Applying Retrospective Life History Methods in the Study of Friendship among the Elderly

        A Paper Presented at the
      Annual Meeting of the Society for the Study of Social Problems
                         by
                         Richard S. Dorr
                        August 24, 1980

Genesis and Evolution of the Study
In mid-1979, the author embarked on a research study of well elderly (age 65 and older) in Great Boston. It was initially a venture without a well-defined objective, except to study this population segment. The early investigation was devoted to identifying factors worthy of more extensive research and to addressing methodological issues.

This exploratory process underscored the fact that, throughout the life course-including old age-social relations outside the family structure consistently play an important role in life decisions, self-perceptions, value structures, and relative sense of well-being. Moreover, whereas many family relationships may be obligatory, the more voluntary nature of other social relationships makes friendship a more sensitive indicator of the social processes influencing the life of an individual.

Perusal of literature on friendship has revealed that it is not a widely studied phenomenon among social scientists, particularly not where the elderly are concerned. The reasoning seems to be that friends are not as relevant as other cohorts among older people. For example, Rosow (1970) states that older people rank friends behind children and neighbors as general reference groups; and he concludes that "this expresses their relative attractiveness and general significance to the aging."

It is not the intent of the author to argue whether family, friends, neighbors or some other cohort is most socially important to elderly individuals. The point is that, as a potential focus for extended study, friendship offers a topic that (1) is not widely dealt with in the literature on older people and (2) has been a variable of some importance throughout the lives of people who currently happen to be old.

Thus the second phase of this research has concentrated on the area of both current and prior friendships as recounted by older people within the multiple contexts of their lives. A primary objective has been to construct a friendship typology that might apply to many chronological life stages. This paper will delineate a preliminary typology and some of its implication. Additional research is planned in order to refine the topology, test the validity of implications already identified, and develop additional hypotheses.

Participants in the Research
It was arbitrarily decided to limit respondents to older people who maintain their own households. The author (hereafter called "the researcher") was fascinated by the fact that this is one of the largest segments of the elderly but is a relatively minor factor in literature on the elderly. Thus, it should be kept in mind that the results of the research refer to well elderly who are sufficiently independent to retain control of their own households.

The exploratory phase of the research involved eight elderly people:
•    Four females (one married, three widowed)
•    Four males (two married, one widowed, and one confirmed bachelor)

Within the constraints of an eight-person sample, they constituted a cross-section of middle-class and upper-middle-class elderly. Extreme poverty was not represented.

Participants were recruited by a combination of random telephoning, cooperation from formal social organizations, and an informal network involving relatives of the researcher's acquaintances. Respondents were told in advance that they would be participating in three interview sessions, lasting approximately one hour each. They were requested to make arrangements to be alone in their homes during this period, to facilitate candid one-on-one conversations.

Not all complied with this last request.. However, even when a spouse was present in the home, it was usually possible for researcher and interviewee to isolate themselves. Several exceptions to this procedure were permitted in the second and third sessions with certain respondents, for reasons to be explained shortly.

Exploratory Interviews and Evolution of the Information-gathering Mechanism
Under the direction of Dr. John B. Williamson of Boston College's Sociology Department, the researcher embarked on the exploratory interview, as did several other individuals. It was decided at the outset that face-to-face interviews were the only interviewing format that would hold any potential for eliciting frank responses from older people. One function of the early interviews was to ascertain what degree of structure should ideally be used. A compromise, semi-structured approach was selected; and the researcher initially followed a list of topics in conversational-type questioning, then asked several structured questions of a scaling or ranking variety.

In the researcher's experience, the conversational portions of his interviews were much more effective than the structured questions. (It should be pointed out that, prior to this study, he had had substantial experience involving more than 2,000 unstructured and structured interviews.)

The most critical difference between the two types of inquiry seemed to involve equivalence of meaning between interviewer and interviewees. An unstructured questioning format allowed greater latitude for the researcher to sense and correct misconceptions on the part of himself or respondents. Few of the conversational responses revealed any inconsistencies with previous replies-often an indicator of potential misunderstanding.

On the other hand, when three men were asked (independently) the reasons for responding to a structured question that their financial situations had been "getting worse" during the past few years, these answers emerged:

    "With my pension and some Social Security and a little 'under the table' work,    I'm actually doing better than before; but I thought that question meant old people in general."

    "We both still work, and we are making at least as much as we did five years ago-probably more-but inflation is tough on us like everyone else."

    "I stopped work last year, so my own income is much less now; but this place is cheaper than our house was. And we get some money from our kids, so we're really better off in the long run; and we have more time together."

One female who initially answered that her financial situation had been "getting better" gave the following rationale for her response:

    "I really have a tougher time making ends meet than I did a few years ago, but so does everyone else; and I really don't like to complain."

Thus it appears that at least some structured questions can pose problems in terms of eliciting a high degree of response validity.

The researcher, out of curiosity, asked four of the initial respondents a few questions used by the Louis Harris organization in a study on aging conducted for an arm of the United States Senate (1975). He concentrated on factors ostensibly grounded in objective measures.

One Harris question had determined that 67% of the total adult public (all ages) believed most people over age 65 spent "a lot of time" watching television, while only 36% of older people concurred with this judgment. The researcher asked his four respondents what they considered to be "a lot" of television viewing; and they gave a mean answer of 25 hours per week.

During the same period, a non-published survey conducted for a Greater Boston television station indicated that the actual median weekly television viewership for adults (all ages) was only 10 hours. It seems highly possible that differences between estimates of older and younger Harris respondents were based more on disagreement as to what is "a lot" than on wide variations in the number of hours each age group would have guessed. The researcher concluded that the possibility of unequivalence of questions' meaning between researcher and respondent was a compelling reason to conduct the remaining interviews on an entirely conversational basis.

Because the initial interviews were entirely exploratory, it was determined that the best means of identifying a topic for more intensive study was to encourage respondents to discuss their lives in retrospect. This was not an easy task, and a key asset of the conversation approach was the flexibility it offered the researcher in probing for the most fruitful discussion sequence.

It might be assumed that the most orderly and least confusing path would be to begin discussing one's past and to progress chronologically to the present. Reversing the process produced much more information in a shorter time span. There were fewer inconsistencies or contradictions in respondent accounts.

Apparently there is something reassuring about beginning with the present before tackling recollections that are difficult to capture. In fact, respondents seemed to relish discussing the present, which eased their immersion in the interview. It became a simple matter for the researcher to establish when the "present" was thought to have begun and later to extend the conversation to the progressively more distant past.

Conversely, when the researcher attempted to initiate the interview by asking for early recollections, the request often appeared to be abrupt or daunting. Some minds almost seemed to freeze, and comparatively little information was forthcoming-even when the conversation reached the present.

While enhancing the quality of the interviews, the "sequential regression" conversation was not foolproof. Some respondents still experienced moments of confusion; but it is interesting that, in attempting to reorient themselves, they used the present rather than the past as a reference point.

It should be noted that this researcher's experience was generally different from those of associates engaged in a similar venture. These researchers felt more comfortable with a substantial amount of structure in the interview; but they also complained of receiving a large percentage of vague and / or mundane information. It was this researcher's conclusion that familiarity with a given information-gathering technique will naturally make a researcher more comfortable using that technique. The associates did not claim much experience in conversational interviewing.

The first session of each interview encounter consisted of working back through the years to identify important events, with an emphasis on timing and duration. If a respondent seemed inclined to discuss a given event in depth, the researcher was encouraging. The researcher concentrated on developing an overview of the life history outline during this session. More detailed information would be sought in ensuing sessions.

After the initial interviews with the first two respondents, the researcher realized that questions using terms such as "events" or "turning points" or "life stages" or "crisis" tended to produce a respondent focus on vital statistics, e.g. births, deaths, residential patterns, marriages and retirement dates. References to people and relationships were almost solely limited to these events.

Consequently, questions were altered by phrases such as "influences in your life" or "anyone or anything that was a part of your life." Conversations rather dramatically took on more of human-relations tone.

The researcher remarked about this phenomenon to the first two respondents. Each indicated that he had assumed, from the way the questions had been worded originally, that the interviewer was principally interested in events and dates. However they were more than willing to talk about people when encouraged to do so.

To avoid over-taxing respondents, the researcher ended each session after one hour and asked the respondent to make notations of any additional insights that occurred prior to the next session. In some cases, interviewees were also given specific "homework" assignments, e.g. looking up records of key dates or photographs that might stimulate memories.

The second session began with the researcher soliciting interim thoughts, which almost every respondent offered. While some of these observations dealt with corrections or accidental omissions, the majority revealed information that had been withheld intentionally during the first session-typically involving family disruptions (e.g. children's divorces).

A key function of the initial session was to nurture the incipient confidant relationship between researcher and interviewee. This was another good reason to defer in-depth discussion until the second session, when conversation would benefit from the "gestation of trust" between sessions.

The researcher brought to the second session a chart entitled "Personal History Time Scan." The left-hand column included all the calendar years of the respondent's life and the respondent's age at each annual checkpoint. The document contained a chronological listing of all items volunteered by the respondent during the first session.

The list was reviewed by the interviewee and was amended as appropriate. This process not only offered an opportunity to recheck chronology, accuracy and completeness; it   also had the effect of reorienting the respondent to the interviewing process and reaffirming the researcher's keen interest in the respondent.

With a chronological framework of the respondent's life having been established, the discussion in the second session began with early years and moved progressively toward the present. This meant that, over the course of the first two sessions, the respondent was exposed to reflections beginning with both the present and the past.

During the second session, most developments on the list were discussed in varying degrees of depth, with emphasis on how and why events occurred and what impact they had imparted on the respondent's life. The third session continued this process and ended with the researcher summarizing, based on his understanding, all the important elements, relationships and influences on the respondent's life. The respondent, by now invariably desiring to be of maximum assistance, usually had some comments, corrections or amplifications.

Each session was somewhat intensive, since much ground had to be covered in a relatively short period. In fact, "intensive interviewing" was a term coined by the author in The Research Craft-a social research text (Williamson et al., 1977). One of the chief requisites and strengths of the method is that the researcher identifies the most efficient conversational paths for eliciting a large amount of information with an economy of time and without communicating a potentially inhibiting sense of urgency to the respondent.

Extension of the Research with a Focus on Friendship
After completion of the exploratory interviews, the researcher decided to continue the interviewing process, but with more of a concentration on the topic of friendship, among a similar mix of eight new respondents:

•    Four females (one married, three widowed)

•    Four males (three married, one widowed)

•    The initial session continued to deal with a present-to-past recounting of all recollected occurrences and relationships that had had an influence on the interviewee's life.

•    The second and third sessions probed for more detailed and in-depth information, but they now focused more narrowly on relationships.

For example, whereas an exploratory interview might have dwelt for some time on the process of retirement and its implications, a "continuation interview" concentrated more on the important people in the respondent's life immediately before, during and after the time of retirement.

The information-gathering mechanism remained highly productive. Respondents were enthusiastic, candid and painstaking in recalling the past with what appeared to be a high degree of accuracy. Any apparent inconsistencies in their stories were explored conversationally in a manner that avoided any potential respondent embarrassment.

In three instances, spouses were allowed to attend the second and third sessions. The researcher initially feared collusion between husband and wife. To the contrary, there was a substantial amount of disagreement-almost an implicit competition to determine which party could be more honest. While a useful check on candor, this phenomenon could inhibit or bias some contributions. The author concluded that dual partner participation should be considered on a case-by-case basis-and never during an initial session.

Preliminary Findings
Among other types of analysis to which the information has been subjected, the researcher has attempted to develop a topology of friends during each of three general periods of life:

•    Early adulthood (roughly ages 20 to 40)

•    Middle-age (roughly ages 40 to 60)

•    Later years (roughly ages 60 onward)

Six types of friends have been very tentatively identified:

Empathic Friends
These are generally people with very similar needs and values. They tend to be carefully selected, cultivated, and guarded almost jealously. They are very likely to be people of the same sex, although the incidence of couple friendships is higher in this category than in any other.

Functional Friends
Functional friends play a specific role in one's life and retain their importance as long as that role continues to have validity. Many functional friends are neighbors. This is the friendship category in which older people seem to feel most legitimized in crossing gender and generation boundaries. There seems to be some security in being able to rationalize such friendships on the basis of their pragmatic utility. Normally, encounters with functional friends are tightly regulated. There may be a scheduled time to "check in" with one another each day. A biweekly shopping trip may become a routine event.

Role-perpetuating Friends
This category includes cohorts who serve as reminders or reinforcements of some role a person holds or held in life. Former co-workers serve as one example. Members of a professional or fraternal association might be another. One respondent revealed that she gravitated toward friends she considered less astute than she. Her explanation was that the relationships reinforced her feelings of intellectual superiority and satisfied her friends' need to be around someone more knowledgeable than they.

Contingency Friends
Contingency friends have latent potential as empathic friends; but, for one reason or another, they are relegated to "reserve" status-in essence an empathic minor league "farm team." They may be elevated to the empathic level if an empathic friend is lost or unavailable; but they are generally regarded as not warranting the emotional commitment invested in empathic friends. They are usually known through organizations and may be identified more with the organizations than as individuals. They may fill many roles, ranging from family substitutes to relief from boredom when empathic friends are not available. This is a very flexible, low-maintenance category that may even be ignored for extended periods without compromising the contingency nature of the relationship.

Holdover Friends
These are friends inherited from an earlier period and with little else to recommend them. An example of a holdover friend might be a previous comrade of a now-deceased spouse. Another might be a former college chum who continues to correspond although the two parties have virtually nothing in common. If a holdover friendship had the same validity as when it had been initiated, it would probably fall into the empathic friendship category. People in the holdover segment persist as friends largely because one or both parties are either afraid or embarrassed to terminate the relationship. While this topology exercise deliberately deals only with non-relatives, it is worth mentioning that family members can sometimes fall into the holdover friend category--forced together by familial bonds but growing increasingly apart with the passage of time, geographic separation, etc.

Peripheral Friends
These are really acquaintances, not friends. As a rule, peripheral friends are forced upon one another by circumstances that are not easily controllable. They are often friends of friends who are tolerated as the price of friendship. Organizational acquaintances frequently occupy this category if they are not deemed to have the substance to be contingency friends. Many co-workers are peripheral friends who may get together for a drink after work but would never see one another again if either leaves the job that links them.

These are very tentative groupings, and there is undoubtedly overlap whereby someone might fit more than one criterion, although the typology structure has not been validated among a large sample of older people. Moreover, the typologies seem to be somewhat fluid; a given individual may fluctuate between friendship types, depending on a variety of circumstances.

During the course of the interviews, the researcher explored ways of ascertaining the relative importance of various friends in the lives of older people. Some of the qualitative differences are reflected in the topology. However, there were also quantitative differences, which proved reasonably easy for respondents to estimate and centered about the amount of time invested in each friend. The concept of time expenditure is paramount in the minds of many older people who feel that the term "investment" is at least as appropriate for time as it is for money.

By asking respondents how much time they spent with various friends and acquaintances, the researcher has been able to derive a very rough ranking of these six friendship categories for each of the three major periods cited earlier: young adulthood, middle-age and advanced age.

For this purpose, the act of "spending time" has been defined as time physically together or on the telephone or even writing letters. There is no intended suggestion of a direct relationship between time invested on a friend and the quality of that relationship. Friendship types have been listed in order of importance among older people.

It is interesting to note that, as time passes, increasing proportions of time are spent on empathic and functional friendships. In stark contrast with their more youthful periods, they spend least time with peripheral and holdover friends. Moreover, most of the respondents indicated that they were more satisfied with their non-familial social structure than in earlier years.

The empathic and functional friendships that older people most embrace are likely to have been formed instinctively and purposefully, whereas many peripheral and holdover friendships emerged by default. It appears that the most socially fulfilled older people are those who spend most of their time on the types of friendships over which they have had greatest control at the initiating and nurturing stages.

These indications lead to a preliminary hypothesis:

Older people exercise greater control over their social destinies than they did when they were younger.

If this hypothesis is supported by more rigorous research, it can have important ramifications.

For example, the older person's narrowing social world, as posited by disengagement theory, is frequently interpreted by some social scientists as negative, confining and even demeaning. In actuality, this narrowing may represent an opportunity for older people to be more selective in forming or perpetuating friendships.

Several respondents recounted tales of relationships they wanted to terminate and indicated that, years ago, they would have been reluctant to do so. Now they have both the motivation and the means. One man stated:

    "I can't afford to waste one minute on someone I don't like much."

A woman commented:

    "I have all the excuses in the world for refusing to see someone I don't like. I'm old and decrepit-right?" (She rose and did a little dance at the end of this sentence.)

Another man said:

    "If my friends don't like what I do, I just change my friends. It's sort of like what some people do with doctors. Old people have a much easier time making new friends than you young people think."

Additional research, if it continues to support the aforementioned hypothesis, may bring more clearly into focus differences between what might be termed entrepreneurial and corporate approaches to friendshipping. The key variable would be the amount of control and selectivity exercised by an individual in shaping his or her social environment. An extension of this theory would be that people of all ages who exercise strongest social control and flexibility will be the most socially satisfied and secure.

In a similar vein, further research should explore the concept of the "social insurance" represented by contingency relationships. This venture should examine the potential value of preparing for friendship disruptions, and the potential importance of contingency relationships, in this preparation.

Overview
This preliminary research investigation among people in their later years offers several reasons for excitement. Methodologically, it appears possible to obtain candid and quite accurate retrospective life-history accounts through a careful, multi-stage conversational interviewing technique.

This method works best when it first deals with the present and works back through the years. It benefits greatly from a chart-type summary which provides an accuracy verification mechanism-a focal point for additional in-depth conversation, and an opportunity to reorient ensuing discussion so it follows a more logical past-to-present format.

Theoretically, the preliminary findings suggest a six-category topology of friendship, which may potentially be applicable to any age group. Further analysis produces an hypothesis that older people exercise greater control over their social destinies than they did when they were younger. Validation of this hypothesis through further research may alter previous notions that the narrowing of one's social world (e.g. through disengagement) is necessarily an involuntary, negative, demeaning, and irreversible phenomenon.