1 SELF-DEFINING AUTOBIOGRAPHICAL ME MORY IN RELATION TO ADULT SELFCONCEPT AND PSYCHOLOGICAL WELL-BEING BY ANGELENIA B. SEMEGON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006
2 Copyright 2006 By Angelenia B. Semegon
3 To my greatest teacher, my son, Shane Aron Semegon.
4 ACKNOWLEDGMENTS Many people were involved in bring this projec t to fruition. Each person contributed in his or her own special way and I deeply appreciate the assistance each of you gave. There are a few people whose contributions were such an integral a nd on-going part of this project that I want to acknowledge them directly here. First, on a professional level, Iâ€™d like to express my appreciation to each of my mentoring committee members; Manfred Diehl, Pat Kricos, Christina McCrae, and Robin West for their guidance and support throughout my te nure as a UF aging trainee . I would also like to thank each of my dissertation committee members; R obin West, Manfred Diehl, Susan Bluck, Jeff Farrar, Greg Neimeyer, and Monika Ardelt. I appreciate your input and guidance throughout the dissertation process and especially your additional input associat ed with the NRSA proposal. In addition to his guidance with the dissertation process, I would also like to express my deepest gratitude to Manfred Diehl for assistance, gui dance and energy associated with the NRSA project. It was an invaluable experience that I learned a great deal from. His guidance was essential in obtaining the grant. I am deeply appreciative of the time and energy Robin West devoted to the dissertation pro cess. Robin was not only a dedicat ed advisor for my dissertation, but also a role model. The support of my friends and family during th is process was invaluab le. I do not believe I would have completed this degree without them by my side. I am forever grateful their love and support. Nicole Alea, Laura Curry, Alissa Da rk-Freudeman and Jacqueline Baron were and always will be my â€œdissertati on sistersâ€ whose support was pricel ess. Kathy Bergâ€™s calm gentle presence was a tremendous source of strength fo r me. Jacqueline Whitmore provided a wealth of encouragement and helped me stay focused. Candy Tierney was there beside me sharing both the
5 discouraging as well as triumpha nt moments . Shane, you were th ere through it all, my light and inspiration.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............11 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW..............................................................13 Self-Concept Theory............................................................................................................ ...13 Development of Self-Concept.........................................................................................14 Self-concept in infancy and toddlerhood.................................................................15 Self-concept in early childhood................................................................................16 Self-concept in middle childhood............................................................................16 Self-concept in adolescence.....................................................................................18 Self-concept in adulthood.........................................................................................20 Autobiographical Memory Theory.........................................................................................25 Development of General Memory...................................................................................27 Development of Autobiographical Memory...................................................................28 Autobiographical memory in childhood..................................................................29 Autobiographical memory in adolescence...............................................................33 Autobiographical memory in adulthood..................................................................34 Functions of Autobiographical Memory Across the Life Span.......................................38 Interpersonal, knowledge-based, an d intrapersonal functions of autobiographical memory.....................................................................................38 The identity function of autobiographical memory..................................................41 Age variation in the functions of autobiographical memory . ...................................43 Central Concepts in the Current Research..............................................................................44 Autobiographical Memory a nd Psychological Well-being.............................................45 Autobiographical Memory, Personality, and Affect.......................................................47 Objectives of the Current Research........................................................................................50 Self-Representations and Self-Defining Memories.........................................................51 Relationships among Valence of Self-def ined Memories, Individual Differences, and Well-Being............................................................................................................52 Valence as a Mediator.....................................................................................................55 Hypotheses of this Study....................................................................................................... .57 Self-Representations and Self-Defining Memories.........................................................57 Valence as a Mediator.....................................................................................................59 2 METHOD......................................................................................................................... ......61
7 Design and Procedure........................................................................................................... ..61 Participants................................................................................................................... ..........61 Sample Size and Power Considerations.................................................................................62 Measures....................................................................................................................... ..........63 Self-Defining Memory Task............................................................................................63 Qualities of Self-Defining Memory.................................................................................64 Coding of Memory Narratives.........................................................................................64 Semantic Representations of Self-Defining Memories . ..................................................66 Semantic Recognition Task.............................................................................................66 Self-Descriptor Rating Task............................................................................................67 Control Measures.............................................................................................................67 Individual Difference Measures......................................................................................70 3 RESULTS........................................................................................................................ .......74 Preliminary Analyses........................................................................................................... ...74 Covariates..................................................................................................................... ..........76 Auditory Verbal Learning Task.......................................................................................76 Reaction Times................................................................................................................77 Vocabulary..................................................................................................................... .77 Hypothesis 1: Relationship between Adultsâ€™ Self-Representations and Self-Defining Autobiographical Memories...............................................................................................78 Hypothesis 2: Relationship between Vale nce of Self-Defining Autobiographical Memories, Individual Difference Variab les, and Psychological Well-being.....................80 Hypothesis 3: The Mediational Effect of Positive Valence of Self-Defining Memories and Individual Difference Variab les on Psychological Well-being....................................82 Individual Difference Variables Entered First................................................................83 Individual Difference Variables Entered Last.................................................................84 4 DISCUSSION..................................................................................................................... ....89 Objective 1Self-Representati ons and Self-Defining Memories...........................................91 Response Latencies of Attributes Associat ed with Self-Defining Autobiographical Memories.....................................................................................................................92 Age Differences and Response Latencies.......................................................................93 Salience of Semantic Representa tions Derived from Self-Defining Autobiographical Memories........................................................................................95 Age Differences and Salience.........................................................................................96 Objective 2: Relationships among Valence of Self-defined Memories, Individual Differences, and Psychological Well-Being.......................................................................96 Valence of Self-Defining Memories a nd Individual Difference Variables.....................98 Valence of Self-Defining Autobiogra phical Memories and Psychological WellBeing.......................................................................................................................... ..99 Individual Difference Variable s and Psychological Well-Being..................................100 Objective 3Valence as a Mediator......................................................................................101 Study Limitations and Future Directions..............................................................................104 Limitations.................................................................................................................... .104
8 Future Directions...........................................................................................................109 Conclusions.................................................................................................................... .......113 APPENDIX A MEMORY QUESTIONAIRE..............................................................................................117 B CODING SCHEME SELF-DEFINING MEMORIES.........................................................119 Key Terms Associated with Coding.....................................................................................119 Coding the Memory Narratives............................................................................................120 Implicit Self-Attr ibute Coding..............................................................................................122 Creating The Complete Self-Attribute Li st For The Semantic Recognition Task...............125 Key Terms Associated with th e Semantic Recognition Task...............................................125 Using the Self-Attribut es Synonym Dictionary....................................................................126 Preparing the List for Use in Visual Basic...........................................................................127 LIST OF REFERENCES.............................................................................................................128 BIOGRAPHICAL SKETCH.......................................................................................................137
9 LIST OF TABLES Table page 2-1 Means and Standard Deviations of So ciodemographic and Health Variables by Age Group.......................................................................................................................... .......73 2-2 Means and Standard Deviations Indi vidual Difference Variables by Age Group.............73 3-1 Bivariate Correlations betw een the Covariates (N=120)...................................................86 3-2 Mean Response Latencies by Age Group and Attribute Type..........................................86 3-3 Relationship Among Positive Valence of Self-Defining Memories, Individual Difference Variables, and Psyc hological Well-Being (N=120)........................................87 3-4 Summary of Regression An alyses Predicting Psychologica l Well-being by Individual Difference Variables and Positive Valence of Autobiographical Memories.....................88
10 LIST OF FIGURES Figure page 1-1 Proposed Model of the Mediational Re lationship Among Positive Valence of SelfDefining Autobiographical Memory, Indi vidual Difference Variables, and Psychological Well-Being.................................................................................................60 2-1 Order of Test Administration.............................................................................................73
11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SELF-DEFINING AUTOBIOGRAPHICAL ME MORY IN RELATION TO ADULT SELFCONCEPT AND PSYCHOLOGICAL WELL-BEING By ANGELENIA B. SEMEGON December 2006 Chair: Robin West Co-Chair: Manfred Diehl Major Department: Psychology The purpose of this study was to examine th e link between individua lsâ€™ self-concept and autobiographical memory in relation to psyc hological well-being. Ther e were three primary objectives: 1) to explore the rela tionship between adults â€™ self-representations and self-defining autobiographical memories, in relation to age; 2) to investigate the relationship of valence of self-defining autobiographical memories to psyc hological well-being, personality and affect; and 3) to test the potential role of valence of self-defining memori es as a mediator of the relationship between individual differe nces and well-being. One hundred twenty-three adults participated in the study. Participants were tested individually in two separate te sting sessions. During the first sessi on, participants narrated three self-defining autobiographical memories, from which personal attri butes were identified (memory-derived attributes). Participants rate d the valence of these memories. In addition, several individual difference variables and a m easure of psychological well-being were also administered in the first session. During a s econd session, a semantic recognition task was administered in which response latencies were ob tained for the memory-derived attributes and general self-attributes. In additi on, participants rated the salien ce of both types of attributes.
12 Both younger and older adults responded more qu ickly to memory-derived attributes than to general self-attributes. Als o, older adults responded more qu ickly to general self-attributes than younger adults. Additionally, the memory-derived attributes were more salient than general self-attributes for both younger and older adults, a nd memory-derived attributes were higher in salience for older adults than younger adults were. In contrast, the salience of general selfattributes were comparable for younger and older adults. Several notable relationships were found among the valence of autobiographical memories, individual difference variables and psychologi cal well-being. The positive valence of selfdefining autobiographical memories predicted a small portion of the variance in scores on psychological well-being. These findings show promise for future i nvestigations of th e connection between autobiographical memory and self-concept. Additionally, examining the relationship among autobiographical memory, individual differences in personality and a ffect, and psychological well-being may be a fruitful ave nue for future research explori ng the role of i ndividualsâ€™ selfconcept as a resource that allows older adults to compensate for age decrements in ways that optimize well-being.
13 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW The purpose of this study was to examin e the link between two major theoretical frameworks, self-concept theory and autobiographical memory th eory. Researchers have studied self-concept and autobiographical memory extensiv ely, yet separately. In this study, a semantic recognition paradigm is describe d as an innovative method to inve stigate the relationship of autobiographical memory and self -concept across the life span. A life span perspective and a multidimensional view of self-concept provide the overarching theoretical approaches for this review . In addition, a review of the autobiographical memory literature is included, highlighting the deve lopmental trajectory, as well as the functions of autobiographical memory. Specifically, the f unctions of autobiographical memory will be emphasized. This review will provide the f oundation necessary to further consider the interconnected relationship of these two perspectives. Self-Concept Theory Theorists often use the terms â€œself-conceptâ€ , â€œself-perceptionâ€, â€œself-knowledgeâ€, â€œselfesteemâ€, and â€œself-representationâ€ interchangeably despite subtle differences in their meaning (Baumeister, 1998). In the context of this study, I have adopted the conceptualization offered by a number of prominent theorists according to which the self-concept is viewed as a multidimensional and dynamic cognitive structure that serves a number of adaptive and selfregulatory functions (Brandtstdter & Greve, 1994; Higgins, 1996; Markus & Wurf, 1987). This conceptualization implies that individualsâ€™ self-concept is an â€œorganized knowledge structure that contains traits, values, epis odic and semantic memories about the self and that controls the processing of self-relevant informationâ€ (Cam pbell et al., 1996, p.141). Theorists have most often conceptualized these multi-faceted aspects of the self as self-schemas (Markus, 1977), self-
14 theories (Epstein, 1973), or self -digest (Higgins, 1996). Each of th ese conceptualizations focuses on the cognitive representation of the self that em bodies the sum of an i ndividualâ€™s collection of knowledge about his or her own person. This selfknowledge arises through social interaction as well as interaction with the environment resulting in a personal theory or mental concept of what the self is like. Regardless of how theorists conceptualize the se lf-concept, they seem to agree that it is not static and that it is constructed and reconstructe d throughout an individualâ€™s lifetime to meet the personâ€™s changing psychological need s and the demands of varying life contexts. The evolving multidimensional nature of the self-concept is particularly apparent during the formative years of childhood and adolescence (Damon & Hart, 1988; Harter, 1998); however, the multidimensional nature of the self-concept continues to be evident across the entire lifespan (Baumeister, 1998; Brandstdter & Greve, 1994). Development of Self-Concept The sequence of self-concept development fr om infancy through adolescence has been well documented (Case, 1992; Damon & Hart, 1 988; Fischer, 1992; Ha rter, 1998; Harter & Monsour, 1992; Lewis & Brooks-Gunn, 1991). An ex tant body of research (Damon & Hart, 1988) has shown that during early childhood, the self-concept is mostly based on a childâ€™s physical characteristics. As the child matures, social characteristics are inco rporated into the selfconcept and, increasingly, psyc hological characteristics are drawn upon in later childhood and adolescence (Damon & Hart, 1988; Harter, 1998). Thus, by the end of adolescence, individualsâ€™ self-concept has developed into role-specific multiple selves (Harter & Monsour, 1992). One of the challenges of early adulthood relates to the coordination of these multiple selves into a coherent overall self-representation (Dona hue, Robins, Roberts, & John, 1993). Given that self-concept is based on cognitive representations, developmental changes in childrenâ€™s self-concept ar e intimately connected to cognitive development. Therefore, it is
15 imperative to consider cognitive developmen t as an underlying foundation upon which selfrepresentations are constructed. Likewise, developmental changes in cognitive resources are also associated with changes in the structure and content of individualsâ€™ self-concept. The following section focuses on the role of cognitive developm ent with regard to developmental differences related to both the content and structure of children and adolescentsâ€™ conceptions of self. Self-concept in infancy and toddlerhood The earliest conceptualizations of self as an object become apparent as infants evince the ability to recognize themselves. Lewis and Br ooks-Gunn (1979) noted lim ited self-recognition in 9-12 month-old infants when provided with conti ngency cues, those that rely on the babiesâ€™ own movements. However, when the cues were noncontingent, such as delayed videotape or photographs, the infants in that age group did not demonstrate self-recognition. Furthermore, Lewis and Brooks-Gunn (1979) found that, by 18 months of age, children be came less interested in contingency cues and more attentive to nonc ontingent cues. At 21 months of age, most children no longer rely on contingency cues for self-recognition. A lthough inter-individual variability is evident in self-recognition tasks, nonetheless self-recognition is well established in children by two years of age (Lewis & Brooks-G unn, 1979). Establishing the hallmark of selfrecognition is integral to estab lishing the emergence of the indivi dualâ€™s sense of self. Lewis and Brooks-Gunn (1979) assert, â€œknowledge of oneself in an existential se nse is necessary for knowledge of oneself in a categorical senseâ€ ( p. 318). That means, when a child recognizes him/herself in photographs, or videos the child is displaying knowledge th at he or she exists outside of the parameters of the present moment . This knowledge of self allows the child to incorporate more complex ideas and information about himself or herself. From these early beginnings, the complexity of th e childâ€™s self-concept develops along a predictable trajectory over the course of childhood.
16 Self-concept in early childhood During early childhood years, a distinct shift in the childâ€™s understand ing of self occurs. Specifically, the child becomes in creasingly aware of his or her individual characteristics that can be directly observed (Damon & Hart, 1988) . These observable characteristics are the foundation of the young childâ€™s conception of se lf. The young childâ€™s self-concept is very concrete (Damon & Hart, 1988; Harter, 1996) and cons ists of discrete characteristics related to his or her physical (i.e., â€œI have curly hairâ€), social (i.e., â€œJamie is my friendâ€), activity related (i.e., â€œI like to play soccerâ€) and psychological (i.e., â€œIâ€™m kindâ€ ) attributes. Children include the physical, social, and activity related traits before they begin to include rudimentary psychological characteristics (Harter, 1983). In addition to having very concrete concep tualizations of themselves, young children are incapable of maintaining more than one categor y or behavioral event simultaneously (Case, 1992; Fisher, 1980; Harter, 1988). Therefore, children tend to describe themselves in a global allor-none fashion and are unable to recognize or include in their self -descriptions traits that may co-occur. For example, children evince this all-or -none conception of self when they describe themselves in terms of being all good or all bad. Because of this tendency to globalize, childrenâ€™s conceptions of self are often generalized acro ss many contexts. Similarly, Fisher (1980) noted that young children were incapable of holding tw o opposing attributes si multaneously. Studies that are more recent have also shown that children are unable to acknowledge having more than one emotion at the same time (Ruble & Dweck, 1995) , even if the emotions are of the same valence (Harter, 1998). Self-concept in middle childhood During middle childhood, children continue to ma intain a relatively concrete conception of self (Harter, 1998). However, the self-concept undergoes changes in both its structure and
17 content (Fisher, 1980; Hart er, 1988; Ruble & Dweck, 1995) . One of the structural changes that occur during middle childhood is seen when children create unidirectional a ssociations that are connected to one another (Fisher, 1980). Fisher (1980) calls this proc ess representational mapping. Harter (1988) expounds upon the role of representational mapping in middle childhood by noting that children often engage in repres entational mapping that involves the linking of various forms of opposites within a given concept. For example, according to Harter (1998) â€œin the domain of physical concepts, children can oppose up and down, taller and shorter, thinner and wider, although they cannot yet demonstrat e the reversible operations necessary for conservationâ€ (p.569) . Representational mappings also play an importa nt role with regard to the content of childrenâ€™s self-concept (Harter, 1998). The nature of representational mappings, particularly those associated with categorizi ng opposites, combined with the childâ€™s inability to hold two opposing attributes simultaneously, contributes to young childrenâ€™s overly optimistic views of their own person (Harter, 1998). Children exagge rate the opposing catego ries â€œgoodâ€ and â€œbadâ€ and demonstrate a proclivity to see themselves and others in a positiv e light (Harter, 1988, 1998). Essentially, during middle childhood, children ar e only capable of ackno wledging the positive aspects of their self-concept. Th is one-dimensional thinking conti nues to prevent the child from incorporating opposing attributes into the self-concept (Harter, 1998). Cognitive development in late childhood is reflected in the increasin gly more complex conceptions of self. Self-concept in late childhood During late childhood, changes in the struct ure and content of the self-concept are associated primarily with cogni tive developments that allo w children to engage in twodimensional thinking. Harter (1998) asserts that the â€œcoordination of sel f-representations that
18 were previously differentiated or considered to be oppositesâ€ (p.571) is a principal advantage of two-dimensional thinking . Children begin to combine various aspects of themselves as well as generalize behavioral instances into broader definitions of self. This allows the child to begin to create trait labels . Incorporating behavioral instances from two or more related activities allows the child to create a generalized self-attribute or knowledge about a trait. For example, a child may combine beliefs about successe s as a soccer player along w ith successes playing tennis to form a generalized belief that he or she is a talented athlete. In a similar manner, twodimensional thinking allows children to acknow ledge both their success as well as their limitations. This permits children to create more realistic evaluations of their own abilities (Harter, 1998). During late childhood, two-dimensi onal thinking also allows children to recognize their capacity to have more than one emotion at the same time (Harter, 1983). For example, children visiting grandparents may acknowledge feelings of happiness to see grandparents and at the same time express feelings of sadness about missing a pet left for the duration of the visit. Unlike in early and middle childhood, during late chil dhood, children incorporate positive as well as negative emotions, evaluations, and aspects of personality into th eir sense of self. As children enter adolescence, cognitive development allows adolescents to conceptu alize increasingly more abstract views of self-concept. Self-concept in adolescence During adolescence, changes in the structure a nd content of the self-concept are primarily associated with cognitive developments that allow adolescents to engage in abstract thinking (Case, 1985; Damon & Hart, 1988; Harter, 1998). Adolescents apply abst ractions not only to themselves but also to others and their envir onment (Harter, 1998). This significant change in
19 cognitive development provides the foundation for an increasingly complex vi ew of self (Harter, 1988). Cognitive advances allow adolescents to hold opposing views of themselves and to understand that the way they view themselves may vary as the situation changes. For example, adolescents may see themselves as both introverted and extroverted by thinki ng of situations that led to feelings of shyness along with other instances in which they felt outgoing (Damon & Hart, 1988). Young adolescents do not tend to feel distre ssed by these discrepant views of self (Harter & Monsour, 1992). Harter and Monsour (1992) assert that although young adolescents are capable of creating abstract constructs of them selves, they are not yet capable of holding and comparing multiple self-constructs concurrently. Therefore, they do not yet experience conflict in thinking about these dispar ate self-attributes. Adolescents become capable, during midadolescence, of conceptualizing multiple self-constructs and also of comparing these different views of self (Harter & Monsour, 1992). A tendency to expe rience feelings of psychological distress often accompanies this aw areness of discrepant views of self (Harter & Monsour, 1992). In some cases, this distress may result in the adolescent reverting to the all-or-none thinking typical of earlier ages (Harter, 1998). The ability to integrate th ese multiple and often disparate conceptions of self does not fully develop unt il late in adolescence (Harter & Monsour, 1992). Over time, adolescents begin to understand that although they may have different views of themselves according to the situation, a core part of the self remains consistent across situations (Harter, 1998). This awareness allows older adolescents to acknowledge their discrepant conceptions of self in a manner that helps th em resolve any psychological conflict associated with the previous stage. As this developmental process occurs, individual sâ€™ self-concepts become
20 progressively more differentiated as well (Har ter & Monsour, 1992), as one would expect for adults. Self-concept in adulthood Although the different steps of self-concept developmen t are well documented from childhood through late adolescen ce and early adulthood (Damon & Hart, 1988; Harter, 1998), similar knowledge is still limited for the year s beyond early adulthood. However, there has been an increased emphasis on the role of the se lf-concept in adult development and aging (Brandtstdter & Greve, 1994; Diehl, Hastings , & Stanton, 2001; Fre und & Smith, 1999; Markus & Herzog, 1991). Several theorists have suggested that, as a dynamic cognitive structure, adultsâ€™ self-concept may become an important resource in negotiating the challenges associated with the aging process with respect to maintaining psyc hological well-being (Brandtstdter & Greve, 1994; Cross & Markus, 1991). Variations in the content, structure, and organization of the selfconcept have increasingly been examined to document the resourcefulness of the self in adjusting to the challenges of adult development and aging (Brandtstdter & Greve, 1994; Campbell et al., 1996; Cross & Markus, 1992; Diehl et al., 2001; Donahue et al., 1993; Freund & Smith, 1999; Higgins, 1996; Kling, Ry ff, & Essex, 1997; Linville, 1987). For instance, Brandtstdter and Greve (1994) ha ve proposed that, as adults are faced with threats and challenges to their construal of self , the adaptive nature of the self-concept plays a crucial role in normal aging. One way in which in dividuals avoid the negativ e impact of aging is directly linked to the maintenance of a more positive self-concept (Brandtstdter & Greve, 1994). In particular, Brandtst dter and Greve (1994) have described assimilative, accommodative, and immunizing processes that they believe are at work as individuals negotiate the challenges associated with the aging proce ss. These three processes allow the individual to
21 preserve a positive sense of self by minimizing discre pancies between an individualâ€™s objective self and his or her normative sense of self. Assimilative processes are those associated w ith action-oriented steps an individual takes in order to achieve and/or maintain a positive sense of self. Adults employ assimilative processes such as modifying their activitie s in a manner that allows them to continue preferred current roles (e.g., playing a sport for shorter periods of time). Productive use of assimilation is dependent on both the resources av ailable as well as the individua lâ€™s perceptions of control and self-efficacy with regard to outcomes. Accommodative processes are those that involve changing oneâ€™s goals and aspirations to accommodate the current situation, including m odifying expectations for success. These processes rely on the individualâ€™s cognitive flexib ility with regard to making adaptive changes that reflect the individualâ€™s current abilities as well as his or her beliefs about the desirability of particular goals. Campbell and her colleagues (1 996) have also suggest ed that decreases in aspirations may account for the increase in lif e satisfaction found in most older adults. Immunizing processes are those that offe r the individual protection from possible discrepancies between the actual self and the self the person be lieves he or she should have achieved. In order to maintain a positive view of self, individual s use immunizing processes such as reevaluating the diagnosticity of particular skills as indicators of success in a particular social arena or cognitive domain. These processes rely on an individualâ€™s willingness to alter beliefs. Immunizing processes are also affected by the st rength of the social norms that drive selfreferential beliefs (Greve & Wentura, 2003). Cross and Markus (1991) have also examined the dynamic nature of individualsâ€™ selfconcept via the reorganization of self-representations in the form of possible selves. Cross and
22 Markus (1991) use the term â€œpossible selvesâ€ to describe individuals â€™ representations of themselves in the future. Possible selves are re presentations of what a person might want to become or characteristics the person might want to develop, as well as roles or characteristics that a person is afraid of developing in the fu ture. Adults reporting hi gh life satisfaction depict different possible selves than those reporting lower life satisfaction (Cross & Markus, 1991). Cross and Markus (1991) assert that individualsâ€™ possible selves act as a psychological resource by guiding behavior. Individuals at tempt to achieve desired states and avoid undesirable states. An individualâ€™s conception of his or her possibl e self also provides an evaluative context allowing the individual to either protect or defend present self -representations (Cross & Markus, 1991). Structural organizations of th e self-concept, such as self-c oncept clarity (Campbell et al., 1996), self-concept differentiation (D iehl et al., 2001; Donahue, et al., 1993), or self-complexity (Linville, 1987), provide the indivi dual with valuable resources in the processes of normal aging. Constructs such as self-concept differentiation (Diehl et al., 2001; Donahue et al., 1993) and selfcomplexity (Linville, 1987) have been li nked to psychological we ll-being. Diehl and his colleagues (2001) found that psychological wellbeing was associated with self-concept differentiation for adults. Specifically, higher levels of self-concept differentiation were associated with lower levels of psychological well-being. Self-complexity, as defined by Linville (1987), is an individualâ€™s ability to hold numerous â€œselfaspectsâ€ while preserving â€œgreater distinction among those self-aspectsâ€. Greater de grees of self-complexity act to safeguard the individual faced with stressful circumstances against depression as well as other stress related illnesses (Linville, 1987).
23 In a similar way, Campbell and her colleagues (1996) found evidence of a strong relationship between self-c oncept clarity and a number of other psychological variables, such as self-esteem, anxiety, insightfulness into internal processes, Neuroticism, Conscientiousness, and Agreeableness (as measured by the NEO-FFI, Co sta & McCrae, 1989). Specifically, Campbell, Assanand, and di Paula (2003) found that structural organizations of self that reflect a united view of oneself, such as self -concept clarity and se lf-concept differentiati on, were especially important to psychological well-being. Furthermore, several studies suggest that older adultsâ€™ self-rep resentations may be distinctively different from the self -representations of other age groups. Specifically, these studies document the relationship between variations in self-repre sentations at different points across the lifespan and psychologi cal well-being. Diehl and colleag ues (2001) provided evidence of the resourcefulness of the self-concept in regards to ag ing by examining self-concept differentiation and role-specifi c self-representations in young, middle-aged, and older adults. Diehl and his colleagues (2001) found that althou gh lower levels of self-concept differentiation were associated with higher levels of positive psychological well-being for both older and younger adults, this effect was more pronounced for older adults. In a study examining the content and function of older adultsâ€™ self-defining attribut es, Freund and Smith (1999) found more similarities than differences. Among 24 vari ous categories, significant differences were noted in only five domains in the content of young-old and old-old adultsâ€™ self-defining attributes . Freund and Smith noted that old-old adults defined themselves using the following categories â€“ family/relatives, interests and hobbies outdoors, interest reflecting social participation, and interpersonal style â€“ less frequently than yo ung-old adults. In addition, in terms of self-definition, the young-old adults used daily living routines less often than the old-old
24 adults. Freund and Smith also noted that life review played a role in older adultsâ€™ self-definition (1999). Additionally, Ryffâ€™s (1991) resear ch showed that adults of different ages use their possible selves to adjust their evaluations and aspirations in accordance with the developmental tasks of a specific age period, including adjustments to age-related loss. Ryff (1991) found a marked difference in the way younger, middle-aged and older adults view themselves. As adults make evaluations of their current levels of functioning, and compare them to their memories of earlier evaluations, they gain an overa ll sense of decline or improvement (Ryff, 1991). Personal progress is often reflected in the memories of younger and middle-aged adults, with maintenance of previous levels of functions more often re flected in older adults â€™ memories of earlier evaluations (Ryff, 1991). Taken together, these studies show the adaptive role of adultsâ€™ self-representations and their relations to various psychological outcomes. Notably, older adults may reconstruct previous self-representations in order to maintain a po sitive sense of self, functioning, and psychological well-being. Although empirically the role of the self-concept as a resource in adult development and aging has been established, (Campbell et al., 1996; Cross & Markus , 1992; Diehl et al., 2001; Donahue et al., 1993; Freund & Smith, 1999; Linville, 1987) taking such a perspective raises the question whether all self-re presentations are equally important. In accordance with Brandtstdter and Greve (1994), it seems reasonable to assume that not all self-representations are of equal importan ce. Specifically, Brandtst dter and Greve (1994) identified three criteria that determine whether self-related cognitions and attributes become part of a personâ€™s self-concept and, hence, a resour ce in the aging proce ss. These criteria are continuity and permanence, discriminative relevance, and biographical meaningfulness
25 (Brandtstdter & Greve, 1994). Cont inuity and permanence refer to those attributes that give a person a sense of sameness and identity over time. Discriminative relevance refers to selfattributes that provide the indi vidual with a sense of uniqueness a nd distinguish him or her from others. Finally, only self-attributes that the person considers impor tant and meaningful for his or her biography will become relevant in negotiating age-related challenges. These qualifications imply that not all self-attributes are equally important in the ag ing process, but that only selfattributes that meet these three criteria will play a prominent role in adult development and aging (cf. Markus & Wurf, 1987). Furt hermore, Brandtstdter and Grev e (1994) maintain that although continuity and permanence as well as discrimi native relevance are both necessary features, no feature is sufficient by itself. Brandtstdter and Greve (1994) assert that it is biographical meaningfulness that is the primary factor in determining whether attributes become part of a personâ€™s core sense of self. Therefore, it follows that autobiographical memori es, those memories that are part of a personâ€™s biography, become an essential component of individualsâ€™ self-repres entations over time (Campbell et al., 1996). In the following section, I will review the litera ture on autobiographical memory with particular emphasis on the link betw een autobiographical memo ry and adultsâ€™ selfrepresentations. Autobiographical Memory Theory This section focuses on relevant literature detailing the developmental trajectory of autobiographical memory, includi ng a brief review of the development of general memory as a precursor to autobiographical memory. This is followed by a review of the autobiographical memory literature as it pertains specifically to adults. Age-related differences in autobiographical memory as well as differences in the charac teristics of autobiogra phical attributed to methodology, are included. The functions of auto biographical memory are highlighted, with a
26 particular emphasis on the identity functi on. Although there are ma ny other aspects of autobiographical memory, a complete review of all aspects of autobiographical memory, such as the veracity of autobiographica l memories, or the point in ti me at which autobiographical memory emerges, is beyond the scope of this thesis. Autobiographical memory is a form of epis odic memory that incorporates personally experienced events to create an individualâ€™s personal past (Nei sser, 1988). A number of scholars have conceptualized multiple layers of autobiographical memory. This multilayered conception is important because these layers may be relate d to the self-concept in different ways. Brewer (1986) identified four sub-categories of autobiographical memory: personal memory, autobiographical facts, generic personal memory , and memories that make up a personâ€™s selfschema. According to Brewer (1986), an indivi dualâ€™s personal memory includes recollections about specific events and experi ences in his or her past. This type of memory may include affective components along with visual imagery a ssociated with a particular event. For the individual, these personal memories are unique in that they refer not only to a particular event, but also to a specific occasion. Ge neric personal memories are pe rsonal memories of events or experiences that are repeated often, although they vary somewhat and eventually become less distinct. Brewer (1986) asserts th at individuals maintain a large body of generic autobiographical information that they use to define their beliefs about themselves. These beliefs are not endemic to a particular event or experi ence. A personally experienced even t that results in an individual having information that is known about the event that does not include visual imagery, or may not even include explicit recall of the experi ence, per se, falls under the classification of autobiographical fact.
27 Like generic personal memories, memories that make up the self-schema arise from events that are repeated often. However, like autobiog raphical facts, memories that make up a personâ€™s self-schema do not include visual imagery. Therefor e, multiple instantiations of experiences that are very similar and do not include imagery re sult in a general body of knowledge about oneself. It is clear that such memories begin forming at a very early age and this process continues to form throughout an individualâ€™s entire life. In or der to examine the developmental trajectory of autobiographical memory, it is im portant to consider the development of general memory as an antecedent to autobiographical memory. Development of General Memory Even infants show the ability to process and later recall inform ation about objects and events in their environment (Meltzoff, 1988; Rovee-Collier & Shyi, 1992). Rovee-Collier and her associates (1992) conducted a series of stud ies using a mobile reinforcement paradigm to investigate memory development in infants between two and six months of age. In these studies, a ribbon was tied to an infantâ€™s ankle and a mobile. The childâ€™s kicking motions created movement in the mobile. Once the child made an association between its movement and the movement of the mobile, the infant was tested at various times to determine if the child had processed the information in a way that allowed fo r recall. Infants as young as 2 months of age, remembered specific movements that had resu lted in rewards (Rovee-Collier & Hayne, 2000). Additional studies in which the mobile reinforcem ent paradigm was utilized revealed that infants also responded to cues that aided memory for act ion patterns that result ed in rewards days or even months after the original act was re inforced (Rovee-Collier & Shyi, 1992). Further evidence that the basic memory system is availa ble from the earliest months of life include the infantâ€™s ability to recognize familiar objects as shown by habituation/dishabituation studies (Borenstein & Ludemann, 1989), their capacity fo r facial recognition, (Legerstee, Anderson, &
28 Schaffer, 1998), and their awareness of size a nd shape constancy (Small, 1990). The evidence provided by these studies suggests th at general memory is availabl e at a very early age. Although memory is clearly available at an ea rly age, it neverthele ss undergoes significant developmental changes. For example, encoding a nd search strategies improve and become more complex with age (Deloache & Brown, 1983). Some researchers attribute th ese gains in memory performance to maturation (Siegler, 1996). Adm ittedly there are neurological advances that contribute to increased memory f unctioning in older children and a dults compared to infants and younger children (Greenough, Black, & Wallace, 1987). Nonetheless, even infants and young children show evidence of using se lf-referent information in a variety of ways that enhance their memory capacity. For example, strategies infants use when s earching for objects increase in complexity (Deloache & Brown, 1983). Heth (1983) suggested that changes in in fantsâ€™ search strategies are the result of the childâ€™s ability to use himself or herself as a reference point when engaged in a search for an object. When searching for a hidden object, very young children normally began their search at the last place the object was found. In contrast, three and five year olds began their search at the location closest to themselves. That is, they use themselves as reference points to begin their search. This increase in the level of sophistication of the organizational structure allows the child to begin using more complex sc hemas, based on personal experience, to aid not only in the retrieval of information but also in the construction of memories (Bransford, Goldman, & Vye, 1991). Howe and Courage (1993, 1996) assert that it is this â€œestablishment of a cognitive selfâ€ that provides the f oundation for autobiographical memory. Development of Autobiographical Memory Why is it necessary to consid er the developmental origins of autobiographical memory? The experiences of earlier years continue to be reflected as important themes in our
29 autobiographical narratives. McAdams (1996) a sserts, â€œchildren are collecting and processing experiences that will eventually make their way or have some influence on the integrative life story they later constructâ€(p. 106). Both the social constructiv ist view (Fivush, 1988; Haden, Haine, & Fivush, 1997; Nelson, 1989) and the cognitive-self view advanced by Howe and Courage (1993, 1996) offer compelling arguments with regard to the em ergence of autobiographical memory and its relationship to the self, a lthough the cognitive-self view places the emergence of autobiographical memory much earlier. There is an on-going debate with regard to whether or not autobiographical memory emerges at the end of the second year with the emergence of the objective self, as theorists embraci ng the cognitive-self view assert , or a few years later, as a result of social exchanges, as the social constructivists ma intain. Although this debate is interesting, it is not directly relevant to the adult development focus of this paper . Thus, I will only briefly describe the cogni tive-self viewpoint. I will present the social constructivist viewpoint in more detail because it has additional implications with regard to the functions of autobiographical memory in adulthood a nd the bi-directional relationship between autobiographical memory and self-concept. Autobiographical memory in childhood The cognitive-self view postulated by Howe and Courage (1993, 1996) contends that autobiographical memory, by defi nition, requires a person to alre ady have a sense of â€œselfâ€. These authors assert that autobiographical memory develops in a continuous fashion and that it is dependent on the budding sense of self as a foundational element (H owe & Courage, 1996). This viewpoint rests heavily on the research out lined earlier regarding increases in complexity and sophistication in general memory organizati on as the child begins using his or her own person as a point of reference. It also rests on evidence of a stab le sense of self represented by
30 the language usage of toddlers (Fraiberg, 1977; Howe & Courage, 1993). Children, around two years old, quickly learn pronouns â€œIâ€ and â€œmeâ€ and apply them appropriately. Listening to the conversations of others would not provide a child with the unique perspective necessary to use these self-referential pr onouns correctly, without a well-formed sense of self (Fraiberg, 1977; Nelson, 1989). Proper usage of pronouns calls for i nversions from â€œmeâ€ to â€œyouâ€ and â€œyouâ€ toâ€ meâ€. Some theorists regard childrenâ€™s ability to comprehend and apply th is conversion of form as evidence of an adequately formed sense self (Fraiberg, 1977; Howe & Courage, 1993). The social constructionist view of autobi ographical memory development in children draws on the model introduced by Vygotsky (1985) . This model focuses on the role of social interaction and the developmen t of cognitive skills as a foundation for autobiographical memories. Vygotsky maintained that cognitive skills are developed in exchanges with the environment and in particular through social interaction. According to Vygotsky, social interactions that take place w ith a more skilled partner guiding the interaction are believed to create a model that the child internalizes for fu ture use. Vygotsky believ ed these exchanges also taught the child what aspects of an event or ex perience were relevant . This process occurs through interactions in which the child and an adult (or more sk illed partner) are involved in social exchanges that are slightly beyond the child â€™s current level of ability. Through these social exchanges, the child becomes more and more co mpetent; meanwhile the adult adjusts his or her level of assistance to match the childâ€™s increasin g ability. Eventually, the child will incorporate these skills into an internal repertoire that will allow him or her to master the task without adult assistance. Examination of childrenâ€™s narratives w ith a more experienced social partner offers support for the social construc tivist view of autobiographical memory (Fivush, 1988; Fivush & Fromhoff, 1988; Haden et al., 1997; Welch-Ross, 1997).
31 Fivush (1988) noted the role of parental involvement in sh aping the narrative structure children adopt when relating autobiographical me mories. The interaction between a parent and child when the child narrates an event, demonstrat es the parentâ€™s ideal of structure for narratives as well as the importance of par ticular aspects of narratives. Earl y conversations between parent and child about past events call for the parent to provide most of the details of the conversation while the child participates minimally. As th e child becomes more accomplished in recounting an event, the parent, optimally, will reduce the am ount of support he or she provides. In turn, the child begins to use the parental structure to fo rmulate and organize inform ation about his or her experiences (Fivush, 1988). It is through these convers ations that the adult is modeling the childâ€™s narrative style (Haden et al., 1997). In dyadic interactions between parents and their children, the children internalize the parentâ€™s style of narration (Haden et al., 1997). When conversations between mothers and their 2 1/2 to 3 1/2 year old children we re taped, two distinct styles of interaction were identified for discussions of past events â€“ elaborative and repetitive (Fivush & Fromhoff, 1988; Haden et al., 1997; Welch-Ross, 1997). A parent who provides rich details and offers contextual information, while conveying interest in the childâ€™s comments, is using an elaborativ e style of communication (Fivush & Fromhoff, 1988). Using this style, parents encourage their children to be active partners in the conversation. A parent using a repetitive style of comm unication often repeats direct questions and highlight s the accuracy of the information the child reports without requesting additional details, or elaborating on what the child is saying by offering supplementary information (Fivush & Fromhoff, 1988). Parents using an elaborative style accentuate the process of communication more th an the details of the information itself.
32 Conversely, parents using a repetit ive style focus on the details of the information while ignoring the importance of the exchange (Fivush & Fromhoff, 1988). Haden and her colleagues (1997) studied the im pact of parents narra tive style used in conversing with their children about past events. Narrative style of the parent (elaborative or repetitive) was highly correlated with the childâ€™ s competence in recounting the experiences. The narrative style of childre n whose parents engaged in an el aborative style of communication included richer and more detailed contextual information (Haden et al., 1997). When these children were interviewed two and a half years later, the children with parents who used elaborative-narrative styles were fo und to have more detailed memo ries of the original event. Other studies have documented the relationshi p between parental narrative style and the childâ€™s ability to recount the past in narrative form (Fivush & Fromhoff, 1988). The evidence clearly supports the notion of the childâ€™s internalization of parental style of narration. The importance of the childâ€™s internal ization of narrative st yle is evident in that memories produced and shared may be retained and recalled more readily (Fivush, 1988; Nelson, 1993). The narrative style then guides future memories and serves to organize information about events and acts as the bedrock for determining what inform ation is most salient to the child (Fivush & Fromhoff, 1988). Although narrative style and structure are established in middle childhood, autobiographical memory remains a collection of unrelated single discrete episodes (Habermas & Bluck, 2000; Habermas & Paha, 2001) . The cognitive resources necessary for organizing and establishing links between numerous events to form a coherent life story is not firmly established until adolescence (Habermas & Bluck, 2000; Habermas & Paha, 2001).
33 Autobiographical memory in adolescence. Childhood experiences influence the life story. Fo r example, the quality of attachment in the first few years of life may eventually help to influence the overall t one of individualsâ€™ life stories (McAdams, 1993). Secure attachment may leave a legacy of trust and optimism setting a more trusting/optimistic tone fo r the individualâ€™s narratives. C onversely, insecure attachment may have the reverse effect. Just as young childre n internalize the narrative structure provided by parental interaction in earlier year s, attachment styles can also be expected to be internalized and to shape adolescentsâ€™ autobiographical memories as well as the narratives they create (McAdams, 1993). However, it is not until adolescen ce, that numerous even ts can be integrated in a way that allows individuals to understand their experiences in a meaningful way (Habermas & Bluck, 2000; Habermas & Paha, 2001). During adolescence and young adulthood, individua ls are faced with the challenge of creating a self that provides their lives w ith unity, purpose, and meaning (Erikson, 1980). McAdams (1996) asserts that people accomplish this task by developing their life story. The life story is comprised of narratives based on au tobiographical memories that have become internalized and reflect an indi vidualâ€™s sense of self (McAdams , 1996). One of the functions of these life stories is to coordinate various experi ences and aspects of the self in a manner that provides individuals with a sense of continu ity over time and brings meaning to these experiences (McAdams, 1996). Abstr act thinking is required in or der to extract meaning from experiences (Thorne, McL ean, & Lawrence, 2004). Cognitive advances provide adolescents the mean s to see themselves in a time other than the present (Harter, 1998). Abst raction allows adolescents to look back on childhood and realize they are no longer the children they once were. At the same time, abstraction allows adolescents to imagine themselves in a variety of different futures (Harter, 1998). Additionally, adolescents
34 are also capable of making distin ctions between their actual or current view of self and their possible or hoped for future self (Harter, 1998). Th is level of awareness es tablishes the setting so that adolescents can begin to understand not only what they were in earlier years, in the context of who they are now, but also who they may b ecome in the future (McAdams, 2000). In late adolescence, looking to the past in order to explain the present b ecomes more prevalent (McAdams, 1996). As individuals enter adulthood, autobiogra phical memory continues to interact with their deve loping sense of self. Autobiographical memory in adulthood In the literature on autobiographical memory in adulthood, researchers have focused on several issues (Bluck, Levine, & Laulhere , 1999; Brewer, 1986; Cohen, Conway, & Maylor, 1994; Pillemer, 1998). Researchers have examined th e events that give rise to autobiographical memories as well as various qua lities of the memories themse lves (Brewer, 1986; Pillemer, 1998). Characteristics of the individu als reporting the memories have also been studied (Brewer, 1986; Fitzgerald, 1984; Wageneaar, 1986). In partic ular, many researchers have turned their attention to the role the age of the individual pl ays with regard to vari ous characteristics of autobiographical memory (Bluck et al., 1999; Cohen et al., 1994). A dditionally, the methods employed to elicit autobiographical memories have also been shown to influence the findings (Bluck & Alea, 2002). In the following section each of these characteristics of autobiographical memory are explored. Declines in cognitive functions associated with aging have been well documented (Schaie, 1996), especially in the domain of speed of processing and fluid inte lligence (Schaie, 1996). Autobiographical memory does not rely on speed , and is associated with crystallized intelligence, thus one would expect autobiographi cal memory to remain stable across the life
35 span. Generally speaking, research findings s upport this assumption (Bluck et al., 1999). However, age differences in several aspects of autobiographical memory have been reported (Bluck et al., 1999; Cohen et al., 1994; C onway & Holmes, 2004; Linton, 1982). Throughout the lifespan, individualsâ€™ ability to cr eate narratives continues to evol ve (Singer, 2004). For example, Linton (1982) noted a general loss of detail in th e autobiographical memory narratives of older adults. This phenomenon, known as overgenerali zation, involves the condensation of details from several different occasions of similar ev ents into a singular ge neric representation. Overgeneralization occurs primar ily with more mundane activities that are a part of peopleâ€™s lives. Events such as holidays, trips to the same location, as well as everyday activities like shopping or attending church, ar e all susceptible to overgener alization. In many cases, any memories that are not both rehe arsed frequently and rather uni que tend to be overgeneralized (Linton, 1982). In addition to age differences in general aspects of autobiographical memory, certain characteristics of events that impact indi vidualsâ€™ memory (Brewer, 1986; Pillemer, 1998; Wagenaar, 1986) have also been identifie d. The importance of some of these event characteristics may also differ with age (Cohe n, 1996; Pillemer, 1998). Some characteristics of an event may be equally important to younger and ol der adults. For example, events that contain a high degree of emotional investment are re membered for longer periods of time (Brewer & White, 1982). Additionally, people exhibit a high de gree of accuracy in identifying events they experienced directly if the events were memo rable, as opposed to generic everyday events (Pillemer, 1998; Wagenaar, 1986). Likewise, the details of events that occur more frequently are not remembered as well as those of uni que events (Wagenaar, 1986). Conversely, other characteristics of an event may vary in importanc e with the age of the in dividual. For example,
36 Cohen (1996) found that predictors of the vividness of an aut obiographical memory differed for different points across the life span. Elapsed time from the event to recollection, and the emotional impact of the event, were associated with the vividness of the memory for younger adults. However, older adults were affected most by the regularity and the number of times the memory had been recalled. In addition, the degree of personal importance of an event was also a significant factor in affecti ng vividness (Cohen, 1996). Other age differences in autobiographical memory have been noted with regard to personally relevant information and reporting errors. Bluck and her colleagues (1999) found evidence of an increase in the amount of correct information older adults remembered when asked repeatedly to recall details of a personally relevant even t. In this study, participants who had watched the outcome of the O.J. Simpson murd er trial eight months earlier were asked to recall the event and report what they remember ed. Specifically, participants were asked their whereabouts, how they had heard about the verdic t, their reactions, and with whom they had spoken about the verdict. They we re also asked to relate what they remembered from watching the verdict as it was being read on TV. Participants were probed three separate times after they reported that they had recounted everything they could remember about the event. They were also asked about their confidence in their memo ry reports. Both younger and older adults were able to recall additional, correct details after pr obing (Bluck et al., 1999). Although Bluck and her colleagues (1999) found no differences between younger and older adults in terms of providi ng additional false information in response to the probes for more information, Cohen and her colleagues (1994) found significant differences in the number of errors made by older and younger adults. Cohen and colleagues (1994), asked adults ranging in age from 18 to 84 years old to write down detailed descriptions of a recent memory. When they
37 were asked eleven months later to share their reco llections of the event ag ain, the recollections of the older adults (64-84 year ol ds) varied significantly from the recollections of the younger adults (Cohen et al., 1994). The memory errors the older group made in cluded both errors of omission and errors of commission. Cohen and he r colleagues (1994) noted that older adults tended to rely heavily on common schemas to he lp reproduce the event. In addition, a general vagueness permeated their memories, as accuracy and detail were lost. These differences in recall have been linked to differences in what is both necessary and sufficient to create lasting impressions for older adults (Cohe n et al., 1994). In a ddition, some researchers suggest that the methods used to elicit autobiographica l memories may influence these findings. Differences in the various quali ties and characteristics found in studies of autobiographical memories may be determined by whether the memo ry to be recalled is generated solely by the participant or is designated by the experimenter (Bluck & Alea, 2002). In st udies that require the participants to generate memori es without any specific cues, re searchers have concluded that there is little or no difference in the availabi lity, vividness, or deta il of autobiographical memories of older adults when compared to those of younger adults (Cohen & Faulkner, 1988; Hoffman & Hoffman, 1990). Additionally, Rabbitt and Winthorpe (1988) found that frequent unplanned rehearsal was the best predictor of me mory vividness and degree of detail in the elderly. Memories that are rehear sed frequently comprise the to tal available pool of memories accessible to older adults (Rabbitt & Winthorpe , 1988). However, no notable differences have been found in the degree of vividness, complexit y, or detail in the self-s elected autobiographical memories of older adults (Bluck & Alea, 2002). Further, it is evident that not all autobiographical memories are of equal importance to the individual. Fitzgerald (1984) ex amined forgetting curves or te mporal ordering of events and
38 discovered a â€œreminiscence bumpâ€. In a study in wh ich middle-aged and older adults were given cues and asked to report a memory associated with the cue, Fitzgerald (1984) found that a majority of the memories were associated with a particular time period, between late teens and late twenties. This pattern wa s true for all adult memories. In addition to these phenomenological charact eristics of autobiographical memories, the functions of autobiographical memo ries have also been investigat ed. Potential age differences in these functions will be considered after pres enting a general conception of the roles of autobiographical memory. Functions of Autobiographical Memory Across the Life Span More recently, researchers have turned their at tention to the purposes that autobiographical memories may serve (Bluck & Alea, 2002; C ohen, 1998; Singer & Salovey, 1993). Although the debate about the number and nature of different functions of autobiogra phical memory is still ongoing, theorists have generally accepted the thr ee broad functions of autobiographical memory suggested by Cohen (1998): interp ersonal, intrapersonal, and knowledge-based functions. These functions are conceptualized as those that occur between people and those that occur within the person. Cohen (1998) suggests there are additional s ubcategories within the interpersonal (social interaction, self-disclosure, empat hy) and intrapersonal functions. Interpersonal, knowledge-based, and intraper sonal functions of au tobiographical memory The interpersonal functions of autobiographical memory are those that occur in a social context, such as memories described in convers ations as a means to establish relationships, friendships and intimacy, or for empathic underst anding of others. For ex ample, autobiographical memory serves the interpersonal function when individuals remember how they have felt or responded in past similar situations. This empa thic understanding provides individuals with the ability to decipher the meaning behind others be havior, thoughts, and fe elings, and to gain
39 insight into how others might be feeling. Likewise, when individuals remember how they felt in similar situations, they can act more appropria tely toward others with similar experiences, providing an emotional connec tion to important others. The interpersonal function of autobiographical memory is also operating when autobiographical remembering takes the form of self-disclosure with another person through sharing of personally relevant information. Cohen (1998) noted that disclosure of details associated with a negative event, particularly t hose that make a person feel vulnerable, creates deeper intimacy in a relationship. Additionally, th e conversational interaction created by sharing memories allows people to be included in on-go ing conversations, which in turn provide a sense of belonging. Individualsâ€™ autobi ographical memories provide them with a wealth of material upon which they base many of their conversations, thereby serving an inte rpersonal function. Cohen also recognized the knowledge-based functions of autobiographical memory (1998). Autobiographical memories serve knowledge-b ased functions when they are stored as general or situation-specific knowledge that can become relevant in general action planning or specific problem solving (Cohen, 1998). These gene ralized autobiographical memories become organized into a personâ€™s basic or glob al knowledge about the world (Tulving, 1986). Knowledge-based autobiographical memory guides people in everyday processing of information and decision making and provides a foundation for reasonab le expectations in particular circumstances (Pillemer, 2003). For ex ample, when a person remembers his or her decision and the outcome from a previous simila r experience this knowle dge will influence not only the decision the person makes in the present but also his or her expectations of the outcome. Autobiographical memories serve the knowledge -based function when they contribute to and define a personâ€™s general knowledge. This genera lized information is then refined or edited
40 as a result of the individualâ€™s ex pectations as well as the outcom es of a given course of action. This generalized knowledge also contains informa tion that can serve as a guide when addressing possible solutions to various problems. Autobiographical memories serve intrapersona l functions when they help a person to regulate emotions or â€œto constr uct, preserve, or edit the se lf-conceptâ€ (Cohen, 1998, p. 106). In terms of mood regulation, people ma y choose to focus and rehearse memories of the past that elicit either positive emotions or negative emotions. Memories of happy times and successes influence the individualâ€™s view of both the pres ent and the future. Conve rsely, the individualâ€™s focus on experiences associated with generally negative emotions, such as painful experiences, or personal failure can lead to a negative valence that will a ffect the personâ€™s mood. A tendency to select mood-congruent autobiographical memo ries may create a cyclic al pattern in which mood reinforces memory choice, and in turn memory choice reinforces mood (Blaney, 1986). Personal relationships may also be affected by the selection of au tobiographical memories that a person choose to rehearse. Memories associated w ith positive aspects of th e relationship result in a more positive view of the relationship. Likewise, if the individual reiterates the more negative aspects or experience in a given relationship, he or she will f eel more negatively toward the person(s) involved (Rub in & Berntsen, 2003). In terms of self-concept, the intrapersona l function of autobiographical memory helps people create and maintain a coherent sense of self throughout life. For example, as discussed earlier, in early childhood, the deve loping cache of personally relevant memories serves to define a childâ€™s sense of self (Nelson & Fivush, 2004). Similarly, in later year s, autobiographical memories may preserve the self-concept through selection of those memo ries that support a personâ€™s view of himself or herself (Cohen, 1998) . The self-concept may also be revised or
41 transformed depending on which autobiographical me mories the individual chooses to rehearse. Additionally, people may attend to particular asp ects of a memory and assign it more importance than other aspects. For example, the manner in wh ich a person remembers his or her role in any given experience is very subjectiv e, and prone to a high degree of creative omissions as well as additions to the story. These pr ivately-remembered autobiographical memories also allow the individual to evaluate his or her actual self against an imagined fu ture self or a proscribed â€œidealâ€ self (Cross & Markus, 1991; Cohe n, 1998). In this manner, autobiographical memories that are privately remembered serve to construct, pres erve, or transform a personâ€™s sense of self. As more researchers investigate the functions of autobiographical memory, a number of terms have appeared to describe the broad functions Cohen (1998) identified, and several subcategories of autobiographical memory functions have been explored. In the autobiographical memory literature, the broad category Cohen (1998) referred to as the in trapersonal function is more commonly labeled the sel ffunction . One subcategory, in particular, is important in this research; the term identity function will be used here to refer to the subcategory of the intrapersonal function of autobiog raphical memory associated w ith continuity of individualsâ€™ conceptions of self. The identity function of autobiographical memory In addition to the three functions of au tobiographical memory explored above â€“ interpersonal, knowledge-based, and intrapersonal functions -there seems to be growing consensus that autobiographical memory and id entity are intimately intertwined (Conway & Pleydell-Pearce, 2000; McAdams, 1996). For exampl e, McAdams (1996) suggests that as an ongoing life story, autobiographical memory brings both meaning and coherence to individualsâ€™ lives. Similarly, Conway and Pleydell-Pearce (20 00) suggest that identities are forged and changed throughout a personâ€™s lifetime as aut obiographical memories and current personal
42 concerns are modified. In line with this notion, a number of researchers have recently turned their attention specifically to the identity function of autobiographical memories (Conway & Pleydell-Pearce, 2000; McAdams, 1996; Pill emer, 1998; Singer & Salovey, 1993, 1996). These theorists have identified a subset of autobiographical memories that are primarily associated with self-concept continuity. In an early study, McAdams (1996) found that people easily recalled specific autobiographical memories that had resu lted in changes of their identity. Similarly, Pillemer (1998) found that certain life periods, specifically late adolescence and early adulthood, were more frequently associated with self-def ining autobiographical me mories than other life periods. These memories remained important to a personâ€™s sense of self for the remainder of his or her life. Although there is a well-reason ed theoretical foundation, empi rical evidence regarding the identity function of autobiographical memory is still limited. There are a few studies, however, that clearly illustrate the conn ection between the identity func tion of autobiographical memory and self-concept contin uity. For example, Bluck, Haberm as, and Rubin (2001) presented participants with various statements that re presented interpersonal, intrapersonal, and knowledge-based functions of autobiographical memo ry and asked participan ts to rate how often their autobiographical memories fulfilled each of these functions. They found that one of their scales consistently assessed the identity or se lf-continuity function of autobiographical memory (Bluck et al., 2001). Additionally, Pasupathi (2001) analyzed adultsâ€™ conversation using a qualitative coding procedure, and documented the use of autobiographical memory to convey important experiences and to share personal characteristics (i.e., oneâ€™s identity) to a conversational partner.
43 Age variation in the functions of autobiographical memory . Autobiographical memory functions at the in terpersonal, intraper sonal, and knowledgebased level throughout the life span (Bluck & Alea, 2002; Cohen, 1998) however; these functions may not be emphasized equally at different ages (Bluck & Alea, 2002; Bluck & Habermas, 2001; Webster & Cappeliez, 1993; W ilson & Ross, 2003). Although, there is little empirical evidence that the functions of autobi ographical memory vary in importance across the life span, a number of theorists have made comp elling arguments suggesting that the functions of autobiographical memory can be expected to va ry in importance across the life span (Bluck & Habermas, 2001; Webster & Cappeliez, 1993; W ilson & Ross, 2003). For example, Webster and Cappeliez (1993) proposed that the f unctions of autobiographical me mories could be expected to reflect the primary concerns and developmental tasks of a given age. They postulated that autobiographical memories primarily act to promot e social interaction in late adulthood, whereas in early adulthood autobiographical memories ar e more often the basis for problem solving. Likewise, in that self-concept formation is of primary importance in childhood and early adulthood, with less emphasis being placed on identity construction in middle and late adulthood (Erikson, 1980) one would expect a greater empha sis on autobiographical memories that serve the identity function in late adolescence (Webst er & Cappeliez, 1993). Autobiographical memories may also reflect developmental challenges associated with each stage of identity development. Conway and Holmes (2004) asked participants to narrate autobiographical memories from various stages of life. The narratives were then coded for Eriksonian psychosocial themes. Themes associat ed with the developmental challenge of each life period were clearly reflected in the autobiog raphical memory narratives associated with that life period. Conway and Holmes (2004) found additional evidence that autobiographical memories were tied to psychosocial themes. Co nway and Holmes (2004) asked participants to
44 provide autobiographical memories associated with a variety of cue words representative of various psychosocial developmental stages. They found participantsâ€™ age at the time of the cued event memory correlated with the expected developmental stage. Central Concepts in the Current Research Not all autobiographical memories are created equal. Some are more likely to hold an important position in the life story than others are. Narratives contain quite a bit of mundane information that is not part of oneâ€™s sense of self. Only some au tobiographical memories have an impact on a personâ€™s self-concept (Brandtstdter and Greve, 1994). This subset of critically important autobiographical memories has been referred to as self-defining memories (Pillemer, 1998; Singer & Salovey, 1996). Self-defining memo ries are the focus of this research. In particular, the objective of the present research is to examine how self-defining memories are associated with other forms of self-representation , such as personality features, or psychological well-being. Self-defining memories are di stinct from other autobiogra phical memories because they are emotionally charged, remarkable, recurring, and may be related to unresolved psychological conflicts (Singer & Salo vey, 1993, 1996). Memories meeting these cr iteria are often the result of an event in a personâ€™s life that caused a notic eable change (Pillemer, 1998; Singer & Salovey, 1996). Pillemer (1998) referred to such events as â€œmomentous eventsâ€. Moreover, events remembered in self-defining memories are often used as reference points when a person engages in life review or life reflecti on. Although empirical evidence exam ining the relationship between self-representations and self-defining autobiogra phical memory is nascent, many researchers have postulated a theore tical link between these two bodies of inquiry (Brandtstdter & Greve, 1994; McAdams, 1996; Pillemer, 1998; Singer & Sa lovey, 1996) and there is evidence, albeit limited, supporting this notion. Self-defini ng autobiographical memories may become
45 semantically represented as part of an individualâ€™s self-con cept. For example, as a person repeatedly recalls an event in which he or she acted in a manner that was kind, in turn the person may create a semantic representation of this ev ent (i.e., â€œI am kindâ€). Self-defining memories, from this perspective, should be related to othe r aspects of self-represen tation (i.e., self-concept clarity, extraversion, etc.). Additionally, psycholog ical well-being may also be impacted as a person focuses on particular autobiographical me mories while downgrading other memories. In this manner, self-defining autobi ographical memories may play a ro le in mediating the effect of other individual difference variab les on psychological well-being. Autobiographical Memory and Psychological Well-being There are a handful of studies that illustrate the connecti on between certain aspects of autobiographical memories and psychological well-being (Josephson, Singer, & Salovey, 1996; McAdams, Reynolds, Lewis, Patton, & Bowma n, 2001). Particularly, these studies have demonstrated a relationship between themes and emotional valence of individualsâ€™ autobiographical memories and indicators of psychological well-being su ch as agency, selfmastery, and empowerment (Grossbaum & Bates, 2002; McAdams, Hoffman, Mansfield, & Day 1996; Seidlitz and Diener 1997; Woike, Ge rsekovich, Piorkowski, & Polo, 1999). Seidlitz and Diener (1997) as ked participants to complete a daily diary for 43 days in which they recorded the best and worst event of the day. Additionally, participants completed questionnaires measuring mood and genera l life satisfaction each day. On the 44th day participants were asked to reca ll as many positive and negatives autobiographical memories from the past 43 day period. Seidlitz and Diener ( 1997) found participants with high scores on measures of psychological well-being recalled more positive autobiographical memories and fewer negative autobiographical memories.
46 Grossbaum and Bates (2002) found that autobi ographical narratives with themes of agency, communion, redemption and contamination were each associated with life satisfaction and specific domains of psychological well-bei ng. Grossbaum and Bates (2002) asked middleaged adults to report five autobiographical memories from each of the following categories of experience; an early childhood experience, a high and low point in their lives, a turning point, and one additional important memory. Participants also completed a measure of life satisfaction (Diener, Emmons, Larson, & Griffin, 1985) in add ition to a measure of psychological well-being (Ryff, 1989). The autobiographical memory narrative s were coded for emotional valence, themes of agency and communion, redemption, and cont amination. The levels of positive emotional valence of the memory narratives were posi tively correlated with several domains of psychological well-being, such as Self-Accep tance, Positive Relations with Others, Environmental Mastery and Pur pose in Life. Emotional valence of the narratives was also predictive of Life Satisfaction scores (Gro ssbaum and Bates, 2002). Specifically, positive valence was associated with greater satisfaction in life and negative valence was associated with lower levels of life satisfaction. In addition to Life Satisfaction, meaningful associations were noted between themes of the autobiographical narratives and certain dom ains of psychological well-being. Specifically, contamination themes, those in which the affective tone of the narrative shifts from positive to negative, were predictive of the psychological well-being domains of Environmental Mastery and Pe rsonal Growth. Narratives with themes depicting Agency correlated significantly with Po sitive Relations with Others , as well as Personal Growth. Narratives depicting Communion themes were si gnificantly associated with psychological domains of Positive Relations with Others, Envi ronmental Mastery, Purpose in Life and Personal Growth.
47 McAdams, Reynolds, Lewis, Patten, and Bowm an (2001) investigat ed the relationship between themes in autobiographical memori es and psychological well-being among young and middle-aged adults. Middle-aged participants we re asked to give a ve rbal account of eight autobiographical memories from each of the following categories of experience: a high and low point in their lives; a turning point; earliest memory; an impo rtant childhood, adolescent, and early adult experience; and one additional important memory. Part icipants also completed four measures of psychological well-being which include d the Satisfaction with Life Scale (Diener, et al., 1985), the Rosenberg Self-Est eem Scale (Rosenberg, 1965), the Center for Epidemiological Studies Depression Scale (Radloff, 1977) and a Sense of Coherence scale (Antonovsky, 1987). The autobiographical memory narratives were coded for themes of agency and communion, redemption and contamination sequences, and em otional valence. McAdams and his colleagues (2001) found significant relationships between the t ypes of themes depicted in autobiographical narratives and psychological well-being. The them es and in particular, the emotional tone reflected in autobiographical memories, appears to be associated with psychological well-being. As adults focus on the positive aspects of a given autobiographical memory, it may act as a psychological resource that signif icantly influences, or even bo lsters, psychological well-being. Autobiographical Memory, Personality, and Affect McAdams (2001) asserts, â€œpeople differ fr om each other with respect to their life stories in ways that are not unlike how they differ from each other on conventional psychological characteristics, traits and motivesâ€œ(p.101 ). McAdams (1996) suggests that personality should be examined at three distinct levels: dispositional traits (level 1), characteristic adaptations (level 2) and the Life stories (level 3). Numerous studies have examined personality at each of these levels: dispositional traits (Costa & McCrae, 1992; Johnson & Srivastava, 1999), characteristic adaptations (Little, 1999; Misc hel & Shoda, 1995), and life st ories (Habermas & Bluck, 2000;
48 McAdams & Bowman, 2001; Singer, 1995). Theorists ha ve recently begun to turn their attention to the association across these le vels of personality. However, the focus to date has been on either the relationship between dispositional trai ts and characteristic adaptations (Roberts & Robins, 2000) or characteristic adaptations and life stories (M cAdams & Bowman, 2001; Woike, 1995; Woike & Polo, 2001). Studies examining the relationship between dis positional traits and life stories are relatively rare. In one such study examining the relationship across levels of personality, McAdams and his colleagues (2004) found evidence of a rela tionship between life stories and individual differences related to dispositional traits (McAdams, Anyidoho, Brown, Huang, Kaplan, & Machado, 2004). McAdams and his colleagues (2004) found that the comple xity and themes of autobiographical narratives were significantly asso ciated with Big-Five pe rsonality traits (Costa & McCrae, 1992). In this study, McAdams and his co lleagues (2004) asked adult participants to provide written autobiographical memory narrative s associated with specific word cues (e.g., a high point, a moral decision). The narratives were then coded for complexity and themes of agency and communion and emotional valence. Narra tive features and part icipantsâ€™ scores on the Big-Five Inventory were examined â€“ Neurotic ism, Extraversion, Openness, Agreeableness and Conscientiousness (Johnson & Srivastava, 1999). McAdams and his colleagues (2004) found significant positive correlations between Ag reeableness and communion themes as well as Openness and narrative complexity. Emotional vale nce of autobiographical narratives and the Big-Five dispositional traits (Johnson & Srivas tava, 1999) were also examined in this study (McAdams et al., 2004). Although, McAdams and his colleagues (McAdams et al., 2004) did not observe any significant differences in the em otional valence of younger and older adults autobiographical memories, they noted significant correlations between specific traits and the
49 emotional valence of autobiographical narrat ives. Specifically, they reported a negative correlation between Neuroticism and positive emotional valence and a positive correlation between Conscientiousness, Agreeableness and positive emotional valence (McAdams et al., 2004). In the only other study to date explicitly examinin g the relationship between dispositional traits and life stories, significant associati ons between phenomenological characteristics of autobiographical memories and dispositional trai ts were noted (Rubin & Siegler, 2004). Rubin and Siegler (2004) asked particip ants to provide 15 different autobiographical memories in response to 15 different cue-words. Partic ipants rated each memory on a variety of phenomenological characteristics. In addition, participants also completed the NEO-Five Factor Inventory (Costa & McCrae, 1992). Rubin and Siegler (2004) reported significant positive correlations between both the Ex troversion and the Openness do main as measured by the NEOFive Factor Inventory (C osta & McCrae, 1992) and phe nomenological qualities of autobiographical memory associated with r ecollection and beliefs about the memory. Upon closer examination, Rubin and Siegler (2004) no ted that specific facets of each domain were responsible for the significant relationship. Specifically, those facet s of Extraversion and Openness that were associated with positive emotions had the highest correlations with characteristics associated with recollection and beliefs about the autobiographical memory. There is evidence that affect associated w ith autobiographical memo ries not only plays a role in mood regulation but also may be related to stable individual difference variables such as goal pursuit strategy (Moffitt & Singer, 1994), and personality traits (Rubin & Siegler, 2004) . Moffitt and Singer (1994) examined goals and self-defining memories of young adults. They found people who recalled a greater number of self -defining memories associated with important
50 goals also rated their self-defining memories w ith more positive affect. On the other hand, people who recalled self-defini ng memories with more negative vale nce also reported goals aimed at avoiding outcomes. Rubin and Siegler (2004) found evidence of a link between individual differences in personality traits like those meas ured by the NEO-Five Factor Inventory (Costa & McCrae, 1992) and certain aspects of autobiogr aphical narratives. Specifically, Rubin and Siegler (2004) found participants â€™ scores on measures of Extraversion and Openness correlated with certain phenomenological characteristics of the autobiographical memory narratives. For example, participantsâ€™ scores on measures of Extraversion significantly correlated with their reports of how well they can â€œhearâ€ the events depicted in their aut obiographical memory. Upon further inspection, Rubin and Siegler (2004) noted that scores on facets of the Extraversion and Openness scales that related to â€œopenness to emotional experiencesâ€ provided the highest correlations with the greatest number of pheno menological characteristics of autobiographical memories. Objectives of the Current Research The purpose of this study is to examine th e link between individua lsâ€™ self-concept and autobiographical memory, examining not only ag e variations in memory, but also exploring varying aspects of self-represe ntation in relation to autobiog raphical memory. There are three primary objectives: A) to explore the relationship between adultsâ€™ self-representations and selfdefining autobiographical memories, in relation to age; B) to investigate the relationship of valence of self-defining autobiog raphical memories to psychol ogical well-being, personality and affect; and C) to test the potential role of valence of self-defini ng memories as a mediator of the relationship between individual di fferences and well-being. This study uses both quantitative and qualitative methods to a ssess these relationships.
51 Self-Representations and Self-Defining Memories Empirical research on the iden tity function of autobiographi cal memory has been rather scarce; in particular, there are no studies that have empirically investigated the connection between semantic representations in adultsâ€™ se lf-concept and self-defining autobiographical memories. A growing body of literature suggests that individualsâ€™ self-representations may act as an important resource in the aging process (Brandtstdter & Greve, 1994). Correspondingly, theorists in the area of autobiographical memory (AM) contend that au tobiographical memories are used to bring coherence and meaning to indi vidualsâ€™ lives. Specifically , researchers who have adopted a functional perspective of AM (Bluck & Alea, 2002) have argu ed that one of the intrapersonal functions of AM, the identity function, may be of particular importance for individuals as they nego tiate the challenges of adult development and aging. Therefore, one of the primary objectives of this study was to em pirically document the relationship between the identity function of autobiographical memory and adultsâ€™ self-representati ons and to investigate potential age differences in this relationship. Because self-defining memories, by definition, are those memories that contain essential information about the self, it is reasonable to as sume that over time this information becomes semantically represented as part of a personâ€™s self-concept. Cons equently, a semantic recognition paradigm was utilized in this research. Afte r participants had recalled three self-defining autobiographical memories, they responded to self-attributes derive d from their memory narratives (memory-derived attributes). The memory -derived attributes were mixed into a list of self-attributes derived from an established list of general self-attributes used in previous research (Diehl et al., 2001; Sheldon, Ryan, Rawsthorne , & Laird, 1997). This wa s done to examine the special status of attributes derived from self-defining memories. Specifically, the memoryderived attributes were expected to represent central features of the self, whereas only some of
52 the attributes from the general li st were expected to represent important characteristics for the individual. Comparisons of these two types of attributes were conducted to investigate the relationship between adultsâ€™ self-representations and self-defining autobiographical memories. There were several steps necessary to gather memory-derived self-attributes used in the semantic recognition task. First, participants we re asked to share self-defining autobiographical memories. Then, coders extracted semantic repres entations of explicit as well as implicit selfattributes from participantsâ€™ memory narratives. Next, the semantic representations extracted from participantsâ€™ narratives were added to a list of 40 general self-descriptors (Sheldon et al., 1997). The resulting list of general attributes and memory-derived se lf-attributes were used in the semantic recognition task. A semantic recognition task developed for research on self-schemas and motivated reasoning (Markus, 1977; Saniti oso, Fong, & Kunda, 1990) was u tilized to investigate the relationship between self -representations and self-defining au tobiographical memories. Latencies in response to activated self-representations were used to test that se mantic representations resulting from self-defining AMs are accessed fast er than other self-des criptive attributes. The list of general and memory-derived self-attributes were also used in a self-descriptive rating task. In this task, participants were asked to rate the degree to which each attribute was personally descriptive, with the expectation that memory derived attributes would be rated as more personally descriptive. Relationships among Valence of Self-defined Memories, Individual Differences, and WellBeing The relationship between indi vidual differences associat ed with self-concept and psychological well-being has been well established in previous literature (Campbell et al., 1996; Cross & Markus, 1992; Diehl et al., 2001; Donahue et al., 1993; Freund & Smith, 1999; Linville,
53 1987). However, as reviewed earlier, studi es examining the relationship between autobiographical memory as it relates to indivi dual differences and psyc hological well-being are relatively rare. There are indications in the literature that the valence of self-defining autobiographical memories may be related to pe rsonality factors and may also be related to psychological well-being. The second goal of this research was to further examine the relationship of the valence of se lf-defining autobiographical memori es, as it relates to individual differences and psychological well-being. Although many research ers conceptualize psychological well-being as a lack of negative sy mptoms, my interest is on optimal functioning in late adulthood. Specifically, I am interested in the ways in which older adults compensate for age decrements by optimizing well-being. Therefor e, in this study psychological well-being was operationalized through the use of a scale develo ped by Carol Ryff (1989) that measures optimal functioning across multiple domains. In addition, the individual difference measures used in the study were chosen because they have been found to correlate with psychological well-being and are most likely to be an important resource fo r people facing challenges associated with the aging process. Self-defining memories are aut obiographical memories that are especially consequential to a personâ€™s self-concept. A lthough there is little research that explicitly addresses the impact of self-defining autobiographical memories on ps ychological well-being, there are a number of studies that illustrate the conn ection between certain aspects of autobiographical memories and psychological well-being (McAdams et al., 1996; Josephson et al., 1996). Particularly, a number of studies have demonstrated a relationship between emotional valence of individualsâ€™ autobiographical memories and indicators of psychological well-bei ng (Grossbaum & Bates, 2002; McAdams et al., 1996; Seidlitz and Diener 1997; Woike et al., 1999). Similarly, a number
54 of investigators have provided evidence for a re lationship between personality variables and selfrepresentations embedded in life stories (Conway & Pleydell-Pearce, 2000; McAdams, 1993, 1996, 2001). . In order to address the sec ond objective, qualitative characte ristics of autobiographical memory narratives were examined. Specifically, the emotional valence of participantsâ€™ selfdefining memories was noted. Autobiographical memory narratives are often coded for emotional tone and assigned a valence rating. Coding for valence has typically occurred as researchers explore complexity of the narrativ es along with multiple themes (e.g.; communion) illustrated by the narratives. However, in this study emotional valence was the primary quality of interest. Therefore, asking participants to provide their own rating with re gard to the valence of each memory narrative seemed most efficient. When measuring valence of the autobiographical memory, scales measuring positive and negative valence separately were employed to avoid the limitations and possible confounds inherent in a unidimensional scale of affect. A unidimensional scale of affect creates an unnatural response set that forces participants to endorse either positive or negative feelings associated with a particular memory thereby i gnoring the fact that peopl e are capable of having mixed feelings about an event. That is, they may simultaneously experience both positive and negative feelings about an event. Asking participants to rate their affective experience on a unidimensional scale may also effectively serve to d ilute the ratings as participants report median scores lying between the end point s. The focus of this study is on self-defining autobiographical memories as a psychological resource, theref ore I have focused on the positive valence of autobiographical memory.
55 The self is embedded within an overall set of constructs such as personality variables, possible selves, self-esteem, psyc hological well-being, and so on. These constructs may vary in strength across age groups (Die hl et al., 2001). Therefore, self -report measures of individual differences were also administered to allow for exploratory analyses with regard to the relationship between the valence of self-defini ng memories and these ot her individual difference variables: self-concept cl arity, positive affect, neur oticism and extraversion . These variables were selected for these exploratory analyses be cause they have been related to psychological well-being in past research (Campbell, 1996; Diehl et al., 2001; Ha rris & Lightsey, 2005; Schmutte & Ryff, 1997). Additiona lly, neuroticism, extraversion, and affect have been linked with autobiographical memory in two recent st udies (McAdams et al., 2004; Rubin & Siegler, 2004). Valence as a Mediator Theorists have argued that adults may use their self-representations and related selfdefining memories to adjust their self-concept in response to age-related challenges and/or losses. If this assumption is correct, then the positive valence of self-defining AM should mediate the relations between individual differe nce variables (i.e., self-concept clarity, positive affect, extraversion,) and psychological well-being. Many theorists have suggested that individuals â€™ subjective well-being is linked to stable individual differences (Campbell et al., 1996; Co sta & McCrae, 1984; Seidlitz, Wyer, & Diener, 1997). Exploration of the identify function of AM, as it relates to stable individual differences and psychological well-being, may provide insi ght into the underlying process that links individual differences and psyc hological well-being. Specifically, the valence of self-defining autobiographical memories may mediate the imp act of individual differences with regard to psychological well-being. The emotional tone of au tobiographical narratives has been shown to
56 significantly predict psychologi cal well-being (Grossbaum & Bates, 2002; McAdams et al., 2001). McAdams and his colleagues (2001) examined individualsâ€™ capacity for finding benefit from adverse situations and noted that people who were able to pe rceive benefit ultimately show better psychological adjustment. Specifically, pe ople who able to find benefit from adverse events reported higher levels of positive affect than their counte rparts who were less capable of finding benefit from adverse events (McAdams et al., 2001). Kennedy, Mather, and Carstensen (2004) noted that the valence of autobiographical memories reported by older adults tends to be significantly more positive than the valence of younger adults â€™ autobiographical memories. These researchers suggest that ol der adultsâ€™ tendency to remember their experiences in a more positive light may be responsible for the differe nces in subjective well -being reported by older and younger adults (Kennedy et al ., 2004). Additionally, individuals were able to enhance their current feelings of well-being by focusing on vari ous components of autobiographical memories (Kennedy et al., 2004). Given the lack of investigation of the rela tionships between self-d efining memories and well-being, there are a number of different appro aches that might be feasible in examining the extent to which our self-defining memories mediate the relationship between individual differences and well-being. For instance, one ap proach would be to examine the impact of positive valence of autobiographical memories with regard to individual difference variables that affect psychological well-being in a negative ma nner. For example, one might consider the relationship among positive valence of autobiogra phical memories, depressive symptoms and psychological well-being. Using this approach would call for examining the role of positive valence of autobiographical memories in overcom ing depressive symptoms in a manner that would maintain psychological we ll-being. Another way to view this would be to say that high
57 well-being, as a positive feature of the self, is not so strongly c onnected to specific personality variables and positive affect, but that these variables influence well-being only through their impact on self-definition. In pa rticular, the positive va lence of oneâ€™s self-defining memories may influence well-being more so than these othe r individual difference variables. The latter approach is the one selected fo r this particular investigation. Several studies have established a relations hip between psychological well-being and a wide range of individual differen ce variables associated with dis tinct personality traits (Campbell et al., 1996; McAdams, 1996). However, for this study, I focus on those individual difference variables that have been shown to predict variations in psycho logical well-being and that have been shown to vary in their importance for ps ychological well-being. Specifically, self-concept clarity, positive affect, Extraversion and Neur oticism were employed (Adkins, Martin, & Poon, 1996; Mroczek, 2001). In addition Ryffâ€™s (1989) m odel of psychological well-being, focusing on optimal functioning across multiple domains, was al so used to investigate the relationship among these key variables: valence of self-defining autobi ographical memories, individual differences, and psychological well-being. Beca use the intercorrelations for the subscales of Ryffâ€™s (1989) well-being measure are moderate to high, a composite score of all th e subscales is often used as an indicator of positive psychological functioning. Th at is the approach used here. Rather than break down well-being into particular aspects (subscales of Ryffâ€™s well-being measure), the emphasis, for this first investigation of these relationships, is placed on overall well-being as it relates to self-concept an d individual differences. Hypotheses of this Study Self-Representations and Self-Defining Memories To document the link between se mantic representations in adultsâ€™ self-concept and selfdefining autobiographical memories, response latencies from the semantic recognition
58 task and salience ratings of th e self-attributes were examined. Hypothesis 1a : Because of their centrality for a personâ€™s self-concept, we hypothesized that in a semantic recognition task, both younger and older adults would respond significantly faster ( p < .05) to memory-derived attributes (MDA s) than to general self-descriptive selfattributes (GSAs). Hypothesis 1b : We also expected, that after contro lling for normal age-related slowing in reaction time, both younger and older adults woul d respond equally fast to MDAs. In contrast, with regard to GSAs, we hypothesi zed that younger adults would re spond faster than older adults would. Hypothesis 1c : Memory-derived attributes were co mpared against general self-descriptive self-attributes with regard to differences in at tribute salience. Specifica lly, because of their importance for self-definition, we hy pothesized that memory-derived attributes would be rated as significantly ( p < .05) more salient than genera l self-descriptive attributes. Hypothesis 1d : Because some literature suggests th at self-defining autobiographical memories may become more important with rega rd to self-concept main tenance as people age, we hypothesized that the main eff ect of attribute type may vary by age, i.e., we predicted an Attribute Type X Age Group interaction). Specifi cally, we hypothesized th at older adults would rate their MDAs as significantly ( p < .05) more salient than young adults would, and younger adults would rate GSAs significantl y higher than older adults would. Relationships among Valence of Self-defin ed Memories, Personality, and Well-being Participantsâ€™ rated the emotional valence of th eir self-defining memori es. Correlation analyses were performed to examine the associations of the valence of self-defining memories with psychological well-being and individual differen ce variables, such as self-concept clarity, positive affect, extraversion and neuroticism.
59 Hypothesis 2a : The level of positive emotional valence of self-defining autobiographical memories was expected to correlate positively with self-concept clarity, positive affect, and extraversion. Conversely, positive emotional valen ce of self-defining autobiographical memories was expected to correlate nega tively with neuroticism. Hypothesis 2b : Positive valence of autobiographical memories was also expected to be positively correlated with psychological well-being as measured by the Ryff (1991) psychological well-being scale. Valence as a Mediator The role of positive valence of self-defining memories as a mediating variable between individual difference variables and psychologi cal well-being was exam ined using regression analysis. Baron and Kenny (1986) state, â€œa given vari able may be said to function as a mediator to the extent that it accounts for the relati onship between the predic tor and the criterionâ€ (p.1176). Baron and Kennyâ€™s model was used to conduct exploratory analyses. Although relationships have been found between the individu al difference variables used in this study and psychological well-being, and relationships have been found between positive valence and psychological well-being, the inter-related associ ations amongst these va riables have not been investigated. Hypothesis 3 : The positive valence of self-defining memories was expected to mediate the effects of individual difference variables on the outcome variable, psychological well-being (see Figure 1-1).
60 Figure 1-1. Proposed Model of the Mediationa l Relationship Among Positive Valence of SelfDefining Autobiographical Memory, Indi vidual Difference Variables, and Psychological Well-Being. Individual Difference Variable Positive Valence of AM Psychological Well-being
61 CHAPTER 2 METHOD The following section presents the methods propos ed for this study. First, the design of the study and the participants are described. Next, th e measures are described with a particular emphasis on the focal measures of the study, thos e that assess self-def ining autobiographical memories. Testing procedures are detailed next , and then coding procedures are explained. Design and Procedure The study used a mixed-factor design with ag e group as a between-subjects factor (young vs. older adults) and attribute t ype (memory-derived self-attributes vs. general self-attributes) as a within-subjects factor. Participants were tested individually in two sepa rate testing sessions scheduled two weeks apart. A tw o-week interval between testing sessions was implemented in order to minimize possible recognition effects relate d to participantsâ€™ verbal report of their selfdefining autobiographical memories. This interval takes into account the findings of a literature review showing that, for older adults, recogni tion effects are increased as compared to young adults, and can last over several days (Laver & Burke, 1993). Testing t ook place at the Adult Development and Aging Project ( ADAPT) laboratory at the Univer sity of Florida and in the community. Specially trained research assistants conducted the testing sessi ons. The order of test administration is shown in Figure 2-1. All partic ipants were asked to volunteer their time. Participants Participants were recruited through the Univ ersity of Florida Psychology Undergraduate participant pool, an existing ol der adult participant pool, and from community organizations. Participants were screened via the telephone to ensure that they were in good health, without a history of mental illness, living independentl y, and without cognitive impairments that would prevent full participation in the study.
62 One hundred twenty-three participants, incl uding 60 younger adults (18 39 years old), and 63 older adults (60 â€“ 87 year s old) were tested and includ ed equal numbers of men and women in each age group. Three participants we re excluded from the sample. One younger adult and two older adults were excluded due to pr ocedural abnormalities. The young adult age range was chosen so that the sample would not be co mpletely skewed towards college students. Key sociodemographic variables such as age, gender, income, education, marital status, and health status were recorded. The means and sta ndard deviations associated with these sociodemographic variables are presented in Table 2-1. With regard to years of e ducation and life satisfaction, no significant differences were found between the young adults and older adu lts (see Table 2-1). However, significant differences were noted with regard to pa rticipantsâ€™ subjective ratings of health, F (1, 120) = 4.56, p < .05. When comparing themselves to thei r age peers, using a Likert-type scale ( 1 = Very poor , 6 = Very good ), older adults reported be tter relative health than younger adults. However, there were no statistically significant differences, F (1,120) = .48, p > .05, between the number of visits to a doctor (per yea r) reported by older adults ( M = 3.8, SD = 2.55) and the number of visits to a doctor reported by younger adults (M = 3.4, SD = 2.20). Significant differences were also noted with regard to participantsâ€™ reported income, F (1, 120), 16.49, p < .05. Younger adults reported higher levels of income compared to older adu lts. This was likely due to younger adults reporting their pa rentsâ€™ annual income. Sample Size and Power Considerations Estimations of sample size and statistical power were considered for mixed-factors analyses of covariance (ANCOVAs). In the abse nce of previous studies that could provide guidance in terms of expected e ffect size, medium (.25) to larg e (.50) effects size were assumed to be reasonable estimates. These estimates were supported by pilot data. Based on Cohen
63 (1988), adequate power ( > .80) at the .05 significance level to detect an effect size of .50 requires a total of 69 participants ( n = 35 per age group). To detect a medium effect size with < .05 and > .80, a total sample si ze of 120 participants ( n = 60 per age group) was required. Thus, the proposed sample size of N = 120 ( n = 60 per age group) was expe cted to have sufficient power ( > .80) to detect medium and somewhat smaller effects. Measures The following section contains a description of the measures th at were used to examine the relationship between self -representations and autobiographical memory. Because self-defining autobiographical memories were the focus of this study, measures that assess self-defining memories will be described first. Next, measures are described that were included to control for normal age-related differences that may affect na rrative quality or participantsâ€™ ability to recall their autobiographical memories. Finally, indivi dual difference measures are described. The individual difference measures included scales designed to examine self-concept clarity, affect, personality factors, and depres sive symptoms. These measures were included to allow for examination of the relationship of self-defining memories and the resulting self-representations with psychological constructs in a larger nomological network. Self-Defining Memory Task Three self-defining autobiograp hical memories were solicite d from each participant. Participants gave a verbal account of each memory, which was audiotaped. The protocol developed by Moffitt and Singer (1994) was used to solicit the self-defining autobiographical memories. Specifically, participants were instruct ed to recall autobiographical memories that were important with regard to th eir current identity. When particip ants indicated they had shared the entire memory, they were presented with f our standard probes, one at a time, to assure completeness of the narrative. The first time, pa rticipants were asked, â€œIs there anything else you
64 can tell me about that experience/event?â€ The second time participants indicated they had completed the narration they were asked, â€œCan you tell me anything else about how you were feeling or what you were thinking at the time?â€ The third probe was, â€œDo you think that is everything?â€ Finally, each participant was as ked, â€œCan you tell me what you learned about yourself from that experience/event?â€ Qualities of Self-Defining Memory After the narration of each me mory, participants were aske d to rate the self-defining memory. Participants rated the valence of the memory as well as several other dimensions frequently used to describe the qualities of autobiographical memories. Three questions were used to assess the valence of each memory. Participants were asked to rate how positive their feelings connected with the memory were (1= N ot at all , 5= Extremely). In this study, this item was averaged across the three memories and used as the measure of positive valence. Participants were also asked to rate how negative their feelings were (1= Not at all , 5 = Extremely ). Additionally, participants were asked to describe thei r overall feelings connected with the memory (1= Very positive , 6 = Very Negative ). The questions assessing valence were included with questions addres sing other dimensions of memo ry such as: vividness (1 = Not at all , 5 = Extremely ), importance (1 = Not at all , 5 = Extremely ), intensity (1 = Not at all , 5 = Extremely ), and frequency of recall (1 = Rarely, 5 = Very Often ). Participants were also asked to date the event that gave rise to the recalled memo ry. Finally, participants were asked what they learned about themselves from the self-defini ng memories they provide d (see Appendix A). Coding of Memory Narratives Each participantâ€™s audiotaped memory interv iew was transcribed verbatim. Trained coders examined the content of each narrative. Self-descr iptive attributes were then extracted for use in a semantic recognition task administered in Session 2.
65 Coding of memory-derived self-attributes. In order to extract se lf-attributes from the autobiographical memories for later use in th e semantic recognition ta sk, the content of each memory narrative was examined. The training of c oders for this extraction process was extensive (see Appendix B). Coders examined the memory na rratives and extracted a list of self-attributes from each memory narrative that met the establishe d criteria for self-defining attributes on two separate occasions (see Appendix B for coder guidelines). For ease of presentation, these attributes will be called memory derived attribut es or MDAs. The first time the coders examined the protocols, they focused only on explicitly ex pressed self-attributes. The second time coders examined the protocols, they focused on extrac ting implicit self-attribu tes based on behavioral descriptions embedded in the narratives. For exam ple, the trait â€œhonestâ€ w ould be included in the list of self-attributes for a narra tive containing a phrase such as â€œbecause of who I am, I just had to tell the truthâ€. Two coders read each narrative transcription. Inter-r ater reliability for the lists of MDAs identified by the coders was required to reac h a minimum Kappa of .70 based on training with protocols from the pilot sample. Inter-rater reliab ility was calculated frequently and remained between .65 and .85 throughout the study, averaging .79. In instan ces in which the coders disagreed on MDAs, both coders di scussed their findings with an independent expert rater until consensus was reached with regard to the attri butes in question. As an additional method check, the percentage of memory-deriv ed attributes in which partic ipants responded â€œnot meâ€ was calculated. The percentage of memory-derived attr ibutes to which partic ipants responded â€œnot meâ€ was less than one-half of one percent. If a memory-derived self-attribute was identical to an item on the general self-attributes list, it was re placed with a synonym derived from a thesaurus
66 of synonyms specifically created for this purpose . The self-attribute thesaurus was based on the familiarity norms of self-a ttributes validated by Dumas, Johnson, and Lynch (2002). Coders were asked to complete The Facts on Aging Quiz (Palmore, 1998) a measure of age-bias. The four coders were asked to indicate their level of agreement (1 = true , 2 = false ) with a variety of statements that reflect age-bi as. Coderâ€™s responses were summed into a total score, with higher scores suggest ing higher levels of age-bias. Th is measure was used to exclude coders that exhibited high levels of age bias that could potentially impact the coding of the narratives. In fact, none of the coders demonstrated age bias. Semantic Representations of Self-Defining Memories . The semantic representations culled from participantsâ€™ narratives of self-defining autobiographical memories were in serted into a list of 40 general self-descriptors (Sheldon et al., 1997). The same 40 general self-descriptors were used for all participan ts. These general selfdescriptors represented a wide spectrum of trait words and mappe d onto the Big-Five personality traits, Neuroticism, Extroversion, Openness, Conscientiousness, and Agreeableness (as measured by the NEO-FFI, Costa & McCrae, 1989). Sheldon a nd colleagues (1997) established the validity of this list of general self-attributes by exam ining the correlations be tween the general selfattributes and participantsâ€™ responses on the NEO-FFI (Costa & McCrae, 1989). The resulting list of 40 general attributes, with MDAs added, wa s used in the semantic recognition task and the self-attribute rating task. Both ta sks were performed during Session 2. Semantic Recognition Task. A procedure for measuring response latenc ies of activated self -representations was established in research on self-schema and mo tivated reasoning (Markus, 1977; Sanitioso et al., 1990). This procedure was designed to test whether semantic repres entations resul ting from selfdefining autobiographical memories would be ac cessed faster than ge neral self-descriptive
67 attributes. Using the Visual Basi c software, participants were presented with 40 general selfattributes. These attributes have been used in previous studies on adu ltsâ€™ self-representations (Diehl et al., 2001; Sheldon et al ., 1997) and represent a comprehens ive list of self -attributes. In addition, self-attributes taken from the self-def ining autobiographical me mory narratives were randomly incorporated into the list of general se lf-attributes. Each self -attribute was presented for up to 3 seconds or until the participant respon ded, whichever occurred first. Participants were instructed to make a judgment regarding whether the presented attribute was self-descriptive or not by pressing a designated â€œMeâ€ or â€œNot Meâ€ button as quickly as possible. Response latencies were recorded online. Participantsâ€™ self-defin ing memories yielded va rying numbers of selfattributes; therefore, proportiona lized scores were used for all analyses. Only the general selfattributes the participant judged as self-descrip tive were used in the analyses. The semantic recognition task was admi nistered in Session 2. Self-Descriptor Rating Task Participants were also asked to rate all self-attributes, general and memory-derived, with regard to their personal salience. Specifically, e ach self-attribute was ra ted on an 8-point Likerttype scale ranging from â€œ Very uncharacteristic of me â€ (1) to â€œ Very characteristic of me â€ (8) (see Diehl et al., 2001; Donahue et al ., 1993; Sheldon et al., 1997). Control Measures Sociodemographic information was collected. In addition, several measures were used to control for age-related cognitive changes that may have affected the narrative quality of the autobiographical memories or the recall of the autobiographical memories. Specifically, measures of overall verbal ability, perceptu al speed, and general memory function were administered. All control measures we re administered in Session 1.
68 Personal data form . A detailed personal data form (Die hl et al., 2001) was used to collect sociodemographic information (e.g., age, gende r, ethnicity, educa tion, etc.) along with information regarding the partic ipantsâ€™ general health status. This measure has been used extensively in research with adults of all ages and can be scored objectively. Verbal ability. The Advanced Vocabulary Test (V-3) from the Kit of Factor-Referenced Cognitive Tests (Ekstrom, French, Harman, & Dermen, 1976) was used to assess adultsâ€™ general level of verbal comprehension. This test requ ires the selection of a synonym for each given stimulus word from five alternatives. This test is widely used in research with adults of all ages and its psychometric properties are we ll established (Ekstr om et al., 1976). General memory . The Auditory Verbal Learning Ta sk (AVLT; Rey, 1941) was used to assess participantsâ€™ overall semantic memory. Part icipants studied a list of 15 semantically unrelated words for 1 minute and then wrote down as many of the words as they could remember from the list. Reliability and valid ity of the AVLT are well established. The Balanced Inventory of Desirable Responding (BIDR) . The BIDR (Paulhus, 1991) was used to examine participantsâ€™ social desirabi lity biases. Individuals were asked to indicate their level of agreement (1 = Not true , 7 = Very true ) with statements that reflected a need for social desirability. Respondentsâ€™ answers were summed into a to tal score, with higher scores suggesting higher levels of so cial desirability bias ( = .73). Participantsâ€™ scores on the BIDR were used for exclusionary purposes. Reaction time tasks . In order to control for normal ag e-related slowing in reaction time when assessing participantsâ€™ res ponse latencies with regard to se lf-attributes, a simple reaction time task and a choice reaction time task were administered. Some research has shown that simple reaction time and choice reaction time show largely the same age-related slowing (Paul
69 Verhaeghen, personal communica tion, June 30, 2004). However, we decided to include both tasks because the demand characteristics of a choice reaction task are more similar to the semantic recognition task. Including both reaction time tasks will increase the sensitivity of the control variables (Paul Verhaeghen, pe rsonal communication, June 30, 2004). For the test of simple reaction, used for st atistical control of ag e-related differences, participants were seated in front of a 17inch computer screen and instructed to respond using a single key when a symbol (â€œ+â€) appeared on the screen. The interstimulus interval was 0.5 to 1.5 milliseconds and varied randomly. Participantsâ€™ scor es on this task were reaction times measured in milliseconds, averaged across all responses. The sy mbol always appeared in the center of the screen and reaction times were recorded online. This task did not require any prior computer experience and the older adults were able to pe rform it without difficulty. Each participant had 5 warm-up trials before the actua l task started. In order to capture differences in reac tion times between younger and older adults, associated with higher-order processing, a choi ce reaction time task was also administered. Participants, seated in front of a 17-inch comput er screen, were instructed to respond using one of two keys that corresp onded with one of two words (â€œMeâ€ or â€œNot Meâ€) that appeared on the screen. When the word â€œMEâ€ appeared on the scr een, participants were instructed to press the key corresponding to the word â€œMEâ€. Likewise , when the words â€œNOT MEâ€ appeared on the screen participants were instructed to press the key corresponding to th e words â€œNOT MEâ€. The same keys were used later in the study to represent an actual â€œMEâ€ or â€œNot MEâ€ choice in response to the various self-attributes. In this way, the motor response demands of the semantic recognition task were matched as much as possi ble. The interstimulus interval was 0.5 to 1.5 seconds and varied randomly. Participantsâ€™ score on this task was the response latency averaged
70 across all responses. In preparation for the task, a â€œ+â€ appeared in th e center of the screen so that participants knew where to expect the stimulus to appear. Reaction times were recorded online. Each participant had 5 warm-up trials be fore the actual task began. Individual Difference Measures A number of individual diffe rence measures were administ ered during Session 1. These measures were included because they assess theore tically related concepts, which we planned to examine in the analyses or statistic ally control for, if necessary. Psychological well-being . The Short Psychological Well-be ing Scale (SPWB; Ryff, 1989, 1995) was used to assess participantsâ€™ psyc hological well-being. The SPWB measures six dimensions of psychological well-being that have been derived from the literature on lifespan development, mental health, and personal grow th (Ryff, 1989, 1995). The six dimensions are Self-Acceptance, Environmental Mastery, Purpos e in Life, Positive Relations with Others, Personal Growth, and Autonomy. Each dimension was measured with a 14-item scale of positively and negatively phrased items, which were alternated in their order ac ross dimensions. Participants responded to each item on a 6-point scale (1 = S trongly disagree , 6 = Strongly agree ). The psychometric properties of the SPWB have been examined in terms of inte rnal consistency and test-retest reliability (Ryff & Keyes, 1995). Ryff and Keyes (1 995) have also reported findings from confirmatory factor analyses supporting the six-factor structure of the questionnaire. The alpha coefficients for internal consistency in this study we re .84, .83, .74, .83, .82, and .81 for Autonomy, Environmental Mastery, Personal Growth, Positi ve Relations, Purpose in Life, and SelfAcceptance respectively, and the alpha coefficien t was .90 for the total scale score. Self-Concept Clarity Scale (SCC ). The SCC consists of 12 items designed to measure â€œthe extent to which the contents of an individualâ€™s self-conce pt were clearly and confidently
71 defined, internally consistent, and temporally stableâ€ (Campbell et al., 1996, p. 141). Items were presented as 5-point scales (1 = Strongly disagree , 5 = Strongly agree ). Scores were summed across the 12 items ( = .87) and higher scores indicated a more clearly defined self-concept. Reliability and validity for the SCC are we ll established (Campbell et al., 1996). Positive and Negative Affect Scale (PANAS). The PANAS (Watson, Clark & Tellegen, 1988) assessed two primary dimensions of mood. Positive affect (PA) re flects the extent to which a person feels enthusiastic, active, and alert. In contrast, negative a ffect (NA) is a general dimension of aversive mood states such as a nger, contempt, guilt, fear, and nervousness. The PANAS has 20 items with 10 items each for PA ( = .79) and NA ( = .84). Respondents rated each item on a 5-point Likert-type scale (1 = Very slightly or not at all , 5 = Extremely) indicating to what extent they felt this way in the past few weeks. The PANAS has excellent psychometric properties (Watson et al., 1988). NEO Five-Factor Inventory (NEO-FFI). The NEO-FFI (Costa & McCrae, 1992) was used to measure the two broad personality factor s of neuroticism (i.e., emotional lability) and extraversion (i.e., sociability and outgoingness). The scales a ssessing these two dimensions consisted of 24 items that are presented in a Likert-type format (1 = Strongly disagree , 5 = Strongly agree ). The coefficients of internal consiste ncy in this study were .84 for Neuroticism and .80 for Extraversion. The NEO-FFI is one of the most widely used personality inventories for healthy adults and its psychometric propertie s are well established (Costa & McCrae, 1992). Center for Epidemiological Studi es Depression Scale (CES-D) . The CES-D (Radloff, 1977) was used to assess the frequency of partic ipantsâ€™ depressive symptoms for exclusionary purposes. Individuals were aske d to indicate how frequently they experienced the listed symptoms within the past week (1 = Rarely or none of the time , 4 = Most or all of the time ).
72 Respondentsâ€™ answers were summed into a total score, with higher scores indicating a higher frequency of depressive symp toms. The CES-D has been specifically recommended for use in nonclinical community-based samples and its psychometric properties are well established (Radloff, 1977). No one in the sample was excluded for moderate or severe depression. Internal consistency reliability in this sample was .91.
73 Table 2-1. Means and Standard Deviations of Sociodemographic and Health Variables by Age Group ________________________________________________________________________ Young Adults Older Adults ( n = 59) ( n = 61) Variable M SD M SD _______________________________________________________________________ Age (yrs) 22.06 2.49 74.81 6.17 Education (yrs) 14.97 1.46 16.50 3.55 Income (1,000s) 71.00 14.35 41.30 8.03 Health 5.06 .79 5.66 .72 Life Satisfaction 4.68 .80 4.88 .72 _______________________________________________________________________ Table 2-2. Means and Standa rd Deviations Individual Di fference Variables by Age Group ____________________________________________________________________________________ Young Adults Older Adults ( n = 59) ( n = 61) Variable M SD M SD _____________________________________________________________________________________ _____________________________________________________________________________________ Note. SCC=Self-Concept Clarity Scale; NEO-N = NEO Neuroticism Scale; NEO-E = NEO Extraversion Scale; PA = Positive Affect Scale. Session 1 Session 2 Session 2 Figure 2-1. Order of Test Administration SCC 40.60 8.33 44.78 6.51 PA 30.56 6.34 26.62 7.17 NEO-N 30.52 7.14 25.52 7.28 NEO-E 44.42 6.40 41.50 6.80 Sociodemographic Info Self-Defining Memory Task Control Measures Individual Difference Measures Reaction Time Tasks Semantic Recognition Task Self-Descriptor Rating Task
74 CHAPTER 3 RESULTS The results of this study are pr esented in four main sections. The first section describes the steps taken to assure that none of the assumptions related to the st atistical analyses were violated. The second section presents the results of the analysis of co variance used to examine the relationship between adultsâ€™ self-representations and self-defining autobiographical memories (Objective A). In the third sec tion, the results are described fr om the correlation analyses examining the relationships between the vala nce of autobiographical memories, individual difference variables, and psychological well-being (Objective B). Finally, in the fourth section a mediation model is tested. This model propos ed that the relations hip between individual difference variables and psychological well-being is mediated by the positive valence of selfdefining memories (Objective C). Preliminary Analyses Preliminary analyses were conduc ted to ascertain that there were no obvious errors in the data set, and that none of the assumptions (i.e., normality, homogeneity, multicollinearity) associated with the planned statistical analyses were violated, To identify data points lying significantly ou tside the normal range, descriptive analyses were conducted that included an examination of extreme values, along with box plots and stem and leaf plots. There were three data points as sociated with one of the covariates, the wordchoice reaction time task, that were of particul ar concern initially. T hus, all analyses were subsequently conducted twice; the analyses we re conducted first with the data from all participants included, and then repeated, excludi ng those participants as sociated with these extreme data points. Comparison of the analyses revealed trivial differences, none of which
75 affected the significance of the analyses being conducted. Therefore, the reported findings are based on all of the data, in cluding these outliers. The means and standard deviations of the i ndividual difference va riables; self-concept clarity, positive affect, neuroticism, and extraver sion were examined to assure the sample was typical (see Table 2-2). Significant age differences were noted with regard to participantsâ€™ scores on Self-Concept Clarity scale F (1, 119), 10.94, p < .05. Older adults reported higher levels of self-concept clarity compared to younger adults, consistent with past research. Typically, older adultsâ€™ scores on measures of Neuroticism a nd Extraversion vary significantly from younger adults. In this sample, significant differences we re noted on participants scores on scales of Neuroticism, F (1, 119) = 14.16, p < .05, and Extraversion F (1, 119) = 5.76, p < .05. Older adultsâ€™ scores on scales of Neuroticism as well as Extraversion were lower compared to younger adults. However, there were no st atistically significant differences, F (1,119) = .03, p > .05, between younger and older participantsâ€™ scores on positive affect The means and standard deviations of the vari ables of interest were examined for potential gender differences. There were no significant gender differences noted on participantsâ€™ scores on the individual difference measures; self-concept clarity, F (1,119) = 2.96, p > .05, positive affect, F (1,119) = .25, p > .05, Neuroticism, F (1,119) = .27, p > .05, and Extraversion, F (1,119) = .19, p > .05. Additionally, there were no significant gender differences noted in mean response latencies for MDA, F (1,119) = 2.04, p > .05, nor GSA, F (1,119) = .13, p > .05. Furthermore, no significant gender differences were noted in the mean salience ratings for MDA, F (1,119) = .85, p > .05, nor GSA, F (1, 119) = .06, p > .05. Tabachnick and Fidell (2001) recommend that to avoid issues of multicollinearity, one should not include variables with a bivariate correlation abov e .70 in the same analysis. A
76 bivariate correlation analysis was conducted for the covariates (word-choice reaction times, simple reaction time, general memory, and vocabul ary) to ensure that multicollinearity was not present. The results are shown in Table 3-1. The results revealed a correlation ( r = .78, p < .001) between word-choice reaction times and simple reaction times. Therefore, simple reacti on times were not included in subsequent data analyses. The decision to continue to use the wo rd-choice reaction times, rather than simple reaction times, was based on the observation that th e word-choice reaction task was similar in its task demands to the semantic recognition task used in this study. Th e bivariate correlation analyses showed that multicollinearity was not a concern for the remainder of the potential covariates. Covariates A number of age-related differe nces were expected that would conceivably influence the variables of interest. Specifically, age-related differences in cognitive function and reaction times were expected. In that, age-related differences in cognitive functioning and response times could lead to an erroneous interpretation of the data , the planned analyses called for using these variables as covariates if age-related differences were significant for these measures. In order to explore the influence of these age-related variations , an analysis of variance was conducted. Auditory Verbal Learning Task Participantsâ€™ general memory was assessed using the Auditory Verbal Learning Task (AVLT; Rey, 1941). Typically, younger adults scor e better on measures of general immediate memory than older adults. As was expected, si gnificant differences in performance were noted, F (1,119) = 7.34, p < .01. Younger adults correctly remembered more words ( M = 9.55, SD = 2.19) than older adults ( M = 8.41, SD = 2.43). However, there were no significant differences between the two groups in the number of word repetitions, F (1,119) =.96, p > .05, nor in the number of
77 errors of commission, F (1,119) = .27, p > .05. In order to contro l for these age-related differences in the general memory of younger and ol der adults, scores from the Auditory Verbal Learning Task (AVLT; Rey, 1941) were includ ed as a covariate in all analyses. Reaction Times Typically, age-related slowing in reaction times occurs. In order to control for differences in age-related slowing, reaction times were a ssessed using a word-choi ce reaction time task. Significant differences in reaction times of younger and older adults were found, F (1,119) = 34.73, p < .001. Young adultsâ€™ reaction times were faster ( M = .67 seconds, SD = .31) than older adults reaction times ( M = 1.08 seconds, SD = .44). Therefore, word-c hoice reaction times were used as a covariate in many analyses, as noted in the analyses. Vocabulary Participantsâ€™ general level of verbal comp rehension was assessed using the Advanced Vocabulary Test (V-3) from the Kit of Factor-R eferenced Cognitive Tests (Ekstrom et al., 1976). Older adults typically score bett er on measures of vocabulary th an younger adults. Differences in verbal ability may affect the complexity a nd coherence of participantsâ€™ narratives. Higher complexity and coherence of a narrative could feasib ly lead to lengthier narratives, with a greater number of implicit self-attributes embedded in the narrative. Add itionally, more advanced verbal ability could lead to significant differences in the number of explicit MDAs identified by the participants in their narratives. Therefore, verbal ability was examined as a potential covariate. No significant vocabulary differences were found between older ( M = 13.23, SD = 1.86) and younger adults ( M = 12.78, SD = 1.53), F (1, 119) = 2.07, p > .05. Therefore, part icipantsâ€™ scores from the Advanced Vocabulary Test were not incl uded as a covariate factor in the analyses. In summary, a number of preliminary analys es were conducted. A simple frequency check was conducted to assure that the values associat ed with each of the measures were within the
78 expected range and all anomalies were addres sed. Additionally, descri ptive analyses were conducted that included an examination of extr eme values, along with box plots and stem and leaf plots in order to identify data points lying significantly outside the normal range. After the outliers were identified, all anal yses were conducted with the outli ers removed as well as with them included in the data set. Comparison of th e outcome of the analys es revealed only minor differences. Therefore, findings were reported that represent al l of the data including these outliers. Analyses of variance ascertained that participantsâ€™ scores on individual difference measures exhibited the typical age-related patterns. Bivariate correlation analyses revealed multicollinearity between simple reaction time a nd word choice reaction. Therefore, simple reaction time responses were removed from subsequent analyses. The remainder of the covariates proved to be suitable variables for in clusion; however, the only covariates that were necessary, based on age differences, were the AVLT and word choice reaction time. Additionally, no significant sociode mographic differences related to ethnicity or age that might confound the interpretation of the data were noted. Finally, no signifi cant gender differences were noted that might obscure th e interpretation of the data. Thus , analyses designed to test the hypotheses of interest were conducted. Hypothesis 1: Relationship between Adults â€™ Self-Representations and Self-Defining Autobiographical Memories A 2 (age group: young adults vs. older adults) by 2 (attribute type : memory-derived vs. general-self) analysis of covariance (ANCOVA) was performed to examine age and attribute type differences with regard to participantsâ€™ re sponse latencies to person al attributes in the semantic recognition task. In or der to control for normal age-re lated slowing in reaction time when assessing participantsâ€™ respons e latencies with regard to self -attributes, participantsâ€™ wordchoice reaction times were included as a covariat e. In order to control for normal age-related
79 differences in general memory, participantsâ€™ scor es on the Auditory Verbal Learning Task (Rey, 1941) were also include d as a covariate. The results showed a significant main e ffect for attribute type (Hypothesis 1a). Specifically, the mean response latencies for MDAs ( M =1.14, SD = .27) were significantly smaller than response latencies for general self-attributes ( M =1.42, SD = .29), F (1, 119) = 8.28, p < .001. As hypothesized, the main effect of attr ibute type was qualified by a significant Age Group x Attribute Type interaction, F (1, 119 ) = 18.19, p < .001. No age differences in mean response latencies occurred for MDAs. However, age differences in mean response latencies were significant for general self -attributes (see Table 3-2). As predicted, no significant mean differences in the response latenc ies of young adults and older adults for the MDAs were noted afte r controlling for reac tion time (Hypothesis 1b). However, with regard to genera l self-attributes, a significant di fference in response latencies of younger adults and older adults was noted. Specifi cally, the mean response latencies for generalself attributes of older adults were significantly smaller than the latencies of younger adults, F (1,119) = 8.28, p < .001 (Hypothesis 1b). To examine differences in salience ratings of MDAs and ratings of ge neral self-attributes a 2 (age group: young adults vs. older adults) by 2 (attribute type: memory-derived vs. general) ANCOVA was performed. The results showed a si gnificant main effect for attribute type (Hypothesis 1c). Specifically, the me an salience ratings for MDAs ( M = 6.90, SD = .05) were significantly higher than th e salience ratings for ge neral self-attributes ( M = 4.41, SD = .77), t (1,119) = 105.16, p < .001. The main effect of attribute type was also qualified by a significant Age Group x Attribute Type interaction, F (1,119) = 6.60, p < .05 (Hypothesis 1d). Older adults rated their MDAs ( M = 7.10, SD = .46) as significantly mo re salient than young adults (M = 6.80,
80 SD = .64), F (1,119) = 8.82, p < .01. Mean salience ratings for ge neral-self attributes were not significantly different across age groups. In summary, results showed that the mean re sponse latencies for MDAs were significantly smaller than response latencies for general self-d escriptive self-attributes . With regard to age differences in response latencies, a signi ficant attribute by age interaction was found. Specifically, no age differences were noted for the response latencies associated with MDAs. However, age differences were noted with regard to response latencies a ssociated with general self-attributes. Additionally, with re gard to salience ratings, a significant main effect for attribute type was noted. The mean salience ratings for MDAs were significantly higher than the salience ratings for general self-attributes. In addition, a significant Age Gr oup x Attribute Type interaction qualified the main effect of attribute type. Olde r adults rated their MDAs as significantly more salient than young adults did. However, salience ra tings for general-self attributes were not significantly different for younger and older adults Hypothesis 2: Relationship between Valence of Self-Defining Autobiographical Memories, Individual Difference Variables, and Psychological Well-being Correlation analyses were performed to examin e the associations of the valence of selfdefining memories with psychological well-being and individual difference variables, such as self-concept clarity, positive affect and personality factors. In or der to perform the correlation analyses, total scale scores were calculated for each of the individual difference variables as well as for the measure of psychologi cal well-being. To create an overa ll positive valence rating, the question addressing the degree of positive fee lings connected with each autobiographical memory was used. These scores were then averaged for a total positive score for each
81 participant. Pearson-product moment correlations were th en calculated. The resulting correlations are summarized in Table 3-3. A significant positive correlation was noted be tween positive valen ce of self-defining autobiographical memories and self-concept clar ity. A significant negativ e correlation was found between positive valence of self-defining memori es and neuroticism (Hypothesis 2a). Although not statistically significant, the relationshi ps between positive valence of AM and other individual differences variables, and the corr elation between positive affect and extraversion were in the anticipated direc tion (Hypothesis 2a). The correla tion analyses also revealed a significant positive correlation between positive valence of autobiographical memories and psychological well-being (Hypothesis 2b). Although not directly related to the predicted hypotheses, a number of other significant correlations worthy of mention were also not ed. Specifically, significant positive correlations between self-concept clarity a nd psychological well-being were found, as well as between selfconcept clarity and positive affect were found. Co rrelation analyses also revealed a significant negative correlation between self-concept clarity and neuroticism. Additionally, a significant negative correlation was noted between psychol ogical well-being and neuroticism. These correlations support previous lite rature on self-concept clarity. In summary, several significant correlati ons between the valenc e of self-defining autobiographical memory, individual difference variables, and psychological well-being were found. Specifically, a positive relationship be tween positive valence of self-defining autobiographical memories and self-concep t clarity was noted. Conversely, a negative relationship between positive valence of self-def ining memories and neuroticism was also found. Additionally, a positive associati on between positive valence of autobiographical memories and
82 psychological well-being was noted. Finally, a significant positive relationship between psychological well-being and se lf-concept clarity was found. Hypothesis 3: The Mediational Effect of Po sitive Valence of Self-Defining Memories and Individual Difference Variables on Psychological Well-being. A series of hierarchical multiple regression an alyses was conducted to examine the role of valence of self-defining autobiog raphical memories as a mediat ing variable between individual difference variables and psychologi cal well-being. Mediation occurs as a sequence in which one variable affects a second variable that, in turn, affects a third variable. Specifically, the mediating variable was expected to have an impact on the relationship betw een a predictor variable and the dependent variable. Baron and Kenny (1986) outlined certain conditions that must be met in order for a variable to operate as a mediating vari able. Specifically, three conditions must be met. First, variations in the levels of the independent variable need to account for significant variations in the mediating vari able. Second, variations in the me diating variable need to account for significant variation in the de pendent variable, and third, when va riations in the levels of the mediating variable are included, the relations hip between the independent and dependent variable will no longer be significant. According to Baron and Kenny (1986), this appr oach allows for an examination of total mediation as well as partial mediation. Total mediation occurs when the primary predictor variable no longer significantly adds to the prediction of the dependent variable when the mediating variable is included (Baron & Kenny, 1986). Likewise, Baron and Kenny (1986) explain that partial mediation takes place if the me diator plays a significant role, but the primary predictor variable cont inues to predict signifi cant amounts of variance re lated to the dependent variable after the mediator is included. In order to examine th e mediating relationship between the valence of self-defining au tobiographical memories, and indi vidual difference variables on
83 psychological well-being, a series of regression analyses was conducted following the model set forth by Baron and Kenny (1986). To assure that none of the independent variable s predicts one another at a level that would be unacceptable, multicollinearity diagnostics we re conducted. Results of the multicollinearity diagnostics revealed the suitabil ity of the independent variables (individual difference variables and valence) for use in the regr ession analyses. Specifically, all tolerance values were above .88. Additionally, the values associated with the Vari ance Inflation Factor (VIF) were all well below 10. Therefore, all of the independe nt variables were used in subs equent regression analyses. In reporting the regressions, both the Beta values and the unstandardized b values were examined (see Table 3-4). Additionally, in order to ascer tain the amount of variance accounted for by the different models, the change in R2 for each model was also examined. Two separate regressions were conducted to examine the impact of the mediat ing variable. As outlined by Baron and Kenney (1986), the first regression analysis calls for the individual differen ce variables to be added first, followed by a second step in which the mediating variable is added to the model. The second regression calls for the regression to be conducted with the mediating variable alone, as a first step, followed by the addition of the individual difference variables to the model (Baron & Kenny, 1986). Individual Difference Variables Entered First. The next regression analysis in the seri es was conducted usin g both the individual difference variables and positive valence to predict psychological well-being in which the individual difference variables were entered simu ltaneously as a block followed by a second step in which positive valence was entered.
84 The individual difference variables, se lf-concept clarity, PANAS-positive affect, neuroticism, and extraversion, alone accounted fo r a significant percenta ge of the variance, F (4,116) = 15.93, p < .001, R2 = .36. However, adding positive vale nce to the model did not result in a significant increase in the variance accounted for, R2 change = .004, p > .25. This indicates that adding positive valence to the model did not in crease the predictive power of the model (See Table 3-4). Individual Difference Variables Entered Last Finally, the last regression analysis in the se ries was conducted. In this regression model to predict psychological well-being, positive valenc e was entered alone followed by a second step in which the individual difference variables were all entered simultaneously. Positive valence accounted for a significant pe rcentage of the variance in psychological well-being, F (1,119) = 7.07, p < .01, R2 = .06. Adding the individual difference va riables to the model resulted in a significant increase in the variance in psychological well-being accounted for, R2 change =.30, F (4,116) = 13.06, p < .001. This indicates that adding the individual difference variables to the model significantly increased the predictive power of the model (See Table 3-4). After testing the individual difference vari ables as a block, the relationship between positive valence and each of the individual predictor variables was also examined. Separate regression analyses were conducted to examine the role of positive valence as a mediator of psychological well-being, with each of the indivi dual difference variables: self-concept clarity, PANAS-positive affect, neuroticism, and extraver sion. Positive valence was not a mediator of well-being in any of these analyses. The results indicated that the effect of positive valence did not account for a significant portion of the vari ance in psychological well-being beyond what was already accounted for independently by each of the critical individual difference variables: self-concept clarity, PANAS-positive affect, neur oticism, and extraversion. To confirm these
85 results, a Sobel (1988) test was ut ilized to test significance of the partial mediating effect of positive valence. In no case was there any significant mediation by positive valence. In sum, regression analyses showed that the positive valence of self-defining autobiographical memories predicted a small porti on of the variance in scores on psychological well-being when entered as a sole individua l variable. Additionally, adding the individual difference variables, self-concept clarity, PANASpositive affect, neuroticism, and extraversion, accounted for a significant increase in the predic tive power of the model. Positive valence did not contribute more variance after the individua l differences variables were entered; positive valence did not mediate the relationship betw een the individual di fference variables and psychological well-being.
86 Table 3-1 Bivariate Correlations between the Covariates (N=120) ______________________________________________________________________________ Variables 1 2 3 4 ______________________________________________________________________________ 1. Word-Choice Rt --.78** -.12 -.07 2. Simple Rt ---.13 -.05 3. AVLT --.07 4. Verbal --______________________________________________________________________________ Note. AVLT = The Auditory Verbal Learning Ta sk; Verbal = The Advanced Vocabulary Test (V-3) from the Kit of Factor-R eferenced Cognitive Tests. ** p < .01. Table 3-2. Mean Response Latencie s by Age Group and Attribute Type _____________________________________________________________________________ Young Adults Older Adults ( n = 59) ( n = 61) Attribute Type M SD M SD ______________________________________________________________________________ MDAs 1.09 .03 1.18 .03 General 1.49 .03 1.35 .04 ______________________________________________________________________________
87 Table 3-3. Relationship Among Positive Valence of Self-Defining Memories, Individual Difference Variables, and Psyc hological Well-Being (N=120) Variables 1 2 3 4 5 6 ______________________________________________________________________________ Note. Valence = positive va lence of self-defining memories; SPWB = Short Psychological WellBeing Scale; SCC = Self-Concept Clarity Scale; NEO-N = NEO Neur oticism Scale; NEO_E = NEO Extraversion Scale; PA = Positive Affect Scale. * p < .05. ** p < .01. 1. Valence --.25** .28** -.28** -.06 .10 2. SPWB ----.55** -.37 .11 .26** 3. SCC -------.50** .11 .06 4. NEO_N --------.09 -.11 5. NEO-E ----------.10 6. PA -------------
88 Table 3-4. Summary of Regression Analyses Pr edicting Psychological We ll-being by Individual Difference Variables and Positive Valence of Autobiographical Memories __________________________________________________________________ B SE B Beta _ Step 1 SCC .04 .01 .47** PA .18 .06 .21** NEO-N -.10 .08 -.11 NEO-E .04 .08 .04 R2 .36** Step 2 SCC .04 .01 .46** PA .17 .06 .21** NEO-N -.09 .09 -.09 NEO-E .05 .08 .04 Valence .04 .05 .07 R2 .36 R2 change .00 ____________________________________________________________________________________________________ ___________________________________________________________________ B SE B Beta ___ ______ Step 1 Valence .14 .05 .24** R2 .06** Step 2 Valence .04 .05 .07 SCC .03 .01 .46** PA .17 .06 .20** NEO-N -.09 .08 -.10 NEO-E .05 .08 .04 R2 .36 R2 change .30** _____________________________________________________________________ Note. Valence = positive valence of self-defining memories; SCC=Self-Concept Clarity Scale; NEO-N = NEO Neuroticism Scale; NEO-E = NEO Extraversion Scale; PA = Positive Affect Scale. * p < .05. ** p < .01.
89 CHAPTER 4 DISCUSSION Understanding the relationships between a dultsâ€™ self-concept and autobiographical memory provides valuable insight into the role that self-concept and autobiographical memory may play as psychological resources in the aging process (Brandtst dter & Greve, 1994). Theorizing about self-concept as a resource conver ges with recent theorizi ng on the functions of autobiographical memory. There seems to be consensus that individualsâ€™ autobiographical memory and personal identity are intimately intertwined (Conway & Pleydell-Pearce, 2000; McAdams, 1996), and this research was designed to tease out several aspects of this relationship. How can we expect self-representations and au tobiographical memory to be linked with each other? Following the arguments of Brandtstdter and Greve (1994) regarding the importance of specific self-representations, only certain autobiog raphical memories would be expected to be relevant for a personâ€™s self-concept. This subset of autobiographical memories has been referred to as self-defining memories (Pillemer, 1998; Singer & Salove y, 1996). Self-defining memories contain essential information abou t the self and it is reasonable to assume that, over time, this information becomes semantically represented as part of a personâ€™s self-concept. Self-defining memories act as a resource because they provide a person with a sense of self-continuity across time that directly contributes to successful aging (Costa, Metter, & McCrae, 2005). These key memories that a person chooses to review freque ntly may also play a ro le in self-regulation, affect management, and well-being. In terms of mood regulation and psychological well-being, people may choose to focus on memories of the pa st that elicit either positive emotions or negative emotions which could have an impact on the personâ€™s current affective state as well as his or her view of the future (Cohen, 1998). The focus of this study is on those self-defining autobiographical memories that ac t to preserve adultsâ€™ identity.
90 Empirical research on the iden tity function of autobiographi cal memory has been rather scarce and there are no studies that have empiri cally investigated the connection between selfdefining autobiographical memories and semantic representations in adultsâ€™ self-concept. There have been a few studies that ha ve explored the relationship betw een valence of autobiographical memories and individual differences in personali ty traits such as neuroticism and extraversion. Likewise, a few studies ha ve explored the relationship of va lence of autobiographical memories and psychological well-being. However, to date th ere have been no studies that have examined the interconnected relationship of positive valence of autobiographical memories, individual differences, and psychological well-being simu ltaneously. This study expands upon those few earlier studies, first, by using an experimental design to exam ine the relationship between the identity function of self-defining autobiographica l memories and their semantic representations across adulthood, and secondly, by exploring the pos itive valence of self-d efining memories in relation to individual differences and psychological well-being. The first section of the discussion reviews the results associated with the first objective of this investigation, exploring se lf-representations and autobiog raphical memory across adulthood. In the second section, the relationships among ke y variables are discusse d. These key variables are positive valence of self-def ining autobiographical memories , individual differences, and psychological well-being. The third section revi ews the findings associated with the third objective â€“ the role of positive valence as a medi ator between individual difference variables and psychological well-being. Next, I discuss limitations of the study and offer suggestions for future research. Finally, the conclusions focus on the extent to which this study offers a new perspective on important issues related to self-concept and autobiographical memory.
91 Objective 1Self-Representation s and Self-Defining Memories One of the recognized functions of autobiograph ical memories is their role in the creation and maintenance of the self-concept (Cohe n, 1998). Because self-defining memories, by definition, contain essen tial information about th e self, it is reasonable to assume that this information becomes semantically represented in the self-concept over ti me. Consequently, one of the primary objectives of the study was to examine the relationship between aging, selfdefining autobiographical memories, and semantic representations in a dultsâ€™ self-concept. To document the link between se mantic representations in adultsâ€™ self-concept and selfdefining autobiographical memories we derived self-attributes fr om participantsâ€™ self-defining autobiographical memories (i.e., memory-der ived attributes, MDAs). These MDAs were expected to represent central features of an i ndividualâ€™s self-concept. Comparisons of MDAs and general self-attributes were conducted, compari ng response latencies and salience ratings as a function of attribute type. When asked to make decisions about the self-descriptive traits, people were expected to respond much more quickly, in terms of response latencies, to highly selfdescriptive traits than to tra its that may or may not be self -descriptive (Siem, 1992; Fekken & Holden, 1992; Markus, 1977; Saniti oso et al., 1990). Specifically, re sponse latencies to attributes derived from autobiographical memory and general self-attributes were e xpected to vary as a function of attribute type and ag e. Numerous studies have demons trated age-related slowing in processing semantic information (Hertzog, Raskind, & Cannon, 1986: Mueller & Johnson, 1990). Nonetheless, due to their centrality to self-concept, younger and older adults were expected to respond equally fast to MDAs. However, age differen ce in response times to general self-attributes were expected to mirror the ge neral literature on age differences in semantic recognition studies (Hertzog et al., 1986; Mueller & Johnson , 1990). Specifically, larger
92 response latencies to general self -attributes were anticip ated for older adults due to age-related slowing in retrieval of general semantic information. In addition, examination of salience ratings served two purposes. Firs t, the salience rating of MDAs provided a manipulation check to assure that the attributes coders extracted from participantsâ€™ self-defining au tobiographical memories were appropriate for both age groups. Secondly, the salience ratings prov ided an additional measure of the attributesâ€™ prominence with regard to the participantâ€™s self-concept, with hi gher salience expected for MDAs than general self-attributes, especially for older adults. Thus, we predicte d an Age Group x Attribute Type interaction. Response Latencies of Attributes Associ ated with Self-Defining Autobiographical Memories As expected, participants responded faster to MDAs than to general self-attributes during the semantic recognition task. This indicates that MDAs we re more readily accessed, and therefore more clearly defined in adultsâ€™ self-concept. These findi ngs are consistent with earlier work documenting the link betw een adultsâ€™ self-concept and autobiographical memory. Specifically, this finding converges with earlie r work documenting the identity function of autobiographical memory (Bluck, Haberm as, & Rubin, 2001). Sanitioso, Fong, and Kunda (1990) also found that individuals accessed autobiographical memori es reflecting self-descriptive attributes more quickly than memories that were not self-descripti ve. Sanitioso and his colleagues (1990) utilized a method in which respons e latencies were measured as indicators of variations in the centrality of self-attributes. Adding to this gr owing evidence, the current study validates the heuristic value of us ing response latencies to assess th e centrality of self-descriptive information.
93 Age Differences and Response Latencies Response latencies were also used to examin e potential age differences in access to key aspects of the self-concept. The results showed no significant differen ces in the response latencies of young and older adults for the MDAs, supporting the second hypothesis that attributes derived from self-defining memories would be accessed quickly and equally fast by younger and older adults. This suggests that seman tic representations derived from self-defining autobiographical memories may hold a special (i.e., more central) place in the semantic space that makes up individualsâ€™ se lf-concept. We expected no ag e difference here because young people in late adolescence become capable of integrating multiple and more complex conceptions of self (Harter & Monsour, 1992). We have documented that young as well as older adults respond more quickly to attributes deri ved from self-defining me mories for weeks after those attributes were originally derived from their memories. This supports the notion that like older adults, young adults have also devel oped a consistent core sense of self. Contrary to our expectations, older adults responded more qu ickly than younger adults to general self-attributes. Older adu lts made quicker decisions about the relevance of an attribute regardless of whether it was a gene ral self-attribute or MDA. Although this was not the original hypothesis, upon reflection, the most lik ely explanation for this result is that the se lf-concepts of older adults are more clearly defined than thos e of younger adults, and this difference may have affected both general attributes and MDAs. Recall, preliminary analyses revealed that older adultsâ€™ scores on the measure of selfconcept clarity were significantly higher than yo unger adultsâ€™ scores indi cating that self-concept clarity was greater for older adults. With greater self-concept clarity, older adults should show greater clarity about who they are, but also gr eater clarity about who th ey are not. In keeping with this, older adults were quick to respond to attributes that were central to their conceptions of
94 self; in turn, they also responded qui ckly to attributes that were not part of their core sense of self or that were only peripherally rela ted to their conceptions of self . For example, if a person has a clear sense of being extroverted, then that person would respond quick ly to make a â€œMEâ€ decision to that attribute, but also would decide quickly, in evaluating th e general self-attributes, that the attribute â€œintrovertâ€ di d not apply to his or her self-c oncept. Based on this result we should expect, in the future, that an individual wi th a clearly defined sens e of self would respond quickly both to self-defining attrib utes and to general descriptors. These findings converge with theoretical a ssumptions addressing developmental tasks associated with various stages across the life span. Theorists propose that the developmental tasks associated with adoles cence and young adulthood focus on defi ning oneâ€™s personal identity (Erikson, 1980). McAdams (1996) asserts that pe ople accomplish this task by developing their life story, starting in childhood and coming together in adolescence. The individual's sense of self continues to evolve thr oughout adulthood as the li fe story is further developed (McAdams, 1996) which may result in a more clearly defined se nse of self in later life (Bauer, McAdams, & Sakaeda, 2005; King, Scollon, Ramsey, & Williams, 2000), although this has rarely been examined. Self-concept development has b een well documented throughout childhood and adolescence (Damon & Hart, 1988; Ha rter, 1998), but not in later li fe. Numerous theorists have suggested that adults' self-concepts are dynami c and change to meet developmental demands across the life span, and have suggested a pe rson's self-concept may become an important resource in facing the challenges associated w ith the aging process (Brandtstdter & Greve, 1994; Cross & Markus, 1991). When faced with age -related challenges, the flexible adjustment of self-representations in a manne r that also allows for self-con tinuity and stability of selfconcept is one way in which a personâ€™s self-con cept may act as a psycho logical resource (Greve
95 & Wentura, 2003). The finding of this study, adds to this previous literature by explicitly examining those central aspects of the self-concept e xpected to play a role in self-continuity as a person ages. Salience of Semantic Representations Deri ved from Self-Defining Autobiographical Memories Attributes derived from a personâ€™s self-defin ing autobiographical memories, by definition, are expected to be related to central aspects of a personâ€™s self-concept. Central aspects of a personâ€™s self-concept should be mo re salient than attributes that are more peripheral to that personâ€™s self-concept. Therefore, the salience ratings for MDAs were expected to be significantly higher than the salience ratings for general self -attributes, and that ex pectation was confirmed. The attributes rated as most sa lient were those attributes deri ved from the autobiographical memory narratives. This finding supports the view that self-defining au tobiographical memories become represented semantically in memory as a part of a personâ€™s self-concept, and it lends further support to the growing be lief that autobiographical memory and identity are undeniably coupled (Conway & Pleydell-Pearce, 2000; McAdams, 1996). To date there have been no studies requiring participants to rate the salience of self-attributes related to their self-defining autobiographical memories. However, the findings in this study converge with similar work that relates participantsâ€™ response latencies to attri butes as an indication of salience (Fekken & Holden, 1992; Markus, 1977; Sanitioso et al., 199 0). The high salience ratings given to MDAs also substantiate the value of this methodology of deriving attributes from autobiographical memories. In this respect, this study makes an important methodological contribution to the current literature.
96 Age Differences and Salience If self-concept continues to de velop across the life span as many theorists have suggested, one would anticipate differences in older and younge r adultsâ€™ conceptions of the self. In that the developmental task associ ated with adolescence and young adulthood is creating a multidimensional self-concept, one would expect the self-concepts of younger adults to be somewhat less defined than those of older adults who have had considerably more time to think about and refine their self-concepts. This deve lopmental pattern should be reflected in the salience ratings of self-attributes. As expected, compared to younger adults, olde r adults rated their MDAs as significantly more salient. These findings conve rge with previous studies that demonstrate variations in selfrepresentations at different point s across the lifespan in the content, structure, and organization of the self-concept (Campbell et al., 1996; Cros s & Markus, 1992; Diehl et al., 2001; Donahue et al., 1993; Freund & Smith, 1999; Kling et al., 1997). These findings also sugge st that the central position of MDAs seems to make them robust agains t the age-related decrements that are usually observed with regard to the speed of accessibility of other kinds of semantic content (Kausler, 1994; Perfect & Rabbitt; 1993; Ri by, Perfect, Stollery, 2004). Objective 2: Relationships among Valen ce of Self-defined Memories, Individual Differences, and Psychological Well-Being A number of scholars have suggested li nkages between autobiographical memory, individual difference variables, and psychol ogical well-being (Grossb aum & Bates, 2002; McAdams et al., 2004; Moffitt & Singer, 1994; Rubi n & Siegler, 2004). Here, we have placed an emphasis on the valence of self-defining autobi ographical memories. We expected valence to play a central role in the rela tionships among variables representing aspects of the self. Valence seems to be a promising aspect of autobiographi cal memory to act as a psychological resource
97 (Grossbaum & Bates, 2002). First, the valence of an autobiog raphical memory that a person selects to rehearse or reflect upon may infl uence his or her current mood. By focusing on personal memories that impart positive affect, in dividuals may ultimately enhance their overall psychological well-being. In addition, the va lence of autobiographical memories shows associations with other key aspect s of the self. Scholars have noted that the emotional valence of autobiographical narratives is related to pe rsonality (McAdams et al., 2004) and to well-being (Grossbaum & Bates, 2002). In this study, we c hose to focus on participantsâ€™ ratings of the positive valence associated with their self-definin g memories. Although this approach is slightly different from earlier studies that examine overall valence, we believe our measure adequately captures the positive valence of autobiographical memory narratives. In an attempt to further understand these inte rrelationships, correlati ons were calculated among variables: positive valenc e of self-defining autobiogra phical memories, individual difference variables (self-concept clarity, posi tive affect, extraversion and neuroticism) and psychological well-being. The positive emotional valence of self-defin ing autobiographical memories was expected to correlate positively with self -concept clarity, positive affect, and extraversion. Conversely, a negative association between posit ive valence of autobiographi cal memories and neuroticism was expected. Neuroticism tends to be positiv ely associated with a nxiety and negatively associated with subjective well-being. Theref ore, finding an inverse relationship between positive valence and Neuroticism would be expected. We also examined valence and other individual difference variables in relation to psychological well-being. A positive association between positive valence of autobiographical memories and positive aspects of psychological well-being was expected. Additiona lly, positive associations were also expected amongst self-
98 concept clarity, positive affect, extraversi on and psychological well-being. Conversely, a negative relationship was expected between neuroticism and psyc hological well-being. Valence of Self-Defining Memories and Individual Difference Variables As was expected, higher positive valence scores (associated with the self-defining autobiographical memories) tended to be associat ed with higher levels of self-concept clarity. This suggests the valence individuals assign to self-defining experiences is linked to stronger conceptions of self, an issue left unexplored in previous research. Higher levels of self-concept clarity reflect higher levels of coherence and better integratio n of a personâ€™s self-concept. (Campbell et al., 2003). Interestingl y, the central events that peopl e used for their self-defining memories were often very negative. Overall, it was not triumph and success at the core of peopleâ€™s self-defining memories; rather tragedy, pe rsonal crises and challenge were at the heart of these narratives. To ultimately portray these ev ents in a positive light, that is, to maintain a high positive valence rating, indivi duals would need to integrate the positive as well as negative aspects of such events, or they would need to hold in mind multiple, and quite possibly disparate, conceptions of the self simultaneously. In this way, positive valence for self-defining autobiographical memories could be related to the unity of oneâ€™s self-concept; a higher positive valence may reflect the fact that this integrat ion process has occurred. Such a unified or wellintegrated self-concept should be associ ated with higher self-concept clarity. Correlation analyses also supported the association between positive valence of autobiographical memories and neuroticism. Higher levels of positive valence associated with the self-defining autobiographical memories tended to be associated with lower levels of neuroticism, suggesting that th e valence individuals assign to self-defining experiences is linked to aspects of the self. This study supported ea rlier findings related to positive valence of autobiographical memories and neuroticism (M cAdams et al., 2004; Rubin & Siegler, 2004).
99 Contrary to the stated hypothesis, a relationship between positive valence of autobiographical memories and overall positive a ffect was not evident; nor was the relationship between positive valence of autobiographical memory and extraversion significant. Although earlier studies examined overall emotional t one, based on these findings, positive valence of autobiographical memories was expected to co rrelate with both positive affect and extraversion (Rubin & Siegler, 2004). Specifi cally, Rubin and Siegler (2004) found participantsâ€™ scores on measures of Extraversion correlated with phenomenological characteristics of the autobiographical memory narratives related to emotional tone of the narrative. This unexpected finding may be related to differen ces in the administration order of the respective measures in these two studies. The measure of affective state in this stud y was given some time after participants completed narrating their autobiogr aphical memories. Therefore, the measure of positive affect may not have adequately captured participantsâ€™ affective state at the time that the autobiographical memory was shared. Valence of Self-Defining Autobiographical Memories and Psychological Well-Being An association between valence of autobiogr aphical memory and psychological well-being has been established in previous research (M cAdams, Reynolds, Lewis, Patton, & Bowman, 2001; Josephson et al., 1996). Psychological we ll-being, in a broader sense, is intimately connected to affect. The valence of autobiogra phical memories influences a personâ€™s affective state. Thus, the valence of self-defining autobiographical memories should influence psychological wellbeing. The expected relationship between positive valence and psychological well-being was supported by the data of this study. Participants who reported higher levels of positive emotions associated with their self-defin ing autobiographical memories al so reported higher levels of psychological well-being. This finding could have multiple interpretations and deserves further
100 exploration. Perhaps, focusing on the positive asp ects of oneâ€™s AMs may allow people to bolster important self-representations in a manner that facilitates psyc hological well-being. It may also be the case that a personâ€™s ability to integrate positive and negative aspects of life-changing events is associated with structural aspects of self-concept that are li nked to psychological wellbeing. Individual Difference Variables and Psychological Well-Being Although not directly related to the predicted hypotheses, seve ral significant associations among the individual difference variables and psychological well-being were noted. A strong relationship between self-concep t clarity and psychological wellbeing was in evidence. Higher levels of self-concept clarity were associated with higher levels of psychological well-being. This suggests a united view of oneself is particul arly important to psychological well-being. This finding converges with previous work by Camp bell and her colleagues (Campbell et al., 1996; Campbell et al., 2003). Campbell and her coll eagues (1996) found evidence of a strong relationship between self-concept clarity and se veral other psychological variables associated with psychological well-being. Using a different measure of self-conc ept unity, se lf-concept differentiation, Diehl and his colleagues (2001) al so found a relationship between the unity of adultsâ€™ self-concept and ps ychological well-being. Participants who reported higher le vels of self-concept clarity al so reported higher levels of positive affect. This finding converges with previous work by Campbell, Assanand, and di Paula (2003). They found moderate correl ations between lower levels of self-concept clarity and higher levels of negative affect. Although negative affect and positive aff ect should not be construed as opposing constructs, the findings fr om this study further clarify the relationship between selfconcept clarity and affect.
101 An association was noted between self-concept clarity and neuroticism. Participants who reported higher levels of self-concept clarity re ported lower levels of neuroticism, supporting earlier work (Campbell et al., 1996 ). Lower levels of neuroticis m are normally associated with higher levels of psychological well-being (K eyes et al., 2002; McCrae, 1983). Thus, the relationship noted in this study between self-concept clarity and neuroticism can be seen as providing additional evidence of the link between self-concept clarity and measures of psychological well-being. Objective 3Valence as a Mediator The third major objective of the study was to explore the role of positive valence of selfdefining memories as a mediating variable between individual di fference variables and psychological well-being. Seve ral theorists have proposed that adults may use their autobiographical memories to adjust their se lf-concept in response to age-related challenges and/or losses. The autobiographical memories pe ople focus on or rehearse can influence their current affective states. In this manner, a longterm focus on autobiographical memories with positive valence may lead to higher psychological well-being. If self-d efining memories are central to the self-concept, the valence of these memories may be more powerful than individual differences, such as personality traits, in thei r impact on well-being. Valence may influence the way in which other individual diffe rence variables, such as self -concept clarity, positive affect, extraversion, and neuroticism affect well-being, even though these variab les should also show a significant relationship with ps ychological well-being (Campbell, Assanand, and di Paula, 2003; McAdams et al., 2004; Rubin & Siegler, 2004). He re, we explored the possibility that the positive valence of self-defining autobiographical memories would mediate this relationship, at least partially.
102 As expected, the individual difference vari ables (self-concept clarity, positive affect, extraversion, and neuroticism) significantly pred icted the positive valence of the self-defining autobiographical memories. This converges with previous research s uggesting that certain individual difference variables are related to specific phenomenologica l characteristics of autobiographical memory, specifically emotiona l valence of the memory (McAdams et al., 2004). For instance, Rubin and Siegler (2004) al so found evidence of a link between individual differences in personality traits, such as extraversion, and ratings of emotionality for autobiographical memory narratives. The pres ent study also expands upon earlier research. Neither positive affect nor self-concept clarity has been included in ea rlier studies examining individual differences and valen ce of autobiographical memory. As expected, the individual difference vari ables (self-concept clarity, positive affect, extraversion, and neuroticism) reliably pred icted psychological well-being. These findings support previous research. Keyes, Shmotkin, a nd Ryff (2002) showed that psychological wellbeing increased as extraversion increased and neuroticism decreased. Likewise, Campbell and her colleagues (Campbell et al., 1996; Campbell et al., 2003) also demonstrated the relationship between self-concept clarity a nd psychological well-being. Here, a set of individual difference variables were entered as a block, so that their interrelationships were considered in predicting well-being â€“ the impact of each variable was eval uated while the others were controlled for. Using this procedure, it was apparent that self-concept clarity was the most important predictor of well-being. I will discuss this issue more in the following section. The positive valence associated with individualsâ€™ autobiograph ical memories emerged as a significant predictor of psychol ogical well-being, although it acc ounted for only a small portion of the variance. This finding suppor ts previous research showing that variations in the valence of
103 individualsâ€™ autobiograph ical memory narratives may be associ ated with levels of psychological well-being (Grossbaum & Bates, 2002; Seidlitz & Diener, 1997) . Grossbaum and Bates (2002) found levels of positive emotional valence of autobiographical memory narratives to be correlated with several domains of psychologi cal wellbeing, includ ing Self-Acceptance, Positive Relations with Others, Environmental Mastery and Purpose in Life as measured by Ryffâ€™s Short Psychological Well-being Scale (R yff, 1989). Additionally, Seidlitz and Diener (1997) found that individuals who recalled more positive autobiographical memories also had higher scores on measures of psychological well-being. Positive valence failed to mediate the impact of the individual difference variables on psychological well-being. Although positive valence di d not play the expected mediational role in this sample, future research may yet discove r that this approach ha s validity. There are a number of different ways to assess the valenc e of an autobiographical memory. The present study used a particular approach, focusing on the positive ratings for self-defined memories. This was the first attempt to test this mediation, and alternative methodologies may yet prove more fruitful than this particular a pproach (see future directions). In stead, the strongest finding in these regression analyses was the powerful impact of self-concept clarity on psychological well-being, accounting for over 25% of the variance. Given that self-concept clarity re flects an integrated unified conception of self, this finding supports earlier studies that have demonstrated the relationship between psychological well-being and self-concept unity, in the form of self-concept clarity (Campbell et al., 2003) and se lf-concept differentiation (Diehl et al., 2001; Donahue et al., 1993).
104 Study Limitations and Future Directions Limitations There were 3 primary limitations of this st udy. There were two limitations associated with the methodology. A third primary limitation was a ssociated with the sample characteristics. Creating unique lists of self-attributes for each participant had advantages and disadvantages. There were two primary issues that must be taken into consideration when faced with such a task. The first issue is associated w ith the general self-attributes, the second with the MDAs. Although the list of general self-a ttributes used in this study was carefully chosen based on empirical evidence of its psychometric properties and use in previous studies, the list may nevertheless suffer from certain limitations. There were several issues considered when choosing this particular list of traits. Th e list of self-attributes used in this study (Sheldon et al., 1997) was a subset of attributes from a more comprehensiv e list of traits originally compiled by Goldberg (1992). Sheldon and his colleagues (1997) scruti nized the subset of self-attributes and demonstrated that the list reflected attributes a ssociated with the five major personality factors captured by the NEO-FFI (Costa & McCrae, 1989). Although any valid li st of attributes will contain general traits that will adequately capture individual traits, such a list could also fail to include one specific self-attribute that might be central to a partic ular individual. By adding those important central self-a ttributes derived from individualsâ€™ self-defining memories, we created lists of self-attributes that were unique for each participant, lists that we knew included individual traits that would be endorsed. In doing so, we no l onger had a single general list of traits that all participants used for their ratin gs. Nonetheless, I believe the advantage of adding self-attributes derived from participantsâ€™ memo ries outweighed the inhe rent disadvantages of using slightly modified trait lists across participants.
105 The second challenge was to cr eate valid lists of MDAs ba sed on participantsâ€™ memory narratives. Although carefully cont rolled, culling self-attributes from the memory narratives proved to be a daunting and complicated task, on e in which certain decisions with regard to practicality had to be made. Precauti ons were taken to assure that the lists of MDAs reflected the meaning the participants intended. First, the coders were trained extensivel y and were required to maintain a minimum level of agreement with regard to the attributes taken from each narrative. Nonetheless, both coders were young adults of a similar age and background. Although the coders were in agreement with each other with regard to th e MDAs, it is possible that the salience of specific words chosen as MDAs could vary for different age groups. For example, an older adult may assign a slightly higher salience rating to the word â€œfeistyâ€ compared to a more contemporary synonym chosen by the coders, such as â€œspunkyâ€. In order to addr ess this potential limitation, salience ratings were used as a manipulat ion check to assure that the self-attributes derived by coders were accurate. The salience rati ngs showed that partic ipants rated the MDAs as highly salient, certainly signi ficantly more so than general se lf-attributes, demonstrating that the codersâ€™ words correctly repres ented important features of the participantsâ€™ self-concept. An alternative approach might have been to ask pa rticipants to supply their own explicit selfattributes. This approach may have more accura tely reflected participantsâ€™ actual semantic representation of their self-concept. However, this could possibly have resulted in a list of MDAs that overwhelming represented positive traits, rather than both positive and negative traits. Complicated as it was, coder derivation of MDAs seemed like the best solution, and the obtained salience ratings verify the value of this approach. Another issue related to generating an accurate list of MDAs is that the coders were not blind to the age and gender of the participantsâ€™ narratives. Initial attempts to mask age and
106 gender resulted in meaningless incoherent narrat ives, therefore the narrati ves were left intact. This could have introduced an age or gender bias with regard to the MDAs selected by coders. If such biases existed, the selected attributes may not have been the most accurate reflection of the semantic representations embedded in particip ants self-defining aut obiographical memories. Given a behavioral description em bedded in the narrative that calle d for extraction of an implicit self-attribute, coders could have assigned di fferent MDAs based on the age and gender of the participant. This could potentia lly result in lists of MDAs that reflected these biases of the coders. In order to address this potential bias, coders completed a scale measuring age bias (Palmore, 1998) and none of the code rs displayed any age bias. Another methodological limitation of this study involves differences in the age of the memories supplied by younger and older adults. By vi rtue of differences in their age, the amount of time that had passed since an event occurr ed could be markedly different for younger and older adults. For example, two participants na rrating memories from the same developmental age, letâ€™s say adolescence, may represent a time passage of only a few years if both participants are young adults, yet for the older adult it may be 60 or more years since the event occurred. It could be that memories of important life events become more firmly entrenched in memory with time. This may result in the memory becoming mo re easily accessible, not because it is integral to a personâ€™s self-concept, but because it has been rehearsed more frequently. This potential issue is inherent in all autobiog raphical memory research. The sample may have been the source of two additional limitations to the study. First, the diversity reflected in the sample was rather limited in three primary ways. Although the age range of the young adults was more than adequate, the majority of the young adults in the sample were college students rather than young adults working in the community. Bluck and Glck
107 (2004) noted differences in the degree to whic h adolescents, younger and older adults reported learning lessons from autobiographical memories that might impact their self-concepts. In the study conducted by Bluck and Glck (2004), the a dolescent group consisted of people between the 15 years and 20 years of age and the young a dult group consisted of people 30-40 years of age. Given that the young adult sample in this study consisted of indivi duals primarily between 18 and 22 years of age, the findings may not ge neralize to a young adult population consisting of a greater age range. Additionally, in this sample the education le vel of both the older and younger adults was higher than one would expect to find in the general population. The rela tionship between levels of education and autobiographical memory is ambiguous. To date studies have not focused on the relationship between levels of education and self-defining autobiographical memories. However, a few studies have doc umented differences associated with levels of education and autobiographical memory (Borrini, Dall'Ora, della Sala & Marinelli, 1 989; Cohen & Java, 1995). For example, Cohen and Java (1995) noted differen ces in higher degrees of accuracy and greater clarity of autobiographical recall was associated with higher levels of education. C onversely, in a study explicitly examining the re lationship between levels of education and autobiographical memory, Janssen, Chessa, and Muerre (2005) di d not find a significant impact of education on various aspects of autobiographical memory. Th e relationship between levels of education and autobiographical memory remains unclear. Like wise, the potential impact of education on autobiographical memory is not understood. Therefor e, given that this sample is comprised of a relatively highly educated group of people, cautio n should be exercised when extending these findings beyond this study.
108 Finally, with regard to the sample, the et hnic diversity was limited. The young adults in the sample adequately reflect th e ethnic diversity of the Alachua county census. However, even though a great deal of effort was made to recrui t minorities for this study, the majority of older adults in the sample were Caucasian. The difficu lty of obtaining an ethnica lly diverse sample of older adults is due in part to the fact that ol der adults make up less than 10% of the population of Gainesville. Recruiting an ethnically diverse sample in Gainesville requires much more time than that available for a dissertation project. Cultural variations in the development of autobiographical memory in children have b een widely noted (Bauer et al., 2005; Nelson, & Fivush, 2004). However, recently a few studies have also noted ethnic variations in autobiographical memory narratives of adu lts (Janssen, Chessa, & Muerre, 2005; Wang & Conway, 2004). Given that these ethnic differences in adultsâ€™ autobiographical memory have been documented and yet have recei ved little attention to date, cau tion should be exercised when extending the findings from this study to ethnically diverse groups. A second limitation related to the sample is that only two age groups, younger and older adults, were sampled. Clear differences were noted between response latencies and salience ratings associated with MDAs of younger and ol der adults. However, without including middleaged adults in the study, conclusions about the ch anges in adultsâ€™ self-defining autobiographical memories and self-concept acro ss the life span cannot be substantiated. Including middle-aged adults in the study would provide a continuous age distribution and allow for more definitive statements to be made with regard to the re lationship between self-d efining autobiographical memories and self-concept across the life span Similarly, the use of a crosssectional approach limits c onclusions about age-related patterns to a discussion of age di fferences. The presence of age differences in a cross-sectional
109 design could be due to cohort fact ors, and not just age-related change. A longitudinal or crosssequential study would be necessary to properly a ddress changes that may occur in self-concept and autobiographical memory over the course of the adult life span. Therefore, beyond the confines of this samp le and a cross-sectional design, one should exercise caution when drawing conclusions. Likewi se, further exploration of the accuracy of the self-attributes drawn from self-defining autobi ographical narratives should be conducted before definitive conclusions can be drawn. These limita tions provide some direction for future study. Future Directions The study of the functions of autobiographical me mory is still in its early stages. For the most part, what we know of the identity function of autobiographical memory is based in theory and as of yet, remains largely empirically untested . This study represents a first-step in exploring several aspects of autobiographical memory and self-concept. Specificall y, this study empirically investigated the connection betw een self-defining AMs and seman tic representations in adultsâ€™ self-concept and demonstrated the effectiveness of using a semantic recognition paradigm to examine this connection. Th ere are a number of dire ctions for future study. First, the semantic recognition paradigm allowe d us to substantiate the link between adultsâ€™ self-defining memories and self-re presentation. Now that the heuristic value of this paradigm has been established, it can be uti lized in new studies exploring the relationship between adultsâ€™ semantic representations and au tobiographical memory. Specifical ly, age differences could be explored in the following manner. In this study, participantsâ€™ response la tencies were measured as they responded â€œMEâ€ or â€œNot Meâ€ to a variet y of self-attributes. It may be informative to examine participantsâ€™ response la tencies to â€œNot Meâ€ decisions in future studies. In earlier studies, examinations of response latencies to se lf-attributes that were either schematic or aschematic have been valuable (Markus, 1977; Sanitioso et al., 1990). In a similar manner,
110 examining response latencies of the â€œNot Meâ€ d ecisions of younger and olde r adults would likely shed more light on the extent to which the self-concept is well defined in younger and older adults. That is to say, if younger adults have a less well-defined self-concept compared to older adults, younger adults would likely take longer to respond to â€œNot Meâ€ attributes that are not associated with self-defining memories. It might also be interesting to include attributes that represent opposites of MDAs in a semantic reco gnition task. For example, if a person responds â€œMeâ€ very quickly to the MDA â€œextrovertedâ€, on e would also expect that person to respond quickly â€œNot Meâ€ to the opposing attribute in troverted. Significant differences in response latencies would be expected for older and younger adults when responding to attributes opposite to MDAs. Such an investigation may be able to provide additional support to the assumption that central aspects of a personâ€™s self-concept are semantically represented in autobiographical memory. This study also offers the first exploration of the relationships among the valence of selfdefining autobiographical memories, several in dividual difference vari ables and psychological well-being. Another logical step would be to include middle-aged adults in a comparable study. Earlier studies (Diehl et al., 2001) have demonstrated a non-lin ear relationship between certain individual differences related to unity of the self-concept and psychological well-being. Specifically, Diehl and his colleagues (2001) found a U-shaped relationship when examining the relationship between self-concep t differentiation and psychological well-being across the adult life span, suggesting that the a ssociation between self-concept differentiation and psychological well-being varies in a systematic age-relate d way. Likewise, Ryff and Keyes (1999) noted differences in the salience for various dimensi ons of psychological well-being across age groups.
111 Although, including middle-aged adu lts in the study may prove enlig htening, longitudinal studies would be necessary to truly unde rstand variations in the self-c oncept across the life span. It might also be informative to explore the various dimensions of psychological well-being individually for all age groups as opposed to examining a composite score of psychological wellbeing as done in this study. Ryffâ€™s (1989) meas ure of psychological wellbeing consists of six separate dimensions; autonomy, environmenta l mastery, purpose in life, personal growth, positive relations with others, and self-acceptance. These dimensions are interconnected in some ways, and many researchers use a composite score of the six individual dimensions as an indicator of psychological well-b eing. However, a number of studies have shown variations between these individual dimensions of psyc hological well-being and other psychological constructs (Campbell et al., 1996; Diehl et al., 2001). In add ition, a few studies have found a relationship between various themes in autobiogra phical memory narratives, including those that incorporate emotional tone, and other measures of psychological well-being (McAdams et al., 1996; Seidlitz & Diener 1997; Woik e et al., 1999). To date, only one other study has examined valence of autobiographical memory and indivi dual dimensions of psychological well-being. Although not the primary focus of this study, Gro ssbaum and Bates (2002) found that the levels of positive emotional valence of the memory narratives were positively correlated with the following individual dimensions of psychologi cal well-being: Self -Acceptance, Positive Relations with Others, Environmental Mastery and Purpose in Life. Ryff and Keyes (1995) assert that the salience of these dimensions of psychological well-being vary across age groups. Therefore, including adults of all ages in an examination of emotional valence of autobiographical memories and individual dimensions of psyc hological well-being may provide additional insight
112 In this research, we focused on the positive valence of autobiographical memory as a mediator. Another direction future studies might take would be to explore negative affect, in particular to examine the relationship of vale nce of self-defining me mories as a protective resource for maintaining psychological well-being in the face of negative affect. If individuals can maintain an overall positive view of what they have learned from life, that is, a more positive valence for their self-defining autobiographical memories, they may be able to retain higher well-being even if they are currently experienci ng negative affect. In th is case, valence would serve as a mediator of well-being, mediating th e influence of negative emotions (neuroticism, negative affect, depression). To examine this issu e, we would look at valence balance (wherein negative valence scores are subtracted from pos itive valence scales), and predict that negative affect, by itself, would be asso ciated with lower well-being, but that valence balance would mediate this relationship. Those individuals with a positive valence balance for their selfdefining memories would be expected to mainta in moderate to high well-being, whereas those with a negative valence balance for their autobiographical memories would be expected to have lower well-being. We already established that in dividual difference variables su ch as self-concept clarity and affect have been shown to have an impact on psychological well-being (Biggler, Neimeyer, & Brown, 2003; Campbell et al., 1996; Diehl et al ., 2001). Constantino, Wilson, and Horowitz (2006) did not find a significant association between self-concept clarity and depressive symptoms. However, evidence of an association between self-concept clarity and participantsâ€™ perception of stress was noted. Speci fically, higher levels of self-c oncept clarity were associated with decreased percep tion of stress levels (Constantino et al., 2006). Upon examining daily measures of affect and self-con cept clarity, Nezlek and Plesko ( 2001) found co-variations in self-
113 concept clarity and affect. Perhaps negative vale nce of autobiographical memories influences self-concept clarity and other i ndividual difference variables that in turn negatively influence psychological well-being. Conclusions The purpose of this research was to explor e the relationship between autobiographical memory and self-concept. Specifically, the sema ntic recognition task employed in this study provided insight into the relati onship between the identity func tion of autobiographical memory and individualsâ€™ self-concept. Re sponse latencies and salience ra tings to MDAs and general selfattributes were used to investigate the relations hip between adultsâ€™ seman tic representations of self and self-defining autobiographical memo ry. Both young and older participants responded more quickly to MDAs than to general self-des criptive self-attributes. Additionally, younger and older adults responded equally fast to MDAs. No previous studies have compared response latencies to MDAs with response latencies to general self-attributes . Through the use of a semantic recognition task, this study documen ted the link between adultsâ€™ self-defining autobiographical memory and self-concept. In addition, salience ratings of MDAs were al so examined. Older adu lts rated their MDAs as significantly more salient than young adults . However, salience ratings for general-self attributes were not significantly different for younger and older adults, as expected. Salience ratings of MDAs have also not been examined heretofore. Therefore, these findings make a unique contribution to the existing literature by illustrating the relationship between adultsâ€™ selfdefining autobiographical memories and semantic representations that constitute a personâ€™s selfconcept. By continuing to examine age-related differe nces in response latencies to MDAs and general self-attributes, as well as salience ratings, we can ach ieve greater understanding of
114 changes related to adultsâ€™ conceptions of self that occur across the life span. Semantic representations of individualsâ€™ self-concept are clearly present in individualsâ€™ self-defining autobiographical memories and a semantic reco gnition paradigm can be used effectively to examine the link between aging a nd the nature of thes e self-representations. The computerized semantic recognition task allowed for more sophisticated measurement of semantic representations of adults â€™ self-concept compared to standard paper-and-pencil measures or cardsort tasks used in past research (Linv ille, 1987, 1992; Showers, 1992.) The methodological paradigm selected for this research â€“ examin ing salience ratings and response latencies to attributes â€“ proved to be a valuable tool for understanding the identity function of autobiographical memo ries in adults. The relationship between positive valence of autobiographical memories, individual difference variables and psychological well-being was also examined. Positive valence of selfdefining autobiographical memory predicted a small portion of the variance in scores on psychological well-being. However, the individual differences variables, Self-Concept Clarity, PANAS-Positive Affect, Neuroticism, and Extrav ersion, accounted for a much greater portion of the variance in well-being. Unexpectedly, posi tive valence did not me diate the relationship between the individual difference variables a nd psychological well-bei ng in this study. The strongest factor predicting wellbeing was self-concept clarity, demonstrating that a personâ€™s ability to conceptualize and maintain a unifi ed self-concept plays an important role in psychological well-being. This study extends earlier findi ngs related to valence of autobiographical memory and psychological well-being in two ways. First, with one exception (Grossbaum & Bates, 2004), researchers investigating these relationships in the past have conceptua lized psychological well-
115 being as a lack of negative symptoms. For exampl e, low levels of depressive symptoms are used to indicate psychological well-being (Costa, Metter, & McCrae, 1994; Keyes et al., 2002; McCrae, 1983). Researchers invest igating the relationship between psychological well-being and autobiographical memory have also conceptua lized subjective well-being as an indicator of psychological well-being, using measures of life satisfaction and self-esteem (Seidlitz & Diener, 1997). Ryff and her colleagues (2002) assert that alt hough subjective well-being and psychological well-being may be conceptually similar, they are empirically distinct . Recall in this study, psychological well-being was conceptualiz ed using Ryffâ€™s (1989) scale that measures eudemonic aspects of psychological well-be ing. These findings expand upon the existing literature by examining optimal overall psycholo gical well-being, as opposed to a lack of negative symptoms Another way in which this study expands upon ear lier research is through the examination of a specific subset of autobiographical memories , those that are self-def ining. In this way, we may draw some more specific conclusions a bout the role of the identity function of autobiographical memory as it relates to psychol ogical well-being. I believe this is an important distinction; it may be that not all types autobi ographical memories play a role in psychological well-being. For example, autobiographical memori es that serve the ge neral-knowledge function may have no impact on psychological well-being. Alternatively, other type s of autobiographical memories may augment psychological well-b eing. For example, of the plethora of autobiographical memories a person may choose to re hearse, share, or review , if he or she selects those self-defining memories associated with positive affect or affirming views of self, this should lead to enhanced psychological well-being.
116 Overall, these findings show promise for futu re investigation of the connection between the identity function of autobi ographical memories and their e ffects on individualsâ€™ self-concept across the life span. The study of the self-concep t in relation to individualsâ€™ autobiographical narratives provides a unique window of insight into the contin ued development of the selfconcept across the adult life span. In particular, a semantic recogn ition paradigm is well suited to examining the link between self-defining autobiographical memories and the semantic representation of individualsâ€™ self-concept. Additionally, exam ining the relationship among the valence of self-defining autobi ographical memories, individu al difference variables, and psychological well-being is a fruitful avenue for exploring the role of in dividualsâ€™ self-concept as a resource in negotiating th e challenges associated with the aging process.
117 APPENDIX A MEMORY QUESTIONAIRE 1. How old were you when this event occurred? ____________________ 2. How important was this event in your life? Not at all A little Somewhat Quite a Bit Extremely 3. How well can you see the events connected with this memory in your mind? Not at all A little Somewhat Quite a Bit Extremely 3. How often do you think about this memory? Rarely Once in a while Sometim es Fairly Often Very Often 4. How clear is your memory about this event? Not at all A little Somewhat Quite a Bit Extremely 5. How often do you tell this memory to someone else? Never Once in a while Sometimes Fairly Often Very Often 6. How vivid is this memory? Not at all A little Somewhat Quite a Bit Extremely 7. How well can you hear the events linked with this memory in your mind? Not at all A little Somewhat Quite a Bit Extremely 8. How strong are the feeli ngs that are connected with this memory? Not at all A little Somewhat Quite a Bit Extremely 9. How would you describe the feelings connected with this memory? Very Mostly Somewhat Some what Mostly Very Positive Positive Positive Negative Negative Negative
118 10. How much impact did this event/experien ce have in terms of who you are today? Not at all A little Somewhat Quite a Bit Extremely What is the main thing you learned about yourself from this event/experience? ______________________________________________________________________________ ______________________________________________________________________________
119 APPENDIX B CODING SCHEME SELF-DEFINING MEMORIES The purpose of this section of the manual is to specify the rules and procedures associated with extracting the self-defining attributes from participantsâ€™ me mory narratives and creating the memory-derived self-attribute list. First, key terms associated with memory coding will be defined. Next, you will find, instructions for coding the memory narratives in order to extract self-attributes. Finally, the procedures for prep aring the self-attribute list for the semantic recognition task will be explained. Key Terms Associated with Coding Self-Defining Memories. Self-defining memories are those memories that contain essential information about the self. Self-defining memories are distinct from other memories in that they are very vivid, emotionally charged, frequently recalled, and often revisited to aid selfunderstanding. Self-defining memories often occur as the result of an event in a personâ€™s life that caused a noticeable change or resulted in an extraordinary insight. Memories of this type become semantically represented as part of a personâ€™s self -concept; that is, self-defining memories become the â€œbuilding blocksâ€ of a personâ€™s identity and self-definition and are the integral parts of that personâ€™s self-concept. It is the semantic representations associated with self-defining memories that are the focal point of this study. Self-Attributes. Self-attributes are adjectives or characteristics (hone st, cranky, joyful, careless) that describe who a pe rson is. Self-attributes are generated when we ask an individual the question â€œWho are you?â€. In this project, we are looking for self-attributes embedded within self-defining memory narratives. For the purpose of this coding task, we distinguish between two major categories of self-attributes: explicit self-attributes vs. implicit self-attributes.
120 Explicit Self-Attributes. Explicit self-attributes are usua lly expressed in the narrative as self-descriptive adjectives (self-attributes ). For example, â€œI was really rebellious in high schoolâ€. In this case, the participant is providing a clear and obvious self -descriptive attribute. You will find additional examples in the section on explicit coding on page 4. Implicit Self-Attributes . In contrast to explicit self-attributes, implicit self-attributes are usually not expressed in a dire ct and obvious way. Rather, they are somewhere â€œbetween the linesâ€ and require that you understand the meaning of the narrative. To say it differently, implicit self-attributes are embedded in the behavioral desc ription and it is the job of the coder to uncover them. Example : â€œ Okay, I was making this map of Florida for school. And, you know, I really wasnâ€™t that into it I guess. I put it together, a nd I thought it looked okay. I didnâ€™t really think that much about it. There was this big blob of glue on it and some fuzz or something got stuck to it. I wasnâ€™t really paying attention and it didnâ€™ t seem like that big of a deal to me..â€ In this example, the participant provides a be havioral description that equates with being careless or perhaps messy. Therefore, there would be two implicit self-attr ibutes that should be extracted from this narrative. You will find additional examples in th e implicit coding section on page 6. Coding the Memory Narratives The content of each memory narra tive will be examined and code d in order to create a list of self-attributes for later use in the semantic recognition task. Transcripts will often contain a clearly stated, single-word self -descriptive attribute (explicit self-attribute). In addition, many narratives will also contain behavioral exemplars of self-attributes (implicit self-attributes). Most memory narratives will contain se veral self-attributes, explicit as well as implicit. Therefore, thorough coding will require the coders to carefu lly examine the narratives several times,
121 concentrating first on explicit self-attributes a nd then on implicit self-attributes. The following section specifies the procedures for coding the memories along with examples of both explicit and implicit self-attributes. Pleas e concentrate on coding only one memory and one type of selfattribute at a time. Explicit Self-Attribute Coding. The first time you examine a protocol, focus only on coding explic it self-attributes . Please follow these steps in the order they are presented. Step 1 : Write the participantâ€™s ID number in th e appropriate blank on the coding sheet. Step 2 : Read the entire memory and pay attention to possible explicit self-attributes. Step 3 : Read the memory again, but now underline any explicit se lf-attributes the participant mentions in his/her narrative. Step 4 : Print each explicit self-att ribute on your coding sheet. Step 5 : Write the total number of explicit self-attributes in the appropriate line on your coding sheet. Examples of Explicit Self-Attributes Example 1Explicit. In the following passage , the participant derived multiple messages about himself or herself from a single episode. Also note, that participants often narrate an entire episode ending with what they learned about themselves, or in what way the event was meaningful with regard to their sense of self. Therefore, it is important that you need to read the entire memory narrative. â€œI needed help moving the tricycle back up the onto the concrete slab and the same time that happened my mother came out the front door and saw me and chided me for not doing
122 it myself. I guess what comes to mind is, I wasnâ€™t competent . I felt, well I felt like Iâ€™m not very smart .â€ Implicit Self-A ttribute Coding After the transcripts have been coded for exp licit self-attributes, coders should focus on extracting implicit self-attributes, those based on behavioral descriptions embedded in the narratives. Often times a participant will provide a phrase or paragraph detailing his/her behavior as an example of a time that he or she displayed a certain self-attribute. Likewise, there are times when a given sentence or phrase does not contai n information about a particular self-attribute, however, when the entire section is read, a particular self-attribute becomes apparent. Therefore, when reading the transcript it is important to ask yourself what pe rsonal characteristics or selfattributes the text conveys. Many times participants will actually offer multip le self-descriptive attributes in a single passage. When this happens, it is important fo r you, the coder, to captu re all of the traits conveyed. There may be behavioral exemplars or implicit self-attributes in the same narrative passage that contains an explicit se lf-attribute. Therefore, it is im portant not to disregard sections of the narrative in which an explicit self-attri bute has already been identified. Each section should be read carefully focusi ng specifically on implicit self-att ributes during the second phase of coding. In order to create the list of implicit self-att ributes, it is necessary to come up with one word that captures the essence of each self-attribu te the participant is trying to convey. Although coding requires you to ascertain the meaning behind the narratives, at the same time is important to avoid over interpreting. Step 1 : Re-read the entire memory.
123 Step 2 : Scan each section carefully and highlight with a yellow highlighter pen any phrases containing behavioral descriptions of implicit self-attributes. Step 3 : Write down the self-attribute reflect ed in the passage on your coding sheet. Step 4: In the event that you encounter a behavioral description that cannot be reduced to a single self-descriptive ad jective that adequately captures the participantâ€™s intention, you should try to summarize the meaning of the passage in a few words or a short sentence. Then highlight the passage with a green highlighter pen and circle the passage. In the â€œNotesâ€ section of your coding sheet, write â€œNeeds Additional Attentionâ€ along with the page number(s). When youâ€™ve completed coding the narrative, DO NOT FILE IT. Instead, place the narratives including the one that needed additional attention, with the coding sheet in the basket marked â€œAdditional Coding Neededâ€. Step 5 : When you have completed the coding, write the total number of implicit selfattributes in the appropriate line on your coding sheet. DO NOT write in a total number if you have behavioral descriptions withou t self-attributes (as in Step 4). Step 6 : Clip the narratives (not the coding sheet) together and place in the â€œnarrativesâ€ file. Examples of Narratives Cont aining Implicit Self-Attributes . Example 1Implicit Self-Attribute. In this example, the part icipant does not explicitly mention a single attribute. However, he or she describes the attribute implicitly. â€œ. The summer before my senior year in hi gh school, I got into this scholars program. There were people from all over Florida. The program was supposed to give you the experience of college. It was an introduction for college life. We had courses. But we also had all these activities. Things that I would never have done before. So, there were all these opportunities.
124 Like rock climbing, swing dancing, horse riding. Things Iâ€™d never done, stuff I just wouldnâ€™t really have tried before. I made up my mind that I was going to try things. In that program, I just felt like I should try things, do thi ngs. And I did. I tried all these new things. And it was great. I really learned that there are so many things out there to do, to tr y. It was just the best So, it got easier and easier to tr y different activities. So now Iâ€™ll try things, you know, Iâ€™ll just try them and see if I like it or not ..â€ In this example, the participant describes le arning to be more adventurous and more selfconfident. Thus, â€œadventurousâ€ and â€œself-confidentâ€ would be the two implicit self-attributes that should be extracted from this narrative. Example 2_ Implicit Self-Attribute . In the following example, again the participant does not specifically mention an attribute. Notice how the participant sets the stage and provides background information in the beginning of the narrativ e. He or she then culminates with the part of the event that had an impact with regard to his/ her identity. This narrative nicely illustrates the importance of reading the entire memory in order to understand the meaning the participant derived from the event. â€œI was a junior in college, and I was supposed to start applying for Law school. I really didnâ€™t want to go to Law school. I REALLY didnâ€™t want to. Iâ€™d just al ways knew I was going to be an attorney, like my dad. It was just assume d. Anyway, things were coming to a head. I knew one day, in the back of mind, I knew, I was going to have to tell my parents I wasnâ€™t going. So, finally I told my Mom and she told my Dad and we got into this big fight. He was saying he knew all along Iâ€™d never finish college. He said he wasnâ€™t going to pay for my apartment or my car insurance. He was done. At first, I was de vastated, well and angry. Then, I thought I donâ€™t care what he says, Iâ€™m going to finish just to show him. Heâ€™s wrong. I WILL finish. After that,
125 they did cut me off for the summer. But after th at, I knew I could put my mind to something and do it. If itâ€™s something I want, Iâ€™ll keep trying.â€ In this example, the participant shares a memory that lead him or her to think of himself/herself as â€œpersistentâ€, â€œdeterminedâ€ and â€œcapableâ€. Other self-attributes may also come to mind, like â€œindependentâ€. However, when the entire narrative is read (ins tead of the fragment I given here), you would realize that other se lf-attributes, like â€œinde pendentâ€ would not be included on the self-attribute list. Creating The Complete Self-Attribute List For The Semantic Recognition Task After you have extracted and made a list of th e self-attributes expre ssed in the narratives, you will create the final complete list of self-att ributes to be used in the semantic recognition task. Each participantâ€™s memory-derived self-attri butes will be added to a list of 40 general selfattributes. Therefore, each participant will have his/her own unique list of self-attributes. The following section contains definiti ons of key terms and details the procedure for creating the selfattribute list for the se mantic recognition task. Key Terms Associated with th e Semantic Recognition Task General Self-Attributes . The general self-attributes are 40 self-descriptive adjectives chosen to represent a comprehensive list of general self-attributes. That is, these are characteristics that every person may, more or less, identify with. All participants will be presented with this list of 40 self-attributes. Memory-Derived Self-Attributes . Memory derived self-attribu tes are the self-attributes extracted from the self-defining autobiographical memories. Each participantâ€™s unique memoryderived self-attributes will be incorporated into the list of general self-attributes. Semantic Recognition Task. The semantic recognition task is designed to examine participantsâ€™ reaction time to general self-attri butes and memory-derived self-attributes. Using
126 Visual Basic software, participants will be presen ted with their own unique list of self-attributes containing the list of 40 general self-attributes along with their own memory-derived selfattributes. Participants will be asked to ma ke a judgment regarding whether the presented attribute is self-descriptive or not by pressing a designated â€œMeâ€ or â€œNot Meâ€ button as quickly as possible. Using the Self-Attributes Synonym Dictionary Because we cannot use any of the self-attributes twice, each memory-derived selfattribute must be checked agains t the master list of 40 genera l self-attributes. Any memoryderived self-attribute that is iden tical to any of the general self-a ttributes should be replaced with a synonym taken from the Self-Attributes Synonym Dictionary . Likewise, if a participant mentions the same self-attribute more than once in his/her memory narrative, you should replace the self-attribute with a synonym from the Self-Attr ibute Synonym Dictionary. Example 1 . A memory narrative contai ning the self-attribute â€œ adventurous â€. Adventurous is on the master list of ge neral self-attributes. Theref ore, â€œadventurousâ€ cannot be used in the final self-attribute list. Upon examining the Self-Attribute Synonym dictionary, you note next to the word â€œadventurous â€ is the synonym â€œ daring â€. You add the word â€œ daring â€ to the final self-attribute list in place of the word â€œadventurousâ€. Example 2 . A memory narrative containi ng the self-attribute â€œ considerate â€. â€œ Considerate â€ is on the master list of general self-attributes. Therefore, â€œ considerate â€ cannot be used in the final self-attribute list. Upon examining the Self-Attri butes Synonym dictionary, yo u note next to the word â€œconsiderate â€ is the synonym â€œ thoughtful â€. You add the word â€œ thoughtful â€ to the final selfattribute list in place of the word â€œ considerateâ€ .
127 Preparing the List for Use in Visual Basic After you have completed the memory-derived self-attribute list, the list must be formatted according to the criteria set forth by the Visual Basic program. Only the memoryderived attributes are included on th is list. Visual Basic will not read the attributes from the list unless the list is in a specific format. Therefore, please perform this task with care. The following section provides specific instructions with re gard to formatting. Step 1: On the computer, open the program called â€œNotepadâ€. To access Notepad, click on the â€œStartâ€ button, and then go to â€œAccessories â€. You find â€œNotepad â€ under â€œAccessories â€. Step 2 : Type the participantâ€™s id number in the top left ha nd corner (Id 1507). Step 3 : Move the cursor down two lines before beginning the list of attributes. Step 4 : Type one attribute per line in all capital letters (e.g., HONEST). The font must be in â€œTimes-Romanâ€ 12-point. Step 5 : After you have typed in all the memory-derived self-attributes, save th e file under the file called Notepad Rt. This file can be found in the â€œAMSR Studyâ€ under â€œMy Documentsâ€. The file name should be the same as the participantâ€™s ID number. Step 6 : Place the coding sheet in the file mark ed â€œNarratives-Coding Completedâ€ and place in the file cabinet. Please file in numerical order. Example_ Visual Basic Formatting . Id 1507 FORGETFUL BASHFUL RESERVED SENSIBLE
128 LIST OF REFERENCES Adkins, G., Martin, P., & Poon, L. W. (1996). Pers onality traits and stat es as predictors of subjective well being in centenarians, octogena rians, and sexagenarians. Psychology and Aging, 11 , 408-416. Bauer, J., & McAdams, D.P. (2004). Personal gr owth in adults storie s of life transitions. Journal of Personality, 72 , 573-604. Bauer, J., McAdams, D.P., & Sakaeda, A. (2005). Interpreting the good life: Growth memories in the lives of mature, happy people. Journal of Personality and Social Psychology, 88 , 203-217. Baumeister, R. R. (1998). The self. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds), The handbook of social psychology ( 4th ed., pp. 680-740). New York: McGraw-Hill. Blaney, P.H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229-246. Bluck, S., & Alea, N. (2002). Exploring the f unctions of autobiograp hical memory: Why do I remember the autumn? In J.D. Webster & B.K. Haight (Eds.), Critical advances in reminiscence work: From theory to application (pp. 61-75). New Yor k, NY, US: Springer Publishing Co. Bluck, S., & Glck , J. (2004) Making things better and learning a lesson: Experiencing wisdom across the lifespan. Journal of Personality, 72 , 543-572. Bluck, S., & Habermas,T. (2001). The life story schema. Motivation and Emotion, 24 , 121-147. Bluck, S., Habermas, T., & Rubin, D.C. (2001) . Extending the study of autobiographical memory: Thinking back about life across the life span. Review of General Psychology . Bluck, S., Levine, L., & Laulhere, T. (1999) Autobiographical remembering and hypermnemia: A comparison of older and younger adults. Psychology and Aging, 14, 671-682. Borenstein, M.H., & Ludemann, P. M. (1989). Habituation at home . Infant Behavior and Development, 12, 525-529. Borrini , G., Dall'Ora, P., & della Sala, S. (1989). Autobiographical memory: Sensitivity to age and education of a st andardized enquiry. Psychological Medicine, 19 , 215-224. Brandstdter, J., & Greve, W. (1994). The agi ng self: Stabilizing and protective processes. Developmental Review , 14 , 52-80. Bransford, J.D., Goldman, S. R., & Vye, N. J ( 1991).Making a difference in peopleâ€™s abilities to think: Reflections on a decade of work and some hope for the future . In Lynn Okagaki, Robert Sternberg (Eds.), Directors of development: Infl uences on the development of children's thinking , (pp. 147-180). Hillsdale, NJ : Lawr ence Erlbaum Associates, Inc,.
129 Brewer, W.F. (1986). What is autobiogr aphical memory? In D.C. Rubin (Ed.), Autobiographical memory , (pp.25-49). Cambridge: Cambridge University Press. Campbell, J., Assanand, S., & Di Paula , A. (2003). The structure of the self-concept and its relation to psychological adjustment. Journal of Personality, 71 , 115-140. Campbell, J.D., Trapnell, P.D., Heine, S.J., Ka tz, I.M., Lavallee, L.F., & Lehman, D.R. (1996). Self-concept clarity: Measurement, persona lity correlates, a nd cultural boundaries. Journal of Personality and Social Psychology, 70 , 141-156. Case, R. (1992). The mindâ€™s staircase . Hillsdale, NJ: Erlbaum. Cohen, G. (1998). The effects of aging on auto biographical memory. In C.P. Thompson, D. J. Herrmann, D. Bruce, J.D. Read, D.G. Payne, & M.P. Toglia (Eds.), Autobiographical memory: Theoretical and applied perspectives (pp. 105-123). Mahwah, NJ: Erlbaum. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Cohen, G., Conway, M.A., & Maylor, E.A.( 1994). Flashbulb memories in older adults. Psychology and Aging ,9 , 454-463. Cohen, G. & Faulkner, D., (1988). The effects of aging on perceived and generated memories. In L. Poon, D.C. Rubin, and B. Wilson (Eds.) Everyday cognition in adulthood and late life ( 222-243). New York: Cambridge University Press. Cohen, G., & Java , R. (1995). Memory for medical history: Accuracy of recall. Applied Cognitive Psychology, 9, 273-288. Constantino, M. J., Wilson, K. R., & Horowitz, L. M. (2006). The direct and stress-buffering effects of self-organization on psychological adjustment. Journal of Social & Clinical Psychology, 25, 333-360. Conway, M. A., & Pleydell-Pearce, C.W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261-288. Costa, P. T., Jr., & McCrae, R. R. (1992). NEO PI-R: Revised NEO Personality Inventory and NEO Five Factor Inventory (NEO-FFI). Odessa, FL: Psychological Assessment Resources Costa, P. T., Metter , E. J., & McCrae, R. R. (1994). Personality stability and its contribution to successful aging. Journal of Geriatric Psychiatry, 27 , 41-59. Cross, S., & Markus, H. (1991). Po ssible selves acro ss the life span. Human Development, 34 , 230-255. Damon, W., & Hart, D. (1988). Self-understanding in childhood and adolescence . New York: Cambridge University Press.
130 DeLoache, J.S., & Brown, A.L. (1983). Very young childrenâ€™s memory for the location of objects in a large-scale environment. Child Development , 54 , 888-897 Diehl, M., Hastings, C. T., & Stanton, J. M. ( 2001). Self-concept differentiation across the adult life span. Psychology and Aging, 16 , 643-654. Diener, E., Emmons, R.A., & Larson, R.J., (1985). Journal of Personality Assessment, 49 , 7175. Donahue, E. M., Robins, R. W., Roberts, B.W., & John, O.P. (1993). The divided self: Concurrent and longitudinal eff ects of psychological adjustment and social roles on selfconcept differentiation. Journal of Personality and Social Psychology, 64 , 834-846. Ekstrom, R.B., French, J.W., Ha rman, H.H., & Dermen, D. (1976). Manual for kit of factorreferenced cognitive tests . Princeton, NJ: Educational Testing Service. Epstein, S. (1973). The self-concept revisited. American Psychologist, 28 , 404-416. Fekken, G. C., & Holden, R.R., (1992). Response latency evidence for viewing personality traits as schema indicators. Journal of Research in Personality, 26 , 103-120. Fischer, K.W. (1980). A theory of cognitive development: The contro l and construction of hierarchies of skills. Psychological Review, 87 , 477-531. Fitzgerald, J.M., & Lawerence R., (1984) Aut obiographical memories across the life span. Journal of Gerontology , 39 , 692-699. Fivush, R., & Fromoff, F.A. (1988) . Style and structure in mother -child conversations about the past. Discourse Processes , 11 , 337-355. Freidman, W. J., (1993). Memory for the time of past events. Psychological Bulletin, 113 , 44-66. Greenough, W.T., Black, J.E., & Wallace, C.S., (1987). Experience and brain development . Child Development, 58, 539-559. Greve, W., & Wentura, D. (2003). Immunizi ng the self: Self-concept stabilization through realityâ€“adaptive self-definitions. Personality and Social Psychology Bulletin , 29 , 39-50. Grossbaum, M., & Bates, G. (2002). Correlates of Psychological well-being at midlife: The role of generativity, agency and communion in narratives themes. International Journal of Behavioral Development, 26, 120-127. Habermas, T., & Paha, C. (2001). The development of coherence in adolescent's life narratives. Narrative Inquiry, 11 , 35-54. Haden, C.A., Haine, R.A., & Fivush, R. (1997). Developing narrative structure in parent-child reminiscing across the pre-school years. Developmental Psychology , 33 , 295-307
131 Harris, P. R., & Lightsey, O. R. (2005). Constructive Thinking as a Mediator of the Relationship Between Extraversion, Neurotic ism and Subjective Well-Being. European Journal of Personality, 19, 409-426. Harter, S. (1983). Developmental perspectives on the self-system. In P.H. Mussen (series Ed.) & E.M. Hetherington (vol. Ed.), Handbook of child psychology : Vol. 4 Social and Personality development (4th ed., pp. 275-385). New York: Wiley. Harter, S. (1998). The development of self-repres entations. In W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol 3. Social, emotional, and personality development (5th ed., pp. 553-617). New York: Wiley. Harter, S. (1999). The construction of self: A developmental perspective . New York: Guilford. Harter, S., & Monsour, A. (1992) . Developmental analysis of conflict caused by opposing attributes in the adol escent self-portrait. Developmental Psychology, 28 , 251-260. Hertzog, C., Raskind, C. L., & Cannon, C. J. (1986). Age-related slowing in semantic information processing speed: An individual differences analysis. Journal of Gerontology, 41 , 500-502. Higgins, E. T. (1996). The "self-digestâ€: Self -knowledge serving self -regulatory functions. Journal of Personality and Social Psychology, 71 , 1062-1083. Howe, M.. L., & Courage, M. L., (1993). On re solving the enigma of infantile amnesia. Psychological Bulletin, 113 , 305-326. Howe, M.L., & Courage, M.L. (1997). The emergence and early development of autobiographical memory. Psychological Review, 1 04 , 499-523 Howe, M.L., Courage, M.L., & Edison, S.C., ( 2003). When autobiographical memory begins. Developmental Review, 23 , 471-494. Janssen , S., Chessa, A. G., Murre, & Jaap M. J. (2005). The reminiscence bump in autobiographical memory: Effects of age, gender, education, and culture. Memory, 13 , 658-668. Johnson, O., & Srivastava, S. (1999) The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. Pervin & O. John (Eds .) Handbook of personality: Theory and research ( 2nd ed, pp.102-138). New York, NY, US: Guilford Press. Josephson, B.R., Singer, J.A., & Salovey, P. (1996) . Mood regulation and memory: Repairing sad moods with happy memori es. Cognition & Emotion, 10 , 437-444. Kausler , D. (1994). Learning and memory in normal aging. San Diego: Academic Press, Inc. Kennedy , Q., Mather, M., & Carstensen , L. (2004) The Role of Motivation in the Age-Related Positivity Effect in Autobiographical Memory. Psychological Science, 15 , 208-214.
132 Keyes , C., Shmotkin , D., & Ryff, C. (2002). Optimizing well-being: The empirical encounter of two traditions. Journal of Personality and Social Psychology, 82 , 1007-1022. King , L. A., Scollon , C. K., & Ramsey, C. (2000). Stories of life transition: Subjective wellbeing and ego development in parents of children with Down Syndrome. Journal of Research in Personality, 34 , 509-536 Kling, K.C., Ryff, C.D., & Essex, M.J. (1997). Ad aptive changes in the se lf-concept during a life transition. Personality and Social Psychology Bulletin , 23, 981-990. Laver, G.D., & Burke, D.M. (1993). Why do sema ntic priming effects increase in old age? A meta-analysis. Psychology and Aging, 8 , 34-43. Legerstee, M., Anderson, D., & Schaffer, A. ( 1998). Five and eight mont h-old infants recognize their faces and voices as familiar and social stimuli. Child Development , 69, 37-50. Lewis, M., & Brooks-Gunn, J. (1979). Social cognition and acqui sition of the self . New York: Plenum Press. Linton, M., (1987). Ways of search ing and the content of memory. Memory and Cognition. 11 , 114-120. Linville, P. W. (1987). Self-complexity as a cognitive buffer against stress-related illness and depression. Journal of Personality and Social Psychology, 52 , 663-676. Little, B.R., (1999). Free traits, personal projects and vide o-tapes: Three tiers for personality psychology. Psychological Inquiry, 7 , 1996. Markus, H. (1977). Self-schemata and pro cessing information about the self. Journal of Personality and Social Psychology , 35 , 63-78. Markus, H. R., & Herzog, R. A. (1991). The role of self-concept in agin g. In K. W. Schaie (Ed.), Annual review of ger ontology and geriatrics (Vol. 11, pp. 110-143). New York: Springer. Markus, H., & Wurf, E. (1987). The dynamic self -concept: A social psychological perspective. Annual Review of Psychology, 38 , 299-337. McAdams, D.p., (1993). Accentuating methodologies: When to find the personality differences that already exist. Psychological Inquiry, 4 , 297-300. McAdams, D. (1996). Narrating the self in a dulthood. In J. E. Birren & G. Kenyon (Eds.), Aging and biography: Explorations in adult development (pp. 131-148). New York: Springer. McAdams, D.P., Anyidoho, N.A., Brown, C., Hu ang, Y.T., Kaplan, M., & Machado, M.A.. (2004). Traits and stories: Links between dispositional and narr ative features of personality. Journal of Personality , 72 , 762-784.
133 McAdams, D.P., & Bowman, P.J. (2001).Narrating life's turning points: Redemption and contamination. In D.P. McAdams, R. Josselson, & A. Lieblich (Eds.) Turns in the road: Narrative studies of lives in transition (pp.3-34). Washington, DC, US: American Psychological Association. McAdams, D.P., Hoffman, B.J ., & Mansfield, E.D. (1996).Themes of agency and communion in significant autobiographical scenes. Journal of Personality, 64 , 339-377. McAdams, D.P., Reynolds, J., Lewis, M., Patton, A., & Bowman, P. (2001). When bad things turn good and good things turn bad: Sequences of redemption and contamination in life narrative and their relati on to psychosocial adaptation in midlife adults and in students. P ersonality and Social Psychology Bulletin, 27 , 474-485. Meltzoff, A. N. (1995). What infant memory tell s us about infantile amnesia: Long-term recall and deferred imitation. Journal of Experime ntal Child Psychology , 59 , 497-515. Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality: Reconceptualizing situations , dispositions, dynamics, and invariance in personality structure. Psychological Review, 102 , 246-268. Moffitt, K.H., & Singer, J.A. (1994). Continuity in the life story: Self-defining memories, affect, and approach/avoidance personal strivings. Journal of Personality, 62 , 21-43. Mroczek, D. K. (2001). Age and emotion in adulthood. Current Directions in Psychological Science, 10 , 87-90. Mueller, J. H., & Johnson, W. C., (1990). Trait distinctiv eness and age specificity in selfreferent information processing. Bulletin of the Psychonomic Society, 28 , 1990. 119-122. Neisser, U. (1988). What is Ordinary Memory the memory of? In U.Neisser & E. Winograd (eds.), Remembering reconsidered: ecological an d traditional approaches to the study of memory (pp.356-373). New York: Cambridge University Press. Nelson, K., (1993) The psychological and soci al origins of autobiographical memory. American Psychological Society, 4, 7-14. Nelson, K. & Fivush, R., (2004). The emergence of autobiographical memory: A sociocultural developmental theory. Psychological Review, 111 , 486-511. Nezlek, J. B., & Plesko, R. M. (2001). Day-to-day relationships am ong self-concept clarity, selfesteem, daily events, and mood. Personality and Social Psychology Bulletin, 27 , 201-211. Palmore, E.B., (1998). The Facts on Aging Quiz (2nd ed.). New York: Springer Publishing Company. Pasupathi, M. (2001). The social construction of the personal past and its implications for adult development. Psychological Bulletin, 127, 651-672 .
134 Perfect , T. J., & Rabbitt , P. M. (1993). Speed and accuracy of memory decisions in older adults. Psychological Reports, 73 , 607-61 Pillemer, D. (1998). Momentous event, vivid memories: Ho w unforgettable moments help us understand the meaning of our lives. Cambridge, MA: Harvard University Press. Rabbitt, P., & Winthorpe, C. (1988) . What do old people remember? Cognitive Psychology , 8 , 578-595 Radloff, L.S. (1977). The CES-D scale: A self-repor t depression scale for research in the general population. Applied Psychological Measurement, 1 , 385-401. Rey, A. (1941). Lâ€™examen psychologique dans les cas dâ€™encephalopathie tramatique. Archives de Psychologie, 28 , 21. Riby, L., Perfect, T. J., & Stollery, B. T. (2004). Evidence for disproportiona te dual-task costs in older adults for episodic but not semantic memory. Quarterly Journal of Experimental Psychology A: Human Ex perimental Psychology , 57 , 241-267. Rovee-Collier, C. (1997). Dissociations in in fant memory: Rethinking the development of implicit and explicit memory. Psychological Review , 104 , 467-498. Rovee-Collier,C.k., & Hayne, H. (2000). Memory in infancy and childhood. In E. Tulving and C. Fergus (Eds.) The Oxford handbook of memory ( pp. 267-282). New York: Oxford University Press Rovee-Collier, C., K., & Shyi, C.W.G., (1992). A functional and cognitive analysis of infant long-term retenetion. In M.L. Howe, C. J. Brainerd, & V.F. Reyna (Eds.) Development of long-term retention (pp. 456-494). New York: Springer-Verlag. Ruble, D. N., & Dweck, C. (1995). Self-concepti ons, person conception, and their development. In N. Eisenberg (Ed.) Review of personality and social psychology: The interface (Vol. 15, pp. 109-139). Thousand Oaks, CA: Sage. Rubin D.C, & Berntsen, D. (2003). Life scripts help to maintain autobiographical memories of highly positive, but not highly negative, events. Memory & Cognition, 31 , 1-14. Rubin, D., & Siegler, I. (2004). Facets of Personality and the Phenomenology of Autobiographical Memory. Applied Cognitive Psychology, 18 , 913-930. Ryff, C.D., (1989). Happiness is everything or is it?: Explorations on the meaning of psychological well-being. Journal of Personality and Social Psychology, 57 , 1069-1081. Ryff, C. D. (1991). Possible selves in adu lthood and old age: A tale of shifting horizons. Psychology and Aging, 6, 286-295. Ryff, C. D. & Keyes, C.L. M. (1995). T he structure of psychologi cal well being revisited. Journal of Personality and Social Psychology, 69, 719-727.
135 Sanitioso, R., Fong, G.T., & Kunda, Z. (1990). Motivated recruitment of autobiographical memories. Journal of Personality and Social Psychology, 59 , 229-241. Seidlitz, L., Wyer, R., & Diener, E. (1997).Cognitive correlates of subjective well-being: The processing of valenced life events by happy and unhappy persons. Journal of Research in Personality, 31 , 240-256. Sheldon, K.M., Ryan, R.M., Rawsthorne, L.J., & Lair d, J. (1997). Trait self and true self: Crossrole variation in the big-five personality tr aits and its relations with psychological authenticity and subjective well-being. Journal of Personality and Social Psychology, 73 , 1380-1393. Siem, F., (1992). The use of response latencies to enhance self-report personality measures. Military Psychology, 8, 15-22. Singer, J.A. (2004). Narrative identity and meaning making across the adult lifespan: An introduction. Journal of Personality, 72, 437-459. Singer, J.A. (1995). Seeing oneself: Locating narr ative memory in a framework of personality. Journal of Personality, 63 , 429-457 Singer, J.A., & Salovey, P. (1993). The remembered self . New York: Free Press. Singer, J.A., & Salovey, P. (1996). Motivated memo ry: Self-defining memori es, goals, and affect regulation. In M. Leonard & A. Tesser (Eds.), Striving and feeling: Interactions among goals, affect, and self-regulation (pp. 229-250). Hillsdale, NJ: Erlbaum. Sobel, M.E. (1988) Direct and indirect effect in linear structural equati on models. In J.S. Long (ed.), Common problems/ Proper solutions: Avoiding error in quantitative research (pp. 46-64) Beverly Hills, CA: Sage. Tulving, E. (1986). How many memory systems are there? American Psychologist, 40 , 385-398. Vygotsky, L.S. (1978). Mind in society: The development of higher psychological processes . Cambridge University Press. Watson, D., Clark, L.A., & Tellegen, A. (1988). De velopment and validation of brief measures of positive and negative affect: The PANAS scale. Journal of Personality and Social Psychology, 54 , 1063-1070. Webster, J.D. (1995). Adult age differences in reminiscence functions. In B.K. Haight & J.D. Webster (Eds.), The art and science of reminisci ng: theory, research, methods, and applications (pp.89-102). Washington, DC : Taylor & Francis. Webster, J., & Cappeliez, P. (1993). Re miniscence and autobiographical memory: Complementary contexts for cognitive aging research. Developmental Review , 13 , 54-91.
136 Welch-Ross, M.K. (1997). Mother-child part icipation in conversation about the past: Relationships to preschoolersâ€™ theory of mind . Developmental Psychology , 33 , 618-629. Wilson, A.E . & Ross, M., (2003). Autobiographical memory and conceptions of self: Getting better all the time. Directions in Psychological Science , 12 , 66-69. Woike, B., Gersekovich, I., Piorkowski, R., & Polo. R. (1999). The role of motives in the content and structure of autobiographical memory. Journal of Personality and Social Psychology , 76, 600-612.
137 BIOGRAPHICAL SKETCH Angelenia Semegonâ€™s academic and research interest in aging began as an undergraduate at the University of North Florida. She gr aduated with honors, earni ng a B.A. in psychology, with a minor in behavioral science. She remained at the University of North Florida for Masterâ€™s level graduate studies. After completion of her M.A. in psychology, she began her doctoral career at the University of Fl orida in 1998 where she joined the developmental area in the Department of Psychology. She was also a traine e in the NIA-funded Aging Training Program at the University of Florida and earned a Graduate Certificate in Gerontology. Angelenia completed the requirements her Ph.D in December, 2006. Angelenia Semegon has been funded through a va riety of mechanisms during her graduate studies, including a Pre-doctoral Fellow-National Institute on Hea lth Research Service Award. A Ruth L. Kirschstein National Research Servic e Award from the National Institute on Aging funded Angeleniaâ€™s dissertation project.