (-.700, .567) Likelihood of referring for psychiatric assessment Number Variable of Cases Mean SD SE of Mean middle class 31 3.3548 1.762 .316 lower class 30 3.7667 1.977 .361 Mean Difference = -.4118 Levene's Test for Equality of Variances: F= 1.146 P= .289 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -.86 59 .393 .479 (-1.371, .547) Unequal -.86 57.73 .394 .480 (-1.373, .549) 158 Appendix C Independent T-Tests on Male Respondents' Evaluations of the Problem Presenting Problem Number Variable of Cases Mean SD SE of mean middle class 25 2.28 .61 .12 lower class 28 2.46 .74 .14 Mean Difference = .18 Levene's test for Equality of Variances: F = 2.266 P = .138 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal -.976 51 .334 -.18 .19 (-.56,. 19) Unequal -.987 50.700 .328 -.18 .19 (-.56,.19) Motivation for Change Number Variable of Cases Mean SE) SE of mean middle class 25 4.32 1.07 .21 lower class 28 4.00 1.19 .22 Mean Difference = .32 Levene's test for Equality of Variances: F= 040 P=842 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal 1.027 51 .309 .32 .31 (-.31,.95) Unequal 1.033 50.993 .307 .32 .31 (-.30, .94) 159 Appendix C Independent T-Tests on Male Respondents' Evaluations of the Problem Client Self-Concept Number Variable of Cases Mean SD SE of mean middle class 25 5.68 1.03 .21 lower class 27 5.78 .93 .18 Mean Difference = .10 Levene's test for Equality of Variances: F = .063 P = .803 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal -.359 50 .721 -9.78E-02 .27 (-.64, .45) Unequal -358 48.506 .722 -9.78E-02 .27 (-.65, .45) Prognosis Number Variable of Cases Mean SD SE of mean middle class 24 3.58 1.32 .27 lower class 28 3.36 1.03 .19 Mean Difference = .22 Levene's test for Equality of Variances: F = 2.218 P = . 143 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal .696 50 .490 .23 .33 (-.43, .88) Unequal .683 43.219 .498 .23 .33 (-.44, .89) 160 Appendix C Independent T-Tests on Male Respondents' Evaluations of the Problem Likelihood of Referring for Psychiatric Assessment Number Variable of Cases Mean SD SE of mean middle class 25 3.40 1.41 .28 lower class 28 4.00 2.05 .39 Mean Difference = Levene's test for Equality of Variances: F = 6.294 P = .015 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal -1.233 51 .227 -.60 .49 (-1.58, .38) Unequal -1.249 48.037 .218 -.60 .49 (-1.57, .37) Personal Interest in Treating the Client Number Variable of Cases Mean SD SE of mean middle class 25 2.92 1.53 .31 lower class 28 2.64 1.39 .26 Mean Difference = .28 Levene's test for Equality of Variances: F= 484 P = .490 T-Test for Equality of Means Variances t-value df 2-Tail Sign Mean Diff SE of Diff 95% CI for Diff (-.53, 1.08) Equal .691 51 .493 .28 .40 Unequal .688 48.939 .495 .28 .40 (-.53, 1.09) 161 Appendix C Independent T-Tests on Male Respondents' Evaluations of the Problem Presenting Problem Number Variable of Cases Mean SD SE of mean middle class 25 2.28 .61 .12 lower class 28 2.46 .74 .14 Mean Difference = 1 8 Levene's test for Equality of Variances: F = 2.266 P = .138 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal -.976 51 .334 -.18 .19 (-.56, .19) Unequal -.987 50.700 .328 -.18 .19 (-.56, .19) Motivation for Change Number Variable of Cases Mean SD SE of mean middle class 25 4.32 1.07 .21 lower class 28 4.00 1.19 .22 Mean Difference = .32 Levene's test for Equality of Variances: F = .040 P = .842 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal 1.027 51 .309 .32 .31 (-.31,.95) Unequal 1.033 50.993 .307 .32 .31 (-.30, .94) 162 Appendix C Independent T-Tests on Female Respondents' Evaluations of the Problem Client Self-Concept Number Variable of Cases Mean SD SE of mean middle class 31 5.97 .84 .15 lower class 26 6.00 .63 .12 Mean Difference = .03 Levene's test for Equality of Variances: F = .572 P = .500 T-Test for Equality of Means 2-Tail Sign 95% Variances t-value df Mean Diff SE of Diff CI for Diff Equal -.162 55 .872 -3.23 .20 (-.43, .37) Unequal -.166 54.474 .869 -3.23 .19 (-.42, .36) Prognosis Number Variable of Cases Mean SD SE of mean middle class 30 3.30 1.29 .24 lower class 27 4.04 1.16 .22 Mean Difference = .74 Levene's test for Equality of Variances: F= 1.208 P=277 T-Test for Equality of Means 2-Tail Sign 95% Variances t-value df Mean Diff SE of Diff CI for Diff .33 (-1 39,) Equal -2.258 55 .028 -.74 Unequal -2.271 55.000 .027 -.74 .32 (-L39,) 163 Appendix C Independent T-Tests on Female Respondents' Evaluations of the Problem Likelihood of Referring for Psychiatric Assessment Variable Number of Cases Mean SD SE of mean middle class lower class 30 27 3.57 3.59 1.87 1.82 .34 .35 Mean Difference = .02 Levene's test for Equality of Variances: F = .461 P = .500 T-Test for Equality of Means Variances t-value 2-Tail Sign 95% df Mean Diff SE of Diff CI for Diff Equal Unequal .053 .053 55 54.626 .958 .958 -2.59E-02 -2.59E-02 .49 .49 (-1.01, .96) (-1.01, .96) Personal Interest in Treating the Client Variable Number of Cases Mean SD SE of mean middle class lower class 30 27 2.93 3.37 1.39 1.76 .25 .34 Mean Difference = .44 Levene's test for Equality of Variances: F = 2.695 P = . 106 T-Test for Equality of Means Variances t-value df 2-Tail Sign 95% Mean Diff SE of Diff CI for Diff Equal Unequal -1.047 -1.034 55 49.422 .300 .306 -.44 -.44 .42 .42 (-1.27, .40) (-1.29, .41) 164 Appendix D Coding Instructions for Part I, Case History Question # 2 1. Assign a number to each of the questionnaires you have received. Do this in numeric order starting with # 1 (e.g. # 1, 2, 3, and so on). 2. Read the response to Part I, Case History Question #2 on the first questionnaire. 3. In the left-hand row of the code sheet provided, list questionnaire by the assigned number. 4. Use one line per questionnaire. 5. For each response make checkmarks in the boxes under the prearranged categories. 6. As you go along, if certain responses do not appear appropriate to any of the prearranged categories, note any way that the categories might be rearranged, added to or refined to be more accurately descriptive of the issues noted by the respondents. Follow this procedure until you have coded each of the questionnaires. For example: if questionnaire # 1 listed 'refer for psychiatric assessment', 'money problems' and 'assess for suicidal ideation', as responses to the question, you would list the questionnaire number in the left hand row and check off the categories as follows: suicide children family mentalhe genhealt finance 1 / 165 Appendix D ANGER anger or anger management by CL class CL middle lower Count ANGER . 00 no mention 1.00 important issue 41 15 17 Row Total 79 71 . 2 32 28 . 8 Column 56 Total 50.5 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association 55 111 49.5 100.0 Value .22993 . 07288 .23002 .22786 Minimum Expected Frequency - 15.856 Number of Missing Observations: 0 ANXIETY anxiety by CL class CL Count |middle lower ANXIETY . 0 0 no mention 1.00 important issue Column Total Chi-Square 40 16 48 56 50.5 55 49.5 Row Total 88 79 .3 23 20.7 111 100.0 Value DF DF Signi ficance . 63157 .78719 .63151 .63311 Signi f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association 4.24035 3.33069 4.33689 4.20215 . 03947 .06800 . 03730 . 04037 Minimum Expected Frequency - 11.396 Number of Missing Observations: 0 166 Appendix D ALCOHOL by CL class CL Count ALCOHOL no mention middle lower 1 2 . 00 n 33 33 1. 00 issue 23 22 Column 56 55 Total 50.5 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for linear association 49.5 Value Row Total 66 59 . 5 45 40.5 111 100 . 0 . 01321 .00000 .01321 .01310 Minimum Expected Frequency - 22.2 97 Number of Missing Observations: 0 DF Significance . 90848 1. 00000 . 90848 ' .90889 FAMILY family or re l a t i o n s h i p problems by CL class Count FAMILY . 00 no mention 1.00 important issue CL middle lower 1 2 44 42 12 13 Column 56 Total 50.5 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for line a r association Row Total 86 77.5 25 22.5 55 111 49.5 100.0 Value .07751 .00262 .07752 . 07681 Minimum Expected Frequency - 12.387 Number of Missing Observations: 0 DF Signi f i c a n c e .78070 . 95918 .78069 .78167 167 Appendix D CHILDHOO childhood issues by CL class CL Count I P U T T DHOO middle 1 lower 2 Row Total . 00 no mention 48 46 94 84 . 7 1. 00 important issue 8 9 17 15.3 Column 56 55 111 Total 50.5 49.5 100.0 Chi-Square Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Minimum Expected Frequency - 8.423 Number of Missing Observations: 0 DEPRESSI depression by CL class CL Count I DEPPE^^I middle 1 lower 2 Row Total .00 no mention 42 44 86 77.5 1.00 important issue 14 11 25 22 . 5 Column 56 55 111 Total 50.5 49.5 100.0 -Square Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Minimum Expected Frequency - 12.387 .09238 . 00163 . 09240 , 09154 .76118 . 96780 .76114 .76222 •39753 1 .52837 .16263 1 .68674 .39838 1 .52793 .39395 1 .53023 Number of Missing Observations: 0 168 Appendix D FAMVIOL family violence by CL class CL middle lower Count FAMVIOL . 00 no mention 1. 00 important issue Chi-Square 25 31 Column 5 6 Total 50.5 39 16 Row Total 64 57 . 7 47 42 . 3 55 111 49.5 100.0 Value DF Signi ficance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association 7.84136 6.80238 7.95040 7 . 77072 Minimum Expected Frequency - 23.288 Number of Missing Observations: 0 . 00511 .00910 .00481 .00531 FINANCE finances/money spending by CL class CL middle lower Count FINANCE .00 no mention 1.00 important Chi-Square 50 49 Row Total 99 89 .2 12 10.8 Column 56 55 111 Total 50.5 49.5 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association . 00109 . 00000 .00109 .00108 .97364 1. 00000 . 97364 . 97376 Minimum Expected Frequency 5.946 Number of Missing Observations: 0 Appendix D GENHEALT general health issues by CL class CL middle lower Count GENHEALT . 00 no mention 1.00 important issue 49 49 Row Total 98 88 . 3 13 11.7 Column 56 Total 50.5 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association 55 111 49.5 100.0 Value .06792 .00000 .06799 .06731 Minimum Expected Frequency - 6.441 Number of Missing Observations: 0 MENTALH mental health issues by CL class CL middle lower Count MENTALH . 00 no mention 1.00 important issue Column Total Chi-Square 49 45 10 56 50.5 55 49.5 Row Total 94 84.7 17 15.3 111 100.0 Value DF DF 169 Significance . 79439 1. 00000 . 79429 . 79530 Sig n i f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .69067 .32206 .69345 . 68445 .40594 .57037 .40499 .40806 Minimum Expected Frequency 8.423 Number of Missing Observations: 0 170 Appendix D MOOD mood/coping w i t h f e e l i n g s by CL c l a s s CL Count MOOD middle 1 lower 2 Row Total . 00 no mention 54 47 101 91. 0 1. 00 important issue 2 8 10 9 . 0 Column Total 56 50.5 55 49 . 5 111 100 . 0 Chi-Square Value DF S i g n i f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency -Cells with Expected Frequency < 5 -Number of Missing'Observations: 0 4 . 07647 2 . 84766 4 . 33142 4.03975 4 . 955 1 OF 4 ( 25.0%) . 04348 . 09151 . 03741 .04444 .04386 . 05263 171 Appendix D OTHER by CL c l a s s OTHER Count CL middle Column 56 Total 50.5 Chi-Square Pearson Likelihood Ratio Mantel-Haenszel test for linear association lower 1 2 . 00 48 45 1. 00 5 6 2 . 00 2 3 3 . 00 1 4 . 00 1 55 49.5 Row Total 93 83 . 8 11 9 . 9 5 4.5 1 . 9 1 .9 111 100.0 Value 2.37887 3.15276 . 10273 Minimum Expected Frequency - .495 Cells with Expected Frequency < 5 -DF Signi f i c a n c e . 66645 . 53259 .74858 6 OF 10 ( 60.0%) Number of Missing Observations: 0 Appendix D PARAN10 fearfulness/paranoia by CL class CL middle lower Count PARANIO . 00 no mention 1. 00 important issue 36 20 Column 56 Total 50.5 Chi-Square 46 Row Total 82 73 . 9 29 26.1 55 111 49.5 100.0 Value DF 172 Signi ficance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association 5.38335 4.42743 5 . 49226 5.33486 .02033 . 03537 . 01910 . 02 09 0 Minimum Expected Frequency 14.369 Number of Missing Observations: 0 173 Appendix D SELFDIR s e l f - d i r e c t e d by CL class CL Count n P T FDTT! middle 1 lower 2 Row Total O 1—1 J_J 1. LJ J. L\ . 00 no mention 50 51 101 91. 0 1.00 important issue 6 4 10 9 . 0 Column Total 56 50 . 5 55 49 . 5 111 100.0 Chi-Square Value. DF Signi f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Fisher's Exact Test: One-Tail Two-Tail .40092 . 09100 .40360 .39731 .52661 .76291 .52523 .52848 .38260 . 74218 Minimum Expected Frequency - 4.955 Ce l l s with Expected Frequency < 5 - 1 OF 4 ( 25.0%) Number of Missing Observations: 0 174 Appendix D SELFEST s e l f esteem/worth/concept by CL c l a s s CL m i d d l e lower Count SELFEST . 00 no mention 1. 00 i m p o r t a n t i s s u e 41 15 Column 5 6 T o t a l 50.5 Chi-Square 43 12 Row Total 84 75 . 7 27 24 . 3 55 111 49.5 100.0 Value DF Signi ficance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association .37197 .15106 .37264 .36862 Minimum Expected Frequency - 13.37E Number of Missing Observations: 0 .54193 .69753 .54157 .54376 SUICIDE s u i c i d e / s e l f damage by CL class Count ^UICIDE middle 1 lower 2 Row Total W W -L- %w» J - !_/ i—J .00 no mention 49 51 100 90 .1 1.00 •important issue 7 4 11 9.9 Column Total 56 50 . 5 55 49.5 111 100.0 Chi-Square Value DF Significance Pearson .84924 1 .35677 Continuity Correction .36466 1 .54593 Likelihood Ratio .85963 1 .35384 Mantel-Haenszel test for .84159 1 .35894 li n e a r association Minimum Expected Frequency - 5.450 Number of Missing Observations: 0 i7H-a Appendix D UNEMPLOY unemployment by CL c l a s s CL m i d d l e lower' Count UNEMPLOY . 00 no mention 1. 00 important issue Chi-Square 45 11 Column 5 6 T o t a l 50.5 41 14 55 49 . 5 Row T o t a l 86 77 . 5 25 22 . 5 111 100 . 0 Value DF S i g n i f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association .53708 .25566 . 53797 . 53224 Minimum Expected Frequency - .12.387 Number of Missing Observations: 0 VOCATION vocational assistance by CL class CL middle lower Count VOCATION .00 no mention 1.00 important issue 54 Column 56 Total 5 0.5 Chi-Square 47 55 49 . 5 Row Total 101 91. 0 10 9 . 0 111 100. 0 Value DF .46364 . 61312 .46327 .46567 Significance 4.07647 2 . 84766 4.33142 4.03975 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel t e s t ' f o r l i n e a r a s s o c i a t i o n Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.955 Ce l l s with Expected Frequency < 5 -. 04348 . 09151 - 03741 . 04444 . 04386 .05263 1 OF 4 ( 25.0%) Number of Missing Observations: 0 175 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Alcohol Count CL middle lower ALCOHOL no mention 00 1.00 important issue 15 .15 11 Column 23 Total 46.9 26 53.1 Row Total 30 61.2 19 38 . 8 49 100.0 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Value .29110 . 06041 .29188 .28516 Minimum Expected Frequency - 8.918 Number of Missing Observations: 0 Anger CL middle lower Count ANGER .00 no mention 1.00 important issue Column Total Chi-Square 16 18 23 46.9 26 53.1 Row Total 34 69.4 15 30.6 49 100.0 Value DF DF Significance .58952 .80584 .58902 . 59334 Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .00064 .00000 .00064 .00063 .97978 1.00000 .97977 .97998 Minimum Expected Frequency 7.041 Number of Missing Observations: 0 176 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Anxiety Count CL middle lower ANXIETY . 00 no mention 1. 00 important issue Column Total Chi-Square 14 23 23 46.9 26 53 .1 Row Total 37 75.5 12 24 . 5 49 100 . 0 Value Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association 5 . 02435 3 . 64305 5 .16701 4 . 92181 Minimum Expected Frequency - 5.633 Number of Missing Observations: 0 Depression Count DEPRESSI .00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 lower 17 19 23 46.9 26 53.1 Row Total 36 73 . 5 13 26.5 49 100.0 Value DF DF Significance . 02499 .05630 .02302 . 02652 Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .00438 .00000 . 00438 . 00429 . 94725 1.00000 . 94724 .94779 Minimum Expected Frequency - 6.102 Number of Missing Observations: 0 177 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Childhood issues Count CHILDHOO . 00 no mention 1. 00 important issue Column Total Chi-Square CL middle lower 21 23 46.9 24 26 53 .1 Row Total 45 91. 8 4 8.2 49 100 . 0 Value DF Significance Pearson .01639 Continuity Correction .00000 Likelihood Ratio .01636 Mantel-Haenszel test for .01605 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 1.878 C e l l s with Expected Frequency < 5 - 2 OF .89814 1.00000 .89822 .89918 .64720 1.00000 4 ( 50.0%) Number of Missing Observations: 0 178 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Family or relationship problems Count FAMILY , 00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 23 46.9 lower 21 26 53.1 Row Total 39 79.6 10 20.4 49 100.0 Value DF Significance Pearson .04727 Continuity Correction .00000 Likelihood Ratio .04721 Mantel-Haenszel test for .04631 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.694 Ce l l s with Expected Frequency < 5 - 1 OF .82788 1.00000 .82799 .82962 .55267 1.00000 4 ( 25.0%) Number of Missing Observations: 0 179 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Family violence Count FAMVIOL . 00 no mention 1.00 important issue Chi-Square CL middle lower 13 10 Column .23 Total 46.9 19 26 53.1 Row Total 32 65.3 17 34.7 49 100 . 0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 1.47627 .83600 1.48009 1.44614 .22436 .36054 .22376 .22915 Minimum Expected Frequency - 7.980 Number of Missing Observations: 0 Mental health issues CL middle lower Count MENTALH .00 no mention 1.00 important issue Column Total Chi-Square 23 46.9 20 26 53.1 Row Total 38 77 . 6 11 22 . 4 49 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .01255 .00000 . 01256 . 01229 .91082 1.00000 . 91077 .91173 Minimum Expected Frequency 5.163 Number of Missing Observations: 0 180 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Finances/money spending Count FINANCE . 00 no mention 1. 00 important Chi-Square CL middle 1 20 Column 23 Total 4 6.9 lower 25 26 53.1 Row Total 45 91.8 4 8.2 49 100.0 Value DF Significance 1.37704 .42347 1.41941 1.34894 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 1.878 Ce l l s with Expected Frequency < 5 -.24061 . 51521 .23350 .24546 .25912 . 32968 2 OF 4 ( 50.0%) Number of Missing Observations: 0 181 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class General health issues Count GENHEALT . 00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 19 23 46.9 lower 23 26 53 .1 Row Total 42 85 . 7 7 14.3 49 100.0 Value DF Significance Pearson .34142 Continuity Correction .03073 Likelihood Ratio .34109 Mantel-Haenszel test for .33445 lin e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.286 Cells with Expected Frequency < 5 - 2 OF . 55901 . 86085 . 55920 .56305 . 42874 . 69178 4 ( 50.0%) Number of Missing Observations: 0 182 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Mood Count MOOD . 00 no mention 1. 00 important issue Column Total Chi-Square CL middle lower 22 23 46.9 26 26 53 .1 Row Total 48 98.0 1 2.0 49 100.0 Value DF Significance Pearson 1.15399 Continuity Correction .00384 Likelihood Ratio 1.53623 Mantel-Haenszel test for 1.13043 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - .469 Ce l l s with Expected Frequency < 5 - 2 OF Number of Missing Observations: 0 4 ( 50.0%) .28272 .95058 .21518 .28768 .46939 .46939 183 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Other CL OTHER Chi-Square Count middle 1 lower 2 Row Total . 00 20 22 42 85.7 1. 00 2 2 4 8.2 2 . 00 1 1 2 4.1 3 . 00 1 1 2 . 0 Column 23 Total 46.9 26 49 53.1 100.0 Value DF Significance Pearson Likelihood Ratio Mantel-Haenszel test for l i n e a r association .91499 1.29778 .28721 Minimum Expected Frequency - .469 Ce l l s with Expected Frequency < 5 - 6 OF 8 ( 75.0%) .82181 .72966 .59201 Number of Missing Observations: 0 184 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Paranoia Count PARAN10 no mention 00 1. 00 important issue Chi-Square CL middle lower Column 23 Total 4 6.9 22 Row Total 40 81. 6 9 18 . 4 26 49 53.1 100.0 V a l u e DF Signi ficance Pearson .32867 Continuity Correction .04148 Likelihood Ratio .32822 Mantel-Haenszel test for .32196 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.224 Cell.s with Expected Frequency < 5 - 2 OF Number of Missing Observations: 0 4 ( 50.0%) .56644 .83861 .56671 .57043 .41800 .71648 185 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Self-directed Count SELFDIR , 00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 22 23 46.9 lower 25 26 53.1 Row Total 47 95.9 2 4 . 1 49 100.0 Value DF Significance Pearson .00785 Continuity Correction .00000 Likelihood Ratio .00783 Mantel-Haenszel test for .00769 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - .939 C e l l s with Expected Frequency < 5 - 2 OF .92942 1.00000 . 92948 . 93014 .72364 1.00000 4 ( 50.0%) Number of Missing Observations: 0 186 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Self-esteem CL middle lower Count SELFEST . 00 no mention 1. 00 important issue Chi-Square 18 Column 23 Total 46.9 22 26 53 .1 Row Total 40 81.6 9 18 . 4 49 100. 0 Value DF Significance Pearson .32867 Continuity Correction .04148 Likelihood Ratio .32822 Mantel-Haenszel test for .32196 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.224 Ce l l s with Expected Frequency < 5 - 2 OF .56644 . 83861 . 56671 . 57043 .41800 .71648 4 ( 50.0%) Number of Missing Observations: 0 187 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Suicide Count SUICIDE .00 no mention 1.00 important issue Chi-Square CL middle lower 20 Column 23 Total 46.9 23 26 53 .1 Row Total 43 87.8 6 12 .2 49 100.0 Value DF Significance Pearson .02573 Continuity Correction .00000 Likelihood Ratio .02568 Mantel-Haenszel test for .02520 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.816 Ce l l s with Expected Frequency < 5 - 2 OF .87257 1.00000 .87267 .87387 .60486 1.00000 4 ( 50.0%) Number of Missing Observations: 0 188 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Vocation Count VOCATION . 00 no mention 1.00 important issue Chi-Square CL middle lower 1 2 22 24 1 2 Column 23 Total 46.9 Row Total 46 93.9 3 6.1 26 49 53.1 100.0 Value DF Significance Pearson .237 51 Continuity Correction .00000 Likelihood Ratio .24299 Mantel-Haenszel test for .23266 li n e a r association Fisher's Exact Test: One-Tail , Two-Tail Minimum Expected Frequency - 1.408 Ce l l s with Expected Frequency < 5 - 2 OF . 62601 1.00000 .62205 .62956 .54684 1.00000 4 ( 50.0%) Number of Missing Observations: 0 189 Appendix D Chi Square Analyses of BCPA Respondents' Choices of Important Issues by Client Class Unemployment Count UNEMPLOY no mention . 00 1.00 important issue Column Total Chi-Square CL middle 1 20 23 46.9 lower 20 26 53.1 Row Total 40 81.6 9 18.4 49 100.0 Value DF Significance Pearson .81940 Continuity Correction .28685 Likelihood Ratio .83561 Mantel-Haenszel test for .80268 lin e a r association Fisher\s Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.224 Ce l l s with Expected Frequency < 5 - 2 OF .36536 .59225 .36066 .37029 .29848 .47162 4 ( 50.0%) Number of Missing Observations: 0 190 Appendix D Chi Square Analyses of C G C A Respondents' Choices of Important Issues by Client Class Alcohol Count CL middle 1 ALCOHOL . 00 no mention 1. 00 important issue 15 Column 3 3 Total 53.2 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for line a r association lower 11 Row Total 36 58 .1 26 41 . 9 29 62 46.8 100.0 Value .35881 .11635 .35959 .35303 Minimum Expected Frequency - 12.161 Number of Missing Observations: 0 DF S i g n i f i c a n c e . 54917 .73303 . 54873 .55240 Anger/anger management Count ANGER no mention .00 1.00 important issue Column Total Chi-Square CL middle 1 25 lower 20 Row Total 45 72 . 6 17 27.4 33 29 62 53.2 46.8 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r a s s o c i a t i o n .35780 . 09790 .35732 .35203 . 54973 .75437 . 55000 .55296 Minimum Expected Frequency 7.952 Number of Missing Observations: 0 191 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Anxiety Count ANXIETY . 00 no mention 1. 00 important issue Column Total Chi-Square CL middle lower 26 Row 2 j Total 25 51 82.3 11 17.7 33 29 62 53.2 46.8 100.0 Value DF Si g n i f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .58215 .18477 .59000 .57276 .44547 .66730 .44242 .44916 Minimum Expected Frequency - 5.145 Number of Missing Observations: 0 Childhood issues Count CHILDHOO .00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 27 33 53 .2 lower 22 29 46.8 Row Total 49 79.0 13 21.0 62 100.0 Value DF S i g n i f i c a n c e Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association .33044 .06875 .32985 .32511 . 56540 .79316 .56575 .56855 Minimum Expected Frequency 6.081 Number of Missing Observations: 0 192 Appendix D Chi Square Analyses of.CGCA Respondents' Choices of Important Issues by Client Class Depression Count PEPFESSI CL middle 1 lower 2 Row Total . 00 no mention 25 25 50 80 . 6 1 . 00 important issue 8 4 12 19 .4 Column Total 33 53 .2 29 46.8 62 100.0 Chi-Square Value DF Signi ficance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Minimum Expected Frequency -Number of Missing Observations: 1. 07976 .51408 1.10095 1.06235 5 . 613 0 .29875 .47338 .29406 .30268 Family or relationship problems Count FAMILY CL middle 1 lower 2 Row Total . 00 no mention 26 21 47 75.8 1.00 important issue 7 8 15 24 .2 Column Total 33 53 .2 29 46.8 62 100.0 Chi-Square Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association .34194 .08271 .34140 .33643 .55871 .77366 . 55903 . 56190 Minimum Expected Frequency - 7.016 Number of Missing Observations: 0 193 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Finances/money spending Count FINANCE .00 no mention 1.00 important Chi-Square CL middle 1 30 Column 33 Total 53.2 lower 24 29 46.8 Row Total 54 87 .1 12.9 62 100.0 Value DF Significance Pearson .91240 Continuity Correction .33128 Likelihood Ratio .91514 Mantel-Haenszel test for .89768 linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.742 Cel l s with Expected Frequency < 5 - 2 OF .33948 .56491 .33875 .34340 .28232 .45606 4 ( 50.0%) Number of Missing Observations: 0 194 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Family violence Count FAMVIOL .00 no mention 1. 00 important issue Column Total Chi-Square CL middle lower 12 21 33 53.2 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for linear a s s o c i a t i o n Row 2 ! Total 20 29 46.8 32 51. 6 30 48.4 62 100.0 Value 6.56928 5 . 32870 6.70010 6.46332 Minimum Expected Frequency - 14.032 Number of Missing Observations: 0 DF Significance .01038 . 02098 . 00964 . 01101 Fearfulness/paranoia Count PARANIO ,00 no mention 1.00 important issue Column Total Chi-Square CL middle 1 lower 2 Row Total 18 24 42 67 .7 15 5 20 32.3 33 29 62 53.2 46.8 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel t e s t for lin e a r a s s o c i a t i o n 5.62248 4.40551 5.83432 5.53180 .01773 .03582 .01572 .01867 Minimum Expected Frequency - 9.355 Number of Missing Observations: 0 195 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class General health issues Count GENHEALT . 00 no mention 1.00 important issue CL middle 1 30 Column 33 Total 53.2 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency lower 26 29 46.8 Row Total 56 90.3 6 9.7 62 100.0 Value . 02777 .00000 . 02771 .02732 2.806 Ce l l s with Expected Frequency < 5 - 2 OF DF Significance .86766 1.00000 .86778 .86872 .59977 1.00000 4 ( 50.0%) Number of Missing Observations: 0 196 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Mental health issues Count MENTALH . 00 no mention 1. 00 important issue Column Total Chi-Square CL middle 1 31 33 53 .2 lower 25 29 46.8 Row Total 56 90.3 6 9.7 62 100.0 Value DF Significance 1.05585 .35651 1. 06545 1.03882 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.806 Ce l l s with Expected Frequency < 5 -.30416 . 55045 .30198 .30810 .27547 .40545 2 OF 4 ( 50.0%) Number of Missing Observations: 0 197 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Mood Count MOOD . 00 no mention 1.00 important issue CL middle lower 32 Column 33 Total 53.2 Chi-Square 21 Row Total 53 85 . 5 9 14 . 5 29 62 46.8 100.0 Value DF Significance Pearson 7.50062 Continuity Correction 5.65225 Likelihood Ratio 8.23913 Mantel-Haenszel test for 7.37964 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.210 C e l l s with Expected Frequency < 5 - 2 OF Number of Missing Observations: 0 4 ( 50.0%) . 00617 .01743 .00410 .00660 .00748 .00938 198 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Other OTHER Chi-Square CL Count middle 1 lower 2 Row Total . 00 28 23 51 82.3 1. 00 3 4 7 11. 3 2 . 00 1 2 3 4.8 4.00 1 • 1 1.6 Column 33 Total 53.2 29 62 46.8 100.0 Value DF Significance Pearson Likelihood Ratio Mantel-Haenszel test for li n e a r association 1.71546 2 .10218 . 00030 Minimum Expected Frequency - .468 Ce l l s with Expected Frequency < 5 - 6 OF 8 ( 75.0%) . 63350 .55147 .98607 Number of Missing Observations: 0 199 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Self-directed Count SELFDIR , 00 no mention 1. 00 important issue Chi-Square CL middle lower 28 Column 33 Total 53.2 26 Row Total 54 87 . 1 12 . 9 29 62 46.8 100.0 Value DF Significance Pearson .31733 Continuity Correction .03374 Likelihood Ratio .32119 Mantel-Haenszel test for .31221 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.742 Cells with Expected Frequency < 5 - 2 OF .57322 . 85426 . 57089 .57633 .43024 .71255 4 ( 50.0%) Number of Missing Observations: 0 200 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Self-esteem Count SELFEST 00 no mention 1. 00 important issue Column Total Chi-Square C L middle 1 23 10 lower 21 Row Total 44 71 . 0 18 29 . 0 33 29 62 53.2 46.8 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association . 05530 .00000 .05538 .05441 Minimum Expected Frequency - 8.419 Number of Missing Observations: 0 Unemployment Count CL middle lower UNEMPLOY .00 no mention 1.00 important issue 25 Column 33 Total 53.2 Chi-Square 21 29 46.8 Row Total 46 74.2 16 25.8 62 100. 0 Value DF . 81409 1.00000 .81396 .81557 Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel t e s t for l i n e a r association .09014 .00009 . 09002 .08868 .76400 .99251 .76415 .76586 Minimum Expected Frequency 7.484 Number of Missing Observations: 0 201 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Suicide Count SUICIDE . 00 no mention 1 . 00 important issue Column Total Chi-Square CL middle 1 lower 29 33 53 .2 28 29 46. 8 Row Total 57 91.9 5 8 .1 62 100.0 Value DF Sig n i f i c a n c e Pearson 1.56600 Continuity Correction .61467 Likelihood Ratio 1.68675 Mantel-Haenszel test for 1.54074 linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.339 Ce l l s with Expected Frequency < 5 - 2 OF .21079 .43304 . 19403 .21451 .22006 .35954 4 ( 50.0%) Number of Missing Observations: 0 202 Appendix D Chi Square Analyses of CGCA Respondents' Choices of Important Issues by Client Class Vocation Count VOCATION . 00 no mention 1.00 important issue Column Total Chi-Square CL middle lower 32 33 53 .2 23 29 46.8 Row Total 55 88 .7 7 11.3 62 100.0 Value DF Significance 4.80610 3 .20462 5.18356 4.72858 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.274 Cel l s with Expected Frequency < 5 -.02836 .07343 . 02280 . 02967 .03505 .04373 2 OF 4 ( 50.0%) Number of Missing Observations: 0 203 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Alcohol CL ALCOHOL Count middle lower 1 2 . 00 15 19 1.00 10 9 Row Total 34 64.2 19 35.8 Column 25 Total 47.2 Chi-Square 28 53 52.8 100.0 Value Pearson .35454 Continuity Correction .09520 Likelihood Ratio .35443 Mantel-Haenszel test for .34785 linear association Minimum Expected Frequency - 8.962 Number of Missing Observations: 0 Anger CL ANGER Count middle lower 1 2 .00 19 23 1.00 6 5 Column Total Chi-Square 25 47.2 28 52.8 Row Total 42 79.2 11 20.8 53 100.0 Value DF DF Significance .55155 .75767 .55161 .55533 Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for linear association .30302 .04462 .30266 .29730 .58199 .83271 .58222 .58558 Minimum Expected Frequency - 5.189 Number of Missing Observations: 0 204 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Anxiety ANXIETY Column Tota l Chi-Square CL Count middle lower 1 2 .00 18 26 1.00 7 2 25 47 . 2 28 52.8 Row Total 44 83 . 0 9 17.0 53 100.0 Value DF S ign i f i cance 4.07557 2.73035 4 .23470 3 . 99867 Pearson Cont inui ty Correc t ion L ike l ihood Ratio Mantel-Haenszel test for l i n e a r as soc ia t ion F i s h e r ' s Exact Test: One-Tai l Two-Tai l Minimum Expected Frequency - 4.245 C e l l s with Expected Frequency < 5 -.04351 .09846 .03961 . 04554 .04830 .06739 2 OF 4 ( 50.0%) Number of Miss ing Observations: 0 205 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Childhood issues CHILDHOO Chi-Square CL Count middle lower 1 2 . 00 23 24 1.00 2 4 Column 25 Total 47.2 Row Total 47 88.7 6 11.3 28 53 52.8 100.0 Value DF Significance Pearson .51980 Continuity Correction .08223 Likelihood Ratio .53097 Mantel-Haenszel test for .50999 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.830 Cells with Expected Frequency < 5 - 2 OF Number of Missing Observations: 0 4 ( 50.0%) .47093 .77430 .46620 .47514 ,39099 ,67179 DEPRESSI depression by CL class 206 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Depression CL DEPRESSI Count middle lower 1 2 . 00 21 24 1.00 4 4 Column Total Chi-Square Row Total 45 84.9 15.1 25 28 53 47.2 52.8 100.0 Value DF Significance Pearson .03029 Continuity Correction .00000 Likelihood Ratio .03025 Mantel-Haenszel test for .02971 lin e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.774 Cel l s with Expected Frequency < 5 - 2 OF .86184 1. 00000 .86193 .86314 .58054 1.00000 4 ( 50.0%) Number of Missing Observations: 0 207 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Family or relationship issues CL RELFAM Count middle lower 1 2 .00 20 23 1.00 5 5 Column Total Chi-Square 25 47 .2 28 52 . 8 Row Total 43 81.1 10 18 . 9 53 100 . 0 Value DF Significance Pearson .03962 Continuity Correction .00000 Likelihood Ratio .03957 Mantel-Haenszel test for .03887 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.717 Ce l l s with Expected Frequency < 5 - 1 OF . 84223 1.00000 .84232 . 84371 . 55858 1.00000 4 ( 25.0%) Number of Missing Observations: 0 208 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Family violence CL FAMVIOL Count middle lower 1 2 . 00 11 19 1. 00 14 9 Chi-Square Column 25 Total 47.2 Row Total 30 56.6 23 43.4 28 53 52.8 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 3 . 06028 2.16611 3 . 08515 3 . 00254 . 08023 . 14108 .07901 .08313 Minimum Expected Frequency 10.849 Number of Missing Observations: 0 209 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Finances/money spending F I N A N C E Total Chi-Square CL Count middle lower 1 2 .00 21 25 1.00 4 3 Column 25 28 47.2 52.8 Value Row Tota l 46 86.8 7 13.2 53 100.0 DF S ign i f i cance Pearson .32190 Cont inui ty Correc t ion .02592 Like l ihood Ratio .32171 Mantel-Haenszel test for .31583 l i n e a r assoc ia t ion F i sher ' s Exact Test: One-Tai l Two-Tai l Minimum Expected Frequency - 3.302 C e l l s with Expected Frequency < 5 - 2 OF .57047 .87209 .57058 .57412 .43443 .69449 4 ( 50.0%) Number of Miss ing Observations: 0 6 210 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class General health issues GENHEALT Count CL middle 1 lower 2 Row Total .00 51 21 25 46 86 . 8 1.00 4 3 7 13.2 Chi-Square Column 25 Total 47.2 28 53 52.8 100.0 Value DF Significance Pearson .32190 Continuity Correction .02592 Likelihood Ratio .32171 Mantel-Haenszel test for .31583 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3.302 C e l l s with Expected Frequency < 5 - 2 OF . 57047 .87209 .57058 .57412 .43443 .69449 4 ( 50.0%) Number of Missing Observations: 0 211 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Mood MOOD Chi-Square CL Count middle lower 1 2 .00 24 25 1.00 1 3 Column 25 Total 47.2 Row Total 49 92 . 5 4 7 . 5 28 53 52.8 100.0 Value DF Significance Pearson .85333 Continuity Correction .16234 Likelihood Ratio .89700 Mantel-Haenszel test for .83723 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 1.887 Ce l l s with Expected Frequency < 5 - 2 OF .35561 . 68701 .34359 .36019 .34961 . 61274 4 ( 50.0%) Number of Missing Observations: 0 212 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Mental health issues MENTALH Count CL middle lower 1 2 Row Total .00 21 23 44 83 . 0 1.00 4 5 9 17 . 0 Column 25 28 53 Total 47.2 Chi-Square 52.8 100.0 Value DF Significance Pearson .03231 Continuity Correction .00000 Likelihood Ratio .03238 Mantel-Haenszel test for .03170 linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.245 Ce l l s with Expected Frequency < 5 - 2 OF . 85734 1.00000 . 85720 .85868 . 57531 1. 00000 4 ( 50.0%) Number of Missing Observations: 0 213 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Paranoia CL PARANOIA Count middle lower 1 2 . 00 19 25 1. 00 6 3 Chi-Square Column 25 Total 47.2 Row Total 44 83 . 0 9 17 . 0 28 53 52.8 100.0 Value DF Significance 1.65367 .84552 1.67023 1.62247 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.245 Ce l l s with Expected Frequency < 5 -.19846 .35782 .19623 .20275 . 17921 .27846 2 OF 4 ( 50.0%) Number of Missing Observations: 0 214 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Self-directed CL SELFDIR Count middle lower 1 2 . 00 21 27 1 . 0 0 4 1 Column 25 28 Total 4 7 . 2 Chi-Square Row Total 48 9 0 . 6 5 9 . 4 53 5 2 . 8 1 0 0 . 0 Value DF Significance 2 . 3 8 7 8 4 1 . 1 5 4 7 2 2 . 5 0 9 5 1 2 . 34279 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2 . 3 5 c C e l l s with Expected Frequency < 5 -. 12228 . 2 8 2 5 6 . 11316 . 1 2 5 8 6 . 1 4 1 9 4 . 1 7 6 1 9 2 OF 4 ( 50.0%) Number of Missing Observations: 0 215 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Self-esteem CL SELFEST Count middle lower 1 2 . 00 19 20 1.00 6 8 Column 25 Total 47.2 Chi-Square Row Total 39 73 . 6 14 26.4 28 53 52.8 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .14200 .00419 . 14244 .13932 Minimum Expected Frequency - 6.604 Number of Missing Observations: 0 .70630 .94836 .70587 .70896 216 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Suicide CL SUICIDE Count middle lower 1 2 . 00 21 27 1.00 4 1 Column Total Chi-Square 25 47 .2 28 52.8 Row Total 48 90.6 5 9.4 53 100 . 0 Value DF Significance 2.38784 1.15472 2.50951 2 . 34279 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.358 Ce l l s with Expected Frequency < 5 -.12228 .28256 .11316 .12586 .14194 . 17619 2 OF 4 ( 50.0%) Number of Missing Observations: 0 217 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Unemployment CL UNEMPLOY Count middle lower 1 2 . 00 19 24 1.00 6 4 Column 25 Total 47.2 Chi-Square Row Total 43 81.1 10 18.9 28 53 52.8 100.0 Value DF Signi ficance Pearson .81419 Continuity Correction .30325 Likelihood Ratio .81552 Mantel-Haenszel test for .79883 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 4.717 Ce l l s with Expected Frequency < 5 - 1 OF .36688 . 58185 .36649 .37144 .29079 .48780 4 ( 25.0%) Number of Missing Observations: 0 218 Appendix D Chi Square Analyses of Male Respondents' Choices of Important Issues by Client Class Vocation CL VOCATION Count middle lower 1 2 . 00 2 3 25 1. 00 2 3 Column 25 Total 47.2 Chi-Square Row Total 48 90.6 5 9.4 28 53 52.8 100.0 Value DF Significance Pearson .11389 1 Continuity Correction .00000 1 Likelihood Ratio .11481 1 Mantel-Haenszel test for .11174 1 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.358 C e l l s with Expected Frequency < 5 - 2 OF 4 ( 50.0%) .73576 1.00000 .73473 .73817 .55510 1.00000 Number of Missing Observations: 0 219 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Alcohol CL ALCOHOL Count middle lower 1 2 . 00 18 14 1. 00 13 13 Row Total 32 55.2 26 44 . 8 Column 31 Total 53.4 Chi-Square Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association 27 58 46.6 100.0 Value .22521 .04406 .22523 .22133 Minimum Expected Frequency - 12.103 Number of Missing Observations: 0 Anger CL ANGER Count middle lower 1 2 . 00 23 15 1.00 8 12 Row Total 38 65.5 20 34.5 Column 31 Total 53.4 Chi-Square 27 58 46.6 100.0 Value DF Significance .63510 . 83374 . 63508 . 63803 DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 2 .21890 1.47061 2 .22622 2.18065 .13633 .22525 .13569 .13976 Minimum Expected Frequency - 9.310 Number of Missing Observations: 0 220 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Anxiety CL ANXIETY Count middle lower 1 2 . 00 22 22 1.00 9 5 Row Total 44 75.9 14 24 . 1 Column 31 Total 53.4 Chi-Square 27 58 46.6 100.0 Value Pearson .87114 Continuity Correction .39158 Likelihood Ratio .88286 Mantel-Haenszel test for .85612 lin e a r association Minimum Expected Frequency - 6.517 Number of Missing Observations: 0 Childhood issues CHILDHOO Count CL middle lower 1 2 .00 25 22 1.00 6 5 Column Total Chi-Square 31 53.4 27 46.6 Row Total 47 81. 0 11 19 . 0 58 100 . 0 Value D F DF Significance .35064 .53147 . 34742 .35483 Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association .00657 . 00000 .00657 .00645 .93541 1. 00000 .93538 .93597 Minimum Expected Frequency 5 .121 Number of Missing Observations: 0 221 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Depression CL DEPRESSI Count middle lower 1 2 .00 21 20 1.00 10 7 Chi-Square Column 31 Total 53.4 Row Total 41 70.7 17 29 . 3 27 58 46.6 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for lin e a r association .27927 .05727 .28051 .27445 .59718 . 81087 .59637 . 60036 Minimum Expected Frequency 7.914 Number of Missing Observations: 0 222 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Family or relationship problems RELFAM Column Total Chi-Square CL Count middle lower 1 2 .00 24 19 1. 00 7 8 31 53.4 27 46.6 Row Total 43 74 . 1 15 25.9 58 100.0 Value DF Signi ficance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association . 37398 . 09669 .37335 . 36753 .54084 .75584 .54119 . 54435 Minimum Expected Frequency - 6.983 Number of Missing Observations: 0 223 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Family violence CL FAMVIOL Count m i d d l e lower 1 2 . 00 14 20 1. 00 17 7 Column T o t a l C h i - S q u a r e 31 53.4 27 46.6 Row T o t a l 34 58.6 24 41.4 58 100 . 0 Va l u e DF S i g n i f i c a n c e Pearson C o n t i n u i t y C o r r e c t i o n L i k e l i h o o d R a t i o M a n t e l - H a e n s z e l t e s t f o r l i n e a r a s s o c i a t i o n 4 . 97328 3 . 85276 5 . 08478 4 . 88754 . 02574 . 04966 . 02414 . 02705 Minimum E x p e c t e d Frequency 11.172 Number o f M i s s i n g O b s e r v a t i o n s : 0 224 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Finances/money spending FINANCE Count CL middle 1 lower 2 Row Total . 00 29 24 53 91.4 1. 00 2 3 5 8.6 Column 31 27 58 Total Chi-Square 53.4 46.6 100.0 Value DF Significance Pearson .39773 1 Continuity Correction .02615 1 Likelihood Ratio .39767 1 Mantel-Haenszel test for .39087 1 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.328 Cells with Expected Frequency < 5 - 2 OF 4 ( 50.0%) . 52827 .87154 .52829 . 53184 .43319 , 65567 Number of Missing Observations: 0 225 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class General health issues CL GENHEALT Count middle lower 1 2 .00 27 23 1.00 4 4 Chi-Square Column 31 Total 53.4 Row Total 50 86.2 13 .1 27 58 46.6 100.0 Value DF Significance Pearson .04435 Continuity Correction .00000 Likelihood Ratio .04426 Mantel-Haenszel test for .04358 lin e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 3 . 724 Ce l l s with Expected Frequency < 5 - 2 OF .83321 1.00000 . 83337 . 83463 .56474 1.00000 4 ( 50.0%) Number of Missing Observations: 0 226 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Mental health issues MENTALH Chi-Square CL Count middle lower 1 2 . 00 29 23 1. 00 2 4 Column 31 Total 53.4 Row Total 52 89.7 6 10 . 3 27 58 46.6 100.0 Value DF Significance 1.08829 . 37335 1.09737 1.06952 Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.793 Ce l l s with Expected Frequency < 5 -.29685 . 54118 .29484 .30105 .27077 .40230 2 OF 4 ( 50.0%) Number of Missing Observations: 0 227 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Mood MOOD Chi-Square CL Column 31 Total 53.4 Count middle lower 1 2 . 00 30 22 1. 00 1 5 Row Total 52 89 .7 6 10.3 27 58 46.6 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for li n e a r association Fisher's Exact Test: One-Tail Two-Tail 3 . 63888 2 .17680 3.87062 3.57614 Minimum Expected Frequency - 2.793 C e l l s with Expected Frequency < 5 -Number of Missing Observations: 0 2 OF .05644 . 14011 .04914 . 05862 .06914 . 08734 4 ( 50.0%) 228 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Paranoia PARANOIA Chi-Square CL Count middle lower 1 2 . 00 17 21 1. 00 14 6 Column 31 Total 53.4 Row Total 38 65.5 20 34 . 5 27 58 46.6 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 3 .36118 2 .42250 3.43707 3 . 30323 . 06675 .11960 . 06375 . 06914 Minimum Expected Frequency 9.310 Number of Missing Observations: 0 229 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Self-directed C L S E L F D I R Count middle 1 lower 2 Row Total .00 29 24 53 91.4 1.00 2 3 5 8.6 Column 31 27 58 Total 5 3.4 Chi-Square 46.6 100.0 Value Pearson .39773 Continuity Correction .02615 Likelihood Ratio .39767 Mantel-Haenszel test for .39087 li n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.328 Ce l l s with Expected Frequency < 5 - 2 OF DF Significance .52827 .87154 .52829 .53184 .43319 .65567 4 ( 50.0%) Number of Missing Observations: 0 230 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Self-esteem CL SELFEST Count middle lower 1 2 . 00 21 23 1.00 10 4 Column Total Chi-Square 31 53 .4 27 46.6 Row Total 44 75.9 14 24 .1 58 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 2 .39788 1.53990 2.47143 2.35654 Minimum Expected Frequency - 6.517 Number of Missing Observations: 0 .12150 .21463 .11593 .12476 Unemployment UNEMPLOY CL Count middle lower 1 2 . 00 26 16 1.00 5 11 Column 31 27 Total 53.4 Chi-Square Row Total 42 72 . 4 16 27.6 58 46.6 100.0 Value DF Significance Pearson Continuity Correction Likelihood Ratio Mantel-Haenszel test for l i n e a r association 4.37590 3 .23057 4.43384 4.30046 . 03645 . 07228 . 03523 .03810 Minimum Expected Frequency 7 .448 Number of Missing Observations: 0 231 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Suicide CL SUICIDE Count middle lower 1 2 .00 28 24 1.00 3 3 Chi-Square Column 31 Total 53.4 Row Total 52 89.7 6 10.3 27 58 46.6 100.0 Value DF Significance Pearson .03198 Continuity Correction .00000 Likelihood Ratio .03192 Mantel-Haenszel test for .03143 l i n e a r association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.793 Ce l l s with Expected Frequency < 5 - 2 OF .85807 1.00000 .85821 .85928 .59560 1.00000 4 ( 50.0%) Number of Missing Observations: 0 232 Appendix D Chi Square Analyses of Female Respondents' Choices of Important Issues by Client Class Vocation CL VOCATION Count middle lower 1 2 .00 30 23 1.00 1 4 Row Total 53 91.4 5 8.6 Column 31 Total 53.4 Chi-Square 27 58 46.6 100.0 Value DF Significance Pearson 2.46037 1 Continuity Correction 1.20913 1 Likelihood Ratio 2.57860 1 Mantel-Haenszel test for 2.41795 1 linear association Fisher's Exact Test: One-Tail Two-Tail Minimum Expected Frequency - 2.328 Cel l s with Expected Frequency < 5 - 2 OF 4 ( 50.0%) .11675 .27150 .10832 . 11995 .13635 .17343 Number of Missing Observations: 0 233 Appendix DI Independent Samples T-Tests for Proposed Length of Therapy by Respondent Gender Female Respondents Number Variable of Cases Mean SD SE of mean middle class 24 13.04 7.80 1.59 lower class 20 14.40 8.43 1.89 Mean Difference = 1.36 Levene's test for Equality of Variances: F = .026 P = .873 T-Test for Equality of Means 2-Tail 95% Variances t-value df Sign Mean Diff SE of Diff CI for Diff Equal -.554 42 .582 -1.36 2.45 (-6.30, 3.59) Unequal -.550 39.263 .585 -1.36 2.47 (-6.35, 3.63) Male Respondents Number Variable of Cases Mean SD SE of mean middle class 23 17.52 12.76 2.66 lower class 20 11.95 7.61 1.70 Mean Difference = 5.57 Levene's test for Equality of Variances: F = 3.457 P = .070 T-Test for Equality of Means Variances t-value df 2-Tail Sign Mean Diff SE of Diff 95% CI for Diff (-1.03, 12.17) Equal 1.706 41 .096 5.57 3.27 Unequal 1.765 3.579 .086 5.57 3.16 (-.83, 11.97) 234 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) CI: Expressive/Inexpressive Number Variable of Cases Mean SD SE of Mean middle class 55 4.4364 1 873 253 lower class 56 4.3750 2.171 .290 Mean Difference = .0614 Levene's Test for Equality of Variances: F= 2.579 P= .111 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal .16 109 .874 .385 (-.702, .825) Unequal .16 107.24 .874 .385 (-.701, .824) C2: Pessimistic/Optimistic Number Variable of Cases Mean SD SE of Mean middle class 55 2.0364 1.319 .178 lower cl a s s 56 2.2857 1.615 .216 Mean Difference = -.2494 Levene's Test for Equality of Variances: F= 3.430 P= .067 t-test for Equality of Means 95% Variances t-value df 2 - T a i l Sig SE of D i f f CI for D i f f Equal -.89 109 .375 .280 (-.805, .306) Unequal -.89 105.51 .375 .280 (-.804, .305) 235 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C3: Trustworthy/Untrustworthy Variable Number of Cases Mean SD SE of Mean middle class lower class 55 55 3.9273 3 .8000 1 . 665 1 . 532 . 225 .207 Mean Difference = .1273 Levene's Test for Equality of Variances: F= .442 P= .508 t-test for Equality of Means Variances t-value df 2 - T a i l Sig SE of Diff 95% CI for D i f f Equal Unequal . 42 . 42 108 107.26 . 677 . 677 .305 .305 (-.478, .732) (-.478, .732) C4: Competent/Incompetent Number Variable of Cases Mean SD SE of Mean middle class 54 3.8889 1.436 .195 lower class 56 4.2321 1.640 .219 Mean Difference = -.3433 Levene's Test for Equality of Variances: F= .279 P= .598 t- t e s t for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -1.17 108 .246 .294 (-.927, .240) Unequal -1.17 107.02 .245 .294 (-.926, .239) 236 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C5: Unintelligent/Intelligent Number Variable of Cases Mean SD SE of Mean middle class 55 5.2545 1.443 .195 lower class 55 3.4545 1.525 .206 Mean Difference = 1.8000 Levene's Test for Equality of Variances: F= .099 P= .754 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of Di f f CI for D i f f Equal 6.36 108 .000 .283 (1.239, 2.361) Unequal 6.36 107.67 .000 .283 (1.239, 2.361) C6: Involved/Withdrawn Number Variable of Cases Mean SjD SE of mean middle class 56 5.04 1.66 .22 lower class 55 4.67 1.88 .25 Mean Difference = .37 Levene's test for Equality of Variances: F = 1.567 P = .213 t-test for Equality of Means Variances t-value df 2-Tail Sign Mean Diff Equal 1.079 109 .283 .36 Unequal 1.078 106.942 .283 .36 237 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C7: Industrious/Lazy Number Variable of Cases Mean SD SE of Mean middle class 55 3.5818 1.410 .190 lower class 56 3.8214 1.177 .157 Mean Difference = -.2396 Levene's Test for Equality of Variances: F= 2.723 P= .102 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal -.97 109 .333 .246 (-.728, .249) Unequal -.97 104.95 .334 .247 (-.729, .250) C8: Impulsive/Reliable Variable Number of Cases Mean SD SE of Mean middle class lower class 55 55 2.2364 2 .4909 1 .261 1. 550 . 170 .209 Mean Difference = -.2545 Levene's Test for Equality of Variances: F= 3.412 P= 067 t-test for Equality of Means 95% Variances t-value df 2 - T a i l Sig SE of D i f f CI for D i f f Equal -.94 108 .347 .269 (-.789, .280) Unequal -.94 103.72 .347 .269 (-.789, .280) Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C9: Responsible/Irresponsible Number Variable of Cases Mean SD SE of Mean middle class 55 4.4909 1.814 .245 lower class 55 4.3818 1.748 .236 Mean Difference = .1091 Levene's Test for Equality of Variances: F= .075 P= .785 t-test for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal .32 108 .749 .340 (-.'564, .783) Unequal .32- 107.85 .749 .340 (-.564, .783) C10: Ignorant/Knowledgeable Number Variable of Cases Mean SD SE of Mean middle class 55 4.2909 1.560 .210 lower class 56 3.3929 1.397 .187 Mean Difference = .8981 Levene's Test for Equality of Variances: F= .498 P= .482 t-t e s t for Equality of Means 95% Variances t-value df 2-Tail Sig SE of D i f f CI for D i f f Equal 3.20 109 .002 .281 (.341, 1.455) Unequal 3.19 107.25 .002 .281 (.340, 1.456) 239 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) CI 1: Friendly/Hostile Number V a r i a b l e of Cases Mean SD SE of Mean m i d d l e c l a s s 55 4.7636 1.201 .162 low e r c l a s s 56 4.2143 1.486 .199 Mean D i f f e r e n c e = .5494 Levene's T e s t f o r E q u a l i t y of V a r i a n c e s : F= .906 P= .343 t - t e s t f o r E q u a l i t y o f Means 95% V a r i a n c e s t - v a l u e d f 2 - T a i l S i g SE of D i f f CI f o r D i f f E q u a l 2.14 109 .035 .257 ( . 0 4 0 1 058) Unequal 2.14 105.14 .034 ' .256 (.041, 1.058) C12: Loud/Quiet Number Variable of Cases Mean SD SE of Mean middle cl a s s 55 4.6545 1.506 203 lower cl a s s 56 3.9643 1.629 '.218 Mean Difference = .6903 . Levene's Test for Equality of Variances: F= .088 P= .767 t-test f o r Equality of Means 95% Variances t-value df 2-Tail Sig SE of Diff CI for D i f f Equal 2.32 109 .022 .298 (.100, 1.281) Unequal 2.32 108.61 .022 .298 (.100, 1.280) 240 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C13: Clean/Dirty Variable Number of Cases Mean SD middle class lower class 55 56 2.9818 3.7143 1 .367 1.303 SE of Mean . 184 . 174 Mean Difference = -.7325 Levene's Test for Equality of Variances: F= 5.188 P= .025 t-test for Equality of Means Variances t-value df 2-Tail Sig Equal Unequal -2 . 89 -2 . 89 109 108.52 . 005 . 005 SE of D i f f 95% CI for D i f f .254 .254 (-1.235, -.230) (-1.235, -.230) C14: Emotional/Controlled Variable middle class lower class Number of Cases Mean 55 56 3.8909 2 . 9464 SD 2 . 096 2 . 004 SE of Mean .283 .268 Mean Difference = .9445 Levene's Test for Equality of Variances: F= 2.676 P= .105 t-test for Equality of Means Variances t-value df 2- T a i l Sig SE of D i f f 95% CI for D i f f Equal Unequal 2.43 2.43 109 108.56 . 017 . 017 . 389 .389 (.173, 1.716) (.173, 1.716) Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C15: Moral/Immoral 241 Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 3 . 3273 3.3393 1 .428 1.392 193 186 Mean Difference = -.0120 Levene's Test for Equality of Variances: F= .177 P= .675 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of D i f f 95% CI for D i f f Equal Unequal - . 04 -.04 109 108.79 . 964 . 964 .268 .268 (-.543, .519) (-.543, .519) C16: Incapable/Capable Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 4 .2000 3 . 5536 1.660 1.560 .224 . 208 Mean Difference = .6464 Levene's Test for Equality of Variances: F= .383 P= .537 t-test for Equality of Means Variances t-value df 2-Tail Sig SE of D i f f 95% CI for D i f f Equal 2.11 109 Unequal 2.11 108.30 . 037 . 037 . 306 .306 (.040, 1.252) (.040, 1.253) 242 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C17: Defensive/Receptive Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 2 .4727 2.7857 1.317 1. 615 . 178 .216 Mean Difference = -.3130 Levene's Test for Equality of Variances: F= 2.608 P= .109 t-test for Equality of Means Variances t-value df 2-Tail Sig Equal Unequal -1.12 -1.12 109 105.47 .266 .265 SE of Diff .280 . 279 95% CI for D i f f (-.868, .242) (-.867, .241) C18: Aggressive/Peaceful Variable Number of Cases Mean SD SE of Mean middle class lower class 55 56 2 . 5636 2.5714 1.229 1.189 t-test for Equality of Means Variances t-value df 2-Tail Sig .166 .159 Mean Difference = -.0078 Levene's Test for Equality of Variances: F= .030 P= .863 SE of D i f f 95% CI for D i f f Equal Unequal .03 .03 109 108.71 .973 .973 .229 .230 (-.463, .447) (-.463, .447) 243 Appendix E Client Characteristics: T-Tests for Independent Samples of Social Class (N = 111) C19: Dishonest/Honest Number Variable of Cases Mean SD SE of Mean middle class 55 4.2182 1.524 lower class 56 3.8393 1.558 t-test for Equality of Means .205 .208 Mean Difference = .3789 Levene's Test for Equality of Variances: F= .272 P= .603 95% Variances t-value df 2-Tail Sig SE of D i f f c i for D i f f Equal 1.29 109 Unequal 1.30 109.00 . 198 . 198 .293 .293 (-.201, .959) (-.201, .959) 244 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) CI : Expressive/inexpressive T e s t s of S i g n i f i c a n c e f o r CI u s i n g UNIQUE sums of squares Source of V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 391 83 89 4 40 CL 58 1 58 . 13 .717 DEG 12 1 12 . 03 . 872 CL BY DEG 7 55 1 7 55 1 . 72 . 194 (Model) 9 20 3 3 07 . 70 . 557 ( T o t a l ) 401 03 92 4 36 R-Squared = .023 Adjusted R-Squared = .000 C2: Pessimistic/optimistic Tests of Sig n i f i c a n c e for C2 using UNIQUE sums of squares Source of V a r i a t i o n SS DF MS F Sig of WITHIN+RESIDUAL 202 .55 89 2 28 CL .11 1 11 . 05 . 827 DEG 4 .05 1 4 05 1.78 . 186 CL BY DEG 3 . 51 1 3 51 1 . 54 .217 (Model) 7 .96 3 2 . 65 1.17 . 327 (Total) 210 . 52 92 2 29 R-Squared = .038 Adjusted R-Squared = .005 C3: Trustworthy/untrustworthy Tests of Si g n i f i c a n c e for C3 using UNIQUE sums of squares Source of V a r i a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 242 72 89 2 73 CL 99 1 99 . 36 . 549 DEG 1 65 1 1 65 . 60 . 439 CL BY DEG 00 1 00 . 00 .966 (Model) 2 . 98 3 99 .36 . 779 (Total) 245. 70 92 2 67 R-Squared = .012 Adjusted R-Squared - .000 245 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) C4: Competent/incompetent T e s t s of S i g n i f i c a n c e f o r C4 u s i n g UNIQUE sums Of squares Source of V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 228 41 89 2 57 CL 1 49 1 1 49 . 58 . 448 DEG 3 10 1 3 10 1 .21 . 274 CL BY DEG 62 1 62 . 24 . 623 (Model) 4 49 3 1 50 . 58 . 627 ( T o t a l ) 232 90 92 2 53 R-Squared = .019 Adjusted R-Squared = .000 C5: Unintelligent/intelligent Tests of Significance for C5 using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 211 89 89 2 38 CL 67 55 1 67 55 28.37 . 000 DEG 1 43 1 1 43 . 60 .440 CL BY DEG 3 64 1 3 64 1.53 .220 (Model) 78 77 3 26 26 11.03 .000 (Total) 290 67 92 3 16 R-Squared = .271 Adjusted R-Squared = .246 C6: involved/withdrawn Tests of Significance for C6 using UNIQUE sums of squares Source of Vari a t i o n SS. DF MS F Sig of F WITHIN+RESIDUAL 262 88 89 2 95 CL 1 18 1 1 18 40 . 529 DEG 04 1 04 01 .909 CL BY DEG 33 98 1 33 98 11 51 . 001 (Model) 38 36 3 12 79 4 33 . 007 (Total) 301 25 92 3 27 R-Squared = Adjusted R-Squared- = . 127 . 098 246 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) C7: Industrious/lazy Tests of Significance for C7 using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 149 84 89 1 68 CL 1 86 1 1 86 1 .11 .296 DEG 63 1 63 . 37 . 542 CL BY DEG 3 37 1 3 37 2 . 00 . 161 (Model) 4 89 3 1 63 . 97 .411 (Total) 154 73 92 1 68 R-Squared = .032 Adjusted R-Squared = .000 C8: Impulsive/reliable Tests of Significance for C8 using UNIQUE sums of s Source of Vari a t i o n WITHIN+RESIDUAL CL DEG CL BY DEG (Model) (Total) R-Squared = Adjusted R-Squared = SS DF MS F Sig of F 179.96 89 2.02 5.02 1 5.02 2.48 .119 4.04 1 4.04 2.00 ' .161 2.87 1 2.87 1.42 .237 9.87 3 3.29 1.63 .189 189.83 92 2.06 . 052 020 C9: Responsible/irresponsible Source of V a r i a t i o n WITHIN+RESIDUAL CL DEG CL BY DEG (Model) (Total) r C9 using UNIQUE sums of squares SS DF MS F Sig of F 285 20 89 3 20 17 1 17 . 05 .821 5 59 1 5 59 1.74 . 190 3 23 1 3 23 1.01 . 318 9 44 3 3 15 .98 . 405 294. 65 92 3 . 20 R-Squared = Adjusted R-Squared = . 032 . 000 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) CIO: Ignorant/knowledgeable Tests of Significance for CIO using UNIQUE sums of squares Source of Va r i a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 201 . 35 89 2 . 26 CL 19 . 46 1 19.46 8 . 60 . 004 DEG 01 1 . 01 . 00 . 945 CL BY DEG 5 . 76 1 5 . 76 2 . 55 . 114 (Model) 29 . 81 3 9 . 94 4 . 39 . 006 (Total) 231 . 16 92 2 .51 R-Squared = .129 Adjusted R-Squared = .100 CI 1: Friendly/hostile Tests of Significance for C l l using UNIQUE sums of squares Source of V a r i a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 159.90 89 1. 80 CL 4 . 30 1 4 . 30 2 . 39 . 126 DEG 2 .37 1 2 .37 1 . 32 .254 CL BY DEG 5.47 1 5.47 3 . 04 . 085 (Model) 14.57 3 4.86 2.70 . 050 (Total) 174.47 92 1.90 R-Squared = .084 Adjusted R-Squared = .053 C12: Loud/quiet Tests of Si g n i f i c a n c e Source of V a r i a t i o n for C12 using SS UNIQUE DF sums of MS squares F Sig of F WITHIN+RESIDUAL CL DEG CL BY DEG 223.62 17.01 . 10 13 . 93 89 1 1 1 2 . 51 17.01 . 10 13 . 93 6 . 77 . 04 5 . 54 . 011 . 842 . 021 (Model) (Total) 37 . 50 261.12 3 92 12.50 2 . 84 4 . 97 . 003 R-Squared = Adjusted R-Squared = . 144 .115 248 SS DF MS F Sig of F 172 91 89 1 94 10 46 1 10 46 5 38 . 023 70 1 70 36 . 549 1 01 1 1 01 52 . 474 10. 92 3 . 3 64 1 87 . 140 183 . 83 92 2 . 00 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) C13: Clean/dirty Tests of Significance for C13 using UNIQUE sums of squares Source of Variation WITHIN+RESIDUAL CL DEG CL BY DEG (Model) (Total) R-Squared = .059 Adjusted R-Squared = .028 C14: Emotional/controlled Tests of Significance for C14 using UNIQUE sums of squares source of Variation WITHIN+RESIDUAL CL DEG CL BY DEG (Model) (Total) R-Squared = .173 Adjusted R-Squared = .145 SS DF MS F Sig of F 329 .73 89 3 .70 19 . 64 1 19 64 5 .30 . 024 . 00 1 00 00 .999 38 . 16 1 38 16 10 30 . 002 68 92 3 22 97 6 20 . 001 398 65 92 4. 33 C15: Moral/immoral Tests of Significance for C15 u s i Source of Variation WITHIN+RESIDUAL CL DEG CL BY DEG (Model) (Total) C15 using UNIQUE sums of SS DF MS 178 .28 89 2 00 05 1 05 08 1 08 5 46 1 5 46 5. 53 3 1. 84 183 . 81 92 2 . 00 Sig of F .03 .871 .04 .843 2.73 .102 .92 .435 R-Squared = Adjusted R-Squared = . 030 .000 249 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) C16: Incapable/capable T e s t s of S i g n i f i c a n c e f o r C16 u s i n g UNIQUE sums of squares Source of V a r i a t i o n WITHIN+RESIDUAL CL DEG CL BY DEG (Model) ( T o t a l ) R-Squared = .090 A d j u s t e d R-Squared = .059 SS DF MS F S i g o f F 223 80 89 2 51 6 94 1 6 94 2 76 . 100 1 08 1 1 08 43 . 514 10 92 1 10 92 4 34 . 040 22. 10 3 7 37 2 93 . 038 245. 89 92 2. 67 C17: Defensive/receptive Tests of Significance for C17 using UNIQUE sums of squares Source of Variation WITHIN+RESIDUAL CL DEG CL BY DEG SS DF MS F 162 . 17 89 1 82 2 .46 1 2 46 1.35 06 1 06 . 03 1 08 1 1 08 .59 3 08 3 1. 03 . 56 165. 25 92 1. 80 Sig of F . 249 . 855 .444 (Model) _,.uo j ±.US .56 .640 (Total) 165.2S Q-5 i R-Squared = .019 Adjusted R-Squared = .000 C18: Aggressive/peaceful Tests of Significance for CI 8 u Source of Var i a t i o n SS WITHIN+RESIDUAL 136 . 43 CL 09 DEG 42 CL BY DEG 3 31 (Model) 3 . 85 (Total) 140. 28 UNIQUE sums of squares DF MS F Sig of F 89 1 .53 1 . 09 . 06 . 814 1 42 .27 . 602 1 3 31 2.16 . 145 3 1 28 ' .84 . 477 92 1 52 R-Squared = Adjusted R-Squared = .027 .000 Appendix F Client Characteristics: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) C19: Dishonest/honest T e s t s o f S i g n i f i c a n c e f o r C19 u s i n g UNIQUE sums of squares Source of V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 204 87 89 2.30 CL 1 . 19 1 1 . 19 . 51 .475 DEG 41 1 .41 .18 . 676 CL BY DEG 13 . 74 1 13 .74 5 . 97 .017 (Model) 17 . 13 3 5.71 2.48 . 066 ( T o t a l ) 222 . 00 92 2 .41 R-Squared = Adjusted R-Squared = . 077 . 046 251 Appendix F Problem Evaluation: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) Perceived Client Self-Concept Tests of Significance for T using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 68 . 63 87 .79 CL .33 1 .33 42 . 518 DEG . 33 1 . 33 42 . 518 CL BY DEG . 80 1 .80 1 02 .316 (Model) 1 .48 3 .49 62 . 602 (Total) 70 . 11 90 .78 R-Squared = .021 Adjusted R-Squared = .000 Prognosis Tests of Significance for I using UNIQUE sums of squares Source of Vari a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 121.74 88 1.38 CL .71 1 .71 .52 .475 DEG 2.58 1 2.58 1.86 .176 CL BY DEG .41 1 .41 .30 .587 (Model) 3.94 3 1 3 1 .95 4 2 o (Total) 125.68 91 1.38 R-Squared = .031 Adjusted R-Squared = .000 252 Appendix F Problem Evaluation: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) Likelihood of Referring for Psychiatric Assessment Tests of Significance for PSYC2 using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 296 15 88 3 37 CL 66 1 66 .20 . 660 DEG 1 31 1 1 31 .39 . 535 CL BY DEG 1 52 1 1 52 .45 .504 (Model) 4 05 3 1 35 . 40 .752 (Total) 300 21 91 3 30 R-Squared = .014 Adjusted R-Squared = .000 Personal Interest in Treating This Client Tests of Significance for G using UNIQUE sums of squares Source of Vari a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 193 06 88 2 19 CL 1 19 1 1 19 54 .464 DEG 20 62 1 20 62 9 40 . 003 CL BY DEG 7 96 1 7 96 3 63 . 060 (Model) 27 85 3 9 28 4 23 . 008 (Total) 220 91 91 2 43 R-Squared = .12 6 Adjusted R-Squared = .096 Appendix F Problem Evaluation: Analyses of Variance by Client Social Class and Respondent Level of Education (N = 89) Severity of Presenting Problem Tests of Significance for PRESPR2 using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 59 . 72 89 . 67 CL . 98 1 .98 1. 47 . 229 DEG . 57 1 . 57 . 85 .359 CL BY DEG . 04 1 . 04 . 06 . 800 (Model) 1 . 57 3 . 52 .78 . 508 (Total) 61 . 29 92 . 67 R-Squared = .026 Adjusted R-Squared = .000 Perceived Client Motivation for Change Tests of Significance for A using UNIQUE sums of squares Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 132 . 13 89 1 .48 CL 1 . 97 1 1 . 97 1. .33 .253 DEG 4 . 01 1 4 . 01 2. .70 . 104 CL BY DEG 1 .41 1 1 .41 .95 .332 (Model) 7 . 67 3 2 . 56 1. .72 . 168 (Total) 139 .81 92 1 . 52 R-Squared = Adjusted R-Squared = . 055 . 023 254 Appendix G Problem Evaluation: Analyses of Variance by Respondents' Years of Experience and Client Social Class (N = 99) Severity of Presenting Problem: Source of V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 64 . 84 99 .65 EXPERYRS 1 . 04 4 .26 .40 .811 CL . 63 1 . 63 . 97 . 328 EXPERYRS BY CL 2 . 93 4 .73 1 . 12 .352 (Model) 4 93 9 .55 . 84 . 585 (T o t a l ) 69 76 108 . 65 R-Squared = 071 A d j u s t e d R-Squared = 000 Client Motivation for Chan ge: Source of V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 145 .29 99 1 . 47 EXPERYRS 3 . 74 4 . 94 .'64 . 637 CL 3 . 08 1 3 . 08 2 . 10 . 151 EXPERYRS BY CL 23 . 48 4 5 . 87 4 .00 . 005 (Model) 29 . 32 9 3.26 2 .22 . 027 ( T o t a l ) 174 . 61 108 1. 62 R-Squared = . 168 A d j u s t e d R-Squared = . 092 Client Self-Concept: Source o f V a r i a t i o n SS DF MS F S i g of F WITHIN+RESIDUAL 71 . 44 97 . 74 EXPERYRS 2 . 01 4 . 50 68 . 607 CL 14 1 . 14 20 . 658 EXPERYRS BY CL 5 . 12 4 1 .28 1. 74 . 148 (Model) 8 . 73 9 . 97 1. 32 .238 ( T o t a l ) 80. 17 106 .76 R-Squared = .109 A d j u s t e d R-Squared = .026 255 Appendix G Problem Evaluation: Analyses of Variance by Respondents' Years of Experience and Client Social Class (N = 99) Prognosis: Source of V a r i a t i o n SS DF . MS F S i g of F WITHIN+RESIDUAL 133 60 97 1 38 EXPERYRS 5 57 4 1 39 1 01 .405 CL 2 09 1 2 09 1 52 . 221 EXPERYRS BY CL 17 66 4 4 41 3 21 .016 (Model) 26 75 9 2 97 2 16 . 032 (T o t a l ) 160 36 106 1 51 R-Squared = .167 Adjusted R-Squared = .090 Likelihood of Referring Client for Psychiatric Assessment: Source of Variation SS DF MS F Sig of F WITHIN+RESIDUAL 334 90 98 3 . 42 EXPERYRS 2 32 4 . 58 . 17 .953 CL 3 48 1 3 .48 1 . 02 .315 EXPERYRS BY CL 10 57 4 2 . 64 . 77 . 545 (Model) 19 42 9 2 . 16 . 63 .768 (Total) .354 32 107 3 .31 R-Squared = .055 Adjusted R-Squared = .000 Personal Interest in Treating the Client: Source of Vari a t i o n SS DF MS F Sig of F WITHIN+RESIDUAL 225 . 64 98 2 30 EXPERYRS 12 95 ' 4 3 24 1 41 .238 CL 54 1 54 24 . 628 EXPERYRS BY CL 13 20 4 3 30 1 43 .229 (Model) 25 28 9 • 2 81 1 22 .292 (Total) 250 92 107 2 35 R-Squared = Adjusted R-Squared = . 101 . 018 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053932/manifest

Comment

Related Items