UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Burnout among high school counsellors Hooper, James T. 1998

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

Item Metadata


831-ubc_1998-0472.pdf [ 4.52MB ]
JSON: 831-1.0053929.json
JSON-LD: 831-1.0053929-ld.json
RDF/XML (Pretty): 831-1.0053929-rdf.xml
RDF/JSON: 831-1.0053929-rdf.json
Turtle: 831-1.0053929-turtle.txt
N-Triples: 831-1.0053929-rdf-ntriples.txt
Original Record: 831-1.0053929-source.json
Full Text

Full Text

BURNOUT AMONG HIGH S C H O O L COUNSELLORS by JAMES T. HOOPER B.A. (Hons.), University of Manitoba, 1969 B.J. (Hons.), Carleton University, 1972 M.A., McGill University, 1972 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER O F ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Counsell ing Psychology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1998 © James T. Hooper, 1998 In p resent ing this thesis in partial fu l f i lment of t h e requ i remen ts f o r an advanced d e g r e e at t h e Universi ty o f Brit ish C o l u m b i a , I agree that t h e Library shall make it f ree ly available f o r re ference and study. I fu r ther agree that permiss ion f o r extens ive c o p y i n g o f this thesis f o r scholar ly purposes may be g ran ted by the head of m y d e p a r t m e n t o r by his or 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 i on o f this thesis fo r f inancial gain shall n o t be a l l o w e d w i t h o u t my w r i t t e n pe rmiss ion . D e p a r t m e n t of The Univers i ty of Brit ish C o l u m b i a Vancouver , Canada Date Jc^TJ^d^ 2 9. / ? 9 f DE-6 (2/88) A b s t r a c t The psychological well-being of those who counsel adolescents is an important issue, but there has been little research on the topic. Burnout f rom job-related stress in the helping professions has been shown to influence negatively the professionals' job satisfaction and performance by eroding their benevolence and commitment. Three aspects of burnout — emotional exhaust ion, depersonal izat ion, and impaired personal accompl ishment — were measured in this study. The High School Counsellor Questionnaire, designed for this investigation, was mailed to 265 members of the British Columbia School Counsel lors Association who were working in high schools. The return rate was 6 1 . 5 1 % ; the usable N was 157. The questionnaire measured the extent of counsellors' burnout and their perception of their own social support (from family, fr iends and others) and administrative support (defined as support from the principal). The questionnaire also gathered information on selected personal and job characterist ics of the counsellor: age, gender, counsell ing experience, level of education, school size, and proportion of work t ime devoted to counsel l ing. Burnout levels were shown by t -tests to be low in relation to Maslach Burnout Inventory norms for mental health professionals except on the emotional exhaust ion scale, where counsellors' scores were significantly higher (t =4.26; p < . 0 0 1 ) . This result may reflect the ambiguity of the counsellors' role, and ever-increasing demands on their t ime and energy. ii The association of burnout with the independent variables was explored by correlation, multiple regression analysis, t -tests, and one-way ANOVA. As expected, burnout was generally negatively correlated with perceived social and administrative support. Gender, age, educat ion, counsell ing experience, and school size were not significantly associated with burnout. Percentage of work t ime designated for counsell ing was significantly ( r= .26 ; p < 0 1 ) correlated with feel ings of personal accompl ishment. Future research might consider the principal's gender, and counsellors' specific work duties, case loads, paperwork burden, role ambiguity, collegial support, marital status, caregiver role at home, and coping strategies for dealing with stress. iii Table of Contents Abstract ii Table of Contents iv List of Tables vi List of Figures vii Foreword viii Chapter I The Problem 1 Introduction to the Problem 1 Statement of the Problem 2 Rationale for the Study 2 Defini t ions 3 Del imitat ions 5 Limitat ions 5 Chapter II Review of the Literature 8 Burnout 8 Perceived Social Support 14 Perceived Administrat ive Support 17 Article Review 21 Chapter III Hypotheses and Methodology 28 Hypotheses 28 Methodo logy 31 Design 31 Data Collection and Procedures 31 Populat ion and Sample 31 Procedures 32 Measures 33 Data Analysis, Organization and Interpretation 38 iv Chapter IV Results Participants Overview of the Data Levels of Burnout Multivariate Tests Regression Burnout and Exploratory Variables Relationships among Independent Variables 41 41 46 46 49 49 53 56 Chapter V Discussion Summary and Implications Social and Administrat ive Support Other Relevant Variables Regression Exploratory Variables Relationships among Independent Variables Levels of Burnout Further Discussion Recommendat ions for Future Research 59 59 59 61 64 67 69 73 75 79 References 85 Appendix A. Notes on Returns and Scoring B. Covering Letter C. High School Counsellor Questionnaire 90 92 94 v List of Tables 1. Means, Standard Deviations and Correlations among All Variables 47 2. Maslach Burnout Inventory (MBI) Means and Standard Deviations 48 3. Correlat ions between Burnout and Independent Variables Employed in the Multiple Regression Analysis 50 4. Stepwise Multiple Regression of Independent Variables on Emotional Exhaustion 51 5. Stepwise Multiple Regression of Independent Variables on Depersonalization 52 6. Stepwise Multiple Regression of Independent Variables on Personal Accomplishment 53 7. Means, Standard Deviations, and F Values of Burnout Scores by Level of Education 54 8. Means, Standard Deviations, and F Values of Burnout Scores by School Size 55 9. Means, Standard Deviations, and F Values of Administrative Support and Percentage Counsell ing by School Size 57 10. Means, Standard Deviations, and t Values of Administrative Support and Social Support by Gender 58 vi List of Figures 1. Participants by Age 42 2. Participants by Gender 42 3. Participants by Experience 43 4. Participants by Education 44 5. Participants by School Size 44 6. Participants by Percentage of T ime Counsell ing 45 vii F o r e w o r d For their help with this research, I thank my thesis committee: my supervisor, Dr. Bill Borgen; Dr. Beth Haverkamp; and Dr. Frank Echols, of the Department of Educational Studies. Thanks also to Dr. Chris Lovato, then of the Institute for Health Promotion Research, in whose Counsel l ing Psychology 579 class this research was envisaged; the British Columbia School Counsel lors Associat ion for their co-operat ion; and the many high school counsellors who helped me with suggest ions and in the piloting of my questionnaire. viii Chapter I The Problem Introduction to the Problem "Who will care for the caregivers?" By the t ime the expression burnout first appeared in the psychological literature in the mid-1970s, this compell ing question was decades old, perhaps centuries. But the concept of burnout focused anew the attention of theorists and researchers on the often intense personal strain suffered by counsellors, priests, doctors, nurses, social workers and others who deal regularly with people who are in physical or emotional pain. Stress has been shown in a myriad of studies to have an adverse effect on physical and mental health (Phillips, 1982; Long & Kahn, 1993; Capner & Caltabiano, 1993; Dryden, 1995). It is ironic but understandable that although counsel lors have devised programs to help persons manage stress, many counsel lors themselves have trouble handling stress, particularly job-related stress. Workers, including counsellors, in person-oriented occupations have been shown to be particularly at risk for burnout, which has been shown to have a negative influence on job satisfaction and performance (Brady, Healy, Norcross & Guy, 1995, pp. 8-10). Burnout has been shown to rob human service workers of their idealism and benevolence and thus of much of their effectiveness. "When helping professionals become less caring and compassionate, the quality of care 1 declines" (Cherniss, 1995, p. 197). The toll on workers and organizations in terms of health problems, absenteeism, and inefficiency is considerable, and the importance of f inding solutions is widely acknowledged. Statement of the Problem Although considerable research has been done in several separate, and somet imes linked, areas — work stress, social support, burnout, supervisory behaviour, counsellor attitudes, and characterist ics of the counsellor and of the job — the relationships among some of these elements have not been establ ished. In particular, little is known about burnout in high school counsellors. A recent, comprehensive book presenting research f indings about stress in counsell ing (Dryden, 1995) has chapters on just about every conceivable counsell ing sett ing except the school. Interventions for burnout in counsell ing and other helping professions have been designed, but their effect iveness is often hampered by lacunae in the research (for a review, see Kahili, 1988). Counsellor self-care is another potential remedy, but its methods and effectiveness are subjects of controversy (for a review, see Will iams, 1995). Rationale for the Study Before meaningful action can be taken to combat burnout and its negative consequences on high school counsellors, factors that have an impact on burnout need to be identified. As no article on burnout among high school professionals in Canada — 2 or high school counsellors anywhere — was found in the literature, this is a pioneering study. It a ims to offer researchers some information on which future investigations and interventions can be based. Counsel lors themselves also need to know about elements of their personal and work situations that are related to burnout. The results of this study will, it is hoped, be disseminated among British Columbia high school counsel lors through their professional associat ion, the British Columbia School Counsel lors Associat ion, and perhaps more widely through publication of the results. Such information will aid them in understanding burnout and may reduce the feelings of isolation experienced by many counsellors, especially those affected by burnout. This study focuses on two major potential influences on burnout among these counsel lors: perceived social support and perceived administrative support. The influence on burnout of counsellors' gender, age, counsell ing experience, level of education, school size, and percentage of work t ime designated for counsell ing will be considered as well. Three aspects of burnout — emotional exhaust ion, depersonal izat ion, and impaired personal accompl ishment — will be measured. Definitions This investigation's three key concepts are burnout, perceived social support, and perceived administrative support. These terms are defined succinctly in this section according to the respective models chosen for use in the study. Elaborations on the three terms will be found in the review of the literature in the next chapter. 3 Burnout: One of the leading theoreticians and researchers in the burnout field is the American social psychologist Christ ina Maslach. Her definition of the burnout phenomenon is adopted for this study: Burnout is a syndrome of emotional exhaust ion, depersonalization, and reduced personal accompl ishment that can occur among individuals who do "people work" . . . It is a response to the chronic emotional strain of dealing extensively with other human beings, particularly when they are troubled or having problems . . . the stress arises f rom the social interaction between helper and recipient (Maslach, 1982, p. 3). Emotional exhaustion is defined as feelings of depletion, overextension, fatigue — a sense that one can no longer give at a psychological level. Depersonalization refers to unfeeling, impersonal and somet imes cynical responses to other people and their problems. Reduced personal accomplishment denotes a tendency to evaluate one's own job performance negatively, and to be dissatisfied and unfulfil led in one's work with others (Maslach & Jackson, 1981, p. 1). Perceived social support: Cobb (1976), in a seminal article on the subject, def ined perceived social support as information leading people to one or more of three beliefs: that they are cared for and loved, that they are esteemed and valued, and that they belong to a network of communicat ion and social obligation. This conceptual ization, as developed by later researchers including Cutrona and Russell (1987) and Vaux, Phillips, Holly, Thomson, Wil l iams and Stewart (1986), is used in the present investigation. 4 Perceived administrative support: According to Arends (1982), whose model of administrative support is employed in this inquiry, the principal's support is crucial to the work of the other professionals in the school. Arends describes perceived administrative support as teachers ' and counsel lors' belief that their principal offers them the professional, material and emotional support they need to do their jobs in an effective and fulfilling way. Such support includes, for instance, the principal's explicit statements of support for programs and workers, smoothing the way for workers to clarify their work roles and to resolve conflict, defending employees from crit icism, giving them something of value (such as money, t ime, comfortable office space), and offering t ime and attention to employees' work and concerns. Delimitations This study will make no attempt to investigate counsellors in any sett ing other than the high school. Similarly, it will measure burnout only, not neurotic symptoms, depression, physical problems, absenteeism, or other manifestations of stress. Moreover, the significant issue of identifying specific sources of counsellor stress must be left to other researchers (for reviews of this work, see Cohen, 1988; Wessel ls, 1989). Limitations Conclusions derived from this inquiry will of course be limited by the sample. There is a need for caution in generalizing beyond the province where the research was conducted, and beyond the participants' constellation of 5 personal and job characterist ics. Additionally, the participants were volunteers; it is difficult to know how this self-selection influenced the data. Findings are also limited by the written format of the survey. Moreover, they need to be interpreted in the light of validity concerns associated with self-report measures. Observat ions, record searches, and peer and supervisory ratings could be employed in later studies to develop further both our understanding of counsellor burnout and the validity of its measurement. There are a number of specific limitations. One is the study's use of the Administrative Support Measure, which was devised for the purposes of this research and which, al though piloted, does not benefit f rom documented evidence of its reliability or validity. Unfortunately, there was no valid instrument available to measure administrative support in the sense it is defined for this investigation. A second limitation is that there may be some overlap between social support and administrative support as measured, respectively, by the Social Support Appraisals Scale and the Administrat ive Support Measure. This multicoll inearity may arise because the Social Support Appraisals Scale does hot restrict respondents f rom considering their principal as a fr iend while responding to the survey; some respondents may, indeed, think of their principals as fr iends and think of them when considering the support of fr iends and "other people". However, this is likely to be only a minor problem. The same difficulty would arise f rom using any of the instruments evaluating social support that were considered for use in the present research. The Social Support Appraisals Scale was chosen for its ease of use and because considerable information about its properties is available. A third limitation of the study is that some high school counsellors are not 6 members of their provincial specialist associat ion, the B. C. School Counsel lors Associat ion, and thus, for reasons to be explained, were far less likely than were members to participate in the study. It is impossible to know the effect of this difference on the results. Perhaps being a member of the association would heighten a counsellor 's perceived social support; however, it is possible too that counsellors who feel well supported socially would be less inclined to join the associat ion. A fourth limitation is the possibility that a few counsellors might have been reluctant to respond in a way that seemed negative to some items on administrative support. This reluctance, if indeed it did exist, might have s temmed from loyalty to the principal or f rom hesitation about making an unfavourable appraisal of a principal's performance when the two had not discussed the counsellor 's discontent. However, this problem is not likely to be major, and in any case is inherent in this type of research. A fifth, and more important, limitation of the study is similarly present in all other research using correlational methods: No inference about causation may be drawn. Although terms such as predict and explain may be used to link the independent variables to the dependent ones, the word cause must be avoided because the direction of influence among the variables cannot be proven. 7 Chapter II Review of the Literature In this chapter, there is further information about this study's three key concepts. Then, there is a review of the literature on researchers' various attempts to link two or more of the variables explored in the present invest igat ion. Burnout Sisyphus, the mythical ancient Greek condemned to a dreadful eternity of repeatedly pushing a stone up a mountain only to have it roll all the way down again, is the archetypal candidate for a modern syndrome. His frustrating, futile, hopeless and unaided labour would quickly induce in him the emotional, spiritual and physical depletion that has come to be known as burnout. This word has long been used popularly to describe fat igue or mild depression, but the term entered the literature of psychology in 1974. Herbert Freudenberger, an Amer ican psychoanalyst who had worked alongside strained col leagues in the free-clinic movement, conceived of burnout as the exhaust ion of the energy and resources of social service workers and helping professionals as a result of the incessant pressures of working with demanding and needy individuals. He descr ibed burnout as physical or emotional depletion, apathy, disil lusionment, or low morale caused by the stress of work (Freudenberger, 1974). Various conceptual izat ions of the phenomenon developed, with the 8 result that present-day dictionaries give diverse but interlocking definitions of burnout. For example: "extinction of energy, motivation or incentive" (Third Barnhart Dictionary of New English); "a total loss of energy and interest and an inability to function effectively, experienced as a result of excessive demands upon one's resources or chronic overwork" (Coll ins English Dictionary); and "emotional exhaust ion, somet imes accompanied by depression, caused by attempting to help mentally disordered people or others under severe stress" (The International Dictionary of Psychology). The last definition encapsulates the two most prominent and enduring conceptual izations of burnout and those on which the most research is based — those of three social psychologists working in the United States. Christ ina Maslach defines burnout as "a syndrome of emotional exhaust ion, depersonalization, and reduced personal accompl ishment that can occur among individuals who do 'people w o r k ' . . . a response to the chronic emotional strain of dealing extensively with other human beings, particularly when they are troubled or having problems . . . t h e stress arises from the social interaction between helper and recipient" (Maslach, 1982, p. 3). Ayala Pines and Elliot Aronson say burnout is "subjectively experienced as a state of physical, emotional and mental exhaustion caused by long-term involvement in situations that are emotionally demanding" (Pines & Aronson, 1988, p. 9). These are the two predominant definitions, but there is a welter of other ones, and they vary in style as well as in content. Some definitions include causes, some mention consequences, and some give only symptoms. Partly as a result of this variation, the borderline separating burnout f rom such disorders as depression and anxiety is somet imes blurred. While an understanding of 9 depression, anxiety and related problems may be very helpful to burnout researchers, a specific emphasis on burnout is desirable. There appears to be a gap between the theoretical work on burnout and the various anti-burnout programs that have been implemented. In reviewing interventions designed to forestall or reduce burnout, one is struck by how tentative and piecemeal they are and how their developers appear to be grasping to make connections to the theory rather than, as one might expect, to be lifting the interventions from a theoretical base (e.g. Pines & Maslach, 1980; Corcoran & Bryce, 1983; Carrilio & Eisenberg, 1984; Pines & Aronson, 1988; Hyman, 1993). Often it appears they are trying to devise something that works, and theory is an afterthought. There is an unmistakeable air of, "Let's see what works" and even of the twisted adage, "It may work in practice, but it will never work in theory." Two noted psychotherapists put it more eloquently: Unfortunately, much of the advice [about dealing with burnout] seems f ragmented, pointing to theoretical f rames of reference that are incomplete and underdeveloped in the literature. W e believe that in order to deal with the problem of professional exhaust ion, an integrated theory must be constructed which encompasses the various underlying causes. Only after such an integrated paradigm is developed can valid and effective solutions for prevention and treatment of burnout be formulated (Grosch & Olsen, 1994, p. xii). Al though the word burnout is occasionally extended, in popular and even professional writ ing, to describe psychological strain unrelated to work, the original, job-related sense of the term is used here. Also, our attention is on research on counsellors, therapists, social workers and the like rather than on medical or other personnel. 10 Burnout is absent as a diagnosis f rom the Diagnostic and Statistical Manual of Mental Disorders — Fourth Edition (DSM-IV), the bible of many of the mental health professionals who design research into, and treatment for, stress-related difficulties. Burnout was likewise absent f rom previous versions of the manual. The social psychologists' burnout research seems not to have permeated the world view of the psychiatrists who determine the contents of the DSM. Perhaps the psychiatrists see burnout as a set of t roublesome work-related attitudes rather than as a clinical syndrome. In DSM-IV the closest category to burnout is V62.2 — Occupational Problem; two examples given are job dissatisfaction and uncertainty about career choices (American Psychiatric Associat ion, 1994, p. 685). In devising research and interventions, professionals may want to base their work explicitly on categories in the D S M for any number of reasons: Possibly they f ind it easier to get funding, or burnout somehow seems an unofficial or casual diagnosis, or psychiatrists don't want to hear about a syndrome if it is not in "the book", or other factors are involved. Whatever the case, the absence of burnout f rom the D S M may be one reason for (or one consequence of) the lack of an integrated view of the syndrome across various disciplines. Burnout has been recognized, though, in the latest edition of the international diagnostic manual , the International Classif ication of Diseases — Tenth Edition (ICD-10). Burnout is a new diagnosis, coded Z73.0, in a section called Supplementary Classif ication of Factors Influencing Health Status and Contact with Health Services. No criteria are given, but at least the inclusion of burnout represents a change f rom previous editions, where burnout was not mentioned. In ICD-9, for instance, "adverse effects of work environment" and "other occupational c i rcumstances or maladjustment" were the categories 11 closest to burnout. The recent inclusion of burnout in the ICD-10 may be a significant step toward specific measures to fight the syndrome. (World Health Organization, 1992, p. 310; Grosch & Olsen, 1994, p. 136). At all events, much of the scholarly literature on burnout is devoted to the conceptual ization, prevalence, prevention and treatment of burnout rather than to its etiology. Most articles, both the descriptive and the prescriptive ones, are either impressionistic, personal, anecdotal , philosophical, or small-scale, or some combinat ion of these. Empirical studies are scarce. A glance at the elements found to be associated with burnout provides a useful context for our discussion. In a meta-analysis, Soderfeldt, Soderfeldt, and Warg (1995, p. 642) summar ize factors shown by various researchers to be related to the three scales of the Maslach Burnout Inventory, the most commonly used measure. Factors related to emotional exhaust ion scores included role conflict, boredom, low job satisfaction, intention to change jobs, and working with chronically ill patients. Related to depersonalization scores were value conflict, high education level, caseload, and lack of supervisory support. Factors related to personal accompl ishment scores included role ambiguity, bad working condit ions, unfair promotional practices, public sector work, Type A personality, and a dearth of team support. Other perspectives include that of Freudenberger, who is a psychoanalyst. He sees burnout mainly as an internal, personal matter, a result of the dichotomy between a person's expectat ions and reality, al though he does not entirely rule out the importance of organizational factors in creating or preventing burnout (Freudenberger, 1983, p. 27). By contrast, some behaviour-oriented stress consultants see burnout as a type of strain lodged in the body, though they are not clear about how the process works; they somet imes favour 12 t reatments such as systematic desensit ization, biofeedback and yoga (Farber, 1983, p. 15). A more philosophical approach is taken by certain communi ty psychologists, who see burnout resulting from a weak commitment to socially relevant work rather than from excessive commitment to a job. "Burnout is not a problem of individual coping or adaptation. . . Burnout results f rom the loss of moral purpose and commitment in work" (Cherniss & Krantz, 1983, p. 211). They say the cure is on a social level and call for formation of "ideological communit ies" of l ike-minded persons working toward common goals. Cherniss and Krantz give the examples of Montessori workers and nuns, who work long hours but have a mission; however, the authors claim that the concept can be extended widely into the work world. In a similar vein, counsellor educator J. M. Day asserts that the motivation that made someone into a counsellor in the first place must be maintained and continually refreshed if both well-being and effectiveness are to endure (Day, 1994). O n the other hand, a number of family systems theorists (e.g. Ulrich & Dunne, 1986) see burnout as an effect of unresolved conflict in a person's current family or family of origin. Some organizational consultants favour yet another approach: "Since w e believe that cont inuous and prolonged uncertainty is the basis for burnout, the most effective interventions are those that are directed at reducing uncertainty" (Selder & Paustian, 1989, p. 79). It is important to note that with burnout, the cure may come from a different direction than the cause. T o explain: When one's shoulder is separated, the way to fix it is to put it back into place — that is, simply to reverse the dislocation. In the case of burnout, the problem may not be reversible; the cause may be in the environment and absolutely unalterable. Therefore, 13 burned-out people may have to take compensatory action which apparent ly has nothing to do with the cause of their distress but which may build them up so they may handle job stress better. In fact the burnout literature, and particularly the research on treatment approaches, tends to be based on the extensive research by social psychologists on coping strategies. This tendency is not surprising, as social psychologists focus on the relationship between personal and environmental factors; moreover, their research forms the basis of treatment programs designed by counsel lors and others. Coping strategies can be conceptual ized in several ways, but the most useful distinction seems to be tripartite: 1. individual strategies, things that can be done on one's own; 2. social strategies, things that need the co-operat ion of several people, typically co-workers, family and fr iends; and 3. organizational strategies, related to administrative policies and actions. From the second category emerges the theme of social support, which is the touchstone of much burnout research and the cornerstone of some interventions. From the third category above emerges the theme of administrative support, which is also seen as a key element in the research. These themes are now considered in turn. Perceived Social Support While helping professionals generally agree that social support is an important element in promoting health, there is less agreement about exactly what such support is, how it works, and how it can be evaluated. The literature that has developed in the last quarter-century on social support is vast, and it is made complicated by the fact there are several fundamental ly different ways of 14 conceptual izing the phenomenon, and, in consequence, many different ways of measuring it. Social support has been variously defined, for example, as support available to a person through social ties to other individuals, groups, and the community; an interpersonal transaction in which concern, aid and information about oneself and the environment are transmit ted; an internal state of met needs; and the availability of psychosocial resources. In empirical studies, social support has been evaluated, for instance, by the number of people offering it, by various classifications of sources, by its f requency and intensity, and by satisfaction with it. Diverse researchers have measured social networks, social bonds, meaningful social contacts, and the number of persons in whom one can confide. Most important for this investigation is the distinction between enacted social support (specific support-giving events) and perceived social support (the belief that one is or will be supported). The latter concept, as formulated by Cobb (1976) in a seminal article on the subject and developed by others including Cutrona and Russell (1987) and Vaux, Phillips, Holly, Thomson, Wil l iams and Stewart (1986), is adopted in this study for the fol lowing reason. Cobb's thesis, that perception of social support — i.e., information leading a person to believe he or she is cared for, valued, or part of a network of social obligation — is the crucial variable, seems particularly powerful in an investigation of counsellors. Such persons often work in relative isolation and with troubled individuals; to maintain their perspective and their effectiveness, it may be they need to know they have social support available. If they in fact have such support but are not fully aware of having it, its allegedly helpful effect may be lost on them. Cobb's concept of social support emphasizes the degree 15 of confidence a person has in being valued, cared for, and / or able to count on others in t ime of need. According to Barrera: [ K n o w l e d g e of people's subjective appraisals of the adequacy of support is more critical to the prediction of their well-being than simply collecting information about the number of supporters or the quantity of support ive behaviours to which they have access (Barrera, 1981, p. 85). This phenomenological view prompted much further research. Cutrona and Russell (1987), working on the basis of Cobb's model, conceived of six "provisions" or classes of perceived support that may be obtained from relationships with others. All six are v iewed as important and even essential to individuals if they are to feel properly supported and avoid isolation. More than one provision may be obtained from the same person. The provisions are divided into two categories, assistance-related and non-assistance-related. The assistance-related category includes two provisions, guidance (advice or information) and reliable alliance (the belief that others can be counted on for tangible assistance). The non-assistance-related category comprises the other four provisions: reassurance of worth (acknowledgment of one's abilities and value by others), opportunity for nurturance (the sense that others depend on one for their wel l-being), attachment (emotional c loseness providing a sense of security), and social integration (a sense of belonging to a group that shares interests and recreational activit ies). These provisions capture (albeit somet imes in different words) the major forms of social support as delineated by several leading theorists. Vaux, Phillips, Holly, Thomson, Wil l iams and Stewart (1986) see social 16 support as a meta-construct comprising several constructs: 1. support network resources (the size, structure and characteristics of networks); 2. supportive acts (e.g. l istening, advising, comfort ing, lending money, assisting with tasks); and 3. subjective appraisals of support (i.e. beliefs that one is cared for, respected, and having one's needs met). Like Cobb and Cutrona and Russell, they see perceptions of support as the most important of the three aspects: Subjective appraisals of support appear to be especially important in regard to psychological well-being. In several studies, for instance, satisfaction with support or perceived adequacy of support has shown a stronger relationship to distress or well-being than did social support network measures (Vaux et al. , 1986, p. 197). Perceived Administrative Support Support within the school system for counsellors and teachers is widely perceived as a crucial e lement in developing a healthy school c l i m a t e — an atmosphere that al lows adults to work effectively and that promotes student learning and growth — as well as in maintaining professional morale. There are several approaches to evaluating such support. Some researchers (e.g. Donaldson & Coladarci, 1987) emphasize the importance of school cl imate in general and endeavour to devise ways of measuring it. Others (e.g. Sutton & Fall, 1995) make a distinction between supervisory support and collegia! support. In a number of longitudinal studies of burnout (in school settings, among others), lack of supervisory support was shown to be an antecedent of burnout (for a review, see Maslach & Schaufel i , 1993, p. 8). Other investigators have approached the matter in a wide-ranging way, attempting to consider the 17 whole social and political context, including the influence of governmental att itudes to education, the role of school boards and their officials, and the effect of parental involvement in the school. Because this broad approach is beyond the scope of this study, because there is considerable overlap between social support and collegial support, and because the school 's principal is seen to set the tone in the school, a model of administrative support focusing on the principal was selected for this study. The model is that of Arends (1982). Much of his work was focused on efforts at educational change and improvement in schools and colleges, but the principles he developed can be appl ied more widely. After studying the development and maintenance of educational programs and analyzing many other reports of school cl imate and the administrat ion's role in developing it, Arends did extensive interviewing of school staff members. In developing a model of administrative support, he relied on the perceptions of the interviewees as to what behaviours were and were not supportive rather than on signs of program success or on other external measures. Arends did not arrive at a concise definition of administrative support but isolated several factors that together represent the concept: 1. Using verbal statements to support programs and workers. Arends says this strategy f inds support in research in social psychology showing that ideas suggested by persons in positions of authority get a different, more attentive hearing than do ideas coming f rom other persons within the system or f rom outsiders. The principal's letting people know that he or she thinks a project or course of action is worthwhile is highly salutary. Arends describes a support ive principal in these terms: "He actively supported the project by 18 speaking out on its virtues at faculty meetings and in front of parent groups. He would explain the new program in 'glowing terms' and in ways that showed 'he was really concerned about the kids'" (Arends, 1982, p. 83). 2. Providing workers with a sense of role clarity and steadiness. Structural changes, new ideas, and different behavioural expectat ions can all be difficult to handle, particularly at a t ime of rapid and incessant change in schools. New programs "introduce noise into a system and unsettle the steadiness that comes to be expected" (Arends, 1982, p. 83). Role clarity means that the principal dist inguishes the work roles of teachers, counsellors, administrators and others f rom one another so as to reduce uncertainty and increase confidence in all staff members. Steadiness in this context means keeping projects and procedures on track by meeting with staff for such reasons as reviewing budgets and progress, developing objectives and t imelines, and setting policy. In sum, Arends found that supportive administrators smoothe the way for school workers to clarify their own roles, resolve conflicts, and create consistency for themselves and for their students. 3. Defending activities f rom dissenting voices and opposit ion. Counsel lors and teachers need this kind of support in the face of an increasingly critical public. Working with children can be particularly onerous because the responsibil ity for helping them grow and learn is divided among home, school, and others. Administrators must make courageous decisions in order to avoid undermining staff. In addition, " [s jometimes this meant quelling or facing squarely resistance and hostility f rom other members of the administration or f rom faculty members" who opposed some work done in the school (Arends, 1982, p. 85). This form of support might also mean cutt ing through "red tape" for school professionals, or hiring the best qualif ied people to 19 promote the school 's goals despite compet ing interests. 4. Giving something of value to workers. Specific, valued gestures of support may involve t ime, money, status, space, recognit ion, professional growth opportunit ies, or personal affection. Two examples would be a principal's attending an in-service workshop along with his or her counsellors, and providing some extra money so counsellors could plan a retreat a imed at developing collegiality and cohesiveness. This style of support may be especially relevant for counsellors, who often f ind themselves working in isolation because, for example, there is only one counsellor in the school, or the personal problems being dealt with are so overwhelming as to make him or her withdraw from others. 5. Clearing his or her schedule to deal with workers ' projects and concerns. Arends says that this, for many school workers, is the essence of perceived administrative support. Comments about principals such as "he left our meeting early" and "she was unwilling to come to staff development events" indicate lack of support. On the other hand, principals who supervise activities personally rather than delegate them to someone else are likely to be seen as supportive. Generosi ty of t ime and attention provides staff members with a good example of help-giving behaviour. Such gestures of support "not only satisfy socio-emotional needs of faculty and staff, but also enhance the status and influence of the recipient in the eyes of other organizational members" (Arends, 1982, p. 87). According to Arends, understanding perceived administrative support is important in two ways. First, administrators themselves can benefit f rom it. Learning what teachers, counsellors and others see as support ive is a first step in developing 20 ways of interacting more effectively with faculty. This is particularly helpful for beginning administrators, who might not otherwise attend carefully to seemingly minor matters such as room allocation and "pats on the back". Additionally, principals can use their knowledge of administrative support to withhold supportive gestures f rom programs or f rom staff actions that principals see as detrimental to the school. Second, an enhanced knowledge of administrative support can be applied to the development of training programs for administrators. Arends notes that such programs have traditionally focused either on administrators as educational leaders (performing such tasks as inservice training, instructional consultation, and curriculum development) or on their role as simple managers or functionaries working out the details of daily life in schools. Arends believes that both roles are important and that a greater knowledge of support ive behaviour can enhance administrators' performance in both of them. Article Review No major article focusing on burnout among high school counsel lors was found in the research literature, but there are many reports each dealing with two or more of the variables explored in the present study. These articles, which deal with burnout or related phenomena among teachers, college counsellors, psychologists, social workers, and counsellors in other settings, suggest aspects of the problem worth investigating. Al though the research of Blase (1982) focused on burnout among teachers, not counsellors, it highlighted a path toward burnout that may also be taken by counsellors. Many stressful elements are common to the work of 21 counsellors and teachers: for example, student indifference, disciplinary problems, unsupport ive or critical parents, f inancial pressures on education, overcrowding, unproductive paperwork, and inadequate administrative support. Blase developed a social-psychological grounded theory of teacher stress and burnout. He found that a lack of social, administrative and other support in dealing with job-related stressors may initially incite teachers to "go it alone" and attempt to make up for the deficits by increasing their involvement with students. However, in the continuing absence of valued and satisfying outcomes, teachers may disengage somewhat from their students. As teachers "give less" to their work, the likelihood of future ineffective and unsatisfying encounters with students increases. Thus, despite what the teacher may see as a protective lowering of involvement, a degenerat ive cycle leads to burnout (Blase, 1982, p. 109). Blase's theory is based on extensive qualitative research and seems well developed, but he does not report the data in a way that allows the reader to see how the theory derived from it. Nagy and Davis (1985) studied teachers in an unidentified urban communi ty in the northwest U.S. They found strong environmental influences on burnout, such as the specific grade or subject taught. There was no evidence of a relationship between years of experience and burnout. Type A personality and high work-orientat ion predicted burnout. The authors emphasize the importance of the person-environment fit, but are unable to draw strong conclusions about how a better fit can be developed. They considered some school climate factors but not administrative or social support; this omission is startling in view of the widely acknowledged signif icance of these inf luences. Russell, Altmaier and Van Velzen (1987) led a wide-scale study of 22 burnout among teachers in the state of Iowa using the Maslach Burnout Inventory and measures of social and administrative support. Support f rom a supervisor, reassurance of worth, and reliable alliance (the sense that others can be counted on for help) were negatively related to burnout. "Teachers who indicated that other people respected their skills and abilities reported less emotional exhaustion, more positive attitudes toward students, and greater personal accomplishment" (p. 272). This suggests that acknowledging an employee's skills and offering assistance are important interventions for principals and others to make. Russell et al. say their results support the suggest ion that intervention programs against burnout should be aimed at supervisors, who should be taught to reassure teachers of their worth. The researchers also found that fr iendship and family relationships also play a significant role in forestall ing burnout, especially the depersonalization type. As to personal and job characteristics, male teachers reported higher levels of depersonalization than did females; younger teachers reported greater emotional exhaust ion than did older ones; married teachers reported greater feel ings of personal accomplishment. Ross, Altmaier and Russell (1989) surveyed 169 doctoral-level counsellors who were working at least half-time in U.S. university counsell ing centres. Ross et al. note that earlier research on people in general had found a positive impact of social support on health; however, they state that this connection called out for further investigation in the case of counsellors, particularly in view of conflicting evidence about the buffering hypothesis, i.e. the idea that the Stress X Social Support interaction predicts physical and mental health. Ross et al. gathered data on personal and job-sett ing characteristics, burnout, stressful job-related incidents, and social support (with 23 supervisory support being a subcategory of this last variable). The researchers produced, by multiple regression procedures, three sets of statistical analyses: (a) descriptive information on job-related stress and burnout; (b) prediction of job-related stress and burnout by personal and job-sett ing characterist ics; and (c) prediction of job-related stress and burnout by social support (with supervisory support as a subcategory). Among the more significant f indings are the fol lowing: 1. Counsel lors showed moderate levels of burnout on two scales of the Maslach Burnout Inventory (depersonalization and personal accompl ishment) and low levels of burnout on the third scale, emotional exhaust ion; 2. counsellors who had fewer years of post-doctoral experience reported a greater number of stressful events and higher emotional exhaust ion; 3. supervisory support was significantly related to burnout on all three Maslach scales; 4. the predicted relationship of social support to burnout was absent; 5. supervising another counsellor was positively related to emotional exhaust ion; 6. married counsellors reported greater emotional exhaustion than unmarr ied ones. On the last two points, Ross et al. speculate that playing multiple roles may be related to greater burnout. As to supervisory support, they note that the strong relationship found between it and the three classes of burnout is in accordance with many other studies. Overall , the results suggest the usefulness of the three categories of burnout provided by the Maslach Burnout Inventory. The researchers did not study counsellors who have a lower level of educat ion than the doctorate, and they state that studying less-educated and/or less-experienced counsellors might provide very different results. In its thoroughness, wel l- formed methodological structure, and careful analysis, this study is indispensable to the literature on burnout in the helping professions. A study of counsellors at junior colleges in New South Wales, Austral ia, 24 found that longer employment in the colleges predicted a higher degree of burnout; this f inding was "consistent with the proposition that burnout results f rom an accumulat ion of stress associated with interpersonal professional con tac t . . . rather than that it occurs early in this kind of employment and then declines over t ime" (Jupp & Shaul, 1991, p. 163). Female counsellors reported somewhat more support than did males f rom family and fr iends and from the bureaucracy, but women and men registered similar burnout levels. The researchers suggest that men got their support needs met in other ways, but these are not specif ied. Although the authors state that job security may be a crucial element in the study of counsellor burnout, they did not measure this variable or its influence. In her study of psychologists in Ontario, Kahili (1986) found a strong relationship between low degrees of perceived social support and burnout, but made tentative conclusions only: "It may be that the harmful effects of stress are reduced by the presence of a supportive social network. It is also possible that individuals who are burned out perceive the world more negatively and thus do not see or seek out available social support or perhaps they drive away supporters because they are negative or make excessive demands for support" (p. 1048). A study of social workers in a community mental health clinic in the U.S. (Beemsterboer & Baum, 1984) found that a group of highly experienced workers appeared immune to any of the effects of burnout although there was a high turnover rate of employees. This suggests experience as an important factor. However, the research is reported without enough evidence of scientific rigour to permit f irm conclusions. One is left to speculate whether experience, age, or another factor might be the most relevant one. 25 Investigating social support and burnout among counsellors at a residential centre for emotionally disturbed children and adolescents, Kruger, Botman and Goodenow (1991) found that high levels of co-worker support (team cohesion and perceived quality of fr iendships) were positively related to higher personal accompl ishment scores (an indication of low burnout) on the Maslach Burnout Inventory. Emotional exhaustion decreased as staff gained more experience. However, women counsellors exper ienced higher levels of emotional exhaustion than did the men. No significant relationship was found between supervisor support and burnout; the authors suggest this is because the supervisors were all hired f rom that type of environment and thus could be expected to be especially knowledgeable and support ive (Kruger et al. , 1991 , p. 347) . One influential study of coping strategies (Shinn, Rosario, Morch, & Chestnut, 1984) examined the relationship of burnout in group workers to such elements as focusing on family and fr iends, building one's competence, taking more breaks from work, and improving communicat ion within the agency. A multiple regression analysis showed that individual and agency-level coping mechanisms were not related to burnout. However, social support activities were closely related to burnout; a plethora of other studies conf irmed this latter conclusion. Sowa, May and Niles (1994) surveyed counsellors working in various settings in the state of Virginia. They discovered that counsellors reporting high occupational stress had lower levels of recreation, self-care, and social support. They recommended that counsellor education programs include stress-management courses providing training in physical and psychological coping skills. However, Sowa et al. did not propose content for such courses, and they 26 did not offer support for their contention that such teaching would have the desired effects. In an investigation of school counsellors in Maine, Sutton and Fall (1995) considered the influence of school cl imate on counsellors' self-efficacy — the belief that one has certain skills and knowledge, as well as the ability to take action and to succeed amid the stresses of life. They found that supportive staff and administrators predicted both high eff icacy expectancy and high outcome expectancy (i.e. the possibility of good results) among counsellors. As to administrators, " [Pjerhaps when principals act as support ive resources for the exchange of ideas and material resources, counselors gain a sense of empowerment and efficacy" and the school 's counsell ing program is enhanced (Sutton et al. , 1995, p. 335). In addition, the level of non-counsell ing-related activities performed by the counsellors predicted outcome expectancy; counsellors who performed duties outside the area of their specialty and training had lower expectancy for the outcome of their counsell ing-related behaviour. Sutton and Fall assert the importance of considering environmental variables, such as size of school and education level of counsellors, in future research. Sutton and Fall did not consider burnout, but one may speculate whether lack of self-efficacy might lead to burnout or at least be related to it. 27 Chapter III Hypotheses and Methodology This chapter presents the study's hypotheses and methods, the latter including a description of the research design, population and sample, data-gathering procedures, instruments, and data analysis, organization and interpretation. Hypotheses The major goal of this study was to identify relationships between several variables and burnout. Because little or no research on high school counsellor burnout had previously been undertaken, a great variety of relationships among variables targeted were considered in the analysis. However, the major questions are simply stated: What is the relationship of perceived social support to burnout among high school counsellors? What is the relationship of perceived administrative support to burnout among these counsel lors? On theoretical grounds, an inverse relationship was expected as an answer to both these questions; in other words, stronger perceived social and administrative support would be associated with lower burnout levels. The next step was to extract, f rom the literature, some characteristics of the counsellors' personal life and job situation, and to see whether these variables added anything to the expected relationship between social and/or administrative support and burnout. Some variables worth exploring were raised in the article 28 review in the previous chapter. Further ideas about variables worth investigating are outl ined here, along with a brief explanation of why the questions suggested themselves. Each of the three variables discussed here was found three t imes in the literature — in each case, either confirmed as an important element of burnout studies or mentioned as a variable that called out for further investigation. In a study of teacher burnout, Schwab and Iwanicki (1982) found that role conflict and role ambiguity explained a statistically significant amount of variance in two aspects of burnout measured by the Maslach Burnout Inventory, namely depersonalization and emotional exhaust ion. They found that teachers whose role in the school was ambiguous or mixed (for example, a mixed teacher/counsellor or teacher/administrator role) were more subject to burnout. Sutton and Fall (1995) found that school counsellors who performed roles outside their specialty and training as counsellors had lower outcome expectancy for their counsell ing-related behaviour. Therefore, the present study asked counsellors for information about their role in the school. The question: Does the percentage of work t ime spent on counsell ing predict burnout? There is evidence of gender-related differences in both the need for social support and the forms of social support that are most helpful. Greenglass (1993) raised a number of pertinent gender-dif ference issues, including whether families alleviate work-related strain for working women as much as they do for working men. The quest ion: Does gender predict burnout? Studies including that of Jupp and Shaul (1991), cited in the article review in Chapter II, point to the significance of length of t ime employed as a counsellor. The quest ion: Does counsell ing experience predict burnout? In addition to these three variables, three further predictors were chosen 29 as exploratory variables. Although not well established in the literature, they all appear there as potentially fruitful areas of inquiry. They are the counsellors' age, the level of education they have obtained, and the size of their schools. The quest ion: Are age, education level and school size related to burnout? This study also considered some ancillary issues — relationships between non-burnout variables. As this study covers new ground, it was considered worthwhile to examine relationships between certain pairs of independent variables. These selected relationships were chosen before the data was entered; they were identified partly by mentions in the literature, and partly by the researcher and his supervisors, as being of potential interest. For example, one issue is whether male and female counsellors feel different degrees of administrative support. Heaney (1993) studied how perceived control at work affects men's and women's psychological health. Her results were ambiguous but suggested there may be a significant gender difference in how counsellors perceive administrative support. An additional factor is that the vast majority of high school principals are men, and therefore the support they offer may be more easily seen or accepted by one gender than the other. The quest ion: Do male and female counsellors perceive different levels of administrative support? Also chosen for analysis were: — The relationships of administrative support with social support, with school size and with percentage of t ime counsell ing. — The relationships of social support with gender and with age. — The relationship between school size and percentage of t ime counsel l ing. 30 Methodology Design This is a cross-sectional, non-experimental, descriptive field study. Based as it is on correlational analysis, it does not permit conclusions to be drawn about causation. However, it al lows us to examine associat ions found among variables and to consider the meaning of these relationships. The objective was to develop a picture of elements related to burnout in high school counsellors in British Columbia. To this end, a diversified sample of such counsellors was surveyed for information on their perceived social and administrative support, gender, age, counsell ing experience, education, school size, percentage of work t ime designated for counsell ing, and degree of burnout. Data Collection and Procedures Population and Sample: The precise number of high school counsellors in British Columbia at any given t ime is unknown. The number varies from school year to school year and is est imated to be several hundred. The British Columbia School Counsel lors Associat ion, the professional associat ion representing both elementary and high school counsellors, has a membership ranging roughly f rom 400 to 750, depending on the season, on membership renewal dates, and on when counsellors renew their memberships (often they do so at the time of conferences). The majority of members counsel in high schools. Some school counsellors are not members of the associat ion, while some individuals who are members may not be counsell ing in any given 31 school year. The sample of 265 potential participants was drawn from the membership rolls of the British Columbia School Counsel lors Associat ion. All current members listed as working in high schools (about 200 counsellors) were selected. The remaining 65 or so participants were high school counsellors who were still on the membership lists although they had not renewed their membership when it expired during the preceding five months (since September 1997); these participants were deemed likely to be counsell ing still. A large initial sample was selected to ensure adequate returns and to improve the likelihood of obtaining a representative sample. Procedures: This study's proposed procedures were approved by the Behavioural Research Ethics Board of the University of British Columbia. The 265 potential participants were sent a mailing including the fol lowing: a covering letter explaining in general terms the purpose of the study and including information about anonymity and confidentiality, a four-part questionnaire, and a s tamped return envelope addressed to the researcher in his department. His phone number and e-mail address were provided in case participants had questions. The questionnaires were given no identifying number or mark; it was believed that complete anonymity would encourage some counsellors to respond to the survey who might not if there was an identifier. No reminder was sent, and there was no other fol low-up with part icipants. Counsel lors who were sent the survey but did not wish to complete it were invited, in the covering letter, to hand the questionnaire to a col league or acquaintance who was counsell ing in a B.C. high school, regardless of whether 32 that person was a member of the B.C. School Counsel lors Associat ion. In light of the assurance of confidentiality and anonymity, and because the questionnaire was relatively brief and worded straightforwardly, an adequate response rate was anticipated by the researcher. Measures: The survey instrument, which is tit led High School Counsellor Questionnaire, is included as Appendix C. The questionnaire was designed so as to take roughly 20 minutes to complete. It consists of four parts: I. Counsellor and Job Characterist ics: Participants are asked for the fol lowing information: a. Age b. Gender c. Years of experience as a part-t ime and/or full-t ime counsellor d. Highest degree they have obtained (in any field of study) e. The student population of their school f. Whether they counsel only, or have other duties such as administration or teaching; if they have duties other than counsell ing, they indicate the percentage of their t ime that is designated for counsell ing II. Social Support Appraisals Scale: This measure of perceived social support (Vaux, Phillips, Holly, Thomson, Wil l iams, & Stewart, 1986) encompasses the major components of models of perceived social support developed by Cobb (1976) and others. It is designed to gauge the extent to which a person believes he or she is loved by, esteemed by, and involved with family, fr iends and others. 33 Participants respond to 23 support-related items on a four-point scale indicating how strongly they agree or disagree with each statement. Five items are reverse-scored. Scoring is accompl ished by adding up the scores for a total score, with higher scores indicating a stronger subjective appraisal of social support. Eight i tems relate to support f rom family, seven i tems relate to support from friends, and the remaining eight i tems relate to support f rom others or f rom people in general. There is considerable evidence of the reliability of the Social Support Appraisals Scale. The instrument was tested with 10 samples totalling almost 1,000 respondents f rom a variety of populations. Internal consistency (coefficient alpha) reported for the scale ranged from .80 to .90 (Vaux et al., 1986, p. 206). No data on stability were reported. The Social Support Appraisals Scale was subjected to intensive evaluation of its validity. It showed very good concurrent, predictive, known-groups, and construct validity. Its scores have been reported as being significantly correlated, in predicted ways, to other measures of social support and of psychological well-being. Such measures included social network satisfaction, family environment, depression, positive affect, negative affect, life satisfaction, loneliness, happiness, and the Symptom Check List 90 (SCL-90). These correlations "are as strong or stronger than those reported for other support appraisal measures" (Vaux et al., 1986, p. 216). The simplicity and brevity of the scale added to its appeal for use in the present study. Furthermore, respondents "appear interested and motivated, and missing data are rare" (Vaux et al., 1986, p. 216). 34 III. Administrative Support Measure: In the absence of any published or unpubl ished instrument measuring perceived administrative support, a 10-item questionnaire was devised for the purposes of the present research. This new measure is intended to be consonant with the notion of administrative support as conceptual ized by Arends (1982), and is based on the five factors isolated by him. There are two questions per factor. Participants reply on a f ive-point scale indicating their perception of how often their principal behaves in certain ways. Scores on the 10 items are added up to make a total score, with higher scores indicating a higher subjective appraisal of administrative support. Al though no reliability or validity data are available, the Administrat ive Support Measure was piloted with high school counsellors who commented on the clarity and comprehensiveness of the statements and on how the instrument might be made easier to complete. The measure was revised in accordance with their remarks, with academic advice, and with its original orientation to Arends's work. Scores during the final stage of piloting corresponded strongly and positively with the respondents' answer to an extra item administered separately as a check: "I am satisfied with the support I get from my principal." However, no correlation was calculated. If this measure were to be used again, internal consistency could be determined using Cronbach's coefficient alpha. IV. Maslach Burnout Inventory (MBI): This study employed the Second Edition (1986) of the Human Services Survey form of the MBI. This form is identical to the other one published, the Educators Survey, with one exception: The human services form uses the term recipient to refer to clients, whereas the educators form uses the term student. 35 Despite the cumbersome nature of the word recipient, the human services form seems preferable because many counsel lors perceive young people at school primarily as clients, though they are obviously students too. The inventory consists of 22 items to which participants respond on a 0-6 scale indicating the frequency of their experiencing job-related feelings. There is no total score, but scores are obtained for the inventory's three scales, which are designed to measure three domains of response to stresses: emotional exhaust ion (9 items), feelings of being emotional ly drained; depersonal izat ion (5 items), negative or impersonal att itudes to clients; and personal accompl ishment (8 items), feel ings of ability and achievement in working with people. Higher scores on the first two scales, and lower scores on the third scale, reflect greater levels of burnout. The MBI is the dominant instrument in its field, and commended itself to use in this study for several reasons. First, the MBI operationalizes burnout as defined by Maslach. She designed the MBI on the basis of her own pioneering conceptual ization and definition of burnout. As her concept of burnout is the one employed in this study, the instrument is ideally suited to measure the construct. Second, the MBI's format is simple, its explanations and items are clear, and it takes only about 10 minutes to complete, including t ime to read instructions (Maslach and Jackson, 1986). The MBI was designed specifically to measure human service workers, and it is worded to reflect their vocabulary. These qualities make the instrument well suited for individual complet ion by counsel lors. Third, the division of scores into three scales permits a more detailed analysis of the specific aspects of burnout that may be in play. Analysis might 36 show, for example, that counsell ing experience is related strongly to emotional exhaust ion, weakly to depersonalization and not at all to personal accomplishment. Such results may be suggestive to later researchers or to persons devising interventions to combat burnout. Fourth, the MBI has been used so frequently in burnout research that it forms a touchstone in the field. The inventory's wide acceptabil ity al lows us to make useful comparisons among research studies, including this one, without concern about variations across instruments. The MBI has been shown to be a reliable and valid measure of burnout. The MBI's reliability, especially for research purposes, has been well establ ished (Sandoval, 1989). Reliability coefficients for the subscales ranged from .71 to .90 (N = 1,316) in development trials. In smaller studies, test-retest reliability coefficients ranged from .60 to .82 after two to four weeks and from .54 to .60 after two years (Hargrove, 1989, p. 474). The emotional exhaustion scale tends to show higher reliability than the other two subscales. In one set of development studies, the alpha coefficient for emotional exhaust ion was .90, and the stability co-efficient over a two-to-four-week period was .82. The alpha coefficient of the depersonalization scale was .79, and the stability was .60. In the personal accompl ishment scale, the alpha was .71 and stability was .80 (Sandoval, 1989, p. 475). Personal accompl ishment correlates weakly with depersonal izat ion (.26), and weakly with emotional exhaust ion (.22). Depersonalization and emotional exhaustion are moderately intercorrelated (.52) (Maslach & Jackson, 1981). Considerable evidence endorses the validity of MBI scores. In studies of convergent validity, "the MBI scores have been correlated with behavior ratings 37 made by knowledgeable informants, with job characterist ics that are expected to contribute to burnout, and with other measures of outcome related to burnout" (Sandoval, 1989, p. 475). The MBI 's discriminant validity is supported by evidence that measures of job satisfaction yield results markedly different from the MBI ; in other words, burnout and job dissatisfaction are not the same thing. They are related but are distinct enough to be seen as separate constructs. Additionally, the MBI has been shown not to be influenced by social desirability considerat ions and not to measure clinical depression. The inventory has been shown to be flexible and to be applicable over a broad geographical and occupational range. "The MBI is certainly the instrument of choice to use in research and evaluation endeavors studying the phenomenon of burnout" (Sandoval, 1989, p. 476). Data Analysis, Organization and Interpretation: T h e questionnaires were scored by the researcher. Scores were entered first into a data entry sheet, then into the computer statistical program SPSS 7.5 for Windows. The data was analyzed to produce results, including f igures and tables, describing the data and showing relationships between the predictor or independent variables (perceived social support, perceived administrat ive support, and the six counsellor and job characteristics), and the criterion or dependent variables (the three burnout scales). First, to get an overview of the data, f requency distributions of all variables were examined. Means, standard deviations, and ranges were found. Correlations among all variables were calculated, with signif icance at the .05 and .01 levels f lagged. 38 Then a series of six bar charts was produced to illustrate the distribution of the six counsellor and job characteristics of the participants. Next, t -tests were employed to compare means of participants' scores on the three MBI scales with the most recent normative data provided by the publishers of the MBI for mental health workers. Signif icance at the .05, .01 and .001 levels was noted. Stepwise multiple regression procedures were performed with each of the three burnout scores as the dependent variable. In each of the three analyses, the five theory-grounded independent variables — social support, administrative support, gender, experience, and percentage of t ime counsell ing — were entered in the order of the strength of their correlation with the dependent variable, from highest to lowest. Independent variables that added nothing to R2 were removed. Next, the effects of the three exploratory variables — age, education, and school size — on burnout were examined. The correlation of age with each of the three classes of burnout was considered. The influence of education on each of the three burnout categories was examined by a one-way ANOVA, as was the impact of school size. Finally, a series of tests investigated the relationships among certain independent variables, as listed in the Hypotheses section earlier: Administrative support 's relationship with social support was found by correlation, as was administrative support 's relationship with percentage of t ime counsell ing. Gender differences in administrative support were explored by a t-test. Differences in administrative support in schools of varying sizes were examined by one-way ANOVA. Gender differences in social support were explored by a t -test, and 39 social support 's relationship with age was found by correlation. Differences in percentage of t ime counsell ing, having regard to school size, were investigated by one-way ANOVA. 40 Chapter IV Results This chapter comprises: a profile of the participants, including graphs of their personal and job characterist ics; an overview of the data, with all variables' means, standard deviations, and correlations; a compar ison of participants' burnout levels with norms; and results of the statistical tests performed on the data to examine the relationships among variables. Participants Of the 265 surveys mailed out to high school counsellors, 163 were answered and returned, for a return rate of 6 1 . 5 1 % . Six questionnaires could not be used, leaving an N of 157. These 157 surveys were complete except that seven lacked information on years of counsell ing experience, and three lacked an answer on the percentage of t ime designated for counsell ing. Details of returns and the scoring of questionnaires are in Appendix A. The job and counsellor characterist ics of the participants are displayed in Figures 1-6; percentages, rather than numbers, of counsellors in the sample are used. Since no study of burnout among high school counsellors was found in the literature, this sample cannot be compared with any other. As Figure 1 indicates, the vast majority of counsellors were aged in their 40s or early 50s. The mean age was 46 .31 ; the standard deviation was 7.05 years. The youngest counsellor was 26, the oldest 60, for a range of 34 years. 41 Figure 1 Participants by Age 40 30 20-^ 10 c o o 1 0- o 26-30 3K35 36^40 41-45 46-50 51-55 56-60 A g e I n t e r v a l Figure 2 Participants by Gender female male G e n d e r 42 Figure 2 shows the breakdown by gender. The sample of counsel lors was 6 3 . 1 % female, 36 .9% male. This imbalance is in line with the proportion of women and men listed as high school counsel lors on the British Columbia School Counsel lors Associat ion membership rolls, f rom which the sample was taken. The participants' years of experience as full-t ime or part-t ime counsellors are displayed in Figure 3. Experience varied from one year to 32 years, for a range of 31 years. The mean was 12.55 years, and the standard deviation was 7.50 years. Figure 3 Participants bv Experience Missing 1-8 Y e a r s of E x p e r i e n c e 17-24 25-32 The counsellors' highest degree obtained is shown in Figure 4. The master's degree is the highest one attained by 66 .9% of the counsellors. 43 Figure 4 80 T —— 70 60-50-40 30 20 8 1 0 Q_ 0 bachelor's graduate diploma master's E d u c a t i o n doctorate Figure 5 40-Participants by School Size 30 20 10 c 0 o CD Q . I I | | | ™ B i l j l l l^^H ^^gH||[l ^^ ^^ ^^ ^^  • • l l n less than 500 1000-1499 2000 or more 500-999 1500-1999 S c h o o l S i z e 44 As Figure 5 illustrates, the great majority of high school counsellors who were surveyed work in medium-sized schools whose student population is between 500 and 1499. Figure 6 displays the percentage of counsellors' work t ime that was designated for counsell ing rather than for other duties such as teaching or administration. Half the participants (50.3%) were full-t ime counsellors, while another 2 2 . 1 % were counsell ing close to full-t ime (73-90% of their work t ime was designated for counsell ing). The mean was 80.66%, and the standard deviation was 25.49%. The lowest percentage designated for counsell ing was 5%, the highest 100%, for a range of 95%. Figure 6 Participants bv Percentage of Time Counselling 60 50 40 30 Missing 5-21% 22-38% 39-55% 56-72% 73-90% 100% % of Time Counselling 45 Overview of the Data Only the relationships mentioned in the Hypotheses section above were chosen for analysis in this study. Other relationships, found incidentally, were not analyzed, in light of the danger that such relationships might simply result from peculiarities of the sample. However, as this study appears to be the first into burnout among high school counsellors, the researcher felt it important to offer a gl impse of all relationships, for the potential benefit of later investigators. Therefore, a full correlational matrix, including means and standard deviations, is presented as Table 1. Levels of Burnout Scores on the three burnout scales were compared with normative data in two ways: by degree of burnout, and by comparing the means of the current and normative samples. As to degree of burnout, the Maslach Burnout Inventory (MBI) divides scores on each subscale into three categories: The lowest one-third of the normative sample was assigned to the low category, the next one-third to the moderate category, and the uppermost third to the high category. For emotional exhaust ion, the current sample of high school counsellors fell in the moderate range, but at the top end of that range. For depersonalization, the present sample fell in the low range, but at the high end of that range. Finally, for personal accomplishment, the counsellors scored in the middle of the low range. The second method of interpretation is by comparing means. Means, 46 * z CO o C Q ' CD 3 ' o ^ 1 Q> II S 3 <*> m b m O l || CD" m r—t-ro o' CD m a. X Q) C }§-. co§ CQ ' 13 Eft O O TJ « II -* a 0J CD ~ -a = f CD CD CO O O 13 m N ' < S ® 5 ' -~s Z5 ro • i r-h 0) TJ cfT > • S II TJ CD CO O DJ. > O O o 3 "o co 3 CD 3 TJ > O CJl o CO o 4^ o 4^ ro ro * ro CO ro CO I ro 4^ co 4^ ro a TJ i b ro o o -j o ro cn * I b co m rn o co I b co o o o co o CD .1 o •a 3 T3 = • O —^ < CD O CO CO CO c o -a Q. o — o 4*. o ro o co o cn ro CO CO CO cn CO CO o ro co o co ro ro " |co co O o c 13 CO CD 5' C O CO cn ro ro * o cn cn CO o =r o o CO N " CD ro cn * ro o o co m a. c: o 0) CO m X T3 CD CD O CD ro Q CD Q . CD > C Q CD < —"I 0)' CT CD ro o ro o co cn ->• 45v 4 ^ co ro co CO cn CO CO C D CO —L C D CO C D CO 4^ CO L o ro ro ro —1 CO CO cn cn co C D ro 4S- cn 4*. CO co fi) CT CD 2 CD &> 3 0) CO ST a D CD < fi)' 5" 3 (0 fi) 3 a O o 3 &> •-ti o" 3 (0 fi) 3 o 3 Q < fi) 2. fi)' cr CD (0 4^ CO CO CO CO CO CD CO CO b - s i CO CO ro cn 4^ CD C D CO cn cn o - s i b cn CO O standard deviations and results of t -tests comparing the current sample with the MBI norms are presented in Table 2. Table 2 Maslach Burnout Inventory (MBI) Means and Standard Deviations Present Sample MBI Normative High School Mental Health Counsellors Sample M (N = SD 157) M (N = SD 730) t P Emot ional Exhaustion 19.96 8.96 16.89 8.90 4.26 < 0 0 1 Depersonal izat ion 4.17 3.67 5.72 4.62 -5.34 < 0 0 1 Personal Accompl ishment * 42.43 4.66 30.87 6.37 31 .24 < 0 0 1 * High scores reflect less burnout As measured by comparing means, the present sample had markedly different levels of burnout from the MBI norm for mental health professionals. The counsellors reported significantly higher ( f =4.26; p < 0 0 1 ) levels of emotional exhaustion than did the normative sample. However, on the other two subscales, the counsellors reported lower burnout. They scored significantly lower (f = -5 .34 ; p < . 0 0 1 ) on depersonalization, and significantly higher (t =31.24; p < .001) on personal accompl ishment — the latter indicating low burnout, as it is interpreted in the opposite direction from the other two scales. 48 The normative sample included counsellors f rom school and other settings, psychologists, psychiatrists and psychotherapists. The suitability of these norms for the present study will be discussed in Chapter V. Multivariate Tests Regression Five independent variables — social support, administrative support, gender, experience and percentage of work t ime designated for counsell ing — were selected, a priori, for inclusion in the multiple regression analysis of each of the three classes of burnout. These variables were chosen, following the review of the literature, because of indications that they were all factors of consequence to burnout. For each of the three classes of burnout, a stepwise multiple regression analysis was executed. In each case, the five independent variables were entered in the order of their correlation (from highest to lowest) with the burnout score. Then, variables that did not add to R2 were eliminated. The relevant correlations are listed in Table 3. 49 Table 3 Correlations between Burnout and Independent Variables Employed in the Multiple Regression Analysis Emotional Personal Exhaustion Depersonalizat ion Accompl ishment Social Support - . 1 0 - . 2 5 * . 4 1 * * Administrat ive - 1 8 * - . 0 7 .26** Suppor t Gender - . 0 8 .12 .08 Exper ience .03 .10 - . 0 4 % Counsel l ing - . 0 6 .01 .26** * p < . 0 5 ** p < . 0 1 Tables 4, 5 and 6 present the results of the multiple regression analyses of the three classes of burnout. The coefficient of determination (R2) indicates the amount of variance in the burnout class that can be predicted from one independent variable or f rom a combinat ion of several. As each step is added, R2 rises to express the new amount of variance predictable from the combinat ion. The R2 increment states the predictive improvement gained by adding each step to the equation. The F -ratio al lows us to conclude whether the change in each step is significant at p <.05, p < .01 , or p < .001 . When the five relevant variables were employed in the regression 50 analysis of emotional exhaust ion (Table 4), percentage of t ime counsell ing was excluded because it added nothing to the predictive value. Administrative support predicted a significant ( p = . 0 1 ) amount, 4.3%, of the variance in emotional exhaustion. The three remaining variables together predicted an additional 0.9% of the variance. The equation accounted, then, for 5 .2% of the variance [F (4 ,145) = 1.972; p = . 1 0 ] ; this equation did not achieve significance at p < . 0 5 . TabTeT Stepwise Multiple Regression of Independent Variables on Emotional Exhaustion Emotional Exhaust ion Step R2 R2 increment F Q 1. Administrative Support .043 .043 6.659 .01 2. Social Support .046 .003 .486 .49 3. Gender .049 .003 .431 .51 4. Experience .052 .003 .392 .53 F(4 ,145) = 1 .972 ;p= .10 The results of the regression analysis of depersonalization are shown in Table 5. Administrative support was excluded because it did not add to the predictive value. Social support, entered first, predicted a significant (p= .01) proportion, 5.2%, of the variance in depersonalization. Gender, entered second, predicted a further 2 .2% of the variance. Together, the two other 51 variables predicted 1.2% of the variance. The equation reached signif icance at p <.05 [F (4,142) = 3.344; p = 01 ] . Table 5 Stepwise Multiple Regression of Independent Variables on Depersonalization Depersonal izat ion Step R2 R2 increment F P 1. Social Support .052 .052 7.983 .01 2. Gender .074 .022 3.388 .07 3. Experience .084 .010 1.599 .21 4. % Counsell ing .086 .002 .292 .59 F ( 4 , 142) = 3.344; p = 01 Table 6 displays the results of the regression analysis of personal accomplishment. This t ime, all f ive independent variables contr ibuted something to predictive power. Social support alone predicted a significant (p < 0 0 1 ) 1 6 . 1 % of the variance in personal accomplishment. Administrative support contributed a further, significant (p =.04) 2.4%. Together, the five variables accounted for 21 .3% of the variance in personal accompl ishment; the equation was significant [F(5 ,141) =7.614; p<.001] . 52 Table 6 Stepwise Multiple Regression of Independent Variables on Personal Accomplishment Personal Accomplishment Step R2 R2 increment F a 1. Social Support .161 .161 27.846 <.001 2. Administrative Support .185 .024 4.178 .04 3. % Counsell ing .195 .010 1.814 .18 4. Gender .196 .001 .211 .65 5. Experience .213 .016 2.944 .09 F(5 ,141) = 7.614; p < . 0 0 1 Burnout and Exploratory Variables The exploratory variables — age, education and school size — were included in the survey as a result of the researcher's curiosity, academic advice, and suggestions in the literature that such factors might be important. These three independent variables were not included in the regression analysis. They were examined separately to determine whether there was a relationship with each of the three classes of burnout. Age correlated very weakly with the burnout scores. The correlation of age with emotional exhaust ion was - . 0 8 , with depersonal izat ion was - . 0 1 , and 53 Table 7 Means, Standard Deviations, and F Values of Burnout Scores bv Level of Education Var iab le M S D F Emotional Exhaust ion: Bachelor 's 20.04 9.64 Graduate Diploma 22.18 9.65 Master 's 19.29 8.63 Doctorate 26.00 * 0.92 .43 Depersonal izat ion: Bachelor 's 4.61 4.19 Graduate Diploma 4.43 3.80 Master 's 4.01 3.56 Doctorate 3.00 * 0.25 .86 Personal Accompl ishment : Bachelor 's 42.87 4.37 Graduate Diploma 41.00 5.79 Master 's 42.68 4.36 Doctorate 46.00 * 1.25 .29 * N = 1 54 Table 8 Means. Standard Deviations, and F Values of Burnout Scores by School Size Var iable M S D F Emotional Exhaust ion: Less than 500 23.12 9.85 500-999 18.88 8.83 1000-1499 20.46 8.80 1500-1999 19.26 8.39 2000 or more 18.33 11.33 0.85 Depersonal izat ion: Less than 500 4.06 3.21 500-999 4.59 3.95 1000-1499 4.24 3.54 1500-1999 2.42 2.97 2000 or more 5.33 4.84 1.44 Personal Accompl ishment : Less than 500 500-999 1000-1499 1500-1999 2000 or more 40.53 5.27 41.82 4.87 42.75 4.72 44.68 2.56 43.17 2.64 with personal accomplishment was .05. None of these correlations was significant at p<.05. Differences in burnout scores by educat ion level were examined by one-way ANOVA; the results are presented as Table 7. Varying education levels accounted for no significant differences in emotional exhaustion ( F = 0 . 9 2 ; p =.43), depersonalization (F=0 .25 ; p =.86) and personal accompl ishment ( F =1.25; p = 2 9 ) . Similarly, an examinat ion by one-way ANOVA was made of dif ferences in burnout scores by school size (Table 8). School size did not account for significant differences in emotional exhaust ion ( F = 0 . 8 5 ; p =.49), in depersonalization ( F = 1 . 4 4 ; p=.23), or in personal accompl ishment ( F = 2 . 2 3 ; p =.07). In the last case the differences approached significance at p<.05; larger school size tended to be associated with greater feelings of personal accompl ishment (an indication of lower burnout). Relationships among Independent Variables Seven pairs of independent variables were selected, a priori, to be analyzed for possible relationships. Administrat ive support was correlated significantly ( r = . 3 1 ; p <.01) with social support and with percentage of t ime counsell ing (r=27; p <.01) (Table 1). A one-way ANOVA showed no significant dif ferences in administrative support by size of school ( F = 0 . 9 3 ; p =.45) (Table 9). A t -test examined differences in administrative support by gender (Table 10). There was no significant difference (t = -0 .52 ; p =.60). Social support was correlated significantly ( r = . 1 6 ; p <.05) with age 56 Table 9 Means, Standard Deviations, and F Values of Administrative Support and Percentage Counselling by School Size Var iab le M S D F p_ Administrat ive Support : Less than 500 38.41 9.35 500-999 38.11 7.28 1000-1499 39.63 8.61 1500-1999 36.79 8.08 2000 or more 42.83 4.83 i t a g e Counsel l ing: Less than 500 60.35 27.23 500-999 78.02 25.18 1000-1499 85.80 23.76 1500-1999 85.32 23.72 2000 or more 100.00 0.00 0.93 .45 4.89 .001 (Table 1); that is, older counsellors tended to report higher perceived social support than did younger counsellors. A t-test showed no significant difference in social support by gender (t =-0 .19 ; p = 85) (Table 10). 57 Table 10 Means/Standard Deviations, and t Values of Administrative Support and Social Support bv Gender Var iab le Administrat ive Support : Female Male M 38.47 39.17 S D t 8.09 8.08 -0 .52 .60 Social Support : Female Male 81.25 81.50 7.47 8.46 -0 .19 .85 Percentage of t ime counsell ing was shown by one-way ANOVA to differ significantly ( F = 4 . 8 9 ; p =.001) by school size (Table 9) ; that is, the larger the school, the higher the percentage of t ime counsellors tended to have designated for counsell ing rather than for other duties such as teaching or administrat ion. 58 Chapter V Discussion This final chapter begins with a summary of the f indings in regard to the hypotheses, together with discussion of the results and their implications (significance is at the .05 level unless otherwise indicated). Then there are suggestions as to the scope, approaches and methods of future research on the topic. Summary and Implications Social and Administrative Support A major hypothesis of this study was that there would be an inverse relationship between each of the three classes of burnout and each of two other factors: perceived social support and perceived administrative support. All correlations between social support and burnout were in the expected direction. Those between social support and depersonal izat ion ( r = - . 2 5 ) , and between social support and personal accompl ishment ( r= .41) , reached signif icance. The correlation between social support and emotional exhaust ion ( r = - . 1 0 ) was not significant. These results generally coincide with those of several other studies (e.g. Sowa, May & Niles, 1994; Shinn, Rosario, Morch & Chestnut, 1984; Blase, 1982) that showed negative relationships between social support and burnout and/or other forms of occupational strain. Flournoy 59 (1990) did not f ind any significant relationships between social support and the three burnout classes. She explained this by noting that her instrument measured the objective availability of family members and others, rather than the perception that there are people on whom one can rely — the latter factor seemingly the most important. The results of Ross, Altmaier and Russell (1989) were mixed. They did not f ind the overall inverse relationship they had hypothesized between social support and burnout. However, using the six-category Social Provisions Scale, they found that three categories were significantly correlated in the expected direction with burnout classes. Higher levels of social integration (a sense of belonging to a group) were related to lower emotional exhaust ion and depersonalization. Reassurance of worth (acknowledgment of one's abilities by others) and guidance (advice or information) were related positively to personal accompl ishment. These results suggest that the specific type or form of social support is a relevant factor. In the case of administrative support and burnout, relationships found were also in the expected direction. The correlations between administrative support and emotional exhaust ion ( r = - . 1 8 ) , and between administrative support and personal accompl ishment ( r= .26) , attained significance. The correlation of administrative support and depersonal izat ion ( r = - . 0 7 ) was not significant. These results reinforce those of other studies. Lack of perceived supervisory support was related to higher levels of all three burnout c lasses in the college counsellors studied by Ross, Altmaier and Russell (1989); the researchers concluded that supervisory support has a positive effect on workers ' physical and mental health. Similarly, Russell, Altmaier and Van Velzen (1987), in their study of Iowa teachers, found that perceived administrative support was negatively related to the three burnout classes. 60 Sutton and Fall (1995), who studied the effects of school cl imate on counsel lors' feel ings about themselves and their work, found that support ive administrators predicted high self-eff icacy among counsellors and high expectat ions for good results in their work. The present results, in general, confirm those of earlier studies associating perceived social and administrat ive support negatively with burnout. Although causation cannot be proven, it is clear that if interventions against burnout among counsellors are created, these support variables will need to be among the factors targeted. Other Relevant Variables Three independent variables — gender, experience, and percentage of work t ime designated for counsell ing — were suggested by the literature as factors potentially relevant to burnout. The direction of such possible relationships did not emerge clearly from the published research studies. Correlations of each of the three variables with the three burnout classes were examined. Gender was not correlated significantly with any of the three burnout classes — emotional exhaust ion ( r = - . 0 8 ) , depersonalization ( r=.12) , or personal accompl ishment ( r= .08) . These nonsignificant f indings coincide with those of three other studies of counsellor burnout in educational institutions. Flournoy (1990) found no gender dif ferences in burnout among college psychotherapists in the Los Angeles metropoli tan area. Jupp and Shaul (1991) found similar burnout levels in female and male counsellors at Austral ian junior colleges. Ross, Altmaier and Russell (1989) discovered similar degrees of 61 burnout in female and male college counsellors in the U.S. T w o other studies each showed gender differences in one burnout class. Russell, Altmaier and Van Velzen (1987), in their study of teacher burnout in Iowa, found that men reported higher depersonalization than did women; the authors offered no explanation. Kruger, Botman and Goodenow (1991) found that female counsel lors of emotional ly disturbed young people reported higher emotional exhaust ion than did their male counterparts; the researchers attr ibuted this in part to the counsellors' need to physically restrain their clients at t imes, a task the women apparently considered more taxing and troubling than did the men. Overall, gender does not appear to be an important factor in burnout. A question ment ioned by Greenglass (1993) is whether family relationships ease job-related strain for working women as much as for working men. Gathering data on familial connect ions and responsibil it ies would permit a more complex analysis of the gender issue. No significant correlation was found between counsel l ing experience and any of the three burnout classes. Experience was weakly correlated with emotional exhaust ion ( r= .03) , depersonal izat ion ( r= .10) , and personal accompl ishment ( r = - . 0 4 ) . Likewise, in their study of teacher burnout, Nagy and Davis (1985) found no relationship between experience and burnout. However, a study of counsellors in Australian colleges found that longer employment there was associated with higher burnout (Jupp and Shaul, 1991). The authors said this suggested that burnout results f rom an accumulat ion of stress rather than occurring early in employment and then declining; however, Jupp and Shaul did not gather data on overall counsell ing experience, leaving open the question whether switching to another job would alleviate burnout. T w o other studies have shown evidence of the opposite tendency: that less-experienced 62 counsel lors are generally more emotional ly exhausted than more-exper ienced ones. Kruger, Botman and Goodenow (1991) found that counsellors of emotional ly disturbed young people reported decreased emotional exhaust ion as they gained experience. Similarly, in their investigation of U.S. university counsell ing centres, Ross, Altmaier and Russell (1989) discovered that less-experienced counsellors were more prone to emotional exhaustion. They suggested that more-experienced counsel lors might be less willing to admit to psychological strain than might less-experienced ones, but also that the latter lacked some of the coping mechanisms required to handle job-related stress. It is likely that, as stresses are accumulat ing over years of experience, counsellors gradually put more emphasis on developing the support and the coping skills that will moderate the damage. If they are unsuccessful at this, they may leave counsell ing; there is a survivor bias in the data. A study of U.S. social workers found that a core group of highly experienced workers seemed immune to burnout although the turnover rate among their col leagues was high (Beemsterboer & Baum, 1984). Percentage of work t ime counsell ing was not correlated significantly with emotional exhaust ion ( r = - . 0 6 ) or with depersonalization ( r= .01) . However, percentage counsell ing was correlated significantly with personal accompl ishment ( r - 2 6 ) ; that is, the higher the proportion of work t ime spent counsell ing, the greater the counsellor 's feel ings of personal accompl ishment were likely to be. Apparently, feel ings of job fulf i lment are greater when the counsellor can concentrate on playing one role. In the literature there is no direct reinforcement of this result, but studies raise related issues. Sutton and Fall (1995) found that school counsellors who were assigned to duties outside their specialty area expected their counsell ing work to be less effective; the 63 authors did not study burnout, but it is reasonable to expect that lower outcome expectancy would be associated with reduced feelings of personal accomplishment. Ross, Altmaier and Russell (1989) found that counsellors who supervised a col league in addition to doing their own counsell ing work were more affected by emotional exhaustion than those who were just counsell ing. The significant positive correlation found in the present study between personal accompl ishment and percentage of t ime devoted to counsell ing suggests it would be helpful for educational administrators to minimize mixed-role assignments in schools: As much as possible, that is, counsellors will counsel, teachers will teach, and administrators will administer. Counsellors would be torn less frequently between the conflicting demands of being a teacher or administrator as well as a counsellor. Because of limited resources and the exigencies of schedul ing, this ideal division of roles would be more easily accompl ished in larger schools. Regression Gender, exper ience and percentage counsel l ing were employed, along with social support and administrative support, in a multiple regression analysis of each of the three burnout classes. The relationships, or lack thereof, between these five independent variables and burnout were discussed above. In this section there is a summary of the regression analyses, along with brief comments on the results. Emotional exhaust ion was correlated significantly with only one of the five variables, administrative support ( r = - . 1 8 ) . In the regression analysis of emotional exhaustion, administrative support predicted 4 .3% of the variance. 64 Social support, gender and experience together predicted only an addit ional 0 .9% of the variance, and percentage counsell ing was excluded as it added nothing to predictive value. The equation was not significant [F (4,145) = 1.972; p = . 1 0 ] . Clearly, factors more relevant to emotional exhaustion need to be identified and evaluated. In their meta-analysis of research on burnout in social work, Soderfeldt, Soderfeldt and Warg (1995) found that emotional exhaust ion had been shown in various studies to be associated with boredom, low job satisfaction, intent to change jobs, and role conflict. None of these factors was considered directly in the current research; however, role conflict may be a significant factor in relation to emotional exhaustion in the present sample, an issue considered below under Levels of Burnout. Of the five independent variables, only social support was correlated significantly with depersonal izat ion. When depersonal izat ion was analyzed by multiple regression, social support predicted 5.2% of the variance. Gender, entered next, accounted for an additional 2 .2% of the variance. As a group, experience and percentage counsell ing added 1.2% to the prediction, while administrative support was excluded. The equation was significant [F(4 ,142) = 3.344; p =.01]. However, it accounted for just 8.6% of the variance. Accordingly, there must be other, more pertinent factors in predicting depersonalization. In the Soderfeldt et al. (1995) meta-analysis of burnout in social workers, factors shown to be associated with depersonalization were value conflict, high education, low age, and lack of supervisory support. The first of these variables was not measured in the present study, and the other three did not prove to be associated with depersonal izat ion. As for personal accomplishment, it was correlated significantly with social support, administrative support, and percentage counsell ing. In the regression 65 analysis, all f ive factors contr ibuted to predict personal accomplishment. Social support alone predicted 1 6 . 1 % of the variance, while administrative support accounted for a further, significant amount, 2.4%, of the variance. The five variables as a group predicted 21 .3% of the variance in personal accomplishment. The equation was significant [F(5 ,141) = 7.614; p < . 0 0 1 ] . Social support, administrative support and percentage counsell ing are undoubtedly related in an important way to personal accompl ishment in the present sample. None of these variables was found in the Soderfeldt et al. (1995) meta-analysis, but it listed other factors shown to be associated negatively with personal accompl ishment: Type A personality, public-sector work, bad working conditions, unfair promotional policies, role ambiguity, and lack of team support. The last two variables may be related to aspects of the present study. In a way, role ambiguity may parallel percentage counsell ing; a lower percentage of counsell ing duties implies the presence of teaching or administrative duties, and this combinat ion of roles might create ambiguity. Lack of team support may be associated with lack of administrative support, as the principal is customari ly expected to develop cohesion among professional staff. In summat ion, the five variables entered into the regression analyses predicted, as a group, only modest proportions of the variance in the three burnout classes. One factor of note in each analysis is multicollinearity; the significant correlation ( r = . 3 1 ; p < . 0 1 ) between social support and administrative support meant that the two factors overlapped in explaining variance. If the exploratory variables had been entered, they would likely have accounted for little additional variance, for they were not significantly associated with burnout. Other factors that might be better predictors of burnout are discussed below in the sections Further Discussion and Recommendat ions for Future Research. 66 Exploratory Variables Three further independent variables — age, educat ion and school size — were added to the study as exploratory factors. Age was not correlated significantly with any of the burnout classes: emotional exhaustion (r=-.08), depersonal izat ion (/-=-.01), and personal accompl ishment ( r= .05) . These results contrast in part with those of Flournoy (1990), who surveyed psychotherapists employed in college counsell ing centres in the Los Angeles area. She found a significant negative relationship between age and emotional exhaustion, and attributed it to two factors: frustrated idealism among younger professionals, and older workers ' more efficient use of t ime and energy. Flournoy found no significant correlation between age and depersonalization, or age and personal accompl ishment. A study of teacher burnout in Iowa found a similar pattern. Younger teachers reported higher emotional exhaust ion than did their older col leagues, but there was no significant relationship between age and either of the other two burnout classes (Russell, Altmaier & Van Velzen, 1987). The lack of a significant f inding in the present study can perhaps be understood in light of the fact that the vast majority of participants were of mature years; only 7 .7% of the counsellors were younger than age 36. This suggests that maturity is an important criterion in deciding which teachers are assigned to counsell ing positions. It is reasonable to suppose the younger professionals were given counsell ing duties because, among other reasons, they displayed the maturity essential to dealing with the job's pressures in a way that would forestall exhaust ion. In a report on their study of self-efficacy among elementary and high school counsellors in Maine, Sutton and Fall (1995) suggested the counsellors' 67 personal level of educat ion as a variable worth investigating. Two of the major inquiries into counsellor burnout (Flournoy, 1990; Ross, Altmaier & Russell, 1989) confined themselves to doctoral-level professionals. There is nothing else in the literature about the possible association of burnout and educat ion level. In the present research, education level did not account for significant differences in any burnout class — emotional exhaustion ( F = 0 . 9 2 ; p =.43), depersonal izat ion ( F = 0 . 2 5 ; p =.86), or personal accompl ishment ( F = 1 . 2 5 ; p =.29). It is difficult to know the meaning of this lack of differences. If it is true, as some burnout theorists say, that burnout arises from difficulties in the personal interactions of helpers with others (Maslach, 1982, p. 3), then education level may simply be irrelevant to this interplay. O n the other hand, one must consider that 67 .5% of the counsellors in the present study had at least a master 's degree, and a further 17.8% had a graduate diploma. While no information was gathered on the disciplines in which these degrees and diplomas were granted, it seems safe to state, f rom anecdotal evidence, that in most cases the fields of study would be counsell ing, counsell ing psychology, or related areas such as educational psychology. In such programs of study, counsel lors generally get information and practice in dealing effectively with others without suffering undue strain. In this light, we might have expected a higher level of education to be associated with lower burnout. Gathering information on fields of study would have permit ted further analysis of this phenomenon. School size was also recommended by Sutton and Fall (1995) as a variable worthy of interest in research on school counsellors. They did not suggest what direction the burnout-size relationship might take. In the present study, school size did not account for significant differences in emotional exhaustion ( F = 0 . 8 5 ; p =.49) or in depersonalization ( F = 1 . 4 4 ; p =.23). There 68 were no significant differences in personal accompl ishment either (F=2 .23 ; p =.07), but this result t rended toward significance at p<.05; larger school size tended to be associated with higher personal accompl ishment (a sign of lower burnout). The explanation of this tendency is not clear, but it may perhaps be found in the significant positive correlation between percentage counsell ing and personal accompl ishment ( r= .26 ; p<.01). As will be noted in the next section, counsellors in larger schools are far more likely than their counterparts in smaller schools to be counsell ing full-time, or nearly; this fact, rather than school size in itself, might account for counsellors' tendency to have stronger feelings of personal accompl ishment in larger schools. Relationships among Independent Variables Seven pairs of independent variables were chosen, a priori, to be examined for relationships. The f indings are outl ined and discussed here. As expected, the correlation between administrative support and social support was positive and significant ( r = . 3 1 ; p<.01) . The explanation has several facets. As mentioned in Chapter I, it is difficult to avoid overlap in measuring the two types of support. Somet imes counsellors consider their principals as belonging to their social networks; in fact, one participant in this study added to the questionnaire a note saying his principal was his best fr iend. Administrative support may be seen as a type of social support and is somet imes measured in that way. For example, Ross, Altmaier and Russell (1989), in studying U.S. college counsellors, appraised support network resources by dividing them into four categories: supervisor(s), co-workers, spouse, and friends/relatives. Another factor helping to explain the correlation 69 found is that both support variables were measured by perception. Counsel lors ' general orientation to the world may have produced part of the correlation. For instance, counsellors who felt lonely may have minimized, consciously or unconsciously, their feelings of being supported by principals, fr iends and family members, while optimistic counsellors may have maximized such feelings. Administrative support was also correlated positively and significantly (r =.27; p < 0 1 ) with percentage of t ime counsell ing; that is, the higher the proport ion of work t ime designated for counsell ing, the higher the counsellor 's perception of administrative support was likely to be. This result may be explained tentatively. Most counsellors who have been assigned to spend all or most of their work t ime counsell ing are likely to feel they have the confidence of the principal. Counsel lors who spend less t ime counsell ing may feel they have not entirely proven themselves to their principals, and would feel less support. The latter group may also be more vulnerable to the role conflict that affects professionals who fill two different roles (in this case, usually teaching in addit ion to counsell ing); this may influence their perception of support as well. Another factor to consider is the pair of significant correlations, discussed above, between percentage counsell ing and personal accompl ishment ( r = . 2 6 ; p < 0 1 ) and between administrative support and personal accompl ishment (also r - 2 6 ; p < . 0 1 ) . It is impossible to determine causation in the interplay of these three variables. It is reasonable to speculate that both higher administrative support and higher percentage counsell ing play a role in promoting feelings of personal accomplishment. However, it may also be true that a sense of enhanced personal accompl ishment is generated in counsellors who are assigned a high percentage of counsell ing duties, and this in turn may predispose them to look upon their principal's behaviour as being more 70 support ive. There were no significant differences in administrative support by size of school ( F = 0 . 9 3 ; p =.45). There is nothing in the literature to indicate that such differences exist. Anecdotal evidence is mixed. Counsel lors and teachers somet imes remark that smaller schools are more closely knit; accordingly, one might expect higher perceived support f rom principals there than in larger schools. O n the other hand (as will be seen below), in larger schools counsellors are more likely to spend all, or almost all, their t ime counsell ing; therefore, they might consider they have the confidence of the principal in their counsell ing role, and this belief would be reflected in their perception of administrative support. Heaney (1993), in a study of perceived control in the workplace, suggested women and men may perceive supervisory support in different ways. In the present study, gender did not account for a significant difference in perceived administrative support (t = -0 .52 ; p =.60). In considering these results, it is important to bear in mind that most high school principals are men; obtaining data on principals' gender would allow exploration of same-gender and dif ferent-gender relationships between counsel lors and principals. Social support was correlated significantly (r=A6) with age, older counsellors tending to report higher perceived social support than did younger ones. This finding may be explained in two ways. First, given that it was counsellors' perception of support that was measured, older counsellors may have a fuller, more reflective awareness than do younger counsellors of their personal social network and its value to them. Second, counsellors' perception may simply coincide with reality; older counsellors may indeed have wider and deeper social and familial connect ions than do their younger col leagues. 71 There was no significant difference in social support by gender (t = -0 .19 ; p =.85). In the vast literature on social support there is some evidence, albeit tentative and often inconsistent, of gender differences in various aspects of social support. For example, Mallinckrodt and Fretz (1988) found that older professional women who had lost their jobs reported greater social support than their male counterparts; however, the participants in that study were not helping professionals, their number was very small , and they were all middle-aged or older, so conclusions are difficult to draw. Other research connecting gender and social support had similar limitations. As for the current study, perception was the sole aspect of social support measured, and it appears to be unaffected by gender. If social support had been measured in other establ ished ways — for instance, by an inventory of fr iends and family members, or by f requency of social contacts with them, or by diagramming social networks — the results might be different. There were significant dif ferences in percentage of t ime counsell ing by school size ( F = 4 . 8 9 ; p =.001); the larger the school, the greater the counsellor 's percentage of work t ime designated for counsell ing tended to be. This f inding was as expected, and is largely attributable to differences in the apport ionment of counsell ing t ime in schools of various sizes. Small schools tend to have two counsellors (administrators prefer to have one female and one male) but often there are insufficient counsell ing hours available for these counsell ing posit ions to be full-time, so the incumbents teach or administer as well . In medium-sized schools there are typically two or three full-t ime counsellors along with one or more part-t ime ones. In very large schools there is plenty of counsell ing t ime to be allotted, and this permits several full-t ime counsellors (frequently one for each grade level). 72 Levels of Burnout The participants in this research reported significantly higher emotional exhaustion than did the normative group of mental health professionals {t =4.26; p < . 0 0 1 ) . However, they reported significantly lower depersonalization [t = -5 .34 ; p < . 0 0 1 ) and significantly higher personal accompl ishment, an indication of low burnout (f =31.24; p < . 0 0 1 ) . Drawing conclusions f rom comparison of this sample with the norms must be done with caution. Although the normative group chosen for the comparison was, in its composit ion, the closest fit to the high school counsellor sample, there is hardly an ideal correspondence between the two. The Maslach Burnout Inventory mental health sample included school counsellors but also several other types of mental health workers including psychologists and psychiatrists. It is possible that working conditions, types of clients, problems addressed, and personal influences may differ greatly between the two groups. Comparisons of burnout levels must be regarded in that light. With this caution in mind, it is still worth considering the relatively high level of emotional exhaustion found in the present sample, especially in view of the low levels of burnout found in the other two categories. If emotional exhaustion were low, one might conclude that burnout appeared not to be a serious issue for high school counsellors; as it is rather high, it merits more attention than the other two burnout classes. Regrettably, the independent variables selected for this study do not allow us to account for much of the comparatively high levels of emotional exhaust ion — the feeling of being emotionally overextended and fat igued by one's work, so that one feels less able to give of oneself. Emotional exhaustion 73 was significantly correlated with administrative support ( r = - . 1 8 ) ; this explained only 4 .3% of the variance in emotional exhaustion. Emotional exhaustion was correlated weakly and not significantly with social support ( r = - . 1 0 ) , and not significantly with any other independent variable. Plainly, we must look elsewhere for explanations, which are necessari ly speculative. In their meta-analysis of studies of burnout in social work, Soderfeldt, Soderfeldt, and Warg (1995) outl ined factors shown by various researchers to be related to emotional exhaust ion: role conflict, boredom, low job satisfaction, and intention to change jobs. The present study did not measure these four factors, and it is possible that some or all of them are associated with the quite high level of emotional exhaustion found; they merit further research. The relatively high emotional exhaustion found in the sample suggests that help should be offered to counsellors who want it. Intervention could be aimed primarily at counsellors' inner world, at the external stress factors, or both. In the first case, programs could help counsellors develop the attitudes, coping strategies, new counsell ing behaviours and other aptitudes that might alleviate job stress and lessen emotional exhaust ion. Such interventions could take forms similar to those offered by counsellors to their clients who have similar problems (for a review, see Pearson, 1990). In the second case, the specific stressors in the job, if identified, could be minimized by counteractions as needed — for instance, cutting the work load, changing or clarifying the role description of counsellors, or providing them with additional free t ime, more comfortable office space, higher prestige, or other forms of support. Some sources of stress in modern high school counsell ing are discussed further in the next section. However, if interventions are to be useful, they will need to be based on further research into factors related to emotional exhaust ion. 74 Further Discussion The results reaffirm the relevance of perceived social and administrative support to burnout. On the other hand, with the exception of the two variables that appear pertinent to personal accomplishment, the six counsellor and job characterist ics showed only weak, nonsignif icant associat ions with burnout. This implies that other factors are more important. Some variables of interest are suggested in this section, which outlines a few contemporary issues in the working lives of British Columbia high school counsellors. Many of these counsellors have made anecdotal reports (some of which were appended to the questionnaire used in this research, though they were not requested) of feeling overwhelmed by one or more aspects of their work: the seemingly ever-increasing demands of their job, higher caseloads, the tedium of the paperwork burden, lack of helpful evaluation of their work, and difficulty created by the lack of a clear role definition for them. Each issue is considered here in turn. As to rising demands on their t ime and talents, counsellors often f ind themselves having to fill the responsibilities of others — for example, to sort out personal problems of students that in an earlier era would have been handled by parents. There are counsellors who are not educated or inclined to do therapeutic work who nevertheless feel forced to take it on because family support and communi ty resources are not available to students. More and more, counsellors do social problem-solving work such as presenting anti-bullying workshops to groups of their student clients. They also serve, with increasing frequency, as unofficial hosts to social workers who visit the school on child-welfare and child-custody investigations, to police personnel, and to a 75 variety of other visitors. These extra duties must be considered deleterious for counsellors in view of evidence that job sett ings that are burnout-prone are characterized by work overload (Flournoy, 1990). Caseloads have been increased sharply in recent years as a result of budget cuts to British Columbia schools. In early 1998, a survey by the British Columbia School Counsel lors Associat ion showed that student/counsel lor ratios in the province's high schools varied from 541/1 to 225 /1 . Reporting the results of the survey, the association's president noted that the average ratio has increased sharply, leading to a wearing away of direct counsell ing t ime with students: Erosion has several faces: elimination of services; increased teaching t ime; increased paper work (particularly at the senior secondary level), meetings, and consultation with staff and parents. In some districts, erosion has been quite subtle. Until the counsellors actually sat down and looked at what counsell ing services they had been providing for students four years ago and what they are doing today, they did not realize how much actual counsell ing t ime they had lost! And while our services are being eroded, we are all seeing more and more needy children in our schools (Clayton, 1998, p. 1-2). Paperwork appears to be a particular annoyance to counsellors, at least in some school districts. Counsel lors have reported their frustration at the mounting paperwork thrust upon them — for instance, government and school district forms, and scholarship applications for senior students. (A few school districts employ clerical workers to perform these tasks.) These comments echo the f inding of Partin (1993) that there is an enormous difference between the ideal and actual t ime designated for school counsell ing, especially in view of 76 the increasingly heavy load of clerical and quasi-administrat ive tasks involved in the job. Construct ive evaluation of counsellors' work is frequently lacking. While the work of high school teachers is evaluated objectively at least somet imes (for example, through provincial exam results and school test results compared with those of other teachers), the effectiveness of counsellors' work is intrinsically hard to measure. Administrators and fellow counsellors are understandably reluctant to intrude on the privacy of much counsell ing work, but the lack of suitable evaluation may contribute to isolation, confusion, or the repetition of pointless or negative counsel l ing behaviour. The role ambiguity inherent in being a counsellor seems troublesome to many counsellors. Expectations for counsellors vary among school districts, among schools in the same district, and even among counsellors in the same school. Role confusion has many causes and takes diverse forms. Often counsellors feel confused or upset because their commitment to the educational, personal and social development of students comes into conflict with the demands of the system; some counsellors are expected to discipline students, to look after attendance monitoring and follow-up, to do t imetabling, to cover teachers' classes, or to submit to various other administrative intrusions on their t ime and preferred role. These matters can have direct, harmful effects on the counsellor; moreover, they can undermine or destroy the student-counsellor relationship, as the student may no longer see the counsellor as a safe and confidential source of support. An additional concern for counsellors at the moment is the potential influence on their work of the province's new Ministry for Children and Families. This ministry was established to play the leading role in providing social services to children and adolescents. Until now, 77 the ministry's work in schools has been mostly confined to the elementary level. High school counsellors have noted little impact on their work, but some are concerned that a projected reorganization of counsell ing services for adolescents might change or even usurp their present role. Incertitude about how the work of the two groups will interact, and to what effect, is causing disquiet (Clayton, 1998, p. 2). One important aspect of the role confusion is that in British Columbia, as in many other places, the high school counsellor 's duties are not specified legally or contractually. Whether the ambiguity of the role helps to create the contractual vagueness, or results from it, is debatable. In any case, counsellors have no specific legal support in saying "no" to having extraneous or contradictory duties imposed on them. Attempts have been made to develop a role description that can be used by counsellors individually or in groups to assert their desires for, and their rights to, a clear and workable role in the schools. The British Columbia School Counsel lors Associat ion outl ined a job description that included four major e lements: counsell ing, consultat ion, co-ordination and education. This f ramework was adopted by the provincial education ministry (British Columbia Ministry of Education, 1995, p. D5-6). However, given the disparities in the expectat ions held for counsellors in different schools, districts and regions, and the relative isolation of many counsellors, little progress has been made in implementing this model. It may be prudent for counsellors to unify themselves more and to press more strongly for the dignity and clarity of their role. If these five issues in the current working lives of British Columbia high school counsellors had been considered in the present research, it is likely that more of the variance in burnout would have been explained. Which of the five variables would have predicted changes in specific burnout classes is a matter 78 of speculation, but a meta-analysis of research on burnout in social work provides clues. Soderfeldt, Soderfeldt, and Warg (1995) found that emotional exhaustion had been associated with role conflict and with boredom. Therefore, role ambiguity and the tedium of paperwork may help to further explain emotional exhaustion. Soderfeldt et al. found depersonalization to be related to value conflict. The role ambiguity often reported anecdotal ly by counsellors may reflect such conflict. Asking counsellors to specify the vagueness, the di lemmas and the contradictions associated with their work might lead us to a deeper understanding of depersonal izat ion. As to personal accompl ishment, role ambiguity again appeared in the Soderfeldt et al. study as an important related factor. In addition, public-sector work and bad working condit ions were related to reduced personal accomplishment. In an era of governmental budget cuts, the increasing demands on counsellors' t ime and the higher caseloads may explain an important proport ion of variance in personal accompl ishment. Clearly, the five elements of high school counsellors' contemporary work situation discussed in this section need to be weighed in any further research that is done, in the possible development of support ive interventions, and in any practical and policy changes designed to enhance counsellors' work. The quickly changing nature and circumstances of counsellors' work point to the need to consult counsellors directly about research variables, interventions, and ongoing changes to their job descript ions and responsibil it ies. Recommendations for Future Research This concluding section comprises some general suggest ions on how other investigators could proceed with research on this topic, and then some 79 more specific recommendat ions of variables to be examined. As noted in the review of the literature, burnout among high school counsellors has attracted scant attention. Burnout studies in general are somet imes hard to evaluate because of the great variety of methods and instruments used, the somet imes vague reporting, and the inadequately def ined samples employed. Investigations need to be designed rigourously, samples need to be chosen carefully and with little regard for convenience, and thorough statistical analysis should be brought to bear. Much remains to be done if a strong understanding of the problem is to be developed. Although burnout is widely v iewed as a menace, little is actually known about its prevalence. W e need more studies employing descriptive statistics to define the prevalence of burnout in the schools of various school districts, provinces, even countries. There is now little empirical justif ication for answers to such questions as: 1. "What percentage of the counsellor population is burned out?" and 2. "Do people get over being burned out, and if so does the problem recur?" This first question raises a central issue: the lack of a clearcut border between burnout cases and non-cases. There is a natural reluctance to label people, and scores on the MBI and on other instruments are usually compared to the relevant norms without any effort to say whether the individual is burned out; the interventions that have been developed are simply aimed at moderat ing the burnout and thus lowering the score. Adopting cutoff points might help researchers focus their attention on dealing with the serious cases and might also promote knowledge about the varying effects of burnout at different levels, from low to high. The second question above underl ines the importance of longitudinal 80 studies. Despite the practical and methodological difficulties inherent in such research, the longitudinal approach will be needed to overcome the limitations of correlational research, in which cause-and-effect inferences cannot be drawn. For example, the present study conf irmed the relationship of administrative support to burnout, but cannot draw conclusions about causation. In contrast, longitudinal studies have already f ingered lack of supervisory support as an antecedent of burnout (Maslach & Schaufel i , 1993). Longitudinal research would also provide information about the phases of burnout. Research into the natural cycles of school life as they affect counsellors could yield interesting results. Although the pattern of such cycles is the subject of much debate among counsellors, many of them would report that their morale and energy are stronger at certain stages of the school year than at others. This research was conducted in February, mid-year for schools on a linear t imetable but the beginning of a semester for schools on a semester t imetable. It is impossible to know how this t iming affected the results. Future research could investigate seasonal f luctuations by surveying counsellors more than once during the year. There is a need for a common language to denote the phenomenon of burnout wherever and whenever it may appear. Using the same measure in all research would provide consistency and comparabil i ty. The Maslach Burnout Inventory (MBI) is by far the most popular instrument and has well-established psychometr ic utility. Moreover, research outside North America shows the MBI's exportability. "Available studies reveal patterns of burnout covariates much like those isolated in U.S. and Canadian settings. Specifically, the three subdomains of burnout seem to be perceived similarly by samples from different nations" (Golembiewski, Scherb, & Boudreau, 1993, p. 235). Therefore, future studies would probably benefit f rom using the MBI. Notwithstanding this suggestion, it is striking that self-report measures such as the MBI are often the only index of burnout used in studies. Researchers seldom consider such items as job performance evaluations, absenteeism, turnover, illness, and ratings by other such as colleagues, clients or family. The relationship between these measures and self-reported burnout needs investigation. Inquiring into relationships among the three classes of burnout considered by the MBI is another potentially rewarding avenue of research. For instance, could feel ings of frustration and reduced personal accompl ishment lead to emotional exhaustion, and could that then lead to depersonalizat ion? Alternatively, might emotional exhaust ion f rom overwork provoke depersonalization, which in turn could lead to impaired feelings of personal accompl ishment? Or could there be a completely different chain of cause-and-effect? If such sequences could be identified, corrective action would be easier to develop and to implement successfully. This study did not inquire into the specific stressful events that counsellors encounter at work. A future investigation could develop a measure of job-related stressful incidents. Perhaps a pilot study could ask a sample of counsellors to list the most stressful events or situations they had experienced in the preceding months or the previous school year — for example, "having to take work home on the weekend" or "argument with a colleague". Researchers then could combine similar replies and make up an inventory of events for use in their research; participants in the second stage of the research would indicate whether they had experienced each type of event and, if so, would indicate on a scale how stressful the event was to them. The researchers could use the 82 number of experienced events as one measure of stress, and the subjective ratings of intensity of stress as another measure. One potential use of such an instrument would be in research into the mutual interaction between stress and burnout — i.e. how stress may create burnout, and how burnout may increase the number of stresses experienced or intensify their effect. There may be patterns of influence that are so far unknown; as a speculative example, disputes with col leagues or administrators might be shown to be more frequent or more damaging to counsellors than are problems in dealing with student clients. If such patterns were known, remedies might be more easily obtained. In addition to the general recommendat ions made so far, here are several more specific ones about variables that might profitably be considered in future investigations. This study found no relationship between the counsellors' gender and their perceptions of administrative support, but the survey did not inquire about the gender of principals. Until recently, this question would have been almost pointless, as virtually all high school principals were men. However, with more women acceeding to the position, future studies might ask for the gender of both counsellors and principals, with a view to determining whether perceived support of same-gender principals differs f rom that of opposite-gender principals. Although the principal is usually seen to set the tone in a school, there is a wider school culture that could be investigated. The importance of collegial support, including that of counsellors from other schools or other counsell ing settings as well as that of counsellors and teachers in the same school, is widely acknowledged, but such support has not been thoroughly examined in 83 connect ion with burnout. The current study did not inquire into the personal lives of counsellors. One interesting approach would be to consider the influence on burnout of a caregiving role at home; in other words, is counsellors' burnout mit igated, or exacerbated, by having responsibility for others? Or, marital status, al though this is somewhat difficult to define at present, could be considered. Ross, Altmaier and Russell (1989), in their study of U.S. university counsell ing centres, found that married counsellors suffered more emotional exhaust ion than did unmarried ones; however, they noted earlier studies showing just the opposite. Another avenue of inquiry would be to f ind out precisely what counsellors do during their working day. Asking them to specify their duties, along with the amount of t ime allotted to each, would garner valuable information. Researchers might discover, for instance, whether a counsellor with a heavy paper burden is more likely to be burned out than one who spends most of the day in group guidance sessions, or one who offers intense individual counsell ing for personal issues. In light of the widely accepted importance of coping strategies in dealing successfully with stress, future burnout studies could investigate the particular ways in which counsellors endeavour to deal with the daily pressures of their lives and jobs. Sowa, May and Niles (1994) found that rational-cognitive problem-solving ability was not associated with reported stress levels among school counsellors in Virginia; the researchers suggested that other coping resources might prove to be of greater importance. Instruments measuring a range of coping processes such as confrontive coping, escape-avoidance, and positive reappraisal (e.g. the Ways of Coping Questionnaire, designed by Folkman and Lazarus) are available and could yield interesting results. 84 References American Psychiatric Associat ion. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, D.C.: Author. Arends, R. I. (1982). The meaning of administrative support. Educational Administration Quarterly, 18, 79-92. Barrera, M., Jr. (1981). Social support of pregnant adolescents: Assessment issues. In B. H. Gottl ieb (Ed.), Social networks and social support. Beverly Hills: Sage. Beemsterboer, J. , & Baum, B. H. (1984). "Burnout": Definitions and health care management. Social Work in Health Care, 10, 97-109. Blase, J. J. (1982). A social-psychological grounded theory of teacher stress and burnout. Educational Administrat ion Quarterly, 18, 93-113. Brady, J. L , Healy, F. C , Norcross, J. C , & Guy, J . D. (1995). Stress in counsellors: An integrative research review. In Dryden, W. (Ed.). The stresses of counsell ing in action. London: Sage. British Columbia Ministry of Education. (1995). Special education services manual of policies, procedures and guidelines. Victoria: Crown Publicat ions. Capner, M., & Caltabiano, M. L. (1993). Factors affecting the progression towardsburnout: A comparison of professional and volunteer counsellors. Psychological Reports. 73, 555-561. Carrilio, T. E., & Eisenberg, D. M. (1984). Using peer support to prevent worker burnout. Social Casework: The Journal of Contemporary Social Work, 307-310. Cherniss, C. (1995). Beyond burnout: Helping teachers, nurses, therapists and lawyers recover from stress and disil lusionment. New York: Routledge. Cherniss, C , & Krantz, D. L. (1983). The ideological communi ty as an antidote to burnout in the human services. In B. A. Farber (Ed.), Stress and burnout in the human service professions, (pp. 198-212). New York: Pergamon. 85 Clayton, S. (1998, Spring). Data summary raises questions. B.C. School Counsel lors Newsletter, 27. 1-2. Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 38. 300-314. Cohen, L H. (Ed.) (1988). Life events and psychological functioning. London: Sage . Corcoran, K. J . , & Bryce, A. K. (1983). Intervention in the experience of burnout: Effects of skill development. Journal of Social Service Research, 7, 7 1 -79. Cutrona, C. E., & Russell, D. W. (1987). The provisions of social relationships and adaptation to stress. In W. H. Jones & D. Perlman (Eds.), Advances in personal relationships (Vol. 1, pp. 37-67). Greenwich, CT: JAI Press. Day, J . M. (1994). Obligation and motivation: Obstacles and resources for counselor well-being and effect iveness. Journal of Counsel ing and Development, 73, 108-110. Donaldson, G. A., & Coladarci, T. (1987, April). Using school cl imate assessments: An approach to collaborative school improvement. Paper presented at the meeting of the American Educational Research Associat ion, Washington, DC. Dryden, W. (Ed.). (1995). The stresses of counsell ing in action. London: Sage. Farber, B. A. (1983). Introduction: A critical perspective on burnout. In B. A. Farber (Ed.), Stress and burnout in the human service professions, (pp. 1-20). New York: Pergamon. Flournoy, S. A. (1990). Burnout among psychotherapists: Effects of work environment, stressful life events, coping, and social support. Unpubl ished doctoral dissertation, University of Southern California, Los Angeles. Freudenberger, H. J. (1983). A psychoanalytic view of burnout. In B. A. Farber (Ed.), Stress and burnout in the human service professions, (pp. 23-28). New York: Pergamon. Freudenberger, H. J. (1974). Staff burn-out. Journal of Social Issues. 30 (1). 159-165. 86 Golembiewski, R. T., Scherb, K., & Boudreau, R. A. (1993). Burnout in cross-national sett ings: Generic and model-specif ic perspectives. In Schaufel i , W. B., Maslach, C , & Marek, T. (Eds.). Professional burnout: Recent developments in theory and research, (pp. 217-236). London: Taylor & Francis. Greenglass, E. R. (1993). Social support and coping of employed women. In B. C. Long & S. E. Kahn (Eds.), Women, work, and coping: A multidisciplinary approach to workplace stress. Montreal: McGil l-Queen's University Press. Grosch, W. N., & Olsen, D.C. (1994). When helping starts to hurt: A new look at burnout among psychotherapists. New York: Norton. Hargrove, D. S. (1989). Review of the Maslach Burnout Inventory, Second Edition. In J. C. Conoley & J . J . K r a m e r , (Eds.), The tenth mental measurements yearbook (pp. 473-475). Lincoln, NE: Buros Institute of Mental Measurements. Heaney, C. A. (1993). Perceived control and employed men and women. In B. C. Long & S. E. Kahn (Eds.), Women, work, and coping: A multidisciplinary approach to workplace stress. Montreal: McGil l-Queen's University Press. Hyman, R. B. (1993). Evaluation of an intervention for staff in a long-term care facility using a retrospective pretest design. Evaluation and the Health Professions, 16. 212-224. Jupp, J . J . , & Shaul, V. (1991). Burn-out in student counsellors. Counsel l ing Psychology Quarterly. 4. 157-167. Kahili, S. (1988). Interventions for burnout in the helping professions: A review of the empirical evidence. Canadian Journal of Counsel l ing. 22, 162-169. Kahili, S. (1986). Relationship of burnout among professional psychologists to professional expectations and social support. Psychological Reports. 59. 1 0 4 3 - 1 0 5 1 . Kruger, L. J. , Botman, H. I., & Goodenow, C. (1991). An investigation of social support and burnout among residential counselors. Child and Youth Care Forum. 20. 335-352. 87 Long, B.C., & Kahn, S.E. (Eds.). (1993). Women, work, and coping: A multidisciplinary approach to workplace stress. Montreal: McGil l-Queen's University Press. Maslach, C. (1982). Burnout: The cost of caring. Englewood Cliffs, NJ: Prent ice-Hal l . Maslach, C , & Jackson, S. E. (1986). Maslach Burnout Inventory Manual (2nd ed.). Palo Alto, CA: Consult ing Psychologists Press. Maslach, C , & Jackson, S. E. (1981). Maslach Burnout Inventory Research Edition M a n u a l . Palo Alto, CA: Consult ing Psychologists Press. Maslach, C , & Schaufeli , W. B. (1993). Historical and conceptual development of burnout. In W. B. Schaufeli, C. Maslach, & T. Marek (Eds.), Professional burnout: Recent developments in theory and research (pp. 1-16). London: Taylor & Francis. Nagy, S., & Davis, L. G. (1985). Burnout: A comparative analysis of personality and environmental variables. Psychological Reports. 57, 1319-1326. Partin, R. L. (1993). School counselors' t ime: Where does it go? The School Counselor. 40. 274-281. Pearson, R. E. (1990). Counsel ing and social support: Perspectives and practice. London: Sage. Phillips, E. L. (1982). Stress, health and psychological problems in the major professions. Washington, DC: University Press of America. Pines, A., & Aronson, E. (1988). Career burnout: Causes and cures. New York: Free Press. Pines, A., & Maslach, C. (1980). Combatt ing staff burn-out in a day care center: A case study. Child Care Quarterly, 9. 5-16. Ross, R. R., Altmaier, E. M., & Russell, D. W. (1989). Job stress, social support, and burnout among counseling centre staff. Journal of Counsel ing Psychology, 36. 464-470. Russell, D., Altmaier, E. M., & Van Velzen, D. (1987). Job-related stress, social support, and burnout among classroom teachers. Journal of Applied Psychology. 72. 269-274. 88 Sandoval, J . (1989). Review of the Maslach Burnout Inventory, Second Edition. In J. C. Conoley & J . J . Kramer, (Eds.), The tenth mental measurements yearbook (pp. 475-476). Lincoln, NE: Buros Institute of Mental Measurements. Schwab, R. L , & Iwanicki, E. F. (1982). Perceived role conflict, role ambiguity, and teacher burnout. Educational Administrat ion Quarterly, 18. 60-74. Selder, F. E., & Paustian, A. (1989). Burnout: Absence of vision. In D. T. Wessells, Jr., et al. (Eds.). Professional burnout in medicine and the helping professions, (pp. 73-82). New York: Haworth Press. Shinn, M., Rosario, M., Morch, H., & Chestnut, D. E. (1984). Coping with job stress and burnout in the human services. Journal of Personality and Social Psychology. 46. 864-876. Soderfeldt, M., Soderfeldt, B., & Warg, L.-E. (1995). Burnout in social work. Social Work. 40. 638-646. Sowa, C. J . , May, K. M., & Niles, S. G. (1994). Occupational stress within the counsel ing profession: Implications for counselor training. Counselor Education and Supervision. 34. 19-29. Sutton, J. M., Jr., & Fall, M. (1995). The relationship of school climate factors to counselor self-efficacy. Journal of Counsel ing and Development. 73. 331-336. Ulrich, D. N., & Dunne, H. P., Jr. (1986). To love and work: A systemic interlocking of family, workplace, and career. New York: Brunner/Mazel. Vaux, A., Phillips, J. , Holly, L , Thomson, B., Will iams, D., & Stewart, D. (1986). The Social Support Appraisals (SS-A) Scale: Studies of reliability and validity. American Journal of Communi ty Psychology, 14, 195-219. Wessells, D. T., Jr., et al. (Eds.). (1989). Professional burnout in medicine and the helping professions. New York: Haworth Press. Wil l iams, A. (1995). Self-care and the counsellor: A literary review. The B.C. Counsellor. 17. 26-32. World Health Organization. (1992). The ICD-10 classification of mental and behavioural disorders: Clinical descript ions and diagnostic guidelines. Geneva: Author. 89 Appendix A Notes on Returns and Scoring Returns Of the 265 surveys mailed out on February 1, 1998, 163 were filled in and returned, for a return rate of 6 1 . 5 1 % . Included in the return rate were six questionnaires that were not used: — One was discarded because only scattered parts of it were completed, and the respondent appeared to have seriously misunderstood the directions. — Another was set aside because Part IV (the burnout measure) had been detached and was not returned. — A third questionnaire was not used because the respondent, normally a counsellor, was teaching full t ime that school year and not counsell ing at all. — Three questionnaires arrived too late to be included in the analysis. Five additional questionnaires were returned without being filled in; these were not included in the return rate. The 157 surveys used in this study were complete except for seven that lacked a response on years of counsell ing experience, and three that lacked an answer on the percentage of t ime designated for counsell ing. Scoring In consultation with Dr. Beth Haverkamp, a member of the research committee, decisions were made about how to deal consistently with occasional responses that did not entirely fit the directions on the questionnaire. 90 O n the burnout measure, f rom time to t ime answers such as "always" and "once in a while" were entered instead of the required numbers. These replies were converted into the most closely corresponding numbers. Also on the burnout measure, answers on individual i tems were somet imes omitted. Each missing answer was scored as the average of the other scores on the subscale to which the item belonged. In all parts of the questionnaire, whenever a participant gave two answers to an item, or an answer fell between two scores, the higher score was used. 91 Appendix B Covering Letter Appendix C High School Counsellor Questionnaire High School Counsellor Questionnaire In Part I of this survey, you are asked for information about you and your job. Then, in Part II, you are asked to comment on a set of statements about your relationships with family and friends. Part III invites you to respond to statements about your principal. Finally, in Part IV, you are asked to say how frequently you exper ience some job-related feelings. Part I Please provide the fol lowing information by filling the blank or using a • , as needed. a. Age: b. Gender: Female Male c. Years of experience as a part-t ime and / or full-time counsellor: d. Highest degree you hold (in any field of study): Bachelor's Graduate Diploma Master's Doctorate e. The student population of your school is: Less than 500 500-999 1000-1499 1500-1999 2000 or more f. Specify your present regular duties in the school: Counselling only Counsell ing and other duties such as administration or teaching * * If you have duties other than counsell ing, please indicate the percentage of your t ime that is designated for counsell ing: % 1 95" Part II Below is a list of statements about your relationships with family and friends. By circling one number in each row, please indicate how much you disagree or agree with each statement. Strongly Disagree Agree Strongly disagree agree 1. My friends respect me. 1 2 3 4 2. My family cares for me very much. 1 2 3 4 3. I am not important to others. 1 2 3 4 4. My family holds me in high esteem. 1 2 3 4 5. I am well liked. 1 2 3 4 6. I can rely on my friends. 1 2 3 4 7. I am really admired by my family. 1 2 3 4 8. I am respected by other people. 1 2 3 4 9. I am loved dearly by my family. 1 2 3 4 10. My friends don't care about my welfare. 1 2 3 4 11 . Members of my family rely on me. 1 2 3 4 12. I am held in high esteem. 1 2 3 4 13. I can't rely on my family for support. 1 2 3 4 14. People admire me. 1 2 3 4 15. I feel a strong bond with my friends. 1 2 3 4 16. My friends look out for me. 1 2 3 4 17. I feel valued by other people. 1 2 3 4 18. My family really respects me. 1 2 3 4 2 16 Strongly Disagree Agree Strongly disagree agree 19. My friends and I are really important to each other. 20. I feel like I belong. 2 1 . If I died tomorrow, very few people would miss me. 22. I don't feel close to members of my family. 23. My friends and I have done a lot for one another. Part III Here are some statements about your principal. Please indicate the relative f requency with which these behaviours occur (or would occur) by circling one number after each statement. 1. My principal makes supportive statements about my work. Never Rarely Sometimes Usually Always 1 2 3 4 5 2. My principal helps me to be clear about my role. 1 3. My principal would defend me against dissenting voices. 3 97 Never Rarely Sometimes Usually Always 4. My principal gives me resources I value (e.g. free time, clerical support, professional growth opportunities, desirable office space). 5. My principal makes t ime available to deal with my projects and concerns. 6. My principal knows how I contribute to the school. 7. My principal encourages me to express my ideas. 8. My principal would help relieve the pressure if I felt overwhelmed by the demands of the job. My principal would consider my needs when changes were being made that would affect me and my work. 10. My principal's style of leadership is one with which I am comfortable. Part IV Please see the next sheet. On it, recipients refers to the students you counsel. 98 Christina Maslach • Susan E. Jackson MBI Human Services Survey T h e purpose of th is survey is to d iscover how var ious persons in the h u m a n services or helping profess ions v iew thei r jobs and the people wi th w h o m they w o r k closely. Because persons in a w ide var iety of occupat ions wil l answer this survey, it uses the t e r m recipients to refer to the people for w h o m you prov ide your se rv ice , care , t reatment, or instruct ion. W h e n answer ing this survey please think of these people as recipients of the service you provide, even though you may use another te rm in your work. O n the fol lowing page there are 22 s ta tements of job- re la ted feel ings. Please read each statement careful ly a n d decide if you ever feel this way about your job. If you have never had this feel ing, wri te a "0 " (zero) before the s ta tement . If you have had this feel ing, indicate how often you feel it by writ ing the number (from 1 to 6) that best descr ibes how frequently you feel that way. An example is shown below. Example: HOW OFTEN: 0 1 2 3 4 5 6 Never A few times Once a A few Once A few Every a year month times a a times day or less or less month week a week HOW OFTEN 0 - 6 Statement: I feel depressed at work. If you never feel depressed at work , you wou ld wri te the number "0 " (zero) under the heading " H O W O F T E N . " If you rarely feel depressed at w o r k (a few t imes a year or less), you would wri te the number " 1 . " If your feel ings of depress ion are fairly f requent (a few t imes a week , but not daily) you wou ld write a "5 . " Consulting Psychologists Press, Inc. 3803 E. Bayshore Road • Palo Alto, CA 94303 Published by Consult ing Psychologists Press, Inc. ("CPP"), 3803 E. Bayshore Road, Palo Alto, California 94303. MBI Human Services Survey® 1986 by Consult ing Psychologists Press, Inc. This copyrighted publication is not offered for sale; it is for l icensed use only, and then only by qualif ied professionals whose qualif ications are on file with and have been accepted by CPP. CPP reserves all rights beyond the limited scope of this l icense, includ-ing, without limitation, all rights under U.S. and international copyright and trademark laws. No portion of this publication may be reproduced, stored in a retrieval sys-tem, or transmitted in any form or media or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior wri t ten permission of CPP. This copyrighted publication may not be resold, subl icensed, exported, redistributed, otherwise transferred, or used in any manner by any party other than the person or entity to whom it is l icensed for use by CPP; any violation of these restrictions may infringe CPP's copyright under 17 U.S.C.§106(3), and any such violation shall automatically terminate any l icense to use this publication. Printed in the United States of Amer ica. 01 00 99 98 97 28 27 26 25 2 4 MBI Human Services Survey HOW OFTEN: 0 i 2 3 4 5 6~~ Never A few times Once a A few Once A few Every a year month times a a times day or less or less month week a week HOW OFTEN 0 - 6 Sta tements : 1 . I feel emotional ly drained f rom my work. 2. I feel used up at the end of the workday. 3. I feel fat igued when I get up in the morning and have to face another day on the job. 4. I can easily understand how my recipients feel about th ings. 5. I feel I treat some recipients as if they were impersonal objects. 6. Working wi th people all day is really a strain for me. 7. I deal very effectively with the problems of my recipients. 8. I feel burned out f rom my work. 9. I feel I'm posit ively inf luencing other people 's lives through my work. 10. I've become more cal lous toward people s ince I took this job. 1 1 . . I worry that this job is hardening me emotional ly. 12. I feel very energetic. 13. I feel f rustrated by my job. 14. I feel I'm work ing too hard on my job. 15. _ _ _ _ _ I don't really care what happens to some recipients. 16. Working with people directly puts too much stress on me. 17. I can easily create a relaxed atmosphere wi th my recipients. 18. I feel exhi larated after work ing closely with my recipients. 19. I have accompl ished many worthwhi le th ings in this job. 20. I feel like I'm at the end of my rope. 2 1 . - In my work, I deal with emot ional problems very calmly. 22 . I feel recipients blame me for some of their problems. (Administrative use only) cat. cat. cat. EE: DP: PA: (OO 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items