Structured and Unstructured

Reminiscence Training and Depression

among the Elderly
P. S. F r y , PhD

ABSTRACT. One hundred and sixty:two depressed Caucasian

elderly were treatment and control subjects for a therapeutic in-
tervention designed to test the efficacy of structured and unstruc-
tured reminiscence training for subjects' depression. Pre-and post-
treatment measures of depression, ego-strength and self-assessment
ratings were obtained. Consistent with hypotheses, a multivariate
analysis revealed that subjects trained in structured reminiscence
showed significantly greater improvement on the dcpcndent
measures than did subjects trained in unstructured reminiscence.
Overall, subjects in both structured and unstructured reminiscence
conditions reported more improvement than the no-treatment control
subjects. Contents of subjects' reminiscence were factor-analysed
for sex differences and yielded four factors with high factor
loadings. Results of the factor analysis and the effects of the two
forms of reminiscence training were discussed in terms of their im-
plications for the treatment of depression in the elderly.

The connections between important life events, predispositions in

terms of coping and defensive styles, and reactions such as adapta-
tion, disengagement or depression in the elderly, encompass an area
of keen scientific interest in recent years (Larson, 1978). A number
of prevention studies (e.g., Bakes & Zerbe, 1976; Berger & Rose,
1977; Blackman, Howe, & Pinkston, 1975; Citron & Dixon, 1977;

Thc study was supported by a 1981 grant to the author from National Health and Welfare
(Welfare Branch) Canada. This financial assistance is gratefully acknowlcdgcd. Thc author
gratefully acknowledges the assistance of several graduate students in Developmental
Psychology who assisted in data collection. The author is gratcful to the elderly subjects (who
asked to remain unidentified) for their cooperation in this study and for the valuable insights
they provided.
Requests for reprints should be sent to P. S. Fry, Department of Educational Psychology,
The University of Calgary. Calgary. Alberta. Canada T2N IN4.
Clinical Gerontologist, Vol. 1(3), Spring 1983
@ 1983 by Thc Haworth Press, Inc. All rights reserved. 15
 CLINICAL GERONTOLOGIST

Gottesman, 1970) have been undertaken in which the purpose is to

assess the level of depression and distress prior and subsequent to
the occurrence of the serious life events, and to evaluate the impact
of some form of prophylactic intervention on depression or similar
form of morbidity. Most of the preventive studies concerned with
improving the psychological well-being of the elderly are derived
from psychosocial theories of aging such as activity theory (Brad-
burn, 1969; Cavan et al., 1949) and disengagement theory (Cumm-
ings et al., 1960), and are predicated on the assumption that the ag-
ing process precipitates events which impact in a negative manner
on elderly individuals' perceptions and attributions of life events.
There is general agreement that (a) depression is the most com-
mon functional psychological disorder among the elderly, affecting
7-1 1% of the population, (Gurland, 1976) and (b) the elderly who
have experienced losses and rejection from negative life events will
most likely respond with affects of increased depression (Lowenthal
& Chiriboga, 1973; Mueller, Edwards, & Jarvis, 1977). These
authors contend that some form of reminiscence training or treat-
ment would lead the depressed elderly to focus on meaningful
dimensions of past life events. The strategy of reminiscence presup-
poses that there exists in the elderly person's long-term memory a
fund of associations, recollections and images that form a part of the
individual's conscious experience. Previous researchers advocating
reminiscence (e.g., Horowitz, 1976; Horowitz, Schaefer, &
Cooney, 1974; Horowitz et al., 1977; Lunberg & Theorell, 1976)
have argued that the elderly, more than other populations, have
tended to use patterns of intrusion and avoidance which have in-
hibited clear recollection of past events. Intrusions are characterized
by the emergence of unbidden and unrelated thoughts and images,
troubled dreams and strong pangs of nonspecific feelings; avoidance
responses include ideational constriction, behavioral inhibition;
counterphobic activity; inhibition of affect and inhibition of sensa-
tion (Horowitz & Wilner, 1980). With respect to the elderly, it is
generally recognized that multiple losses associated with aging and
the psychological consequences of not having either the skills or op-
portunity to enage in adaptation often lead the elderly to become
more depressed and morbid after serious life events. Thus it may be
argued that an intervention procedure which (a) encourages the
elderly towards some purposive and spontaneous retrospection of
serious life events and (b) assists them to overcome tendencies of in-
trusion and avoidance in retrospection would have the effect of
alleviating their depression and morbidity. Thus the assumptions are
that reminiscence training would have the effect of encouraging
clients to let go of losses associated with past events, recognizing
their avoidance tendencies in recalling past events and thus adapting
more readily to their current stage of development. A number of
writers and researchers (e.g., Altrocchi, 1980; Boylin, Gordon, &
Nehrke, 1976; Kastenbaum, 1979) have emphasized the value of
some form of reminiscence training for the elderly. The assumption
is that most elderly subjects lack sophisticated cognitive and intellec-
tual skills required in therapy. Reminiscence training, by contrast,
involves simply the ability of the client to elicit memory of past
events (cf. Romaniuk, Note 1) in integrating present events.
However, as concluded by Levy et al. (1980) and Boylin, Gordon,
and Nehrke (1976) many important parameters of reminiscence
need to be explored and "current understanding of this therapy
technique is tentative at best."
Thus one of the major purposes of the present study was to study
the effects of reminiscense training on depression among the elder-
ly. The use of this counseling strategy with the elderly is based upon
the assumption that the skill for reminiscing is a personal resource
that is available to the elderly in almost all environmental, physical,
and psychological conditions. As described by Havighurst and
Glasser (1972) it is the act of recalling to memory past events, and
may be either oral or silent. While reminiscence as a counseling
strategy is intended to be "purposive," it is deliberately not viewed
as a marshalling of events in order to make a decision, nor is it viewed
as a special skill distinguishing certain individuals from others
(Havighurst & Glasser, 1972). One of the major hypotheses of this
study was that reminiscence training would lead to a significant
reduction in the self-reported depression scores of the elderly and in
their self-ratings of their ego-strength.
Since two formats of reminiscence training were planned it was
hypothesized additionally that "structured" reminiscence would be
more effective than "unstructured" reminiscence in alleviating
depression and improving ego-strength.

Components of Depression as Conceptualized in This Project

The focus in this study is upon depressive mood and the

behavioral symptomatology (Beck, 1967; 1970) associated with it.
There is no attempt here to distinguish between endogenous versus
18 CLINICAL GERONTOLOGIST

reactive depressions (Mendels, 1970; Partridge, 1949) nor between

clinical versus subclinical manifestations of such depressive mood
states. It is recognized that depressive symptomatology can be
associated with a range of diagnostic categories and pathological and
non-pathological conditions.
While the distinctions between various depressive clinical types
or syndromes are ambiguous and controversial, the criteria for
establishing depressive mood are much less so. Depressive mood, it
is here held, is signified by a variety of factors which have been
traditionally used as clinical mood indicators on clinical scales and
in clinical interviews. Beck (1967) and Mendels (1970) have shown
furthermore, that a variety of symptoms such as lack of reactivity to
the environment; loss of interest in the environment, overreaction to
stressful events, self-pity, blue mood, loss of perceived control over
events, uncertainty and fear of recurrence of events, lack of self-
confidence-are common to both reactive and endogenous depres-
sion.

The Model of Depressive Mood Underlying the Study

The model of depressive symptomatology which most relates to

the present study is that of depression as learned helplessness
(Engel, 1968; Melges & Bowlby, 1969; Seligman, 1975). Learned
helplessness is thought to be associated with a perceived loss of con-
trol over life situations. Depressive mood in the elderly can be
usefully viewed from this perspective in that a number of factors
would seem to lead to a potential loss of perceived control in this
group; for example, personal losses incurred through death and per-
sonal incompetence in handling past events, diminishing personal
skills in reaching future goals, to mention a few. It appears then that
the effects of learned helplessness are cumulative and global. The
model implies that depression as associated with learned helpless-
ness can be unlearned by enhancing the individual's sense of control
over past and present life events (see Fry & Grover, 1981).

Model of Reminiscence Underlying the Study

The essential concept of reminiscence is adapted from Coleman

(1974), Havighurst and Glasser (1972), and Lieberman and Falk,
(1971). These authors define reminiscence as the general behavior
of dwelling upon the past and reported systematically by people
through their own introspection and related to their life experience.
The content of reminiscence (both oral and silent) is studied by rely-
ing on the self-reports of individuals. The assumption is that this
memorial phenomenon which encourages individuals to reminisce
and elicit memories of past events has adaptive value for the elderly
(Havighurst & Glasser, 1972; Lewis, 1971; McMahon & Rhudick,
1967).
The model of reminiscence as used in this study was derived from
a broad dynamic theory that incorporated ideas from Sherman (Note
3) and relied on Horowitz (1975; 1976), Horowitz and Wilner
(1980), Janis (1969), and Lazarus (1966) who examined the mental
processes at work in the conscious and unconscious human ex-
perience. Some of these authors (e.g., Horowitz, 1975; 1976;
Horowitz & Wilner, 1980) abstracted. from psychotherapy inter-
views clients' systematic response patterns of intrusion and
avoidance. For purposes of the present study the systematic patterns
of intrusion and avoidance identified by these authors were deemed
to be important for the content analyses of subjects' responses.
Reminiscence was encouraged by asking respondents to review
five life events which had had the most stressful effect on them. Of
the two broad categories of reminiscence-oral and silent-the ma-
jor focus was on oral reminiscence since it was expected to be a bet-
ter indicator of the affective quality of the reminiscence and allowed
for the exchange of a wide variety of examples, questions and clues
between the respondent and the trainer.

METHOD
Subjects were initially 400 elderly Caucasians between the ages of
65 and 82 years who were invited to participate in a study in which
the objective, as defined, was to obtain subjects' appraisals of their
past life events. The initial pool of subjects was drawn from major
Canadian cities (i.e., Vancouver, British Columbia; Calgary,
Alberta; Toronto and Ottawa, Ontario) and United States cities
(i.e., Philadelphia, Pennsylvania; Washington, D.C., San Fran-
cisco, California, and Spokane, Washington). All subjects con-
sented to respond to the Beck Depression Inventory (Beck et al.,
1967) and an adapted version of Barron's (1953) Ego Strength
Scale. Subjects were grouped into High and Low Depression condi-
tions based upon a median split of the Beck Depression Inventory
Scores. Thus 162 high depression subjects (Mean Beck Score = 19;
20 CLINICAL CERONTOLOGlST

Mdn age = 68.5 years; 96 females and 66 males) were identified.

All subjects, prior to retirement, had been in managerial and ad-
ministrative positions, including technicians, engineers, medical
doctors, university teachers, school teachers, accountants and
nurses.
Seventy-five percent of the subjects lived alone (i.e., were
widowed, single, separated from their families) in bachelor units,
apartments and single-dwelling homes and had some paid help in
taking care of the house. People living in institutional settings were
not selected since the objective was to get subjects who though
depressed, were functioning, at least marginally, in the com-
munities.

Procedure

The subjects' self-ratings on the depression scale, the ego-

strength scale and self-assessment scale were used as preinterven-
tion measures for purposes of the present study.
The Ego-Strength Scale adapted from Barron (1953), and Parit-
zky and Magoon (1979) measures low medium and high levels of
ego-strength. High scores on the scale indicate, among other fac-
tors, the ability to share feelings, a self-image of personal adequacy,
and a strong sense of reality. Individuals fitting these characteristics
are positively oriented, self-determined, self-motivated and perform
well in specific decision-making activities (Paritzky & Magoon,
1979). The 50 true-false items of this scale were fully comprehensi-
ble to all subjects. Based upon discussions of the reliability and
validity of the Ego-Strength Scale (cf. Barron, 1953; Paritzky &
Magoon, 1979) it was assumed that high scores on this scale would
be negatively associated with high scores on the depression scale.
Self-Assessment Rating Scale. The scale tapped subjects' percep-
tions of their need for isolation, degree of pessimism, need for
assistance from other adults, frequency of obstacles, sense of self-
efficacy and other such related variables. A 0-10 self-anchoring
scale adapted form Cantril (1965) was used in obtaining the sub-
jects' ratings.
Beck Depression Inventory. In order to ascertain the subjects'
level of current depression the 21-item Beck Depression Inventory
(Beck et.al., 1961) was used. This self-reporting scale is a measure
of affective, behavioral, cognitive and somatic symptoms of depres-
sion. The measure has been shown to correlate highly with
psychiatrists' ratings of depression (Bumberry et al., 1978) and with
interview-based scores (Hammen, 1980). All respondents to the in-
itial call for subjects were promised some treatment of intervention.
It was decided that treatment should first be extended to the high
depression subjects.
Intervention. The purpose of the intervention was.to evaluate the
effects of two formats of reminiscence training on subjects' depres-
sion scores, ego-strength scores and self-assessment ratings.
The 162 high depression subjects (identified from the larger pool
of 400 subjects) were randomly assigned to two treatment groups
and one no treatment control group ( ~ = 5 4in each group).
Treatment I : Structured Reminiscence Training. The training
procedure in this condition attempted to make the reminiscence pro-
cess as structured as possible for the client. A set of explicit ques-
tions were designed which encouraged the client to reminisce about
negative life events following a step-by-step trainer-client interac-
tive procedure.
The procedure which was used in this reminiscence training was
based upon the results of a pilot study (n=20 high depression) in
which the objective was to test the efficacy of a structured
reminiscence modality adapted from Butler (1963) and Sherman
(Note 3). A content analysis was done of the interview data and the
findings suggested that factors having the highest loadings (i.e., .70
and above) dealt with strong emotions that subjects had tried to con-
trol, and unresolved conflicts and feelings of guilt.
The findings of the pilot study suggested a rationale consistent
with the views of previous authors (e.g., Horowitz, 1979; Horowitz
I& Wilner, 1980) who had also noted a preponderance of "intrusive
thoughts" and "avoidance" in the cognitions of the elderly.
Thus, the purpose underlying the structured reminiscence train-
ing procedure was to encourage the subject to deal with "intrustive
thoughts" and "avoidance" items that are.often associated with
stressful and/or negative life events. Since many of the stressful life
events about which the elderly subjects chose to reminisce had to do
with death, separations, arguments with important people, major
threats to self, loss of affection, threats to material and financial
well-being, abandonment by family etc., the expectation was that a
strucrured reminiscing procedure would provide to the subject an
outlet for strong affects, unresolved feelings, compulsive thoughts
and anxieties, and major concerns, fears and hopes.
In the structures reminiscence the subject was encouraged 10
22 CLINICAL GERONTOLOGIST

dwell on all eight aspects of each life event as described below, on a

10-minute interval schedule.

1. Relate strong affects, both negative and positive, associated

with the event. (General question was "Tell me how you
felt?" Specific clues given by the trainer were, for example,
"Did you feel happy, lonely, left out, deeply hurt, angry, sad,
numbed by it all, irritated, frustrated, threatened, anxious, un-
safe, violent?")
2. Review objects, other events, persons associated with the
event.
3. Relate images and compulsive thoughts which impinged upon
the subject (e.g., "Were there thoughts that flooded your mind
even when you were tired and didn't want to think? Were there
thoughts that came into your mind over and over again? Were
there thoughts you couldn't get rid of although you tried very
hard? Were there persons whose memory you could not banish
from your mind?").
4. Relate hopes, anxieties and fears associated with the event
(e.g., "Did you wish certain things would happen that would
make you feel better? Tell me about your hopes at that time?
Tell me about the fears you had? Were there things that made
you feel afraid and anxious?").
5. Relate dreams, including unsettling bad dreams that were
associated with the event.
6 . Relate social-interactional preferences (e.g., "Did you wish
people would let you alone? Did you wish people would stay
around you and talk with you? Did you wish people would
make a fuss about you, give you special attention?").
7. Relate unresolved feelings which the subject was aware of but
which (s)he had attempted to keep "under wraps" (e.g., "Did
you feel guilty, responsible, ashamed, embarrassed, self-
conscious, defeated, persecuted, victimized, lost, confused?
Were there people or situations you tried very hard to
avoid?").
8. Relate the manner in which the subject spent his (her) time
(General question: Tell me a bit about how you spent most of
your time when this event occurred? Clues: Did you sleep a
lot? Did you go for walks? Did you cry? Did you talk on the
telephone? Pace up and down?).
 P. S.Fry 23

Treatment 2: Unstructured Reminiscence Training. In this treat-

ment condition, the subject was asked to try and remember and to
reconstruct aspects of the life-event that (s)he thought were impor-
tant. Subjects were asked "to talk about anything you want to about
the upsetting event-Tell me about the things you recall or associate
with the events". Reminiscence was encouraged by the trainer say-
ing "Yes, tell me more about it" on a 10-minute interval schedule.
Unlike the procedure used in the structured reminiscence, the
trainer refrained from asking specific questions regarding objects or
persons related to the life-event although these were often men-
tioned by subjects of their own accord. The trainer avoided giving
cues that would limit or focus the responses of the subjects. The on-
ly restriction on the unstructured free recall of the client was that the
trainer sometimes sought clarification of certain statements made by
the subject (i.e., sorry, I didn't quite understand that, could you ex-
plain that? or give an example). Overall, the trainer encouraged the
subject to dwell on aspects of the event that (s)he perceived to be
significant.
In neither treatment condition did the trainer require the subject in
the reminiscence to marshall or recall facts that led up to a specific
decision on the part of the subject. Nor was the subject at any time
pressed into remembering what decision (s)he had made in the past.
For example, questions such as "how much money did you give to
charity?", "how much were your assets or debts etc?", "What ac-
tion did you take?", "What was the name of your lawyer?", were
carefully avoided by the trainer. The trainer avoided advice-giving,
or making any evaluating statements such as "That's good-or that
was terrible!" although advice or feedback was often sought by sub-
jects.
In both treatment conditions the methodology used was such that
the trainer offered an "empathic listening environment" within
which there was good rapport between the trainer and subject.
Trainer comments and reflections were designed to be nonevalua-
tive, nonjudgmental and very accepting of the scattered reminiscences
of the subject. Subjects' postexperimental evaluations of the trainers
confirmed that they had provided the intended "warm and
listening" orientation.
In both training conditions the reminiscence procedure was such
that subjects were encouraged to be fairly verbal, and to speak
without hesitation about a diversity of life events and experiences.
24 CLINICAL GERONTOLOGIST

The present author's previous experience of relating with the sub-

jects for the present study (see Fry & Grover, 1981) had provided
sufficient evidence that elderly subjects are quite articulate and ac-
customed to thinking and talking about human behavior and about
themselves. An examination of the background of the subjects
showed they were predominantly middle class and had had a diversi-
ty of experiences.
In both training conditions, subjects engaged in reminiscing about
five of the most negative and upsetting events which they had ex-
perienced. The trainer met with each subject for a 9 0 minute weekly
session for five weeks. The subject was encouraged to reminisce
about one negative event per week.
In both training conditions subjects were encouraged to supple-
ment the oral reminiscence by silent reminiscence in private and/or
by writing diaries. For purposes of the present study, however, the
data concerning the reminiscence process and contents analyses of
reminiscence were derived from subjects' oral reminiscence only.

No-treatment Control Condition

In this condition, the trainer met with each subject for a 90-minute
weekly session for five weeks. In order to control for amount of in-
teraction with the trainer and the effects of trainer time, the trainer
got each subject's permission to visit the home. The trainer engaged
the subject in some kind of activity of mutual interest (e.g., watch-
ing T.V. and encouraging the subject to talk about the program;
discussing some topical issue from the newspaper etc., playing
chess or teaching the subject some new activity). On some occa-
sions, subjects spontaneously began to reminisce about past events.
Trainers were alerted to this possibility and were instructed to keep
the subject involved in the "here-and-now" activity. Three control
subjects could not be persuaded to relinquish their reminiscence and
their data were excluded from the study sample.
Trainers were 16 male assistants (median age = 28.5) with a
completed Master's degree in Counseling and/or Clinical Psychol-
ogy. All trainers had extensive prior training in human relations and
communications' skill training and all of them were required to at-
tend a training program organized by four master-trainers. The pur-
pose of the training was to ensure that they would implement the
reminiscence procedures to exact specification and with consisten-
cy.
 P. S. Fry 25

Reliability Assessments

Two master trainers worked with small groups of four trainers

and practiced with master tapes prepared in advance from the
responses of a pilot group of 20 elderly subjects. Eight trainers
received practice for implementing the unstructured reminiscence
procedure and the other eight trainers received practice for im-
plementing the structured reminiscence procedure. In order to en-
sure that there were no carry over effects from the trainers' personal
orientation they were not informed about the specific hypotheses of
the study nor about the treatment or control conditions of the sub-
jects.
Two master trainers rated the reminiscence procedure of the
assistant trainers (functioning with a number of pilot clients) for a
total of 150 minutes until a minimum of 80% agreement in ratings
was obtained between pairs of master trainers. Interrater percentage
of agreement ranged between 80% and 93% during training.

Intetjudge Reliability Assessments

After master trainers were satisfied that assistant trainers were

appropriately trained to implement structured or unstructured
reminiscence one final tape of an assistant trainer functioning with
an elderly pilot client was submitted for interjudge ratings.
Judges were four male university professors with considerable
expertise in, reliability assessments and also communication skills
training. Pairs of judges independently rated on a five-point scale,
each assistant trainer's tape for (1) skill to provide a warm, em-
pathic tone of voice; (2) ability to maintain a "structured" or
"unstructured" format in the interview; (3) ability to maintain a
"listening and attentive" expression throughout the interview. In-
terjudge percentage of agreement for the three measures ranged be-
tween 85% and 97% during training of the assistant trainers.
Reliability checks by judges were performed two more times during
the five week period of the study. The trainer-subject tapes selected
for the reliability check were randomly selected by a predetermined
schedule of dates. Assistant trainers were unaware of the judges'
schedule. Interjudge percentage of agreement for the additional
reliability checks on 16 assistant trainers ranged between 80% and
95% for the three measures described earlier.
Assessment Measures. Subjects' pretest-posttest and follow-up
26 CLINICAL GERONTOLOGlST

change scores were examined for The Beck Depression Scores, the
Ego-Strength Scores and Self-assessment Ratings.

Overall Program Procedure

Data for the preintervention phase of the study were collected in

March 1981. All subjects were pretested on the Depression Inven-
tory, Ego-Strength Scale and Self-assessment rating Scale, and in-
dividual and group assessments on the measures were retained in
order to obtain adequate postintervention data. All subjects in the
treatment and control conditions were posttested on the measures
(employed in the preintervention phase) two weeks after the ter-
mination of the 5-week treatment in June 1981 and 15 weeks subse-
quent to posttesting. Since there was a considerably long gap be-
tween the pre and postintervention and follow-up administration it
was assumed that scores were not assessing possible retesting ef-
fects.

RESULTS

Content Analyses of Reminiscences

Table 1 shows the results of a factor analysis of the frequencies of

variables that entered into the contents of the subjects' reminis-
cences. Four trained graduate students (who had prior experience of
coding and categorizing subjects' cognitive statements and had
assisted in the present author's previous studies of cognitive ap-
praisals e.g., Fry & Grover, 1981; Fry & Ghosh 1980; Fry, 1982)
were employed to do a components analysis of the tapes prepared
from the reminiscence sessions.
Research assistants produced verbatim typescripts and audiotapes
of the reminiscences in the two treatment conditions. Ten-minute
segments of verbalizations randomly selected from the beginning,
middle and end of tapes were assembled on master tapes. The com-
ponents analysis was done by four graduate students who coded sub-
jects' responses under eight broad categories of contents and
themes. Coders were not informed about the purpose and
hypotheses of the study.
The coders first set out to compile a master list of major areas and
topics mentioned by the subjects. Obvious redundancies and areas
mentioned infrequently were eliminated. The factor analysis was
 P. S. Fry

Table 1

Factor Loadings f o r Content Analyses of Subjects' Reminiscence

~ -

Factor Loadings

Factors Male Subjects Female Subjects

n = 81 n = 81

1 Negative Preoccupations

about M a t e r i a l Things

2 H e a l t h Hassles

3 Disharmony

4 Emotionality

5 Family Stress Occupation

6 Stresses of D a i l y Routine

7 Unadaptive and Adaptive

Habits

8 Goals and Aspirations

done by using all input into a principal components analysis, fol-

lowed by a varimax rotation of the factorial axes. Because of the
unwieldiness of the original set of 25 areas listed as input to the final
analysis, preliminary factor analyses were done to reduce the
original set of 25 areas to a more tractable set of 10 areas that seemed
central to the elderly subjects' spheres of concern, agitation or
distress.
Four strong and interpretable factors emerged from the frequency
of reportings and accounted for l o % , 14%, 11% and 22% of the.
variance in the data, for a total of 57%. These factors were labeled
as Negative Preoccupations about Material Objects, Health Hassles,
Disharmony and Emotionality (see discussions by Chiriboga &
Cutler, 1980). There were factor loadings of .60 and above on the
four factors.
The first factor included negative preoccupations with legal af-
fairs, finances, inflation, property losses etc. As Table 1 shows
28 CLINICAL GERONTOLOGIST

there were sex differences in the patterns of reminiscing, with

women reporting greater negative preoccupations about material
objects than men (p < .05).
The second factor loading highest was on hassles dealing with
personal health, doctor's visits, lack of faith in the medical profes-
sion, shortage of nursing staff in hospitals, etc. Table 1 indicates
that women in retirement reported more of this type of general
health hassle than men (p < .05). Men respondents reported hassles
with health but appeared more concerned about general (as opposed
to personal) matters such as the effect of environmental pollution
etc., on physical health, radiation, acid rain etc.
The third factor with the highest loading (i.e., Disharmony) con-
cerned the nonoccurrence of events, feeling empty and restless all
day long, feelings of discrimination, and anticipation of impending
stress. No significant sex differences were observed for this factor.
The fourth factor which, by far, had the highest factor loading
was emotionality which included strong affects such as anger,
frustration, self-pity; feelings of guilt, conflict with loved ones; and
anxiety of inner events such as death, oblivion, etc. Table 1 shows
women reporting these patterns of reminiscence more frequently
than men (p c .01).
One other factor that did not achieve a high factor loading across
sexes but accounted for 10% of the variance in the data for women
subjects was labeled Family Stress Preoccupation. This factor
achieved a loading of .60 for women who reported great personal
stress as arising from family and social activities, (e.g., children and
grandchildren's birthdays, visits by relatives). This was a surprising
finding and contrary to the assumptions and conclusions of the ac-
tivity and disengagement theory of aging which predicts detachment
with increasing age (Bradburn, 1969; Cummings et al., 1960).
Mean preintervention-follow up change scores for Depression,
Ego-Strength and Self-assessment ratings are shown in Table 2.
An overall multivariate analysis revealed a main effect for treat-
ment, F(28, 274) = 3.16, p < .O1 and a main effect for sex (28,
174) = 1.69, p < .05 suggesting that the treatment subjects dif-
fered significantly from the no-treatment control subjects and that
male and female subjects were different.

Depression

Multivariate analyses revealed that structured reminiscence treat-

ment subjects differed significantly from unstructured reminiscence
treatment subjects and no-treatment control subjects, F(28,
274) = 3.77 p < .01. A Newman-Keuls post-hoc test (Winer,
1962) showed that in regard to the dimension of depression the no-
treatment control subjects had higher posttest scores than the treat-
ment groups (p < .01). A comparison of the depression scores of
the two treatment groups showed that the structured reminiscence
subjects had significantly lower posttest scores than the unstructured
reminiscence subjects (p c .05) suggesting the greater effectiveness
of the structured reminiscence training in reducing depression.

Ego-Strength Scores

Means of preintervention-follow up change scores for three

dimensions of ego-strength are reported in Table 2.
Multivariate analyses revealed that treatment subjects were
significantly different from control subjects, F(28, 274) = 4.01,
p < .05. Subsequent univariate analyses showed that the dimensions
which significantly differentiated the groups were feelings of per-
sonal adequacy, F(2, 159 = 4.75, p < .O1 ability to share feel-
ings, F(2, 159 = 3.72, p < .05 and sense of reality, F(2,
159 = 2.89, p c .05 test showed that subjects in the structured
reminiscence condition had higher posttest scores in ability to share
feelings in personal adequacy and sense of reality than subjects in
the unstructured reminiscence condition (p < .05), and subjects in
the control condition (p < .01).

Self-assessment Ratings

Mean preintervention-follow up change scores for five negative

dimensions and five positive dimensions of self-assessment ratings
are presented in Table 2.
Multivariate analyses revealed that treatment subjects were
significantly different from controls, F(28, 274) = 4.11, p < .Ol .
Subsequent univariate analyses showed that negative dimensions
which significantly differentiated treatment from control groups
were pessimism, F(2, 159) = 3.16, p < .05, and fear of obstacles,
F(2, 159) = 4.77, p < .O1 and isolation, F(2, 159) = 3.66,
p < .05. Newman-Keuls tests showed that in regard to these
negative dimensions the structured reminiscence subjects showed
greater reduction in feelings of isolation and pessimism than sub-
jects in the unstructured reminiscence treatment (p < .05) and no-
 Table 2
Mean Preintervention to Follow-up Change Scores for Depression,
Ego-Strength and Self-Assessment Ratings of Subjects in the
Structured and Unstructured Reminiscence Training Conditions
and No-Treatment Control Condition

Variables Structured Reminiscence Unstructured Reminiscence

Subjects' Mean Change Scorer Subjects' Mean Change Scores Control Subjects
(N.54) (Nz54) (Nz54)

M SD M SD M SD

Beck's Depression Inventory

Oeoression Score 7.5 2.25 5.1 2.70 1.2
Ego-Strength Scale
Ability to Share Feelings 3.82 1.69 3.01 1.87 .99

Feelings of Personal
Adequacy 4.65 0.92 3.95 0.65 .98
Sense of Reality 4.37 1.01 1.67 1.11 1.05

Self-Assessment Scale

Isolation
Pessimism
 +
L I

,"
I
Prn n
c
I
-
"- b
m
=L
0 -
m
I
. - 3
" - . -
C
O
C
2
,
e m
L
O
w
O
,
L'
L

r ; .r 2 hY / c i 7i -Z Yi
L
=
U
e r n L O % a

S
a
32 CLINICAL GERONTOLOGIST

treatment controls (p < .01). Regarding the fear of obstacles dimen-

sions, there was a significant difference in the reduction between
treatment and control subjects but not between subjects in the two
treatment conditions. Univariate analyses with respect to positive
dimensions revealed that treatment groups were significantly dif-
ferent in their scores on self-efficacy, F(2, 159) = 4.01, p < .05,
and self confidence, F(2, 159) = 3.11, p < .05, Newman-Keuls
tests showed that the subjects in the structured treatment conditions
made greater gains in positive dimensions than subjects in the
unstructured treatment (p < .05) and no-treatment control condi-
tions (p < .01).

Effects of Sex

The multivariate analysis of variance in the main analysis showed

that there was a main effect for sex. Univariate analyses showed that
men and women in the sample were different in their change scores
for Depression and Ego-Strength, F(2, 159) = 3.26, p < .05 and,
F(2, 159) = 3.71, p < .05, respectively. Newman-Keul tests
showed that women showed less change in Depression (p < .05)
and Ego-Strength (p < .05) both in the two treatment and no-
treatment control conditions.

DISCUSSION

The research data reported introspectively, by a sample of elderly

subjects, points to the efficacy of reminiscence as a therapeutic pro-
cedure. As suggested by Romaniuk (1979) reminiscence is a
phenomenon attributed to a multiplicity of factors in the life ex-
periences of subjects and there is little systematic research to assess
its impact. The present research data confirm that training in
reminiscence has the effect of reducing subjects' self-reportings of
depression and increasing their feelings of self-confidence and per-
sonal adequacy. Thus the results provide additional support for
Lieberman and Falk's (1971) contention that reminiscence has a
general adaptive function for geriatric subjects.
In analyzing the major content factors of subjects' reminiscing,
the emotionality factor which had a very high loading had strong af-
fect components (e.g., anger, helplessness, sadness), unresolved
feelings (e.g., guilt, embarrassment, conflict with loved ones), and
fears (e.g., fear of past stressful events reoccurring, sense of impend-
ing stress). It may be that these factors as described here attributed
directly to the subjects' depression and therefore influenced their
perceptions of life events. It is possible that reminiscence as a
therapy procedure allowing for self-expression and communication,
provided the elderly a formal outlet for the expression of strong af-
fects, unresolved feelings and fears which they had tried to repress.
The fact that therapy time provided "a warm, listening and em-
pathic environment" would suggest that subjects felt free to engage
in overt reminiscence as opposed to the silent reminiscence which is
assumed to be characteristic of the elderly but which is hard to
measure and assess.
When structured and unstructured reminiscence effects were
compared, the analysis shows that subjects in the structured condi-
tion improved more significantly than subjects in the unstructured
condition, relative to control subjects. One possible interpretation of
this finding is that subjects in the structured condition were en-
couraged to "marshall" their introspection in an informative and in-
tegrated manner; they were verbally reinforced for reviewing
various and specific aspects of past life events and for recalling af-
fects, fears and hopes associated with the upsetting events. The
trainer clues (given to subjects in the training process) were for-
mulated such as to encouraee " subiects to intros~ecton the "intru-
sions" and "avoidance" items that were assumed to be inhibiting
spontaneous reminiscence. Thus it is conceivable that structured
reminiscence procedures encouraged subjects to give verbal expres-
sions to unbidden thoughts and images and any strong waves of af-
fect and unresolved feelings. Thus, the overall merit of the struc-
tured procedure may have been that it requires subjects to organize
and integrate both their spontaneous and elicited memories within a
cognitive framework that.was informative for the elderly themselves
and within a social therapeutic context that was very supportive, and
responsive to their communication needs.
By comparison, the merit of the unstructured reminiscence pro-
cedure lies perhaps in the fact that it provided subjects the oppor-
tunity to express their thoughts. The assumption is that normally
"good taste" would prevent a person from talking about negative
events with the average person. As became evident in postinterview
commentary, subjects had previously refrained from reminiscing
about negative events (e.g., I pretend to be happy when I am with
others; My children don't like it when I dwell on past things that
bother me-I can tell that, because they stop inviting me or visiting
34 CLINICAL GERONTOLOGIST

me; When I'm feeling depressed I brood a lot and I don't pay atten-
tion to what is happening around me; I've learned to stay at home
when I'm low-most people don't want to know about my fears and
hurts). Since most of the subjects spent a lot of time alone the oppor-
tunity for oral reminiscence was no doubt limited. It may be that the
unstructured reminiscing provided an acceptable time for the overt
expression of memories. Subjects had control over the recall pro-
cess and could therefore withhold communications andlor accen-
tuate aspects of their life events which were phenomenologically im-
portant to them.
In review, it seems that reminiscence, both structured and
unstructured, as generated in this study, offers certain advantages
over other methods of increasing psychological well-being. Since
reminiscence is a cognitive activity originating from the subject's
memory, there are no special setting conditions, client characteris-
tics of methodological tools that are needed for therapy nor must
clients meet any rigorous personality criteria. The one restriction in
thc use of oral reminiscence is that it is more likely to be useful with
clients that are fairly verbal, and are cooperative. One note of warn-
ing is important: Clinical judgment would suggest that some of the
processes involved in reminiscence which may be most helpful to
the general population of depressed subjects are the ones that would
probably not be safe for use with psychotic clients whose
reminiscences are subject to much distortion (Botwinick, 1978;
Klerman et al. Note 2; Radoff, 1977).
Schultz (1976) and Shultz and Hanusa (1978) have argued that
therapeutic effects in geriatric patients can seldom be attributed to
intervention and may more likely be a function of positive relations
and social contacts. It may be that some of the more stable
therapeutic effects of the reminiscence training procedures can be
demonstrated only by a longitudinal follow up of subjects. The pres-
ent study provided some evidence that the positive effects of treat-
ment wre maintained for a period of at least 15 weeks after trainers'
social contacts with the subjects had been terminated. The subjects
were followed up only for a 15-week period due to a number of
mitigating factors which would have rendered more extensive
follow up data unreliable. The factors included, for instance,
changes in the schedule of subjects, changes in physical health and
attrition in the sample. Follow up over a longer period would cer-
tainly have been valuable in establishing the stability of the trends
noted in the data for treatment subjects in the structured and unstruc-
tured reminiscence conditions. Also, such prospective studies might
usefully include a silent reminiscence training procedure to examine
whether this procedure produces similar effects as does overt remi-
niscence. Such was not used in the present study due to time, man-
power and subject sample constraints.
Future studies should also compare the cost benefit-effects of
reminiscence training and some other kinds of cognitive interven-
tions.

REFERENCES

Altrocchi, J. Abnon~ralbclrrnior. Ncw York: Harcourt Brace Jovanovich. 1980.

Bakes. M. M., & firbe. M . B,Independence training in nursing home residents. Geronrulo-
gist. 1976, 16. 428-432.
Barron, F. An cgostrength scale which predicts response to psychotherapy. Jouruul oJCon-
sulring Psychology. 1953. 17. 327-333.
Beck, A . T. Depression, experime~~rul. nud lheorerical ospecrs. New York: University o f
Pennsylvania Press, 1967.
Bcck, A. T. The core problem in depression: The cognitive triad. Scir~lcea d Psyc11oon~l)'-
sis. 1970, 17. 47-55.
Bcck. A. T.. Rush. A. 1 . . Shaw. B . F.. & Emery. G. Copliri~wrlrrral,)'ofdcpression. New
York: Guildford Press, 1979.
Beck, A. T.. Ward. C. H.. Mendelson. M., Mock, J.. & Erbagh. J. An inventory for
measuring depression. Ar(:lri~~e.s of Grncrol Psyhiarry. 1961, 4, 53-63.
Berger, R. M . & Rose. S. D . Interpersonal skill training with institutionalized elderly pa-
g ) ' . 32. 346-353.
tients. Jounlal of G ~ r ~ ~ l ~ o l o1977.
Blackman, D . K.. Howe, M. & Pinkston. E. M. lncrcasing participation in social interaction
o f the institutionalized elderly. Grronrologirr. 1976. 16. 69-76.
Botwinick, J. Aging a~rdbchavior New York: Springer. 1978.
Boylin, W., Gordon, S. K. & Nehrke, M . F. Reminiscing and cgo infcgrity in institutional-
ized elderly malcs. Gcro~~mlogisr.1976, 16, 118- 124.
Bradburn. N. M. 7hc .s/r!rrnrrcofpsy-lrnlogi~:a/ndl-being. Chicago: Aldine Publishing Co..
1969.
Bumberry. W.. Oliver. J. M.. & McClurc. I. N. Validation o f thc Beck Dcprcssion Inven-
tory using psychiatric cstimatc as the criteria. Joirr~rrrl of Corlslrlring and C l i ~ ~ i c a l
Psychology. 1978, 48. 150-155.
Butler, R. N . The life rcvicw: An interpretation o f reminiscence in Ihc agcd. Psycl~iurry.
1963, 26. 65-76.
Cavan. R. S., Burgess, E. W.. Havighurst. R . J. & Goldhammer. H. Persr~~~uladjusrrne~~l in
old age. Chicago:.Science Research Associatcs. 1949.
Chiriboga. D . A. & Cutlcr, L. Stress and adaptation: Life span perspcclivc. I n L. W. Poon
(Ed.) Aging in rhe 1980s. Washington. D . C . : . American Psychological Association,
1980.
Citrin. R. S. & Dixon. D. N . Reality orientation a milieu therapy used in an institution for
the aged. Gerontologisr. 1977. 17. 39-43.
Coleman. P. G. Measuring reminiscence characteristics from conversation as adaptive fea-
tures of old age. hr~rnorin~tnl Jorrriml of A g i q & Human D c ~ ~ e l o p n ~1974.
e ~ ~ r .5 , 281-
294
Cummings, E.. Dcan, L. P., Newell, D . S. & McCaffrey, 1. Discngagcnlcnt-a tenlative
theory o f aging. Snciomerry. 1960, 23. 23-35.
36 CLINICAL GERONTOLOGIST

Engel, G.'A life setting conducive to illness. The giving up-given up complex. Bullerin of
rhe Mennin~erClinic, 1968. 32, 355-365.
Fawcett, G., Stonner, D. & Zepclin, H. Locus of control, perceived constraint, and morale
among institutionalized aged. lnrernarional Journal of Aging and Human Developmenr,
1980, 11. 13-23.
Fry. P. S. Positive and negative attributions of longevity: A cross-sectional study of the per-
ceptions of the elderly from three socioeconomic conditions. International Journal of
~&chology. 1982, in press.
Fry, P. S. & Ghosh. R. Attributional differences in the life satisfaction of thc elderly. A
cross-cultural comvarison of Asian and United States subiects. International Journal of
Psychology. 1980,'15. 201-212.
Fry, P. S. & Grover. S. Cognitive appraisals of life stress and depression in the elderly: A
crosssultural comparison of Asians and Caucasians. Iuternarional Journal of
Psychology. 1981. in press.
Gibson, R. W. Planning a total treatment program for the hospitalized depressed patient. In
I. Cole, A Schatzberg. & S. Frazier (Eds.), Depression, biology, psychodynamics and
rrearmenr, New York, 1974.
Gattesman, L. E. Organizing rehabilitation services for the elderly. Gerontologisr, 1970, 10,
287-293.
Gurland, B. J. The comparative frequency of depression in various adult age groups. Journal
of Geronrology. 1976, 31. 283-292.
Hammen, C. Beyond the Beck Depression Inventory. Journal of Consulling und Clinical
Psychology. 1980, 48. 126-128.
Havighurst, R. I.. & Glasser, R. An exploratory study of reminiscence. Joarnul of Geron-
rology. 1972. 27. 245-253.
Horowitz. M. Stares of mind. New York: Plenum Press. 1976.
Horowitz. M. Intrusive and repetitive thoughts after experimental stress: A summary. Ar-
chives of General Psychiatry, 1975, 32. 1457.1463.
Horowirz, M.. Schaefer, C.. Hiroto, D., Wilner. N.. & Lcvin, B. Lifeevent questionnaires
for mcasuring presumptive stress. Psychosomufic Medicine, 1977, 39. 413-431.
Horowitz, M. 1.. Schaefer, C., & Cooney, P. Life event scaling for recency of experience.
In E. K. E. Gunderson & R. H. Rahe (Eds.), Li/e stress and illness. Springfield, Ill.:
Charles C Thomas, 1974.
Horowitz, M. I . , & Wilner, N. Life events stress and coping. In L. W. Poon (Ed.) Aging in
rhe 1980s. Washington, D.C.: American Psychological Association, 1980.
Janis. I. dress and frustrarion. New York: ~ a r c o u r i ~ r a Jovanovich.
ce 1969.
Kastenbaum, R. Growing old. Years of~lJillmenr.New York: Harper & Row, 1979.
Larson, R. Thirty years of research on the subjective wcll-being of older Americans. Journal
of Geronrology, 1978, 33, 109-125.
Lazarus, R. S. Psychologicalsrressand rhe coping process. New York: McGraw-Hill, 1966.
Lewis, C. Reminiscing and self-concept in old age. Journal of Geronrology, 1971, 26, 240-
243.
Lieberman, M. A., & Falk, J. M. The remembered past as a source of data for rcsearch on
the life cycle. Human Developnrenr. 1971. 14. 132-141.
Lowenthal, M.. Thurner. M.. & Chiriboga. D. Four srages of life. San Francisco: Jossey-
Bass, 1976.
Lunberg, U., & Theorell. T. Sealing of life changes: Differences between thrcc diagnostic
groups and between recently experienced and noncxperienced events. Journal of Human
Stress. 1976. 2. 7-17.
Levy. G. M.. Derogatis, L. R.. Gallagher, D. & G a b . M. Intervention with older adults and
the evaluation of outcome. In L. W. Poon (Ed.) Aging in rhe 1980s. Washington. D.C.:
American Psychological Association, 1980.
McMahon, A. W. & Rhudick. P. I. Reminiscing in the aged: An adaptational response. In
S. Levin & R. I. Kahana (Eds.), Psychodynamic srudies on aging, creariviry, reminisc-
ing. and dying. New York: International University Press. 1967.
 P. S. Fry 37

Melges. F. T. & Bowlby. J. Types of hopelessness in psychopathological process. Archives

of General Psychiarty, 1969. 20. 690-699.
Mendels. J . Concepts of depression. New York: John Wiley, 1970.
Mueller, D. P.,Edwards, D. W. & Yarvis, R. M. Stressful life events and psychiatric
symptomatology: Change or undesirability? Journal of Heullh and Social Behavior.
1977, 18. 307-317.
Paritzky, R. & Magoon. T. Human potential seminar outcomes as measured by the personal
orientation inventory and goal attainment invcntorics. Journal of Counseling Psychology.
1979, 26. 30-36.
Partridge, M. Some reflections on the nature o f affective disorders arising from the results
o f prcfrontal leucotomy. Journal of Mental Science. 1979, 95. 795-825.
Radoff, L. S. The CES-D Scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement. 1977. 1. 285401.
Schultz, R. The effects o f control and predictability on the physical and psychological wcll-
being of the institutionalized aged. Journal of Personaliry and Social Psychology. 1976,
33, 563-573.
Schultz, R, & Hanusa. B. Long term effects o f control and predictability-enhancing inter-
ventions: Findings and ethical issues. Journal of Personality and Social Psychology,
1978, 36. 1194-1201.
Seligman. M . E. P. Helplessness: 011 depression, development and death. San Francisco:
W. H. Freeman and Company. 1975.
Winer, B. I. Srutisricul principles in experimental design. New York: McGraw-Hill, 1962.