<%BANNER%>

The Relation of the Conceptual Self to Autobiographical Memory

Permanent Link: http://ufdc.ufl.edu/UFE0041917/00001

Material Information

Title: The Relation of the Conceptual Self to Autobiographical Memory A Comparison of Young and Middle-Aged Adults' Earliest and Recent Memories
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Demiray, Burcu
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: adult, autobiographical, childhood, development, earliest, mediation, memory, midlife, self
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The current study combines life span developmental theory (e.g., Baltes, 1997) with the self-memory system model of autobiographical memory (SMS model; Conway, & Pleydell-Pearce, 2000) to examine adult age group differences in the conceptual self, and how individuals? current conceptual self guides their recall of recent and early autobiographical memories. The sample consists of 285 young (ages 19-29) and 135 middle-aged (ages 47-64) adults. Participants completed an online survey assessing their current self-attributes (e.g., autonomy) and future time perspective, and wrote narratives of two memories: their earliest childhood memory and a recent memory from the last year. They then rated the content of their memories for self-attribute themes. Findings show that young and middle-aged adults have different current conceptual selves: Although young adults have a more open-ended and positive future time perspective, they have lower levels of autonomy, environmental mastery, positive relations with others, self-acceptance and purpose in life than middle-aged adults. In addition, future time perspective acts as a suppressor in the relation between age and current self-attributes, bolstering the predictive validity of age. The study also tested the SMS model using both a basic and a more stringent test. As expected, findings showed that the link between current self-attributes and memory content is more evident in recent memories than earliest memories. This effect is similar across young and middle-aged adults, indicating developmental stability in the relation between current self-attributes and memory content. The study contributes to the adult development literature, and extends the SMS model by providing data on different age groups, and different types of memories (i.e., earliest and recent).
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Burcu Demiray.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Bluck, Susan B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041917:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041917/00001

Material Information

Title: The Relation of the Conceptual Self to Autobiographical Memory A Comparison of Young and Middle-Aged Adults' Earliest and Recent Memories
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Demiray, Burcu
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: adult, autobiographical, childhood, development, earliest, mediation, memory, midlife, self
Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The current study combines life span developmental theory (e.g., Baltes, 1997) with the self-memory system model of autobiographical memory (SMS model; Conway, & Pleydell-Pearce, 2000) to examine adult age group differences in the conceptual self, and how individuals? current conceptual self guides their recall of recent and early autobiographical memories. The sample consists of 285 young (ages 19-29) and 135 middle-aged (ages 47-64) adults. Participants completed an online survey assessing their current self-attributes (e.g., autonomy) and future time perspective, and wrote narratives of two memories: their earliest childhood memory and a recent memory from the last year. They then rated the content of their memories for self-attribute themes. Findings show that young and middle-aged adults have different current conceptual selves: Although young adults have a more open-ended and positive future time perspective, they have lower levels of autonomy, environmental mastery, positive relations with others, self-acceptance and purpose in life than middle-aged adults. In addition, future time perspective acts as a suppressor in the relation between age and current self-attributes, bolstering the predictive validity of age. The study also tested the SMS model using both a basic and a more stringent test. As expected, findings showed that the link between current self-attributes and memory content is more evident in recent memories than earliest memories. This effect is similar across young and middle-aged adults, indicating developmental stability in the relation between current self-attributes and memory content. The study contributes to the adult development literature, and extends the SMS model by providing data on different age groups, and different types of memories (i.e., earliest and recent).
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Burcu Demiray.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Bluck, Susan B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041917:00001


This item has the following downloads:


Full Text

PAGE 1

THE RELATION OF THE CONCEPTUAL SELF TO AUTOBIOGRAPHICAL MEMORY: A COMPARISON OF YOUNG AND MIDDLE-AGE D ADULTS EARLIEST AND RECENT MEMORIES By BURCU DEMIRAY 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 2010 1

PAGE 2

2010 Burcu Demiray 2

PAGE 3

To my Dad, mer Faruk Demiray, who has inspired me to become an academic 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank my family for their uncond itional love and support. Special thanks to my dear mother, Belgin Demiray and yol arkada m Gkhan Batur: This could not have been done without their endless patience and support. I would like to express my gratitude to my advisor, Dr. Susan Bluck for her intellectual gui dance and superb mentoring, and for giving me all the necessary tools of the trade. I woul d also like to thank my supervisory committee members for their time and constructive feedback, and the Life Story Lab members for their help through the entire re search process. 4

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........7LIST OF FIGURES.........................................................................................................................8ABSTRACT.....................................................................................................................................9CHAPTER 1 INTRODUCTION................................................................................................................. .11Outline of the Paper........................................................................................................... .....12Theoretical Accounts of the On set of Earliest Memories.......................................................13The Self-Memory System: Effect of the Conceptual Self on Autobiographical Memories...15Adult Development and the Conceptual Self: Future Time Perspective and SelfAttributes.............................................................................................................................17Future Time Perspective: Defining Midlife....................................................................17Self-Attributes of Young and Late Middle-Aged Adults................................................19The Role of Age and Future Time Perspective in Self-Memory Relations............................22Testing the SMS Model: Effects of Self-A ttributes on Earliest and Recent Memory Content................................................................................................................................23Specific Aims and Hypotheses...............................................................................................252 METHODS...................................................................................................................... .......29Participants.............................................................................................................................29Procedure................................................................................................................................31Measures.................................................................................................................................33Self-Attributes Measure...................................................................................................33Memory Content Ratings................................................................................................34Future Time Perspective Measures.................................................................................35Background Measures.....................................................................................................37Use of Online Data Collection: Considerations......................................................................373 RESULTS...................................................................................................................... .........40Preliminary Analyses........................................................................................................... ...40Potential Covariates.........................................................................................................40Sample Dependency Issues.............................................................................................42Sample Selectivity: Attr ition Due to Drop-Out...............................................................43Age at Time of Memory..................................................................................................43Major Analyses: Testing Hypotheses.....................................................................................44Hypothesis 1: Relation of Age Grou p to Future Time Perspective.................................44 5

PAGE 6

Hypothesis 2: Relation of Age Gr oup to Current Self-Attributes...................................45Hypothesis 3: Examining Future Time Pe rspective as a Mediator Between Age Group and Self-Attributes............................................................................................46Analytical approach..................................................................................................46Findings....................................................................................................................48Hypothesis 4: Relation of Current Self-Attributes to Memory Content..........................524 DISCUSSION................................................................................................................... ......67Adult Development and the Conceptual Self: Future Time Perspective and SelfAttributes.............................................................................................................................68Relation of Age Group to Future Time Perspective........................................................68Relation of Age Group to Current Self-Attributes..........................................................70The Role of Age and Future Time Perspective in Predicting Current SelfAttributes......................................................................................................................74Summary: Adult Development and the Conceptual Self................................................79Testing the Self-Memory System Model................................................................................80Relation of Current Self-Attri butes to Memory Content.................................................81The Reconstructive Nature of Autobiographical Memory..............................................81Stringent Test of the SMS Model Using Earliest Childhood Memories.........................83Does the SMS Model Explain Why Adults Recall Their Earliest Memories?................86Expanding the SMS model with a Life Span Development Perspective........................88Limitations.................................................................................................................... ..........90Participant Recruitment: Sampling Adults from Late Midlife........................................91Online Data Collection....................................................................................................92Cross-Sectional Design...................................................................................................93Issues Related to Conceptualizing th e Complexity of Adult Development....................94Issues Related to Conceptualizing the Complexity of the SMS Model..........................94Conclusions.............................................................................................................................96APPENDIX A MEMORY NARRATIVE INSTRUCTIONS........................................................................99B MEMORY NARRATIVE EXAMPLES..............................................................................100C SELF-ATTRIBUTES MEASURE (RYFF, 1989A).............................................................101D MEMORY CONTENT RATINGS FOR EARLIEST AND RECENT MEMORY NARRATIVES..................................................................................................................... 103E FUTURE TIME PERSPECTIVE SCALE (CARSTENSEN & LANG, 1996)....................105F DEMOGRAPHIC QUESTIONS..........................................................................................106REFERENCES............................................................................................................................107BIOGRAPHICAL SKETCH.......................................................................................................118 6

PAGE 7

LIST OF TABLES Table Page 3-1 Descriptive statistics for current self-attributes in young and late middle-aged adults.....583-2 Pearson correlation coefficients between th e six self-attributes, the Future Time Perspective Scale, the Rappaport Time Line and age group.............................................593-3 Mediation model 1: Future Time Perspect ive Scale as a single mediator between age group and current self-attribut es (5000 bootstrap samples)...............................................593-4 Mediation model 2: Rappaport Time Line as a single mediator between age group and current self-attributes (5000 bootstrap samples).........................................................603-5 Mediation model 3: Future Time Persp ective Scale and Rappaport Time Line as multiple mediators between age group and current self-attributes (5000 bootstrap samples)....................................................................................................................... ......603-6 Hierarchical multiple regression analysis predicting positive relations content of all memories (N = 791)...........................................................................................................613-7 Hierarchical multiple regression analysis predicting purpose in life content of all memories (N = 800)...........................................................................................................623-8 Hierarchical multiple regression analysis predicting personal growth content of all memories (N = 795)...........................................................................................................633-9 Hierarchical multiple regression anal ysis predicting autonomy content of all memories (N = 799)...........................................................................................................643-10 Hierarchical multiple regression analysis predicting environmental mastery content of all memories (N = 789)..................................................................................................653-11 Hierarchical multiple regression analysis predicting self-acceptance content of all memories (N = 789)...........................................................................................................66 7

PAGE 8

LIST OF FIGURES Figure Page 1-1 The self-memory system with self-attributes and future time perspective shown as aspects of the conceptual self.............................................................................................271-2 The effect of self-attri butes on autobiographical memory content as predicted by age group and future time perspective......................................................................................283-1 Illustration of effects and thei r weights in mediation models............................................584-1 Illustration of suppression effects in relation to predicted results.....................................98 8

PAGE 9

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 THE RELATION OF THE CONCEPTUAL SELF TO AUTOBIOGRAPHICAL MEMORY: A COMPARISON OF YOUNG AND MIDDLE-AGE D ADULTS EARLIEST AND RECENT MEMORIES By Burcu Demiray August 2010 Chair: Dr. Susan Bluck Major: Psychology The current study combines life span developmental theory (e.g., Baltes, 1997) with the self-memory system model of autobiographica l memory (SMS model; Conway, & PleydellPearce, 2000) to examine adult age group diffe rences in the concep tual self, and how individuals current conceptual self guides th eir recall of recent and early autobiographical memories. The sample consists of 285 young (ages 19-29) and 135 middle-aged (ages 47-64) adults. Participants completed an online surv ey assessing their curren t self-attributes (e.g., autonomy) and future time perspective, and wrot e narratives of two me mories: their earliest childhood memory and a recent memory from the last year. They then rated the content of their memories for self-attribute themes. Findings show that young and middle-aged adults have different current conceptual selves: Although young adults have a more open-ended and positive future time perspective, they have lower le vels of autonomy, environmental mastery, positive relations with others, self-acceptance and purpose in life than middle-aged adults. In addition, future time perspective acts as a suppressor in the relation between age and current selfattributes, bolstering the predictive validity of ag e. The study also tested the SMS model using both a basic and a more stringent test. As e xpected, findings showed that the link between 9

PAGE 10

10 current self-attributes and memory content is more evident in recent memories than earliest memories. This effect is similar across young an d middle-aged adults, indicating developmental stability in the relation between current self -attributes and memory content. The study contributes to the adult development literature, and extends the SMS mo del by providing data on different age groups, and different types of memories (i.e., earliest and recent).

PAGE 11

CHAPTER 1 INTRODUCTION The current study is grounded in and te sts the self-memory system model of autobiographical memory (SMS model; Conway & Pleydell-Pearce, 2000) According to this model, individuals current conceptual self gui des their recall of autobiographical memories. Little research has tested the SMS model by ex amining the link between the current conceptual self and retrieved autobiographical memories. Mo reover, no research to date has empirically examined whether the relation of the conceptual self guides recall of distant memories such as earliest childhood memories. The earliest ag e from which an adult can retrieve an autobiographical memory is between 3 and 4 years (Dudycha & Dudycha, 1941; Howes, Siegel, & Brown, 1993). This represents the offset of th e childhood amnesia period (i.e., adults inability to recall personal memories from the first year s of life) and the onset of autobiographical memory (Davis, Gross, & Hayne, 2008). Though explai ning earliest memories is not the primary focus of the SMS model, the model does suggest th at the impact of the conceptual self during retrieval should extend not only to recent but also to earliest memories. Thus, the present study fills a gap in the literature by testing the SM S model using two types of memories: recent memories that allow a basic test of the SMS mo del and are most likely to show an effect, and earliest memories that allow a novel and stringent test of the SMS model. In doing so, the study also contributes to the literat ure by investigating whether the SMS model provides a convincing explanation for why adults are able to recall their earliest memories. Another contribution of the present study is to bring a life span development perspective to the study of autobiographical memory (Blu ck & Habermas, 2000), by examining adult developmental differences or similarities in th e relation between the conceptual self and the content of memories. The study examines age differe nces in the current conceptual self, and its 11

PAGE 12

effects on recall of autobiographical memories in young adults and those in late midlife. There is a scarcity of research on middle-aged individual s autobiographical memories in general, and particularly their earliest me mories (Dixon, De Frias, & Ma itland, 2001). Studies have been conducted almost exclusively with college stude nts. Two studies have examined middle-aged individuals earliest memories (Demiray, Glgz, & Bluck, 2008; Multhaup, Johnson, & Tetirick, 2005). Thus, further research is needed to gain a fuller understanding of autobiographical memory in midlife. In short, one goal of the presen t study is to examin e adult developmental differences in the current conceptual self (i.e., se lf-attributes and future time pe rspective) by comparing adults from two age groups. The second go al is to test the claim of the SMS model that the current conceptual self (i.e., self-attri butes) guides the conten t of retrieved memories, considering both different age groups and di fferent types of memories. This an alysis also sheds light on whether the SMS model provides a convincin g explanation concerning the re trieval of earliest memories. Outline of the Paper The introduction first compares three theoreti cal accounts that provide explanations for the retrieval of earliest memories, highlighting the self-memory system model as the theoretical framework to be adopted in this research. Next the self-memory system is described with a focus on the conceptual self identifying future time perspective and self-attributes (e.g., sense of purpose in life) as aspects of the conceptual self. Following that, th e conceptual self is described in adult life span deve lopmental terms, as being different in young adulthood and late midlife both in regards to an individuals future time pers pective and in regards to certain self-attributes. Next, a further conceptual specification of the SMS model is presented which suggests that age and future time perspective shape self-attributes wh ich in turn shape the content of both earliest and recent memories. These relations between the c onceptual self and memo ry are tested in two 12

PAGE 13

ways: a basic test with recent memories, and a mo re stringent test with earliest memories. The Methods section describes the sample, measures and procedure that was used to conduct the research. Next, statistical analys es and results are reported. Study findings are discussed in terms of contributions made to the adult development literature (e.g., life sp an development theory, Baltes, 1997), and the autobiographical memory literature with a focus on the SMS model (Conway & Pleydell-Pearce, 2000). Finally, the li mitations of the study are provided as well as some overall conclusions. Theoretical Accounts of the Onset of Earliest Memories Researchers have developed three major expl anations for the offset of childhood amnesia (i.e., onset of autobiographical memory). That is, there are three major explanations for why earliest memories are recalled throughout adulthood. The first two accounts emphasize the conditions under which earliest memories are encoded in childhood, whereas the third account (i.e., SMS model) also emphasizes the condition s under which these memories are retrieved in adulthood. According to the first two accounts, memories are not encoded and organized autobiographically during infancy due to deve lopmental limitations. The SMS model (Conway & Pleydell-Pearce, 2000) was not designed specifi cally to account for childhood amnesia, but it emphasizes that the current conceptual self a ffects retrieval of any autobiographical memory. Failure to retrieve memories from infancy is seen as related not only to early encoding processes but to the current goals of the c onceptual self. Each of these a ccounts is described briefly here. The cognitive self account suggests that the em ergence of a cognitive self is a necessary condition for the onset of earliest memories (Howe & Courage, 1993, 1997). At around 2 years of age, a cognitive self begins to emerge that al lows infants to encode experiences in terms of self and organize them autobiographically. Therefor e, adults cannot recall memories of the self in infancy because they were encoded before the self emerged. The social-interaction account 13

PAGE 14

argues that parent-child memory sharing is an important activity in autobiographical memory development (Haden, Haine, & Fivush, 1997; Ne lson, 1993). As children begin to use language and to share their past, they le arn conventional narrative rules a nd forms. Co-constructed parentchild conversations provide a basis for the tran sition from encoding fragmented memories into encoding coherent episodic narratives (Nelson, 1993). This account, thus suggests that adults cannot recall events that were encoded before th ey began to narrativize them as children. In both of these accounts, the conditions under which memories are encoded during early childhood (up to about age three) prohibit their recall at later stages of life. The account based on the SMS model emphasizes conditions during retrieval. The difference between this account and those previously described is that it considers not only the developmental conditions of the child during en coding but also that of the adult doing the remembering. The SMS model suggests that au tobiographical memories, including earliest memories, are constructed during re trieval in line with the current goals and needs of the self (Conway & Pleydell-Pearce, 2000). Accordi ngly, childhood amnesia occurs due to the discrepancy between the content and organization of goals that were active during the encoding of memories in infancy as compared to the goa ls that are active duri ng retrieval in adulthood (Conway & Pleydell-Pearce, 2000; Schachtel, 1947). Memories in infancy tend to be encoded in line with basic needs that may have little overlap with ones goals in adulthood. If there is a lack of correspondence between the self s goals at encoding and retrie val, effective retrieval cannot occur and specific memories wi ll not be recalled (C onway & Pleydell-Pearce, 2000). Thus, earliest memories represent the first encoded events of infancy that fit with ones current selfconceptions. 14

PAGE 15

Most studies of earliest memo ries to date are based on the two accounts that highlight developmental conditions during encoding in infanc y. For example, researchers argue that adults recall their earliest memories in qualitatively diffe rent ways and from different ages because as children they grew up in different narrative envi ronments and developed different self-construals (Leichtman, Wang, & Pillemer, 2003). Although childhood cognitive development and social context may be important, the current resear ch embraces the SMS account because it includes the current adult self that is doing the rememberi ng of the earliest memory. If the SMS model is correct, ones current conceptual self (e.g., self-att ributes) should affect the memories one recalls of recently experienced events but also of earliest events. This study examines that basic tenet of the SMS model, by examining recent memories but also investigating whether earliest memories show a match with ones current co nceptual self. The next section provides a brief description of the SMS model including its portrayal of how th e conceptual self affects remembering not only of earliest memories but of memories more generally. The Self-Memory System: Effect of the Con ceptual Self on Autobiographical Memories According to the SMS model (Conway & Pley dell-Pearce, 2000) the current conceptual self influences retrieval. Aut obiographical memories are patterns of activation across the three components of the Self-Memory System: the epis odic memory system, the working self and the long-term self (see Figure 1-1). Episodic memori es are event specific composites of sensoryperceptual-cognitive-affective deta il that invoke visual imagery a nd the experience of mentally reliving an event (Conway, Singer, & Tagini, 2004) The working self includes control processes that coordinate and modulate c ognition, affect and behavior thr ough prioritizing the individuals complex hierarchy of goals (Conway & Pleyde ll-Pearce, 2000). Finally, the long-term self includes both the conceptual self and the autobi ographical knowledge base. The conceptual self is of greatest interest in terms of the SMS clai m that recalled memories are shaped at retrieval 15

PAGE 16

and in terms of the account of earliest memories. The conceptual self contains information that one knows about ones self such as self-attri butes, beliefs and attit udes. In contrast, the autobiographical knowledge base includes temporal knowledge of experienced events that is remembered. It is composed of the hierarch ically organized life st ory schema (Bluck & Habermas, 2000), lifetime periods (i.e., chapters in the life story such as When I was in graduate school) and general events (Conway et al., 2004). The present study examines th e relation of the conceptual self to autobiographical memories. The conceptual self c ontains abstract knowledge that one knows about ones self. For example, it includes self-attributes, personality characteristics, attit udes, beliefs, personal motives and possible selves. The conceptual self influences the working self by shaping its current goals, and thereby influencing the retrieva l (i.e., construction) of memories. For example, if ones conceptual self includes a negative attitude towards smoki ng, one is more likely to have anti-smoking goals, and is more likely to re call negative memories related to smoking. There is some previous research concerning ho w particular aspects of the conceptual self affect the construction of autobiographical memories. Woike and her colleagues (Woike, 1994a, 1994b; Woike, Lavezzary, & Barsky, 2001; Woik e, McLeod, & Goggin, 2003) have shown that individuals with agentic motives at a personality level also retrieve memories that are thematically more agentic, wher eas individuals with communal motives retrieve more communal memories (Woike & Polo, 2001). In keeping with the SMS model, Wilson and Ross (2003) also argue that there is a bi-directi onal relation between au tobiographical memory and the conceptual self. They have shown how individuals reconstruc t their past according to how they would like to perceive themselves currently: as different from the past self (Conway & Ross, 1984), superior to the past self (Wilson & Ross, 2001) or distant from a negative past self (Ross & Wilson, 2002). 16

PAGE 17

In sum, there is a small body of research, based on the SMS model, which shows that particular aspects of the conceptual self influence the content of autobiographical memories. None of these studies, however, examine the effect of the conc eptual self particularly on earliest memories, and none have taken an adu lt developmental perspective. The current study suggests that adults from diffe rent age groups (e.g., young versus late middle-aged adults) should have different conceptual selves. Furthermore, none of the previous studies examine future time perspective and self-attri butes as unique aspects of the conc eptual self (see Figure 1-1). The current study suggests that both future time pers pective and self-attributes are important aspects of the conceptual self that should show developm ental variation. The next section describes these two aspects of the conceptual se lf in developmental context. Adult Development and the Conceptual Self: Future Time Perspective and Self-Attributes Though not included in the original SMS model, from a life span developmental perspective it seems clear that the conceptual self should be shaped by ones current age. The current study focuses on two aspects of the concep tual self that, based on previous theory and research, are likely to be affected by age: future time perspective and self-attributes. Future Time Perspective: Defining Midlife Across adulthood, individuals have different pers pectives on how much time is left in their life. Future time perspective is especially meaningful in midlife, and is embedded in theoretical descriptions of midlife. For example, midlife is described as governed mostly by an internal social clock that reminds indi viduals that they are in the middle (Neugarten, 1996). Although life span developmental theory accepts midlife as the period between 40 and 60 with flexible boundaries (Baltes, Staudinger, & Lindenberg er, 1999), there is no clearly agreed-upon demarcation of midlife. It is a complex devel opmental stage that is a time of challenge and potential stress due to multiple roles, responsibi lities and potential declines as well as a time of 17

PAGE 18

achievement, maturity, and generativity. This complexity makes midlife difficult to assess in terms of chronological age, partic ularly as age norms are quite flexible in midlife (Lachman & Bertrand, 2001): adults of the same chronological age are in different lif e stages in terms of career, family and social roles. The limited va lue of chronological age in defining midlife puts emphasis on alternative conceptions such as future time perspective for defi ning this stage of life in relation to young adulthood (Staudinger & Bluck, 2001). Future time perspective refers to where indi viduals place themselves in the life span in relation to time left to live (i.e ., close to death vs. far away from death) and how they perceive their future (i.e., open-ended and positive vs. limited and lacking opportu nities). Age is an important predictor of future tim e perspective: young individuals tend to perceive the future as open-ended, whereas older individuals accept it as more limited with higher awareness of declines and death (Fung & Carstensen, 2006). Some researchers suggest that midlife is composed of two phases: early-midlife (40-50 years) and late-midlife (50-60 years; Helson, So to, & Cate, 2006; Staudinger & Bluck, 2001). The current study embraces this two-phase perspec tive and suggests that th e shift in future time perspective is most clear in late-midlife. The early and late phases of midlife lead to quite different experiences, as one is first leaving young adulthood and then preparing to enter old age. The early phase may include continued growth and multiple roles, whereas one may begin to experience some losses or declines in the latter ph ase or at least to envision them on the horizon. For example, Cate and John (2007) compared wome n in their 20s, 40s and 50s in terms of their focus on opportunities for or limitations in the future. They found that women in their 20s focused more on opportunities than both midlif e groups who did not differ. The late-midlife group, however, focused more on limitations an d losses than both the young and early-midlife 18

PAGE 19

women. These findings were rep licated in a longitudinal stud y showing that women did not change from the age of 43 to 61 in terms of their focus on opportunities, however their focus on limitations and losses significantly increas ed in late-midlife (Cate & John, 2007). In sum, the transition from an open-ended and positive future time perspective to a more limited time perspective in which one begins to acknowledge that life has an ending tends to occur in late-midlife. This is one major developmen tal shift that should be represented as part of the conceptual self. That is, ones conceptual se lf changes across adulthoo d in terms of beliefs and expectations about how much time the self has left to live. The next section discusses developmental shifts in anot her aspect of the conceptu al self, self-attributes. Self-Attributes of Young and Late Middle-Aged Adults Self-attributes are another aspe ct of the conceptual self th at demonstrates developmental change. The current study compares young and late middle-aged adults self-attributes, and how these influence the content of their memories. Self-attributes have been conceptualized as psychological well-being by Ryff (1989a) who integr ated concepts from personality and clinical psychology (Allport, 1961; Maslow, 1968; Rogers 1961), as well as life span developmental psychology (e.g., Erikson, 1959; Neugarten, 1973) to create a parsimonious set of six selfattributes that have been validated in several studies (Ryff & Singer, 2006). Self-acceptance is a central component of self-act ualization (Maslow, 1968), optimal functioning (Rogers, 1961) and maturity (Allport, 1961). High scorers accept bot h their good and bad self characteristics. Positive relations with others refers to feeling empathy and af fection for others (Maslow, 1968), being capable of loving, and formi ng deep friendships (Erikson, 1959). Autonomy refers to making decisions independently and regulating behaviors and emotions from within. Autonomous individuals evaluate themselves by pe rsonal standards rather than others approval. Environmental mastery is the ability to choose or create environments compatible with ones 19

PAGE 20

physical and psychological needs, including the ability to make use of social opportunities. Purpose in life refers to having goals and a sense of direction. Finally, personal growth is the need for continued development and realization of ones potential. Viewing these attributes as part of the conceptual self is novel, but is in line with other researchers who refer to these attributes as aspects of personal ity (Fleeson & Heckhausen, 1997). Previous research shows adult age differences in some but not all of these self-attributes. For example, both young (18-29 years) and middleaged adults (30-64 years) express more personal growth and score higher on purpose in li fe than older adults (over 64 years; Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). Young adults report signi ficantly less environmental mastery than both middle-aged and older adults who do not differ (Ry ff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). Middle-aged adults report more autonomy than both young and old adults, although the difference from old adults is no t significant (Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). There are inconsistent results regarding two self-attributes, positive relations and selfacceptance. In terms of positive relations, some studies find no age differences (Ryff, 1989b; Ryff, 1991) and others show older adults repor t greater positive relati ons than both young and middle-aged adults (Ryff & Keyes, 1995). This is in line with socioemoti onal selectivity theory that argues that as people get older, they shift their attention to emotiona lly meaningful goals and enhancing intimate relationships (C arstensen, 1995). In some studies, older adults report greater self-acceptance than both middle-aged and young adults, although the significant difference is between the old and the young group (Ryff, 1991). Ot her studies have found a constant level of self-acceptance across the life span (Ryff, 1989b; Ryff & Keyes, 1995). 20

PAGE 21

The studies reviewed here did not use the two-phase perspect ive of midlife, but instead used samples with a wide midlife age range (3064). Therefore, their fi ndings may be unclear regarding distinctions betw een the young, midlife, and older groups due to their operationalization of midlife. For example, th ey showed that young and middle-aged adults do not differ in personal growth, self-acceptance or positive relations, whereas old adults score lower in personal growth, and higher in self-acceptan ce and positive relations than both groups (Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). It is unclear whet her this pattern of results would have been obtained if they had separa tely examined the two phases of midlife. Conceptualizing midlife as having two distin ct phases and focusing on late-midlife adds precision to conceptualizing the ex perience of midlife, and may pr event inconsistent results in examining self-attributes. Thus, the present st udy compares young adults with late middle-aged adults, and expects to see clear distinctions between th ese groups in self-a ttributes that have shown inconsistent results in the past (e.g., positive relations and self-acceptance). The present study predicts that late middle-aged adults should be more similar to older adults and will express greater levels of positive relations and self-acceptance, but less personal growth and purpose in life than young adults. In sum, research shows that two aspects of th e conceptual self, future time perspective and self-attributes, are likely to show meaningful developmental variation in adulthood. The next section combines these two age effects with the SMS model by introducing a comprehensive conceptual model (see Figure 1-2) showing future time perspective as a mediator between age group and current self-attributes (i .e., ones age group predicts future time perspective which influences ones self-attributes). The conceptual model also shows how th e conceptual self at different points in adulthood shape the conten t of retrieved autobiographical memories. 21

PAGE 22

The Role of Age and Future Time Perspective in Self-Memory Relations The current study is based on a new conceptual model that extends the SMS model and links four major constructs (i.e., age, future time perspective, curre nt self-attributes and autobiographical memory content). As shown in the model (see Figure 1-2), future time perspective is expected to medi ate the relation between age group and self-attributes. That is, adults have different self-attribut es not only because they are of different chronological ages, but because they have different perspectives of futu re time due to their age. How adults perceive time left to live may be just as meaningful as their chronological age (i .e., time since birth) in shaping their self-attributes. For example, fu ture time perspective is the central tenet of socioemotional selectivity theory which suggests that the amount of time people perceive they have left ahead of them is associated with the extent to which they hold emotional versus futureoriented goals (Carstensen, 1995; Fung & Carstensen, 2006). The model suggests that future time perspectiv e will function as a mediator for all but two self-attributes (i.e., autonomy and environmental mastery). These two self-attributes are expected to be directly shaped by ones age. Ryff and her colleagues show the most consistent results for these two self-attributes such that middle-aged adults alwa ys score higher than young adults (Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). The fact that th ese two self-attributes show more consistent and stronger links with age than other self-attributes may suggest that these links are not mediated by future time perspective. Furthermore, theoretically both autonomy and environmental mastery are expected to be gained through adult development and life experiences. That is, independent of ones future time perspec tive, the process of gaining experience as one masters life challenges and new environments across adulthood provides one with a chance to develop autonomy and environm ental mastery. In sum, the current study is based on a newly developed model which argues that ones age has direct effects on both future 22

PAGE 23

time perspective and self-attributes, but that it also has an indirect effect on four self-attributes through future time perspective as a mediator. Th e proposed model is an el aboration of the basic SMS model: age group and future time perspectiv e shape self-attributes (i.e., the conceptual self), which in turn shape the content of aut obiographical memories. The next section focuses on the effect of self-attributes on young and late mi ddle-aged adults earliest and recent memories. Testing the SMS Model: Effects of Self-Attri butes on Earliest and Recent Memory Content According to the conceptual model used in the current study, ag e and future time perspective shape self-attributes, which in turn influence autobiographical memories. As discussed above, individuals in different stages of adulthood ha ve different self-attributes. Differences in the conceptual self, through working self goals, lead to differences in the content of autobiographical memories. For example, Alea and Bluck (2003) argue th at individuals age, gender and personality have an impact on the go als of their working selves and hence on their memories. Thus, the present study tests the SMS model by examining wh ether self-attributes across different age groups are reflected in th e content of autobiograp hical memories (i.e., earliest and recent memories). According to the SMS model, self-attributes as part of the conceptual self, shape autobiographical memory content regardless of memory type. If earliest memories are the first memories that fit with the current self, they too should have content that matches with current self-attributes. For example, an individual who is currently high on personal growth should be more likely to recall an earliest memory that ha s a personal growth theme than an individual with low personal growth. No research to date has examined this relation between current selfattributes and earliest memories. Conway and Pleydell-Pearces (2000) model did not particularly focus on earliest memories, but simply refers to all autobiographical memories. As such, testing the impact of current self-attributes on the content of earliest memories will be a 23

PAGE 24

stringent test of the SMS model. That is, it should be more difficult to provide support for the SMS model using earliest memories because it is more difficult to show that current selfattributes have an impact on such distant memories. The current study tests the relation of self-att ributes to memories by examining earliest memories in relation to a more recent memory for comparative purposes. In addition to earliest memories, participants will be asked to retrieve a memory from the period between a year ago and three months ago. The last three months are ex cluded in order to avoid recall of very trivial events (e.g., yesterdays breakfast) that would not be comparable to earliest memories, and to avoid recency effects (e.g., remembering what you did just before be ginning the study). The effect of self-attributes is exp ected to be more evident on recent memories, because one is likely to have had the same self-attributes at the time of the event as curre ntly, at the time of retrieval. That is, as these memories have been encoded an d are being retrieved in the same developmental stage, there should be a strong ma tch between their content and one s current self-attributes. In contrast, the fit between ones se lf-attributes and ones earliest memory content should be less evident because these memories were encoded in a very different developmental stage (i.e., childhood) than the stage in whic h they are being retrieved. In conclusion, earliest memories are theorized to be retrieved because they are the first memories that match ones current self-attributes. The current study te sts this and thereby, investigates whether the SMS m odel provides a convincing explanat ion for the recall of earliest memories. In opposition to earliest memories, r ecent memories should show an even closer match between current self-attributes and memory content. As the SMS model does not have a developmental focus, it does not make predic tions about whether this relation between the current conceptual self and memo ry content will be different in different adult age groups. The 24

PAGE 25

current study predicts that th ere will be no differences be tween the two age groups. Although there are changes in the current conceptual self across development, the relation between current conceptual self and memory content should always be evident. That is, on es conceptual self in any developmental stage should guide the recall of th eir memories. In sum, in testing the relation of the conceptual self to memory content, the current study provides a new test of the SMS model (i.e., examining earliest memories), as we ll as one that has been used by a small number of researchers (e.g., examining recent memories; Conway & Holmes, 2004), and also examines whether this relation is stable across the adult life span. Specific Aims and Hypotheses The goals of the study were outlined in the lit erature review. They are summarized here in terms of the specific aims and hypotheses of the study. The first study aim is to examine whether young adults and late middle-aged adults have differe nt conceptual selves (i .e., different sense of future time perspective and self -attributes). The second aim is to examine whether any obtained age differences in self-attributes across the two age groups (as per aim 1, hypothesis 2) are driven by chronological age or are mediated by an individuals sense of future time perspective. Hypothesis 1 : Late middle-aged adults are expected to have a less open-ended future time perspective with a perception of future as limited and lacking opportunities as compared to young adults. Hypothesis 2: Late middle-aged adults are expected to have higher levels of positive relations, self-acceptance, autonomy and enviro nmental mastery than young adults, whereas young adults are expected to show higher purpose in life and personal growth than late middleaged adults. Hypothesis 3: Future time perspective is expected to be a mediator between age and four self-attributes: purpose in life, personal growt h, positive relations and self-acceptance. Future 25

PAGE 26

time perspective is not expected to mediate the relation between age and the remaining two selfattributes: autonomy and environmental mastery. These attributes should be directly linked to chronological age group. The third study aim is to provide a test of the SMS model by examining whether differences in self-attributes are reflected in di fferences in memory content overall. This also includes a stringent and novel test of the model which examines the link between the current conceptual self and earliest ch ildhood memories. According to the SMS model, these relations should occur regardless of age group. These findings will also elucidate whether the SMS model provides a convincing explanation for th e retrieval of earliest memories. Hypothesis 4: Individuals with higher levels (as comp ared to lower levels) of a particular self-attribute are more likely to rate their memori es as including higher levels of the content of that attribute. For example, those who are currently high in autonomy are more likely to rate their memories as including autonomy co ntent than those who are low in autonomy. This is expected for all six self-attributes and fo r both age groups, but the effect should be less evident on earliest memories than recent memories. That is, though bo th earliest and recent memories should reflect self-attributes, earliest memories ar e expected to be weaker represen tatives of self-attributes than recent memories. For example, in a person with high environmental mastery, that mastery should be more clearly represented as content in th eir recent than in their earliest memory. 26

PAGE 27

LONG-TERM SELF EPISODIC MEMORY SYSTEM Sensory I m a g e Autobiographical Knowledge Base Conceptual Self Future Time Pers. Self-Attributes Life Story Schema Life-time Period General Events WORKING SELF Figure 1-1. The self-memory system with self-att ributes and future time perspective shown as aspects of the conceptual self. 27

PAGE 28

CONCEPTUAL SELF SELF-ATTRIBUTES Purpose in life Age Group: Personal growth Future Time Perspective Memory content Young vs. Late middle-aged Positive relations Self-acceptance Autonomy Environmental Mastery Figure 1-2. The effect of self-a ttributes on autobiographical memo ry content as predicted by age group and future time perspective. 28

PAGE 29

CHAPTER 2 METHODS Participants The sample consists of 285 young adults (108 men, 177 women) and 135 middle-aged adults (55 men, 80 women). Y oung adults ranged from 19 to 29 years old (M = 21.13, SD = 1.18) and late middle-aged adults ranged from 47 to 64 years old (M = 55.67, SD = 5.99). The young age range represents the young adulthood life phase. The age range of the midlife sample represents late-midlife (not early-midlife) as di fferentiated in life span developmental theory (Staudinger & Bluck, 2001). Focusing on late mi dlife adds precision to conceptualizing the experience of midlife in terms of both future time perspective and self-attributes: Individuals in late midlife are likely to be distinct from younger adults on bo th of these constructs. Of participants who reported ethni city, 64% of the young adults were Caucasian, 15.2% were Hispanic, 6.4% were Asian, 7.8% were African Am erican, and 6.7% reported his or her race as other. All of the young sample had completed high school and were en rolled in university: There were 142 sophomores, 93 juniors and 49 seni ors. Of the late middle-aged adults, 87% were Caucasian, 8.3% were Hispanic, 1.5% were Asian, and 3% were African American. Ten percent of the late middle-aged adults had a Ph.D. degree, 27.3% had a Masters degree, 37.3% had a Bachelors degree, 9.4% had an Associat es degree and 15.6% had a high school degree. Invitations to complete the online survey that was composed of two parts were sent by email to 398 young and 501 late middle-aged adu lts. Among those, 346 young adults (87%) and 212 late middle-aged adults (4 2%) started the survey, and 309 young adults (78%) and 148 late middle-aged adults (30%) completed it (i.e., th ere were 37 young and 64 late middle-aged adults who did not complete both parts of the survey). To ensure data quality in those who did complete the entire survey, participants who did not follow the survey instructions and those who 29

PAGE 30

answered at least two of the data quality foil items incorrectly were excluded (twelve individuals). Twenty five individuals were excl uded because their reported age was not within the appropriate age ranges fo r the current study. Thus, the fi nal sample consisted of 420 individuals. The young adult sample was recruited from th e Psychology Departments participant pool and two other psychology courses. Students received course cred it for participation. The latemidlife participants were accessed through the yo ung adult sample. Before beginning the survey, young adults were asked whether they would lik e to receive two extra course credits by providing the researcher with names and contac t information of two middle-aged individuals. These middle-aged individuals c ould be family members, friends or neighbors who the student felt might be willing to participate in the study. Th e students were told th at they would receive extra credit whether the middle-aged individuals agr eed to participate or not but that they must provide valid names, email addresses and telepho ne numbers for each referred individual. These potential middle-aged participants were sent an email explaining how they had been contacted and inviting them to participate in the study. Th e email included a link to the online survey. These middle-aged adults were also asked if they would like to provide th e researcher with the contact information of other friends or relatives who might be interested in participating in the study. All middle-aged participants were compensa ted in two ways. First, after completing the survey, they were directed to a research-based informative handout on midlife development created by the researchers. Second, a $1.00 donation was made to one of the following developmentally-relevant and repu table charity organizations of the participants choosing: the American Association of Retired Persons ($ 22.0 0) or the Childrens Health Fund ($ 119.00). 30

PAGE 31

Procedure The online survey (Surveymonkey.com) was compos ed of two parts. Young adults and late middle-aged adults were sent an email including the link to the first pa rt of the survey and inviting them to participate. The email briefly ex plained the focus of the study and signified that participants had one week to complete the first survey. They were told that when they were finished with the first survey, they would be se nt the second survey link in 48 hours. Participants were free to complete the surveys anywhere that had access to the internet. They were, however, asked to choose a quiet location and told that they needed to complete each survey in one sitting. The study consisted of two sessions with a minimum 48-hour and a maximum 72-hour interval in between. The first survey included the informed consent form, the current selfattributes measure (Ryff, 1989a) and the demographi cs and health items presented in that order. In the second survey, individuals were prompted to share two memories and to rate the memory content of each (see description of memory-sha ring below). They then completed the two future time perspective measures: the Rappaport Time Li ne and the Future Time Perspective Scale. Participants wrote their memory narratives and completed the memory content ratings before completing the future time perspective measures to prevent their memory recall from being confounded by having been prompted to think about future time. Having two separate surveys with a 48-hour time lag prevented dependency between the current se lf-attributes measure (Session 1) and the memory-sharing and memory content ratings (Sessi on 2). Spreading the measures across two sessions also helped to prevent participants from becoming bored or exhausted while completing the survey. The firs t survey took about 10 minutes, and the second survey took about 20-30 minutes to complete. Through counterbalancing (within ag e and gender), half of the participants retrieved their earliest childhood memory first, followed by a recen t memory from the last year but excluding 31

PAGE 32

the most recent three months. The other half of the sample retrieved the recent memory first, followed by their earliest memory. Participants completed the first memory narrative and then the memory content ratings for that particular memory. They then proceeded to the second memory narrative and completed the same me mory content ratings for the second memory. The instructions for producing the memory narratives were designed to elicit two specific memories from the participants lives (see Appendix A). Specific memories were defined to them as any event/experience that occurred at a pa rticular place and time (i.e., it may have lasted minutes or hours, but the event itself was not long er than one day). Participants were told that their memories may be quite unique experiences or just everyday events, but what is important is that they should be specific events (i.e., not ge neral life periods) that say something about them as a person. They were asked to report memories that say something about them as a person so as to collect memories that are associated with the self as per the se lf-memory system model (Conway & Pleydell-Pearce, 2000). For both memo ries, participants were presented with a standard text box in which they were required to type approximately 800 characters. The number of characters required was determined through pilot-testing for common length of memory narratives produced. The earliest memory instructions were to take a moment to think back to their childhood and to then try to recall the very earliest even t/experience that reveals something about who they are as a person and to describe it in writing as thoroughly as possibl e. They were instructed to write down everything they were doing, thinking and feeling at the time of that event. Instructions emphasized that the memory shoul d be their own recollection from earliest childhood, not an episode that they had only seen in a picture or heard ab out from someone else, and that it should be the very earliest memory that they could recall. Rece nt memory instructions 32

PAGE 33

were to take a moment to think back over the la st year and to recall something memorable from the period between three months ago and a year ago and write about it. The most recent three months were excluded to avoid a re cency effect in recall. That is, to prevent participants from reporting very recent memories (e.g., things that had happened that same day) that may be about trivial events. Participants were reminded that this recent memory should be one that says something about who they are as a person and we re instructed to write everything they were doing, thinking and feeling at the time of that even t. Thus, the instructi ons were the same for both memories except for the focus on earliest ve rsus recent events. Exam ples of both earliest and recent memory narratives are presented in Appendix B. Measures The first survey included the current self-attributes measure and the background measures including demographics and h ealth status. In the second su rvey, individuals shared two memories and completed self-rated memory conten t ratings. They then completed the two future time perspective measures. Self-Attributes Measure This 54-item scale (Ryff, 1989a) consists of subscales designed to measure six different self-attributes. The subscales are: (1) self-acceptance, (2) posi tive relations with others, (3) autonomy, (4) environmental mastery, (5) purpose in life, and (6) personal growth. The short form, consisting of nine items in each subscale, was used (see Appendix C). Internal consistency for the total score in the current study was very high with a Cronbachs alpha of .95 and the six subscales showed moderate to high consistenc ies with Cronbachs alphas between .75 and .90. Participants were instructed to focus on their current selves while answering the questions, and rated the items using a scale ranging from 1 (str ongly disagree) to 8 (s trongly agree). Scoring 33

PAGE 34

consisted of summing the nine items in each subscale after reversing negatively phrased items. Higher scores are indicative of higher leve ls of that particul ar self-attribute. Memory Content Ratings This measure consists of 24 Likert-type items based on the 54-item version of the Ryff scales (1989a; see Appendix D). The measure wa s developed for the current study to assess the extent to which participants self-attributes (i.e., autonomy, environmental mastery, purpose in life, personal growth, self-acceptance and positive relations with others) are reflected in the content of their autobiographical memory narratives. Content rela ted to each self-attribute is measured with four self-report items, two of which are reversed. These items were selected from the larger scale because they were applicable (w ith minor modification) to rating in relation to memory content, and they had high loadings on the relevant self-attribute factor suggesting that they were representative items fo r that factor (Abbott et al., 2006). Internal consistencies were calculated separately for earliest memory and recent memory content ratings overall (Cronbachs alpha = .91 for both). Internal consistencies for environmental mastery, self-acceptance and posit ive relations were high in both memories (Cronbachs alphas between .82 and .87). Internal consistencies for purpose in life, personal growth and autonomy, however, were low in bot h memories. Therefore, one item from each of these three subscales was excluded to increase in ter-item reliability. When the item In this memory, I was living each day and not thinking about the future was excluded from the purpose in life subscale, Cronbachs alpha increased from .46 to .63 for the recent memory, and from .50 to .58 for the earliest memory. When the item In this memory, I was not trying to make improvements or changes was excluded from the personal growth subscale, Cronbachs alpha increased from .56 to .62 for the recent memory, and from .64 to .68 for the earliest memory. Finally, when the item In this memory, I was influenced by strong others was excluded from 34

PAGE 35

the autonomy subscale, Cronbachs alpha increase d from .46 to .48 in the recent memory, and stayed at .51 in the earliest memory. New comp osite scores were calculated for these three subscales, with three items in each subscale, an d the mean score for each subscale was used in the analyses. Participants responded to each item in rela tion to both their recent and their earliest memory on a scale ranging from 1 (strongly disagr ee) to 8 (strongly agree). The possibility of dependency in responses across the current se lf-attributes measure (Ryff, 1989a) and the memory content ratings was clearly an issue and several things were done to prevent it. The two measures were administered with a minimum 48hour time lag between them to reduce recall of previous responses, and foil ite ms were distributed among the items of both measures. In addition, there were clear instructional set diffe rences between the two measures. Ryffs (1989a) measure of current self attributes instructed participants to focus on themselves in general and in the present, whereas the memory content ratings instructed them to focus on each specific reported memory while answering the questions. While the time lag and the instructional set differences should take care of any dependency issues, additional precautions were taken. Both the content and the order of foil items were different in the two measures to reduce ability to remember how one had previously responded. Th e exact wording of the items measuring the same self-attribute (e.g., autonomy) across the two measures was sli ghtly different, again so that participants were unlikely to recall earlier responses to the items. Aside from the memory content ratings, participants reported the date of each of their memories to the closest month and year, as well as reporting their chronologi cal age during each remembered event. Future Time Perspective Measures Two future time perspective measures were us ed to assess different aspects of this construct. The first is a modified version of the Rappaport Time Line (Rappaport, Enrich, & 35

PAGE 36

Wilson, 1985) in which participants receive a horizontal line representing their life starting from birth and ending at death. They respond by clicking a now point on the line to indicate where they subjectively feel they are in their life at this time. Distance from death marked on the line was represented numerically betw een 16 (birth) and 1 (death). Hi gher scores indicate a longer future time perspective. Carstensen and Langs (1996) Future Time Pers pective Scale was also used (see Appendix E). Participants provided ratings on a scale from 1 (very untrue) to 7 (v ery true) indicating the degree to which they agreed with each of ten items. A sample item is Many opportunities await me in the future. The total score for each partic ipant is the mean across all items. Higher scores indicate a more open-ended and positive future, whereas lower scores indicate a more limited and pessimistic future time perspective. Internal consistency is very high with a Cronbachs alpha of .89. The correlation between these two future time pe rspective measures was moderate, as they assess slightly different aspects of future time perspective, r (417) = .53, p < .05. The Rappaport Time Line is a straightforward measure of time pe rspective that represents ones subjective sense of their position in the life span. The Future Time Perspective Scale assesses subjective time left (similar to Rappaport Time Line), but also includ es ones feelings about what the future may hold. Particularly, those who recei ve higher scores indicate optimistic views of the future and believe that the future holds oppor tunities for them. Given this conceptual dis tinction and the moderate correlation, these two measures are treate d separately rather than collapsed into one composite score. Both future time perspective scales were administered after part icipants had provided narratives of their earliest and recent memories in a counterba lanced order. A multivariate 36

PAGE 37

analysis of variance (MANOVA) showed that th is order manipulation did not have an impact on how participants responded to the Rappaport Time Line or th e Future Time Perspective Scale, F (2, 414) = 0.24, p > .05. Background Measures The background measures includ e demographics (see Appendix F) and a current health status question, assessed for potential use in anal yses as covariates. Current health status was assessed due to its potential to differ across age groups and also to affect ones sense of future time perspective. Individuals completed a single item rating their health on a 6-point scale ranging from 1 (very poor) to 6 (very good): Compa red to other people my age, I believe my health to be (Zelinski, Burnight, & Lane, 2001). Use of Online Data Collection: Considerations Data was collected through an online survey created at SurveyMonkey.com. Internet-based data collection is an increasin gly popular method of conducting ps ychological research. Previous research on the quality of internet-based data collection shows no differen ces between traditional paper and pencil surveys and online surveys in asse ssing such things as student ratings of quality of instruction or reportin g of sexual behaviors, in terms of in ternal consistency, criterion-related validity, factor loadings and mean scores (Chang, 2005; Chuah, Drasgow, & Roberts, 2006; Epstein, Klinkenberg, Wiley, & McKinley, 20 01; King & Miles, 1995; Lonsdale, Hodge, & Rose, 2006). Furthermore, missing data are often more common in the paper and pencil format compared to both supervised and unsupervised online surveys (Lonsda le et al., 2006; Wood, Nosko, Desmarais, Ross, & Irvine, 2006). For open-ended questions such as the memory narratives in the present researc h, studies show that online survey responses are either longer or of the same length as mail responses (Fricker & Schonlau, 2002). Online surveys also lead to higher and faster response rate s than surveys distributed via postal mail (Cobanoglu, Warde, & 37

PAGE 38

Moreo, 2001; Lonsdale et al., 2006). Using undergraduates at th e University of Florida as participants, Pealer, Weiler, Pigg, Miller, and Dorman (2001) found that online surveys showed the same response rates as mail surveys. Furtherm ore, online surveys enc ourage collecting larger samples as they have the potential to eliminate costs and labor-intensive fielding tasks such as survey package preparation, mailing and data entry (Fricker & Schonlau, 2002). The literature suggests that using an online survey can be adva ntageous in terms of data collection speed and quality. In addition, adults in midlife are notorious ly difficult to recruit du e to their busy lives. Use of an online survey is a good tool for reaching this sample of adults (Nosek, Banaji, & Greenwald, 2002). The use of an online survey might also have disadvantages. First, internet samples may under-represent populations that have low levels of access to the internet, however the challenges for internet research are similar to the challenge s of all experimental re search in this regard (Nosek et al., 2002). The current sample of young a dults and individuals in late midlife however were highly likely to have intern et access either at home, at work, or both. Thus, the online format was not expected to create undue bi as in this sample. Second, participants can involuntarily end participation in the case of a computer or server crash, a broken internet connection, a program error or even a power outage (Nosek et al., 2002). This problem was prevented by asking the participants to report such cases to the e xperimenter so that their missing data could be labeled involunt ary in contrast to those w ho decided to end participation voluntarily. One participant in the current sample involuntarily exited the survey and was allowed to begin again. In this study, various measures were taken to increase data quality in the online format. First, the survey was broken up into multiple pages with a small number of questions on each 38

PAGE 39

page to make it clearer for participants as well as preventing technical problems such as pages not loading due to their lengt h. Second, there were foil items embedded in the survey that checked if the participants were really reading the items (e.g., items that simply give an instruction such as Answer Strong ly agree for this item). Third, to control for narrative length in providing their memory narrat ives, participants were given a standard text box in which they were required to type a certain number of characters before they could move on to the next item. Fourth, participants were timed, and those who sp ent less than a certain amount of time and those who spent too much time on the survey were excluded from the study. Finally, they were given one week to complete the survey as Fricker and Schonlau (2002) suggest ed that researchers should leave a survey in the field for about ten days to achieve 70-80 % response rates. 39

PAGE 40

CHAPTER 3 RESULTS The results are divided into two major sections. The first sectio n reports preliminary analyses. The second section reports findings related to the study hypotheses using univariate and multivariate analyses of variance, bootstra pping for testing mediation, and hierarchical regression analyses. Preliminary Analyses Four sets of preliminary analys es were necessary. First, any potential differences in health and demographic variables, or the order of reporting the memories were examined to identify potential covariates that may need to be included in major analyses. Second, the dependence between young and middle-aged participants due to the referral-based recruitment method was tested. Third, sample selectivity issues between participants who complete d the entire survey and those who dropped out were examined. Finally, a manipulation check was conducted analyzing young and late middle-aged adults age during their earliest and recent memories. Potential Covariates Current health status was considered as a possible covariate. A 2 x 2 ANOVA showed that there were no age or gender differences in current health, F ranges from 0.86 2.44, all p > .05.1 One concern was that health status might drive individuals views of time left to live. Current health was not, however, correlated with the Ra ppaport Time Line, and had a weak correlation with the Future Time Perspective Scale, r (415) = .20, p < .05. Thus, current health was not entered as a covariate in further analyses. 1 The sample sizes in the analyses vary due to missing data on some of the criterion variables. 40

PAGE 41

Preliminary analyses also examined potential differences by race and gender in the major study variables. In order to test whether there were differences in the two future time perspective measures, current self-attributes, earliest memory content ratings, or recent memory content ratings, four MANOVAs were run with race and gender entered as the two independent variables. Results showed that race had no si gnificant main effect on any of the dependent variables, F ranges from 0.30 2.02, all p > .05. In contrast, sex had a main effect on current self-attributes, F (6, 326) = 3.22, p < .05. Results showed that women ha d higher levels of personal growth, positive relations, purpose in li fe and self-acceptance than men, t ranges from 2.11 4.44, all p < .05. As sex had a si gnificant effect on an important variable (i.e., current selfattributes), it was entered as an independent variable (i.e ., to test interactions) or as a covariate in all of the major analyses. Finally, in order to test wh ether the order of recalling th e two memories (i.e., earliest memory first vs. recent memory first) had an effect on the content of the memories, two MANOVAs were conducted. The first MANOVA incl uded earliest memory content ratings as the six dependent variables and order as the inde pendent variable. Results showed that order had a significant main effect on earliest memory content, F (6, 390) = 3.22, p < .05. When earliest memories were recalled first followed by the recent memory, adults rated their earliest memories as having higher levels of all six memory content themes than when recent memories were recalled first. The second MANOVA included re cent memory content ratings as the six dependent variables and order as the independent variable. Resu lts showed that order did not have a significant main effect on recent memory content, F (6, 369) = 1.57, p > .05. These findings suggest that only earliest memory c ontent was affected by the order of memory retrieval. As order showed this significant effect it was entered as a predictor in any analyses 41

PAGE 42

involving memory content (i.e., Hypothesis 4 ex amining the relation between current selfattributes and memory content). Sample Dependency Issues As reported earlier, younger adul ts were given the option to refer middle-aged adults to participate in this study. In order to ensure that young participants who referred middle-aged adults were not significantly different from young participants who did not refer middle-aged adults, the two groups of students were compared in terms of background measures (i.e., race, education level and health) and major study variab les (i.e., current self-a ttributes, future time perspective and memory content ratings). An alyses of variance showed no significant differences, F ranges from 0.13 2.21, all p > .05. Similarly, late-midlife participants were also asked if they would like to refer other middle-aged people. Analyses showed that late-midlife participants who referred middleaged adults were not significantly different from late-midlife participants who did not provide referrals, in terms of background measures or major study variables, F ranges from 0.22 2.01, all p > .05. This method of recruitment also led to some young adults being related to the middle-aged individuals who particip ated in the study (e.g., a young student referred his or her middle-aged parent) and some late middle-ag ed adults being related to other middle-aged adults (e.g., a middle-aged person referred his or her friend) Univariate ANOVAs were conducted to examine whether these pairs of related individuals were different from single individuals who did not refer anyone. A dummy variable for group was created so that each group (e.g., each youngmidlife referral pair, or midlife-midlife referral pai r) had a distinct code, as did single individuals (i.e., those who did not refer anyone). ANOVAs were conducted w ith the group dummy variable entered as a random factor. ANOVAs were run for all dependent variables: current selfattributes, future time perspective measures and memory content ratings Groups did not differ 42

PAGE 43

significantly on any variables, F ranges from 0.6 2 1.26, all p > .05. This shows that there were no differences between young-midlife referral pa irs, midlife-midlife referral pairs, and individuals who did not provide referrals, in the way they res ponded to any of the major study measures. Sample Selectivity: Attrition Due to Drop-Out To ensure that participants who dropped out after completi ng the first survey were not significantly different from those who completed the entire study, the tw o groups were compared in terms of available measures on both groups: the background measures (i.e., race, education level, sex and health) and current self-attributes. Chi-square tests showed that race was the only demographic variable that show ed a significant difference between the two groups, and this was the case for only young adults: African American young adults (n = 8) dropped out at a higher rate than the expected value (3), (N = 316, df = 4) = 11.04, p < .05. A MANOVA showed no differences between those who did and did not complete the entire survey in terms of cu rrent self-attributes in either age group, F (6, 420) = 1.08, p > .05. These findings suggest that overa ll there were no major differenc es between adults who did or did not complete the entire survey. Howeve r, young African Americans showed a higher dropout rate. Age at Time of Memory In terms of age at time of earliest memory, young adults ( M = 4.53, SD = 1.51) and late middle-aged adults (M = 4.74, SD = 1.57) showed no differences, t (370) = 1.22, p > .05. On average, people recalled memories from when they were between four and five years old. This is in line with previous research that examined the earliest memories of both young and midlife adults and found that childhood amnesia wanes ar ound the age of 4.7 (Multhaup et al., 2005). In terms of their recent memories, both age gr oups correctly responded to the study task by 43

PAGE 44

reporting memories that were from the previous year: On average, young adults were about 19 years old ( M = 19.31, SD = 2.10), whereas late middle-aged adults were on average about 55 years old in their recent memories ( M = 54.77, SD = 5.50). Major Analyses: Testing Hypotheses Results are provided below for each of the study hypothesis. Hypotheses were addressed using a variety of analytical techniques: An alyses of variance, bootstrapping for testing mediation, and hierarchical regressions w ith appropriate follow-ups were conducted. Hypothesis 1: Relation of Age Group to Future Time Perspective Late middle-aged adults were expected to have a less open-ended perspective on future time, viewing their future as limited time-wise and lacking in opportunities as compared to young adults. To demonstrate the predicted deve lopmental differences between the two age groups, a MANOVA was conducted with age group and sex as the between-subjects factors. Dependent variables were the two future time pe rspective measures. Init ial evaluations of the data showed that Box-M test for the homogeneity of variance-covariance matrices produced a significant result, F (9, 421263.09) = 5.77, p < .001. Levenes test found that the assumption of homogeneity of variance coul d not be supported for the two time perspective measures, F ranges from 4.97 9.79, all p < .05. Therefore, the more conservativ e and robust Pillais trace is used for the estimation of F -statistics in the analysis that follows. The MANOVA showed a significa nt main effect for age F (2, 412) = 450.04, p < .05, p = .69, but not for sex, F (2, 412) = 1.36, p > .05. There was no significant interaction with sex, F (2, 412) = 0.85, p > .05. As predicted, follow-up univariate ANOVAs showed significant differences between young and late middle-aged adults on both the Future Time Perspective Scale, F (1, 413) = 90.19, p < .05, p = .18), and the Rappaport Time Line, F (1, 413) = 901.89, p < .05, p = .69. Individuals in late midlife had a less open-ended future time perspective (M = 44

PAGE 45

47.92, SD = 11.44) than young adults (M = 57.82, SD = 8.72) as measured by the Future Time Perspective Scale, and had lower scores on the Rappaport Time Line (M = 6.83, SD = 2.03) than young adults (M = 12.04, SD = 1.41) indicating a shorter future time perspective. Hypothesis 2: Relation of Age Group to Current Self-Attributes Individuals in late midlife were expected to have higher current le vels of several selfattributes, including positiv e relations with others, self-accep tance, autonomy and environmental mastery than young adults. Young adults were e xpected to show higher purpose in life and personal growth than those in late midlife. To test for the predicted developmental differences on current self-attributes betw een the two age groups, a MANOVA was conducted with age group and sex as the between-subjects factors. Dependen t variables were the six current self-attributes. Initial evaluations of the data showed that Box-M test for the hom ogeneity of variancecovariance matrices produced a significant result, F (63, 110337.2) = 1.73, p < .001. Levenes test showed that the assumption of homogeneity of variance coul d not be supported for four of the six dependent variables, F ranges from 3.35 5.49, all p < .05. Therefore, the more conservative and robust Pillais trac e is used for the estimation of F -statistics in the analyses that follow. The MANOVA found significant main effects for both age, F (6, 349) = 15.03, p < .05, p = .21, and sex, F (6, 349) = 3.98, p < .05, p = .06. There was no significant interaction, F (6, 349) = 1.40, p > .05. Follow-up univariate ANOVAs were conducted. In terms of age, late middle-aged adults had significantly higher auto nomy, environmental mast ery, self-acceptance, positive relations and purpos e in life than young adults, F ranges from 6.79 54.16, all p < .05 (see Table 3-1). There was no significant difference in personal growth between young and late middle-aged adults, F (1, 354) = 2.41, p > .05. These findings mostly supported the hypothesis, 45

PAGE 46

but those in late midlife, instea d of scoring higher than younger adults on four of the six selfattributes scored either on par or high er than the young on all six self-attributes. In examining current self-attributes by sex, women self-reported higher personal growth ( M = 59.34, SD = 7.32) than men (M = 56.46, SD = 9.01), F (1, 354) = 4.67, p < .05, and more positive relations with others ( M = 59.53, SD = 9.20) than men ( M = 54.78, SD = 11.59), F (1, 354) = 12.35, p < .05. Men and women were similar in term s of the other four self-attributes, F ranges from 1.35 3.58, all p > .05. Hypothesis 3: Examining Future Time Persp ective as a Mediator Between Age Group and Self-Attributes Future time perspective was expected to be a mediator between age group and four selfattributes: purpose in life, personal growth, posi tive relations and self-acceptance. That is, some of the variance in these four se lf-attributes explained by particip ants age group was expected to be due to their sense of future time. Autonomy and environmental mastery were expected to be directly linked to chronological age. Analytical approach Mediation was tested using Preacher a nd Hayes (2004) bootstrapping method. This innovative, non-parametric method for testing me diations does not impose the assumption of normality of the distributions and does not require large samples (Preacher & Hayes, 2008). This method was preferred over Baron and Kennys (1986) causal steps ap proach with the Sobel test (1982) due to its higher power and higher control over Type 1 e rror rate (MacKinnon, Lockwood, & Williams, 2004; Hayes, 2009). This method was also preferred particularly for the current study because it allows entering mu ltiple mediating variables into the model simultaneously (i.e., the two future time perspective measures) to see whether they have a combined indirect effect on the dependent variable (Preacher & Hayes, 2008). This method also 46

PAGE 47

allows testing the indirect effect of each of th e mediators individually while controlling for all other variables in the model a nd compares the effects of the mediators against one another (Preacher & Hayes, 2008). As the two future time perspective measures used in the study assess slightly different aspects of the concept of future time, they were tested both separately and simultaneously in the mediation analyses. That is, three mediation mo dels were used. In the first two models, the Future Time Perspective Scale and the Rappaport Time Line were entere d separately as single mediators. The third model was a multiple-mediators model, in which the Future Time Perspective Scale and the Rappaport Time Line were entered simultaneously. Running these analyses with single and then multiple mediation m odels allowed for investigation of each of the time perspective measures as independent mediators, as well as examining both their joint effects and their relative efficacy. It is important to distinguish between vari ous effects and their co rresponding weights in these mediation analyses to interpret the followi ng results. For illustrative purposes, Figure 3-1 depicts the general mediation model used for all of the mediation analyses, illustrating the hypothetical paths through which th e Future Time Perspective S cale and/or the Rappaport Time Line are expected to mediate the relation between age group and current self-attributes. Figure 31A presents the total effect of the independent variable (i.e., IV = age group) on the dependent variable (e.g., DV = autonomy) which is weight c In Figure 3-1B, weights a1 and a2 represent the effect of the IV (i.e., age group) on the me diators (i.e., Future Time Perspective Scale and Rappaport Time Line). Weights b1 ad b2 represent the effects of the mediators on the DV (e.g., current self attributes, such as autonomy) partialling out the effect of the IV and any other mediators entered in the model. All of these pa ths are quantified with unstandardized regression 47

PAGE 48

coefficients. The indirect effect of the IV on the DV through the mediators (i.e., the mediational path) is the product of weights a and b (i.e., a1*b1 or a2*b2). This indirect effect and the direct effect (i.e., weight c' ) of the IV on the DV after control ling for the mediator(s) are the two components that sum to explain the total effect of the IV on the DV (i.e., weight c ). A mediation model is a path model that sp ecifies a causal chain between the IV, DV and the mediating variables (MacK innon, 2008). A significant mediat ion requires obtaining a significant indirect effect, as well as a direct effect (weight c' ) that is of smaller magnitude than the total effect (weight c ). That is, the simple relation be tween the IV and the DV should be reduced with the addition of the mediator (MacKinnon et al., 2000). In contrast, if the magnitude of the direct effect becomes larger than the tota l effect after the addition of a mediator(s), this indicates a suppression effect (i.e., inconsistent mediation; Davis, 1985). Suppression occurs when a mediator increases the magnitude of the relation between the IV and the DV, and thus increases the predictive validity of the IV (MacKinnon et al ., 2000). Within a mediation model, the fact that the direct and the indirect effects of an IV on a DV have opposite signs is also a sign of suppression (Shrout & Bolger 2002). Suppressor variables ar e informative: they suppress variance in the IV that is irrelevant to th e prediction of the DV, and thereby enhance the predictive power of the IV (Tabachnick & Fi dell, 2001). In sum, me diation models and suppression models are closely related, and are tested using the same set of methods. The bootstrapping analyses performed here can reveal whether there is mediation, suppression or no relation between variables in the specified models (Shrout & Bolger, 2002). Findings To begin, bivariate Pearsons correlations we re conducted for all of the variables to be used in mediation analyses: current self-attributes, the future time perspective measures and age group (see Table 3-2). Note that age group is negativ ely correlated with both of the future time 48

PAGE 49

perspective measures, but its asso ciation with the Rappaport Time Li ne is much stronger than its association with the Future Time Perspective Scale. Also note that the Rappaport Time Line is negatively correlated with some of the self-attributes, whereas the Future Time Perspective Scale is positively correlated with self-attributes. Mediation analyses were conducted separately fo r each of the six self-attributes, with three mediation models for each self-attribute as pres ented in Tables 3-3 3-5. The analyses and bootstrap estimates that are pr esented are based on 5000 bootstra p re-samples. Bias corrected confidence intervals with a confidence level of 95 % were computed.2 Point estimates of indirect effects were considered significant when zero wa s not contained in their confidence intervals indicating that the indire ct effects are significantly different from zero. Note that these mediation analyses include the variables examined in hypothesis 1 and 2. MANOVA results were presented above to illust rate mean group differences since those were most appropriate for testing hypothesis 1 and 2. Th e regression results, as expected, show the same pattern of findings for those variables. Supporting hypothesis 1, age group was negatively and significantly associated with both the Fu ture Time Perspective Scale and the Rappaport Time Line ( a weights in Tables 3-3 3-5) repl icating the MANOVA fi ndings. Age group was positively and significantly associated with five of the current self-attributes ( c weights in Tables 3-3 3-5) also replicating the MANOVA fi ndings presented for hypothesis 2 above. Three mediation models were used to test th e separate and combined effects of the two time perspective variables as medi ators. In the first two models, the Future Time Perspective 2 Bias corrected confidence interval s are preferred over percentile confidence intervals or bias corrected and accelerated confidence intervals due to extensive simulation results sup porting bias corrected bootstrapping (Preacher & Hayes, 2008). 49

PAGE 50

Scale and the Rappaport Time Line were each entere d separately. In the third model, they were entered simultaneously. With respect to the effect of the two mediators on current self-attributes ( b weights), although there were no specific hypotheses, negative weights were expected for positive relations and self-acceptance. That is, in keeping with the hypotheses concerning age group differences in current self-attributes, it was expected that (due to the inverse conceptual relation between age and future time perspective) future time perspectiv e would be negatively related to these two self-attributes, but positivel y related to personal growth and purpose in life. Finally, future time perspective was not expected to be related to autonomy and environmental mastery. The first model showed, however, that (con trolling for age group) the Future Time Perspective Scale was significantly and positively re lated to all six current self-attributes (see Table 3-3). That is, as time perspective incr eases (i.e., becomes more open-ended and positive), the level of all self-attributes increases. Similarly, the second model (using the Rappaport Time Line) also revealed positive relations, but for onl y four self-attributes: positive relations, purpose in life, personal growth and self -acceptance (see Table 3-4). That is as predicted, time left to live as measured by the Rappaport Time Line di d not predict adults current autonomy or environmental mastery levels. The third model show ed that when the two measures were entered as simultaneous mediating variables, only the Future Time Perspective Scale significantly predicted the six self-attributes (see Table 3-5) There were significant positive relations between Future Time Perspective and the six self-attri butes, but no significan t relations between the Rappaport Time Line and any of the self-attributes. With respect to the indirect effect of age group on the current self-attributes through the mediators (weights a*b ), the first model showed that the indirect effects through the Future Time 50

PAGE 51

Perspective Scale were significant for all six sel f-attributes (see Table 33). Confidence intervals for the point estimates of these six indirect e ffects did not contain zer o, indicating statistical significance. These significant indirect effects, however, indicated suppression rather than mediation. For all six self-attributes, the indirect effects and direct eff ects had opposite signs. As is usual with a suppressor effect the addition of the Future Ti me Perspective Scale into the model increased the magnitude of the direct effects making them larger than the initial total effects of age group on the self-attr ibutes. That is, the Future Time Perspective Scale acted as a suppressor rather than a mediator in the relati on between age group and the six self-attributes. It suppressed the irrelevant varian ce in age group allowing age group to emerge as an even stronger predictor of each of the self-attributes. The second model showed that the indirect effects through Rappaport Time Line were significant for only three self-attri butes: purpose in life, personal growth and self-acceptance (see Table 3-4). Similar to the findings for the Future Time Perspective Scale, the Rappaport Time Line acted as a suppressor rather than a mediator in the relation betw een age group and current self-attributes: The direct effects of age gr oup on purpose in life, personal growth and selfacceptance were larger than the total effects of age group on these three self -attributes, indicating suppression. The third model that analyzed the Future Time Perspective Scale and Rappaport Time Line simultaneously revealed specific indirect effects separately for the two mediators, and a total indirect effect for both mediators as a combined set. In terms of specific indirect effects, when the Future Time Perspective Scale and Rappaport Time Line were analyzed simultaneously, only the indirect effect through the Future Time Pe rspective Scale was significant (see Table 3-5). Again in line with suppression, the direct eff ects of age group on the six self-attributes were 51

PAGE 52

stronger than the total effects of age group on th ese self-attributes That is, the Future Time Perspective Scale was a significant suppressor in the eff ect of age group on all six self-attributes (controlling for Rappaport Time Line and age group), but Rappaport Time Line was not a significant suppressor (cont rolling for Future Time Perspectiv e Scale and age group). In terms of the total indirect effect of the two mediators (i.e., the specific indirect effect through Future Time Perspective Scale plus the speci fic indirect effect through Rappaport Time Line), the total indirect effect of age group was significant on onl y one self-attribute: personal growth. That is, the total indirect effect of age group on personal growth through the two variables was significant with a point estimate of 0.79, and a 95 % bias corrected bootstrap confidence interval of 1.4539 to 0.1243. This suggests that the two variables worked together as suppressors in the rela tion between age group and personal gr owth, but that the Future Time Perspective Scale, not the Rappaport Time Line, played the suppressor ro le in relation to the other five self-attributes. These findings suggest that the Future Time Perspective Scale and the Rappaport Time Line do not act as mediators of the relation between age group and current self-attributes as hypothesized. They do (particularly the Future Time Perspective Scale) play an important role in explaining the relation of age group to current se lf-attributes. The nature of this role (i.e., suppression), however, is different from the hypothesized expectation of mediation. Hypothesis 4: Relation of Current Self-Attributes to Memory Content The first three hypotheses examined adult develo pmental differences in two aspects of the conceptual self: future time perspective and cu rrent self-attributes. Recall that the final hypothesis had a different focus: testing the relati on of the conceptual self to autobiographical memory content. Hypothesis 4 provided an empi rical test of the self -memory system model (Conway & Pleydell-Pearce, 2000) by examining wh ether current self-attri butes are reflected in 52

PAGE 53

the content of both earliest and recent memories More specifically, i ndividuals with higher current levels of a particular self-attribute were expected to rate their memories as including higher levels of the content of that attribute. This was expected for a ll six self-attributes and across both age groups. Note, however that the effect was predicted to be less evident in earliest memories than in recent memories. To test this hypothesis, autobiog raphical memories rather than participants were used as the unit of analysis. The hypothesis was tested with six hier archical regression analyses conducted separately for the six current self-attributes (see Tables 3-6 3-11). The criterion variable was memory content rating (e.g., autonomy content in memories). Sex was included in the initial step of each regression. In the second step, age group (young vs. late midlife), memory type (earliest vs. recent) and cu rrent self-attribute (e.g., current autonomy) were entered. In the third step, two-way interaction te rms were entered: the interaction between memory type and age group, the interaction between memory type and current self-attribute, and the interaction between age group and current self-attribute. Fina lly, a three-way interact ion term between age group, memory type and current self-attribute wa s entered. Variables in cluded in interaction terms were centered by subtracting th e mean of each variable from the participants score for that variable.3 Four findings were central to testing hypothesis 4: a signific ant effect of current selfattributes on memory content, no interaction between current self-att ributes and age group, a 3 The six hierarchical regression analyses conducted to test hypothesis 4 were replicated with order of reporting the memories (earliest memory first vs. recent memory first) as an additional predictor. This did not change the pattern of effects found in hypothesis 4, therefore regressions without the order variable are presented for simplicity purposes. 53

PAGE 54

significant interaction between current self-attributes and memory type, and no three-way interaction between current self-att ributes, age group and memory type. As predicted, there was a si gnificant and positive relation be tween current self-attributes and memory content for all six self-attributes (see Step 4 in Tables 3-6 3-11). Overall, individuals with higher levels of current positive relations, pu rpose in life, personal growth, autonomy, environmental mastery and self-accepta nce recalled memories that contained higher levels of these six themes than individuals who scored lower in these self-attributes. Also as expected, the interaction between current self-attributes and ag e group was non-significant for all six self-attributes (see St ep 4 in Tables 3-6 3-11), indicating no age differences in the relation between current self-attributes and memory cont ent: The relation between current self-attributes and overall memory content appears to hold for both young and late middle-aged adults. Note that this finding was qualified by a three-way interaction, invo lving only autonomy, as described below. The interaction between current self-attri butes and memory type was significant for environmental mastery and self-acceptance (see Step 4 in Tables 3-10 and 3-11). Pearsons correlations were thus conducted separately for ea rliest and recent memories. Results show that for both variables there is a si gnificant correlation between cu rrent self-attribute and memory content for the recent but not for the earliest memory. That is, there was no relation between current environmental mastery and environmental mastery content of earliest memories, whereas there was a significant positive relation be tween current environmental mastery and environmental mastery content of recent memo ries, r (394) = .25, p < .001. Similarly, there was no relation between current selfacceptance and self-acceptance content of earliest memories, but there was a significant positive relation between current self-acceptance and self-acceptance 54

PAGE 55

content of recent memories, r (395) = .32, p < .001. In sum, these two sign ificant interactions provide partial support for the hypothesis that the re lation between current self-attributes and memory content is less evident in earliest me mories than in recent memories. For both age groups, all six self-attributes showed a significant positive relation with the content of recent memories, however only four self-attributes (i.e., positive relations, purpose in life, personal growth and autonomy) were linked with the content of earliest memories. Finally, the three-way interaction between age group, memory type and current selfattributes was significant, but for only one of the six memory content ratings: autonomy (see Step 4 in Table 3-9). Pearsons correlations were conducted separately for th e earliest and recent memories of both young and late middle-aged adults. Findings showed a positive relation between young adults current le vels of autonomy and the aut onomy content of their recent memories, r (278) = .25, p < .001, but not earliest memories, r (275) = .10, p > .05. In contrast, late middle-aged adults current autonomy wa s related to their earliest memory content, r (126) = .28, p < .05, but was not related to their recent memory content, r (122) = .16, p > .05. That is, young and late middle-aged adults current levels of autonomy related differently to their earliest and recent memory content. Autonomy is the only self-attribute that shows this age difference in the relation between current self -attributes and memory content. The relation of current selfattributes to memory content is the same for young and late middl e-aged adults across the other five self-attributes (i.e., positive relations, pur pose in life, personal growth, environmental mastery and self-acceptance). In sum, these four findings generally support hypothesis 4. They show that current selfattributes are significantly associated with the content of memories, that this relationship does not vary across age groups, and that it is somewhat less evident in earliest memories (i.e., current 55

PAGE 56

levels of environmental mastery and self-accep tance are not linked to these two themes in earliest memories). One difference between the two age groups is that young and late middleaged adults current autonomy is related differently to their earlie st and recent memory content. Aside from this small difference, these findings suggest generally that curre nt self-attributes are associated with the content of autobiographical me mories, and that this pattern is quite similar for young and late middle-aged adults. 4 Other findings that emerged as pa rt of this model but that are not central to this hypothesis will be mentioned briefly. First, age group was a significant predictor of memory content: Younger adults tended to recall memories that included higher levels of positive relations ( B = 1.32, SEB = 0.54), personal growth ( B = 0.94, SEB = 0.37) and environmental mastery content ( B = 1.49, SEB = 0.70) than late middle-aged adults. S econd, the two-way interaction between age group and memory type was significant for positive relations and self-acceptance content of memories (see Step 4 in Tables 3-6 and 3-11). Pearsons correlations between memory type (earliest, recent) and memory content (posi tive relations, self-acce ptance) were conducted separately for young and late middle-aged adults to follow-up these significant interactions. Results showed that late middle-aged adults report higher levels of posit ive relations content in recent than earliest memories, r (256) = .19, p < .05, but there was no such relation for young 4 The six hierarchical regression analys es conducted to test hypothesis 4 were replicated with current self-attribute (e.g., current autonomy) as the criterion variable instead of a predictor. After supporting hypothesis 4 by finding that current self-attributes significantly predict memory cont ent, another aim was to te st whether memory content predicts current self-attributes as well. In these reversed set of regression analyses, sex was included in the initial step of each analysis. In the second step, age group, memory type and memory content (e.g., autonomy content of memories) were entered. In the third st ep, two-way interaction terms were entered, and in the final step the threeway interaction was entered. Results showed that all six memory content themes signifi cantly predicted current selfattributes (e.g., autonomy content in memories significantly predicted current levels of autonomy). This suggests that there is a significant bi-directional relation between cu rrent self-attributes and memory content as predicted by the self-memory system model (Conway et al., 2004). 56

PAGE 57

adults, r (562) = .02, p > .05. Self-acceptance shows the same pattern: Late middle-aged adults self-acceptance content was higher in recent than earliest memories, r (257) = .19, p < .05, but young adults self-acceptance content was not related to memory type, r (564) = .00, p > .05. Other minor findings include the relation of sex and memory type to the content of memories. Sex was a significant predictor of only self-acceptance content (see Step 4 in Table 311). Men tended to have higher self-acceptance c ontent in their memories than women, B = 1.20, SEB = 0.58. Memory type was a significant predictor of purpose in life content ( B = 2.34, SEB = 0.33), personal growth content ( B = 1.03, SEB = 0.34), autonomy content ( B = 1.66, SEB = 0.36) and environmental mastery content ( B = 1.44, SEB = 0.63). Recent memories were more likely to include all of these content themes than earliest memories. 57

PAGE 58

Table 3-1. Descriptive sta tistics for current self-a ttributes in young and late middle-aged adults Young Late Middle-Aged Current Self-Attribute M SD M SD Positive Relations 56.52 10.61 60.25 9.60 Purpose in Life 56.91 9.26 58.83 9.67 Personal Growth 58.07 7.50 58.60 9.33 Autonomy 50.84 8.14 58.02 9.33 Environmental Mastery 51.27 9.29 57.35 9.99 Self-Acceptance 54.64 11.71 58.70 11.05 Note. The maximum possible score for each subscal e was 72 (nine items per subscale measured on an 8-point Likert scale). Middle-aged partic ipants scored significan tly higher on all current self-attributes except Personal Growth. Figure 3-1. Illustration of effects and their weights in mediati on models. (A) Direct effect of age group on the six self-attributes (B) Indirect effect of age group on the six selfattributes through mediators. 58

PAGE 59

Table 3-2. Pearson correlation coefficients between the six self-attributes, the Future Time Perspective Scale, the Rappaport Time Line and age group Measure 1 2 3 4 5 6 7 8 9 1. Positive Relations 2. Purpose in Life .52** 3. Personal Growth .53** .61** 4. Autonomy .32** .37** .42** 5. Environmental Mastery .62** .61** .45** .48** 6. Self-Acceptance .72** .59** .53** .49** .75** 7. Future Time Perspective .28** .31** .33** .04 .25** .35** 8. Rappaport Time Line -.08 -.02 .04 -.27** -.20** -.07 .53** 9. Age group .17** .10 .03 .37** .29** .16** -.43** -.83** Note. ** p < .001. Age Group: 1 = Young, 2 = Late Midlife. Table 3-3. Mediation model 1: Fu ture Time Perspective Scale as a single mediator between age group and current self-attribut es (5000 bootstrap samples) Dependent Effect of IV Effect of M Direct Indirect Total variable (DV) on M ( a) on DV ( b) effects (c' ) effects ( a*b) effects ( c ) Positive Relations 2.43* 0.40* 1.95* 0.97* 0.98* Purpose in Life 2.40* 0.37* 1.42* 0.97* 0.53* Personal Growth 2.48* 0.31* 0.98* 0.77* 0.21 Autonomy 2.46* 0.22* 2.32* 0.54* 1.79* Environmental 2.38* 0.41* 2.50* 0.98* 1.54* Mastery Self-Acceptance 2.32* 0.55* 2.31* 1.28* 1.04* Note. IV = Independent variable (i.e., age group). M = Mediator (i e., Future Time Perspective Scale). Sex was entered as a covari ate in all mediation models. p < .05. 59

PAGE 60

Table 3-4. Mediation model 2: Rappaport Time Line as a si ngle mediator between age group and current self-attributes (5000 bootstrap samples) Dependent Effect of IV Effect of M Direct Indirect Total variable (DV) on M ( a) on DV ( b) effects (c' ) effects ( a*b) effects ( c ) Positive Relations 1.31* 0.63* 1.81* 0.82 0.99* Purpose in Life 1.31* 0.65* 1.38* 0.85* 0.53* Personal Growth 1.31* 0.62* 1.02* 0.81* 0.21 Autonomy 1.30* 0.32 2.21* 0.41 1.80* Environmental 1.29* 0.42 2.08* 0.53 1.54* Mastery Self-Acceptance 1.30* 0.87* 2.18* 1.13* 1.06* Note. IV = Independent variable (i.e., age group). M = Mediator (i. e., Rappaport Time Line). Sex was entered as a covariat e in all mediation models. p < .05. Table 3-5. Mediation model 3: Future Time Perspective Scale and Rappaport Time Line as multiple mediators between age group and current self-attributes (5000 bootstrap samples) Dependent Effect of IV Effect of M Direct Indirect Total variable (DV) on M ( a) on DV ( b) effects (c' ) effects ( a*b) effects ( c ) Positive Relations Future Time 2.43* 0.41* 1.74* 0.97* 0.98* Rappaport 1.30* 0.19 0.25 Purpose in Life Future Time 2.40* 0.38* 1.21* 0.91* 0.53* Rappaport 1.31* 0.18 0.24 Personal Growth Future Time 2.47* 0.31* 1.00* 0.77* 0.21 Rappaport 1.31* 0.01 0.01 Autonomy Future Time 2.46* 0.22* 2.18* 0.54* 1.79* Rappaport 1.30* 0.12 0.16 Environmental Future Time 2.37* 0.43* 2.03* 1.01* 1.54* Rappaport 1.29* 0.41 0.53 Self-Acceptance Future Time 2.32* 0.57* 1.92* 1.32* 1.06 Rappaport 1.30* 0.34 0.44 Note. IV = Independent variable (i.e., age group). M = Mediator (i e., Future Time Perspective Scale, Rappaport Time Line). Sex was entere d as a covariate in all mediation models. p < .05. 60

PAGE 61

Table 3-6. Hierarchical multiple regression analys is predicting positive relations content of all memories (N = 791) B SE B t Step 1 Sex .00 .52 .00 .00 Step 2 Sex .92 .52 .06 1.78 Age 1.22 .54 .08 2.28* Memory Type 1.10 .49 .08 2.26* Positive Relations .17 .02 .26 7.14** Step 3 Sex .92 .52 .06 1.79 Age 1.33 .54 .09 2.46* Memory Type 1.13 .49 .08 2.34* Positive Relations .18 .03 .26 7.30** Memory Type*Positive .05 .05 .04 1.11 Age*Positive .07 .05 .05 1.37 Age*Memory Type 2.50 1.07 .08 2.35* Step 4 Sex .91 .51 .06 1.77 Age 1.32 .54 .09 2.44* Memory Type .96 .49 .07 1.94* Positive Relations .18 .03 .26 7.31** Memory Type*Positive .07 .05 .05 1.37 Age*Positive .07 .05 .05 1.37 Age*Memory Type 2.14 1.08 .07 1.99* Age*Positive*Memory .21 .11 .07 1.94* Note. R = .00 for Step 1 (p > .05); R = .07 for Step 2 (p < .05); R = .01 for Step 3 ( p <.05); R = .00 for Step 4 (p > .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 61

PAGE 62

Table 3-7. Hierarchical multiple regression anal ysis predicting purpose in life content of all memories (N = 800) B SE B t Step 1 Sex .06 .35 .01 .18 Step 2 Sex .21 .34 .02 .62 Age .02 .36 .00 .04 Memory Type 2.29 .33 .24 7.01** Purpose in Life .09 .02 .18 5.08** Step 3 Sex .15 .34 .02 .44 Age .01 .36 .00 .03 Memory Type 2.30 .33 .24 7.05** Purpose in Life .09 .02 .18 5.06** Memory Type*Purpose .03 .04 .03 .82 Age*Purpose .05 .04 .04 -1.25 Age*Memory Type 1.02 .71 .05 1.44 Step 4 Sex .15 .34 .02 .44 Age .01 .36 .00 .04 Memory Type 2.34 .33 .24 7.12** Purpose in Life .09 .02 .18 5.07** Memory Type*Purpose .03 .04 .03 .82 Age*Purpose .05 .04 .04 1.26 Age*Memory Type 1.08 .71 .05 1.52 Age* Purpose*Memory .08 .08 .04 1.09 Note. R = .00 for Step 1 (p > .05); R = .09 for Step 2 (p < .05); R = .00 for Step 3 ( p >.05); R = .00 for Step 4 (p > .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 62

PAGE 63

Table 3-8. Hierarchical multiple regression anal ysis predicting personal growth content of all memories (N = 795) B SE B t Step 1 Sex .30 .36 .03 .80 Step 2 Sex .12 .36 .01 .33 Age .96 .37 .09 2.58* Memory Type 1.03 .34 .10 2.98* Personal Growth .13 .02 .21 5.84** Step 3 Sex .16 .36 .02 .44 Age .94 .37 .09 2.54* Memory Type 1.03 .34 .10 2.99* Personal Growth .13 .02 .22 6.08** Memory Type*Personal .04 .04 .03 .89 Age*Personal .08 .04 .06 1.80 Age*Memory Type .12 .74 .01 .15 Step 4 Sex .16 .36 .02 .44 Age .94 .37 .09 2.54* Memory Type 1.03 .34 .10 2.99* Personal Growth .13 .02 .22 6.08** Memory Type*Personal .04 .04 .03 .89 Age*Personal .08 .04 .06 1.80 Age*Memory Type .12 .74 .01 .17 Age*Personal*Memory .05 .09 .02 .52 Note R = .00 for Step 1 (p > .05); R = .06 for Step 2 (p < .05); R = .00 for Step 3 ( p > .05); R = .00 for Step 4 (p > .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 63

PAGE 64

Table 3-9. Hierarchical multiple regression analysis predicting au tonomy content of all memories (N = 799) B SE B t Step 1 Sex .02 .35 .00 .06 Step 2 Sex .16 .34 .02 .46 Age .64 .39 .06 1.64 Memory Type 1.37 .34 .14 4.10** Autonomy .11 .02 .21 5.58** Step 3 Sex .16 .34 .02 .47 Age .69 .40 .07 1.72 Memory Type 1.38 .34 .14 4.13** Autonomy .11 .02 .20 5.48** Memory Type*Autonomy .02 .04 .02 .62 Age*Autonomy .03 .04 .02 .61 Age*Memory Type 1.05 .78 .05 1.35 Step 4 Sex .16 .34 .02 .46 Age .68 .40 .07 1.70 Memory Type 1.66 .36 .17 4.64** Autonomy .11 .02 .20 5.47** Memory Type*Autonomy .03 .04 .03 .85 Age*Autonomy .02 .04 .02 .60 Age*Memory Type 1.46 .80 .07 1.83 Age*Autonomy*Memory .18 .08 .08 2.16* Note R = .00 for Step 1 (p > .05); R = .06 for Step 2 (p < .05); R = .00 for Step 3 ( p > .05); R = .01 for Step 4 (p < .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 64

PAGE 65

Table 3-10. Hierarchical multiple regression anal ysis predicting environmental mastery content of all memories (N = 789) B SE B t Step 1 Sex .48 .63 .03 .76 Step 2 Sex .76 .62 .04 -1.22 Age 1.39 .69 .07 2.02* Memory Type 1.66 .60 .10 2.74* Environmental Mastery .15 .03 .17 4.66** Step 3 Sex .76 .62 .04 1.22 Age 1.48 .70 .08 2.12* Memory Type 1.68 .60 .10 2.80* Environmental Mastery .15 .03 .17 4.67** Memory Type*Environment .14 .06 .08 2.19* Age*Environment .05 .07 .03 .73 Age*Memory Type 2.65 1.36 .07 1.94 Step 4 Sex .75 .62 .04 1.21 Age 1.49 .70 .08 2.12* Memory Type 1.44 .63 .08 2.29* Environmental Mastery .15 .03 .17 4.68** Memory Type*Environment .14 .06 .08 2.15* Age* Environment .05 .07 .03 .74 Age*Memory Type 2.24 1.40 .06 1.60 Age*Environment*Memory .18 .14 .05 1.32 Note R = .00 for Step 1 (p > .05); R = .04 for Step 2 (p < .05); R = .01 for Step 3 ( p < .05); R = .00 for Step 4 (p > .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 65

PAGE 66

Table 3-11. Hierarchical multiple regression analys is predicting self-acceptance content of all memories (N = 789) B SE B t Step 1 Sex .78 .59 .05 1.33 Step 2 Sex 1.22 .58 .07 2.11* Age 1.04 .62 .06 1.68 Memory Type .89 .56 .06 1.59 Self-Acceptance .15 .03 .21 5.97** Step 3 Sex 1.20 .58 .07 2.09* Age 1.02 .62 .06 1.64 Memory Type .94 .56 .06 1.69 Self-Acceptance .15 .02 .21 6.01** Memory Type*Self .15 .05 .11 3.10* Age*Self .00 .05 .00 .08 Age*Memory Type 2.44 1.23 .07 1.99* Step 4 Sex 1.20 .58 .07 2.09* Age 1.02 .62 .06 1.64 Memory Type .96 .57 .06 1.69 Self-Acceptance .15 .02 .21 6.01** Memory Type*Self .15 .05 .11 3.07* Age*Self .00 .05 .00 .08 Age*Memory Type 2.48 1.24 .07 2.00* Age*Self*Memory .02 .11 .01 .19 Note R = .00 for Step 1 (p > .05); R = .05 for Step 2 (p < .05); R = .02 for Step 3 ( p < .05); R = .00 for Step 4 (p > .05). p < .05, ** p < .001. Age Group: 1 = Late Midlife, 2 = Young. 66

PAGE 67

CHAPTER 4 DISCUSSION The study combined a life span developmenta l perspective (Baltes, 1997) with the selfmemory system model of autobiographical memo ry (SMS model; Conway & Pleydell-Pearce, 2000) to examine how individuals current conceptu al self guides their re call of autobiographical memories. One major goal of the study was to ex amine adult development by demonstrating age differences in young and late middle-aged adults cu rrent conceptual self: Those in late midlife reported higher levels of all self-attributes and also reported that they saw future time as more limited and as holding less opportunity. Further, findings showed that future time perspective plays a significant role as a suppressor in th e relation between age group and current selfattributes. Understanding age di fferences in the conceptual se lf was important in adding a developmental perspective to the original SMS model. Another major goal was to test the SMS model. The study found support for the model using recent memories, but also challenged the model with a nove l and stringent test examining the link between the current conceptual self and earliest childhood memories. The SMS model provides a general consideration of the relation of the conceptual self to memory content, without notice that different types of memories (e.g., very distant vs. recent) may show different patterns of relations. The current study showed that the relation between the current conceptual self and memory content is less evident in earl iest memories than recent memories. Although it is less evident, there is still a link between the curr ent conceptual self and earliest memories. This finding contributed to the debate on the emergence of autobiographical memories (i.e., offset of childhood amnesia). Findings also extended th e SMS model by showing that young and late middle-aged adults, although they have different content in the conceptual self (i.e., future time 67

PAGE 68

perspective and self-a ttributes), show the same pattern of relation between their current conceptual self and content of their autobiographical memories. These findings are discussed in more detail be low in two sections: (i) differences in the conceptual self in young adulthood and late midlif e, and (ii) the relation between current selfattributes and autobiographical memory content as a test of the self-memory system model (Conway & Pleydell-Pearce, 2000) and as a contribution to the de bate on why earliest memories are recalled. Study limitations are then discussed, and a final section provides some overall conclusions. Adult Development and the Conceptual Self: Future Time Perspective and Self-Attributes The conceptual self is a major compone nt of the self-memory system (Conway & Pleydell-Pearce, 2000). The current study examined two aspects of the con ceptual self expected to show developmental variation (Ryff, 1989b; Staudinger & Bluck, 2001): future time perspective and self-attri butes (see Figure 1-1). As well as dem onstrating age differences in these two aspects, the study examined whether future time perspective mediated the relation between ones age group and current self-attributes (see the first three boxes in Figure 1-2 for the conceptual model). Relation of Age Group to Future Time Perspective The study examined future time perspective as an aspect of the current conceptual self likely to show variation across adult age groups. Individuals in late midlife reported a less openended future time perspective th an young adults on both future ti me perspective measures: the Rappaport Time Line (Rappaport et al., 1985) which is a strai ghtforward measure of ones subjective sense of position in the life span, as well as the Future Time Perspective Scale (Carstensen & Lang, 1996) which as sesses subjective time left to liv e (similar to Rappaport Time Line), but also includes optimism about the future Individuals in late midlife are not only aware 68

PAGE 69

that they are temporally closer to the end of th eir lives, but perceive the future as having fewer opportunities than young adults. Thes e findings support classic theories suggesting that midlife is governed by an internal social clock that reminds individuals that they are in the middle and that life has an ending (Jung, 1933; Neugarten, 1996). These findings are also in line with other st udies documenting changes in time perspective across the life span. Midlife is the developmenta l stage when awareness of the finitude of life emerges for the first time, and particularly late midlife is the phase when a shift to a more limited future time perspective occurs (Heckhausen, 2001). For example, Gould (1972) asked adults to rate the item There is still plenty of time to do mo st of the things I want to do and showed that endorsement of this item dropped between ages 35 and 40, and then remained stable throughout the 50s. Cate and John (2007) showed that women in their twenties focused more on opportunities for the future than either women in early-midlife (40-49 years) or late-midlife (5060 years). As such, it appears that late midlife is a time when individuals recognize that the future is limited, even as compared to those in early midlife (Cate & John, 2007). The current study embraced this two-phase perspective on midlif e, focusing on late midlife as a time when individuals are likely to have crossed the boundary of realizing th at there is now more life lived than left to live (Neugarten, 1996). This is in line with socioemoti onal selectivity theory (Carstensen, 1993, 1995). The theory suggests that young adults have an expansive future time perspective leading them to hold knowledge-oriented goals, whereas older individuals have a constrained future time perspective leading th em to focus on emotional goals (Fredrickson & Carstensen, 1990; Lang & Carstensen, 1994). Mid life is the developmental stage when time perspective is changing and the crossover be tween knowledge-oriented and emotion-oriented goals occurs (Carstense n & Turk-Charles, 1994). 69

PAGE 70

All of these findings, including those of the cu rrent study, support the conceptualization of midlife as a period in which time perspective shif ts: Adults realize that time will eventually run out and that life does not present limitless opportun ities. This sense of time perspective is one aspect of the conceptual self that varies acr oss adulthood. The next section focuses on age group differences in current self-attributes. Relation of Age Group to Current Self-Attributes Study findings also demonstrate developmenta l differences in another aspect of the conceptual self: ones current se lf-attributes. Largely as predicted, those in late midlife report higher levels of autonomy, environmental mast ery and purpose in life than young adults. Individuals in young adulthood face the Eriksonian developmental task of forming a coherent identity and resolving the inti macy-isolation crisis. Middle-aged adults have moved beyond a focus on their self and finding a partner, to broa den their commitments to family, work, society and future generations (Erikson, 1959; McAd ams, 2001). Midlife involves many family and work roles and duties that need to be juggled (H eckhausen, 2001). As such, it is unsurprising that middle-aged individuals show greater autonomy in decision-making, stronger feelings that they have an impact on their environment, and a grea ter sense of purpose concerning how their life is proceeding. These findings are also in line with theories of adult personality development (Erikson, 1959; Gould, 1978; Levinson, 1978) and pr evious empirical research (Ryff, 1989a) suggesting that midlife is characterized by consider able complexity. It is a time when adults must move nimbly and competently ahead in a va riety of domains of life (Heckhausen, 2001; Neugarten, 1973). Previous research suggests that midlife is in some ways the peak of life, when individuals have skills such as the ability to handle stress, be productive, show social responsibility (Lachman, Ziff, & Spiro, 1994), and excel in both everyday problem solving and life-problem solving (Heckhausen, 2001; Staudinger & Bluck, 2001). 70

PAGE 71

The current demonstration of higher autonomy, environmental mastery, purpose in life in late midlife fit well with previous theory and l iterature. Past research has been inconsistent, however, about age group differences in two self-attributes: positive relations with others and self-acceptance. Some studies have found no age differences, whereas others showed that older adults report greater positive relations and self-acceptance than both young and middle-aged adults (Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). In order to more clearly examine the differences between young and middle-aged adul ts, the current study purposely sampled a late middle-aged group which was expected to be more similar to older adu lts. Indeed, the late middle-aged group in this study was similar to older adults, showing higher levels of positive relations with others and greater self-acceptance than young adults. This suggests that late middle-aged adults are more positive in both self-evaluation (self-acceptance) and otherevaluation (positive relati ons with others) than young adults. The self-acceptance findings are in line with previous research showing that discrepa ncies between the actual and the ideal self are greater for young adults than for middle-aged individuals (Okun, Dittburner, & Huff, 2006). The positive relations findings are consistent with research showing that middle-aged adults are more likely to focus on emotionally meaningful goals in social relations (Fung, Carstensen, & Lang, 2001) and more likely to attend to and recall positive information and positive events than young adults (Charles, Mather, & Carstensen, 2003). Havi ng positive relations with ones self and with others bodes well for those in midlife. Positive re lationships (especially with family and friends) are the strongest predicto r of middle-aged adults current well-being, wher eas daily activities are the strongest predictor for young a dults (Ryff & Heidrich, 1997). In sum, these findings help to clarify past work that showed mixed results, potentially due to the wide age range previously 71

PAGE 72

used to define midlife. The current study demo nstrates that younger adults report less positive relations with both self and others than do those in late midlife. Note that there were also exceptions to the predicted pattern of developmental differences. Young adults were expected to have higher levels of purpose in life and personal growth than middle-aged people. This was based on life span developmental theory suggesting that young adulthood is the developmental stage when allocation of resources are directed towards growth (Baltes, 1987, 1997). The theory s uggests there are three aspects of life span development: growth, maintenance (including rec overy), and regulation of loss (Baltes, 1997; Staudinger, Marsiske, & Baltes, 1995). Each of these is present in all phases of life. In the second half of life, however, due to an overall age-related decline in biological and sensory-cognitive resources, there is a gradual shift in the proportion of gains to losses such that losses become greater than gains by late adulthood (Baltes, 1987, 1997; Labouvie-Vief, 1981). Thus, young adults were expected to show more purpose in li fe and personal growth because they are in a developmental stage focused on growth and gain, particularly as compared to those in late midlife. Past research shows that young (18-29 years) an d middle-aged adults (30-64 years) express higher personal growth and purpose in life than do older adults (over 64 years; Ryff, 1989b; Ryff, 1991; Ryff & Keyes, 1995). We had expected that, because the study particularly focused on late midlife, the pattern for young as compared to older adults would be replicated in the current sample of late middle-aged adults. This was not the case: this late middle-aged sample reported higher purpose in life than young adults, and there were no age differences in personal growth. Thus, it appears that wh ile older adults may show lowe r levels of purpose and growth orientation, these attributes remain strong across both early and late midlife. Ebner, Freund and 72

PAGE 73

Baltes (2006) showed a similar pa ttern of results for personal goa l orientation: Young adults had a stronger orientation to wards growth than towards maintena nce of resources or regulation of losses and declines. Middle-aged adults show ed an increase in goa l orientation toward maintenance and prevention of losses, but like in the current study, still reported a primary focus on growth. This suggests that although adults may start expecting losses and considering ways to maintain current resources in late midlife, they experience personal growth at levels similar to young individuals and have more purpose in life than the younger generations. In conclusion, current self-attri butes are a part of ones conceptual self that vary according to ones age group. Those in late midlife s how higher scores on a variety of positive and advantageous characteristics including autonomy, self-accep tance, environmental mastery, purpose in life and positive relations (Ryff, 1989a; 1989b; Keyes & Ryff, 1999). Research also shows that life satisfaction conti nues to increase towards the en d of midlife (approximately age 65; Lachman, Rcke, Rosnick, & Ryff, 2008; Mr oczek & Spiro, 2005). Together, these findings suggest that midlife is not a peri od of decline or crisis as per cu rrent stereotypes, but a period in which an individual has developed many positive self-attributes. In contrast, young adulthood is the period of learning and practic ing the necessary knowledge and sk ills to master adult life (Staudinger & Kunzmann, 2005), with the expectation that through effort and experience one can achieve higher future and ideal levels of th e six self-attributes (Ryff, 1991; Ryff & Heidrich, 1997). Findings reviewed thus far show that those in late midlife differ from younger adults in two aspects of the conceptual self (i.e., future time perspective and self-attributes). Note that these aspects provide a complex picture of the con ceptual self in different stages: Future time perspective is becoming less open in time, but individuals are showing high levels of positive 73

PAGE 74

self-attributes in late midlife. In contrast in young adulthood, the future appears open and limitless, but the individual displays lower levels of positive self-attributes. The next section examines this complex set of relations in more detail to further understand development in young adulthood and late midlife. The Role of Age and Future Time Perspectiv e in Predicting Current Self-Attributes. Many psychological outcomes vary by age grou p, and chronological age is an important and widely-used variable in life span developmenta l research. It is also, however, a rather clumsy and abstract variable as it represents only lived time since birth in a numerical sense. For example, although the study showed th at age group is related to curr ent self-attributes, there may be variables related to age that are more proximal to self-attributes and are better able to explain why individuals in different age gr oups have different levels of th ese attributes. In the conceptual model developed for this study (see Figure 1-2), future time perspective was added as a proximal psychological variable that might elucidate why adu lts in late midlife have higher levels of these positive self-attributes than those in young adulthood. Two measures of future time perspective were used to assess both subjective sense of position in the life span (Rappaport Time Line; Rappaport et al., 1985), and feelings about how open-ended the future is and wh at opportunities it may hold (Fut ure Time Perspective Scale; Carstensen & Lang, 1996). As such, three assessment s of ones place in th e life span were made: objective chronological age, sense of life span position, and feelings about the future. Mediation analyses examined whether the tw o future time perspective measures were important in the pathway between objective ch ronological age and se lf-attributes. Indeed, findings showed that time perspective had a sign ificant impact, but it wa s a suppression effect rather than a mediation effect (Davis, 1985). That is, instead of decreasing the validity of age group for predicting current self-attributes, futu re time perspective in creased the predictive 74

PAGE 75

validity of age group (i.e., future time perspective suppressed the criterion-irrelevant variance in age group) and this occurred for all of the six self-attributes. Note that the Future Time Perspective Scale, assessing feelings that th e future is open and o ffers opportunities, best demonstrated this suppression effect. That is, wh en the two future time perspective measures were assessed as multiple mediators in one mode l, the Future Time Perspective Scale attenuated the effect of the Rappaport Time Line and acted as the only suppressor for all six self-attributes. As this multiple mediators model is the most informative and parsimonious model, findings from only this model are interpreted below. This s uggests that subjective lo cation in the life span (Rappaport Time Line) is not cruc ial in the relation betw een age group and self-a ttributes. It is an awareness of ones place in the life span in co mbination with optimistic feelings about the future that shows a meaningful effect on th e relation between age gr oup and self-attributes.5 The suppression effect occurred in the same manner across all six self-attributes: Age group is negatively associated with optimistic fe elings about the future. That is, young adults have a more open-ended and positive view of the future than late middle-aged adults. Holding such positive feelings about the future is related to having higher levels of positive self-attributes independent of ones age That is, regardless of whether one is in young adulthood or midlife, having an optimistic and open future time perspective is positively associated with higher levels of all self-attributes. Finally, age group is also positively related to all six self-attributes, and the inclusion of positive feelings about the future into the model bolsters this relation of age group to 5 When the Future Time Perspective Scale was tested as a single mediator, it acted as a suppressor for all six selfattributes, whereas when the Rappaport Time Line was tested as a single mediator, it did not act as a suppressor for positive relations, autonomy and environmental mastery. For purpose in life, personal growth and self-acceptance, both the Rappaport Time Line and the Future Time Perspectiv e Scale act as single suppressors. That is, these selfattributes are affected by both subjective location in th e life span and positive view of future, whereas positive relations, autonomy and environmental mastery are not affected by subjective location in the life span. 75

PAGE 76

the self-attributes. That is, when optimistic future orientation is accounted for, age group shows an even stronger relation to self-attributes (through the suppr ession of criterion-irrelevant variance in age group which makes age group a purer predictor of self-attributes). These findings suggest that chronological age and future tim e perspective act as opposing forces for both age groups. For late middle-aged adults, chronological age acts as an advantage and provides them with high levels of self -attributes (gained through maturation and life experience), however their limited future time pers pective diminishes the positive effect of their age on their self-attributes. That is, late middleaged adults less open future perspective works against them and attenuates the self-attribute-enhancing effect of their current age. When future time perspective is accounted for, however, the negative effect of their limited perspective disappears, and their age shows its actually st ronger positive effect on self-attributes. This indicates a compensatory effect of chronological age: Late middle-aged adults may benefit from coping strategies to compensate for their limited fu ture view and to eliminate its negative effects on their self-attributes. To the ex tent that late middle-aged adults can compensate for their negative future time perspective, they can enha nce their self-attributes more. Some unmeasured factors associated with being middle-aged may be helping them compensate for the realization that they are closer to death with fewer opportunities ahead of them. This interpretation is in line with the mode l of selective optimization with compensation (SOC model; Baltes & Baltes, 1990; Freund & Ba ltes, 2000; Marsiske, Lang, Baltes, & Baltes, 1995). The SOC model, suggests that selecti on, optimization and compensation are three strategies inherent in any developmental process. Successful development requires an efficient coordination of these three strategies to re spond to available resources by focusing on the maximization of growth (gains) and the minimiza tion of losses across the life span (Marsiske et 76

PAGE 77

al., 1995). Late midlife is a period when selectio n and compensation are important to maintain adequate levels of functioning and to continue to experience gains in some domains (Baltes & Carstensen, 1996; Marsiske et al ., 1995; see also Brandtstdter & Greve, 1994). In the second half of life, adults begin to more freque ntly use compensation or an accommodative coping strategy (Brandtstdter & Renner, 1992) by adjusting their personal goals and standa rds to their actual functioning levels. This strategy is especially used when losses a nd declines related to aging are uncontrollable or irreversible such as the realization of the finitude of life and a shorter time left to live (Wentura & Brandtstdter, 2003 ). Late middle-aged adults may begin to use compensation strategies such as positive rea ppraisals (i.e., seeing the positive sides of developmental losses such as a more limited time perspective) in order to maintain their high levels of positive self-attributes. Research shows that although such secondary control strategies are not preferred by young adults and are negatively related with young adults well-being, they are more frequently used by middle-aged peopl e and are positively related with middle-aged peoples well-being (Wrosch, Heckhausen, & Lachman, 2000). This suggests that such control strategies gained through aging help midlife adults to compensate for their diminishing time perspective and keep their self-attributes high. This notion that individuals can control agerelated changes through psychological mechanisms is in line with previous research showing that a majority of Americans believe that they have control over many do mains of life, including aging and its negative aspects. For example, 84% of Americans believe that there are things they can do to control the agi ng process (Lachman, 2006). For young adults, the same opposing forces are evident, but in th e opposite direction. Young adults have a more open-ended and positive fu ture view than older adults, but they lack the accumulated life experiences that come with age. As chronological age is a much stronger 77

PAGE 78

predictor of current se lf-attributes than posit ive feelings about the future, young adults are disadvantaged and they have lower levels of self-attributes due to their young age. Unlike middle-aged adults, they cannot compensate for th eir future time perspective which is already limitless and positive. They just need to go through the natural stages of development and age in order to reach the self-attribute leve ls of adults in midlife. As pr evious research has shown (Ryff, 1989a, 1991) growing older and potentially gaini ng life experience (Clark-Plaskie & Lachman, 1999) increases ones positive relations with others, self-acceptance, sense of autonomy, environmental mastery and purpose in life. This s uggests that ones curren t age determines their levels of self-attributes, and a positive future vi ew can further enhance thes e attributes. All of the self-attributes accumulate with age and maturation (potentially due to life experience), but in addition, having a positive and optimistic future time perspective is associat ed with higher levels of attributes. Young adults are in an early develo pmental stage in which they have not benefited from such life experiences or maturation, but they realize that they have an open-ended future in which to still grow and enhance their self-attributes. Note however, that regardless of their current ag e, all individuals benefit from feeling that the future is positive and open-ended. This reflects the fact that not every middle-aged person has a narrow and pessimistic time perspective, and similarly not every young person has a very openended time perspective. In sum, chronological age has a stronger im pact on current self-attributes than positive feelings about the future, but bot h young and late middle-aged people benefit from an open-ended and positive future time perspective to enhance their self-attr ibutes even more. It seems like these attributes are strongly shaped by the past and how much has been experienced already, but they are also influen ced by feelings about the future and how much time there is left to live. 78

PAGE 79

Suppression effects in relatio n to predicted results. As discussed above, the suppression effect of future time perspective was evident for all six self-attributes, however in relation to the study predictions, this effect chal lenged the hypotheses in differe nt ways. Thus, having reviewed the actual study findings, it seemed prudent to simply mention how they differ from expected effects (see Figure 4-1). Ther e were no specific study hypothese s concerning the relation of future time perspective to self-attributes, but the study had expected to find a negative relation between future time perspective and positive relati ons and self-acceptance. Note, however, that a more open-ended future time perspective was posi tively (not negatively) as sociated with adults feelings that they have positive relations with others and their degree of self-acceptance (see Figure 4-1A). Regarding purpose in life a nd personal growth, it was hypothesized that chronological age would be nega tively associated with these two self-attributes. It was found, however, that being in the late midlife group was positively related to ha ving a purpose in life and experiencing personal growth (see Figure 4-1B). Finally, autonomy and environmental mastery were not expected to be influenced by futu re time perspective at all, but results revealed a positive relation between future time perspect ive and these two self-attributes (see Figure 41C). Summary: Adult Development and the Conceptual Self Individuals in late midlife report higher levels of a number of positive self attributes, and a shortening view of time left to live in rela tion to their younger counter parts. Although current age is a better predictor of self-attributes than future time perspective, to the extent that individuals feel that the future is open and offers opportunity, this sense of time is also related to more positive self-attributes. In combination, these findings contri bute to the adult development literature, particularly the relatively sparse literature on midlife development (Lachman & Bertrand, 2001). The findings highlight how both maturation in time, and optimism about the 79

PAGE 80

future enhance ones current sense of self. Ryff (1989a) has interpreted these positive selfattributes as components of ps ychological well-being. Indeed, high levels of autonomy, selfacceptance, environmental mastery, purpose in life and positive relations represent advantageous adult characteristics. This sugge sts that adults in late middleage experience higher well-being than young adults (Ryff, 1989a; 1989b; Keyes & Ryff, 1999). Despite stereotypes, research shows that aging is not related to negative aff ect, depression or decreases in well-being or life satisfaction (Brandtstdter, Wentura, & Greve, 1993; Westerhof & Barrett 2005). In particular, middle-aged adults constitute the powerful group in the society who are the norm-bearers and the decision makers who to some extent control th e other age groups (Neugarten, 1973). They have many roles which help to increase their sense of responsibility and self efficacy (Clark-Plaskie & Lachman, 1999). Although midlife is also a time of high stress due to multiple roles and duties and little time for leisure, middle-aged adults peak in competence and ability to handle stress (Clark-Plaskie & Lachman, 1999). In sum, alt hough adults in midlife have more limited and negative views of the future than young adults, they still have more positive self-attributes than young adults, and they can use certain strategies to compensate for their shortened time left to live, and thereby enhance thei r self-attributes even more. Findings discussed thus far have demonstrated that individu als in young adulthood and late midlife show differences in both time perspective and current self-attribute s, important features of the conceptual self. The next section discusse s whether these different conceptual selves are represented in the content of autobiographical me mories that people have about themselves in recent times and in early childhood. Testing the Self-Memory System Model The discussion thus far has focused on adult de velopment, particularly the relation of age to the make-up of adults conceptual self (i.e., future time perspective and self-attributes). In 80

PAGE 81

particular, the outcome variable of interest in the mediation m odels was current self-attributes, which was expected to shape memory content as per the self-memory system model (Conway & Pleydell-Pearce, 2000). Recall that the SMS model predicts that the conceptual self shapes autobiographical remembering (s ee last two boxes in Figure 1-2). This study tested the selfmemory system model by investigating the relati on between the current conceptual self and the memory content of both recent memories and earliest childhood memories. Little previous research has tested the model (Conway & Holmes, 2004; Ross & Wilson, 2002; Woike, 1994a; 1994b). None of the previous stud ies examined the effect of th e conceptual self on earliest childhood memories, and none have taken an adult developmental perspective, comparing the relation of self to memory in different age group s. The current study addressed these gaps in the literature. Relation of Current Self-Attr ibutes to Memory Content The SMS model suggests that i ndividuals current conceptual self guides the retrieval of individual memories, thereby shap ing the content of what is cu rrently recalled (Conway et al., 2004). This claim was supported: people who curre ntly report having higher levels of positive relations, purpose in life, personal growth, auto nomy, environmental mastery and self-acceptance recall memories that contain higher levels of th ese six themes than people who score lower in these self-attributes. Independent of ones curr ent age (i.e., young adult vs. late middle-aged adult) and the type of memory recalled (i.e., ea rliest vs. recent), indi viduals current selfattributes are reflected in the content of their memories. The Reconstructive Nature of Autobiographical Memory These findings support the reconstructive natu re of autobiographical memory (Conway, 2005) on which the self-memory system model (Conway et al., 2004) is based. The model suggests that autobiographical me mories are not stored in the brain as holistic and static 81

PAGE 82

representations of the past, but are constructed at retrieval from abstract representations of our past experiences (Mace, 2007). Memories are constr ucted during retrieval in accordance with the current goals and needs of the rememberer (B luck, Alea, & Demiray, in press). They are produced not just by a memory system but an integrated self-memory system. That is, memories emerge from the intersection of two competi ng demands: the need to record accurate goal activity and reflect reality (i.e., adaptive corres pondence), and the need to maintain a coherent and stable record of the selfs interaction with the world (i.e., self c oherence; Conway et al., 2004). Due to this reconstructive, integrative memory system, human beings have a tendency to gently shape the recall of past ev ents to confirm their current identity, thereby maintaining selfcoherence (Wilson & Ross, 2003). In the current study, for example, a highly autonomous person when asked to recall a memory tends to recall a time in which they felt autonomous. A person currently low in autonomy may have fewer high autonomy memori es to choose from, but also retrieves memories that are consistent with the current conceptual self which has low autonomy. The SMS model illustrates how the conceptual self affects the content and structure of retrieved memories, but why are memories constructed during retrieval in accordance with the current goals and characteristics of the self ? According to the functional approach to autobiographical memory (Bluck, 2003), they ar e reconstructed so that they are useful and functional for the self in the contexts they are re called in daily life (Bluck et al., in press). That is, memories are not retrieved without reason. They are remembered because the self is in a context where it is necessary to bring to mind me mories that are somehow relevant and useful in meeting current environmental demands. According to the functional appr oach, there are three ecological functions of memories: self (i.e., devel opment and continuity of the self), social (i.e., developing, maintaining and enhanc ing social relationships) and di rective (i.e., guiding present 82

PAGE 83

and future behavior; Bluck, Alea, Habermas & Rubin, 2005; Pillemer, 1992). All of these functions serve the current self so that the self can adapt to the environment better and operate better. Thus, in daily life, indivi duals recall memories that fit with their current self-attributes so that they maintain their self coherence, and use these memories for certain purposes such as presenting themselves to others in social intera ctions. The current findings suggested that, for example, a highly autonomous mi ddle-aged person is more lik ely to recall autonomy-related memories in daily life, which wi ll confirm and enhance his or he r current autonomy level. These autonomy-related memories will serve a self functi on and implicitly remind that person that he or she is really an autonomous person. Simila rly, a less autonomous young person might be recalling memories that carry a w eaker autonomy theme, which repres ents the current self of the young person. In sum, the current study found support for th e SMS model by showing an effect of selfattributes on memory content overall (without di stinguishing between different memory types). This effect makes more ecological sense wh en a functional approach is taken towards autobiographical memory (Bluck et al., in press). The SMS model shows how the current conceptual self shapes the retrieval of autobiographical memories. Future research could further integrate the functional approach (Bluck, 2003) to the SMS model to also shed light on why the current conceptual self shapes th e recall of memories (i.e., toward what adaptive ends). The next section provides an interpretation of the more stringent test of the SMS model that distinguished between earliest and recent memories. Doing so revealed a slightly different picture. Stringent Test of the SMS Model Us ing Earliest Childhood Memories Another study aim was to potentially expa nd the SMS model through examining findings from a more stringent test of the model usi ng earliest childhood memories. The study examined whether the general relati on between memory content and curren t self that was observed with all 83

PAGE 84

of the six self-attributes was equally observed for recent and earliest memories. For environmental mastery and self-acceptance, ther e was a relation between current self-attributes and memory content in recent memories but not earliest memories. These findings suggest that the relation between current self-attributes and memory content is more evident in recent memories. Across age groups, all six self-attributes were positively related to the content of recent memories, however only four self-attributes were linked with the content of earliest memories. This difference be tween the relation of the conceptual self to earliest and recent memories both supports and challenges the SMS model. According to the model, earliest memories are theorized to be retr ieved because they are the first memories that match ones current self-attributes. The current fi ndings show that they do match on some selfattributes, but not on as many as recent memories do. That is, the extent of the association between the conceptual self and retrieved memori es is different across memories from different lifetime periods. There are two plausible reasons for these diffe rences. One possibility is that adult selfattributes, particularly environmental mastery and self-acceptance, are less likely to occur during the experience of childhood events. Note that self-attribute themes, at the mean level, are lower in earliest than in recent memo ries. For example, earliest memories are less likely to include purpose in life, personal growth, autonomy and environmental mastery content than recent memories. This suggests that these characterist ics may be less strongly experienced in early childhood events, which tend to be simpler events, and are of course ther eby less evident when these events are recalled. Research shows that earliest memories contain less information in different categories (e.g., emotion, setting, context, color) and fewer details, and that earliest memory narratives are less complete than recent memories (Howes et al., 1993; West & Bauer, 84

PAGE 85

1999; Westman & Orellana, 1996). This is in line with the SMS model that suggests that the goals of the self in childhood are very different than self goals in adul thood. That is, childrens goals are most probably associated with less comp licated needs that might not fit perfectly with adult goals. In particular, environmental mastery and self -acceptance may be characteristics that may not yet have developed in the conceptual self of a four or a five-year-old. According to Erikson (1959), by the time children are three years old, th ey have sufficiently resolved the trust vs. mistrust conflict, and the autonomy vs. self-doubt conflict. These two stages seem to be the stages to develop positive relations, autonomy, gr owth and purpose in life, but it might be too early for one to have environmental mastery a nd self-acceptance skills. Environmental mastery and self-acceptance may still be developing as they might require a mo re advanced, even a reflective self-concept. For example, Levine (200 4) suggests that the cognitive self that emerges at around age two is a very basic fo rm of self awareness that is de monstrated by the mirror test in which children detect the sticker on their head upon looking in a mirror (Amsterdam, 1972). A more advanced temporally-extended self awar eness emerges at around ag e four. For example, although two-year-old children can retain event knowledge for long time periods, this knowledge is fragmentary, cue-dependent and inconsistent. Although their even t knowledge persists, it does not become integrated into their self before the age of four. Due to this primitive self awareness before the age of four, earliest memories that are recalled from around the ag e of four and a half might be lacking self-acceptance and environm ental mastery that may be developing later. The second possibility is simply the age of the memory. The age of the current participants earliest memories range from 17 to 51 years, whereas recent memories are at most one year old. Very old memories tend to become more fixed and stable over time and are more 85

PAGE 86

likely to be reproduced in an unvarying and schematized format (Anderson, Cohen, & Taylor, 2000). In contrast, recent memories may be mo re likely to be dynamically reconstructed according to the current conceptual self. Thus, earliest memories may be less affected by the current conceptual self. Future iterations of th e SMS model will need to elucidate how different types of memories may be dealt with differen tly in the self-memory system, and to further elaborate the role that time and recall across lifetime periods, play in the strength of associations between the conceptual self and retrieved memories. In sum, the relation between current self-attri butes and memory content is less evident in earliest memories than in recent memories. There is less of a match between the current conceptual self and the content of earliest memori es (encoded in a distant developmental stage) as compared to recent memories (encoded and re trieved in the same developmental stage; e.g., young adulthood for young adults) and the same lifet ime period (e.g., the college years for young adults). Thus, the current study provided some refinements to the SMS model by testing it with earliest childhood memories. Does the SMS Model Explain Why Adul ts Recall Their Earliest Memories? Findings on the relation between earliest memories and the cu rrent conceptual self also contribute to solving the puzzle co ncerning why adults are able to recall very distant memories such as their earliest childhood memories. Why a dults recall certain childhood memories and not others remains a mystery (Peterson, Grant, & Boland, 2005). The current study contributes a potential solution to this puzzl e by showing that both young and late middle-aged adults recall earliest memories that are linked with their curren t conceptual self. According to the SMS model, adults cannot recall memories from the first year s of life because of the discrepancy between goals that were active in the self during the encoding of memo ries (i.e., in early childhood) and the goals and characteristics that are currently active in the adult dur ing retrieval (Conway & 86

PAGE 87

Pleydell-Pearce, 2000). Young child rens goals are most probably associated with basic needs for nurturance, healthy attachment as well as gaining autonomy (Erikson, 1959). As implied above, memories in early childhood tend to be enc oded in line with these basic goals which have almost no overlap with ones goals in adulthood. If there is little continuity between encoding and retrieval goals, then effective retrieva l cannot occur (Conway & Pleydell-Pearce, 2000). According to this view, earliest memories represen t the first events in life that fit both with the self at encoding and with ones current conceptual self. Neissers (1962) view is in lin e with the SMS approach to ea rliest memories, and with the current findings. According to Neisser (1962), co gnitive structures change with development; therefore memories from childhood do not fit with adult schemata and cannot be brought to mind. Research shows that children and adults ha ve different retrieval mechanisms such that children recall things that would not seem memorable to adults (Davis et al., 2008; Fivush, Gray, & Fromhoff, 1987; Usher & Neisser, 1993). They might recall uninteresting or trivial details of events such as (What did you do on the camping trip ?) First we eat dinne r, then go to bed, and then wake up and eat breakfast (Fivush & Hamond, 1990, p. 231). In sum, young children and adults perceive and retrieve the world differently (West & Bauer, 1999). The current study found support for this view by assessing the current self during retrieval, as well as retrospectively assessing the conditions duri ng encoding. Conditions during re trieval are not taken into consideration by the cognitive self account (How e & Courage, 1993) or the social-interaction account (Haden et al., 1997; Nelson, 1993). In fact, it is unlikely that any one of these accounts is adequate on its own; therefore recent theoretical formulations have emphasized an eclectic approach to the study of earliest memories (R eese, 2002; Wang, 2003). We contributed to this 87

PAGE 88

approach by finding empirical support for the SMS account of earliest memories for the first time in the literature. Another goal of the study was to examine whet her the SMS model adequately explains the relation of the conceptual self to memory cont ent in individuals from different age groups. The next section compares young and late middle-aged adults in terms of the relation between their current self-attributes and memory content. Expanding the SMS model with a Life Span Development Perspective Though the SMS model does present a complex representation of the relation between the self and autobiographical memory, it does not cons ider how developmental processes may affect this relation. A contribution of the current study was to examine whether theorized processes in the SMS model hold across two adult age groups Given that the model does not rely on processes expected to change fr om young adulthood to late midlif e, the overall relation between the conceptual self and memory content was not expected to, and did not, differ across young and late middle-aged adults. Although this is a cross-sectional study, th is finding suggests that the relation between current con ceptual self and memory may be stable across age groups. As mentioned above, young and late middle-aged adults show differences in the nature of their conceptual self (with middle-aged adults higher in many positive self-attributes). In spite of such age differences in content of the current conceptual self, there are no age differences in the relation between the current self and memory cont ent. So age may shape how adults see themselves in the present, but does not have an impact on the way the conceptual self affects retrieval of autobiographical memories. This is in line with the life span deve lopment theory which describes life span development as multidimensional and multidirecti onal. Baltes (1997) suggests that different domains of development should be examined indi vidually as there might be declines in some 88

PAGE 89

domains, stability in others, and others may show growth with ag e (e.g., Brandtstdter & Wentura, 1995). The current finding s indicate growth in a variety of current self-attributes from young to late midlife, but stabil ity in the relation between curr ent self-attributes and memory content across young adulthood and late midlife. Showing developmental stability in the relati on between current self and memory content is an important extension of the SMS model. Gi ven the overarching constr ucts of the model (e.g., conceptual self, autobiographical knowledge base), these processes are postulated as universal and would in fact be expected across the entire a dult life span. This would require examination in samples that also include additional younger (e.g., adolescence) and older adults. For example, there is evidence showin g that across various samples (e.g., university students, middle-aged individuals, celebrities) and on a variety of dimensions, people believe their past selves to be inferior to their presen t self (Wilson & Ross, 2000, 2001), and they tend to recall past selves that are consonant with their curre nt self (Wilson & Ross, 2003). As mentioned above, one function of autobiographical memories is that they help people maintain self-continuity (Bluck & Alea, 2008; Wilson & Ross, 2003). The fact that the SMS model shows developmenta l stability at least in the first half of life, may provide an unde rpinning for maintaining such continuity. This continuity might begin to break down in very old age, especially the four th age (Baltes, 1997) or under dementia conditions (e.g., Alzheimers dis ease) when the self-memory system is under threat due to neurological change. Future rese arch might focus on the re lation between self and memory in both the normal and impaired aging mind. Note that the above discussion focuses on the age of the participant. Age invariance also occurs, however, for the age of the memory. You ng and late middle-aged adults current selfattributes are similarly related to both recent and earliest memories. Both age groups recent 89

PAGE 90

memories are from the previous year so it make s sense that recent memories would be uniformly related to current self-a ttributes across age groups. Late middl e-aged adults earliest memories, however, are much more distant (i.e., about 34 years earlier) than young adults earliest memories, but are still related to the current c onceptual self. Research shows that by the time children are 14 years old, they ha ve adult-like childhood amnesia whic h is quite stable from that point on (Peterson, Wang, & Hou, 2009). As both th e young and late middle-aged adults in the current sample are much older than 14, they are experiencing childhood am nesia similarly, and it appears that their first memories after amnesia offsets, are similarly related to their current conceptual self. Future research should exam ine memories from other ages (e.g., reminiscence bump memories, i.e., memories from ages 1030; Demiray, Glgz, & Bluck, 2009) to examine whether adults from different ages demonstrat e the same relation between their current selfattributes and these memories. In conclusion, the stringent te st of the SMS model with earliest memories both extended the model by showing that some modifications might be necessary for explaining the retrieval of different memories, and contribut ed to solving the puzzle of w hy earliest memories are recalled. The model was also extended with an adult developmental focus. Young and late middle-aged adults who are currently in di fferent developmental stages an d who show differences in the content of their current conceptual selves, display a similar relation between their current selfattributes and memory content (for both earliest and recent memories). Limitations The current study had several limitations that could be addressed in future research. These were related to participant recruitment, online data collection, and the cross-sectional design of the study. Issues concerning conceptuali zation of the complexity of adult development, 90

PAGE 91

and the ability to test the SMS model (Conway & Pleydell-Pearce, 2000) in a single study are also discussed. Participant Recruitment: Sampling Adults from Late Midlife Recruiting individuals in midlife to participate in research is notoriously difficult (Nosek et al., 2002), and this was no different in the curre nt study. Researchers contacted 501 middle-aged people, but only 42% of them started the survey and of those 70% completed both parts of the study. Basic analyses were conducted to show th at those who began but did not complete the survey were no different from t hose who did complete. It is not possible, however, to know if the obtained sample is different fr om those who were invited but did not participate, or how representative they are of the general late middle-aged population. For example, since this was an online study, those who did not start the survey might not be very familiar or comfortable with computers, though all had an email addr ess through which they were contacted. Such people might not be employed, not have easy access to computers, or might have lower education levels and socio-economic status. These adults might show lower levels of selfattributes (e.g., environmental mastery) and pot entially a less open-ended future time perspective than adults who are highly educated and active in society. For example, Markus, Ryff, Curhan, and Palmersheim (2004) showed that college-educat ed adults in midlife scored higher on nearly every self-attribute th an high school-educated middle-aged a dults. Note that 75% of our latemidlife sample had an education level of a Bachelors degree or higher, and 81% were employed, which suggests that our sample was re latively highly educ ated and active. Though this might be a bias, note that obtained age differences are in comparison with young adults who are also enrolled to obtain a college degree. T hus, education level might account for absolute levels but would not account for age differences in current self-attributes In addition, though it is possible that those who did not pa rticipate were hindered by computer literacy, it is unlikely that 91

PAGE 92

such a large percentage of middleaged individuals are computer i lliterate given current rates of computer ownership and literacy in society and in the workplace. An alternative possibility is that those who did not start the survey might be extremely active, busy individuals who simp ly did not feel they had time to participate in research (Heckhausen, 2001). One could speculate that thes e adults might show ev en higher levels of some self-attributes such as environmental mast ery, autonomy, and purpose in life than those in the current sample. Finally, the type of incentive offered might have been a factor in adults decision to participate. Those adults who were not interested in the compensation offered in the current study (i.e., an informative handout on mid life development, and a donation to either the American Association of Retired Persons or the Childrens Health Fund), or who did not find it sufficient, might have not participated in th e study. These individuals may be intellectually different from those who are curious and enthusiastic about read ing about their own developmental stage, or they may be different in terms of generativity leve ls. Midlife, even when conceptualized in two stages with a focus just on late midlife shows great variation in terms of demographic differences, life circumstances a nd everyday routines (Staudinger & Bluck, 2001). As such, any non-representative sample of midlif e, including the current study, must be cautious in generalizing to the wider population. Online Data Collection Data were collected through an online survey at SurveyMonkey.com. Internet-based data collection is an increasingly popular method which has been co mpared with paper and pencil surveys and shown to be psychometrically sim ilar (e.g., internal consistency; Chang, 2005; Chuah et al., 2006). Procedures outlined earlier were put in place to address foreseeable issues related to the use of this me dium. One limitation, however, might be the lack of researcher presence during survey completion and the lack of standardization of the conditions in which 92

PAGE 93

participants completed the survey. That is, particip ants were free to take the survey wherever and whenever they wished within a certain amount of time. Both of these issues were taken into account to some extent through da ta quality checks. If particip ants took too long to do the survey, it was clear they were not focusing or we re distracted, and they were removed from the sample. Individuals who were identified as not taking the survey seriou sly (e.g., did not provide correct responses to foil items, had large am ounts of missing data, did not provide memory narratives) were also excluded. Ho wever, there may be other aspe cts of the uncontrolled nature of online data collection (e.g., location, lighting, noise, time of day) that had an impact on our findings. Overall, however, online data collection a ppeared to be an efficient way of collecting data, especially from a difficult-to-access sample such as middle-aged people. Another shortcoming of online data collection is that online surveys need to be brief so that participants are willing to complete them in the absence of a researcher who can encourage or prompt them to finish the survey. While co mpleting an online survey at home, participants may be more likely to drop out if the survey is too long. Therefore, some items and scales were excluded from the current study in order to ke ep the survey short. This prevented the measurement of other variables of interest that might be rela ted to the outcomes (e.g., other potential mediators in th e relation between age group and self-attributes). Cross-Sectional Design All age group differences in the current study are cross sectional fi ndings; therefore they might represent cohort differences (Schaie, 1993) rather than actual developmental differences. That is, the age differences in future time pers pective and self-attributes might be cohort-based. Effectively the study compares the me gene ration of baby boomers to young adults born between 1980 and 1990. It is unclear whether coho rt might influence peoples self-conceptions and sense of the future. In addition, our findings on the relation between the current conceptual 93

PAGE 94

self and memory content indicated developmen tal stability across th e two age groups. This relation, however, should be investigated in a longitudinal design in future research to examine whether within-subject stability occurs acr oss time, preferably over the life span. Issues Related to Conceptualizing the Complexity of Adult Development Only one possible mediator between age gr oup and current self-attributes was tested, although there might be many other processes in the complicated relation between a persons age and how they perceive their own self-attributes. Future research should te st the role of other possible mediators such as self-c oncept clarity (i.e., having a self -concept that is clearly defined and internally consistent). Research shows that older people report hi gher self-concept clarity (Bluck & Alea, 2008), and self-c oncept clarity is positively rela ted to psychological well-being (Diehl, 2006), as well as self esteem and positive aff ect (Diehl, Jacobs, & Hastings, 2006). Based on these findings, one might expect that self-c oncept clarity would medi ate the relation between age group and self-attributes. That is, late mi ddle-aged adults may have higher self-concept clarity than young adults, which in turn would lead them to have higher levels of current positive self-attributes than young adults. I nvestigating the role of such po ssible mediators, especially in more complex multiple mediators models or in structural equation modeling analyses, is important in terms of more fully understa nding the dynamics of adult development. Issues Related to Conceptualizing the Complexity of the SMS Model The self-memory system model (Conway & Pl eydell-Pearce, 2000) is complex and even the conceptual self includes many aspects (e.g., pe rsonal scripts, possible selves), most of which were not examined in the current study. The research focused on only two aspects of the conceptual self, chosen because they were likely to show adult developmental variation: future time perspective and self-attributes. Showing that current self-attributes are reflected in memory content provides support for the SMS model, but is clearly not sufficien t to fully validate the 94

PAGE 95

model or to represent th e complex nature of this system. That is, the model proposes that current self-attributes are not the only as pects of the conceptual self th at shape memory content: Other attitudes, traits, possible selves and other self-characteristics are thought to pl ay a role. In addition, during retrieval self-attri butes interact with other aspect s of the conceptual self, with the autobiographical knowledge base, and with th e episodic memory system as well as being regulated by the current goals of the working self. The current study focused on a very specific part of this complex network of relations. Future research may help to more fully elucidate the relation between the self and autobiographica l memory, providing further tests of the SMS model. One clear limitation related to that just described, is that the study did not thoroughly examine the current goals of the working se lf. According to the SMS model, although autobiographical memories are the product of the interaction between all three components of the self-memory system (i.e., episodic memory sy stem, long-term self and working self), the working self is a very important component. Th e working self is similar to Baddeleys (1986) working memory in the sense that it includes control processes which coordinate and modulate other separate systems: in doing so it prioritizes action within the individu als complex hierarchy of goals (Conway & Pleydell-Pearce, 2000). Th at is, the working self organizes the psychological present, which is the current e xperience, through the ca tegorization, evaluation and prioritization of its goals (Conway et al., 200 4) and recruits goal-relevant autobiographical memories that will be of use in the current context. Therefore, it is important to examine the effect of the working self on the conceptual self For example, in this study, one explicit goal of the participants working self was to pay atte ntion to the study instructions and recall two specific autobiographical memories that said something about them as a person. This goal, 95

PAGE 96

however, was not the only active goal in the self-memory system at that time: the system includes many explicit and implicit goals simultaneously (Woike et al., 2001). In addition to this explicit goal that the study instructions imposed on the participants, they might be carrying other explicit goals such as receiving course credit for their participation in th e study, or implicit goals such as being a conscientious person and completing the survey properly. Thus, it is difficult to capture this complex nature of the self-memory system and measure the existing goal hierarchy of a person at any given point in time. Another problem related to the working self is that goals are viewed as processes, which are difficult to measure and quantify. Future research should measure peoples current goals th rough self reports, as well as giving people certain goals through experimental manipulation and analyze their autobiographica l memories to see the effect of experimentally contro lled goals on retrieval. Conclusions The study integrated life span developmenta l theory (e.g. Baltes, 1997) with the selfmemory system model of autobiographical me mory (Conway & Pleydell-Pearce, 2000) to provide a revised conceptual model in whic h adult development was included. The study contributes to the adult development literatur e (especially the scarce midlife development literature) by demonstrating diffe rences in the conceptual self across age groups: Those in late midlife were found to have higher levels of self-attributes such as autonomy, environmental mastery, purpose in life, self-acceptance and pos itive relations with ot hers than young adults. Both adults current age and thei r feelings about the future were shown to be related to their reports of current self-attributes That is, self-attributes are a ffected by both time lived and time left to live, which is a good example of the impo rtance of taking a life span perspective while examining development. Note that although the conceptual self differed across age groups, the relation of the conceptual self to memory c ontent remained the same. These findings provided 96

PAGE 97

important examples of the multidimensionality a nd multidirectionality of life span development (Baltes, 1997): There are gains, losses and stability in all stages of development. In young adulthood, people perceive the future as infinite and full of possibilities to grow, but they do not have an accumulated and rich set of skills and self-attributes yet. In contrast, in late midlife, people have limited and less positive feelings abou t the future, but they have gathered positive self-attributes through maturation and life experien ce. That is, each developmental stage has its own advantages and limitations. The study also provided support for the SMS model (Conway & Pleydell-Pearce, 2000) claiming that the conceptual self guides retrieva l of autobiographical memories, which has been much cited but little research ed to date. Beyond support, the st udy also suggests refinements to the model, such as the focus on development but al so concern for the relati on of the current self to memory when retrieving different types of me mories. Earliest memories were less related to current self-attributes than recent memories, sugg esting that the SMS model could be extended to take into account retrieval of diffe rent types of memories, in this case, distant memories. At the same time, the match between ones current self and ones earliest memory on several attributes suggests that the SMS model provides a reasonab le explanation of why earliest events are recalled. Despite its limitations, the study offers new findings and suggest s new directions for research on the changing self in young adulthood a nd late midlife, and provides new insights on how our current self affects wh at we recall in daily life. 97

PAGE 98

Figure 4-1. Illustration of suppre ssion effects in relation to predicted results. (A) Positive relations and self-acceptance. (B) Purpose in life and personal gr owth. (C) Autonomy and environmental mastery. 98

PAGE 99

APPENDIX A MEMORY NARRATIVE INSTRUCTIONS We are interested in hearing about memories fr om your life! We will be asking you to tell us about TWO SPECIFIC MEMORIES By specific we mean that the memory can be about any event/experience but it must have occurred at a particular place and a particular time. That is, it is not a general time in your life, but must be a memory of one specific time (it may have lasted minutes, or hours, but was not longer than one day ). It may be a unique experience or it might be just an everyday event, but what is important is that it is a memory of something that occurred at a pa rticular time and place. There are no better or worse, or right or wrong memories. We are interested, however, in you sharing specific memories that "SAY SOMETHING ABOUT YOU AS A PERSON". 1) Rest of the Instructions Sp ecific to Earliest Memories: Lets get started. Think about your early childhood, and try to remember the very earliest event/experience of your life. Please make sure this is the very first memory that you can recall (and that you actually remember it, you did not just see it in a phot o or hear the story from a family member). Most people's earliest memories come from BEFORE the age of 7. Also, dont forget that this memory should so mehow "say something about you as a person". 2) Rest of the Instructions Specific to Recent Memories: Please think about your life over the last year and remember an event/experience from your life that occurred over the last year but not ve ry recently, that is NOT in the last three months. So, recall any event that occurred before last summer but was in the last year. Also, dont forget that this memory should someho w "say something about you as a person". Please describe your memory as thoroughly as possible including all you remember about where the event took place, who was present, what happened, and what you were feeling, perceiving and thinking at that time. Please type your memory into the box below. Try to fill the entire box, but when you reach the end of the box (when it starts scrolling dow n) please wrap up your paragraph and do not keep writing. 99

PAGE 100

APPENDIX B MEMORY NARRATIVE EXAMPLES 1) Earliest Memory Narrative of a Young Adult I remember the first Christmas after my litt le brother was born, when I was just about 3 years old. It happened back at our house in Ta mpa, FL and both of my parents, my 2-monthold little brother, and myself were all pres ent. I still remember being very excited about having a little brother and being able to celebra te Christmas with him for the first time. I can remember coming down the stairs that morning dressed in my Teenage Mutant Ninja Turtles pajamas and just being so happy and excited for the day. It was a great moment and really is the earliest moment I can remember where my entire family was together and having a great time. 2) Recent Memory Narrative of a Young Adult This past summer I was in NYC for about a week. I have a lot of different memories but the one I am describing only lasted a couple hours. On my second day of being in NYC I visited my grandma who was recently moved to an assi sted living facility of Manhattan. My mom, little brother, and myself all visited her. I had not seen my grandma in over 4 years so it was just great seeing her again, but it was also sad just seeing her in the type of condition she was in as well. It is always great to be around family and your loved ones, especially after not seeing them for a long time, so it's definitely something I'll remember for a long time. 3) Earliest Memory Narrative of a Late-Midlife Adult My friend and I walked to the corner store al one. We bought some candy and then left to go home. We walked up the street but continued past the street wh ere we were supposed to turn. We walked and walked and were lost. An old la dy invited us in and ga ve us cookies. I didn't eat any because I thought they might be poisoned. I was scared that I would never get back home. 4) Recent Memory Narrative of a Late-Midlife Adult In April of this year I went to visit two of my children who now live in Texas. I had not seen my son for a year or my daughter since the pr evious Christmas. I was looking forward to seeing them but also a little apprehensive be cause my son is a "rec overing addict" but had recently stopped keeping in touch so I wasn't sure what I was going to find when I got there. I was also feeling a lot of excitement because I was bringing my youngest son with me so we could all be together but neith er of them knew that. I was imagining the joy they would get and I would get when the surprise was reveale d. My daughter picked us up at the airport and actually started screaming when her brother ran up behind her as she was hugging me. It was a great moment! 100

PAGE 101

APPENDIX C SELF-ATTRIBUTES MEASURE (RYFF, 1989A) Autonomy Subscale 1. I am not afraid to voice my opinions even wh en they are in opposition to the opinions of most people. 2. My decisions are not usually influenced by what everyone else is doing. 3. I have confidence in my opinions even if th ey are contrary to the general consensus. 4. Being happy with myself is more important than having others approve of me. 5. I tend to worry about what other people think of me. 6. I often change my mind about de cisions if my friends and family disagree. 7. It is difficult for me to voice my own opinions on controversial matters. 8. I tend to be influenced by people with strong opinions. 9. I judge myself by what I think is important, not by the values of what others think is important. Environmental Mastery Subscale 10. I am quite good at managing the many responsibilities of my daily life. 11. I generally do a good job of taking care of my personal finances and affairs. 12. I am good at juggling my time so that I can fit everything in that needs to be done. 13. I have been able to build a home and a lifestyle for myself that is much to my liking. 14. I do not fit very well with the people and the community around me. 15. I often feel overwhelmed by my responsibilities. 16. I have difficulty arranging my life in a way that is satisfying to me. 17. In general, I feel that I am in ch arge of the situation in which I live. 18. The demands of everyday life often get me down. Personal Growth Subscale 19. I think it is important to have new experi ences that challenge how you think about the world. 20. I have the sense that I have developed a lot as a person over time. 21. I am not interested in activities th at will expand my horizons. 22. I don't want to try new ways of doing things my life is fine the way it is. 23. When I think about it, I haven't really im proved much as a person over the years. 24. I do not enjoy being in new situations that require me to change my old familiar ways of doing things. 25. There is a truth in the saying that you can't teach an old dog new tricks. 26. For me, life has been a continuous pro cess of learning, changing and growth. 27. I gave up trying to make big improvements or changes in my life a long time ago. 101

PAGE 102

Positive Relations with Others Subscale 28. Most people see me as loving and affectionate. 29. I enjoy personal and mutual conversations with family members or friends. 30. People would describe me as a giving person, willing to share my time with others. 31. I know that I can trust my friends and they know that they can trust me. 32. I often feel lonely because I have few clos e friends with whom to share my concerns. 33. I don't have many people who want to listen when I need to talk. 34. It seems to me that most other people have more friends than I do. 35. Maintaining close relationships has b een difficult and frustrating for me. 36. I have not experienced many warm and trusting relationships with others. Purpose in Life Subscale 37. I am an active person in carrying out the plans I set for myself. 38. I enjoy making plans for the future and working to make them a reality. 39. I tend to focus on the present, because the future nearly alwa ys brings me problems. 40. My daily activities often seem trivial and unimportant to me. 41. I don't have a good sense of what it is I am tr ying to accomplish in life. 42. I used to set goals for myself, but that now s eems a waste of time. 43. I sometimes feel I have done a ll there is to do in life. 44. Some people wander aimlessly through life, but I am not one of them. 45. I live life one day at a time and dont really think about the future. Self-Acceptance Subscale 46. I have made some mistakes in the past, but f eel that all in all everything has worked out for the best. 47. The past had its ups and downs, but in general I wouldn't want to change it. 48. When I compare myself with friends and acquaintances, it makes me feel good about who I am. 49. In general, I feel confident and positive about myself. 50. I feel that many of the people I know have got more out of life than I have. 51. In many ways, I feel disappointed about my achievements in life. 52. My attitude about myself is probably not as positive as most people feel about themselves. 53. I like most aspects of my personality. 54. When I look at the story of my life, I am pleased with how things have turned out. 102

PAGE 103

APPENDIX D MEMORY CONTENT RATINGS FOR EARLIEST AND RECENT MEMORY NARRATIVES EARLIEST MEMORY NARRATIVE You have just written down your earliest memory. Please think of that sp ecific earliest childhood memory while answering the following ques tions. We dont want you to answer how you generally felt as a child but really to specifi cally think about how you felt in this memory. Go back in your mind to the time of this memor y. Remember what you were feeling, thinking and doing during this specific memory as you answer each of the following questions. Feel free to use the entire scale from 1 8 (1= strongly disagree, 8= strongly agree). There are no wrong answers. How much do you agree with each of these statements about you in this specific memory? 1. In this memory, I was not afraid to voice my thoughts even if othe rs might disagree. 2. In this memory, being happy with myself was more important that getting others approval. 3. In this memory, it was difficult for me to express myself. 4. In this memory, I was influenced by strong others. 5. In this memory, I was feeling in charge of my situation. 6. In this memory, I was feeling overwhelmed. 7. In this memory, I was having difficulty arrang ing things in a way that felt satisfying. 8. In this memory, I was feeling I fit very well with the people or community around me. 9. In this memory, I was learning, changing or growing. 10. In this memory, I was not interested in learning. 11. In this memory, it was important fo r me to have new experiences. 12. In this memory, I was not trying to make improvements or changes. 13. In this memory, I was feeling lonely because there were few peopl e I felt close to. 14. In this memory, I did not have peop le who wanted to listen to me. 15. In this memory, I felt I trusted people and they trusted me. 16. In this memory, other people were se eing me as loving and affectionate. 17. In this memory, I was actively doing something I wanted to do. 18. In this memory, what I was doing seemed trivial and unimportant to me. 19. In this memory, I didnt have a good se nse of what I was trying to accomplish. 20. In this memory, I was living each day and not thinking about the future. 21. In this memory, I was feeling c onfident and positive about myself. 22. In this memory, I felt everything was turning out for the best. 23. In this memory, my attitude about myself was not as positive as most other people. 24. In this memory, I was feeling disappointed with myself. 103

PAGE 104

RECENT MEMORY NARRATIVE You have just written down a memory from the la st year. Please think of that specific memory while answering the following questions. We dont want you to answer how you have generally felt about yourself in the last year, but really to specifically think about how you felt in this memory. Go back in your mind to the time of this memory. Remember what you were feeling, thinking and doing during this specific memory as you answer each of the following questions. Feel free to use the entire scale from 1 8 (1= strongly disagree, 8= strongly agree). There are no wrong answers. How much do you agree with each of these statements about you in this specific memory? 1. In this memory, I was not afraid to voice my thoughts even if othe rs might disagree. 2. In this memory, being happy with myself was more important that getting others approval. 3. In this memory, it was difficult for me to express myself. 4. In this memory, I was influenced by strong others. 5. In this memory, I was feeling in charge of my situation. 6. In this memory, I was feeling overwhelmed. 7. In this memory, I was having difficulty arrang ing things in a way that felt satisfying. 8. In this memory, I was feeling I fit very well with the people or community around me. 9. In this memory, I was learning, changing or growing. 10. In this memory, I was not interested in learning. 11. In this memory, it was important fo r me to have new experiences. 12. In this memory, I was not trying to make improvements or changes. 13. In this memory, I was feeling lonely because there were few peopl e I felt close to. 14. In this memory, I did not have peop le who wanted to listen to me. 15. In this memory, I felt I trusted people and they trusted me. 16. In this memory, other people were se eing me as loving and affectionate. 17. In this memory, I was actively doing something I wanted to do. 18. In this memory, what I was doing seemed trivial and unimportant to me. 19. In this memory, I didnt have a good se nse of what I was trying to accomplish. 20. In this memory, I was living each day and not thinking about the future. 21. In this memory, I was feeling c onfident and positive about myself. 22. In this memory, I felt everything was turning out for the best. 23. In this memory, my attitude about myself was not as positive as most other people. 24. In this memory, I was feeling disappointed with myself. 104

PAGE 105

APPENDIX E FUTURE TIME PERSPECTIVE SCALE (CARSTENSEN & LANG, 1996) In order to indicate your ag reement with the items, pleas e use the following scale: 1 2 3 4 5 6 7 Very Very Untrue true 1. Many opportunities await me in the future. 2. I expect that I will set ma ny new goals in the future. 3. My future is filled with possibilities. 4. Most of my life lies ahead of me. 5. My future seems infinite to me. 6. I could do anything I want in the future. 7. There is plenty of time left in my life to make new plans. 8. I have the sense that time is running out. 9. There are only limited possi bilities in my future. 10. Over the years, I am beginning to experience time as more limited. 105

PAGE 106

APPENDIX F DEMOGRAPHIC QUESTIONS 1) Date of birth: Month _________ Day _________ Year _________ 2) Sex (check one): Male _____ Female ______ 3) Race / Ethnicity (check one): ________ Caucasion ________ African American ________ Asian ________ Other: _______________________ ________ Hispanic 4) What is the highest de gree you have received? _______ Elementary school ________ Masters degree _______ High school diploma / GED ________ Doctorate degree _______ Associates degree ________ Other: ____________________ _______ Bachelors degree 5) Are you currently employed? ________ Full time employment ________ Part time employment ________ Retired ________Unemployed 106

PAGE 107

REFERENCES Abbott, R. A., Ploubidis, G. B., Huppert, F. A., Kuh, D., Wadsworth, M. E., & Croudace, T. J. (2006). Psychometric evaluation and predictive validity of Ryff's psychological wellbeing items in a UK birth cohort sample of women. Health and Quality of Life Outcomes, 4. doi: 10.1186/1477-7525-4-76 Alea, N., & Bluck, S. (2003). Why are you telling me that? A conceptual model of the social function of autobiographical memory. Memory, 11, 165-178. doi:10.1080/741938207 Allport, G. W. (1961). Pattern and growth in personality New York, NY: Holt, Rinehart & Winston. Amsterdam, B. (1972). Mirror self-image reactions before age two. Developmental Psychobiology, 5 (4), 297-305. Anderson, S., Cohen, G., & Taylor, S. (2000). Re writing the past: Some factors affecting the variability of personal memories. Applied Cognitive Psychology, 14 435-454. doi:10.1002/1099-0720(200009)14:5<435::A ID-ACP662>3.0.CO;2-B Baddeley, A. D. (1986). Working memory. Oxford, England: Clarendon Press. Baltes, P. (1987). Theoretical propositions of life-span developmental psychology: On the dynamics between growth and decline. Developmental Psychology, 23 (5), 611-626. Baltes, P. (1997). On the incomplete architectu re of human ontogeny: Selection, optimization, and compensation as foundation of developmental theory. American Psychologist, 52 (4), 366-380. Baltes, P. B., & Baltes, M. M. (1990). Psychol ogical perspectives on successful aging: The model of selective optimizat ion with compensation. In P. B. Baltes & M. M. Baltes (Eds.), Successful aging: Perspectives from the behavioral sciences (pp. 1). New York: Cambridge University Press. Baltes, M., & Carstensen, L. (2003). The proces s of successful aging: Selection, optimization and compensation. In U. M. Staudinger & U. E. R. Lindenberger (Eds.), Understanding human development: Dialogues with lifespan psychology (pp. 81-104). Dordrecht Netherlands: Kluwer Academic Publishers. Baltes, P., Staudinger, U., & Lindenberger, U. (1999). Lifespan psychology: Theory and application to inte llectual functioning. Annual Review of Psychology, 50 471-507. doi:10.1146/annurev.psych.50.1.471 Baron, R. M., & Kenny, D. A. (1986). The moderato r-mediator variable di stinction in social psychological research: con ceptual, strategic, and st atistical considerations. Journal of Personality and Social Psychology, 51 1173-1182. 107

PAGE 108

Bluck, S. (2003). Autobiographical memory: E xploring its functions in everyday life. Memory, 11(2), 113-123. doi:10.1080/741938206 Bluck, S., & Alea, N. (2008). Remembering be ing me: The self continuity function of autobiographical memory in younger a nd older adults. In F. Sani (Ed.), Self continuity: Individual and collective perspectives (pp. 55-70). New York, NY: Psychology Press. Bluck, S., Alea, N., & Demiray, B. (in press). You get what you need: the psychosocial functions of remembering. In J. Mace (Ed.), The act of remembering: Toward an understanding of how we recall the past. Bluck, S., Alea, N., Habermas, T., & Rubin, D. (2005). A tale of three functions: The selfreported uses of autobiographical memory. Social Cognition, 23 (1), 91-117. doi:10.1521/soco.23.1.91.59198 Bluck, S., & Habermas, T. (2000). The life story schema. Motivation and Emotion, 24, 121-147. doi:10.1023/A:1005615331901 Brandtstdter, J., & Greve, W. (1994). The ag ing self: Stabilizing and protective processes. Developmental Review, 14 (1), 52-80. doi:10.1006/drev.1994.1003 Brandtstdter, J., & Renner, G. (1992). Copi ng with discrepancies be tween aspirations and achievements in adult development: A dual-process model. In L. Montada, S. H. Filipp & M. J. Lerner (Eds.), Life crises and experien ces of loss in adulthood (pp. 301-319). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Brandtstdter, J., & Wentura, D. (1995). Adjustment to shifting possibility frontiers in later life: Complementary adaptive modes. Compensating for psychologic al deficits and declines: Managing losses and promoting gains (pp. 83-106). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Brandtstdter, J., Wentura, D., & Greve, W. (1993). Adaptive resources of the aging self: Outlines of an emergent perspective. International Journal of Behavioral Development, 16(2), 323-349. Retrieved from PsycINFO database. Carstensen, L. L. (1993). Motivation for social contact across the life span: A theory of socioemotional selectivity. Nebraska Sy mposium on Motivation, 1992: Developmental perspectives on motivation (pp. 209-254). Lincol n, NE: University of Nebraska Press. Carstensen, L. L. (1995). Evidence for a life-s pan theory of socioemotional selectivity. Current Directions in Psychological Science, 4 151-156. doi:10.1111/1467-8721.ep11512261 Carstensen, L. L., & Lang, F. R. (1996). Future Time Perspective Scale. Unpublished manuscript. Stanford University. Carstensen, L. L., & Turk-Charles, S. (1994). Th e salience of emotion across the adult life span. Psychology and Aging, 9 (2), 259-264. doi:10.1037/0882-7974.9.2.259 108

PAGE 109

Cate, R. A., & John, O. P. (2007). Testing models of the structure and development of future time perspective: Maintaining a focus on opportunities in middle age. Psychology and Aging, 22 186-201. doi:10.1037/0882-7974.22.1.186 Chang, T. (2005). The validity and reliability of student ratings: Comparison between paperpencil and online survey. Chinese Journal of Psychology, 47 113-125. Charles, S., Mather, M., & Carstensen, L. (2003) Aging and emotional memory: The forgettable nature of negative images for older adults. Journal of Experimental Psychology: General, 132(2), 310-324. doi:10.1037/0096-3445.132.2.310 Chuah, S. C., Drasgow, F., & Roberts, B. W. (2006). Personality assessment: Does the medium matter? Journal of Research in Personality, 40 359-376. doi:10.1016/j.jrp.2005.01.006 Clark-Plaskie, M., & Lachman, M. (1999). The sense of control in midlife. Life in the middle: Psychological and social development in middle age (pp. 181-208). San Diego, CA US: Academic Press. Cobanoglu, C., B. Warde, & Moreo, P. J. ( 2001). A comparison of mail, fax and web-based survey methods. International Journal of Market Research, 43 441-455. Conway, M. (2005). Memory and the self. Journal of Memory and Language, 53 (4), 594-628. doi:10.1016/j.jml.2005.08.005 Conway, M., & Holmes, A. (2004). Psychosocial st ages and the accessibility of autobiographical memories across the life cycle. Journal of Personality, 72(3), 461-480. doi:10.1111/j.0022-3506.2004.00269.x Conway, M. A., & Pleydell-Pearce, C. W. ( 2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107 261-288. doi:10.1037/0033-295X.107.2.261 Conway, M. A., & Ross, M. (1984). Getting what you want by revising what you had. Journal of Personality and Social Psychology, 47 738-748. doi:10.1037/0022-3514.47.4.738 Conway, M. A., Singer, J. A., & Tagini, A. (2004). The self and au tobiographical memory: Correspondence and coherence. Social Cognition, 22 491-529. doi:10.1521/soco.22.5.491.50768 Davis, J. A. (1985). The logic of causal order Beverly Hills, CA: Sage Publications Inc. Davis, N., Gross, J., & Hayne, H. (2008). Defining the boundary of childhood amnesia. Memory 16, 465-474. doi:10.1080/09658210802077082 Demiray, B., Glgz, S., & Bluck, S. (2008, November). Why are earliest memories so memorable? Poster session presented at the 49 th Annual Meeting of the Psychonomic Society, Chicago, IL. 109

PAGE 110

Demiray, B., Glgz, S., & Bluck, S. (2009). Examining the life story account of the reminiscence bump: Why we re member more from young adulthood. Memory, 17 (7), 708-723. doi:10.1080/09658210902939322 Diehl, M. (2006). Development of self-representa tions in adulthood. In D. Mroczek & T. Little (Eds.), Handbook of personality development (pp. 373 398). Mahwah, NJ: Lawrence Erlbaum Associates. Diehl, M., Jacobs, L., & Hastings, C. (2006). Temporal stability and authenticity of selfrepresentations in adulthood. Journal of Adult Development, 13 (1), 10-22. doi:10.1007/s10804-006-9001-4 Dixon, R. A., De Frias, C. M., & Maitland, S. B. (2001). Memory in midlife. In M. E. Lachman (Ed.), Handbook of midlife development (pp. 248278). Hoboken, NJ: John Wiley & Sons Inc. Dudycha, G. J., & Dudycha, M. M. (1941). Child hood memories: a review of the literature. Psychological Bulletin, 38 668-682. Ebner, N., Freund, A., & Baltes, P. (2006). Deve lopmental changes in personal goal orientation from young to late adulthood: From striving fo r gains to maintenance and prevention of losses. Psychology and Aging, 21 (4), 664-678. doi:10.1037/0882-7974.21.4.664 Epstein, J., Klinkenberg, W. D., Wiley, D., & McKinley, L. (2001). Insuring sample equivalence across Internet and paper-and-pencil assessments. Computers in Human Behavior, 17 339-346. doi:10.1016/S0747-5632(01)00002-4 Erikson, E. (1959). Identity and the life cycle. Psychological Issues 1, 1-171. Fivush, R., Gray, J., & Fromhoff, F. (1987) Two-year-olds talk about the past. Cognitive Development, 2 393-409. Fivush, R., & Hamond, N. (1990). Autobiographical memory across the preschool years: Toward reconceptualizing childhood amnesia. Knowing and remembering in young children (pp. 223-248). New York, NY: Cambridge University Press. Fleeson, W., & Heckhausen, J. (1997). More or less me in past, present, and future: Perceived lifetime personality during adulthood. Psychology and Aging, 12 125-136. doi:10.1037/0882-7974.12.1.125 Fredrickson, B., & Carstensen, L. (1990). Choosi ng social partners: How old age and anticipated endings make people more selective. Psychology and Aging, 5 (3), 335-347. doi:10.1037/0882-7974.5.3.335 110

PAGE 111

Freund, A., & Baltes, P. (2000). The orchestratio n of selection, optimization and compensation: An actiontheoretical conceptualization of a theory of developmental regulation. Control of human behavior, mental processes, and consciousness: Essays in honor of the 60th birthday of August Flammer (pp. 35-58). Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Fricker, R. D., & Schonlau, M. (2002). Advantages and disadvant ages of internet research surveys: Evidence from the literature. Field Methods, 14 347-367. Fung, H. H., & Carstensen, L. L. (2006). Goals change when life's fragility is primed: Lessons learned from older adults, the September 11 attacks and sars. Social Cognition, 24 248278. doi:10.1521/soco.2006.24.3.248 Fung, H., Carstensen, L., & Lang, F. (2001). Age -related patterns in social networks among European Americans and African Americans: Implications for socioemotional selectivity across the life span. The International Journal of Aging & Human Development, 52 (3), 185-206. doi:10.2190/1ABL-9BE5-M0X2-LR9V Gould, R. (1972). The phases of adult life: A study in developmental psychology. The American Journal of Psychiatry, 129(5), 521-531. Gould, R. (1978). Transformations: Growth and change in adult life Oxford England: Simon & Schuster. Haden, C. A., Haine, R. A., & Fivush, R. (1997). Developing narrative structure in parent-child reminiscing across the preschool years. Developmental Psychology, 33 295-307. doi:10.1037/0012-1649.33.2.295 Hayes, A. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408-420. doi:10.1080/03637750903310360 Heckhausen, J. (2001). Adaptation and resilience in midlife. Handbook of midlife development (pp. 345-391). Hoboken, NJ: John Wiley & Sons Inc. Helson, R., Soto, C. J., & Cate, R. A. (2006) From young adulthood through the middle ages. In T.D. Little & D.K. Mroczek (Eds.), Handbook of personality development (pp. 337-352). Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Howe, M. L., & Courage, M. L. (1993). On resolving the enigma of infantile amnesia. Psychological Bulletin, 113 305-326. doi:10.1037/0033-2909.113.2.305 Howe, M. L., & Courage, M. L. (1997). The emergence and early development of autobiographical memory. Psychological Review, 104 499-523. doi:10.1037/0033295X.104.3.499 Howes, M., Siegel, M., & Brown, F. (1993). Early childhood memories: A ccuracy and affect. Cognition, 47 95-119. doi:10.1016/0010-0277(93)90001-C 111

PAGE 112

Jung, C. G. (1933). The stages of life Modern man in search of a soul (pp. 95-114). Orlando, FL: Harcourt Brace. Keyes, C., & Ryff, C. (1999). Psyc hological well-being in midlife. Life in the middle: Psychological and social development in middle age (pp. 161-180). San Diego, CA: Academic Press. King, W. C., & Miles, E. W. (1995). A quasi-e xperimental assessment of the effect of computerizing noncognitive paper-and-pencil measurements: A test of measurement equivalence. Journal of Applied Psychology, 80 643-651. doi:10.1037/00219010.80.6.643 Labouvie-Vief, G. (1981). Proactive and reactive aspects of constructivism: Growth and aging in life-span perspective New York, NY: Academic Press. Lachman, M. E. (2006). Perceived control over aging-related declines : Adaptive beliefs and behaviors. Current Directions in Psychological Science, 15 (6), 282-286. doi:10.1111/j.1467-8721.2006.00453.x Lachman, M. E., & Bertrand, R. M. (2001). Personality and the self in midlife. In M.E. Lachman (Ed.), Handbook of midlife development (pp. 279-309). Hoboken, NJ: John Wiley & Sons Inc. Lachman, M. E., Rcke, C., Rosnick, C., & Ryff C. (2008). Realism and illusion in Americans' temporal views of their life satisfaction: Ag e differences in recons tructing the past and anticipating the future. Psychological Science, 19 889-897. doi:10.1111/j.14679280.2008.02173.x Lachman, M., Ziff, M., & Spiro, A. (1994). Main taining a sense of control in later life. Aging and quality of life (pp. 216-232). New York, NY: Springer Publishing Co. Lang, F., & Carstensen, L. (1994). Close emotional relationships in late life: Further support for proactive aging in the social domain. Psychology and Aging, 9 (2), 315-324. doi:10.1037/0882-7974.9.2.315 Leichtman, M. D., Wang, Q, & Pillemer, D. B. (2003). Cultural variations in interdependence and autobiographical memory: Lessons from Ko rea, China, India and the United States. In R. Fivush & C.A. Haden (Eds.), Autobiographical memory and the construction of a narrative self: developmental and cultural perspectives (pp. 73-98). Mahwah, NJ: Lawrence Erlbaum Associates. Levine, B. (2004). Autobiographical memory and the self in time: Brain lesion effects, functional neuroanatomy, and lif espan development. Brain and Cognition, 55 (1), 54-68. doi:10.1016/S0278-2626(03)00280-X Levinson, D. J. (1978). The seasons of a man's life New York, NY: Knopf. 112

PAGE 113

Lonsdale, C., Hodge, K., & Rose, E. A. (2006). Pixels vs. paper: Comparing online and traditional survey methods in sport psychology. Journal of Sport & Exercise Psychology, 28, 100-108. Mace, J. H. (2007). Involuntary memory: Concept and theory. In J. H. Mace (Ed.), Involuntary memory (pp. 1-19). Malden, MA: Blackwell Publishing. MacKinnon, D. (2008). Introduction to statistical mediation analysis New York, NY: Taylor & Francis Group/Lawrence Erlbaum Associates. MacKinnon, D. P., Krull, J. L., Lockwood, C. M. (2000). Equivalence of the mediation, confounding and suppression effect. Prevention Science, 1 (4), 173. doi:10.1023/A:1026595011371 MacKinnon, D., Lockwood, C., & Will iams, J. (2004). Confidence Limits for the Indirect Effect: Distribution of the Produc t and Resampling Methods. Multivariate Behavioral Research 39 (1), 99-128. doi:10.1207/s15327906mbr3901_4 Markus, H., Ryff, C., Curhan, K., & Palmersheim, K. (2004). In their own words: Well-Being at midlife among high school-educated and college-educated adults. How healthy are we?: A national study of well-being at midlife (pp. 273-319). Chicago, IL: University of Chicago Press. Retrieved from PsycINFO database. Marsiske, M., Lang, F., Baltes, P., & Baltes, M. (1995). Selective optimization with compensation: Life-span perspectives on successful human development. Compensating for psychological deficits and declines : Managing losses and promoting gains (pp. 3579). Hillsdale, NJ England: Lawrence Erlbaum Associates, Inc. Retrieved from PsycINFO database. Maslow, A. H. (1968). Toward a psychology of being (2nd ed.). New York, NY: Van Nostrand. McAdams, D. (2001). Generativity in midlife. Handbook of midlife development (pp. 395-443). Hoboken, NJ: John Wiley & Sons Inc. Mroczek, D., & Spiro, A. (2005). Change in Life Satisfaction During Adu lthood: Findings From the Veterans Affairs Normative Aging Study. Journal of Personality and Social Psychology, 88(1), 189-202. doi:10.1037/0022-3514.88.1.189 Multhaup, K. S., Johnson, M. D., & Tetirick, J. C. (2005). The wane of childhood amnesia for autobiographical and pu blic event memories. Memory, 13, 161-673. doi:10.1080/09608210344000652 Neisser, U. (1962). Cultural and cognitive discon tinuity. In T. E. Glad win & W. Sturtevant (Eds.), Anthropology and human behavior (pp. 54-71). Washington, D. C.: Anthropological Society of Washington. Nelson, K. (1993). The psychological and social origins of autobiographical memory. Psychological Science, 4 7-14. doi:10.1111/j.1467-9280.1993.tb00548.x 113

PAGE 114

Neugarten, B. L. (1973). Personality change in la te life: A developmental perspective. In C. Eisdorfer & M. P. Lawton (Eds.), The psychology of adul t development and aging (pp. 311-335). Washington, DC: American Psychological Association. Neugarten, B. L. (1996). The middle y ears. In D. A. Neugarten (Ed.), The meanings of age: Selected papers of Bernice L. Neugarten (pp. 135-159). Chicago: University of Chicago Press. Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). E-Research: Ethics, security, design and control in psychological research on the internet. Journal of Social Issues, 58 161 176. doi:10.1111/1540-4560.00254 Okun, M., Dittburner, J., & Huff, B. (2006). Perc eived changes in well -being: the role of chronological age, target ag e, and type of measure. The International Journal of Aging & Human Development, 63 259-278. doi:10.2190/87XT-R7RV-BGLF-7DHW Pealer, L. N., Weiler, R. M., Pigg, R. M. J., Mi ller, D., & Dorman, S. M. (2001). The feasibility of a web-based surveillance system to collect health risk behavior data from college students. Health Education and Behavior, 29 547-559. Peterson, C., Grant, V., & Boland, L. (2005). Ch ildhood amnesia in children and adolescents: Their earliest memories. Memory, 13, 622-637. doi:10.1080/09658210444000278 Peterson, C., Wang, Q., & Hou, Y. (2009). 'When I was little': Childhood recollections in Chinese and European Canadian grade school children. Child Development, 80 506-518. doi:10.1111/j.1467-8624.2009.01275.x Pillemer, D. B. (1992). Remembering personal ci rcumstances: A functional analysis. In E. Winograd, & U. Neisser (Eds.), Affect and accuracy in recall : Studies of flashbulb memories (Emory symposia in cognition 4, pp. 236-264). New York: Cambridge University Press. Preacher, K. J., & Hayes, A. F. (2004). SPSS and SA S procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, In struments & Computers, 36 717-731. Retrieved from PsycINFO database. Preacher, K. J., & Hayes, A. F. (2008). Asympto tic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879-891. doi:10.3758/BRM.40.3.879 Rappaport, H., Enrich, K., & Wilson, A. (1985). Relation between ego identity and temporal perspective. Journal of Personality and Social Psychology, 48 1609-1620. Reese, E. (2002). Social factors in the development of autobiogra phical memory: The state of the art. Social Development, 11 (1), 124-142. doi:10.1111/1467-9507.00190 Rogers, C. R. (1961). On becoming a person. Boston: Houghton Mifflin. 114

PAGE 115

Ross, M., & Wilson, A. E. (2002). It feels like yesterday: Self-esteem valence of personal past experiences, and judgments of subjective distance. Journal of Personality and Social Psychology, 82, 792-803. doi:10.1037/0022-3514.82.5.792 Ryff, C. D. (1989a). Beyond Ponce de Leon and life satisfaction: New directions in quest of successful ageing. International Journal of Be havioral Development, 12 35-55. Ryff, C. D. (1989b). 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 adult hood and old age: A tale of shifting horizons. Psychology and Aging, 6 286-295. Ryff, C. D., & Heidrich, S. (1997). Experience an d well-being: Exploratio ns on domains of life and how they matter. International Journal of Behavioral Development, 20 (2), 193-206. doi:10.1080/016502597385289 Ryff, C. D. & Keyes, C. L. M. (1995). The structure of psychological well-being revisited. Journal of Personality and Social Psychology, 69 719-727. doi:10.1037/00223514.69.4.719 Ryff, C. D., & Singer, B. H. (2006). Best ne ws yet on the six-factor model of well-being. Social Science Research, 35 1103-1119. doi:10.1016/j.ssresearch.2006.01.002 Schachtel, E. (1947). On memory and childhood amnesia. Psychiatry: Journal for the Study of Interpersonal Processes, 10 1-26. Schaie, K. (1993). The Seattle Longitudinal Study: A thirty-five-year inquiry of adult intellectual development. Zeitschrift fr Gerontologie, 26 (3), 129-137. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and none xperimental studies: new procedures and recommendations. Psychological Methods, 7 (4), 422. doi:10.1037/1082-989X.7.4.422 Sobel, M. E. (1982). Asymptotic confidence interval s for indirect effects in structural equations models. In S. Leinhart (Ed.), Sociological methodology 1982 (pp. 290-312). San Francisco, CA: Jossey-Bass. Staudinger, U. M., & Bluck, S. (2001). A view on midlife development from life-span theory. In M. E. Lachman (Ed.), Handbook of midlife development (pp. 3-39). Hoboken, NJ: John Wiley & Sons Inc. Staudinger, U., & Kunzmann, U. (2005). Positive Adult Personality Development: Adjustment and/or Growth?. European Psychologist 10(4), 320-329. 115

PAGE 116

Staudinger, U., Marsiske, M., & Baltes, P. (1995 ). Resilience and reserve capacity in later adulthood: Potentials and limits of development across the life span. Developmental psychopathology, Vol. 2: Risk, disorder, and adaptation (pp. 801-847). Oxford, England: John Wiley & Sons. Tabachnick, B. G., & Fidell, L. S. (2001). Computer-assisted research design and analysis Boston, MA: Allyn and Bacon. Usher, J., & Neisser, U. (1993). Childhood amnesia and the beginnings of memory for four early life events. Journal of Experime ntal Psychology, 122 (2), 155-165. doi:10.1037/00963445.122.2.155 Wang, Q. (2003). Infantile amnesia recons idered: A cross-cu ltural analysis. Memory, 11(1), 6580. doi:10.1080/741938173 Wentura, D., & Brandtstdter, J. (2003). Age stereotypes in younger and older women: Analyses of accommodative shifts with a sentence-priming task. Experimental Psychology, 50 (1), 16-26. doi:10.1026//1618-3169.50.1.16 West, T., & Bauer, P. (1999). Assumptions of in fantile amnesia: Are th ere differences between early and later memories? Memory, 7 (3), 257-278. doi:10.1080/096582199387913 Westerhof, G., & Barrett, A. (2005). Age identity and subjective well-being: A comparison of the United States and Germany. The Journals of Gerontology: Series B: Psychological Sciences and Social Sciences, 60B (3), 129-136. Retrieved from PsycINFO database. Westman, A., & Orellana, C. (1996). Only visual impressions are almost always present in longterm memories, and reported completeness, ac curacy, and verbalizabil ity of recollections increase with age. Perceptual and Motor Skills, 83 (2), 531-539. Wilson, A. E., & Ross, M. (2000). The frequency of temporal-self and social comparisons in people's personal appraisals. Journal of Personality and Social Psychology, 78 (5), 928942. doi:10.1037/0022-3514.78.5.928 Wilson, A. E., & Ross, M. (2001). From chump to ch amp: Peoples appraisals of their earlier and current selves. Journal of Personality and Social Psychology, 80, 572-584. doi:10.1037/0022-3514.80.4.572 Wilson, A.E., & Ross, M. (2003). The identity fu nction of autobiographical memory: Time is on our side. Memory, 11, 137-149. doi:10.1080/741938210 Winograd, E., & Neisser, U. (1992). Affect and accuracy in recall : Studies of 'flashbulb' memories. New York, NY: Cambridge University Press. Retrieved from PsycINFO database. Woike, B. (1994a). The use of differentiation an d integration processes: Empirical studies of separate and connected ways of thinking. Journal of Personality and Social Psychology, 67, 142-150. doi:10.1037/0022-3514.67.1.142 116

PAGE 117

117 Woike, B. (1994b). Vivid recollection as a techni que to arouse implicit motiverelated affect. Motivation and Emotion, 18, 335-349. doi:10.1007/BF02856473 Woike, B., Lavezzary, E., & Barsky, J. (2001). The influence of implicit motives on memory processes. Journal of Personality and Social Psychology, 81 935-945. doi:10.1037/00223514.81.5.935 Woike, B., Mcleod, S., & Goggin, M. (2003). Implicit and explicit motives influence accessibility to different autobiographical knowledge. Personality and Social Psychology Bulletin, 29 1046-1055. doi:10.1177/0146167203254504 Woike, B., & Polo, M. (2001). Motive-related memories: Content, structure, and affect. Journal of Personality, 69 391-415. doi:10.1111/1467-6494.00150 Wood, E., Nosko, A., Desmarais, S., Ross, C., & Irvine, C. (2006). Online and traditional paperand-pencil survey administration: Examining experimenter presence, sensitive material and long surveys. Canadian Journal of Human Sexuality 15(3-4), 147-155. Retrieved from PsycINFO database. Wrosch, C., Heckhausen, J., & L achman, M. (2000). Primary and secondary control strategies for managing health and fina ncial stress across adulthood. Psychology and Aging, 15 (3), 387-399. doi:10.1037/0882-7974.15.3.387 Zelinski, E., Burnight, K., & Lane, C. (2001). The relationship between subjective and objective memory in the oldest old: Comparisons of findings from a representative and a convenience sample. Journal of Aging and Health 13(2), 248-266. doi:10.1177/089826430101300205

PAGE 118

118 BIOGRAPHICAL SKETCH Burcu was born and raised in Istanbul, Turkey. After completing her middle and high school education at American Robert College, she attended Ko University to pursue her undergraduate degree in psychology. During her undergraduate studies she worked as Dr. Sami Glgzs research assistant and became interest ed in autobiographical memory research. She was excited to examine how and why adults recall their own life events across different phases of their lives. She decided to stay at Ko University to continue working with Dr. Sami Glgz and to pursue a masters degree in developmental psychology. Her masters thesis examined middleaged individuals life span di stribution of memories with a focus on the relatively greater frequency of memories recalled from young adul thood (ages 10 to 30). During her graduate training at Ko, she realized that her research interests and go als strongly fit with Dr. Susan Blucks program of research at the University of Florida (UF). After starting the developmental psychology Ph.D. Program at UF, she expanded her research progr am by integrating theory and empirical research on adult develo pment and social cognition. Her training as the manager of Dr. Blucks Life Story Laboratory, a nd as teaching assistant to many courses has provided her with a strong background in life span development, providing a foundation for specialization in adult development and its relation to autobiographical memory. In the future, she hopes to further develop her research program on the bi-directiona l relationship between a dults self conceptions and the autobiographical memories they recall from their lives.