Gender, networks, and adaptation among an inner-city population

GENDER, NETWORKS, AND ADAPTATION AMONG AN INNER-CITY POPULATION CARL I. COHEN SUNY Downstate Medical Center JAY SOKOLOVSKY University of Maryland JEANNE TERESI New York State Psychiatric institute DOUGLAS HOLMES Community Research Applications ABSTRACT: In zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDC varied cultural frameworks men and women in old age have shown distinct differences in their ability to engage in social interaction. This article examines the sex dtfferences in social interaction among 133 inner- city elderly living in SRO hotels. Contrary to earlier reports of inner- city elderly women, there was no evidence of pervasive isolation and longitudinal data indicated that SRO women had dynamic networks that allowed for the replacement of lost contacts with new ones. In comparing the sexes there were similarities in the various social network dimensions and the use of these networks to ful$ll needs and cope with physical illness. M oreover, the similatities in life histories between the sexes would suggest that as women’s life- sty les in the general population come to approximate those of men, the social interactions of aging men and women may more closely resemble each other. INTRODUCTION In varied cultural frameworks men and women in old age have been shown to have quite distinct adaptations to aging over the life cycle (Cool and McCabe 1983; Kerns 1983; Peace 198 1). One noted area of differentiation has been the ability to engage in informal social relations beyond those rigidly ascribed by tradition or one’s place in society. Some studies of non-Western societies, even those strongly dominated by Direct ~11correspondence to Carl I. Cohen, Professor of Psychiatry, SUNY Downstate M edical Center, Box 1203, 450 Clarkson Avenue, Brooklyn, N. Y. I1 203. JOURNAL OF AGING STUDIES, Volume 2, Number 1, pages 45-56. Copyright @ 1988 by JAI Press, Inc. All rights of reproduction in any form reserved. ISSN: 0890-4065. 46 JOURNAL OF AGING STUDIES Vol. ~/NO. l/l 988 males, indicate that the greater ability of women to forge intense long-term personal ties in both household and community domains is linked to greater life satisfaction and support especially in the final stages of old age. (See especially Harrell [ 198 1] for the case of Taiwan, and Cool and McCabe [ 1983 J for the cases of Corsica and Lebanon.) In the United States, data for the aged on this subject appear to be situational to the environmental context. Studies of generalized urban populations point to a greater capacity of older adult females versus males to not only develop more intimate friendships (especially among nonrelatives) but also to be able to replace defunct relationships with new ones (Powers and Bultena 1976). However, various studies of poor, inner-city aged populations have produced quite different findings. Here, older women have been commonly depicted as not only lacking social resources compared to men but having a diminished capacity to compensate for lost relationships or deal with crises such as widowhood or health problems. Lopata (1975) in her study of aged widows in Chicago’s inner-city, describes these women as lacking “the ability to voluntarily engage in supportive relations” (p. 35). Similarly dramatic statements have emerged from the study of aged women living in single-room occupancy (SRO) hotels. Lally, Black, Thronock, and Hawkins (1979) in studying several Seattle hotels found that the elderly female respondents “consistently claimed to neither be friends with or even know other women in the same hotel” (p. 70). In her well-known study of a downtown Detroit hotel Stephens (1976) stated that in general the older residents (90% male) “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” (p. 91). Amidst this dismal picture women residents are said to have a greater difficulty in developing even the few utilitarian ties which the men generate. For the SRO elderly this issue of social interaction is of singular importance as many needs are met on the basis of informal mechanisms of social interaction. Moreover, physical illness can diminish the elderly SRO resident’s prospects for remaining in the community. The availability of social networks to help sustain the person in the community is of paramount importance. In an earlier pilot study of SRO aged using cross-sectional data we postulated that the more unhealthy men tended to lose their network members whereas the less healthy women tended to increase their social ties (Cohen and Sokolovsky, 1979). We suggested that a longitudinal investigation was needed to confirm these findings and to assess what implications this might have regarding gender differences in adaptive patterns among inner-city elderly. Finally, with respect to aging in general, a study of SRO elderly may provide a tentative rejoinder to several questions posed by Leavy (1983) in his recent review of social network studies. He observed that although most investigations have found significant differences between the social networks of men and women (e.g., aging women tend to have larger, more intimate, more stable ties), it would be interesting to know whether feminists or androgynous persons might differ from other women in the support systems that they have and desire. Also, there may be cohort effects on social network patterns that mirror the cultural changes in the options becoming available to women. Midtown Manhattan SRO hotels can serve as an excellent laboratory in which to address these questions because many of the older women have nontraditional work and marital histories that parallel those of the elderly men living in these hotels. Gender, Networks, and Adaptation 47 Among an Inner-City Elderly Population Operationally, we have addressed the issues outlined above through a large-scale, longitudinal analysis of elderly residents living in SRO hotels in midtown Manhattan. This will involve four separate analyses: ANALYSIS 1: We will examine how closely the aging SRO men and women resemble each other with respect to various demographic socioeconomic, health, and biographical data. ANALYSIS 2: We will determine whether there are any differences in social network dimensions that distinguish between the male and female residents. ANALYSIS 3: Using need fulfillment as a measure of adaptation, we will ascertain whether there are any differences between sexes in their use of social networks to meet various needs. ANALYSIS~: We will determine whether there are any differences between the sexes in network size, network losses, and in ability to compensate for network losses under varying states of physical health. METHOD Sample Using a single-stage clustering technique (Van Dalen 1973) interviews were conducted in 2 1 single room occupancy (SRO) hotels in a midtown sector of Manhattan during the TABLE 1 General Characteristics of SRO Men and Women Mean (Time 1) Males (n = 58) Physical symptoms Proportion dead on 2-year follow-up Psychological symptoms Number of days in hospital in past year Unfulfilled needs Stressful events Age Previous occupational level (SES)” Current annual incomeb Proportion married previously Years in SRO hotel Years in current hotel ‘SES: 1 = highest; 8 = lowest b Income: $2600-3899 = 4; $3900-5 *p < .02 199 = 5 -Sl .21 .32 5.14 4.85 2.50 12.66 2.69 4.90 .52 15.82 10.32 Females (n = 75) .23 .07 -.30 7.15 4.01 2.13 71.79 2.71 4.75 .48 16.46 12.19 t-values (df = 737) .82 2.41* 1.47 .36 1.35 .65 .5l .06 .74 .45 .21 1.1 L 48 JOURNAL OF AGING STUDIES Vol. ~/NO. 1 il988 period 1978-1980. Within each selected hotel an effort was made to interview all residents aged 60 and over listed on a roster furnished by the management. Vigorous attempts to recruit respondents were made by providing five dollars as remuneration, initiating a series of letters and telephone calls, using flexible interview hours, and working with neighborhood agencies and hotel personnel. The initial sample (Time 1) consisted of 161 residents. On one-year follow-up (Time 2), 133 individuals were reinterviewed. The final sample consisted of 58 men and 75 women; 93% were white and their mean age was 72.3 years (see Table 1). In examining the representativeness of the sample, we found that our sample in many ways resembled that of a national census survey of 157,000 elderly SRO residents (Haley, Pearson, and Hull 198 1). There were no differences between the two samples with respect to age (72.3 years Manhattan, 72.8 years National) and percentage having spent five years at their current address (53% Manhattan, 5 1% National), and there were only minimal differences in percentages having completed eighth grade (52% Manhattan, 42% National) and in median income ($3900 Manhattan, $4760 National). There were greater differences in sex (56% female Manhattan, 38% female National) and race 93% white Manhattan, 66% white National) that reflected the particular section of Manhattan that we studied. Instruments Physical and mental health and social functioning were determined by the Comprehensive Assessment and Referral Evaluation (CARE) developed by Gurland and associates (1977). Internal consistency reliability was high for most of the scales comprising this instrument. The alphas for the SRO sample ranged from .5 1 to .92 for 22 CARE scales. The interrater reliability for all scales ranged from .56 to 1.0. The outcome variable (Nee&) and moderator variable (Physical Health Symptoms) used in the analyses were derived from the CARE. Needs: .4 40-item scale based on Lawton’s taxonomy of needs list (1970) is used here as a measure of adaptation. The scale comprises six major areas: physical health needs, mental health needs, physical self-maintenance, instrumental self-maintenance “effectance” (e.g., recreation, creativity), and social (e.g., finance, transportation), needs. The higher the score, the fewer needs are being met. Items were based on self-report rather than external objective criteria. Illustrative of the items that were used are “Goes without bath because of lack of help”; “Essential shopping not done”; “Cannot do leisure activity because of mobility problem.” Physical Health Symptoms: This variable is the sum of respondent’s z-scores on 11 health scales that was derived from the CARE at Time 1. These scales are Somatic Symptoms, Heart Disorder, Edema, Sleep Disorder, Arthritis, Stroke, Respiratory Problems, Hypertension, Cancer, Hearing Problems, and Visual Disorders. The sample was divided into two groups: those in relatively good health and those in relatively poor health. “Poor health” was defined as summed z-scores below the 33rd percentile. This cut-off score is based on Gurland and associates’ (1983) finding that one-third of New York City’s elderly had at least minimal physical impairments. Gender, Networks, and Adaptation 49 Among an zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON Inner- CityNderly Population Two other variables derived from the CARE and used in the comparative analyses were Psy chological Sy mptoms and Stress. Psy chological Sy mptoms: This variable is the sum of an individual’s z-scores on four mental health scales derived from the CARE (depression, organic brain syndrome, alcoholism, schizophrenia) at Time 1. The higher the score, the greater the number of symptoms. Stress: A 15-item scale based on categories derived from Dohrenwend and coworkers’ life events list (1978). Events had to occur during the one-year period. Typical items included “deterioration in the neighborhood,” “family illness,” “mugging in the past year, ” “personal injury due to crime,” and “financial problems.” Items that were redundant with social network variables were removed. Various criteria have been used to examine the diverse features of social networks (Barnes 1954: Boissevain 1974; Mitchell 1969). Earlier studies of inner-city populations have focused primarily on only a few social support measures of interaction at one instance in time, e.g., number of contacts, frequency of interaction, intimates (Cantor 1979; Lopata 1975). The notion of social support tends to view “more as better.” We have utilized a broader concept of social networks that encompasses social support variables but includes variables that do not directly imply value, e.g., directional flow of interaction, density, gender distribution of network. Moreover, we have developed a longitudinal portrayal of social networks by assessing a representative number of these variables on one-year follow up. Consequently, we can observe the more proximal effects of network changes on physical health. Measures of social interaction were obtained from the Network Analysis Profile (NAP), which was developed in our pre- zyxwvutsr vio us work with inner-city populations (Sokolovsky and Cohen 198 1). The interrater reliability ranged from .83 to .92 on various subsections of the profile. The NAP comprised six sectors of interaction: self-hotel contact, self-outside non-kin, self-kin, self-hotel staff, self-agency staff, self-social institution. The first three fields are termed the “informal network.” Although the totality of the profile encompasses the respondent’s “activity field,” the present analysis excluded the self-social institution field, which represented linkages to places (e.g., churches, stores) rather than individuals. Any meaningful linkages to individuals in these places were recorded within the appropriate field of interaction. Only those linkages with a minimal frequency of once every three months for hotel contacts and once a year for nonhotel contacts were included. Four dimensions of social interaction were identified that were comprised of 19 variables. The four dimensions examined various material and emotional aspects of exchange (Interactional Data Set), quantitative and morphological features of the overall network (Structural Data Set), characteristics of the network member vis-a-vis the respondent (Member Attribute Set), and those features of the environment which have been found to influence network formation and composition (Environmental Attribute Data Set). These sy nchronic network variables are described in Table 2. Factor analysis of the 19 network variables revealed that, in general, the network characteristics were statistically independent. Only seven of the variables loaded together (numplex, very important, directionality, number of clusters, size, multiplex, large clusters) with an alpha of .76. We did not use the scale in the main analyses 50 JOURNAL OF AGING STUDIES Vol. ~/NO. 1 /1988 TABLE 2 Comparison of Synchronic Network Variables Between Men and Women Mean (Time 1) Males (n=58) Variables Description Numplex Mean number of transactions (e.g., food exchange, advice-giving) per person in respondent’s network. Sustenance Females t-values (n= 75) (df= 131) Interactional data set 2.43 2.33 .07 Proportion of network members providing at least one basic support item to respondent viz. money, food, medical aid. .43 .42 .I1 Very Important Total number of persons rated as very important or most important by respondent; correlates highly with sharing or intimate thoughts, 2.85 2.81 .25 Frequency Mean frequency 7, weekly = I) per week (e.g., daily = 2.89 3.00 .32 Directionality Predominant direction in which aid between respondent and others flow. Mean directionality was calculated for hotel, non-kin, and kin contacts, (Coding: dependent = 3, reciprocal = 2, helping = I) 2.31 2.48 I.43 Multiplex Total number of persons having at least exchanges of activities with respondent. 6.61 7.22 .59 Size Total number of persons in respondent’s 7.03 8.43 I .48 Density Ratio of actual number of linkages between persons in respondent’s network to the potential number of linkages (excluding linkages with respondent) .25 .30 1.02 Degree The average number of linkages each person has with others in respondent’s network (excluding linkages with respondent) I .42 1.55 .50 Clusters Number of subunits of network with 100% density. 2.21 2.17 .I7 Large Clusters Number of clusters members. .45 .29 1.41 Configuration Graphic measure of network interconnectivity (cluster = 1, no cluster = 2, various mixed forms = 3 or 4) 2.78 2.82 .32 Length of linkage Mean number respondent. Geographic density Proportion of network mal) from the hotel members Gender homogeneity Proportion of network sex as the respondent Age homogeneity Proportion older of contact two Structural data set in network network. with 5 or more Member attribute set of years of network contacts have known 14.20 15.30 .5 I (informal or for- .37 .41 .82 members being of the same .62 .73 2.50* members being 60 years or .54 .64 1.97** (continued) Gender, Networks, and Adaptation Among an Inner-City Elderly Population 51 TABLE 2 (continued) Mean (Time 1) Variables Males (n=58) Description Females t-values /n=75) (df=l31) Environmental attribute set Age density Dichotomous variable based on whether percentage of persons aged 60 and over living in respondent’s hotel is 20% or greater (yes = 1, no = 0) .64 .68 .48 Welfare Dichotomous variable dent’s hotel is primarily not(yes= l,no=O) .38 .31 .48 hotel Hotel size Number of rooms based on whether a public assistance in respondent’s hotel responhotel or 339 336 .19 *p < .02 **p < .05 because it has not been cross-validated. Rather, we maintained the variables within the separate data sets in accordance with the theoretical conceptualizations described in the literature and discussed earlier. In addition to the synchronic network variables, eight zyxwvutsrqponmlkjihgfedcbaZYXW network change vanizbles were created. Five of the change variables (numplex, sustenance, very important, frequency, directionality) comprised the Znteructionul Change Set. These variables were formed by subtracting synchronic variable scores at Time 1 from those at Time 2. A Structurul Change Set was also created consisting of three variables: (1) number of linkages lost in past year, (2) number of new linkages in past year, (3) compensatory linkages (i.e., whether there was at least one new linkage to replace a lost linkage over the past year). Analysis 7 In looking at the overall characteristics of the aged SRO men and women there was little that could distinguish between the two sexes (Table 1). They were approximately the same age, had the same income, lived as many years in hotels, had similar marital histories, were equally successful at fulfilling approximately 90% of their material needs, and experienced equal levels of stress. There were few gender differences in previous occupational status. Only 2 1% of women had been housewives, whereas most had worked in clerical, sales, or in semiskilled and unskilled jobs. Consistent with actuarial data in the general population, the men were significantly more likely to die on two-year follow-up despite reporting somewhat better physical health than the women. Interestingly, men had more psychiatric symptoms than women, although this was not statistically significant. Analysis 2 There were also only meager differences between the two sexes with respect to social network variables (Table 2). Women had 1.4 more linkages than did the men, but men compensated by having slightly more material exchanges per linkage (numplex). Women had somewhat more dense networks and more dependent interactions, but none of these differences attained statistical significance. Women exhibited a sig- 52 JOURNAL OF AGING STUDIES Vol. ~/NO. 111988 nificantly greater number of informal linkages with persons of the same sex than did the men. In part, this may have reflected the greater availability of women than men within the hotels that we studied. Women also had a significantly higher percentage of their informal networks composed of persons age 60 and over, although for both sexes more than half of their linkages were age 60 or greater. Analysis 3 Our aim was to determine to what extent social networks affect the ability of aging men and women to fulfill their needs. Utilizing multiple regression as the analytic tool of choice, we were able to examine the amount of variance in need fulfillment that was explained by the social network variables taken as a whole as well as by individual network variables. Two hierarchical regression analyses were performed with need fulfillment at Time 2 as the dependent variable. For both analyses, the synchronic network data sets and network change data were entered in the following order, respectively: (I) Need fulfillment Time 1, Interactional Data Set, Structural Data Set, Member Attribute Data Set, Environmental Attribute Data Set; (2) Need Fulfillment Time 1, Interactional Change Set, Structural Change Set. After controlling the Needs at Time 1, the synchronic network variables (in toto) were able to account for a greater percentage of variance in need fulfillment among the males than among the females (Table 3). However, the differences in explained variance between the two sexes was not statistically significant (p < .21). It should be underscored that the percentage of variance explained by the social networks was high for both sexes: 36% for the women and 46% for the men. The latter attained statistical TABLE 3 Comparison Between Men and Women of Increments in Explained Variance (R2) of Need Fulfillment at Time 2 in Hierarchical Regression Analyses of Synchronic Network and Network Change Variables Me n In = 58) Variable Sets (n = 75) zyxwvutsrqponmlkjihgfedcba Synchronic Network Variables .lO .06 Need fulfillment (Time I) .I1 .09 Interactional data set (Time 1) .I8 .I I Structural data set (Time I) .09 .ll Member attribute data set (Time I) .08 .04 Environmental attribute data set (Time 1) .4@ .35 Total variance explained by network sets zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP Network Change Variables Need fulfillment (Time 1) Interactional data set (Time 2 minus Time 1) Structural data set (Time 2 minus Time 1) Total variance explained by network sets Note: Variable 0009*p < .05 sets were entered in order listed in the table .lO” .16 .04 .20 .06* .17* .02 .19* Gender, Networks, and Adaptation 53 Among an Inner-City Eider/y Population signi~cance. In ~on~asting these groups with respect to inurements in variance explained by each data set, the patterns were approximately the same. The men, however, had somewhat greater increments in the structural and environmental sets. Although we were primarily interested in the effects of the total network and the larger data sets, we did find that several individual variables had significant F-values. For women, having more long-term relationships was associated with greater ability to fulfill needs. For men, more numplex linkages, a greater percentage of network members outside the hotel, more diffuse network patterns, and living in a pubhc assistance hotel were associated with greater need ful~llment. These findings suggest that a broad, less concentrated network was helpful to these men. In looking at the network change variables, we found virtually no differences between the two sexes in the percentage of explained variance in need fulfillment accounted for by the change variables (Table 3). The network change variables accounted for 20% of the explained variance for men and 19% of the explained variance for women. The percentage of explained variance for the women attained statistical significance whereas the percentage of explained variance for men nearly attained statistical signi~c~ce. There were no differences between the sexes in the propo~ionate amounts of variance contributed by the interactional and structural data sets. With regard to single network variables, for both men and women increases in the frequency of interaction with contacts over the one-year study period were significantly associated with ability to fulfill needs, For women, an increase in those linkages rated “very important” was also associated with enhanced need fulfillment. Ana/ysjs 4 In order to examine the relationship between physical health and social networks, we dichotomized the male and female samples at Time 1 into persons in relatively good health and poor health according to the cut-off scores described in the variable section. An examination of the data indicated that those men and women in relatively poor health had larger networks than their more healthy counterparts (Table 4). However, TABLE 4 Comparison Between Men and women of Network Changes Under Varying States of Physical Health Females Males size (Time l)’ Number of linkages lost on 1-year follow-upb Proportion having lost one or more linkages Number of new linkages on l-year follow-up’ Proportion having at least one new linkage to replace lost linkages Network Poorer Health (n=18) Better Health (n=40) 8.1 1.9 .78 1.4 5.6 1.0 .5@ 0.8 .43 .48 Poorer Health (n = 25) Better Health (n = 50) 9.6 2.2 .88 0.7 7.8 2.0 .80 1.5 .41 .55 Note: zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA “F= 1.59,df=3,130,p=n.s.(1-way ANOVA) bF=2.91,df=3,130,p<.05(l-wayANOVA) ‘F = 2.03, df = 3,130, p = ns. (l-way ANOVA) ‘Male Better Health vs. Ail Other Groups Combined, X2 = 7.46, df= l,p < .Ol 54 JOURNAL OF AGING STUDIES Vol. ~/NO. l/l 988 the differences between health categories were not statistically significant. Although a high percentage of individuals in all health categories lost at least one linkage, approximately one-half of the persons in each category were able to compensate with at least one new linkage. Moreover, regardless of health status, all groups showed an ability to forge new linkages over the one-year study period, with the less healthy men and the more healthy women being relatively more successful at making new linkages. Also noteworthy was that the group with the fewest overall linkages at Time 1 (the healthier males) was considerably more adept at maintaining their network size than the other groups, all of whom showed a diminution in network size over the follow-up period. DISCUSSION The findings contradict many of the earlier assumptions regarding the impaired sociability of women living in SRO hotels. There was no evidence of pervasive isolation and the longitudinal data presented here suggested that these women had dynamic network systems that allowed for the replacement of lost contacts with new ones. Moreover, social networks were important in explaining how these women fulfilled their daily needs. It should be noted that the SRO women were relatively isolated versus a sample of middle-class elderly women; the SRO women having approximately half the number of contacts as the community women (C. Cohen, unpublished data). Perhaps this may account for why researchers using normative notions of sociability may have been predisposed to characterize SRO dwellers as isolates. The results have important implications with respect to public policy that has been replacing SRO hotels with office buildings or luxury residential towers. For instance, in New york City more than 100,000 low rent SRO apartments have disappeared since 1971. One-third of homeless individuals in city shelters have listed an SRO hotel as their last address (Hayes 1986). However, our data make it clear that SRO hotels do not house human flotsam who can be shifted from hotel room to public shelter without consequences. The hotels provide a viable social support system that cannot be easily replaced. A second striking finding was the similarities in the ways that the men and women used their networks. For these men and women we found nearly identical proportionate distribution of the explained variance in need fulfillment attributed to each of the four synchronic network data sets and to the two network change data sets. Contrary to our earlier cross-sectional study (Cohen and Sokolovsky 1979), we found no differences between the men and women in poor health and their ability to maintain their social network system. Thus, our data contrasted with earlier impressionistic reports that depicted elderly SRO women as less socially adept than the SRO men (Lally et al. 1979; Stephens 1976). The similarities in social adaptation styles between the sexes may be reflective of the parallel life courses of these men and women. Unlike many of the women described in Stephen’s SRO hotel in Detroit most of whom were more likely to have been married previously and who were “more vulnerable, less successful at coping, and unreconciled to their present status” (Stephens 1976, p. 280), a majority of the SRO women in midtown Manhattan had never married, had been self-supportive, and had been living in SRO hotels for over 16 years. Only one-fifth of SRO women had been housewives, Gender, Networks, and Adaptation 55 Among an zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON Inner- City Etder/y Population the rest had worked, most commonly as clerical, sales, or unskilled manual workers. Many of the women had come to the hotels in search of a place with security (e.g., 24-hour staff coverage), service (e.g., maids, desk clerks), and convenience to work and shopping. For the unattached working women in the 1950s and early 1960s the SRO was a reasonable choice of residence. The life histories of the aging women living in midtown Manhattan SRO hotels mirrored those of the older women living in Seattle’s SRO hotels. Lally and her coworkers (1979) reported that these women more closely identified with their fathers, worked in trades dominated by men, and were fiercely independent. Paralleling the women’s histories, only one-half of the aged men in our sample had married and most had worked in clerical, sales, and skilled or semiskilled occupations. However, nearly one-fourth of the men had erratic work histories, possibly due to alcohol abuse, mental dysfunctions, or wanderlust. Nonetheless, most had settled down in recent years having spent an average of 10 years at their present address. Many of the men had come to SRO hotels for the same reasons as the women, and they were currently living on annual incomes that were comparable to those of the women. Although it may be argued that the findings for an SRO sample cannot be easily generalized to other populations, the data, nevertheless, strongly support the notion that gender congruence in work roles and lifestyles result in gender congruence in the utilization of social relationships. Consequently, the similarities found between the aging SRO men and women in various network dimensions and in the role that networks play in adaptation and survival indeed provide tentative support to the contention that as the lifestyles of women in the general ~pulation come to approximate those of men, the social interactions of aging men and women may more closely resemble each other. This is a revised version of a paper presented at the 39th Annual Meeting of the Gerontological Society of America, Chicago, October 1986. The authors thank Henry Rajkowski and Carole Lefkowitz for their assistance. This research was supported by the Center for Studies of Mental Health of the Aging No. I-ROl-MH3 1745 and 1-KO7-MHOO523 (Dr. Cohen). ACKNOWLEDGMENTS REFERENCES Barnes, John A. 1954. “Class and Committees in Norwegian Island Parish.” Human Relations 7:39-58. Boissevain, Jeremy. 1974. Friem!s of Frkmis. New York: St. Martin’s Press. Cantor, Marjorie. 1979. “Neighbors and Friends: An Overlooked Resource in the Informal Support System.” Research on Aging 1:434- 463. Cohen, Carl I., and Jay Sokolovsky. 1979. “Health-seeking Behavior and Social Networks of the SRO Aged.” Journal of the American Geriatric Society 27:270- 278. Cool, Linda, and Justin McCabe. 1983. “The ‘Scheming Hag’ and the Dear Old Thing: The Anthropology of Aging Women.” Pp. 56-7 1 in Growing Qld in Difirenr Societies, edited by J. Sokolovsky. Belmont: Wadsworth. Dohrenwend, Bruce S., Larry Krasnoff, Alexander R. Askenasy, and Bruce P. Dohremvend. 1978. “Exemplification of a Method for Scaling Life Events: The PER1 Life Events Scale.” Journal of Health and Social Behavior 19:205- 229. 56 JOURNAL OF AGING STUDIES Vol. ~/NO. 1 il988 Gurland, Barry J., John Copeland, Judith Kuriansky, Michael Kelleher, Lawrence Sharpe, and Laura L. Dean. 1983. 77~ Mood and Mind of Aging. New York: Haworth Press. C&and, Barry J., Judith Kuriansky, Lawrence Sharpe, Robert Simon, Pamela Steller, and Peter Berkett. 1977. “The Comprehensive Assessment and Referral Evaluation (CARE)Rationale, Development, and Reliability.” International Journal of Aging and Human Development 8~9-42. Haley, Barbara A., Margaret Pearson, and Doris A. Hull. 1981. “Urban Elderly Residents of a Single Room Occupancy Housing (SRO’s), 1976 and 1980.” Paper presented at the 34th Annual Meeting of the Gerontological Society of America, Toronto. Harrell, Steven. 1981. “Growing Old in Rural Taiwan.” Pp. 193-210 in Other Ways of Grow~g Old, edited by P. Amoss and S. Harrell. Stanford: Stanford University Press. Hayes, Robert M. 1986. “The Issue is Housing.” The New York Times (November 27), p. A2.S. Kerns, Virginia. 1983. Women and the Ancestors. Chicago: University of Illinois Press. Lally, Maureen, Eileen Black, Martha Thornock, and J. David Hawkins. 1979. “Older Women in Single-room Occupancy (SRO) Hotels: A Seattle profile.” Gerontologist 19:67-73. Lawton, M. Powell. 1970. “Assessment, Integration, and Environments for Older People.” 75, Geronto~g~t 10:38-46. Leavy, Richard L. 1983. “Social Support and Psychological Disorder: A Review.” Joum~~ qf Communes Psych&y 1113-2 1. Lopata, Helena. 1975. “Support Systems of Elderly: Chicago of the 1970’s.” The Gerontologkt 15:35-41. Mitchell, John S. 1969. “Concept and Use of Social Networks.” Pp. l-50 in Social Networks in Urban Situations, edited by J. C. Mitchell. Manchester: Manchester University Press. Peace, Sheila. 198 1. An International Perspective on the Status of Older Women. Washington, DC: International Federation on Aging. Powers, Edward A., and Gordon L. Buitena. 1976. “Sex Differences in Intimate F~endships of Old Age.” Journal of cadge and the ~arni~ 38: 739-747. Sokolovsky, Jay, and Carl I. Cohen. i 98 1. “Toward a Resolution of Methodological Dilemmas in Network Mapping.” Schizophrenia Bulletin 7: 109- 116. Stephens, Brenda J. 1976. Loners, Losers, Lovers: Elderly Tenants in a Slum Hotel. Seattle: University of Washington Press. Van Dalen, Diebold B. 1973. Understunding Educational Research. 3d ed. New York: McGraw-Hill.