UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Changing images of self : the efficacy of retirement preparation programmes for women Comish, Sara Elizabeth 1995

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

Item Metadata

Download

Media
831-ubc_1995-982721.pdf [ 5.27MB ]
Metadata
JSON: 831-1.0053960.json
JSON-LD: 831-1.0053960-ld.json
RDF/XML (Pretty): 831-1.0053960-rdf.xml
RDF/JSON: 831-1.0053960-rdf.json
Turtle: 831-1.0053960-turtle.txt
N-Triples: 831-1.0053960-rdf-ntriples.txt
Original Record: 831-1.0053960-source.json
Full Text
831-1.0053960-fulltext.txt
Citation
831-1.0053960.ris

Full Text

CHANGING IMAGES OF SELF: THE EFFICACY OF RETIREMENT PREPARATION PROGRAMMES FOR WOMEN by SARA ELIZABETH COMISH B.A. (Horns.) The University of Alberta, 1983 M.A, The University of Alberta, 1986 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Departnnent of Counselling Psychology) We accept this thesis as confornning to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March, 1995 @Sara Elizabeth Comish, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and stud, I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Signature) Department of Counselling Psychology The University of British Columbia Vancouver, Canada Date 23 March, 1995 DE-6 (2/88) Changing innages ii Abstract Facing retirement can be a stressful experience for sonne older adults. A cognitive-behavioural intervention was designed to help older wonnen cope with retirement anxiety, specifically targeting positive images of self in the future, described by Markus and Nurius (1986) as possible selves. Sixty-one pre-retirement women ranging in age from 49 to 71 (/W = 59.9) were randomly assigned to either a targeted-change group, a structured-discussion group, or a wait-list control group. The targeted-change group and the structured-discussion group met for 2 hours a week for 8 weeks. A modified state form of the State-Trait Anxiety Inventory (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), the Short form of the Multiscore Depression Inventory (Berndt, 1986), the Retirement Self-efficacy questionnaire (Poser & Engels, 1983), and a measure of possible selves were administered pre- and post-treatment, with a 3-month follow-up. Following the post-test, participants in the wait-list were assigned to one of the retirement seminars. It was predicted that from pre- to post-test participants in the targeted-change group would show a greater change in the measures than participants in the structured-discussion groups who would in turn, show a greater change than participants in the wait-list. The predictions were tested using planned orthogonal univariate tests. Retirement self-efficacy and positive possible selves increased more for participants in the targeted-change group and structured-discussion groups than for participants in the wait-list from pre-test to post-test. In addition, positive possible selves increased more for participants in the targeted-change than for participants in the structured-discussion group from pre-test to the average of post-test and follow-up. It was concluded that the interventions were effective in changing how participants saw themselves in the future. Conclusions with regard to treatment efficacy, however, were restricted by observed change in the wait-list group on the Changing images ill anxiety measure following the initial interview, The results are discussed in terms of non-specific treatment effects, along with issues relating to research on possible selves, and implications for retirement programmes. Changing innages iv TABLE OF CONTENTS Abstract Table of Contents List of Tables List of Figures Acknowledgement Dedication Chapter One Chapter Two Chapter Three Introduction Possible Selves Retirement Stress Retirement Preparation Programmes Review of the literature Retirement as Crisis Individual Differences in Retirement Gender Issues Retirement Preparation Programmes Summary Counselling Approaches to Retirement Preparation Programmes Possible Selves Overview of the Study and Hypotheses Hypotheses Method Participants IVlaterials Spielberger State-Trait Anxiety Inventory (STAI-S) ii iv vii ix X xi 1 2 4 5 9 9 15 19 22 27 28 31 37 38 40 40 41 41 Chapter Four Short form of the Multiscore Depression (SMDI) Retirement Self-efficacy (RSE) Possible Selves Pilot Testing Design Procedures Expectations of Treatment Efficacy Group Leaders Treatment Fidelity Clinical Significance Data analysis Type 1 and Type II error Results Sample Characteristics Group Comparability Correlation Matrix Preliminary Analysis Main Analysis Hypothesis Tests Post-hoc Analysis Anxiety and Depression Clinical Significance States of Mind Depression Subscales Retirement Self-efficacy Subscales Evaluation Questions Changing images V Inventory 42 43 45 47 47 48 50 51 51 51 52 55 56 56 56 63 65 67 67 83 83 83 87 89 94 96 Changing Innages vi Preparatory Behaviours 98 Mid-treatment Means 101 Chapter Five Discussion 102 Common Treatment Factors 103 Significance 105 Role of Anxiety in Transition 107 Retirement Counselling for Women 108 Retirement Preparation Programmes 108 Possible Selves 109 Contributions 113 Limitations 114 References 115 Appendix A, Directions used for the modification to the State-Trait Anxiety Inventory 123 Appendix B, Coding Manual for Possible Selves 124 Appendix C, Demographics Sheet 126 Appendix D, Expectancy Questionnaire 127 Appendix E, Manipulation Checl< on Treatment Fidelity 128 Appendix F. Programme Evaluation Questionnaire 130 Appendix G Mid-treatment means and standard deviations 132 Changing Images vll List of Tables Page Table 1. Planned contrasts for three groups pre- to post-test 53 Table 2, Planned contrasts for two groups pre-test, post-test, and follow-up 54 Table 3, Age and years of education for participants In each of the groups 57 Table 4, Frequencies for number of years worked by group 58 Table 5. Frequencies for planning to retire by group 59 Tabled. Marital status frequencies by group 60 Table 7, Occupational status frequencies by group 61 Table 8. Reason for retirement 62 Table 9. Correlation matrix for dependent variables and demographic variables 64 Table 10. Expectations regarding effectiveness and recommending course, means, and standard deviations 66 Table 11. Means and standard deviations for targeted-change, structured-discussion, and wait-list groups pre- and post-test 71 Table 12. Results of planned contrasts for change from pre- to post- to follow-up for combined targeted-change and structured-discussion groups 78 Table 13. Means and standard deviations for dependent measures, pre-post- and follow-up for combined targeted-change and structured-discussion groups 79 Table 14. Reliable change index for STAI scores 84 Table 15. Improvement and deterioration rates for responses on the STAI 84 Table 16. Reliable change index for SMDi scores 85 Table 17, Improvement and deterioration rates for responses on the SMDI 86 Table 18, States of Mind Ratio 88 Changing innoges viii Table 19. States of Mind categories at post-test 88 Table 20. Means and Standard deviation for the SMDI subscales 90 Table 21. Analysis of Variance results for the SMDI subscales 92 Table 22. Analysis of Variance results for the RSE subscales 94 Table 23. Means and standard deviations for the RSE subscales 95 Table 24. Means and standard deviations for the evaluation questions 97 Table 25. Frequencies for retirement preparation behaviours 99 Changing images ix List of Figures Page Figure 1. Design of the study iiiustrating the cross-iagged approach 47 Figure 2, iVIean scores as a function of time of measurement: pre-and post-treatment; and group; targeted-change, structured-discussion, and wait-list 73 Figures, Mean scores as a function of time of measurement: pre-test post-test and foiiow-up; and group: targeted-change versus structured-discussion 80 Changing images X Aci<nowledgement This dissertation would not have been possible without the time and effort of the participants in the project, I am grateful to the women who participated for their sharing and openness. In addition, I would lil<e to thank my co-leaders, Allison Krause, Elaine Young, Donna Roberge, Sally Shamai, and Karen Flood, for their help and commitment, I am indebted to my supervisor, Bonnie Long, for her guidance and support throughout the course of this work. I am particularly grateful for her tolerance of the demands of my life outside of graduate school, I have been very fortunate throughout the dissertation process to have had a committee (Bill Borgen, Brian de Vries, and Nand Kishor) and supervisor, who have been genuinely constructive, helpful, supportive, and kind. Changing images xi Dedication For John, because he made the effort to l<now. And for Evan, because it didn't matter that he didn't l<now, Changing images 1 Chapter One Introduction Retirement has become accepted as a life transition in which older adults exchange the struggles and benefits of work for a more self-directed way of life, Although there is an assumption that most people will retire, there is a surprising lack of information about retirement and about how people cope with the retirement transition, Statistics Canada, for example, does not routinely keep data on the number of people who are retired, nor on the number of people who retire each year, Data on retirement has to be indirectly inferred from labour force participation rates (McDonald & Wanner, 1990) or on specific studies of retired people (Lowe, 1991), In 1986, there were 2.7 million adults in Canada age 65 and over, and 7% were employed full-time in the work-force (Methot, 1987), Presumably, many of the remaining 93%, or 2.5 million consider themselves to be retired or semi-retired, although some of these may be unemployed, disabled, or have never participated in the work-force. This is, however, an underestimate of the number of retired people. Lowe (1991) estimates that in Canada, 63% of retirees leave their work before age 65. For some of those people who retire, regardless of their age, coping with the retirement transition may be easy and successful. For others, however, it is a stressful time accompanied by anxiety and depression. The retirement process can be placed in the framework of life transitions, Placing transitions within a contextual perspective, Schlossberg (1984) describes transitions as dynamic processes that have different meanings for different people depending upon the context of the individual. In addition, within individuals, transitions can have both positive and negative meanings at the same time. The key to understanding the impact of a transition, from this perspective, is the meaning of the transition for the individual, Similarly, Lazarus and Folkman Changing images 2 (1984) have proposed that stress and the coping responses that accompany stress are dynamic experiences which may be initiated by an external event, but whose true significance lies in the appraisal of that event by the individual (Folkman & Lazarus, 1988a, 1988b; Lazarus & Folkman, 1984). Smith, Haynes, Lazarus, and Pope (1993) distinguish appraisals from attributions. Attributions are cognitions about the causes of an event. Appraisals, on the other hand, are evaluations of the significance of an event. They argue that it is appraisal that evokes an emotion (Smith et al., 1993; Smith & Lazarus, 1993). Thus, stress is a transaction arising from the relationship of individual and environment when the individual perceives that the demands of the environment may exceed available resources. In primary appraisal, the event is evaluated for its threat to well-being. In secondary appraisal, the individual evaluates the ability to cope with the event. Thus, individuals appraise an event as stressful when they think that they are lacking the skills needed to cope, and in addition, both the emotion and the appraisal are dynamic, on-going processes (Folkman & Lazarus, 1988b). There are two types of coping functions; problem-focused coping, in which direct action is taken to deal with the stressful event, and emotion-focused coping in which efforts are made to deal with the emotional results of the stressful event (Lazarus & Folkman, 1984), Possible Selves At the centre of both the Lazarus and Folkman, and the Smith and Lazarus models is the appraisal process which entails an evaluation of the well-being of self involving both resources and situational demands. In the case of pre-retirement stress, individuals must anticipate the situation that they will be in and put themselves in that situation. Markus and Nurius (1986, 1987) have recently proposed a concept that is particularly relevant for understanding how people view themselves in the future. They propose that possible selves are images of Changing images 3 what peopie would iil<e themselves to be and what they fear themselves to be at a future point in time, Possible selves thus consist of sets of images about the self, known as self-schemata, and these images may include knowledge of the resources needed for coping, such as skills, competence, persistence in the face of obstacles, and confidence in coping with emotions, On the other hand, these images may contribute to the experience of stress if the resources needed to cope with the future situation are perceived of as lacking or inadequate, Prior research on self-schemata has focused on images of self in the present or past, rather than the future. The self-schema has been defined as an organized body of knowledge about the self in a particular domain that integrates and guides processing of new information (Fiske & Taylor, 1984), Information that is not a part of the schema may be distorted, ignored or discounted (Bower, Black, & Turner, 1979; Fiske & Taylor, 1984), Thus, the schema can result in biases in memory, or mis-perceptions of events in favour of material that is consistent with the schema. For example, individuals who have a self-schema for being gregarious with work mates may Interpret social interactions in a way that Is consistent with the self-schema, and fail to attend to information about inconsistent information, such as declining a number of invitations to attend a seniors club. The addition of the future orientation to the self-schema concept adds certain dimensions, Markusand Nurius (1986,1987) propose that the dynamic images that people have of themselves in the future may operate to motivate behaviour, and that these images operate similarly to a self-schema, A future self can serve as an incentive for initiating behaviour, or delaying gratification as in working on a fund-raising benefit to sponsor a club outing. In addition, the possible self-schema may contain information about how to achieve the goal, for example, the image of the self as an award-wining volunteer also contains Changing innages 4 information about the necessity to spend nnany hours volunteering. In this sense, the use of the possible-selves concept invokes sonne of Bandura's notions of self-efficacy (Banduro, 1977), Self-efficacy describes a person's perceived sense of ability to produce a desired effect and Is thus an expectation about future behaviour, in a sinnllar fashion to possible-selves, self-efficacy expectations can Influence future coping behaviours and can be used to predict behaviour in certain situations (Banduro, 1977), Group counselling can assist in altering seif-efficacy expectations by providing an atnnosphere in which effective behaviour can be modelled, behavioural rehearsal can be encouraged, and feedbaci< can be provided by group leaders and group members. In effect, this process serves to build and strengthen schemata of self as capable of effective action in a given situation, In terms of coping with retirement stress, an intervention can strengthen coping behaviours and change the nature of appraisals of the situation and of the available resources. Assessing the images of possible selves would allow for assessment of the mechanisms by which the treatment operates. Retirement Stress One aspect of retirement stress that has been examined in the literature Is satisfaction with the transition from work to retirement. Life satisfaction Is defined by Crowley (1985) as a multidimensional concept with cognitive, affective, and behavioural components. As Bortner and Hultsch (1970) indicate. It is a concept referring to a general assessment of life quality, This concept Is assessed in several ways ranging from a single-item measure asking people to rate global satisfaction on LIkert scales of 5- to 10-point spreads, to asking people to evaluate aspects of their mood such as anxiety, sadness, joy, and calm, again using similar scale spreads. Changing images 5 In the literature, it is typically concluded that there is no evidence that retirement has a universally negative impact on well being (KasI, 1980; Matthews & Brown, 1987; Matthews & Tindale, 1987; McGoldrick & Cooper, 1985; Palmore, Fillenbaum, & George, 1984), This conclusion rests upon a series of studies that reveal that the majority of retirees indicate on measures of life satisfaction that they are satisfied with their lives. This majority has been estimated as consisting of approximately two-thirds of retirees (KasI, 1980; McGoldricl< & Cooper, 1985), The finding that two-thirds of retirees are satisfied, also implies that one-third are other than satisfied with the retirement experience (KasI, 1980), In addition, Parnes and Nestel (1981) report that 19% of retirees regret their decision to retire. Retirement places a number of new demands on people, and for some, retirement, and its anticipation, are stressful. Concerns around this event include, fear of ageing, lack of ability to motivate oneself, anxiety about loneliness, and the demands of living on a reduced income. Women, in particular, face the prospect of being alone and living below the poverty level (Gee & Kimball, 1987). In addition, women facing retirement today have a different work history from men in their cohort. Typically, these women have experienced extensive periods away from their jobs while caring for their children and have held different types of jobs than men (Gee & Kimball, 1987; O'Rand & Henretta, 1982), In this research, the focus will be on retirement stress in women. Retirement Preparation Programmes For people who are troubled about retirement, pre-retirement programmes have the potential to be particularly beneficial in helping with the transition, and clarifying expectations about the event. In the past 10 years, retirement preparation programmes have become increasingly popular and many have been developed to help people with the retirement transition (Avery & Jablin, 1988), Many of these programmes are operated by corporations as Changing Images 6 part of their employee benefits and are offered to employees who are in the retirement age range. Content and scope of these programmes varies considerably, although the majority appear to focus on financial planning and pension benefits. Typically, these programmes consist of approximately 12 to 14 hours offered either as weel<end workshops or in 1 1/2 to 2 hour sessions over 2 months (Comrie, 1985; Durrant, 1988; Palmore, 1982). Durrant (1988) reports that the usual format is for a speaker to make a presentation on a specific content topic, and that this may be followed by discussion and questions. With regard to the effectiveness of pre-retirement programmes for reducing stress, evaluations of these information programmes have indicated mixed results. Durrant (1988) found that an information programme was effective in reducing anxiety in male participants but not in female participants. Phillipson and Strang (1983) have criticized didactic, information approaches to retirement preparation because of the failure in this format to thoroughly explore participant concerns about retirement. Following an in-depth evaluation of retirement preparation programmes offered by eight British companies, Phillipson and Strang made a series of recommendations. These included restructuring the format of the programmes to allow greater interaction among participants, and focusing upon the counselling element of retirement education. Under a counselling model, exercises could focus on development of skills in communication, life-planning, and problem-solving, while allowing participants to increase awareness of their emotions and goals and to integrate these with their plans for retirement, In effect, Phillipson and Strong are recommending that pre-retirement programmes utilize strategies designed to help participants cope with stress. Counselling has a variety of techniques to offer that have been successful in decreasing anxiety and improving coping, Two cognitive-behavioural approaches that have targeted these coping skills are stress inoculation training Changing images 7 (Meichenbaum & Cameron, 1983) and problem-solving (D'Zurilla, 1988), Therapies in these frameworks have been primarily focused on teaching sl<ills rather than just giving information. The format typically takes the form of small groups within which the skills ore explained, modelled, and practiced. Furthermore, participants are encouraged to use the skills outside of the group and to monitor their own behaviour. Programmes such as these have been effective in reducing anxiety and depression in a variety of age groups (D'Zurilla, 1988; Matheny, Aycock, Pugh, Curlette, & Cannella, 1986), These approaches teach people skills that they can use in future situations and In so doing, may change participants' images of themselves and increase their sense of effectiveness. Creating effective possible-selves may help pre-retirees cope with the stress of retirement. The potential, from a counselling perspective, for the possible-selves perspective lies in its relevance for understanding the change process. Building or strengthening a particular set of possible selves may increase motivation to behave according to that set. Thus, the possible selves concept has the potential to explain how changes in cognition can mediate change in behaviour. With regard to altering self-efficacy, the techniques of discussion, visualization, modeling, and role-playing may build a sense of self that is competent and capable of mastering future situations. In turn, this may increase the probability of taking action and of persisting despite failure experiences and obstacles. In addition, these counselling techniques promote anticipatory coping using mental simulations of future events (see Taylor & Schneider, 1989), Further, the techniques demonstrate to the participants through the use of role plays that they can put the skills into action, thus increasing the frequency with which the positive selves are instantiated, Providing evidence that possible-selves change Changing images 8 over the course of treatment would suggest that these self-schemata play a role in the mechanism by which the treatment operates. The question that was addressed in the present research was whether a cognitive-behavioural pre-retirement programme, compared with a structured-discussion group and a wait-list control group would be effective in changing the way women view themselves in future retirement, in creating a view of self as an effective problem-solver, and in reducing anxiety and depression, Changing Images 9 Chapter Two Review of the literature Retirement, through a system of pension plans and legislation, has become entrenched as an expected process In the lives of most older adults, despite recent challenges to mandatory retirement policies. Retirement has been studied from a psychological perspective for a number of years, although not to the same extent as other life events, such as widowhood, Research on retirement has typically fallen into four categories; exploring retirement as life crisis, examining predictors of life satisfaction, gender issues, and retirement preparation programmes for coping with the retirement transition, Programmes that incorporate counselling techniques to foster problem-solving and increase self-efficacy appear to have most potential to increase coping and decrease retirement anxiety. The concept of possible selves, images of self In the future, fits well within the frameworl< of the retirement transition, and may serve as a mechanism contributing to change in the way pre-retirees see themselves coping in their retirement, Retirement as Crisis Research exploration in the field of retirement has been primarily theory-driven. However, as Palmore et ol, (1984) point out, these theories are very general and not well-developed. The life events theory, borrowed from the adult development area of psychology, is a prime example of such a theory. Under this approach, certain life-events are presumed to provoke crises in which the steady functioning of the individual is disrupted until a return to homeostasis can be accomplished (Hultsch & Plemons, 1979; Whitbourne, 1985). Within this frameworl<, retirement was conceived of as a stressful life event with a negative impact on the psychological adjustment and well-being of the retiree. Some researchers have gone as far as referring to a "retirement disease" (Selye, 1984, Changing Images 10 p, 32) to describe the outcome of this crisis, Thus, retirement is perceived of as a crisis, having a universally negative impact, at least initially until the retirees adapt to their new non-role or find a work replacement (Schnore, 1985), Although the exact extent of the time period Involved in this transition is unspecified, it appears to be limited to 1 or 2 years following the event. In this sense, retirement is the process of transition from a formal work-role; rather than the outcome of the transitional phase representing the entire life-span from the ending of the work role to death. The study of the domain covered by the second definition is typically reserved for psychologists studying old age. It is the first definition that has been the focus of interest for psychologists studying retirement and it is the transition process that is presumed to create crisis. The prediction that follows from the life events framework is that psychological well-being will be adversely affected by the process of retirement, The assumption made here is that one's psychological health will be poorer for a certain time after leaving work than it was when working. Psychological well-being, otherwise known as morale, or life satisfaction, is defined by Crowley (1985) as a multidimensional concept with cognitive, affective, and behavioural components and is assessed using a variety of paper and pencil measures for life satisfaction. Primarily, a single question is used asking "how satisfied are you with your life?". Various Likert scales are used, ranging from 5-point spread to the 11 -point spread of the Cantril self-anchoring ladder (Bortner & Hultsch, 1970), In terms of research questions, the crisis approach has generated one question, "does the retirement transition have a negative impact on well-being for all people?". Presumably, if it does then retirement can be classified as a crisis, Unfortunately, the individual's unique interpretation of events is lost with this perspective. In a contextual perspective, on the other hand, people are conceived of as actively constructing their world, and consequently the Changing Innages 11 presupposition that a single external event such as retirement would have the same Impact on all people is unrealistic. Based on studies indicating that the majority of retirees report high life satisfaction scores, It Is generally concluded that there is no evidence for retirement having a negative impact on well being (KasI, 1980; Matthews & Brown, 1988; Matthews & Tindale, 1987; McGoldrick & Cooper, 1985; Palmore et al., 1984), For example, Atchley (1976) is frequently cited for his research on telephone operators which revealed that approximately two thirds of the women indicated that they were quite satisfied with their life after retirement. There are, however, no control groups in this study; it is possible that these same proportions of life satisfaction are found in all adults, regardless of retirement status. In addition, as Kosi (1980) points out, even If there were a comparison to non-retired adults of the same age, cross-sectional designs are inadequate tests of the retirement question because retired and non-retired people of the same age differ on several important dimensions. These Include health, income, job commitment, occupational status, to name but a few. The question that is being asked in the research on life satisfaction is "does retirement cause a change In life satisfaction." To adequately answer a question of this sort about the change process, longitudinal studies that follow individuals across time are necessary (Baltes, Reese, & Nesselroade, 1977) and thus, it would be necessary to follow the same individuals as they make the retirement transition. In this section, I examine in further detail a study using this more rigorous design, which has been cited as evidence that retirement does not provoke a crisis (KasI, 1980; Matthews & Brown, 1987; McGoldrick & Cooper, 1985). Bell (1975) made a concerted effort to test the crisis theory in a longitudinal study. He developed two specific hypotheses derived from the crisis theory, First, that life satisfaction would decline following retirement, and second, that Changing Images 12 behaviour in a series of domains would be disrupted, These domains were family involvement, measured by hours of contact; community involvement, measured by hours spent with friends and In participation in political activities; and volunteer involvement, again measured by hours of participation. Bell collected data from 114 men prior to retirement, and again, 1 year later, on average 6 months following retirement. He assessed life satisfaction using a single item, "On the whole, how satisfied would you say you are with your way of life today?" (p. 158), and a 5-point scale. T-tests revealed that life satisfaction declined following retirement, but that behaviour, as measured by the above activities was not disrupted. Subsequently, Bell concluded that retirement did not provol<e a crisis. Adding the behavioural components certainly helps to expand the research on retirement. It is not clear, however, why the removal of work should be expected to disrupt these particular sets of behaviours. Research on the meaning of work suggests that work can mean mastery, satisfaction, distraction from family problems, stress, structure, and identity (Benner, 1984). Removing work roles, then, is more likely to disrupt personal fulfillment, sense of identity, satisfaction with marital relationships, and scheduling rather than the number of hours spent on various activities. Consequently, these results are not surprising. It is interesting, though, that the decline in satisfaction Is Interpreted as evidence against the crisis theory. I am unsure why this is the case. The participants of Bell's (1975) study were interviewed a third time, on average, a year and a half following retirement. Bell (1979), however, did not report any analyses of the time three data with the earlier time one and two, so further examination of change over time is not possible. In addition, there Is another alternate explanation of the results. It is possible, given the absence of a control group, that life satisfaction scores always decline over a 1 -year period for people in this age group regardless of whether they retire or not. Changing Images 13 In sum, the above studies, and the data indicating that a majority of retirees report being satisfied with retirement, have been taken to indicate that the retirement as life crisis model should be discounted, IVIany of these studies, however, rest upon the assumption that the life satisfaction questionnaires adequately measure psychological well-being. Unfortunately, there are several limitations related to the use of the life satisfaction measures. First and foremost is the question of whether these measures adequately tap the concept of psychological well-being. The validity of life satisfaction measures has not been explored beyond a rudimentary level (Lohmann, 1980), From its definition, well-being would appear to be a multi-dimensional concept. Using a single-item to measure this appears unwise, Crowley (1985) has argued that single measures of well-being are inadequate to represent the dimensions of the concept. It would appear more reasonable to obtain separate assessments of depression, anxiety, and other domains of well-being and to find out about an individual's meaning and purpose in life. In his study. Bell (1975) defines life satisfaction as "the phenomenal experience of pleasure, with self and others, relative to past or present social circumstances. In essence, satisfaction represented a statement of personal morale with respect to time and place" (Bell, 1975; p, 158), He then goes on to say "life satisfaction was assessed by means of a single item," This appears somewhat simplistic. Second, the reliability of life satisfaction measures has been questioned (Lohmann, 1980), Third, Bortner and Hultsch (1970) indicate that there is a tendency for people to report on their satisfaction with their life as a whole rather than report on their current life situation. Thus, even if experiencing current difficulties, an older adult may report high levels of life satisfaction because of having earlier achieved certain life goals; and as a consequence, the meaning of the measurement instrument to the research participant differs from the interpretations made by the researcher. Fourth, it is Changing Images 14 not clear, from a theoretical stance, why retirement should necessarily affect general well-being. It is more reasonable to suppose that it will impact on specific domains of functioning such as leisure activities, time structure, the quality of the marital relationship, and personal meaning. Some theorists have advocated the need to adopt domain specific measures in addition to the general measures (Crowley, 1985), Taking a somewhat different approach, Parnes and Nestel (1981) asked 1,113 men whether they were disappointed by retirement. The men were port of a larger longitudinal study of 5,020 men age 45 to 59 initially interviewed in 1966, The questions regarding retirement were asked 10 years later in 1976 of those men who remained In the study and had retired. The majority of respondents indicated that they were not disappointed, although 19% indicated that they were disappointed. When those who had taken voluntary retirement were asked if they could do it over again, whether they would make the same decision to retire, 19% indicated that they would not, they would instead choose to continue working. Again, it appears that a certain proportion of retirees are dissatisfied with their retirement experience. In sum, the life sctlsfaction measure has been used as the primary means of examining the retirement as crisis theory. Reviewers have dismissed the theory even though there are a series of methodological problems that make it difficult to draw conclusions regarding the theory. These problems include inappropriate use of cross-sectional designs, lack of control groups, and an over-reliance upon problematic, global measures of life satisfaction to determine psychological well-being. Despite these methodological weaknesses, however, it seems fair to say that although a number of people report satisfaction with retirement, there remains a portion that are dissatisfied and do not fare as well with retirement. Changing Images 15 Individual Differences in Retirement With the crisis theory fading from the forefront as a research issue, an alternative has come forward. The research direction appears to have evolved into identifying predictors of subjective well-being following retirement. The question, then is "what are the predictors of a positive or negative outcome in well-being after retirement?" Underlying this question is the assumption, made explicit by Riddick (1982), that variables of research interest should be those that are mutable, that is, changeable. Mutable variables have the potential to serve as targets for future interventions although interventions in retirement have rarely followed from this research. That issue aside, the research in retirement in the last decade has been primarily focused on the above research question with the goal of identifying predictor variables. Two sets of variables have been the primary focus of psychologists in this area. The first set of variables relates to attitudes towards work. The second set consists of psychosocial variables. Benner (1984) in her phenomenological exploration, has indicated that workers have a variety of attitudes towards their work; with the level of commitment toward the work varying considerably. Researchers have speculated that job commitment contributes to determining satisfaction with retirement. Specifically, Glamser (1981) found that having a negative attitude toward retirement was predicted by missing one's job, and In turn, high job commitment was related to low life satisfaction following retirement. The inference Is that people who are highly committed to their work will have a difficult time adjusting to retirement, which will in turn affect their well-being, Hooker and Ventis (1984), on the other hand, suggested that the need for fulfillment through work could be met even after retirement If retirees perform activities that they perceive to be useful. Hooker and Ventis asked 76 retirees to keep track of the number of activities they performed each day for 7 days and Changing Images 16 the usefulness of each activity, In addition, the Protestant Work Ethic questionnaire (Mirels & Garrett, 1971) was used to assess job commitment. Correlations revealed that the number of activities was related to global satisfaction using the Life Satisfaction Index-A (Neugarten, Havighurst, & Tobin, 1961). With a measure of satisfaction that was more specific for retirement, the Retirement Descriptive Index (Smith, Kendall, & Hulin, 1969), however, a different pattern emerged. High job commitment predicted high satisfaction when the retiree was performing tasl<s judged to be useful, but predicted low satisfaction when the retiree performed tasl<s that were perceived as low in usefulness. The reverse pattern was found for individuals with low job commitment. More recently, conceptual models of retirement have been developed. These models have assigned an important role to psychosocial variables. One of these models was illustrated by Schnore (1985) in a study of the responses of 750 male and female workers and retirees to a series of questionnaires. Schnore found that well-being was predicted by low expectations, a positive evaluation of life in general, and a strong sense of self-competence. The retirees reported higher levels of satisfaction and lower expectations than the workers. In his model, past experiences contributed to perceived competence, which in turn contributes to observed behaviour, thus generating a dynamic interaction of internal variables and external events and behaviours. In addition to acknowledging the complexity of Individual differences in life satisfaction, researchers in this area have been paying greater attention to measurement issues. Floyd et al. (1992) examined the reliability and validity of a 51-item Retirement Satisfaction Inventory. The measure performed well in test-retest reliability and concurrent validity with other measures of life satisfaction, The use of a measure with good psychometric properties would enhance the Changing Images 17 work of researchers interested in assessing individual differences in retirement satisfaction, Personal attributes are assigned an important role in a conceptual model of psychological well-being in retirement developed by Crowley (1985), She suggests that pre-retirement well-being, personal attributes, work attributes, decision to retire, and both pre- and post-retirement resources interact and contribute to post-retirement well-being, Crowley examined some of the predictions of her model in a longitudinal study of men tested before and after retirement. She found strong evidence for individual differences in retirement, with health and financial situation contributing to well-being. In addition, indicators of well-being prior to retirement were significant contributors to well-being after retirement. The factors that are related to reports of low life satisfaction in retirement are low life satisfaction prior to retirement, a lack of self-competence, high expectations, poor health and financial status, high job commitment if performing retirement tasks low in usefulness, and low job commitment if performing retirement tasks high In usefulness. Although the demographic variables do not fit into the category of mutable, the psychosocial variables do fit. In particular, the sense of self-competence and expectations about retirement have the potential to be changed and thus serve as targets for intervention, In an effort to Identify appropriate target domains for interventions, Fretz, Kluge, Ossana, Jones, and Merikangas (1989) examined retirement-related anxiety and depression in employees eligible for retirement. The purpose of the study was to determine what variables were related to retirement anxiety and depression. Anxiety and depression were measured using the anxiety and depression scales of the Multiple Affect Adjective Checklist (Zuckerman & Lubin, 1965), modified by asking participants to think about retirement when responding, Changing Images 18 Information about five other classes of variables was gathered from each of the 139 participants. The first set of measures was demographic information including income, gender, age, and education. The second set was a 26-item questionnaire tapping l<nowledge about ageing and retirement modified from Palmore's (1977, 1981) Facts on Aging studies. The third area of interest in the Fretz et al. (1989) study, was planfulness and attitudes specifically with regard to retirement. Planfulness was assessed with a single item for eight domains relevant to retirement planning, Health and social support were assessed using self-report measures. The final domain assessed was self-efficacy, a sense of competence in dealing with the retirement situation. To measure this, participants were asked to assess their competence in adjusting to retirement and then to rate their confidence that they would adjust successfully. Using correlational analysis, Fretz et al. found that low levels of self-efficacy were related to increased levels of anxiety and depression, suggesting that interventions related to coping with retirement should focus on mastery experiences related to retirement problems. They point to some targets that are accessible for intervention, specifically self-efficacy. Although not explicitly addressing the issue of mutability. Hooker (1991) takes an idiogrophic approach to examining how the self changes as part of the experience of retirement. She focused on the evaluation of the self made daily by four participants over the course of four months during their initial transition to retirement. The participants daily filled in two measures of self-concept, the Index of Adjustment and Values (Bills, Vance, & McLean, 1951) and the Self-Sentiment factor of the 16PF. Hooker then used p-technique factor analysis to examine stability and variability over time in each participant's responses. She found that three participants showed considerable stability in their responses. For each participant, however, there was some variability consistent with changing events Changing Images 19 in their lives. For example, one participant had looked forward to being able to relax in her retirement. She showed variability in a factor relating to serenity. After her retirement, the serenity factor increased. The fourth participant had very little stability in his responses. This was again consistent with events in his life. He was dissatisfied with his retirement and had not established a routine. Hool<er's research suggests that the experience of the retirement transition, and the relationship of retirement to the self, differs for each person. Gender Issues Prior to 1976, the overwhelming majority of studies on retirement had been conducted using male participants (Szinovacz, 1982). The reason for this bias is quite clear; the majority of people retiring from the worl<-force at this time were male, with the exception of certain groups of women in lower SES occupations, such as domestic staff. With changes in the composition of the work force, women are now retiring in greater numbers. Thus, gender bias in the research on women's retirement became an issue in the literature (Szinovacz, 1982), along with the underlying assumption that women's retirement may be unique. Certain differences between men and women have been put forward as reasons why retirement may differ for women. These include different occupational histories, society's attitudes towards women and work, different family roles, increased life expectancy, and higher poverty levels (Perkins, 1992; Szinovacz, 1982). In short, there is some reason to believe that gender may be a relevant issue in the study of retirement. In this section, I examine how gender has been studied with respect to retirement. Several authors have called for an increase in the number of studies on women's retirement to address the question of "what is women's experience of retirement?" (Keith, 1982; Price-Bonham & Johnson, 1982; Szinovacz, 1982). Riddick (1982) made an attempt to answer this question by developing a causal Changing Images 20 model of life satisfaction in older women. On the surface, Riddicl< was asking what factors predict scores on a life satisfaction questionnaire, but in her conclusion she used her study to make inferences about the retirement experience. She asked 753 women over the age of 65 to respond to a series of questions about income; whether they were retired or employed; self-perception of time spent reading, socializing and doing volunteer work (the scale was "hardly any", "some but not a lot", "a lot"); and how serious a problem they hod with health and public transportation. In addition, life satisfaction was assessed using an 18-item version of the life satisfaction index (Wood, Wylie, & Sheafor, 1969). Unfortunately, there are problems with the questionnaire in that it used a somewhat biased phrasing by assuming that problems with health and transportation would exist. In addition, there are no clear anchor points; people are left to judge whether 2 hours a day of reading is a lot or hardly any. Judgements of this kind may be made using criteria that vary from person to person. Putting these concerns aside, however, Riddick concludes from the study that retired women had lower satisfaction scores than employed women, and that activity level, income, and health were the strongest predictors of satisfaction scores with high activity level, high income, and better health being associated with high satisfaction scores. Higher income was related to being employed, suggesting one explanation for lower satisfaction scores for retired women. Seventy-five percent of the variance, of the life satisfaction scores remained unaccounted for, indicating that there are other contributing variables to life satisfaction. Furthermore, at this point the earlier critique of life satisfaction questionnaires as a measure of well-being should be recalled. In short, this study tells us little about the retirement experience of women. Block (1982) also examined predictors of retirement satisfaction; her study, however, differed from Riddick's in several ways. Block used a more extensive Changing Images 21 measure of retirement satisfaction, the 63-item Retirement Descriptive Index (Smith, Kendall, & Hulin 1969) that asks aPout evaluations of specific situations relevant for retirement, The women in Block's study were all professional women with high status. Health, social resources, income, and pre-retirement planning were assessed using questionnaire responses. Good health and high income after retirement were predictors of satisfaction with retirement accounting for 36% of the variance in the Retirement Descriptive Index scores. Due to its specificity, the Retirement Descriptive Index provides more information aPout the retirement experience than the Life Satisfaction Index used Py Riddick. Interestingly, participation in a pre-retirement programme was also a predictor of high life satisfaction, Atchley (1982) examined the retirement experience of men and women further, Py mailing questionnaires regarding health, demographics, self-confidence, and attitudes towards work and retirement to all adults over 50 years of age in a small community in the United States. He had a 73 % return rate to his questionnaires. He explored differences in the responses of men and women and found that employed women were more likely than men to indicate that they would never retire and yet those women who were retired reported higher levels of satisfaction than men. Both men and women had generally high levels of satisfaction with life after retirement. Health and financial status were the Pest predictors of satisfaction scores; with people In financial difficulties and poor health tending to have lower scores, Unmarried women were disadvantaged with respect to income. As with other studies using life satisfaction questionnaires as the outcome measure, I find it difficult to generalize the results of these studies to my understanding of well-Peing. It appears that all people report high level of satisfaction unless something oPvious is wrong with their lives, such as poor health Changing Images 22 or financial difficulties. In some ways, the assessment of global satisfaction tells little about day-to-day struggles with boredom, worries about physical attacic, loneliness, family squabbles, or cognitive appraisals of threat, coping, and stress, A different approach is fallen by Jewson (1982). She attempted to describe the inner experience of retirement for professional women. In order to do this she conducted semi-structured interviews with 32 retired, female professionals. Open-ended questions were used in addition to quantitative measures of retirement satisfaction. Comparison was made to interviews with 16 retired male professionals and 14 retired female non-professionals, Jewson reports that a descriptive analysis was made of the interview results but does not indicate how this was done. Nor does she Indicate the training of the interviewers. As a result, it is difficult to evaluate the methodology, and consequently it is difficult to draw conclusions on the basis of the results. Narrative comments indicated that some of the professional women were able to retain feelings of usefulness by maintaining continuity with their professional roles through volunteer activities. Both groups of professionals indicated that they made creative use of their time. In addition, professional contacts were maintained by the women to a greater extent than by the men. The professionals spoke about personal philosophies regarding the meaning of their retirement. Interestingly, many of the people in the study reported regret at not participating in retirement planning programmes. It is not clear whether this response is the result of the use of leading questions or whether despite indicating high level of satisfaction, the retirees felt that there was still room for improvement in their adjustment to retirement. Retirement Preparation Programmes Retirement preparation programmes have been primarily offered as an employee benefit through large companies. The primary purpose of these Changing Images 23 programmes is to offer information about pension plans (Dennis, 1986), rather than to explore feelings and concerns about retirement problems. Information-oriented models primarily employ a presentation format in which an instructor presents the material and answers questions. Tiberi, Boyack, and Kerschner (1978) have outlined three other formats in addition to the presentation format; these are the interaction format, the discussion format, and the individual model, The interaction format uses group interaction and role plays to explore retirement problems. The discussion format involves the presentation of material followed by discussion among the participants. Finally, the individual format consists of allowing individuals to have access to resources such as books and videotapes. Tiberi et al. (1978) evaluated these formats by assigning participants to one of each of the formats and to a control group, and comparing performance of participants on questionnaires of behaviour, information, and attitude. The behaviour measure consisted of questions about activities taken to promote health, financial planning, and community involvement. The information measure assessed knowledge of health, financial, and emotional issues; and the attitude measure assessed anticipated enjoyment of retirement, rejection of retirement stereotypes, and self-worth. IVlultiple regression revealed several significant effects, with participants in the interactive and discussion groups responding more positively on the behavioural questionnaire than participants in the other groups, Participants in the interactive and presentation groups had higher retirement knowledge scores. The results for the attitudes measure were less clear, with participants in the interactive and discussion groups showing change only in the rejection of stereotypes portion of the questionnaire. Despite the suggestion provided in the study by Tiberi et al. (1978) that an interactive format is most effective in producing change In a variety of domains, most retirement preparation programmes have employed a presentation format. Changing Images 24 In an effort to determine the effectiveness of a retirement programme for teaching people retirement information, Kamouri and Cavanaugh (1986) examined retirement information in an evaluation of a pre-retirement programme that lasted for 6 weeks and used a combination of presentation and discussion formats, Comparison of attendees with a wait list group using analysis of variance revealed that knowledge of retirement, assessed by a modified version of the Retirement Descriptive Index (Smith et al„ 1969), was higher in the programme participants. This suggests that pre-retirement preparation programmes can be effective in their goal of providing information. The impact of these programmes on behaviour and retirement adjustment, however, has received little in the way of rigorous evaluation as many of the evaluations have had serious design weaknesses (Palmore, 1982). One exception is a study by Turnquist (1986) which examined participation In retirement preparation courses offered by five companies to their employees, The primary focus of the courses was on financial information and a presentation model was used, Turnqui§t used the Retirement Descriptive Index (Smith et al„ 1969) to assess satisfaction with four domains of life: finances, health, activities, and friends. T-tests Indicated that participants in these programmes did not report higher retirement satisfaction levels in health, activities, or friends than retirees whose companies did not offer the retirement preparation programmes. Participation in the programme did appear to increase satisfaction with finances, Durrant (1988) had a somewhat different approach to evaluating retirement preparation. She examined the effectiveness of a retirement preparation programme not for providing information, but instead, for reducing anxiety and increasing retirement planning in women, Male and female participants attending a weekend workshop on retirement were divided into two groups and were asked to complete three questionnaires. The first group Changing Images 25 completed the questionnaires prior to participation in the 2 day workshop and the second group completed the questionnaires after participating, Anxiety was assessed using the A-state form of the State-trait Anxiety Inventory (Splelberger, Gorsuch, & Lushene, 1970). In addition, assessment was made of retirement knowledge using the Retirement Knowledge Index (Glamser & DeJong, 1975) and of retirement preparation behaviour using a check list of activities and a time frame of intention to complete each activity. The workshops consisted of 2, 7-hour days of lectures, presentations, paper and pencil tasks, and small group discussions. Although Durrant does not indicate the number of participants per workshop, she had a total of 129 participants and four workshops, leaving an average of 32 participants per seminar. The workshops focused primarily on financial planning and pension benefits (6 out of 10 sections) and included sections on role changes, living arrangements, leisure, and health care. Using t-tests, Durrant found that for men, anxiety levels were lower in the group measured post-workshop relative to the group assessed pre-workshop, Workshop participation, however, did not appear to reduce anxiety levels for women. Retirement knowledge was higher for both genders following participation, although preparation behaviours did not appear to change. A breakdown of the subjects into demographic subgroups, however, revealed that preparation behaviour was reported to be higher following the workshop for those women who were single, in managerial positions, or planning an earlier retirement (there were no differences between groups on planning prior to the workshop). This study raises a number of interesting issues regarding why the workshops were not effective in reducing anxiety for the women. In some ways, it is not completely surprising that the workshop did not reduce anxiety for all Changing Images 26 participants, as it appeared to focus on providing information rather than targeting anxiety reduction. Further, retirement preparation programs for women have been criticized for falling to place the retirement process for women in soclo-hlstoric context in order that an atmosphere can be created in which women can explore their life choices (Meade & Walker, 1989). Creating such an atmosphere may be easier in groups of women. Thus, allowing women to explore their anxiety and plans in groups with other women may be more effective than with groups of men and women, Poser and Engels (1983) offered their retirement preparation programme targeting self-efficacy to women only, The programme used an interactive format and involved exposure to audio tapes of problematic retirement scenarios in which some of the characters in the taped material responded positively, taking control and being creative. Thus, the taped material modelled constructive cognitions and behaviours. Role playing and rehearsal of the scenarios was included in the programme along with discussion of retirement experiences and plans. Self-efficacy was assessed prior to the programme and at 3-months follow-up by asking participants to rate their expectations regarding their ability to perform nine retirement related tasks. Analysis of the results of the self-efficacy assessments using analysis of variance indicated that participation in the training programme Increased self-efficacy relative to a wait-list control group on two of the retirement tasks (keeping busy and total efficacy with retirement), As Poser and Engels themselves point out, their study is suggestive rather than conclusive regarding the ability of the counselling techniques to increase self-efficacy with regard to retirement tasks. The participants of their study were all volunteers, who had not reported prior difficulties coping with the transition to retirement. It is more than possible that these participants had little Changing images 27 more to learn about the retirement experience, and consequently could be expected to gain little in the way of increased efficacy scores. It appears that some retirement preparation programmes can be effective in increasing l<nowledge about finances and satisfaction with one's financial situation. Others, such as the one offered by Poser and Engels (1983) may be more effective at increasing self-efficacy and thus have the potential to increase coping and decrease retirement anxiety. Summary Previous research on retirement indicates that although many retirees report being satisfied with their life, there remains a proportion that are dissatisfied, and indeed, even regret their decision to retire. Poor health and poor financial situation are associated with these lower levels of life satisfaction scores, as are certain psychosocial variables such as unrealistic expectations and a low sense of self-competence and self-efficacy. These psychosocial variables have the potential to serve as targets for interventions designed to reduce anxiety and depression surrounding retirement, The study by Poser and Engels (1983) found that self-efficacy was increased following a week-end seminar involving small group discussion, modeling, and role play of situations requiring competence and mastery. Pre-retirement programs have been developed to prepare people for retirement. Typically, these programs are provided by companies for their employees and are designed to provide information on the financial aspects of retirement. Evaluations of these programs have shown that they are effective in teaching people this information, but are less effective in helping participants cope with retirement stress, particularly for women. One reason why the needs of women may not be met with the standard format is that typically these programs are offered to a mixed group of men and women, and the women Changing images 28 may feel less comfortable discussing their issues in these groups, Further, women are more likely to have a shared history of career disruptions and of coping with additional non-work roles such as being a mother. Allowing women to meet with other women may help to create an atmosphere where they con explore their relationship to the cultural and social climate surrounding both women's participation in, and retirement from, the work-force. Counselling Approaches to Retirement Preparation Proarammes In the stress and coping model outlined by Lazarus and Folkman (1984), the central components are the appraisals made by the individual of both the situation and of personal resources. Unlike information-oriented presentation formats, counselling approaches can more directly target appraisals and expectations about future possible selves. Two complementary theories that fit within the Lazarus and Folkman framework are Meichenbaum's stress inoculation training (Meichenbaum & Cameron, 1983) and D'Zurilla's problem-solving model (D'Zurilla, 1988). Stress inoculation training targets the feelings of stress and problem-solving provides a method to deal directly with finding problem solutions, Although the name, stress inoculation, implies a focus on stress, it has been applied to a variety of emotional disorders, including anger and pain. The thrust of the approach is to train clients in a variety of skills designed to ameliorate their emotional discomfort in future situations (Meichenbaum & Cameron, 1983), This is achieved from the base of a collaborative, therapeutic relationship. As a result of the future orientation, it is particularly appropriate for targeting changes in possible selves. There are three phases to the therapy in the stress inoculation model, The first phase is the assessment phase in which the therapist assesses the client's situation and teaches assessment skills to the client to enable the client to monitor progress on an on-going basis. Assessment provides a detailed analysis of the Changing images 29 client's problem, including relevant thoughts feelings and behaviours. In addition, the client is provided with an explanatory model of stress reactions and on outline of therapy. From this point the therapy moves to phase two, in which specific sl<ills are trained and rehearsed. The list of skills that can be dealt with at this point is comprehensive, ranging from emotion-focused relaxation, to problem-focused assertiveness training. In practice, however, the emphasis appears to be on coping with the emotional experience of stress. Relaxation, self-statements, and attention diversion are all designed to alter the client's perception of, and response to, a stressful event rather than taking action to effect change within a problem situation (although Meichenbaum and Cameron argue that relaxation skills are a form of taking action). Assertiveness skills operate more on the action end. Problem solving skills have the potential for a more instrumental direction. Regarding these skills, Meichenbaum and Cameron refer therapists to D'Zurilla's problem-solving theory. Phase three involves guiding clients to put the skills into action through use of in-session anticipatory imagery, role plays, and graded in-vivo stressors. The goal with this phase is to aid clients in generalizing these skills from the therapy context to other contexts in their lives. Jaremko (1979) has argued that cognitive restructuring is one of the critical components of stress inoculation. In their review, IVIatheny et al. (1986) found that cognitive restructuring and relaxation were effective in reducing stress and recommended that these techniques be taught in situations where the problem could not be directly resolved. In brief, the stress inoculation model fits into the stress and coping framework due to the emphasis placed upon the individual's appraisal of both the situation and of skills and ability to cope with the situation. There are several points at which Meichenbaum's model goes beyond the stress and coping framework, by virtue of the therapeutic aspects. For example, the emphasis on Changing images 30 the therapeutic relationship that is part of stress inoculation is not referred to explicitly in the Lazarus and Folkman framework, although it is still consistent with the framework. Problem-solving therapy Is also consistent with the stress and coping framework, in that it addresses problem-focused coping skills that are primarily relevant for those situations over which the Individual has some control, D'Zurilla (1988) places an emphasis on training clients in the skills needed to solve a variety of problems. Whether a situation is defined as a problem or not, depends upon the Interaction of the Individual, the situation, and the individual's appraisal of the situation, D'Zurilla draws attention to the similarity in his model of the relationship of individual to environment to the model of Lazarus, Therapy is accomplished by teaching a 5-stage model to clients. The first stage is problem appraisal which requires that a problem is identified and that an orientation toward solving problems be adopted. The second stage involves clarifying the problem and goals and gathering information, Alternative solutions are generated in the third stage and the choice of solution made in the fourth stage. In the final stage, the solution is put into action and evaluated for effectiveness. Outcome studies with both problem-solving and approaches similar to stress inoculation have indicated that these are effective with older adults in increasing social skills (Petty, Moeller, & Campbell, 1976; Toseland, 1977), In a meta-analysis of interventions designed to reduce stress, Matheny et al. (1986) found that interventions that addressed problem solving had a higher effect size (ES=,74, in combination with other treatments) than other forms of treatment, including relaxation, cognitive re-structuring, and stress monitoring for reducing stress and anxiety (ES =,61, ,61, and ,62, respectively, in combination with other treatments), Changing images 31 These counselling approaches are particularly relevant for adults who are anxious about retirement for at least three reasons. First, the effectiveness of these techniques has been demonstrated with older adults (Petty et a!,, 1976; Toseland, 1977). Second, both approaches have a future orientation that is particularly relevant for people concerned with future retirement and people who are anticipating problems in the future. Thus, the approaches work to alter the images that participants have of themselves in the future. Third, the techniques are designed to reduce anxiety and depression by promoting problem-focused coping with those aspects of the environment that are controllable, and promoting emotion-focussed coping for those aspects that are not controllable. Possible Selves One of the mechanisms that may be important in contributing to change is the set of images that clients have of themselves in the future. These images have been labelled possible selves by Markus and Nurius (1986), They contain both hoped for, and feared, future selves, A woman about to retire, for example, may have images of herself as the matriarch of her family, being the president of a charity organization, and being a cheerful companion to her spouse on their travels abroad. In addition, she may have images of being a lonely widow, being abandoned by family and friends, and turning cantankerous and spiteful, Markus and Nurius (1986) proposed that possible selves are important motivators of behaviour and can contain information about how to achieve a desired end-state. The image of being president of a charity, for example, contains information about being visible within the organization, networking, attending meetings, and running for president. These images of self in the process of becoming an end-state can promote persistence in the face of obstacles and encourage delayed gratification (Oyserman & Markus, 1990), In Changing images 32 addition, possible selves can contribute to self-efficacy expectations and perseverance, Feared selves can also contribute to motivating behaviour if they contain information about how to avoid a certain outcome. The woman who fears being alone may know that she can avoid this outcome if she actively maintains her social network. In this way, possible selves can contribute to the change process. Moreover, the building and strengthening of appropriate possible selves can be used in counselling for points of leverage to promote change (Nurius, 1986), The literature on possible selves is a recent one, and has primarily been examined from a conceptual point of view. Research, on possible selves, unfortunately, has lagged behind somewhat, possibly due to difficulties in measuring people's possible selves (Nystedt, Smari, & Boman, 1991), Nurius (1986) has suggested that the aspect of possible selves that should be assessed is the working future-self concept, that subset of possible selves that is most functional and likely to be used in guiding behaviour and processing information about the self, Markus and Nurius (1986) described a method that could be used for measuring possible selves. They developed a list of 150 items from the responses of college students to an open-ended question asking them to describe possibilities for the self. The items were classified into six categories, including personality descriptors, physical descriptors, and occupational descriptors. One third of the items in each category were rated as neutral, another third as positive, and a final third as negative, Markus and Nurius then asked 210 college students to indicate whether each item had described them in the past, whether it described them in the present, whether they had ever considered it in a possible self, and whether it was probable that the item would describe them in the future. They found a high correlation between past and present selves (r = Changing images 33 ,68) and a much lower correlation between present and possible selves (r = ,21) Indicating that the students' possible selves differed from their present selves. Interestingly, a higher number of students reported a possible self of being physically paralyzed than reported a possible self of having wrinkles, perhaps indicating something about students attitudes towards ageing. In addition, the students completed standardized measures of positive affect, negative affect, locus of control, esteem, and optimism. These were each regressed in a multiple regression with the measures of past, present, possible, and probable selves. All of the measures of self, including the possible and probable selves, contributed significantly to the models. This indicates that each aspect of the self can contribute to affective state. One study examining the relationship of possible selves to behaviour was conducted by Oyserman and Markus (1990), The domain of behavior that they were interested in was delinquency, however, their study is relevant here because of the methodology that they employed to study possible selves. They studied 141 youths from four groups classified according to level of delinquent behaviour. Two methods were used to assess the possible selves of the youths, In the open-ended procedure, the participants listed three each of their hoped-for selves, their expected selves, and their feared selves. Responses were classified by coders into content categories such as references to school performance, crime, negative selves, and positive selves. Analysis of covariance revealed that the content of the possible selves differed significantly across the levels of official delinquency. Youths in the more delinquent groups were less likely than the less delinquent youths to make references to good school performance in their expected or hoped for selves. In addition, poor school performance figured less prominently in their feared selves than it did for the less delinquent youth. Changing images 34 The closed-ended questions employed a technique similar to that used by Marl<us and Nurius (1986), Participants were asked to rate a series of items for how well the items described them in the present, how well the items will probably describe them in the future, and how much they would lil<e the items to describe them in the future, Youth in the more delinquent group were less likely than the less delinquent youth to endorse positive items in their expected selves and more likely to endorse negative items as descriptions of their current and expected selves. The results of this study indicate that there is a relationship between possible selves and behaviour, although, as Oyserman and Markus point out, their study is not designed to assess a causal relationship. They suggest, however, that constructing a possible self of achieving a certain end-state is necessary for a person to engage in behaviour leading to the desired outcome. The same open-ended measurement approach used by Oyserman and Markus (1990) was also used by Carver, Reynolds, and Scheier (1994) to explore the relationship of possible selves to optimism/pessimism, They found that the possible selves generated by optimistic participants were more positive than those generated by pessimistic participants. Ratings for possible selves, both hoped for and feared did not, however, differ for optimistic and pessimistic participants. The number of items that participants can list is limited in this methodology, and it was not possible for the researchers to assess whether there were a greater number of selves generated for hoped for or for feared possible selves for either optimists or pessimists, The measurement approaches that Oyserman and Markus (1990) employed to assess possible selves are interesting, and provide information about how the participants in the study perceived themselves. Neither the closed-ended nor open-ended approaches, however, fulfill the mandate of Nurius Changing innages 35 (1986) to tap the worl<lng-set of possible selves. The closed-ended approach provides only a limited, pre-determined set of stimulus items for participants to endorse. The open-ended approach that was used forced participants to give three images for each of the hoped for and feared possible selves, and thus, disregards the relative strength of the images. It would be more indicative of the working set if participants were allowed to freely describe their future selves, whether hoped for or feared. This would also allow for an examination of the balance between the positive and negative possible selves. Indeed, it has been argued that it is the balance of positive to negative thoughts, rather than the absolute level of either that Is important for overall mental health (Schwartz, 1993). Schwartz and Garamoni (1989) have developed a ratio, called the States of Mind (SOIVI) ratio for assessing the balance of positive to negative thoughts. A more open-ended approach to measuring possible selves was used by Cross and Markus (1991) to assess differences in possible selves between older and younger adults, participants were asked to list feared and hoped-for possible selves and then to indicate both their capability of accomplishing each and the likelihood of each becoming true on a 7-point scale. The items were coded into 11 categories such as family, occupation, and physical. Cross and Markus found that participants in the older age groups generated fewer possible selves than participants in the younger age groups. In addition, participants in the older age groups generated fewer hoped for selves in the occupation category than younger age groups and more in the life-style category. This particular approach to gathering possible selves allows participants to respond freely within each of the categories of hoped-for and feared possible selves but does prompt participants to think of these two polarities. Ryff (1991) explored differences between older, middle-aged, and younger adults on perceptions of present, past, ideal, and future selves. She used Changing images 36 a closed approach in which participants rated each of the four aspects of thennselves on six dimensions of well-being; self-acceptance, positive relations with others, autonomy, environmental mastery, purpose in life, and personal growth, Ryff found overall, that older adults perceived stability between past and present selves, Younger and middle-aged adults perceived improvement from past to present. With regard to future selves, older adults foresaw decline for themselves. Younger and middle aged adults saw improvements. The discrepancy between present and ideal self was less for older adults than for younger and middle aged adults. With the use of the closed-ended technique, however, it is not possible to make conclusions about whether the research taps the working self. The role of counselling in changing Images of possible selves was examined by Nurius, Lovell, and Edgar (1988). They asked 11 mothers of abused children to generate words that describe positive and negative parenting. The participants each chose eight negative words that described the aspects of themselves that they wanted to change and eight positive words that described their parenting goals. Participants then rated the words using a similar approach to the closed-ended method employed by Oyserman and IVIarkus, for how much the words described them in the present, and how likely they were to describe them in the future. The mothers then participated in a 26-week group-counselling programme involving stress, anger management, child management, self-talk, assertiveness, and problem solving. Following the programme, participants rated the self-generated items again, Nurius et al, do not indicate the type of statistics that they employed but reported that the results were significant in indicating that following the intervention, mothers rated fewer of the negative items and more of the positive items as descriptive of themselves in the present. The results for the ratings of expectations about the future were not significant but suggested Changing images 37 that both positive and negative items were ranked as more likely to describe the mothers in the future at the post-test than at the pre-test, The self-generation of items in the Nurius et al, (1988) study is an improvement over providing a pre-determined set of items for participants, Unfortunately, it fails to take into account that the range of items used to describe the self may change following intervention. In addition, the argument remains that this approach does not tap the working future-self concept that participants may have. It is not particularly surprising, however, that methodological difficulties are present in the research on possible selves given the limited number of empirical studies in this area. The concept itself offers considerable promise for increasing our understanding of the change process. Overview of the Study and Hypotheses The goal of the present study was to determine whether images of positive and effective possible selves in retirement along with a sense of self-efficacy could be built and strengthened by participation in a stress inoculation training and problem-solving programme. In addition, the effectiveness of the programme for reducing retirement-related anxiety and depression was determined. To examine these questions, the counselling approaches of problem-solving and stress inoculation were combined around retirement issues in an 8-week programme. The programme was targeted at women who were anxious about retirement. The effectiveness of the cognitive behavioural targeted-change programme was compared to that of a wait-list control group and a structured-discussion group in which participants met for the same length of time but were not taught the problem-solving and self-monitoring skills. The structured-discussion group was Included in the design of the study in order to determine the effectiveness of the active cognitive components of the treatment by controlling Changing innages 38 for non-specific factors such as treatnnent expectations, structured activities, homeworl<, and peer group support, Hypotheses Question One. The targeted-change group ennpioyed counselling techniques designed to decrease anxiety and depression, in order to confirnn the effectiveness of this treatment approach, question one asi<s whether anxiety and depression concerning retirement could be reduced following participation in a pre-retirement programme. 1. It was predicted that from pre- to post-treatment, participants in the targeted-change group and the structured-discussion group would show a greater reduction in (a) anxiety scores and (b) depression scores; compared with participants in the wait-list control group. In turn. It was predicted that from pre-to post-treatment, participants in the targeted-change group would show a greater reduction in (a) anxiety scores and (b) depression scores, compared with participants in the structured-discussion group. The differences between the treatment and structured-discussion groups were predicted to be maintained at follow-up. Question two. It was asked whether images of positive and effective possible selves could be built and strengthened by participation in the stress inoculation and problem-solving programme. Participation in a skills-oriented retirement preparation workshop was expected to have a direct impact on the concept of self in retirement, creating and reinforcing a "possible self" that is effective and competent. 1. It was predicted that from pre- to post-treatment, the participants in the targeted-change group and the structured-discussion group would show a greater increase in (a) effective possible-self items and (b) positive possible-self items, compared with participants in the wait-list group. In turn, it was predicted Changing images 39 that from pre- to post-treatment participants in the targeted-change group wouid show a greater increase in (a) effective and (b) positive possibie-seives, compared with participants in the structured-discussion group, The differences between the treatment and structured-discussion groups at post-test were predicted to be maintained at foiiow-up, 2, The above pattern was predicted to be repeated for the Retirement Seif-efficacy score, 3, it was predicted that from pre- to post-treatment, there wouid be a decrease in negative possible-seif items for participants in the targeted-change group and the structured-discussion group, compared with participants in the wait- list control group. In turn, from pre- to post-treatment participants in the targeted-change group would show a decrease in negative possible-self items, compared with participants in the structured-discussion group, The differences between the treatment and structured-discussion groups were predicted to be maintained at follow-up. 4, Effectiveness in the possible selves measure was predicted to be positively and moderately correlated to Retirement self-efficacy scores for participants in the targeted-change group and the structured-discussion group at post-test and at follow-up. Changing Images 40 Chapter Three Method Participants Sixty-one employed women were recruited following stories about the research project in local newspapers and in union newsletters. The mean age of the women was 59.9, range 49 to 71, and they had a mean of 15.1 years of education. Fifty-eight of the participants had worked for more than 10 years. Participants met the following criteria; (a) they placed in the upper quartile of scores on the Spielberger State-Trait Anxiety Inventory, state version (Spielberger etal., 1983), (b) they wanted to participate In a programme dealing with anxiety about retirement, (c) they provided written consent to participate in the study, (d) they agreed to attend the 8-week retirement seminars and the follow-up session, and (e) they were not involved in other counselling programs at the time of the study. It was initially intended that the women meet the criterion of being less than 2 years from retirement in order to ensure that the women were engaged in the process of retiring. During the interviews it became clear that those women who had volunteered were quite engaged and anxious even when they reported being more than 2 years away from retiring and so this criterion was dropped. In addition, many participants were unsure about when they would be retiring. Forty participants reported being less than 2 years from retirement and 21 were more than 2 years from retirement. An additional eight participants were interviewed but were unable to participate in a group due to time or location conflicts. Twenty-eight further people responded to the advertising, but were not interviewed. Ten of these participants were interested in financial information only, two were male, seven were not interested in participating in a group, six were unwilling to travel to the groups, and three were already retired, Changing Images 41 IS/laterials Four questionnaires, a measure of possible selves, the Retirement Self-efficacy measure used by Poser and Engels (1983), a modified form of the state version of the Spielberger State-Trait Anxiety Inventory, and the short form of the Multiscore Depression Inventory (Berndt, 1986) were administered. Demographic information regarding age, marital status, and occupation were collected, in addition to information about participants expectations for treatment efficacy, information about behavioural responses to retirement preparation, and material that was used for a descriptive programme evaluation. Spielberaer State-Trait Anxietv inventory (STAI-SI Definitions of anxiety abound, with the distinction typically being made between state anxiety, an emotional state of tension with an accompanying physiological response, and trait anxiety, a personal characteristic or predisposition toward the emotional state of anxiety. Matheny et al. (1986) conducted a meta-analytic review of the literature on stress and concluded that the outcome measure that was used most extensively was the State-Trait Anxiety Inventory (STAI; Spielberger et al,, 1983). Use of this particular measure, then, facilitates comparison to other studies. Following instructions to researchers in the STAI manual, the state version of the STAI was modified by instructing participants to think about retirement prior to completing the questionnaire. The modified instructions are in appendix A. The state version of the STAI (STAI-S) is a 20-item questionnaire using a 4-point Likert scale (Spielberger et al„ 1983), Scores range from 20 to 80, with higher scores indicating higher levels of anxiety, The questionnaire has been used extensively in previous research, and there is ample data regarding reliability and validity. The four normative samples consisted of 1,838 working adults (ages 19-69), 855 college students, 424 high school students, and 1,964 military recruits. Coefficient Changing Images 42 alphas for the normative samples were high, ranging from ,86 to ,95 (Spielberger et al,) indicating high internal consistency, Test-retest reliability is, as expected for a measure of emotional state, low;, 16 and ,33 for females and males, respectively, in the college sample after one hour, ,27 and ,54 after 20 days, and ,31 and ,33 after 104 days. In the high school sample, test-retest reliabilities were somewhat higher; ,34 and ,62 for females and males after 30 days and ,36 and ,51 after 60 days. Overall, the median test-retest reliability is ,33. Validity has been investigated thoroughly. Regarding construct validity, the state version of the STAI is moderately correlated with the Trait version ranging from r = .59 to .75 across the four samples. Further, state scores have been compared under stressful and non-stressful situations revealing that the state scores were higher in the stress condition (Gaudry, Vagg, & Spielberger, 1975; Rule & Trover, 1983). Factor analysis has revealed a factor structure consisting of two Trait factors, negative and positive descriptors and two State factors, negative and positive descriptors (Bernstein & Eveland, 1982). Short form of the Multiscore Depression Inventory (SMDn Depression is the name given to a cluster of symptoms including sad mood and low energy (Berndt, Petzel, & Berndt, 1980). In this study, the short version of the IVlultiscore Depression Inventory (SMDI; Berndt, Berndt & Kaiser, 1984; Berndt, 1986) was used for three reasons. First, it provides more information than other more commonly used depression scales such as the Beck Depression Inventory (Beck, 1967) because, in addition to a global rating of depression, the SMDI provides nine subscaies that cover the individual symptoms of depression (Berndt et al,, 1980), These subscaies are low energy level, cognitive difficulty, guilt, low self-esteem, social introversion, pessimism, irritability, sod mood and instrumental helplessness. The Inclusion of a subscale specifically for low energy level may be particularly relevant for an older population. Second, the SMDI is more sensitive Changing Images 43 to lower levels of depression found In the normal population than the more clinically oriented scales (Berndt et al., 1984), Third, the scale has good psychometric properties (Lanyon, 1984). The SMDI is a 47-item questionnaire using a true-false format with a range of 0 to 47 with higher scores indicating higher levels of depression. In addition, separate scores are available for each of the nine subscales. The total score was used as the dependent variable for the planned contrasts in this research. Post-hoc analyses were conducted on the subscales. According to Lanyon (1984), the IVIultiscore Depression Inventory has been constructed with care and attention to psychometric properties. Regarding reliability, coefficient alphas have been reported for two samples (n = 108 and 133) ranging from .67 to ,87 for each subtest and .92 and .88 for the total score (Berndt et al., 1984). Immediate test-retest reliabilities are good, r = .95 for the total scale and .83 to ,95 for the subscales. Concurrent validity of the SIVIDI has been assessed. The short form correlates highly with the full form corrected for item overlap by removal of short form items, r (n = 162) = .91 (Berndt et al., 1984). In addition, the relationship of SMDI scores with the Beck Depression Inventory (Beck, 1967) has been examined, r(n = 131) = .63, and with the Depression Adjective Checklist (Lubin, 1967), r = .76 (Berndt, Petzel, & Kaiser, 1983). Factor analysis of the full scale reveals eight stable factors and this was confirmed in a second sample (Berndt, 1981). Retirement Self-efflcacv fRSEI Central to this study is the individual's appraisal of a future event, in this case retirement. One construct that is important in understanding how people perceive themselves in the future is self-efficacy, defined by Bandura (1977) as a belief that an individual has about his or her ability to behave in order to produce a given result. Self-efficacy is a sense of mastery, a sense of one's ability to be Changing images 44 competent in future situations. There are three dimensions to seif-efficacy; ievel, strength, and generality. Level refers to the individual's perception of the ? ^ n £. o standard of performance th ? confidence of the individua I refers to the specificity of th 1988). Typically, the first twc participants to indicate wh they can, to rate their coni hi and Engels (1983) modifiec ^ about competence and c g Engels (1983) Retirement J H . g ^ ft • - S g three domains; specificall' ^ activity management. Th H p o ^ are asl<ed to Indicate whi they were that they couk Examples include "I can p day," and "I can name a-The total score consists c and ranges from 0 to 10( efficacy. In addition, su domains. The total scor< contrasts. Post-hoc one developed with 35 fem( (1984) reports that scor Satisfaction Index (Neu w 3' D-w < ft ^ &-5' cro •fl o •n 0 H^ JJ; n' n G ft i: 3 lu EL b ft o O p O a> c lO 10 13 l~ IM I I I -1 ^ n I it> I3> ! • » 1-^ lo K I It) io lO i> |3 I S iTi |(o I I c • 1 l > IO 1?^  Im I z IM | i - i im M 13: I ICO I l;x; M l-l im M Im I l-n im 100 l-H l > 3 a s - 2 O n " a "T X • CT-O a o g-D. c-o: P ? the ^nerality lade (Lift, 3 Hour and if t>ur, Poser nt by asking I Poser and fficacy in en o H d it^eand O p a a~ ^ ( T o 3 S-participants 1^  certain ibo. S o *^ cr tr^  ^ fc ^— P O s ^ a n ^ ^ n O ^ S t—• ft 2 * ;::< -' m n> " w ft r^  tr O •-t u 3 n I - ihe whole n ° >e retired." T cf. for each item a-n Of self-three ie planned ie scale was adults. Engels the Life Changing Images 45 Possible Selves Safran, Segal, Hill, and Whiffen (1990) point out that in schema research there is a conflict between satisfying experimental requirements by pre-determining the stimuli that will be used in a study, and examining the unique qualities of an individual's self-schema. A measure of possible selves should provide a balance of these two factors. In addition, the possible selves measure should assess the functional set of possible selves, that set of possible selves that is readily employed by the Individual (Nurlus, 1986). In this study, these requirements were met by the construction of a measure of possible selves that allowed the participants to respond freely regarding their images of future selves in retirement. Participants received the following instructions; In this section, we are asking you to describe your Images of yourself in the future. Think about yourself in retirement, You may have one image of yourself, or several Images. Some images may be positive, others may be negative. Take a moment to think about the first of these images that comes to mind. When that Image is clear in your mind, write down short sentences that describe yourself in that image. When you have finished, go on to the next page, [next page] If you have another image of yourself in retirement take time to think about that image. When the Image is clear in your mind, write down short sentences that describe yourself in that image. Repeat this process until there are no more images of yourself that are clear to you. Participant responses were coded by two independent raters on four dimensions; positive self-references, negative self-references, effective self-references and negative effective self-references. Positive self-references include "I will be relaxed." Negative self-references are items such as "I see myself bored with nothing to do." Effectiveness was defined as references to the self taking action, CInanging Images 46 or being in control of a situation through problem-solving, such as "I will meet people by volunteering at the art gallery," Negative effectiveness was defined as references to the self tal<ing action in a negative way, examples included "I will eat and drinl< too much." There were insufficient items in this category and it was collapsed into the negative self-references category. The measure, then, produces total frequencies for each of the positive, negative, and effective possible self-items and these scores were used as the dependent variables for examining the hypotheses. Raters met and went over the coding manual (see appendix B). They then coded responses from participants in the pilot work as practice and discussed any difficulties. The practice session was repeated a second time and inter-rater agreement was found to be greater than 80%. The raters then worl<ed with responses from participants in the dissertation. After coding the responses individually, inter-rater agreement was calculated and agreement reached between raters on any items that they had disagreed on. For the calculation of inter-rater agreement, it was necessary for the raters to agree on the idea unit, in addition to the content of the idea unit. Thus if a sentence was coded by one rater as containing one idea unit, and by another rater as containing two idea units, then the raters were coded as disagreeing on two items. Inter-rater agreement was calculated for each participant's response by dividing the number of Items that the coders had agreed upon by the total number of items coded for that response. For example, one participant's response contained 4 ideas and coders agreed on the coding of 3 ideas, thus agreement was 75%, Another participant's response contained 9 ideas and coders agreed on the coding of 8 ideas, thus agreement was 88,89%. Overall percentage agreement between raters was 82.14%, Changing Images 47 Pilot Testing Pilot testing of the material was conducted with six participants to assess the comprehensibility of the Possible Selves questionnaire and the adequacy of the coding scheme. In addition, pilot participants were asked about issues that they would like to see covered in a retirement seminar. Desian A time lagged cross-over design was employed. This is a 3 (targeted-change group, structured-discussion group, wait-list control group) x 3 (pre-test, post-test, follow-up) design, with random assignment to groups. The first factor was manipulated between subjects and the last factor was within-subjects. This design is presented in Figure 1. Targeted-change Follow-up Group Discussion group Follow-up Targeted-change Walt-list < I ^ Group wait-list wait-list post-test follow-up Discussion group Pre-test Post-test Figure 1. Design of the study illustrating the cross-lagged approach. Changing images 48 At pre-test, the measures were administered to all participants, Participants in the targeted-change and structured-discussion groups then participated in the respective 8-week seminars. At post-test, the measures were again administered to all participants. Three months after the seminars ended, participants in the targeted-change and structured-discussion groups participated in a follow-up measurement session, Following the post-test, the participants in the wait-list were randomly assigned to either the targeted-change or the structured-discussion group. In turn, following the 8-weel< sessions, at the wait-list post-test, the measures were administered to participants in the wait-list targeted-change group and the wait-list structured-discussion group. Again, 3 months following the seminars, the measures were administered at the wait-list follow-up session. Procedures All participants were interviewed individually prior to the start of the study in order to determine their expectations about the study, to explain the project in more detail, and to obtain consent. Younger women who did not meet the criteria for the study and were primarily interested in financial counselling were referred to appropriate agencies. At the end of the interview, participants were asked to complete the possible selves items and the instructions were read to them. The other three questionnaires were then given to the participants to complete and additional demographic data was gathered (see appendix C), Participants were randomly assigned to one of the three groups in sets of three, A random numbers table was used to assign the first participant to one of the three groups, A random numbers table was then used to assign the second participant to one of the two remaining groups, and the third participant In the set was assigned to the remaining group, Changing Images 49 McGhee (1987) recommended an 8-10 week, 2-hour per session format for pre-retirement programmes because this time frame allows material to be assimilated and allows time for participants to start applying material outside of the sessions, and the 8-week format was used in this study. Participants in the taraeted-chanae aroup were asked to attend eight 2-hour sessions. The sessions were run according to the detailed outline in the treatment manual (available from the author) by two co-counsellors. The co-counsellors endeavoured to foster a supportive, collaborative atmosphere through the use of empathy and counselling skills. In addition, the use of empathy served to keep the group focussed on feelings about retirement, The main emphasis of the skills sessions was to apply problem-solving and stress Inoculation techniques to the issues involved in retirement. Participants were presented with a model of stress inoculation. Small group, group discussion, and individual exercises were focused on changing the images that participants have of themselves in the future. Activities specifically related to challenging negative self-statements with positive self-statements, and practicing problem-solving skills. Mid-way through the sessions, at the end of session four, the participants were given the Retirement Self-efficacy and Possible Selves measures to complete, The four questionnaires were re-administered at the end of session eight. At the follow-up, 3 months later, participants were asked to return in order to complete the four questionnaires a final time. Participants in the structured-discussion aroup were asked to attend eight two-hour sessions. The sessions consisted of support and discussion on issues relating to retirement according to the outline in the treatment manual (available from the author). The activities included explored the meaning of leisure activities, ageing, and housing options. The co-counsellors endeavoured to foster a supportive, collaborative atmosphere through the use of empathy and Changing Images 50 counselling skills. In addition, the use of empathy served to keep the group focussed on feelings about retirement. Mid-way through the sessions, at the end of session four, the participants were given the Retirement Self-efficacy and Possible Selves measures to complete, The four questionnaires were re-administered at the end of session eight. At the follow-up, 3 months later, participants were asked to return in order to complete the four questionnaires a final time. Participants in the wait-list control aroup were told that space was available for them In a future seminar. Participants waited a minimum of 8 weeks, and an average of 15 weeks before a group was available for them in the location that they wanted. Prior to their participation in the seminar, participants in the wait-list control group were asked to complete all four questionnaires. They were then randomly assigned to either the targeted-change or the structured-discussion group, using a random numbers table was used to assign the first participant to one of the two groups. The second participant in the set was assigned to the remaining group. All participants were thanked for their participation and debriefed at the end of the study. The results of the project were made available to them at the completion of the study. Expectations of Treatment Efficacy In order to determine whether the targeted-change and structured-discussion groups were equally plausible to participants, information about treatment expectations were gathered, At the end of the first session, participants in the targeted-change and structured-discussion groups were asked to indicate on a 5-point scale how effective they believed the sessions would be for helping them to cope with retirement, and how likely they would be to recommend the programme to a friend (see appendix D) Changing Images 51 Group leaders There were 10 groups of participants, 5 targeted-change groups and 5 structured-discussion groups ranging in size from 4 participants to 10 participants. The groups were led by female therapists experienced in leading adult groups who had graduate training in adult counselling interventions. In addition, the therapists were trained in the targeted-change approach or the structured-discussion group approach outlined in the treatment manuals, as appropriate. I co-led 6 of the 10 groups and led 2 of the smaller groups (both structured-discussion groups) by myself. The training sessions in the two treatment approaches consisted of reading the treatment manuals followed by two meetings to discuss the particular approach, and issues relating to retirement and ageing. Co-leaders discussed session plans and group dynamics once a week prior to each session. Treatment Fidelity The treatment manuals and session plans enabled the therapists to maintain treatment fidelity for the two types of treatment by ensuring that the activities used in the two groups were distinct. In addition, the sessions were video-taped. A manipulation check on treatment fidelity was made by using expert judgement of a sample of tapes to assess whether the two groups could be identified (see appendix E for details). Programme evaluation forms were used to assess participants experiences of the two groups (see appendix F). Behaviours performed in preparation for retirement were assessed using the Preparatory Behaviour form (Durrant, 1988) as a further evaluation of treatment fidelity. Clinical Sianificance It has been argued that statistically significant results are not necessarily on indicator of clinically meaningful results. Consequently, the concept of clinical Changing Images 52 significance has been proposed along with a variety of methods of calculating clinical significance (Jacobson & Revenstorf, 1988), In an effort to provide a standardized approach to calculating clinical significance, Jacobson and Truax (1991) outline a formulation in which a measure of reliable change is calculated that takes into account the reliability of the measure. Although, Speer (1992) has criticized this formulation of clinical significance for not taking into account regression to the mean and has in turn proposed another formula, it must be noted that Speer's formula is particularly conservative. As Jacobson and Revenstorf state, a method that is overly conservative runs the risk of a Type II error, in which an effective treatment is overlooked. In the current study. Reliable Change Index was calculated according to Jacobson and Truax as RC= 7i. y , The pre-test mean is subtracted from the post-test mean and divided by the square root of 2 times the pre-test standard error of measurement squared. This method was selected in order to balance the risks of Type I and Type II error. Further, effect sizes were calculated for each of the planned analyses as an additional method of assessing the significance of the results. Data Analysis In question one it was predicted that there would be greater change over time from pre-test to post-test in anxiety and depression for participants in the structured-discussion and targeted-change groups than in the wait-list group. Further, it was predicted that there would be greater change over time for participants in the targeted-change group than in the structured-discussion group. Following the decision tree of Glass and Hopkins (1984, p. 393), the hypotheses were tested using 2 planned orthogonal contrasts. The contrasts for the three groups, targeted-change, structured-discussion, and wait-list are listed Changing Innages 53 in Table 1. The null hypotheses is that there would be no differences between the groups in change over time, Table 1 Planned contrasts for three groups pre-to post-test Contrast number 1 2 Targeted-change pre-test +1 +] Targeted-change post-test -1 -1 Structured- Structured-discussion discussion pre-test post-test +1 -1 -1 +1 Wait-list pre-test -2 0 Wait-list post-test +2 0 Contrast number one compares the change from pre- to post-test for the torgeted-change and structured-discussion groups with the change from pre-test to post-test for the wait-list group. It is predicted that there will be change for the two interventions but not for the wait-list. Contrast number two compares the change from pre- to post-test for the targeted-change group with the change for the structured-discussion group. It is predicted that the change for the targeted-change group will be greater than the change for the structured-discussion group. It was further predicted that the greater change over time for participants in the targeted-change group than the structured-discussion group on the anxiety and depression variables would be maintained at follow-up. The use of the cross-over design permits a larger n as participants from the wait-list were randomly assigned to the targeted-change or structured-discussion group, The hypotheses were tested for these larger targeted-change and structured-discussion groups in two stages using orthogonal planned contrasts on the time Changing Images 54 factor and are illustrated in Table 2, Contrast number three compares the change over time from post-test to follow-up for the targeted change group with the change over time for the structured discussion group, Contrast number four compares the change over time from pre-test to the average of post-test plus follow-up for the targeted change group with the change over time for the structured discussion group. Again, the null hypothesis is that there would be no differences between the groups in change over time, Table 2 Planned contrasts for two groups pre-test, post-test, and follow-up Contrast number 3 4 Targeted Targeted- Targeted--change change change pre-test post-test follow up 0 -1 -Hi 2 -1 . -1 Structured Structured Structured -discussion -discussion -discussion pre-test post-test follow up 0 +1 -1 -2 +] +] In question two it was predicted that there would be greater change over time from pre-test to post-test in positive possible selves, effective possible selves, Retirement Self-efficacy, and negative possible selves, for participants in the structured-discussion and targeted-change groups than in the wait-list group, Further, it was predicted that there would be greater change over time for participants in the targeted-change group than in the structured-discussion group. Again, following the decision tree of Glass and Hopicins (1984, p, 393), the hypotheses were tested using 2 planned orthogonal contrasts. The contrasts for the three groups, targeted-change, structured-discussion, and wait-list are listed in Table 1, The null hypotheses is that there would be no differences between the groups in change over time. Changing innages 55 It was further predicted that the greater change over time for participants in the targeted-change group than the structured-discussion group on the positive possible selves, effective possible selves. Retirement Self-efficacy, and negative possible selves variables would be maintained at follow-up. The use of the cross-over design permits a larger n as participants from the wait-list were randomly assigned to the targeted-change or structured-discussion group, The hypotheses were tested for these larger targeted-change and structured-discussion groups in two stages using orthogonal planned contrasts on the time factor and are illustrated In Table 2. Again, the null hypotheses is that there would be no differences between the groups in change over time. Finally, It was predicted that following the intervention, the correlation between retirement self-efficacy and effective possible selves would be positive and moderate, in the range of .5 to .7, Type I and Type II error In terms of balancing Type I and Type II error, this study was seen as exploratory, Consequently, Type II error was seen as important, while at the same time it was also important not to incorrectly reject the null hypothesis and make a Type I error. In order to balance the two, the alpha level for each contrast was set at .05 rather than setting the level for the overall experiment and using a Bonferroni procedure. In addition, effect size was calculated for each of the planned contrasts as a method for assessing the significance of the effects, Changing innages 56 Results Sample Characteristics Five participants dropped out before the end of the sessions: One participant from the wait-list condition dropped out after the second session; one participant from the targeted-change condition dropped out after the first session, and three participants from the structured-discussion condition dropped out, one after the first session, one after the second session, and one after the third session. Reasons for dropping out included illness and death of family members, relocation, and time conflicts. Data from these participants are included in the pre-treatment analysis and later as missing data. Four participants did not return questionnaires from the follow-up session, one from the targeted-change condition, and three from the structured-discussion condition, due to participants' other time commitments. Group Comparability To assess group comparability, the demographic variables were examined. The data from the 10 groups of participants were collapsed into the targeted-change, or structured-discussion groups as appropriate. A one-way analysis of variance (ANOVA) on the three groups (targeted-change, structured-discussion, or wait-list) did not reveal any significant differences in the age of participants assigned to the three groups, F < 1. Similarly, a one-way ANOVA on the three groups did not reveal any significant differences in years of education of the participants, F < 1, Means and standard deviations for both variables are presented in Table 3, Changing innages 57 Tabie 3 Aae and years of Education for participants in each of the groups Group Targeted-change Structured-discussion Walt-iist M 59.22 59.33 60.65 Age SD 3.11 5.37 3.98 n 16 17 28 Educational M 15.33 14.87 15.11 SD 2.82 2.13 2.76 n 15 16 27 ^ IVIissing data due to two participants leaving answer blank. Changing images 58 The remaining demographic data were nominal and were analyzed using chi-square, In each case, the number of cells with an expected frequency of less than five was too high for a chi-square analysis, and the data had to be collapsed into a smaller number of categories. Participants were asl<ed to indicate the number of years they had worl<ed by classifying themselves into one of four categories; 1 (less than 5 years), 2( 5-10 years), 3(10-20 years), and , 4(20+ years). The data are presented in Table 4. The data were collapsed into two groups, those who had worked less than 20 years and those who had worked more than 20 years. Chi-square analysis did not reveal any significant differences between the groups, %^(1, n=61) = 2.60, n.s, Table 4 Freauencies for number of vears worked bv group Group Targeted-change Structured-discussion Wait-list CI c 1 missing case. 5-10 yec irs 1 1 1 10-20 years 1 5 8 20+ years 14 11 19 Total 16 17 29 Changing images 59 Similarly, participants were asked to indicate when they planned to retire by classifying themselves into one of four categories: 1 (0-3 months), 2( 3 months -1 year), 3(1 -2 years), and , 4(2+ years). The data are presented in Table 5. The data were collapsed Into two groups, those who planned to retire in less than two years and those who planned to retire in more than two years. Chi-square analysis did not reveal any significant differences between the groups, x^O-n=61) = .14, n.s. Table 5 Frequencies for plannina to retire by group Group Targeted-chonge Structured-discussion Walt-list a c 1 missing case. 0-3 months 1 0 4 3 months -1 year 3 3 7 1-2 years 6 8 8 2+ years 6 6 9 Total 16 17 29 Changing images 60 The data for marital status are presented in Table 6. Marital status was collapsed into two groups, those who were married and those who were not married. Chi-square analysis did not reveal any significant differences between the groups, x^O • n=59) = 1.36, n.s. Table 6 Marital status freauencies by group Group Targeted-change Structured-discussion CI Wait-list b Single Married Separated Divorced Widowed other Total 10 0 0 2 0 16 0 17 1 29 c 1 missing case; t> 2 missing cases. Changing innages 61 Participants were asl<ed to classify their occupational status into one of seven categories. The data are presented in Table 7, The data were collapsed into two groups. Professional (level 2 and 3) and administrative/sl<illed (level 3,4,5, and 6). Chi-square analysis did not reveal any significant differences, x^O-n=59) = 1.76, n.s. Table 7 Occupational status frequencies by group Occupational category Group 1 2 3 4 5 6 Total Targeted-change 1 1 5 5 3 1 16 Structured-discussion CI 0 3 7 2 3 1 17 Wait-list Q 0 11 5 5 7 0 29 Note, Occupation; 1 = not ennployed (retired); 2= executive/advanced professional; 3= business manager/lower professional; 4= administrative personnel/snnall business owner; 5= sl<illed nnanual; 6= senni-skilled/nnachine operator; 7= unskilled; 8= student, c" 1 nnisslng case Changing Images 62 In addition, participants were aslced to indicate their prinnary reason for retiring using 7 categories. The data are presented In Table 8. The data were collapsed Into voluntary (categories 1, 4, and 5) and Involuntary (categories 2, 3, and 6) categories. There were no significant differences, x^O- n=57) = ,50, n.s. Table 8 Reason for retirement Group Reason for retiring 1 4 7 10 2 0 1 0 3 1 2 2 4 5 0 1 6 4 4 10 7 TOTAL 1 16 2 17 4 29 Targeted-change ci Structured-discussion ^ Wait-list b Note. Reason for retirement: 1 = more leisure time; 2= pressure from employer; 3= health reasons; 4= spouse retiring; 5= more volunteer time; 6= age 65; 7= other. c 1 missing case t" 2 missing cases Changing Images 63 In sum, there appear to be no pre-treatment differences between the groups In sample characteristics. Assessing sample comparability Is somewhat difficult due to the limited number of studies on pre-retirement programs. Educational levels appear similar to that reported by Fretz et al. (1989) who Indicated that over half of the women in his study had some college courses, and to Durrant (1988), although she reports information on education for a combined group of both men and women. The majority of women In her study had worked for more than 30 years. Correlation Matrix The correlation matrix for the main dependent variables and demographic variables Is presented In Table 9. Bonferroni probabilities were used to establish significant relationships, From these it can be seen that Retirement Self-efficacy Is significantly correlated to anxiety with higher self-efficacy being related to lower anxiety. Other relationships are non-signlflcant. The relationships, however, are In the expected directions. Retirement self-efficacy Is positively correlated with positive possible selves, and negatively correlated with depression, and negative possible selves. Anxiety Is negatively correlated with positive possible selves but positively correlated with depression. Changing images 64 Table 9 Correlation Matrix of main dependent variables and demographic variables Age Education Positive'^ Effective^ Negatived RSE© STAlf SMDig Age --.14 -.10 -.10 -.01 -.08 .08 .18 Educ-at ion^ --.10 .19 .06 -.16 .19 -.15 Positive -.10 .03 .37 -.34 -.07 Effective -.00 .19 .04 .04 Negative --.36 .29 .29 RSE --.50* -.40 STAi SMDI -.39 '-' Years of education '-' Positive possible selves ^ Effective possible selves ^ Negative possible selves ® Retirement Self-efficacy ^ State-trait anxiety inventory 9 Short form Multiscore Depression Inventory * Q_<.05 Note: n=50 Changing innages 65 Preiiminarv Analysis In order to evaluate the treatment implementation across the groups, participants' expectations about the seminars were assessed towards the end of the first session, following the description of the treatment, They were asked to indicate on a 5-polnt scale how effective they thought the sessions would be in helping them to cope with retirement (l="highly ineffective," and 5="highly effective"). Note, for participants in the wait-list group, expectancy was measured after post-test and subsequent re-assignment to one of the two treatment groups. The results are presented in Table 10. A one-way ANOVA on the effectiveness question revealed significant differences between the groups, F(2,53)=4,49, Q_ < ,05. Scheff§ contrasts at the .05 level revealed, as expected, that participants in the targeted-change group had higher expectancies than the participants in the wait-list group. The difference between the structured-discussion group and the wait-list group was not statistically significant, nor was the difference between the structured-discussion group and the targeted-change group. In addition, participants were asl<ed how likely they would be to recommend the course to a friend (1 ="highly unlikely," 5="highly likely"). The means are presented in Table 10. A one-way ANOVA on the recommend to friend question again revealed statistically significant differences between the groups, F(2, 53) = 6,24, Q_ < ,05, Similarly, Scheff6 contrasts indicated that the targeted-change and structured-discussion groups had higher scores than the wait-list group, but did not differ from each other. Thus, it appears that there are no differences in expectations between the targeted-change and structured-discussion groups that might interfere with treatment. Changing images 66 TabielO Expectations reggrding effectiveness end recommending course, means, and standard deviations Group Targeted-cliange Structured-discussion Wait-iist IVI 4,67 4,50 3,81 Effective SD ,49 .65 1,13 n 16 14 26 Recommend IVI 4.73 4,79 3,96 SD ,49 ,43 1,00 n 16 14 26 Changing images 67 Main Anaiyses in order to determine whether the assumptions of the analysis of variance (ANOVA) were met, a stem-ieaf plot, Cochran's C, and Bartlett Box-F were performed for the positive possible selves, effective possible selves. Retirement self-efficacy, negative possible selves, STAI-S, and SMDI. With the exception of the negative possible selves, the data appeared to meet the assumptions. The negative possible selves data appeared sicewed and non-normal. As a result, this variable was transformed using a log transformation (Myers, 1979, p. 72-3). Pre-treatment differences were assessed with one-way ANOVA across the three groups on the pre-test scores on the main dependent variables. There were no pre-treatment differences between the three groups on the main dependent variables; F (2,58) = 1.18, n.s., for positive possible selves; F < 1, for effective possible selves; F < 1, for Retirement Self Efficacy; F < 1, for negative possible selves; F < 1, for STAI-S; and, F (2,53) = 1.58, n.s.; for SMDI. Note, five participants left missing items for the SMDI and the questionnaire was considered invalid for those participants as indicated in the manual for the Multiscore Depression Inventory and was reported as missing data. Thus, there is no indication that pre-treatment differences will interact differentially with treatment effects. Hypothesis Tests Due to the use of the cross-lagged design, the data analyses fall into two clusters; pre- and post-treatment for the three groups (targeted-change, structured discussion, and wait-list) and pre- post- and follow-up for the two groups (targeted-change and structured-discussion) combined with participants re-assigned from the wait-list. The results are presented for each cluster, Pre- to post-test analysis of taraeted-chanae, structured-discussion, and wait-list aroups. Following the results of these initial analyses, the hypotheses Changing images 68 described in chapter two were examined in more depth. In question one, the first hypothesis was that participants in the targeted-change and structured discussion groups would show a greater change in anxiety and depression compared with participants in the wait-list group from pre-to post-treatment. Second, participants in the targeted-change group were predicted to show a greater change in anxiety and depression, compared with participants in the structured-discussion group from pre- to post-treatment, A 3 (group) by 2 (time) repeated measures analysis of variance was performed for each variable and 2 orthogonal planned contrasts on the groups were used to test the hypotheses. Only the results of the planned contrasts are reported. Effect size was calculated for each planned contrast using the following formulas adapted from Cohen (1977); 1, ES^ (ni-1)sf^(r^-1)^-H(f|-1)^ 2 FS (^ i - " i2 ) - ( /n i -^2) r\ + i\-Z For STAI-S, the hypotheses were not supported [targeted-change group and structured-discussion group versus wait-list, F (1,52)=1,34, n,s„ ES=,48, and targeted-change versus structured-discussion, F < 1, ES=,09], Similarly, for SMDI there were no differences, F < 1, for each comparison, ES=-,29 and ,02 respectively. The means and standard deviations are presented in Table 11 and Figure 2, For question two, the first hypothesis was that participants in the targeted-change and structured discussion groups would show a greater change in Changing images 69 positive possibie selves, effective possible selves. Retirement self-efficacy, and negative possible selves compared with participants in the wait-list group from pre-to post-treatment, Second, participants in the targeted-change group were predicted to show a greater change in positive possible selves, effective possible selves. Retirement self-efficacy, and negative possible selves compared with participants in the structured-discussion group from pre- to post-treatment. A 3 (group) by 2 (time) repeated measures analysis of variance was performed for each variable and 2 orthogonal planned contrasts on the groups were used to test the hypotheses. Effect size was calculated for each planned contrast using the previous formulas adapted from Cohen (1977), For positive possible selves, only the hypothesis of change over time for the targeted-change group and structured-discussion group versus wait-list was supported, F(l,53)= 12.61, e<.05, ES=-2,17, The hypothesis regarding change over time for the structured-discussion group versus targeted-change group was not supported, F(l,53)= 2.79, n,s,, ES=,68, The increase in positive possible selves is greater for participants in the targeted-change group and the structured-discussion group than for participants in the wait-list group. There were no differences for effective possible selves, F < 1, for each comparison, ES=,05 and -,28, respectively. For neoative possible selves the hypotheses were not supported [targeted-change group and structured-discussion group versus wait-list, F < 1, ES=,32; targeted-change versus structured-discussion, F(l,53)= 1.88, n,s„ ES=,50], For Retirement Self-efficacv, the hypothesis was supported for change over time for the targeted-change group and structured-discussion group versus wait-list, F(l ,49)= 9,87, e<,05, ES=-1,90, but not for targeted-change versus structured-discussion, F < 1, ES=,17, The Increase in Retirement Self-efficacy is greater for participants in the targeted-change group and the structured-Changing innages 70 discussion group than for participants in the wait-iist group. The nneans and standard deviations are presented in Table 11 and Figure 2, Changing innages 71 Table 11 Means and standard deviations for targeted-chgnge, structured-discussion, ond wgit-list groups pre- ond Group Positive Possible Selves Torgeted-chgnge Structured-discussion Wgit-list Effective Possible Selves Torgeted-change Structured-discussion Wait-list post-test. nci 15 13 28 15 13 28 1 M 2,73 3.77 3.75 2.80 2.62 1.75 Neaative Possible Selves ftransfornned) Targeted-change Structured-discussion Wait-list Retlrennent Self-efficacv Targeted-change Structured-discussion Wait-list State Anxietv Targeted-change Structured-discussion Wait-list 15 13 28 14 13 28 15 12 28 .27 .24 .28 77,02 74.30 73,41 39,60 40,42 41,53 P re-test SD 2.15 2,83 2,82 2.70 3.55 2.05 .27 .32 ,30 13.09 9.80 13.44 13.54 12.18 12.33 M 6.40 5.46 3.46 3.47 2.31 2.00 .18 .30 ,31 86.48 81,99 68.92 34,09 35,83 39,29 Post-test SD 3,09 3,60 2,46 4,39 3,55 2,28 ,26 ,35 ,23 8,74 10,76 18.76 7,80 9,50 12,69 (table continues) Group Depression Targeted-change Structured-discussion Wait-iist n 13 9 26 M 7.54 7.61 11.23 Pre-test SD 5.40 6.36 7.34 Changing innages 72 Post-test M SD 7.31 6,56 7.56 5.96 10.12 6.65 a ns differ due to missing data Clnanging images 73 7--6--5--Mean 4--Responses 3--2 •• 1 •• 0 Positive possible selves Pre-test Post-test targeted (15) discussion (13) wait-l ist (28) Mean Responses 4 -3-2-1 • U"* Effective possible selves [ ^^^ " • " • " ^ - - - - - - . - ^ Pre-test Post-test • targeted (15) @ discussion (1 3) m wait-l ist (28) Figure 2. iVIean scores as a function of time of measurement; pre-and post-treatment; and group; targeted-change, structured-discussion, and wait-iist. CInanging images 74 Mean Responses 90 80 70 60 50 40 30 20 10 0 Retirement Self-efficacy targeted (14) discussion (1 3) wait-list (28) Pre-test Post-test Mean Responses 0.41 0.3 • 0.2-0.1 • U i Negative possible selves !>*<c;;] ] ]*^ Pre-test Post-test • targeted (15) S I discussion (13) Oni wait-list (28) Figure 2. (Continued) Changing images 75 Mean Responses 50-] 40-30-20-10-n. U 1 State Anxiety • -^^ssss Pre-test Post-test • targeted (15) ^ discussion (12) m wait-list (28) Mean Responses 20 T 15--Depression 10--5 -Pre-test Post-test targeted (13) discussion (9) wait-l ist (26) Figure 2. (Continued) Changing innages 76 Pre- post-treatnnent and follow-up for targeted-change and structured-discussion groups. In guestion one it wos predicted thct there would be greoter chgnge fronn pre-treatment to post-treatment between participants in the targeted-change group and structured-discussion group and that this would be maintained from post-test to follow-up for anxiety, and depression. Participants from the wait-list were randomly assigned to the targeted-change or structured-discussion group and the results from these participants were combined with the results from participants in these two groups. A 3 (time) by 2 (group) repeated measures ANOVA was performed for each variable and 2 planned orthogonal contrasts were used to test the hypotheses of change over time for the two groups. Again, effect size was calculated for each contrast using the following formulas: 3 £5^ ( j ( ^ r ^ 2 - ^3) - (2 /n i -^2- /TJa) / i} + i\-2 4 ES " (^2"^3) + ( ^ 2 - ^ 3 ) 2 /(A^-1)sf-H( There were no significant differences, The results are reported in Table 12, The means are presented in Table 13 and plotted in Figure 3, In guestion two it wos predicted that there would be greater change from pre-treatment to post-treatment between participants in the targeted-change group and structured-discussion group and that this would be maintained from post-test to follow-up for positive possible selves, effective possible selves. Retirement self-efficacy and negative possible selves. Participants from the wait-list were randomly assigned to the targeted-change or structured-Changing images 77 discussion group and the results from these participants were combined with the results from participants in these two groups, A 3 (time) by 2 (group) repeated measures ANOVA was performed for each variable and 2 planned orthogonal contrasts were used to test the hypotheses of change over time for the two groups, Again, effect size was calculated for each contrast using the previous formulas. There was only one significant difference, and that was for positive possible selves. There was greater change over time from pre-test to the average of post-test and follow-up for participants in the targeted change group, F(l, 53) = 6.12, 2 < •05, ES= -1,58. This supports the hypothesis that the differences between the two groups were maintained. The other results were non-significant and are reported in Table 12. The means are presented in Table 13 and plotted in Figure 3, Chionging innages 78 Table 12. Results of planned contrasts for chanae fronn pre-test to post-test to follow-up for connblned targeted-chanae and structured-discussion aroups Planned contrast Measure pre to post and follow-up post to follow-up Positive possible F (1,36) =6,12^ F (1,36) =1,17 selves ES=-1,58 ES=,39 Effective possible F (1,36) = 3,26 F < 1 ES=.09 f<l ES=-,62 f <1 ES=-,33 f <1 ES=-,13 f <1 ES=-.42 ^ p < ,05 selves Negative possible selves Retirement Self-Efficacy STAI-S SMDI ES=-1,09 f<l ES=-,07 £<1 ES=-,10 F<1 ES=-,30 F(l,33) = 1,98 ES=-,80 Changing images 79 Table 13 Means and standard deviations for dependent measures, pre-test, post-test, and follow-up for combined taraeted-change gnd structured-discussion groups Pre-test Post-test Follow up Group n^ Positive Possible Selves Targeted-ctiange 22 Structured-discussion 22 Effective Possible Selves Targeted-change 22 Structured-discussion 22 M 2.82 3.68 2.18 2.59 SD 1.99 2.92 2.46 3.14 Neaative possible selves Ctransformed") Targeted-change 22 Structured-discussion 22 .24 .27 .26 .28 M 5.73 5.05 3.41 2.23 .29 .22 SD 2.68 3.54 4.08 3.29 .28 .27 M 6.05 4.32 3.32 1.82 .24 .35 SD 2.57 1,96 4.36 1.74 .28 .33 Retirement self-efficacv Targeted-change 22 76.22 12.17 84.30 9.28 84.64 10.82 Structured-discussion 22 71.11 11.92 76.36 20.77 81.12 11.09 State Anxietv Targeted-change 22 41.09 12,64 34.79 Structured-discussion 22 39.54 10.56 31,64 Depression Targeted-change 20 9,50 6,42 8,50 Structured-discussion 19 9.26 7.45 6.68 ^n's differ due to missing data 8.40 8.81 6,57 5,31 33.96 31,96 5.30 5.84 8.86 9.36 4.99 5.58 Changing Images 80 Mean Responses Positive possible selves Pre-treatment Post-treatment Follow up Targeted (22) Discussion (22) Mean Responses 4-2-1 • U"! Effective possible selves Pre-treatment Post-treatment Follow up • Targeted (22) S Discussion (22) Figure 3 Mean scores as a function of time of measurement, pre-test, post-test, and follow-up; and group, targeted-change versus structured-discussion, Changing innages 81 90 80 70 60 Mean Responses 50 40 30 20 10 0 Retirement Self-efficacy Pre-treatment Post-treatment Follow up Targeted (22) Discussion (22)1 Mean Responses 0.4 T 0.3--0.2 •• 0.1 •• Negative possible selves Pre-treatment Post-treatment Follow up Targeted (22) Discussion (22) Figure 3 (continued) CInanging innages 82 State Anxiety Mean Responses 50 T 40--30--20--10--Targeted (22) Discussion (22) H h Pre-treatment Pcst-treatment Follow up Depression 20 T 15--Mean Responses 10 • • 5--Targeted (20) Discussion (19) -I 1-Pre-treatment Post-treatment Follow up Figure 3 (continued) Changing innages 83 Correlationai hypothesis. Finally, the relationship between the effective possible selves nneasure and the Retirement Self-efficacy scores for participants in the targeted-change group and the structured-discussion group was assessed using a Pearson product-nnonnent correlation. It was not significant at post-test, r =. 18, n.s. It was, however, significant at follow-up, r =.30, 2 < •05. This does not meet the criteria of .5 set out in the hypotheses. Post-hoc Analvses A series of post-hoc analyses were conducted. These were further analysis of the anxiety and depression measures, evaluation of clinical significance, an analysis of the balance of positive to negative thoughts, examination of the subscales of the depression measure, and examination of the evaluation and preparatory behaviours questionnaires. Anxiety and Depression In order to further evaluate the clinical effectiveness of the retirement preparation programmes, a post-hoc ANOVA, pre- and post-test by wait-list versus combined targeted-change and structured-discussion group was performed on the anxiety and depression measures. For the STAI-S, this revealed an effect of time, £(1,54)=9.49 £<.05, but not of group, F < 1, nor a group by time interaction, F(l,54)=l .59, n.s. The wait-list and the combined targeted-change and structured-discussion groups improved with time. This indicates that regression to the mean may have occurred. For the SMDI, there were no significant effects: for group, F(l,46)=3.2, n.s., for time, f < 1, nor for the group by time interaction, F < 1. Clinical Sianificance The means for the Reliable Change Index for pre- to post-test change for the modified STAI-S were calculated using the test-retest reliability of r=,72, The means are presented in Table 14. A one-way ANOVA on the Reliable Change Changing images 84 Index for the three groups (targeted-change, structured-discussion, and wait-list) was not significant, F < 1, The improvement and deterioration rates are presented in Table 15. Table 14 Reliable Chonae Index for STAI scores Group M Targeted-change 1,51 Structured-discussion 1,25 Wait-list ,61 SD 3,38 2,69 1,61 Table 15 Improvement and deterioration rates for responses on the STAI Group Improved No change Deteriorated Targeted-change Structured-discussion Wait-list 4 5 4 11 10 25 1 2 0 Changing innages 85 Similarly, the Reliable Change index was calculated for the SMDI using the test-retest reliability of r =,77, The nunnber of participants who con be categorized as having experienced significant change is still low. Again, the one-way ANOVA for the SIVIDI Reliable Change Index revealed no significant differences between the groups, F < 1. The results are presented in Table 16 and Table 17, Table 16 Reliable Chanae Index for SMDI scores Group M Targeted-change ,09 Structured-discussion ,04 Wait-list ,43 SD 1,31 2,99 1,44 CInanging images 86 Tabie 17 improvement and deterioration rates for responses on the SMDi Group Improved No change Deteriorated Targeted-change Structured-discussion Wait-list 0 1 1 15 15 27 1 1 0 Changing innages 87 States of Mind To assess the balance of positive thoughts to negative thoughts, the States of iVlind ratio (Schwartz & Garannoni, 1989) was calculated as positive possible selves divided by positive possible selves plus negative possible selves, The nneans for the three groups at post-test, targeted-change, structured-discussion and wait-list ore presented in Table 18, In addition the categories used by Schwartz and Garamoni (1989); positive dialogue, negative dialogue, conflict, positive monologue, and negative monologue were used to categorize the post-test results and the results are presented in Table 19, In order to conduct a chi-square analysis the results had to be collapsed into two categories, dialogue, representing the first four categories, and positive monologue. There was a significant difference between the groups, x^Q., n=61) = 8,78, 2 < '05, It appears, from the States of Mind categories, that participants in the targeted-change group had more positive than negative images of themselves, than participants in the wait-list group, who in turn hod more positive than negative images of themselves than participants in the structured-discussion group. Although this finding is somewhat unclear, it does add to the results from the hypothesis tests in confirming that the targeted-change group worl<ed to change participants' images of themselves in a positive direction. In addition, the ratio for participants In the structured-discussion group appears to decrease from pre-test to post-test suggesting that the ratio of positive to negative possible selves is lower following the intervention. Changing innages 88 Table 18 States of IVIind Ratio Group Targeted-change Structured-discussion Wait-list Note. The hiaher the ratio, the ar M .70 ,70 .72 eater P re-test SD .36 .36 .34 M .88 .61 .80 Post-test SD .25 .42 ,15 the nunnber of positive possible selves to the total of positive possible selves and negative possible selves. Table 19 States of IVIind Cateaories at post-test Negative Monologue Negative Dialogue Internal Conflict Positive Dialogue Positive Monologue Targeted-change 1 0 0 0 15 Structured-discussion 5 0 2 2 8 Wait-list 0 0 2 6 21 Changing images 89 Depression Subscales The SMDi has nine subscaies. One reason that the scaie was selected was in order to ailow an examination of the subscales. It was thought that some of the subscales might be particularly relevant for older adults. A series of repeated measures ANOVA were performed on each of the subscales pre- and post-test for each of the three groups. The means and standard deviations are presented In Table 20. There was only one significant difference between the groups. For the Pessimism subscaie, there was a group effect, with scores being lower for participants in the targeted-change group. There were, however, no significant changes over time. The results are presented in Table 21. Changing images 90 Tabie 20 IVIeans and Standard deviations for tlie SMDi subscales Subscale Group Low energy Targeted-change Structured-discussion Wait-list Cognitive Difficulty Targeted-change Structured-discussion Wait-list Guilt Targeted-change Structured-discussion Wait-list Low self esteem Targeted-change Structured-discussion Wait-list Social introversion Targeted-change Structured-discussion Wait-list Pre-treatment M 1.00 1,47 2,11 1,14 1,20 1,64 ,86 1,07 1,46 ,64 ,40 ,71 1,64 1,07 1,21 SD 1,84 2,03 2,18 1,29 1.21 1,85 1.10 .89 1,48 1.01 .63 1.12 1.69 1.28 1.26 Post-treatment M .71 1.47 2.21 .71 1.47 1,61 1,21 1,13 ,82 ,50 ,53 ,82 1,64 1,00 1,36 SD ,73 1,89 2,31 ,91 1,36 1,76 1,53 1,41 1,16 ,76 1,06 1,16 1,82 1,56 1,45 (table continues) Changing Images 91 Group Pre-treatment Post-treatment M SD M SD Pessimism Irritability Sad Mood Targeted-change Structured-discussion Wait-list Targeted-change Structured-discussion Walt-list Targeted-change Structured-discussion Walt-list Instrumental helplessness Targeted-change Structured-discussion Wait-list .57 1,87 1.25 .93 .20 .57 .21 .60 .86 00 .27 .50 1,34 2.13 1.60 .92 .41 1.03 .58 .74 1.11 00 .59 .96 ,57 2.07 1.04 .86 .47 .79 .21 .67 .61 00 .60 .18 1.09 1.94 1,50 1,23 .83 1.17 .58 ,98 ,99 00 .91 ,55 Changing images 92 Table 21 Analysis of Variance results for the SMDI subscales Subscale Effect Low energy Group Time Group X Time Cognitive difficulty Group Time Group X Time Guilt Group Time Group X Time Low Self-esteem Group Time Group X Time Social Introversion Group Time Group X Time Lvalue F(2,54)=2.8 f < 1 f <1 F(2,54)=l ,53 £<1 £<1 f < l f <1 F(2,54)=l,77 E<"l £ < • ! £<1 f < 1 f < 1 E<1 Probability level n.s n,s, n,s, (table continuesl Changing images 93 Subscale Effect F value Probability level Pessimism Irritability Sad Mood Group Time Group X Time Group Time Group X Time Group Time Group X Time Instrumental helplessness Group Time Group X Time F(2,54)=3.78 £<1 F<1 F(2,54)=2,ll f <1 f < 1 F(2,54)= 3,03 f <1 f < 1 F(2,54)=2,7 f < 1 F(2.54)=3.09 e< .05 n.s. n.s, n.s. n.s. Changing images 94 Retirement Self-efficacv subscales The RSE has three subscales, time and activity management, social interaction, and self-esteem. Analysis of the subscales revealed a pattern of results similar to that found for the total scale scores. A 3 (group) by 2 (time) repeated measures ANOVA revealed main effects for time and group by time interactions. The results and means for the subscales are presented in Table 22 and Table 23, Table 22 Analysis of Variance results for the RSE subscales Subscale Time and activity management Social interaction Self-esteem Effect Group Time Group X Time Group Time Group X Time Group Time Group X Time F value F(2,52)= 1,21 F (1,52)= 13,63 F(2,52)=3.74 F(2,51)=2,30 F(l,51)= 12,58 F(2,51)=7,09 F(2,53)= 2,21 F(2,53)= 12.45 F(2,53)= 8,37 Probability level n,s. e><,05 2<.05 n,s. e<,05 e<.05 n.s. e<.05 2<.05 CInanging innages 95 Table 23 IVIeans and standard deviations for the RSE subscaies Subscale Time and activity nnanagennent Sociai interaction Seif-esteem Group Targeted-change Structured-discussion Wait-iist Targeted-change Structured-discussion Wait-list Targeted-change Structured-discussion Wait-list Pre-treatnnent M 76,70 82.04 75,96 76,70 74.09 71,40 77.36 69.93 72,56 SD 14,86 13,39 18,14 21,02 21,67 21,70 13,95 15.90 15,90 Post-treatnnent M 86,91 85,69 77,50 88,67 85,62 69.12 84,11 770,09 70.46 SD 18,57 14,81 17,05 16.22 11,67 23,08 11,19 12,37 17,20 Changing innages 96 Evaluation Questions In the last session, participants were asl<ed to evaluate specific aspects of the course, They indicated on a 10 point scale (1 = 'not useful' and 10 = 'very useful') their perceptions of the groups on a series of dinnensions. There were no significant differences between the responses of participants in the targeted-chonge group and the structured-discussion group. The nneans and standard deviations are presented in Table 24. It must be noted that the retirennent model, problem-solving, and self-statements were not topics included in the structured-discussion group. That no differences were found between groups on these items suggests a halo effect in responding, Ciionging images 97 Tobie 24 IVIeans and standard deviations for tlie evaluation questions. Group a M SD b M SD c M SD d M SD e IVI SD Targeted-change 9,00 1,12 9,18 1.31 9,18 1,09 6.34 2,29 8.13 2,34 Structured-discussion 9.10 1.26 9.38 1.02 9.20 1.20 6.77 1.82 8.12 1,50 h Group M S D M S D M S D M S D M S D Targeted-cliange 7.93 2.40 7,15 2.13 7.89 1.83 7.50 2.08 7.80 1.71 Structured-discussion 7.89 1.68 8.12 1.36 8.00 1.83 7.65 1.77 7.50 2,73 Note. a= group support; b=group ieaders; c= group structured-discussion; d=fionnewori<; e=retirennent model; f=probiem-soiving; g=seif-statements; h=booi<s and materials; i=ageism/stereotypes; j=famiiy relationships Changing innages 98 Preparatory Behaviours At the follow-up, participants were asked to indicate whether they had performed a series of behaviours in preparation for their retirement, using the Preparatory Behaviour form (Durrant, 1988), If they had not performed these behaviours they were asked to indicate when, if ever, they intended to do so, it was necessary to collapse the cells in order to perform the chi-square analysis. For item 11, "spoken with ,,, insurance agent," the responses were collapsed into two categories; have completed or intend to complete the task, and intend never to complete the task. For all other items, the responses were collapsed into two categories; have completed the task, and have not yet completed the task. There were three significant differences between the targeted-change groups and the structured-discussion groups. For the statement 'Located available retirement resources', there was a significant difference, %2(1, n= 51) = 4,31, 2 <,05; for the statement 'prepared or revised my will' there was a significant difference, ^2(1- n=51) =4,1, ^ < ,05, and for the statement, 'Spoke with the following key advisors regarding my retirement plans: insurance agent' there was a significant difference, ^2(1, n= 45) = 4,98, Q_ < ,05, The frequencies are presented in Table 25, A clear pattern of results does not emerge. Although participants in the targeted-change group have taken more action than participants in the structured-discussion group with regard to retirement resources and speaking with their insurance agent, participants in the structured-discussion group hove taken more action with respect to preparing their wills. Ctianging Images 99 Table 25 Freauencies for retirement preparation behaviours Intend to take action In; Group Completed 6 mos Identified possible retirement problems Targeted-change 20 Structured-discussion 26 Located available retirement resources Targeted-change 18 Structured-discussion 14 Identified my retirement-planning goals Targeted-change 15 Structured-discussion 20 Made plans for my leisure time Targeted-change 15 Structured-discussion 21 Analyzed my present financial position Targeted-change 17 Structured-discussion 23 Developed a financial plan for retirement Targeted-change 14 Structured-discussion 17 Prepared or revised my will Targeted-change 14 Structured-discussion 24 0 0 3 3 4 0 3 2 4 1 3 4 9 2 1 year 2 0 0 5 2 3 0 0 2 1 5 2 0 2 2 years 0 1 2 3 1 3 2 3 0 2 0 4 0 0 3 years 0 0 0 2 1 0 1 0 0 0 0 0 0 0 Never 0 1 0 1 0 0 2 0 0 0 0 0 0 0 (table continues) Changing images 100 Intend to take action in; Group Completed 6mos 1 year Investigated options for retirement-living arrangements Targeted-change 12 Structured-discussion 17 2 1 0 0 Developed or maintained a physical health program Targeted-change 12 Structured-discussion 19 Spoken with my financial planner Targeted-change 14 Structured-discussion 17 Spoken with my insurance agent Targeted-change 9 Structured-discussion 3 Spoken with my accountant Targeted-change 10 Structured-discussion 7 Spoken with my lawyer Targeted-change 6 Structured-discussion 5 9 6 3 4 3 0 3 1 5 0 1 2 1 0 0 2 0 3 0 2 2 years 1 3 0 1 0 0 0 1 2 1 1 1 1 3 years 3 4 0 0 0 0 2 1 1 1 0 1 Never 2 3 0 0 4 4 8 16 6 10 9 15 ftable continues') Changing Images 101 Intend to take action in: Group Completed 6 mos 1 year 2 years 3 years Never Spoken with my employer Targeted-change Structured-discussion Spoken with my physician Targeted-change Structured-discussion Mid-treatment means Possible selves and Retirement self-efficacy measures were collected at the mid-point of the treatment programmes, The means are presented in Appendix G, It should be noted that the means appear approximately half-way between pre- and post-treatment scores and were not analyzed separately. 9 12 10 7 3 2 4 3 3 1 1 1 0 4 0 1 2 3 3 2 2 4 3 11 Changing images 102 Discussion The present research examined whether a cognitive-behavioural pre-retirement programme, compared with a structured-discussion group and a wait-list control group would be effective in reducing anxiety about retirement and depression. This set of hypotheses was not supported, Participants in the retirement groups did not show a greater decrease in the measures of anxiety and depression than participants in the wait-list group. In addition, this research examined whether a cognitive-behavioural pre-retirement programme, compared with a structured-discussion group and a wait-list control group would be effective in changing the way women view themselves in future retirement, and In creating a view of self as an effective problem-solver. Partial support was found for these hypotheses. There was a greater increase in positive possible selves and retirement self-efficacy from pre-to post-test for participants in the cognitive-behavioural and structured-discussion groups than for participants in the wait-list. Thus, participating in the retirement groups increased participants' sense of their own effectiveness in coping with their future retirement and increased their positive images of themselves in the future. In addition, positive possible selves increased more from pre-test to the average of post-test and follow-up for participants in the targeted-change group than in the structured-discussion group. The effect sizes for these findings are quite respectable. It should be noted that the cognitive-behavioural intervention specifically targeted participants' positive images of themselves in the future and was more effective at increasing these images than a discussion intervention that did not target these images. Participants did not, however, show a greater increase in effective possible selves nor did they show a decrease in negative possible selves. Changing images 103 The finding of the decrease in anxiety for participants in the wait-list group from pre- to post-measurement raises concerns about regression to the mean. If statistical regression to the mean occurs then the explanation for any effect lies not with the treatment but with measurement error, It is fortunate that the wait-list condition was included in the current study, as without this condition, erroneous conclusions may have been drown regarding treatment effectiveness for reducing anxiety, Common Treatment Factors There ore several alternate explanations that can account for the change in the wait-list group. These explanations relate to the potential effectiveness of common treatment factors (Lambert, Shapiro, & Bergin, 1986), The Initial interview may itself have had a therapeutic effect. Participants had the opportunity to express some of their concerns regarding retirement during the interview. Some of the participants indicated that it was difficult to talk to other people about retirement fears because other people tended to belittle the fears. Consequently, being in an environment where they could talk about their concerns, and have those fears token seriously, may hove been validating and helped to allay the fears. Second, wait-list participants knew that they would be attending a group in the near future and this knowledge may have acted as a treatment itself in developing a positive expectation about outcome and building a sense of hope and optimism. Third, the initial interview may hove started participants thinking and talking about retirement outside of the groups and thus precipitated change. In all of these scenarios, the Initial interview may hove worked to reduce the feelings of isolation that the women experienced, and to hove increased hope and a positive expectation for improvement. As Lambert et ol. point out, these factors should be recognized as having a crucial role in producing change, rather than being seen as extraneous to treatment, Changing images 104 This point has been reinforced by Kirsch (1990) who has argued that expectations play a critical role in producing changes in behaviour. Indeed, he has developed an approach to counselling that specifically targets clients' expectations regarding their own behaviour and builds into the therapy opportunities to develop positive expectations regarding treatment effectiveness. Similarly, the lack of differences between the targeted-change group and the structured-discussion group on the behavioural measures can be attributed to common factors. In all groups, the co-leaders employed counselling sicills, in particular empathy, to promote a supportive atmosphere. In addition, the leaders were attentive to group process. In the groups, there was a focus on exploring feelings about retirement, both positive and negative. Further, there was a sharing of information about retirement and retirement preparation. Bool<s, pamphlets, information on local services, and resources were available in both groups. In addition, both group situations provided opportunities for participants to discuss their own plans for retirement, to receive feedback, and to refine their plans. Sometimes the best discussions went on during the coffee breaks. Thus all participants had access to information and were able to take steps to act accordingly. The unfortunate consequence of discovering change on the anxiety measure in the wait-list group is that it is not possible to conclude that the reduction in anxiety relates to treatment without conducting further research using a design to avoid statistical regression, This could be accomplished by eliminating the use of a cut-off score for participation, and including participants with the full range of anxiety scores in the groups. This is likely, however, to alter the dynamics in the groups. A further alternative is to add an additional time of measurement, using a design in which all participants are measured twice prior Changing images 105 to the intervention. The first time of measurement is used for participant selection, Comparison of treatment effectiveness is made using the second time of measurement as the pre-treatment score, This wiii not disentangle change following the first interview from statistical regression but wiii allow an examination of the treatment effectiveness of the groups separate from any initial change. Significance The significance of the results of this study was evaluated in three ways. Statistical significance was calculated for the hypotheses regarding the mean differences between groups of participants in change over time. Clinical significance was evaluated using a formula to calculate whether change for individual subjects was significant given the reliability of the measure. And finally, effect size for the mean difference for change over time was calculated. The effect sizes for the treatment groups compared with the wait list group were large for positive possible selves and retirement self-efficacy. In addition, the effect size for the targeted-change group compared with the structured-discussion group was also large for postive possible selves. For Retirement Self-efficacy, the effect size was moderate for the comparison of the two treatment groups. The two forms of significance, statistical and effect size, confirm the findings of differential change over time for positive possible selves and Retirement Self-efficacy, Thus, I can feel confident about these findings. For anxiety, however, the different forms of significance suggest somewhat different conclusions. From the results of clinical significance, one might conclude that participants did not improve. Change over time for anxiety, however, was statistically significant, with anxiety decreasing from pre-test to post-test, although there were no differences between the groups in change over time leading to a conclusion that on average, anxiety decreased over time. The Changing images 106 effect size for the change over time is, however, moderate for a comparison of the treatment groups with the wait-iist group. For anxiety, it would appear that there may be an effect. Further research wouid be heipfui here to disentangle the findings and to examine retirement anxiety in greater depth. The lack of any major clinical effect is supported by the clinical significance results. In the anxiety and depression measures there were only a few participants who showed signs of mal<ing clinical improvements. There were also no significant differences between groups. Did the groups then have no impact on retirement anxiety beyond the initial interviews? Responses from the participants themselves may be revealing at this point. As could be expected from a demand characteristic effect, a number of comments were very positive in a global sense. For example, in the evaluation comments participants said "this has been one of the most effective groups I've participated in terms of bringing about personal change and gaining insights," "an excellent experience," "helpful," "I have learned so much," "wonderful chance to share," etc. More revealing, however, are the more specific comments regarding change as a result of the group. One participant reported that the retirement group "reassured me it was OK to feel apprehensive about retirement." Similarly, for another participant the group affirmed "that my concerns are shared by other women of my age." Sharing of concerns was reported by a number of the women as a benefit of the group. For others, the group was an opportunity to start coping with their anxiety. One participant reported that "the group has made me face up to the reality of retirement, to seriously think about it and start making plans." Similarly for another participant, the group was "a time to focus on the positives and negatives of retirement and think through what retirement Changing images 107 will actually be like." For another "it helped me to discover some changes I will have to make in my lifestyle to replace some of the things I will miss," Many participants reported that the group was an opportunity to focus on retirement goals and feelings, For some this meant opening up areas that they had been avoiding; "has made me face retirement which I have feared," One participant revealed that the group "has sparked some introspection," Role of Anxiety In Transitions The women in this study were pre-retirees. They were at the start of their transition from work. Schlossberg (1984) recommends focusing on the meaning of the transition for the individual and proposes that transitions be described as having beginning, middle, and end phases. The phases are to be thought of a fuzzy set rather than as a sequence of step-wise stages and are for the purpose of communicating information about prototype members. The women in this project can thus be viewed as being in the beginning phase of the retirement transition. The task for many of them at this point is to explore what retirement means to them. Or, to put it in the framework used by Lazarus, they are in primary appraisal, evaluating the threat of the transition to themselves. Indeed, exploring retirement issues, as the women did in the groups, has the potential to heighten anxiety for those Individuals who have only minimally worked on primary appraisal. The targeted change Intervention was designed to provide participants with skills to cope with retirement stress, to change secondary appraisal, their appraisal of their skills. It is possible that the targeted-change intervention did not reduce anxiety because it was not meeting the needs of participants at that point in their transition. Equally, it is possible that the intervention would have been measured as successful with a different instrument, one that focussed on skills for coping with retirement. Perhaps the impact of the intervention will be seen after the women cease work when they Changing images 108 have the opportunity to test their new skills and resources for coping with retirement. Retirement Counsellina for Women The current study did not compare retirement stress in men and women. Nor did it assess the need for pre-retirement counselling in men. What was clear, was the appreciation of the women in the project for a retirement group composed only of women. Clearly there was a selection process in this appreciation. There may also be demand for a mixed gender group but those women who wanted only a mixed gender group are unlikely to have volunteered for this project. Additionally, many retirement seminars encourage spouse participation and It is interesting to speculate on whether having spouses attend would increase or decrease the effectiveness of a retirement group. This would have to be addressed empirically, Similarly, the study was not designed to assess whether there was greater retirement stress in women than in men, Again, the demand among men for a retirement group that addresses feelings about retirement is unknown. In this project, the women were supportive of each other and worked at reminding each other of their strengths as women. The shared work histories that the women had, and their shared histories of having met multiple demands of having focussed on nurturing others, appeared to work for the groups. The women seemed to draw upon this shared knowledge to point out each others' hidden strengths, something that may not hove happened as readily in a mixed gender group. Retirement Preparation Programmes In retirement preparation, there are two main models; educational models, In which pre-retlrees are given information, and counselling models with an emphasis on facilitating the individual's personal growth (Tinsley & Changing images 109 Schwendener-Holt, 1992), The emphasis has been on the former, and there are far fewer counselling options available for retirees. Perhaps what needs to happen is to ask pre- and post-retirees what they feel they need In terms of retirement counselling. What should be the goals of retirement counselling? What programs should be offered? For some individuals, reduction of retirement anxiety will be the major goal of a retirement programme. For these individuals, a programme that effectively reduces that anxiety will be necessary. The results of the current study, in line with recent work in treatment effectiveness (Kirsch, 1990; Lambert et al., 1986) suggest that in order to develop and evaluate such a programme it will be necessary to move beyond the current counselling models and focus on an examination of non-specific treatment effects. What may be relevant for professionals planning interventions for retirement is to Incorporate an understanding of the phase of the retirement transition that individuals are in, and the meaning of the transition to the individual, Pre-retirees may have the need to explore the meaning of retirement for their lives and to examine their fears before taking inventory of their strengths for coping with retirement and working on building skills. An intervention that works more specifically to build these specific skills for coping with retirement may work better following a thorough exploration of retirement issues. Possible Selves Of interest from a theoretical perspective in this study was women's perceptions of themselves in their future retirement. Anticipating future events has been thought to play an important role in the coping process (Taylor & Schneider, 1989). These anticipatory images have the potential to serve as a basis for problem-solving and for motivating behaviour. Moreover, women's appraisals are open to modification, hence they serve as targets for intervention. The cognitive-behavioural approaches involved in stress Changing images 110 inoculation training and problem-solving seemed particularly appropriate for this modification due to the emphasis upon appraisal in these models. Examining changes in women's Images of their possible selves as they participate in the treatment approaches, was thought to illuminate the appraisal process and its role in coping with retirement stress, Further, assessing cognitions, in this case using the possible-selves measure, allowed for an examination of possible treatment mechanisms (Kendall & Korgeski, 1979). In order to assess the claims made for possible selves in this study, it is first necessary to determine whether possible selves were measured adequately. It would seem that the measure of possible selves used in this study was adequate on a number of points. From an administration perspective, it was manageable. Participants appeared to understand the instructions. Some participants responded that they had no future images, which is why they wanted to attend the groups. The responses that were given were readily coded and an adequate level of reliability could be reached. The free-response format, while requiring two coders to read through the responses, has the advantage of not constraining the respondents. Further, there was change in the positive possible selves following the cognitive-behavioural intervention that did indeed target positive images of the self. A further indication was the small but significant correlation between the effective possible selves and the retirement self-efficacy. What was the process by which the possible selves were modified? Nurius (1986) provides a framework for the role of a group in producing change in possible selves. In her model, the group has seven mechanisms for producing change in the self-concept. These are (a) providing a rationale ,(b) feedback from other group members giving alternate perceptions, (c) enhanced awareness of patterns of behaviour from group feedback and observations of Changing images 111 others in the group, (d) group support and reinforcement for change, (e) information, (f) reduced isolation, and (g) new social setting, In the current study, the mechanisms of group support. Information, reduced isolation, and new social setting were present in both the targeted-change and the structured discussion group. The targeted change group however had additional components in that it provided an interpretive frameworl< that focussed on self-monitoring, and may have focussed the feedbacl< of the group on self-talk. What is not clear from this study, given the difficulty in drawing conclusions regarding anxiety, is the role of possible selves in retirement anxiety, It has been demonstrated in previous research that hopes and fears for future selves can be identified by people in different generations (Cross & Markus, 1991; Dittman-Kohli, 1990; Roberts, 1992). It is repeatedly assumed in the literature that future selves impact on present functioning (Markus & Nurius, 1987; Nurius & Majerus, 1988; Roberts, 1992), Is there support for this? The current study offers no clear insights. There is evidence that possible selves can be modified, Nurius and Majerus (1988) provide a case study in which possible selves are shown to be modified during counselling. Similarly, in the current study, positive possible selves were increased. What remains less clear, is the relationship to current functioning. Cross and Markus (1991) have found a relationship between life satisfaction scores and the types of possible selves generated, with participants with low life satisfaction having fewer possible selves relating to personal well-being. The need exists for an investigation of the relationship between future selves and current functioning, Possible selves (and self schema) are viewed as being involved in dynamic, reciprocal interactions with the environment and through this interaction, to contribute to behaviour, Cross and Markus (1991) point to the role of possible selves in structuring an individual's processing of information, For Changing images 112 example, a woman who sees herself in the future as lonely and isolated may be primed to attend to certain aspects of social interactions such as on off-hand comment and misattribute this to the other person not valuing the relationship, Behaviour In further situations can be affected by these misottributions and contribute to a self-fulfilling prophecy effect. Additionally, Cross and Markus suggest that discrepancies between current and future selves may initiate behaviour. For example, an Individual may choose to start an exercise programme to achieve a possible self of slender and healthy. Future selves can also be a source of reinforcement for current behaviour that has short-term costs but long-term rewards. Creating, strengthening, and elaborating on possible selves has the potential to alter behavioural pathways. For example, a woman with a possible self as a student will be more likely to notice information in her newspaper on continuing education courses, and then more likely to register in the course, Thus, for the participant whose possible selves changed from "I am very confused at this time-There are no images at this time-Good health is very important" at pre-test, to "I think about myself very healthy, enjoying life-Helping the elderly-Being more with family and friends-. Gardening-, Croft-, Aerobics" at post-test, there is the potential for on impact on behaviour in several ways. First, from an information processing perspective, there is a more elaborate possible self-structure. The potential to be alert to opportunities to act is greater, as is the potential for greater memory of possible-self relevant material (Koto & Markus, 1993), Second, any discrepancy between present and future selves may initiate behaviour. Third, the more positive images may help the woman to endure short-term costs in order to obtain the long-term goals, such as being healthy, Effective possible selves were separated from positive possible selves in this study in order to assess whether there was a relationship between images of self Changing images 113 being effective in the future and Retirement Self-efficacy, a sense of mastery with retirement situations. The relationship was not as strong as predicted. One reason for this may lie in the nature of the questions used to assess Retirement Self-efficacy and effective possible selves. The questionnaire assessed mastery with respect to global retirement situations. Participants answering the Possible Selves question, on the other hand, described specific situations in which they envisioned themselves being effective. For example, the response, "I will exercise regularly" was coded as effective. It is possible that a participant could see herself as effective in a specific future activity but not necessarily effective in finding things to do in retirement. Contributions This research contributes on several levels. On a methodological level, the research provides a methodology for coding possible selves by employing open-ended responses. By using an open-ended approach, participants' responses are not constrained by the researchers expectations, and changes over time in the set of possible selves can be evaluated. On a theoretical level, the research demonstrates that possible selves in older women can be modified, Previous research has demonstrated that older adults possess possible selves (Cross & Markus, 1991) but modification of possible selves in this population has not been demonstrated previously, Further, the research examines possible selves in a counselling setting. Although the role of possible selves and anxiety was not clear in the present research, the research still provides information that can be relevant for further research in this area. Finally, on a practical level, the research contributes to a focus on retirement as a significant life transition by highlighting issues facing a portion of women facing retirement, The research contributes to demonstrating the relevance of counselling for retirement preparation. Changing images 114 Limitations Certain restrictions have been applied to this study that limit generalizability in two main areas. First, the sample is select, Participants were predominately middle-class, Caucasian women who have responded to the request for volunteers and who were experiencing anxiety. The results, then, may not generalize to all pre-retirees. Second, the time of measurement is restricted, A 3 month follow-up period was selected, but It is not l<nown whether the results generalize beyond that point in time to a year or even 2 years following retirement. Indeed it would be relevant to examine change following retirement in a further study. In addition, the study was seen as exploratory for the decision made In determining experiment-wise error. As a result, a large number of contrasts was permitted, thus inflating the possibility of a Type I error that a hypothesis could be accepted when the null hypothesis of no difference is in fact correct. Replication of the study is important before mal<ing policy decisions. Changing images 115 References Atchley, R. C. (1976). The sociology of retirement. Cambridge, MA: Schenl<man Publishing. Atchley, R. C. (1982). The process of retirement: Comparing women and men. In M. Szinovacz (Ed.), Women's retirement: Policy implications of recent research (pp. 153-168). Beverly Hills, CA: Sage. Avery, C. M., & Jablin, F. M. (1988). Retirement preparation programs and organizational communication. Communication Education, 37, 68-80. Baltes, P. B., Reese, H. W., & Nesselroade, J. R. (1977). Life-span developmental psvcholoav: Introduction to research methods. Monterey, CA; Brooks/Cole . Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychoioaical Review, 84, 191-215. Becl<, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. New York: Hoeber. BelL B. D. (1975). The limitations of crisis theory as an explanatory mechanism in social gerontology. International Journal of Agina and IHuman Development, 6, 153-168. Bell, B. D. (1979). Life satisfaction and occupational retirement: Beyond the impact year. International Journal of Agina and Human Development, 9, 31 -50. Benner, P. E. (1984). Stress and satisfaction on the job: Work meanings and coping of mid-career men. New York; Proeger Publishers. Berndt, D. J. (1981). How voiid gre the subscgles of the Multiscore Depression Inventory? Journol of Clinicol Psvchologv, 37, 564-570. Berndt, D.J. (1986). Multiscore Depression Inventory ("MDII Monual. Los Angeles, CA; Western Psychological Services. Berndt, D. J., Berndt, S. M., & Kaiser, C. F. (1984). Multidimensional assessment of depression. Journal of Personality Assessment, 48, 489-494. Berndt, D. J., PetzeL T. P., & Berndt, S. M. (1980), Development and initial evaluation of a multiscore depression inventory. Journal of Personality Assessment, 44- 396-403. Berndt, D. J., Petzel, T. P., & Kaiser, C. F, (1983), Evaluation of a short form of the Multiscore Depression Inventory, Journal of Consuitina and Clinical Psychology, 51, 790-791, Changing images 116 Bernstein, I, H., & Eveland, D, C, (1982), State vs trait anxiety: A case study in confirmatory factor analysis. Personality and Individual Differences, 3, 361 -372, Bills, R,E., Vance, E.L„ & Mclean, D.S, (1951), Anindexof adjustment and values. Journal of Consulting Psvchoiogy, 15, 257-261. Block, M, (1982), Professional women: Work pattern as a correlate of retirement satisfaction. In M, Szinovacz(Ed,), Women's retirement: Policy implications of recent research fpp, 183-194V Beverly Hills, CA: Sage, Bortner, R, W„ & Hultsch, D. F. (1970), A multivariate analysis of correlates of life satisfaction in adulthood. Journal of Gerontoloav, 25, 41 -47, Bower, G„ Black, J„ & Turner, T, (1979), Scripts in memory for text. Cognitive Psvcholoav, 11, 177-220. Carver, C,S,, Reynolds, S. L., & Scheier, M, F, (1994), The possible selves of optimists and pessimists. Journal of Research in Personality, 28, 133-141. Cohen, J, (1977). Statistical power analysis for the behavioral sciences. New York: Academic Press, Comrie, S, M, (1985), Training: Teach employees to approach retirement as a new career. Personnel Journal, 64, 106-108. Cross, S., & Markus, H. (1991). Possible selves across the life span. Human Development, 34- 230-255. Crowley, J, E, (1985), Longitudinal effects of retirement on men's psychological and physical well-being. In H, S, Parnes, J. E, Crowley, R. J. Haurin, L, J. Less, W, R, Morgan, F, L, Mott, & G, Nestel, Retirement amona American men (pp. 147-173), Lexington, MA: D, C, Heath, Dennis, H, (1986), Retirement preparation programs: Issues in planning and selection. Journal of Career Development, 13, 30-38. Dittmann-Kohli, F. (1990). The construction of meaning in old age: Possibilities and constraints, Ageing and Society, 10, 279-294, Durrant, A, E. (1988), The effects of preretirement workshop participation on women's retirement knowledge, anxiety levels and planning behaviors, (Doctoral dissertation, Tennessee State University, 1985) Dissertation Abstracts International, 48(12-A), 3074-3075, D'Zurilla, T, J. (1988). Problem-solving therapies, In K, Dobson (Ed,), Handbook of cognitive-behavioral therapies (pp, 85-135), New York: The Guildford Press, Engels, M-L, (1984), The assessment and modification of social skills in older women. Unpublished doctoral dissertation, McGill University, Montreal. Changing images 117 Fisl<e, S. T„ & Tayior, S, E, (1984). Social coanition. New Yorl<: Random House, Fioyd, F, J„ Haynes, S. N., Doii, E, R„ Winemiiier, D„ Lemsi<y, C„ Burgy, T, B„ Werie, M„ & Heilman, N, (1992), Assessing retirement satisfaction and perceptions of retirement experiences, Psvchoioav and Aaina, 7, 609-621, Foii<man, S., & Lazarus, R, S, (1988a), The relationship between coping and emotion: implications for theory and research. Social Sciences and iVIedicine, 26,309-317, Foikman, S., & Lazarus, R, S, (1988b), Coping as a mediator of emotion. Journal of Personality and Social Psvchoioav, 54, 466-475. Fretz, B, R„ Kiuge, N, A., Ossana, S, M., Jones, S. M„ & Merikangas, M, W, (1989). intervention targets for reducing preretirement anxiety and depression. Journal of Counselling Psvchoioav, 36, 301 -307. Gaudry, E„ Vagg, P„ & Spielberger, C, D, (1975). Validation of the state-trait distinction in anxiety research, Multivariate Behavioral Research, iO, 331 -341, Gee, E, M„ & Kimball, M, M, (1987), Women and aaina, Toronto: Butterworths, Glamser, F, D, (1981). Predictors of retirement attitudes. Aaina and Work, 4, 23-29. Glamser, F. D., & DeJong, G. F. (1975). The efficacy of preretirement preparation programs for industrial workers. Journal of Gerontoloav, 30, 595-600. Gloss, G., & Hopkins, K. D. (1984). Statistical methods in Education and Psvchoioav. New Jersey, NY: Prentice Hall. Hooker, K. (1991). Change and stability in self during the transition to retirement: An intraindividuoi study using P-technique factor analysis. International Journal of Behavioral Development, 14, 209-233. Hooker, K., & Ventis, D. G. (1984). Work ethic, daily activities and retirement satisfaction. Journal of Gerontoloav, 39, 478-484. Hultsch, D. F., & Plemons, J. K, (1979), Life events and life-span development, Life-span Development and Behaviour, 2, 1 -35, Jacobson, N. S,, & Revenstorf, D, (1988), Statistics for assessing the clinical significance of psychotherapy techniques: Issues, problems, and new developments. Behavioral Assessment, 10, 133-145. Jacobson, N.S., & Truax, P, (1991), Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psvchoioav, 59, 12-19. Changing images 118 Jaremko, M. E. (1979). A component analysis of stress Inoculation; Review and prospectus. Cognitive Therapy and Research, 3, 35-48. Jewson, R. H. (1982). After retirement; An exploratory study of the professional woman. In M. Szinovacz(Ed.), Women's retirement; Policy implications of recent research (pp. 169-181). Beverly Hills, CA; Sage. Kamouri, A. L., & Cavanaugh, J. C. (1986). The impact of preretirement education programmes on worl<ers, preretirement socialization. Journal of Occupational Behaviour, 7, 245-256. KasI, S. V. (1980). The impact of retirement. In C. L. Cooper & R. Payne (Eds.), Current concerns in occupational stress (pp. 136-185). New Yorl<; Wiley. Kato, K., & Markus, H. R. (1993). The role of possible selves in memory. Psvcholoaia, 36- 73-83. Keith, P. M. (1982). Working women versus homemakers; Retirement resources and correlates of well-being. In M. Szinovacz (Ed.), Women's retirement: Policy implications of recent research (pp. 77-91). Beverly Hills, CA; Sage. Kendall, P. C, & Korgeski, G, P. (1979), Assessment and cognitive-behavioral interventions. Coanitive Therapy and Research, 3, 1-21. Kirsch, I. (1990). Chanaing expectations; A key to effective psychotherapy. Pacific Grove, CA; Brooks/Cole. Lambert, M. J., Shgpiro, D.A., & Bergin, A. E. (1986). The effectiveness of psychotheropy. In S. L. Garfield & A. E. Bergin (Eds.), Handbook of psychotherapv gnd behgvior chonge (pp. 157-211). New York; Wiley. Lanyon, R. L. (1984). Personality assessment. Annual Review of Psvchoiogy, 35, 667-701. Lgzarus, R., & Folkman, S. (1984). Stress, appraisal, and coping. New York; Springer. Lift, IVI. D. (1988). Cognitive mediotors of stressful experience: Self-efficocy and perceived control. Coanitive Therapy and Research, 12, 241 -260. Lohmann, N. (1980). Life satisfaction research in gging: Implications for policy development. In N. Datan & N. Lohmann (Eds.), Transitions of Aging (pp. 27-40). New York: Acodemic Press. Lowe, G. S. (1991). Retirement ottitudes, pigns gnd behgviour. Perspectives, Autumn, 8-17, Lubin, B. (1967). Depression Adiective Checklists; Monugi, Son Diego, CA: EducgtiongI gnd Industrigl Testing Service, Clianging images 119 IVIarkus, H.. & Nurius, P, (1986), Possible selves, American Psychologist, 41, 954-969, Maricus, H„ & Nurius, P, (1987), Motivation and the self-concept, in K. Yardiey & T, Honess (Eds.), Self and identity: Psychosocial perspectives fpp, 157-172), New York; Wiley. Matheny, K, B„ Aycock, D, W„ Pugh, J, L„ Curiette, W, L„ & Cannelia, K, A, S, (1986), Stress coping: A qualitative and quantitative synthesis with implications for treatment. The Counselina Psvchoiogist, 14, 499-549. Matthews, A. M., & Brown, K, H. (1988), Retirement as a critical life event: The differential experiences of women and men. Research on Aaina, 9, 548-571, Matthews, A, M„ &TindaleJ, A, (1987), Retirement in Canada. In K. S, Markides& C. L, Cooper (Eds.), Retirement in Industrialized societies (pp. 43-75). New York: Wiley, McDonald, P, L., & Wanner, R, A, (1990), Retirement in Canada. Toronto: Butterworths, McGhee, V, L, (1987), Preretirement education: Challenge or crisis? In R. D, MacNeil & M, L, Teague (Eds,), Aaing and leisure: Vitality in later life (pp. 322-348). Englewood Cliffs, NJ: Prentice Hall. McGoldrick, A. E., & Cooper, C. L. (1985), Stress at the decline of one's career: The act of retirement. In T, A, Beehr & R. S, Bhagat (Eds,), Human stress and coanition in organizations: An integrated perspective (pp, 177-2011, New York: Wiley. Meade, K., & Walker, J, (1989), Gender equality: Issues and challenges for retirement education, Educational Gerontology, 15, 171-185, Meichenbaum, D., & Cameron, R, (1983), Stress inoculation training: Toward a general paradigm for training coping skills. In D, Meichenbaum & M, E, Jaremko (Eds.), Stress reduction and prevention (pp. 115-154), New York: Plenum. Methot, S. (1987), Employment patterns of elderly Canadians, Canadian Social Trends, Autumn, 7-11, Mirels, J, L,, & Garrett, J, B, (1971), The protestant ethic as a personality variable. Journal of Consulting and Clinical Psychoiogv, 36, 40-44. Myers, J, L, (1979), Fundamentals of experimental desian. Boston, MA: Aiiynand Bacon, Neugarten, B, L, Havighurst, R, J„ & Tobin, S, S, (1961), The measurement of life satisfaction. Journal of Gerontoioav, 16, 134-143. Changing innages 120 Nurius, P. S. (1986). Reappraisal of the self-concept and innplications for counseling. Journal of Counselina Psycholoav, 33, 429-438. Nurius, P. S., Lovell, M., & Edgar, M. (1988). Self-appraisals of abusive parents. Journal of Interpersonal Violence, 3, 458-467. Nurius, P. S., & Majerus, D. (1988). Rethinking the self in self-talk: A theoretical note and case example. Journal of Social and Clinical Psychology, 6, 335-345, Nystedt, L., Smori, J., & Bonnan, M. (1991). Self-schennata: Annbiguous operationalizations of an important concept. European Journal of Personality, 5, 1-14. O'Rand, A., & Henretta,J. C. (1982). Midlife work history and retirement income. In M. Szinovacz (Ed.), Women's retirement: Policy implications of recent research (pp. 123-138). Beverly Hills, CA: Sage. Oyserman, D., & Markus, H. R. (1990). Possible selves and delinquency. Journal of Personality and Social Psychology, 59, 112-125. Palmore, E. (1977). Foctsonaging. Gerontologist, 1 7, 315-320. Poimore, E, (1981), The facts on aging quiz: Parttwo, Gerontoloqist, 21, 431 -437, Palmore, E, (1982), Preparation for retirement. In N, Osgood (Ed,), Life after work (pp, 330-341), New York: Praeger, Palmore, E, B„ Rllenbaum, G, G„ & George, L K, (1984). Consequences of retirement. Journal of Gerontology, 39, 109-116. Pgrnes, H. S., & Nestel, G. (1981). The retirement experience. In H. S. Parnes (Ed.), Work and retirement: A lonaitudinal study of men CPP. 155-197). Cambridge, MA: The MIT Press. Perkins, K. (1992). Psychosocial implications of women and retirement. Social Work, 37, 526-532. Petty, B. J., Moeller, T. P., & Campbell, R. Z. (1976). Support groups for elderly persons in the community. The Gerontoloaist, 6, 522-528. Phillipson, C, & Strang, P. (1983). The impact of pre-retirement education: A lonaitudinal evaluation. Keele, U.K.: Departmentof Adult Education, University of Keele. Poser, E. G., & Engels, M-L. (1983). Self-efficacy assessment and peer group assistance in a preretirement intervention. Educational Gerontoioav, 9 159-169, Changing images 121 Price-Bonham, S., & Johnson, C. K. (1982). Attitudes toward retirement; A comparison of professional and nonprofessional married women. In M. Szinovacz (Ed.), Women's retirement; Policy implications of recent research (pp. 123-138). Beverly Hills, CA; Sage. Riddick, C. C. (1982), Life satisfaction among aging women; A causal model. In M. Szinovacz (Ed.), Women's retirement: Policy implications of recent research (pp. 45-59). Beverly Hills, CA; Sage. RoPerts, P, (1992). I think of Ronald Reagan; Future selves in the present. International Journal of Aaing and Human Development, 34, 91-107. Rule, W. R., & Trover, M. D. (1983). Test-retest reliaPilitiesof State-Trait Anxiety Inventory in a stressful social analogue situation. Journal of Personality assessment, 47, 276-277. Ryff, C. D. (1991). Possible selves in adulthood and old age; A tale of shifting horizons. Psythology and Aging, 6, 286-295. Safran, J. D„ Segal, Z. V., Hill, C, & Whiffen, V. (1990). Refining strategies for research on self-representations in emotional disorders. Cognitive Thergpy and Research, 14, 143-160. Schnore, M. M. (1985). Retirement; Bone or blessing? Wgterloo, ONT; Wilfred Laurler Press. Schlossberg, N. K. (1984). Counseiina adults in transition. Linking prgctice with theory. New York; Springer. Schwartz, R. M. (1993). The idea of balance and integrative psychotherapy. Journal of Psychotherapy Integration, 3, 159-179. Schwartz, R, M„ & Garamoni, G, L. (1989). Cognitive balance and psychopathology: Evaluation of an information processing model of positive and negative states of mind. Clinical Psychology Review, 9, 271 -294. Selye, H. (1984), Stress and health; A perspective on aging and retirement. In A. S. Sethi & R. S. Schuler (Eds.), Handbook of oraonizational stress coping strgtegies (pp. 15-34). Cgmbridge, IVIA; Bgllinger. Smith, C. A., Hgynes, K. N., Lazarus, R. S., & Pope, L. K. (1993). In search of the "hot" cognitions; Attributions, appraisals, and their relation to emotion. Journal of Personality and Social Psychology, 65, 916-929. Smith, C. A„ & Lgzgrus, R. S. (1993). Apprgisgl components, core relgtional themes, and the emotions. Cognition and Emotion, 7, 233-269. Smith, P, C, Kendall, L. M., & Hulin, C. L, (1969). The measurement of satisfaction in work and retirement. Chicago; Rand McNally. Changing images 122 Smith, P. C, Kendall, L M., & Hulin, C. L. (1975). The measurement of satisfaction in work and retirement. Chicago; Rand McNally. Speer, D. C. (1992). Clinically significant change: Jacobson and Truax (1991) revisited. Journal of Consulting and Clinical Psychology, 60, 402-408. Splelberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). IVIanual for the State-Trait Anxiety Inventory ("Self-Evaluation Questionnaire). Palo Alto CA; Consulting Psychologists Press, Splelberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Szinovacz, M. (1982). Introduction: Research on women's retirement, in M. Szinovacz (Ed.), Women's retirement: Policy implications of recent research (pp. 13-21). Beverly Hills, CA; Sage. Taylor, S. E., & Schneider, S. K. (1989). Coping and the simulation of events. Social Cognition, 7, 174-194. Tiberi, D. M., Boyock, V. L., & Kerschner, P. A. (1978). A comporgtive gnalysis of four preretirement education models. Educational Gerontoioav, 3, 355-374. Tinsley, D. J., & Schwendener-Holt, IVI. J. (1992). Retirement and Leisure. In S. D, Brown, & R. Whent (Eds.), IHandbook of Counseling Psychology (pp. 627-662), New York: John Wiley. Toseland, R. (1977). A problem-solving group workshop for older persons. Social Work, 22, 325-236. Turnquist, P. H. (1986). An empirical investigation of pre-retirement planning programs' effect on retirement life. (Doctoral Dissertation, Mississippi State University, 1986). Dissertation Abstracts International, 47 f4-A), 1405-1406. Whitbourne, S. K. (1985). The psychological construction of the life span. In J. E, Birren & K. W. Schaie (Eds.), Handbook of the psychology of oging (2nd ed.) (pp. 594-618). New York; Von Nostrond Reinhold. Wood, v., Wylie, M., & Shegfor, B. (1969). An gnglysis of g short self-report meosure of life sotisfaction: Correlation with rater judgments. Journal of Gerontology, 24, 466-469. Zuckermgn, M., & Lubin, B. (1965). Mgnual for the Multiple Affect Adjective Check List. Son Diego, CA: EDITS. Changing images 123 Appendix A Directions used for the modification to the State-Trait Anxiety inventory DIRECTIONS: A number of statements which people have used to describe themselves are given below. Read each statement and then blacken in the appropriate circle to the right of the statement to indicate how you feel right now, that is, at this moment when thinl<ing about retirement. There are no right or wrong answers, Do not spend too much time on any one statement but give the answer which seems to describe your present feelings best. Changing images 124 Appendix B Coding Manual for Possible Selves Possible selves are images of self in the future, In the current coding scheme, participants' future images of themselves will be coded into five categories; positive, effective, negative, negative active, and miscellaneous, Codina First, read through the response. Second, marl< off each idea unit. An idea unit consists of a complete idea describing an action or category, it may be a whole or part of a sentence, Ideas can be both general and can be elaborations on a general theme. For example, / will travel./1 see myself travelling to Hawaii / and travelling to see my daughter In Kelowna is three ideas as marl<ed. Third, determine which category the idea unit belongs in and mark the unit with P, E, N, NA, or M as appropriate. Beware the danger of over-interpretation. Try to work with the surface meaning of the ideas, The categories are defined as follows: Positive: A desired outcome, or event, includes travelling (this is an outcome rather than working towards a goal) unless specified in more active terms such as travelling through the Andes on a hiking holiday and similarly volunteering rather than volunteering to drive bus for my church group, Note: items in this category are more passive than those in the effective category e.g. happy, relaxed, healthy, active, busy are states rather than actions as is having time for all the things I want. Also includes, reading, being entertained. Effective: Acting upon the world, producing a result or working towards a goal, Implying a sense of agency, e,g„ / see myself breeding dogs or / will work at events in my local community. Also includes entertaining, exercising, playing tennis, taking a specific course. Negative: An undesired outcome or event, e,g,, I fear being In poor health. Negative active: Acting upon the world, producing a negative result or working towards a negative goal, e,g„ / see myself drinking too much for my own good. Miscellaneous: A complete idea that does not fit into the above categories Fourth, sum the number of Ps, Es, Ns, NAs, and Ms, Mark on the sheet. Inter-rater reliability. With a second coder, go over the responses and codes and indicate the number of agreements out of the total number of idea units. Changing images 125 For disagreements, reach consensus and mark any corrections on the master copy. IVIake any changes necessary to the coding manuai. Changing innages 126 Appendix C Demographic S h e e t 1 . What i s y o u r d a t e of b i r t h ? D M Y 2. How long have you worked? (Please check one) Less than 5 years 5 to 10 years 10 to 20 years More than 20 years 3. When do you plan to retire? (Please check one) In 0 to 3 months In 3 months to 1 year In 1 to 2 years In more than 2 years 4. What is your reason for retiring? (Please check one) Want more time for leisure activities Feel pressure from employer to retire Health reasons Spouse is retiring Want more time for volunteer activities Reaching age 65 Other (please specify) 5. Are you currently participating in a counselling program? Yes No Changing images 127 Appendix D Code: Expectancy Questionnaire How effective do you believe these sessions will be for helping you to cope with retirement? (Please circle one item) 1. Very ineffective 2. Somewhat ineffective 3. Uncertain 4. Somewhat effective 5. Very effective How likely are you to recommend this programme to a friend? 1. very unlikely to recommend it 2. Somewhat unlikely to recommend it 3. Uncertain 4. Somewhat likely to recommend it 5. Very likely to recommend it Changing images 128 Appendix E Manipuiation checl< on treatment Fidelity Treatment fidelity was further assessed by obtaining expert judgements on a sample of videotapes of treatment sessions. A counsellor who had not been involved with the project was given the treatment manuals from the targeted change and structured discussion seminars. She then viewed two tapes from each group for session three and made a judgement regarding which of the two groups the taped sessions were from. Each of the tapes was correctly identified. Changing images 129 Appendix F CODE: Programme Evaluation Questionnaire RETIREMENT PLANS 1. What are your goals for retirement? (Responses included the following: To maintain health To continue to be mentally active and challenged To promote continuing close relationships. To be productively active ) 2. In what ways will you cope with retirement? (Responses included the following: Try to fill my time with useful + enjoyable hobbies & activities) 3. Using the following scale, rate how you view retirement. 1 2 3 4 5 Very Somewhat Very Somewhat positive positive neutral negative negative (58.89%) (41.12%) (0%) (0%) (0%) 4. What has the retirement group meant to you? (Responses included the following: -reassured me it was OK to feel apprehensive about retirement. -helped me overcome some of my apprehension -helped me focus more positively about retirement, -gave me a good feeling to belong to a group that was made up of such nice people) Changing images 130 RETIREMENT GROUP: EVALUATION What aspects of the group were useful to you? (Responses included the following: -sharing feelings -presentation of the group leaders) 2. What changes would you recommend for the next group? (Responses included the following: -fewer total sessions) 3. How effective do you believe these sessions were for helping you to cope with retirement? (Please circle one item) 1 Very ineffective 2 Somewhat ineffective 3 Uncertain 4 Somewhat effective 5 Very effective 4. How likely are you to recommend this programme to a friend? 1 Very unlikely to recommend it 2 Somewhat unlikely to recommend it 3 Uncertain 4 Somewhat likely to recommend it 5 Very likely to recommend it Changing images 131 5. Using the following scale, how would you rate the following aspects of the group? 1 2 3 4 5 6 7 8 9 10 Not Very Useful Useful Group support Group leaders Group discussion 'Homework' Retirement model Specific topics: Problem-solving Self-statements Books and materials Ageism/stereotypes Family relationships 6. Do you have any comments about the group, the material, the leaders ? (Responses inc luded the f o l l o w i n g : very vocal participation by the group. The leaders set up a comfortable atmosphere which made it easy to talk about our concerns and helped us to gain insight into a sensible self-actualizing approach to problem solving.) Changing images 132 Appendix G IVI id-treatment means and standard deviations Variable Group M Positive possible selves Effective possible selves Negative possible selves Retirement self-efficacy Targeted-change Structured-discussion Targeted-change Structured-discussion Targeted-change Structured-discussion Targeted-change Structured-discussion 4,89 4,78 2.63 2,89 ,26 .43 78,98 74,20 2,89 3,07 2.65 2.99 .28 .37 12.52 13.91 27 27 27 27 27 27 28 28 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0053960/manifest

Comment

Related Items