Jay Sokolovsky and Carl Cohen
[Originally published in C. Fry, ed., Dimensions: Aging, Culture and Health,  Bergin, 1981.]

1.   Introduction

Recent research on America's aged population has centered on the extent
and nature of support provided by their interpersonal relationships. This
concern has flowed from two interrelated social science perspectives dealing
with the lives of older adults. The first theoretical perspective generated
mostly from sociology and psychology has long sought to understand the
impact of growing old on social interaction, the performance of roles and
the impact on psychological states. A second perspective, emanating from
social work has tried to understand the practical implications of informal
social support for delivery of services to the aged living in non-
institutionalized environments.

In both sets of research, the issue of social isolation has been a central point
of discussion. This is apparent in research describing the poor urban
elderly, especially residents of inner-city single room occupancy (SRO)
hotels. Much of the literature has generated an axiomatic perception of such
elderly as being almost total isolates, lacking personal networks of any
significance and possessing a "nominal culture" (Hertz and Hutheesing
1975). It will be the contention of this chapter that certain aspects of these
assertions are a convenient myth, fostered by both (1) a failure to examine
the totality of interaction of the aged and (2) the lack of concern for the
cultural significance of their social linkages. To demonstrate this, we have
used an anthropologically based network analysis to examine quantitative
and qualitative aspects of social interaction for a particular aged center-city

We will also briefly suggest how the knowledge gained from such analysis
can assist community agencies and other professionals in maximizing their
assistance to the elderly.

2.   Social Isolation, Networks and Growing Old

The coming of old age presents some universal problems for maintaining
interpersonal relations. As death removes coevals and older persons from
the various linkages in which a person was active, some shifts are expected
in the interactional relation between and within age groups. Naturally
occurring decrements in health and mobility also potentially limit the
physical activity necessary for many forms of human contact. Social
isolation in the latter stages of life may thus be considered a logical
possibility for many entering the last stages of the life course.

Most studies of isolation in old age have centered on the extent of
interpersonal isolation from family, work and friends. The largely
discredited disengagement theory (Cumming and Henry 1961) argued that a
universal, progressive, interpersonal isolation of the aged occurred, and was
functionally linked to the needs of both the aging individual and the ever-
changing society. While at first glance the process of growing old may seem
to necessitate a great decrease in the size and complexity of personal
linkages, this is not necessarily the case. Anthropological studies of certain
non-Western societies show that the elderly can often retain and even
intensify their cultural linkages into familial, socio-political and sacred
spheres of activity by: (1) creating new younger-generation contacts through
adoption; (2) marriage of a young spouse; (3) sponsorship of initiation rites
(Guemple 1969; Hart and Pilling 1960); or (4) changing the content of
behavior in pre-existing role relationships, such as from executor to
mediator (see especially Guttman 1974, Turnbull 1965). In certain societies--
such as the Kapauku Papuans (Pospisil 1971) of Melanesia, or the
Abkhasians (Benet 1974) of the Caucasus--for women especially the onset of
old age means the lifting of taboos on various types of social intercourse
and a consequent increase in interaction beyond the household. Yet It Is
also clear that those aged, particularly in certain hunting and gathering and
horticultural societies, have relatively small, weak, interpersonal support
systems, are in tenuous social and economic positions, and may be the focus
of intergenerational conflict via accusations of witchcraft. (Pospisil 1971;
Holmberg 1969).

However, it is in industrialized, capitalistic societies like our own that the
aging process is most often equated in the gerontological literature with
irreparable disjunctures in networks, especially in familial and occupational
role relationships. It is often presumed that our culture--with high levels of
industrialization and an emphasis on youth and the isolated nuclear family--
has deprived our aged of interaction and support available in more
traditional societies. The overall impact of these social losses is summed up
by Rosow, who asserts that "our current institutions do not offer a choice
between marginality and integration of the aged but simply between
alternative forms of alienation" (1974:168). Given such a statement, there is
little surprise in noting that perhaps the most distinctive social
characteristic attributed to America's urban elderly is social "isolation."
This appears to be most emphasized in studies of the inner-city elderly, who
have been identified as a uniquely troubled category of our aging
population. (Carp 1972; Clark 1971; Clark and Anderson 1967; Lawton and
Kleban 1971; Lopata 1975; Stephens 1976). In studies of poor Jewish aged
in a Philadelphia slum area called "Strawberry Mansion," Lawton and
Kleban (1971) describe them as having poorer health, lower morale, very
limited mobility, and low levels of interaction with family and friends, in
comparison with other groups of aged. Clark, an anthropological pioneer in
studies of the inner-city aged, remarks that the social interaction of inner-
city aged is "simple, seldom intimate and narrowly circumscribed"
(1971:64). However, Clark is appropriately cautious, and notes that
virtually nothing is known about these social networks, and that our
culture's preoccupation with productivity "has trained us to dismiss such
people from our range of perception" (p.64).

By far the strongest statements about isolation in the later years stem from
the growing number of studies of dilapidated center-city hotels, to which the
elderly poor have flocked. Starting with the early studies of Zorbough
(1929) and Hayner (1936) center-city hotels have been depicted as
representing the ultimate negative effects of urbanism on social life. These
studies report pervasive levels of extreme anonymity and isolation from
personal contacts and general social activity. Some recent studies centering
more specifically on the elderly have come to similar conclusions. The most
emphatically stated is the work of Stephens, based on a research sample of
100 tenants in a downtown Detroit SRO, the Guinevere. She views the
elderly as living in a "society of the alone," dominated by social "atomism"
and norms of freedom, privacy and utilitarianism. She describes virtually
the entire aged population (90 percent male) as strongly alienated life-long
"1oners," who "have broken all ties to family, friends and for the most part
do not attempt to replenish what was already an impoverished repertoire of
social relations" (1976:91). The few ties that do develop are claimed to be
non-intimate and utilitarian, based on either economic exchange or leisure-
time activities in public places. It is also maintained that those few aged
women who live in the Guinevere have a particularly hard time in
generating even the few utilitarian ties which the men maintain. Another
recent study echoes Stephens' analysis, particularly with regard to female
SRO residents. Lally et al. (1979), studying a number of Seattle hotels,
focused on a small sample (16) of older women who had higher than
average education for their age cohort, and a majority of whom had
worked in traditionally male occupations. Based on interviews, the authors
depict these women as having only limited functional ties with hotel staff,
and nearby business proprietors as respondents "consistently claimed to
neither be friends with or even know other women in their hotels" (p.70).

Despite the emphatic denials of much personal contact occurring for the
urban aged, when studies have been designed to measure the extent of
social interaction, the results have been surprising. Starting with the studies
of intergenerational family relations by researchers such as Sussman and
Burchinal (1962), Rosenmayer (1977) and especially Shanas et al. (1968,
Shanas 1979) it has been found that despite the high levels of mobility,
diversity of interests and preponderance of separate dwelling patterns for
urban families that "family help, particularly in time of illness, exchange of
services, and regular visits are common among old people and their
children and relatives whether or not these live under a single roof"
(Shanas 1979, Shanas et al. 1968). What has essentially been found is that
the structure of the family network, rather than having disintegrated, has
adapted to high levels of change and diversity by maintaining generational
"intimacy" at a distance through the maintenance of frequent contact and
the exchange of mutual services.

In similar fashion, a limited number of large-scale studies concerning the
social linkages of the inner-city aged have indicated low levels of extreme
isolation. Marjorie Cantor's study (1975) of a cross-section of New York's
poorest residents (sample of 1552 persons) provides data to strongly
indicate the "importance of friends and neighbors in the life of urban
elderly" (p.25). More than 60 percent know at least one neighbor well, and,
importantly, "live alones" are significantly more likely to report knowing at
least one neighbor well. In strong contrast to previously discussed studies,
50 percent of the sample maintained intimate contacts and two thirds saw
their children monthly (p.25-26). The study also notes that only eight
percent appear to have no significant personal support system, although
some of these "true" isolates have contact with community organizations or
religious institutions.

Other studies of center-city aged -- Tissue in Sacramento (1971), Lawton,
Kleban and Singer (1971) in Philadelphia, Bild and Havighurst (1976) in
Chicago and Ehrlich's (1976) St. Louis hotel research--while not indicating
the high interaction rates found by Cantor, show that many elderly
maintain at least a few close personal ties. Anthropologically oriented
studies of hotels In San Diego (Erickson and Eckert, 1977), New York's
Upper West side (Siegel, 1977) and the Mid-town area (Sokolovsky et al.,
1978; Sokolovsky and Cohen, 1978), provide cultural and empirical data
disputing the claims of total isolation for the SRO aged.

What is emerging from a growing body of data on the inner-city aged is a
confounding picture bearing on the question of the extent of their social
isolation. It is the contention here that the glaring contradictions are not
simple matters of variations in sampled populations. Rather, the
inconsistencies stem more from a failure to both secure reliable data on
total social networks and to place this data in its proper cultural context.

3.   Social Interaction and Gerontology

The methodological issue of empirically defining social networks is of great
importance in gerontological studies. Measures of social interaction have
been utilized in studies of the elderly in testing various theories relating
levels of sociability with: (1) the aging process itself (Cumming and Henry
1961); (2) community creation (Johnson 1971; Hochschild 1973; Ross 1977;
Keith 1979, 1980a); (3) loneliness (Townsend 1957, 1968); (4) life
satisfaction (Rosow 1967; Lemon, Bengtson and Peterson 1972); (5)
adaptations to total institutions (Granick and Nahemow 1961); and (6)
measures of mental health (Lowenthal 1964, Clark and Anderson 1967).
While researchers in social gerontology strongly advocate the need to
understand the social networks of the elderly (see especially Lowenthal and
Robinson 1976) virtually no research has investigated the total extent of
personal ties maintained by the aged.2

Moreover, in the limited number of attempts to research the entire scope of
elderly social interaction, little concern has been taken to delineate the
relative importance of given relations. For example Townsend (1957),
followed by Tunstall (1968) has taken a behavioral approach in measuring
the societal integration of his British elderly sample by a scale based on the
number of social contacts per week (social contact score). Although these
excellent studies are some of the few attempts to measure the total extent of
elderly interaction, the significance of each social link is essentially
determined by the frequency of contact.

Granick and Nahemow (1961), trying to measure the isolation of Jewish
Home elderly populations during adulthood and just prior to entering the
home, devised a more complex scale. This is based on a restricted measure
of contacts in a person's prior job career, marital status, kinship
connections, friendships and organizational activity. They considered anyone
under the mean figure for each scale to be relatively isolated. However,
such a scale can underrepresent the social interaction of those not having a
large family. This is particularly problematic for many inner-city elderly,
who have life histories showing early disjunctures from much of their
kinship networks. By their scale, for example, a person, all of whose close
kin members were dead, but who maintained 20 intense non-kin
friendships, would be considered isolated.

A more ambitious approach was taken by Clark and Anderson in a study of
elderly persons in San Francisco. They sought to establish a measure called
"social interaction level" which intended to determine "to what extent does
the subject extend himself into the social world around him?" (1%7:49-50).
This scale combines levels of personal, economic and cultural activity. Yet,
as the authors note "no effort was made to distinguish these levels on the
basis of the duration, content or frequency of the social interaction"
(1967:151). Still, it was clearly shown that for the community elderly gross
measures of social interaction, rather than social roles, were most important
in explaining variations in mental health.

Other approaches have used simple enumerative sociometric techniques,
eliciting data by questions such as, "How many good friends do you have?"
or, "How many relatives have you seen in the last week?" However, as we
have argued in greater detail elsewhere, (Sokolovsky and Cohen, In Press)
such methods not only fail to illuminate the complexities of social in-
teraction but tend to inaccurately portray actual behavior.

This observation is replicated by the work of Killworth and Bernard (1976),
which casts serious doubt on simple enumerative methods. In an intriguing
study of teletype communication among the deaf elite of Washington, D.C.,
persons were asked to rank others they "talked" to (via teletype) on a
"most to least" scale. In comparing the cognitive perception (who do you
talk to, most-least?) with behavioral action (recorded frequency and length
of communication), it is shown that neither the amount of communication
nor the number of alters chosen relates to the ability to rank communicants
accurately (p.226). The authors' conclusions confirm what we discovered
impressionistically, "group structure based on one-choice socio-grams are
unlikely ever to yield significant results" (p.281).

4.   Social Networks and the Network Profile

As a solution to this problem, anthropologists building on the earlier work
of Barnes (1954), Bott (1957), Epstein (1969) and Mitchell (1969) have
begun to see networks analysis as a useful tool, especially in urban settings
in which the informal ties of friendship, kinship, and neighborliness are not
easily subsumed by formally bounded institutionalized groups or categories,
(see especially Boissevain and Mitchell 1973, Boissevain 1974, Sanjek 1978,
Sokolovsky and Cohen, In Press). Network analysis has also become a
major research strategy in allied social sciences, with the work of Pattison
(1977) and Caplan (1974) in psychology, Fisher (1977) in sociology, and
Collins and Pancoast (1976) in social work.

In our work studying elderly hotel-dwellers in mid-Manhattan, we at-
tempted to avoid the pitfalls of simple sociometrics by using a hybrid form
of network analysis based on an anthropological approach to social
linkages. Such an approach focuses on behaviorally analyzing egocentric
social networks--"persons with whom a given person is in actual contact,
and their interconnections" (Boissevain 1974:25)--within an
ethnographically described cultural context.

In doing our analysis, we combined: (1) participant observation; (2) a 100-
item questionnaire, consisting of standard biographical/demographic items
and several health/psychosocial inventories; and (3) a revised "network
analysis profile" which had been developed in previous research
(Sokolovsky et. al. 1978). The "network profile" is an interview schedule
which consecutively structures questions around separate networks sectors
(hotel resident, kin); characteristics of each contact (sex, age); and the
specific behavioral (recruitment, frequency, type and directional flow of
transaction), environmental (in hotel lobby, park) and emotionally salient
(intimacy, overall importance) features of interaction. The profile
distinguished five network sectors involving social ties; ego-hotel tenant;
ego-outside non-kin; ego-kin; ego-hotel staff; ego-agency staff. Only those
links behaviorally active (including mail and telephone contact) within the
prior year are included as a part of the networks.

It is important to realize that we developed the network profile only after
four months of field work and casual discussion of interaction with SRO
dwellers. Only after such ethnographic work began to reveal the settings
and contexts in which social behavior was likely were we able to ask
culturally meaningful questions that revealed information corresponding to
our observations.

In total, the resultant methodology provides what anthropologists have
called the "first order zone" (Barnes 1972) network, and what others define
as the "support system" involving "an enduring pattern of continuous or
intermittent ties that play a significant part in maintaining the psychological
and physical integrity of the individual over time" (Caplan 1974:7). The
network profile enables us to distinguish further the "informal support
system" (also referred to here as the "personal network") from the "formal
support system." In the former support system--measured by the first three
network profile sectors--personal ties are selected by the elderly from
among residential neighbors, outside acquaintances and relatives. This
leaves for the latter support system the last two network sectors developed
within the context of a formal or bureaucratic relationship, such as tenant--
management, client-social worker or patient-psychiatrist.

The data collected on the total support system allows us to compare the
interactional and morphological variation in our respondents' networks.
Links to a given network member can be either single-stranded (uniplex)--
involving only one transactional content (e.g., casual conversation)--or
multistranded (multiplex)--having multiple contents. They may also vary in
terms of the frequency of activation, directionality (the direction in which
aid in a relationship flows), and cognitive importance. Variation can also be
noted in the interconnectedness of an ego's social relations represented by
density (ratio of actual to potential links) or by overall structural design. In
this way support systems may be understood in their quantitative and
qualitative diversity, varying in size, intensity, intimacy and morphology.
What then does this methodology show for a particular set of persons
growing old in SRO hotels?

5.   The Sample-Its Setting and Characteristics

The elderly that we studied resided in 11 SRO hotels in the eastern section
of midtown Manhattan. These hotels--as is the case of other cities where
they are numerous--are found mostly down narrow, commercially-zoned
side streets, surrounded by bars, office buildings, sandwich shops, and small
retail and grocery stores. The hotels themselves range in size from six to 12
stories, and house between 90 and 400 persons of whom 15 to 25 percent
are over 60 years of age.

Our sample consisted of 96 persons (47 males and 49 females) ranging in
age from 60 to 93 years (mean: 72 years). In similar proportion to the
elderly in general living in New York's SRO hotels, 90 percent of our
sample are white, nine percent black and one percent hispanic. The
educational level was surprisingly high, with just under a third having some
years of college and just over a third having at least some high school
education. However, this high educational attainment is also a characteristic
noted in the studies by Lally et al. (1979) and Bild and Havighurst (1976)
for the hotel populations they studied.

Perhaps the most dramatic demographic characteristic of our sample is the
fact that 39 percent have never been married. This compares with only ten
percent for the rest of the country's aged 60 or over. Almost identical
figures were found for the hotel and inner-city populations studied by
Ehrlich (1976) and Tissue (1971). Thus much of our sample has had a long
history of residing alone, with 97 percent now residing by themselves and
the mean number of years living alone being 25. On the average, 17 years
have been spent in SROs in general and 11 years have been lived in their
present hotels.

5.1. Informal Support System

The total extent of personal contacts ranged from zero to 26 persons, with a
mean network size of 7.5 individuals (see table 9:1). Males averaged only
one person more than females in their total informal support systems (n.s.),
and no correlations were found between support system size and age or
education. While only one percent had 20 or more personal ties, 89 percent
had at least three persons in their personal networks. Importantly, only two
persons fit the myth of total isolation by having no personal ties, with the
majority of residents maintaining viable and often complex networks which
allows them to live independently in the urban community.

While the average number of personal ties are not impressive, more than
half of these links involved at least minimally multiplex relations. Only six
percent of our sample lacked such relationships, with 71 percent of the
networks having three or more multiplex ties. Moreover, more than a third
of the sample had at least three very complex relationships consisting of
three or more transactions, and on the average just over a quarter (28
percent) of a respondent's total informal support system contained such ties.

The gamut of human relations is seen in these complex network components
involving casual conversing, advising, and providing emotional support;
visiting or the passing of time in a local park or by the exchanging of food,
cigarettes, drink, money and medicine; or helping in securing benefits from
intricate welfare bureaucracies. When asked to rate contacts in terms of
their importance, it turned out that about a third of all ties were considered
"very important," and 45.8 percent had three or more such relationships.
On the average, whole networks contained between two and three
relationships which were either cognitively very important or functionally
complex. A considerable smaller number of informal support systems
contained intimate ties (mean .9) with only five percent having three or
more persons to share personal thoughts with. No intimate links existed in
41 percent of the personal networks, about double the number which lacked
any very complex or very important ties. A more optimistic way to view the
extent of isolation from emotionally and instrumentally important support is
to consider that while 59 percent of the residents maintained intimate ties,
82 percent had some "very important" ties and 75 percent had at least one
very complex relationship.

One of the strongest contentions made about SRO dwellers, regardless of
age, has been that contact beyond the hotel walls is virtually non-existent
(Shapiro 1971, Stephens 1976, Hertz and Hutheesing 1975, Lally et. al.
1979). For the population under study, examining how personal ties are
distributed throughout the components of the total networks gives a
somewhat different picture.

                                 TABLE 9.1

                       TOTAL INFORMAL SUPPORT SYSTEM
                OF A SRO HOTEL ELDERLY POPULATION (N=96) *

Network                                             Percent 3     Percent
Characteristics          Mean     Range     or More       None

Total Contacts          7.5           0-26           2             89

Multistrand                 4.5          0-18           6             71

Total Contacts
Involving Three          2.6           0-16          25             36
or More

Total Contacts
Rated "Very                 2.4           0-8           18              46

Total Contacts
Considered           .        . 9           0-3           41                5

*Table 9.1 is adapted from Sokolovsky and Cohen 1978:332.

TABLE 9.2 A & B

9.2A Informal Support System: Hotel, Outside Non-Kin, Kin*

Network   Network          Mean       Range    Percent       Percent None
Percent     Characteristic                                3 or More
Hotel         Total                  2.         70-15         44                26

                  Total Three
                  or More             .8            0-6            8                 57

                  Rated "Very
                  Important"           .4            0-5            4                76
Outside       Total Contacts     4.8        0-23         68                   6
and Kin
                     Total Three
                      or More              1.8         0-16        --                   --

                      Rated "Very
                      Important"             2.0        0-12        --                   --
Outside           Total Contacts      2.7        0-17         40                 24
                        Total Three
                         or More               1.1        0-16       12                   61

                          Rated "Very
                          Important               .8        0-4           9                   51
Outside               Total Contacts     2.1       1-12         28                   29
                            Total Three
                            or More
                            Transactions          .7       0-10           8                  66

                              Rated "Very
                              Important           1.2       0-10         14                  48

      9.2B Formal Support System: Hotel Staff and Institutional Ties

  Network                 Mean      Range    Percent 3     Percent
  Sector                                                  or More       None

  Hotel Staff                 1.8           0-7        30                26

  Ties                            1.4           0-6        15                32

  *Table 9.2A is adapted from Sokolovsky and Cohen 1978:333.

  While the number of personal ties (Table 9:2A) is fairly well distributed
  among the three sectors of informal support--hotel, outside non-kin, kin-
  -almost two-thirds (64 percent) of all network links occur outside the
  hotel walls. In the hotel-network sector there was a mean of 2.7
  contacts; less than a majority had three or more hotel ties. In terms of
  whole, informal support systems, on the average 40 percent of total net-
  works were composed of hotel contacts.

  Nonetheless, just over one-fourth were totally shut off from personal ties
  in their hotel, as compared to six percent who were shut off totally from
  the outside sectors. Importantly, although the frequency of interaction
  (see below) with residential neighbors is considerably higher than that
  for outside contacts far fewer of such ties are significant for our sample.
  Less than one-third (30 percent) of the very complex and fewer than
  one-fifth (17 percent) of the "very important" links generated were
  found among other hotel residents. Moreover, the vast majority (76
  percent) did not consider any of their hotel contacts as very important
  to themselves.

  In looking beyond the hotel, the mean number of ties is slightly higher
  for non-relatives (2.7) than that for kin (2.1), and a higher percent (29
  per-cent versus 24 percent) were totally isolated from communicating
  with kin in the prior year. The highest level of functionally-complex ties
  occur among outside non-kin personnel, who include not only recently
  made acquaintances in nearby hotels, and workers and owners of the
  many small businesses in the neighborhood, but also long-term friends
  who have been known well before the respondents had entered old age.
  This is reflected in the fact that these non-kin contacts outside of the
  hotel have remained part of the informal support system for an average
  of 19 years.

  Turning to the statistics for ties to kin, it is seen that this network sector
  included the fewest number of contacts (28 percent of the total) and
  showed the highest levels of isolation from links providing instrumental
  support. Despite an extremely low frequency of interaction with kin and
  the lowest figures for three or more contacts in a total network, relatives
  are cognitively the most important part of the informal support system.
  This is seen in the fact that just about double the amount of "very
  important" relationships exist with relatives as is the case for either the
  hotel or outside non-kin sectors.

  5.2.    Formal Support Systems

  Much attention in the literature on the SRO elderly has drawn attention
  to the importance of the "anonymous service fringe," those links
  generated by formal contact with various types of institutions or
  professional care givers. We have divided up such contacts between ties
  with hotel staff and representatives of service-giving agencies, such as
  doctors, nurses, social workers, or staff of senior centers, etc. (see table
  9:2B). Lally (1979) even suggests that such links provide virtually the
  sum of all contacts her female sample of inner-city aged maintained. It
  should be clear by now that this is not the case here. However,
  considering the support system as a totality, 30 percent of all contacts
  are with these formally designated service givers. The most frequent
  formal source of contact is with hotel staff, usually the desk clerk or a
  maid, with an average of about two such contacts. Approximately the
  same amount of SRO aged (26 percent) have nothing to do with hotel
  staff, as was the case with regard to any one sector of the informal
  support system. For a small number of extremely isolated individuals
  with virtually no informal support, the hotel staff can be crucial for
  their survival. Despite the large signs at some hotel desks stating
  "Absolutely No Credit," and denials of such aid by the staff, many
  tenants rely at least partially on small monthly loans to maintain them
  through another cycle of woefully inadequate income. With rent as high
  as $200 a month, many elderly tenants have between $6 and $10 per day
  to spend on food, medicine and other essentials. As reported in other
  SRO studies (Shapiro 1971, Stephens 1976, Eckert 1979b), tenant-staff
  relationships--although potentially exploitative on the part of the staff--
  are often agreeably symbiotic. The aged tenants form the economic
  backbone of steady rental income for hotel owners, while the staff
  provides practical assistance with a minimum of emotional investment.
  In certain cases individuals told us that they would rather borrow
  money from a hotel desk clerk than endanger a close personal
  relationship by the potential conflict loaning can generate. Occasionally
  these formal ties may crystallize into informal personal ones, when
  exchange goes beyond the accepted norms of worker-tenant.

  Among aged hotel residents there is generally manifested a great deal of
  distrust of certain service or care-giving institutions, especially hospitals.
  Many tenants perceive that if they are hospitalized for even a short
  period of time they will most likely lose their independence and end up
  in a nursing home or other long-term institution. Much of this fear is
  born out of ignorance, and the fact that with understaffed outreach
  services tenants often undertake costly travel only to confront large,
  impersonal bureaucracies. Nonetheless, with a mean of 1.4 institutional
  contacts, two-thirds of the tenants dealt with, on a person to person
  basis, some professional care giver over the last year.

  6.      Comparing Interaction

  While this is the only study we know of attempting to quantify in detail
  the total support system of the aged, some of the data we have collected
  can be fruitfully compared to other studies. In terms of the size of the
  total informal support system, network sizes of 20-30 persons reported
  for a general urban population (Pattison 1977) and a younger group of
  SRO residents (Sokolovsky et. al. 1978) are several times higher than
  our findings for the SRO elderly. In the latter study of an SRO with an
  experimental program including an on-site recreational, lunch and social
  services program, such services appeared to stimulate considerably
  higher levels of sociability than is found in other hotels. In a study of
  San Diego SRO resident over age 50, Erickson and Eckert (1977) found
  residents (mostly male) have an average of 14 "acquaintances."
  Unfortunately, it is difficult to compare our figures with this study, as
  no indication is provided as to how the number of "acquaintances" was
  determined. However, it may be that the much lower rents in the San
  Diego hotels (almost half that in New York) and a benign climate may
  provide more spendable money and a more hospitable environment for

  Another way of comparing the degree of isolation is in terms of the fre-
  quency of interaction. In Table 9:3, we have applied the Townsend
  interaction scale to our sample in order to compare it to Townsend's
  (1957) working-class, urban Londoners (Bethnal Green) and the four-
  area survey study by Tunstall (1966) which includes both urban and
  rural areas of England. The results indicate relative isolation for the
  SRO population in terms of absolute numbers of contacts. Here 72
  percent of the SRO elderly fell within Townsend's "rather isolated" and
  'isolated" categories, versus 23 percent and 44 percent for the other two
  British samples. Yet for that part of the Tunstall survey that live alone,
  as almost all of our sample did, relatively more were found to be at least
  'rather isolated" (87 percent) than in the Manhattan hotels. However,
  again we would note that the comparison must be viewed quite
  cautiously, as the scale has a strong bias in favor of those working, or
  those living in family households and being in close proximity to
  relatives; the scale gives substantial positive weighting to Interaction in
  these settings.

  This indeed becomes problematic in considering the variable frequency
  of interaction by the SRO elderly within the different sectors of their
  informal support system. In comparing (Table 9:4) the once-a-week or
  better interaction from two different hotel studies and two general
  surveys, one sees that the most dramatic delimiting of social interaction
  in the SROs occurs in the kin sector. While 49 percent in Exelrod's
  (1964) Detroit survey and 82 percent of the Shanas et al. (1968) survey
  respondents communicated with relatives at least weekly, this occurred
  for 17 percent of the New York aged and for only three percent of the
  San Diego hotel population. Nevertheless a different picture emerges
  when considering the network sectors of residents/neighbors and friends.
  In both these segments of the informal support system, a reversal of
  sociability occurs and the SROs appear to equal or exceed the two
  general population surveys, although the pattern in the hotel samples
  vary in terms of which sector is most frequently interacted with. In the
  New York SROs, almost three-quarters (72 percent) interacted with at
  least one fellow resident, and under a majority (45 percent) had similar
  contact with non-kin outside the hotel, over a week's time. Of course this
  does not provide a relative notion of how many in a given network each
  ego communicates with weekly, but it shows that comparatively large
  segments of the SRO are not without any frequent contact both in and
  outside of their residences.

                                TABLE 9.3

                        British Four Area Survey**

Contact           New York   Bethnal Green*     Total         Living
Score              SROs           Sample                 Sample    Alone

35 or More
 per week        28%                 77%                  54%         12%
(not isolated)

per week
(isolated/           72%                 23%                  44%         87%
rather isolated)

*These figures are taken from Townsend (1957)
**These figures are taken from Tunstall (1966)

                                                 TABLE 9.4


                       Percent Once-a-Week Contacts
                                Residents/     Friends/
                          Kin           Neighbors         Acquaintances

New York
SROs                  17              72                      45

San Diego              3              50                      73
SROs*                 (8,33)      (38,92)             (73,96)

Population**          49              29                     28

Survey***             82               --                       40

*The figures come from Erickson and Eckert (1977); the numbers not in
parentheses are for the working class hotels which seem to be closest to
those studied in New York. The numbers in parentheses are for "skid row"
and middle class hotels, respectively.
**These figures are taken from Axelrod 1964.
***These figures are taken from Shanas et al. (1968).

7.   Network Structure, Function and Meaning

Although in this study we have stressed the sometimes drastic differences
with other statements made about the SRO aged, our research concurs with
others (see especially Stephens 1976, Eckert 1979a) who detail an ex-
aggerated normative concern for self-reliance, privacy and instrumental
social relations. A constant refrain one hears in speaking to SRO tenants
are such statements as "I'm a loner, always have been, always will be, don't
need no one, I don't bother no one, don't let anyone bother me, I just mind
my own business." One hears such statements so often in such similar form
that one soon realizes that this is a commonly accepted description of
cultural and personal identity. It is a standard defense against further
penetration by prying social scientists into the social world of the hotel tenant.

This proclamation of loner" status represents a certain understanding of the
cultural reality of being old in a SRO. Residing alone, these elderly must
negotiate daily tasks of survival independently to prevent being sent to a
nursing home. On very meager monetary resources--which are habitually
depleted before the end of a given month--they must feed, clothe and
medically provide for themselves. Group interaction and organization is
typically discouraged by hotel managers, who are afraid of any resident
groups emerging who could effectively protest the very poor living condi-
tions in many hotels. The rooms themselves--by their very tiny size--barely
provide private living space for the individual tenant. Indeed, for many of
these poor elderly, who spend a good deal of time in their rooms, intrusions
into one's quarters can take on the nature of a violation of one's physical
being. Moreover, they have little control over who will be surrounding them
in other rooms, and with the great human heterogeneity found in most
hotels tenants often have little actual basis for easy mutuality in social
relationships.  Nevertheless most of the SRO aged have developed informal
support systems by erecting personal networks to help them negotiate
survival within the constraints of their small urban world.

                                TABLE 9.5
                      OF AID FOR EACH NETWORK SECTOR

                              Directional Flow of Aid
Network Sector  Ego to Other  Reciprocal   Dependent-Other to Ego

Hotel                    10                    67                        23

Non-Kin                 6                    70                        25

Kin                         8                    63                        29

When one penetrates beyond the emic portrayal of social life by observing
and participating in it, one finds that frequent socialization (at least in the
hotels), room visiting, the exchange of food, money, practical services and
emotional support occur at least within some of the social bonds which
crystallize. The nature and structure of these bonds reflect the cultural
ideals and environmental constraints of the population. One way of seeing
this is by examining the degree of dependency manifested in the transac-
tions that tie people together.

In analyzing (see Table 9.5) the number of network sectors in which
relationships were predominantly manifested by either dependence,
reciprocity, or instrumental transactions, it is seen that in all three areas of
informal support about three-fourths of the networks studied lack a
predominantly dependent nature. The data indicates that the provision of
aid in all sectors is dominated by a reciprocal give-and-take, with no
statistically significant variation existing for the population as a whole.

                              TABLE 9.6

                 60 to 70 Years Old                    70 Years or Older

                  Ego to          Dependent Aid        Ego to                 Dependent
                  Other or                                      Other or                 Aid
              Reciprocal Aid                                 Reciprocal Aid

Poor             91                    9                             50                    50


Good            90                   10                            89                     11


                             N.S.                                                  P=.02

However, one does see (Table 9.6) an important relationship between flow
of aid and the factors of age and health. Especially, within the hotel sector
those who combine very old age (over 70) and relatively poor health are
most likely to generate dependent support within their personal networks.
Table 9:6 shows that those who are in the eldest and least healthy category
are five times more likely to have a predominantly dependent flow of aid in
the hotel segment of their informal support system than their younger,
healthier counterparts. Similar trends occur in the other parts of networks,
but they are not statistically significant for the combined effects of age and
health. While this data cannot inform us about a causal relationship
between networks, age, health and dependence, it does indicate that despite
the ideal of "self-reliance," those entering very old age in poor health
manifest considerable levels of support within their hotels.

Another important dimension of networks is their morphological structure,
and here we distinguish four types of configurations among our sample. In
Table 9:7 is shown the general distribution of these network configurations,
which we have called: diffuse-cluster; cluster; diffuse; and kin-cluster. A
majority of our respondents (51.6 percent) formed diffuse-cluster networks,
in which there existed at least two other persons who formed a cluster of
mutual interaction, while in the network as a whole less than 75 percent of
the members were in contact with each other. Such networks tend to be the
largest in total size, with the interconnected clusters typically containing two
to four persons. These maximally dense parts of a network are most likely
to be in the hotel, although they are also likely to occur among a set of
outside friends who may meet frequently at a local bar, off-track betting
parlor or inexpensive cafeteria. Yet, considered as a whole, diffuse-cluster
networks are structurally dispersed, with linkages rarely (about ten percent
of the time) transcending the boundaries of the three sectors we have
distinguished previously.

                                TABLE 9.7

                 Diffuse/      Cluster      Kin-Cluster    Diffuse        Totals
                 Cluster                        Diffuse
Total           51.6          20               14                 13.7        100
Population   (49)          (19)             (14)               (13)        (95)

Besides being structurally dispersed, most such networks show a highly
variable frequency in activating the ties that are maintained. Besides the
general variation in interactive rates between network sectors noted above,
there is an important diurnal cycle related mostly to a blighted and often
dangerous urban zone. Although two-thirds of the residents had social
contacts both day and night within the hotel, over three-fourths did not see
anyone outside the hotel after dark. There is also an important monthly
cycle, geared to the arrival of government or pension checks at the
beginning of the month. It is during the prior week, when funds invariably
run out or low, that the full extent of one's personal support system is

Within such networks--especially in the hotel sector--there may be included
"natural helpers" who serve as direct and indirect conduits of useful
information or physical aid, and may serve as a point of contact between
different informal support systems. One such man, 63 years old, known to
tenants in one hotel as "Captain" (due to his prior  seafaring occupation),
had appeared to have mastered the intricacies of New York's social service
bureaucracy. Although he tended to be reclusive, and directly helped only a
few other persons in the hotel, through these other friends and even via a
hotel maid several other tenants obliquely requested and received help from
the Captain.

The next type of configuration, cluster network, encompasses one-fifth of
the sample and involves an instance where at least 75 percent of all persons
ego is in contact with interact with each other. The majority of such
networks are focused exclusively in the hotels, with about 20 percent
exclusively located among outside non-kin, and only one example of an ex-
clusive, kin-focused cluster. As is the case with the other configurations, in
very few instances (three) did the crossing of intra-network boundaries oc-
cur. In each such case, a hotel contact was linked with another sector of the
informal support system. These cluster networks often have at least three
members besides ego, and typically represent what Boissevain has called a
quasi-group: "a coalition of persons, recruIted according to structurally
diverse principles by one or more existing members, between some of whom
there is a degree of patterned interaction and organization" (1968:550). In
some cases this may simply involve getting together at a regular time in a
public place outside the hotel to eat, socialize or to play cards. However,
especially for some elderly in poorer health and lacking mobility, cluster
configurations can provide their means of survival in the hotel.

Such was the case of Ms. Austin, a 66-year-old black woman who, due to a
bad back, retired from working as a housekeeper at the age of sixty. Ms.
Austin finds it difficult to walk, spending most of the day in her small but
neatly kept room. Once a day she goes to the hotel lobby for a short time,
but returns to her room, where she waits for the daily visits from her three
elderly friends or less frequently goes to visit in their rooms. The exchanges
involved are quite varied and include visiting, food preparation and food
borrowing, money loans, shopping, providing care during illness, and, in the
case of one of Ms. Austin's friends, the exchange of personal thoughts. Her
three hotel friends do not stay for long together in Ms. Austin's room, but
rather float in and out during the day, gossiping, sharing meals cooked on a
hot plate, or drinking some beer. Ms. Austin also maintains contact with a
female cousin in Maryland who she either phones or writes once a month,
although she eagerly looks forward each year to a two-week summer visit
down South. It is important to note that despite a great fluidity in hotel
cluster formations, marked by the changing of personnel, in general ties are
not especially transient, with the length of relationship averaging seven
years with 90 percent being maintained three years or longer. In the case of
Ms. Austin, all the members of her hotel network have been known to her
for five or six years, and she has been drawn closer to them by the death of
two other friends in the past two years.

The third type of configuration, kin-cluster, describes persons with at least
75 percent of their kin ties in contact with each other, but where non-kin
segments of a network had no mutual interaction. Found in almost 15
percent of the sample, such networks typically display the highest levels of
exchange and frequency of contact with relatives. Conversely, there is a
tendency for kin-cluster support systems to maintain few or no hotel con-
tacts, thereby enhancing even more the importance of relatives. Yet it must
be recalled that kin ties In general-while having great "Importance" com-
pared to other parts of the network-are activated least frequently, with 80
percent of the respondents having no weekly contacts with kin. Moreover,
with very few relatives either living In the respondents' neighborhood or In
weekly contact with them, such configurations can provide very tenuous
support in times of crisis.

Finally, we also have distinguished diffuse network configurations in which
no interconnections link members of ego's network with each other. Such
network structures occur in only 13.7 percent of the SRO elderly we
studied, and represent in many cases the picture of severe isolation other
researchers have noted. In a majority of such configurations, persons have
three or fewer personal ties in their entire network, typically having no
more than one contact in a given network sector. More than half of the
persons with such networks have no social relationships in their hotel, and
of those that do only two have multiplex ties which go beyond the exchange
of conversation. In fact, three-fourths of those with diffuse network
configurations have no relationship which we have defined as "very
complex." Many of these persons appear to be the "true" loners in etic,
sociological sense, whose relationships are not only minimal in quantity but
also lack the type of transactional complexity typical of strong human

8.     Summary, Conclusions and Suggestions

This chapter has examined several aspects of social interaction for an
elderly urban population whom many consider to epitomize the isolation of
America's aged. In considering this we have used network analysis to deter-
mine the total extent of the "social support system" for our sample of SRO
aged. The results do not provide an easy answer to whether or not and to
what extent the SRO aged are really isolated.

It was found that a negligible number of persons were without any personal
ties, and while the frequency of contact with kin was exceedingly low in the
comparative sense, contact with residential neighbors and outside contacts
occurred at levels comparable with the general aged population. The total
"support system" (informal and formal) averaged close to 11 members;
however, there is virtually no data available to determine how this
compares with other aged populations. In terms of the qualitative nature of
ties, while persons averaged only two or three functionally or cognitively
important links and about one person to share personal thoughts with, work
by Lowenthal and Haven (1968) indicates that even a single intimate tie can
be an important factor in deterring mental illness among the aged.

On the other hand, substantial portions of the SRO aged were devoid of
instrumentally and emotionally important ties. This element of social
isolation was particularly glaring with ties to residential neighbors and kin;
in the former case, three-fourths did not consider any hotel contacts "very
important" and in the latter instance two-thirds lacked even one kin tie in-
volving several types of transactions. While this might appear to correspond
very strongly to previous findings, a consideration of networks as a whole
show that, far from the assertions of Stephens (1976) and Lally et al. (1979),
personal ties outside the hotels are more likely to be considered of
importance by the elderly tenants.

One must note, however, that the combination of the network profile and
ethnographic research we employed appears far more sensitive to extant but
weak ties than is the case in other studies. This points out the necessity of
ethnographically examining social networks, especially among the most
socially invisible elderly. Gubrium (1974), for example, in his study of a
nursing home, notes that while the staff there perceived that many residents
totally sever relationships with the outside world, he shows that the move
from a private residence to "Murray Manor" does not necessarily eliminate
social relationships as much as change their nature. While face-to-face
interaction is sharply reduced, "hidden ties" (especially kin links) are
maintained through knowledge about previously active network members
facilitated by telephone calls and visits of friends who know the famIly and
other friends (p.104).

In understanding the lifestyle adaptation of the SRO aged, it was also
necessary to explore the sometimes elusive features of network configura-
tions that link them to their social environment. It was shown that network
morphology varied considerably, from single, dense clusters to small, un-
connected, dispersed sets of ties. But the prevalent pattern was one that
combined both these characteristics. Such diffuse/cluster networks epitomize
the cultural identity of these aged, generating a basis for support and self-
sufficiency through the provision of a relatively complex support system
which selectively maintains a few crucial ties structurally dispersed through
the urban environment.

Finally, we may ask how such information can be of use in understanding
and aiding the survival of the inner city elderly. It is relevant to note first
that community care workers are becoming more sensitive to the need to
recognize interpersonal systems developed autonomously by a client
population. A recent book on social work notes that social workers are
often trained to see the lack of pre-existing support systems and become
absorbed in organizing and maintaining formal services; therefore they may
be blind to informal, positive, helping activities that go on outside the
confines of formal services" (Collins and Pancoast 1976:25). In this light, we
believe that the precise nature of the social support system can best be
comprehended by the incorporation of a network analysis profile into
clients' charts. This will add a new dimension to the provision of services
for the urban elderly.

After mapping all personal linkages, where lack of support is noted,
organizations can focus their activities. Where networks are strong,
superfluous services can be avoided and resources directed where they are
most needed. In times of stress, agencies will be able to utilize the strengths
of the total support system to help the individual remain in the community.


1. The authors wish to thank Alan Laskow, Lynne Sten, John Stern, Ann
Avitobile, David Burns, Mary Beckor, Mike Braverman, and the Murray
Hill SRO Project for their assistance. Special thanks to the residents of the
SRO hotels we studied, who really made this research possible. The
research was supported by grants from the Ittleson Foundation, the New
York Foundation, the van Ameringin Foundation, Commonwealth
Foundation, NIMH Center on Aging Grant No. 1-RO1-MH31745, and by
computer time provided by the University of Maryland, Baltimore County.

2. Possible exceptions to this statement are the work of Cantor (1976) and
Babchuck (1978-9). However, both works focus on either very frequent ties
or those considered emotionally very important.

3. After interactional data concerning a given contact were elicited, the
respondent was asked to rate the overall "importance" of the contact based
on a four point scale: 1-"not important", 2-"important"; 3-"very
important"; 4-"the most important a person can be to you."