Reminiscence, Personality, and Psychological Functioning in Older Adults

Abstract

Purpose: The present study examined the relationships between the frequency and functions of reminiscence, personality styles, and psychological functioning. There is little research on the psychological factors that correlate with reminiscence, especially in relationship to clinical constructs such as depression and anxiety. Research in the area of reminiscence functions may facilitate a better understanding of the factors affecting change in reminiscence therapies. Design and Methods: Seventy-seven healthy older adults completed the following self-report scales: Reminiscence Functions Scale, NEO Five Factor Personality Inventory, Beck Depression Inventory—Second Edition, State–Trait Anxiety Inventory, and the Templer-McMordie Death Anxiety Scale. Results:Using canonical correlation techniques, results indicated that individuals with negative psychological functioning frequently reminisce as a way to refresh bitter memories, reduce boredom, and prepare for death. Implications: The present study provides implications for both researchers and clinicians. Contrary to previous studies, results indicate that depressed and anxious older adults commonly use reminiscence and therefore may be appropriate candidates for reminiscence treatments.

Laurence G. Branch, PhD

Reminiscence, the recalling of memories from one's past experiences, is a multidimensional process that occurs in people of all ages. Butler 1963 described reminiscence as a part of the normal life review process brought about by the realization of death and viewed the use of the past as a mechanism that aids individuals in preparing for death by mitigating fear and anxiety. Contrary to prior assumptions, age does not necessarily predict the frequency of reminiscence (Romaniuk and Romaniuk 1983; Webster 1994). However, older adults, compared with younger reminiscers, use past experiences for different purposes. Molinari and Reichlin 1984 suggested that reminiscence for older adults reflects qualitative rather than quantitative differences. For example, older adults tend to use their memories primarily as a source of entertainment, whereas young people use their memories mainly as a problem-solving tool (Revere and Tobin 1980). In addition, research has shown that older adults fluctuate with respect to the frequency and degree of pleasantness associated with reminiscence (Fry 1991).

Much of the reminiscence literature has focused on the clinical applications of reminiscence and life-review groups in relation to outcome factors such as depression and anxiety. Although literature supports the use of such treatments through group psychotherapy (Goldwasser, Auerbach, and Harkins 1987; Haight 1991; Rattenberg and Stones 1989), much of this information is anecdotal or involves group comparison outcome methodologies, which have traditionally neglected an examination of the factors affecting symptom relief or behavioral change. Haight 1991 comprehensively examined reminiscence treatments and noted that there is a continued need to define the factors that lead to successful reminiscing.

Wong and Watt 1991, in one of the few studies examining reminiscence functions in relation to adaptive aging, developed a taxonomy of six reminiscence types to differentiate between successful and unsuccessful aging. Their model consisted of integrative, instrumental, transmissive, narrative, escapist, and obsessive forms of reminiscence. Results indicated that successful agers showed significantly more integrative and instrumental forms of reminiscence, which primarily function to reconcile the past, attain a sense of self-worth, and problem solve. Similarly, Wong and Watt found that successful agers had lower levels of obsessive reminiscence (Lo Gerfo 1980), which included a ruminative style of accessing the past resulting from guilt, stress, or grief and evidenced by bitterness or despair.

Preliminary research has indicated that certain personality traits may facilitate reminiscence, and others may inhibit it (Lieberman and Falk 1971). Fry 1991 found that specific personality factors are strong predictors of the frequency and pleasantness of reminiscence activity, postulating that individuals who have a "delicate balance" of such personality traits as sentience and openness may have a greater capacity for, and be more responsive to, reminiscence as a therapeutic process.

Several studies have explored the relationship between reminiscence and personality, using both Webster 1993 Reminiscence Functions Scale (RFS) and the NEO Five Factor Personality Inventory (NEO–FFI). Webster 1993 found that bitterness revival reminiscence correlated with Neuroticism, conversation reminiscence correlated with Extraversion, and both identity and problem-solving reminiscence correlated with Openness. Webster 1994 indicated that the personality traits of openness and neuroticism were found to be significantly correlated with overall reminiscence frequency, whereas persons high in neuroticism were more likely to experience negative emotions when they reminisced. However, the generalizability of these findings to an older adult population appears limited, due to the small sample size of adults age 63 and up (n = 13). Finally, Molinari, Cully, Kendjelic, and Kunik in press examined the relationship between personality factors and reminiscence functions in an outpatient geriatric psychiatry population. Results indicated positive correlations between Extraversion and conversation and Openness with identity and problem-solving reminiscence functions.

Regarding reminiscence and depression, several studies have found that depressed individuals may be prone to reminisce using memories with a negative emotional valence (Fromholt, Larsen, and Larsen 1995; Mathews and MacLeod 1994; Matt, Vazques, and Campbell 1992). Other research with older adults has noted that depressed participants may have difficulty in reminiscing owing to repeated interruptions by anxious concerns about their current physical and emotional condition (McMahon and Rhudick 1967). Levy and Mineka 1998 examined the effects of anxiety on autobiographical memory within a college population. Results indicated that low-anxious participants recalled more memories overall than did high-anxious participants. Using an older adult sample, Lieberman and Falk 1971 found that individuals under stress tended to reminisce less frequently, reflecting on more painful and negative aspects in recounting their life stories. In addition, results indicated that individuals under stress tended to emphasize childhood memories in lieu of adult memories.

Previous findings related to the effect of death anxiety on life review and reminiscence have been mixed, with few conclusions drawn (Fishman 1992; Lieberman and Falk 1971; Quackenbush and Barnett 1995). The conflicting findings may be due in part to differences in construct definitions and a lack of well-validated measures.

To summarize, reminiscence research has not carefully examined the therapeutic components of reminiscence treatments. Researchers have suggested that reminiscence and life-review therapies will be more valuable if studies target and assess the factors associated with positive change. A logical step in using reminiscence as a treatment modality would be to identify the salient characteristics and relationships between reminiscence and psychological constructs that are targeted for treatment, such as depression, anxiety, and death anxiety, conditions common among older adults.

Research Questions/Hypotheses

The current study was primarily exploratory in nature; we sought to examine the relationships between reminiscence functions, personality, and psychological functioning. On the basis of previous literature, we proposed the following hypotheses:

1. Negative reminiscence functions (e.g., bitterness revival and boredom reduction) were predicted to correlate positively with psychological distress. Positive forms of reminiscence (e.g., death preparation, identity, and problem solving) would negatively correlate with psychological distress.

2. Bitterness revival was predicted to be positively correlated with depression and anxiety.

3. Higher levels of psychological distress were predicted to correlate negatively with reminiscence frequency.

Methods

Participants

Eighty-three participants age 65 and older were recruited from several independent living facilities (n = 71) and from a private university alumni foundation (n = 12) in a Midwestern metropolitan city. These individuals were generally well educated (mean years of education = 13) and reported few major medical conditions (e.g., stroke or heart attack). Potential participants completed the Mini-Mental State Exam (MMSE; Folstein, Folstein, and McHugh 1975) to determine if they were cognitively intact. Participants with scores of 24 and above were entered into the final analysis. Of the 83 independent living facility participants, 6 scored below the MMSE cutoff of 24, leaving a final sample size of 77. Table 1 provides a complete list of sample means and standard deviations for demographic variables and psychological measures.

Procedures

Participants were recruited through personal contact, flyers, and scheduled group administration sessions. The majority of measures were completed during group administrations in the presence of an examiner. Participants were placed in an environment where they could complete the protocol in private, free from distractions. To avoid potential order and response bias effects, measures were counterbalanced for order of presentation.

Measures

Reminiscence: The RFS.

The RFS is a 43-item questionnaire in which participants are asked to rate themselves on a 6-point Likert-type scale assessing how often they reminisce with a particular function in mind (Webster 1993). Items in the questionnaire are presented as completions to the stems "How often do you reminisce…" and "When I reminisce it is…". Responses range from 1 = never [reminisce for the stated purpose] to 6 = very frequently [reminisce for the stated purpose]. The completed measure yields scores on eight reminiscence functions, and a total composite reminiscence score may also be computed. Webster offered the following factor definitions:

1. Boredom Reduction is people's propensity to reminisce when their environment is understimulating and they lack engagement in goal-directed activities.

2. Death Preparation assesses the way people use their past when thoughts of their own mortality are salient, and it may contribute to a sense of closure and calmness.

3. Identity refers to how people use their pasts in an existential manner to discover, clarify, and crystallize important dimensions of their sense of who they are.

4. Problem Solving is defined as how people use reminiscence as a constructive coping mechanism whereby the remembrance of past problem-solving strategies may be used again in the present.

5. Conversation is people's natural inclination to invoke the past as a means of connecting or reconnecting with others.

6. Intimacy Maintenance is a process whereby cognitive and emotional representations of important persons in people's lives are resurrected in lieu of the remembered person's physical presence.

7. Bitterness Revival is the extent to which memories are used to affectively charge recalled episodes in which people perceive themselves as having been unjustly treated.

8. The Teach/Inform factor measures the ways in which people use reminiscence to relay to others important information about life and/or themselves. It is an instructional type of narrative.

The RFS has adequate to good psychometric properties. Factor internal consistency scores range from a low of .74 for Teach/Inform to a high of .86 for Identity and Problem Solving (Webster 1997). Research regarding the criterion validity of the measure has been positive. Webster 1993 found strong and expected correlations between NEO–FFI factors and multiple reminiscence functions. In addition, Webster 1993 examined the RFS developmentally and found significance for an increase in death preparation reminiscence with increases in age.

Personality: NEO–FFI.

The NEO–FFI is a shorter version of the well-known, well-validated NEO Personality Inventory (Costa and McCrae 1992) and consists of 60 items presented on a 5-point Likert-type scale. The inventory takes approximately 15 min to complete and yields scores on the Big Five personality factors: Neuroticism (tense, anxious, worrying, and moody), Extraversion (talkative, assertive, active, and dominant), Openness (imaginative, curious, and possessing a wide range of interests), Agreeableness (kind, affectionate, generous, and friendly), and Conscientiousness (responsible, reliable, dependable, and organized).

Depression: Beck Depression Inventory—Revised (BDI–II).

The BDI–II (Beck, Steer, and Brown 1996) is the recently revised version of the well-known BDI (Beck, Ward, Mendelson, Mock, and Erbaugh 1961). Items on the BDI–II are presented under subject "headings," with each item containing four sentence options. Participants are instructed to select one statement in each group that best describes the way they have been feeling for the past 2 weeks, including the present day. As the BDI–II is a relatively new inventory, little research exists for its use with older adults. However, the original BDI is a reliable and well-validated measure that has shown adequate psychometric properties with older adults in both clinical and research settings (Gallagher, Breckenridge, Steinmetz, and Thompson 1983; Gallagher, Nies, and Thompson 1982).

Anxiety: State–Trait Anxiety Inventory (STAI).

The STAI is the reliable and well-validated measure created by Spielberger, Gorsuch, and Lushene 1970; it has been found to be valid with older adults (Himmelfarb and Murrell 1983; Kabacoff, Segal, Hersen, and Van Hasslet 1997). The STAI is an objective personality test, designed for the measurement of current (state) and general (trait) anxieties. State anxiety is measured under the heading "How do you feel right now?" whereas trait anxiety, a more stable attribute, is described as the way an individual typically reacts to stressful or anxious situations.

Death Anxiety: Templer-McMordie Death Anxiety Scale (TDAS).

The TDAS is a widely used 15-item scale that is both a reliable and a valid measure of death anxiety (Templer 1970). A modified version of the TDAS was used for the current research, which uses a 7-point Likert-style format, ranging from strongly agree to strongly disagree, in lieu of the original true–false version (McMordie 1979; Templer 1970). Internal consistency and test–retest reliability scores for the modified version of the TDAS are .84 and .83, respectively (McMordie 1979). On the basis of a sample of 320 undergraduates, McMordie 1979 indicated that the mean and standard deviation for the TDAS were approximately 58 and 13, respectively.

Results

The primary analyses were aimed at understanding the relationship between reminiscence functions and personality factors, anxiety, depression, and death anxiety. Holding the level of significance constant at p < .05, it can be seen from the two-tailed Pearson correlations in Table 2 that Neuroticism scores correlated positively and significantly with bitterness revival, boredom reduction, death preparation, and total reminiscence. Extraversion correlated positively and significantly with conversation and teach/inform. Also significant, Agreeableness correlated negatively with bitterness revival and boredom reduction. With regard to psychological outcomes, a review of Table 2 indicates a pattern of positive relationships between depression, anxiety, and death anxiety and the reminiscence functions of bitterness revival, boredom reduction, and death preparation, as well as total reminiscence.

Canonical Correlation

The canonical analysis provided a multivariate assessment of the overall relationship between the two sets of variables (psychological functioning and RFS factors), controlling for interrelationships. Canonical correlation attempts to generate a maximum correlation between the two sets of variables in the form of a variate that allows for the interpretation of those variables that contribute most significantly to the relationship. In conducting such an analysis, there may exist additional, separate variates; however, each subsequent variate produced is orthogonal to those previously generated. The current canonical analysis included a set of reminiscence variables (boredom reduction, death preparation, identity, problem solving, conversation, intimacy maintenance, bitterness revival, and teach/inform) and a collection of psychological variables that included the five NEO–FFI personality factors, depression (BDI–II), state and trait anxiety (STAI), and death anxiety (TDAS).

Three of the seven canonical variates from the analyses were statistically significant beyond the .05 level. As indicated by a redundancy analysis, a total of 24% of the variance in the reminiscence factors was explained by the psychological measures, with the first variate explaining the majority (18%). An examination of the structural coefficients (see Table 3 ) indicated that the reminiscence functions with the highest structure coefficients for Variate 1 are bitterness revival (−.95), boredom reduction (−.64), and death preparation (−.51). Structure coefficients from psychological functioning revealed that trait anxiety (−.75), neuroticism (−.76), state anxiety (−.75), depression (−.59), death anxiety (−.55), and agreeableness (.50) had the highest loadings. This first variate appears to suggest that as psychological functioning becomes "healthier" (e.g., lower levels of anxiety, depression, and death anxiety), individuals tend to use reminiscence less frequently for the purposes of bitterness revival, boredom reduction, and death preparation.

The second canonical variate accounted for 4% of the variance and consisted of the RFS factors Conversation (.56) and Teach/Inform (.66) and the psychological variables extraversion (.66), openness (.63), and depression (−.63). The second variate suggests that individuals high on extraversion and openness with lower levels of depression tend to frequently reminisce for the purposes of teaching and conversing. Although the third variate generated was statistically significant, it accounted for only 2% of the variance and, accordingly, is not reported.

Discussion

Butler 1963 contended that if individuals used reminiscence in a negative fashion, they would be more likely to experience heightened emotional distress as evidenced by increased symptoms of both depression and anxiety. The results of the present study lend empirical support to Butler's postulate. Our analyses suggest two general themes: (a) Negative psychological functioning is positively associated with the bitterness revival, boredom reduction, and death preparation functions of reminiscence; and (b) higher levels of psychological distress (e.g., state anxiety, trait anxiety, depression, and death anxiety) are associated with greater reminiscence frequency.

Findings indicate that as psychological distress increases, older adults may be more likely to recall bitter memories and reminisce in order to alleviate feelings of boredom as well as to deal with issues of mortality. However, due to the correlational nature of the study, the opposite may also exist: As individuals more frequently use bitterness revival, boredom reduction, and death preparation functions of reminiscence, they may be prone to psychological distress.

In addition, because of the robust nature of the correlation between bitterness revival and psychological distress, one could argue that the relationships between boredom reduction and death preparation may be only indirectly related, mediated by high levels of bitterness revival. Therefore, it may be the interaction of these reminiscence functions, mediated by bitterness revival, that ultimately relates to psychological distress.

In comparing the current findings with prior research, we should note that a significant difference exists in the expected level of reminiscence frequency with depressed and anxious participants. The current study found that higher levels of depression and anxiety were positively correlated with overall reminiscence frequency. These findings differ from those reported previously (Levy and Mineka 1998; Lieberman and Falk 1971; McMahon and Rhudick 1967). Previous researchers have suggested that higher levels of psychological distress would contribute to a difficulty in reminiscing. The current findings did not support this hypothesis but indicated that individuals with greater psychological distress exhibited to noticeable decrements in their ability to recall and use information from their past. The current findings suggested that individuals with higher psychological distress appeared to use reminiscence more frequently than their psychologically healthier counterparts. However, we should address two caveats. First, higher levels of reminiscence activity associated with "distressed" reminiscers may be due to increases in the levels of negative reminiscing, which in turn create higher levels of overall reminiscence frequency. Second, as indicated, cause cannot be inferred as to whether negative forms of reminiscing are due to high psychological distress or whether high psychological distress is influenced by negative forms of reminiscing.

The current results strongly support Butler 1963 proposition that personality factors play a salient role in the types and frequencies of reminiscence activity. The results suggest that the personality variables neuroticism and extraversion appeared to be the best predictors of overall reminiscence frequency. The extraversion findings showed mixed consistency with previous studies. On the other hand, higher levels of neuroticism have been consistently associated with higher levels of reminiscence activity across studies (Molinari and associates in press; Webster 1994). Therefore, individuals higher in neuroticism may be good candidates for reminiscence group treatments.

In addition to predicting overall reminiscence, we also found that several personality factors accounted for a moderate amount of variance within specific reminiscence typologies. For example, higher levels of extraversion and openness were associated with both conversation and teach/inform reminiscence. Consistent with our initial hypothesis, neuroticism was found to be positively associated with bitterness revival and boredom reduction. Finally, agreeableness was inversely related with the bitterness revival and boredom reduction forms of reminiscence.

Although the results from the present study are not robust, there was a significant trend indicating that increased death anxiety is positively associated with higher levels of reminiscence geared at preparing for death. In addition, death preparation forms of reminiscence were strongly correlated with higher levels of depression and anxiety, possibly suggesting that death preparation may have both positive and negative aspects. For example, a moderate amount of death preparation may be beneficial to psychologically accept one's mortality; however, greater amounts of death preparation may begin to resemble a ruminative style of thought processes, creating distress instead of functioning as a coping mechanism.

Clinical Applications

Reminiscence therapy may be a clinically viable treatment modality, especially for patients suffering from depression and anxiety. Our results indicate that older adults who reported more depression and anxiety used reminiscence more frequently than their psychologically "healthier" counterparts. These individuals may have felt quite comfortable with a reminiscence intervention, as it was an activity in which they already engaged. Examining the ways in which depressed and anxious individuals used reminiscence, it appears that these individuals tended to use negative forms of reminiscence as indicated by the RFS factors of Bitterness Revival and Boredom Reduction. Therefore, clinicians may wish to focus their reminiscence treatments in the area of providing patients with a balanced look at the past, especially in the area of bitter memories. In addition, owing to the presence of the seemingly maladaptive boredom reduction strategies, clinicians may teach or focus on other, more adaptive aspects of reminiscence (e.g., problem solving, conversation, intimacy maintenance, etc.). It is believed that the RFS may ultimately aid clinicians in matching patients on their personal uses of reminiscence with treatment modalities that emphasize such functions.

Future Research

Replication of these findings with persons experiencing clinically significant anxiety and depression would be useful. Future research should continue to formulate an answer to the question "Why do reminiscence and life review therapies work?" specifically focusing on those factors that are suggestive of treatment efficacy. The current study can be viewed as a first step in a series of research questions. Long-term goals of such a program of research include (a) addressing the RFS's factors and their utility in helping clinicians identify individuals who will benefit from reminiscence and life-review therapies, (b) identifying the curative factors of reminiscence group treatments, and (c) identifying those techniques or treatment modalities that will be most effective within a therapeutic context, as determined by the individual's use of reminiscence.

There continues to be a need to examine the RFS in terms of test–retest reliability over a sustained period of time. Similarly, it would be beneficial to assess the stability of such reminiscence typologies, especially with relation to increases or decreases in psychological distress. One critique of the RFS, in relation to past research findings, is the scale's inability to temporally categorize reminisced memories. Previous research efforts suggest that individuals who are bitter or tend to reminisce for "negative" purposes also tend to rely on early memories in lieu of their recent past. If intended for use in assessment and clinical applications, it would appear beneficial for the RFS to include items assessing the temporal sequencing of recalled memories.

We thank the thoughtful review of Victor A. Molinari, PhD, and Jeffrey Webster, MS. We also gratefully acknowledge Honore M. Hughes, PhD, and her research team for their support and guidance.

References