| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Nan L. Stevens, Center for Psychogerontology, Radboud University, P.O. Box 9104, 6500 HE Nijmegen, the Netherlands. E-mail: stevens{at}psych.ru.nl
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Friendship • Intervention • Loneliness • Older women
The Friendship Enrichment Program is an educational program that is based on theories of successful aging that assume that an aging individual regulates the quality of life by setting goals, striving to achieve them, and accumulating resources that are useful in adapting to change with age (Schulz & Heckhausen, 1996; Steverink, Lindenberg, & Ormel, 1998). Friends are helpful in adapting to change with age in various ways: as sources of information and as role models undergoing similar changes; for confirmation of identity and self-esteem; as companions for pleasurable activities; and as providers of emotional support (Hartup & Stevens, 1997). There is evidence from cross-sectional and longitudinal studies that friendship networks decline over the life course (Kalmijn, 2003; Van Tilburg, 1998), which is one reason why enhanced activity to develop and maintain friendship may be desirable. Older women, in contrast to older men, continue to be interested in making new friends in advanced old age (Field, 1999). The main question in this article is whether older adult women are able to develop new friendships, improve existing friendships, and reduce loneliness when they receive encouragement, training, and guidance in this domain.
Theories on Social Adaptation
Various theories are relevant for our understanding of how aging individuals deal with changes in personal relationships and meet social needs under changing life circumstances. Baumeister and Leary (1995, p. 497) have proposed a fundamental need to belong, "a pervasive drive to form and maintain a minimum quantity of lasting positive and significant interpersonal relationships." To satisfy the need to belong it is not necessary to maintain particular types of relationships. Involvement in regular, pleasant interactions with a certain number of people with whom one shares affective concern and that are likely to continue in the foreseeable future is sufficient. Theoretically, social contacts are substitutable for one another; however, Baumeister and Leary acknowledge that the accumulation of shared experience and intimacy takes time. Therefore, a new relationship cannot immediately serve as a substitute for a long-lasting relationship that is no longer available; it must develop into a stable affectionate relationship before it becomes effective in fulfilling the need to belong.
Satisfaction of the need to belong is associated with subjective well-being, happiness, and positive affect in general. When individuals experience deprivation in belongingness, they are more prone to experiencing negative affect, depression, loneliness, and anxiety. Furthermore, they are motivated to seek interpersonal contact and cultivate possible relationships until they have reached a minimum level of relatedness (Baumeister & Leary, 1995). This would explain why some older adults are motivated to participate in interventions focused on developing new relationships. However, there is a satiation point beyond which individuals are less likely to be interested in expanding their social networks.
The social compensation model (Ferraro & Farmer, 1995) focuses explicitly on adaptation in social relationships as individuals age. According to this model, compensation is the primary strategy for dealing with loss and change in social relationships with age. An underlying assumption is that people strive to maintain continuity in their social involvements by replacing relationships or social activities that have been lost. Various studies on widowhood indicate that widows tend to intensify their friendship ties with other widows to compensate for the loss of their husbands (Ferraro & Farmer; Stevens, 1995). Therefore, those women who have lost a partner as a result of widowhood or divorce may be strongly motivated to develop new friendships in later life.
Another theory on social adaptation in later life is the socioemotional selectivity theory (Carstensen, Isaacowitz, & Charles, 1999). According to this theory, the goals that individuals strive to achieve in relationships are similar throughout the life course, although in later life there is a change in the salience of goals that is due to the decrease in future time perspective. The awareness that time is running out results in a stronger preference for spending time with those with whom interactions are predictable and most satisfying emotionally. In other words, people become more selective with age, favoring interactions with their closest friends and family members. As they age, people may therefore be less willing to invest time and energy in new relationships because of the uncertain rewards involved.
The Friendship Enrichment Program
The Friendship Enrichment Program is designed to empower older women in the process of meeting personal goals in friendship (Stevens & Albrecht, 1995). It is based on feminist therapy and reevaluation counseling (Stevens, 2001). The program assumes that women possess an expertise in relationships based on a lifetime of experience. It is designed to help participants (a) clarify their needs, desires, and expectations in friendship; (b) analyze their current relationship networks in order to identify actual and potential friends; (c) formulate goals that involve the improvement of existing friendships or the development of new friendships; and (d) develop strategies to achieve these goals. Throughout the program, women are encouraged to affirm their own self-worth as the basis for planning and realizing their goals in friendship. The program consists of 12 weekly lessons that focus on self-esteem, relational competence, and various phases in friendship formation, utilization, and maintenance, as well as the practice of relevant social skills. For the analysis of social networks, participants fill in a personal convoy of relationships (Kahn & Antonucci, 1980). Reflection on their convoys is the basis for setting goals in friendship, whereas group discussion on strategies to achieve them provides participants with extensive material for social learning. There is a follow-up meeting after 6 months to evaluate each woman's success in achieving her goals in friendship.
Hypotheses
The review of theories on social needs and adaptation to age-related changes led us to formulate the following hypotheses.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Study 2 was conducted as part of a national program of research on successful aging. This study involved a pretest–post-test control group design with a follow-up measurement. Senior service agencies in four communities that offered the Friendship Enrichment Program were involved in the study. Among the participants in the program (N = 84), 69 respondents were recruited for the study; the response rate was 82%. The control group of 60 women was recruited from waiting lists for the program, through announcements on a website for older adults, and articles on the friendship program in women's magazines. The women in the control group were all interested in enhancing friendship.
In Study 2, both the intervention and control groups had a baseline measurement prior to the program (Time 0, or T0), a measurement immediately following the course or 3 months after baseline (T1), and a follow-up measurement 6 months after the program ended or 9 to 10 months after baseline (T2). A questionnaire was mailed to respondents 1 year after the course (Time 3, or T3). Excluding women who dropped out as a result of natural causes (5%), the nonresponse rate for both groups at T1 was 6%; at T2 the nonresponse rate was 2%. There were 60 women in the intervention group and 55 women in the control group who completed three measurements. The mailed questionnaire was returned by 54 women in the intervention group and 54 women in the control group.
For this article, we drew an additional comparison group from the Dutch Aging Survey (Steverink, Westerhof, Bode, & Dittmann-Kohli, 2001), which involved a nationwide representative sample of 983 individuals aged between 40 and 85 years. The comparison group included 226 women aged 55 to 75; the age range was restricted to attain an average age similar to that of program participants. Table 1 provides an overview of the characteristics of the four samples that we compare in this article.
|
2(3) = 1.869, p =.600. The percentage of women living alone is high in both groups, at 69% and 68%. There is a difference in education (years completed). The women in Study 1 are slightly better educated than those in Study 2; t(110) = 2.006, p =.047. Unfortunately, different questions were used to measure subjective health in the two studies, so we cannot compare the two samples directly on subjective health. However, when the intervention groups are compared with a representative sample of women from the Dutch Aging Survey on two measures of subjective health, the participants in the interventions consider themselves to be healthier. Women in Study 1 report fewer restrictions caused by health problems, 2(2) = 8.481, p =.010, and those in Study 2 describe their health as excellent or good more often, 2(4) = 10.730, p =.029, than women from the survey. Loneliness at T1 was also quite similar for the two intervention groups, 6.9 (SD = 3.9) for Study 1 and 7.2 (SD = 3.4) for Study 2, with t(110) = –0.562, p =.575. Because of the high degree of similarity in age, marital status and living situation, loneliness, and rather positive subjective health, we find that combining the two samples in the analyses for this article is justifiable. Here we have employed the data on friendship at T1 and T2 from both studies and the scores on loneliness 1 year after the program had ended.
The control group in Study 2 demonstrated no significant differences in age, marital status, living situation, and subjective health when compared with the intervention group in the same study (Martina & Stevens, in press); however, the control group was less lonely than the intervention group, with t(113) = 3.463, p =.001. We have used it as a control group for the combined intervention group, controlling for differences in loneliness when doing comparative analyses.
Procedure
In Study 1, researchers collected data by means of a questionnaire passed out during the last meeting of the program and by conducting interviews at the end of the program (T1) and 10 to 12 months later (T2). The questionnaire included questions on age, marital status, educational level, and subjective health, as well as a loneliness scale (De Jong Gierveld & Kamphuis, 1985). This scale consists of five positive items (e.g., I can rely on my friends whenever I need them) and six negative items (e.g., I experience a sense of emptiness around me). Scores range from 0 (not lonely) to 11 (extremely lonely). It is a valid and reliable instrument that has been used extensively in the Netherlands (De Jong Gierveld & Van Tilburg, 1999). In Study 1, Cronbach's alpha was 
= 0.86 for this scale.
During the first interview, an important source of information was the personal convoy that women had filled in during the fifth meeting of the program. The personal convoy is visually represented by three circles, in which individuals are assigned positions according to their importance and closeness to the respondent, located in the center. Interviewers asked respondents to identify the role of each person. For those identified as friends, respondents reported age, marital status, and shared activities. The location of friends in the convoy was noted. In the second interview, interviewers asked respondents the following questions: "Have you undertaken any activity to renew old friendships or to develop new friendships since the program ended?" and "Has there been any improvement in friendship?" Furthermore, interviewers asked them to reexamine the convoy and describe any changes that had taken place in their convoys since the first interview. They were asked, "Have new people been added to your convoy? Do you consider them to be friends? Have some people changed positions or left the convoy? Does this represent improvement or deterioration in friendship?" We scored changes in friendship by using content analysis; relevant here are scores for new and improved friendships. We scored new friendships when a new person in the convoy was identified as a friend or when women reported having met or made new friends. We scored improvements when friends were moved closer to the respondent in the convoy at T2, or when descriptions of friendship change included key words such as closer, deeper, more personal, more open, or better contact. The interrater reliability for coding new and improved friendship was good (Cohen's kappa was 
= 0.80 on both measures). Respondents also completed the loneliness scale a second time at T2.
In Study 2, interviewers collected data by means of semistructured interviews and written questionnaires. The interview included questions on background information, friendship, and evaluation of the program. The personal convoy model was filled in during the first interview. For those identified as friends, respondents then described the nature of the friendship in terms of shared activities, confiding, and frequency of contact. During the third interview (T2), interviewers asked participants to draw a new convoy. Then they compared the new convoy with the first version and identified any changes involved represented new friendships, improvements, or deteriorations in friendship. The interviewers scored these directly.
The written questionnaires that respondents in Study 2 completed at each measurement point included standardized instruments for loneliness (De Jong Gierveld & Kamphuis, 1985; 11 items; Cronbach's = 0.93) and the Positive and Negative Affect Scale (Watson, Clark, & Tellegen, 1988) for subjective well-being. This scale includes 10 positive and 10 negative items (Cronbach's = 0.80 and = 0.84 for positive and negative items, respectively).
Although there were some differences in the ways in which information on friendship was collected in Study 1 and Study 2, the questions on the convoy and changes in friendships were similar. In both studies the same scale was used to measure loneliness; when we combined the samples, the loneliness scale maintained its high reliability (Cronbach's = 0.91). The Dutch Aging Survey covered a wide range of topics. It was useful for comparison because it included the loneliness scale (Cronbach's = 0.89 for this subsample) and the Positive and Negative Affect Scale (Cronbach's = 0.81 for positive items and = 0.87 for negative items) that were used in Study 2.
Analyses
To compare participants in the Friendship Enrichment Program with women in the Dutch Aging Survey on the indicators of deprivation, we used chi-square tests and the t test for independent samples (Hypothesis 1). We applied binary logistic regression to examine the effects of early levels of loneliness (T1) and condition (intervention group vs control group) on specific friendship changes within a year (Hypothesis 2). Whether or not women reported the development of new friendships, improved friendships, or a combination of these changes were the dependent variables in these analyses.
We used standardized regression on loneliness 1 year after the course began (T2 for the combined sample) to test Hypothesis 3, with a dummy variable for change in friendship. No change in friendship was the reference category. We entered loneliness at T1 and condition (Study 1 vs Study 2) into the equation as controls. To test hypothesis 4, we used binary logistic regression; for this hypothesis, the reported use of the combination of friendship strategies was the dependent variable, whereas the presence of a partner, presence of a friend in the inner circle of the convoy, and loneliness at T1 were the covariates. To test the fifth and sixth hypotheses, we added partner loss and age to the logistic regression equation, but we removed presence of a partner.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
2(3) = 117.73, p <.001.
Other indicators of deprivation are measures of subjective well-being. Among the program participants, 12% scored as not lonely (0–2), 45% were rather lonely (3–8), and 43% were very lonely (9–11); among DAS women, 61% were not lonely, 33% were rather lonely, and 7% were very lonely. The program participants were significantly lonelier than DAS women at 2(2) = 240.896, p <.001.
For friendship program participants in Study 2, the average score for positive and negative affect was 30.83 (SD = 4.19) and 29.46 (SD = 5.37), respectively, on the Positive and Negative Affect Scale at T1. The DAS women had average scores of 35.91 (SD = 5.16) and 23.23 (SD = 6.15) for positive and negative affect, respectively, on this scale. The differences between participants and DAS women were significant, t(268) = –7.00, p <.001 and t(278) = 7.11, p <.001; friendship program participants reported less frequent positive emotions and more frequent negative emotions.
These results support our first hypothesis that predicted that there would be more evidence of deprivation in belongingness among participants in the Friendship Enrichment Program than among women from the DAS.
The second hypothesis concerned the greater likelihood that participants in the program would report positive developments in friendship than would members of the control group. Many participants reported that they had developed new friendships (63%) or had improved existing friendships (61%), although less than half (43%) reported both changes in friendship.
In the control group, only 33% reported that they had developed new friendships whereas 46% reported that they had improved friendships during a similar period. Only 16% reported both developments in friendship.
We categorized changes in friendship as experiencing no change, developing new friendships only, improving friendships only, and experiencing a combination of these two changes. The difference between the participants in the friendship program and the control group on friendship development was significant, 2(3) = 15.447, p =.001; participants in the program reported more positive developments in friendship.
To control for the effect of higher levels of loneliness among the participants in the Friendship Enrichment Program, we did three logistic regression analyses in which reporting either one or both friendship developments were the dependent variables, and loneliness at T1 and condition (intervention vs control groups) were covariates. The results are presented in Table 2. The intervention group was about twice as likely to make new friends (1.8), improve friendships (2.2), and report the use of both friendship strategies (1.9) compared with the control group. Loneliness at T1 did not have a significant impact on friendship development. The prediction that women who had participated in the Friendship Enrichment Program would report more success in developing new friendships and improving existing friendships was thus supported.
|
|
In the fourth hypothesis, we predicted that satiation of the need to belong would decrease the likelihood of reporting positive change in friendship during the year after the program. Indicators of satiation included the presence of a partner, the presence of a close friend in the inner circle of the convoy, and low levels of loneliness at the first measurement point (T1). The results of the binary logistic regression on the combined use of the friendship strategies are reported in Table 4 (Model 1). None of the variables had a significant impact on the likelihood of using a combination of friendship strategies. Thus, there was no evidence of satiation among those who did not report positive changes in friendship within a year after following the program.
|
|
Additional logistic regression analyses on use of the friendship strategies separately revealed that age had no effect on the likelihood of making new friends (Model 1, Table 5); however, it had a negative effect on the likelihood of improving friendships (Model 2, Table 5). The older participants were, the less likely they were to report improving friendships. This is in contrast to our prediction.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
It was rather surprising that a combination of developing new friendships and improving existing friendships was necessary for participants to reduce loneliness significantly. When the Friendship Enrichment Program was designed as a preventive intervention, the authors assumed that either strategy might be useful for reducing loneliness (Stevens & Albrecht, 1995), depending on the situation of the participant, the level of loneliness experienced, and the social resources that were available. However, the average loneliness of participants was quite high, at 7.1 (SD = 3.3) on an 11-point scale (De Jong Gierveld & Van Tilburg, 1999). Apparently both the social skills and opportunities for developing new friendships, as well as more advanced skills for improving friendship (Rook, 1991), are necessary to effect significant change in higher levels of loneliness.
Theory on the need to belong was less useful in explaining failure to report positive developments in friendship. There was no evidence for the predicted effect of satiation in belongingness; in fact, there was some evidence that having at least one close friend was an advantage to further developments in friendship.
During the program and the interviews, we discovered that the action-oriented approach to friendship does not appeal to everyone. Some women were quite hesitant to formulate any goals at all. The romantic notion that friendship happens when one meets the right person is difficult to dispel, perhaps because it does not expose one to the risk of social failure. Other women did formulate goals but failed to achieve them. For women with a history of negative experiences in friendship, the action-oriented approach to friendship development is probably not sufficient. Some women need individual counseling to deal with issues of low self-esteem, social anxiety, distrust, or fear of rejection before they are able to benefit from a group program. An example of a more individual approach is the cognitive behavioral approach to the treatment of friendship disorders that has been described by Young (1986). In his therapeutic approach, Young describes how to diagnose specific problems in friendship and develop strategies to change cognitive schemas related to the self and social relationships. One way to improve the Friendship Enrichment Program might be to spend more time on diagnosis of specific problems in friendship. Another option is to offer the Friendship Enrichment Program in a series of interventions that are planned for individuals in specific situations. For example, a trajectory might involve individual counseling, and then a support group followed by the Friendship Enrichment Program.
The hypothesis concerning compensation was not supported. There are many biographical circumstances beside partner loss that motivate older women to try to develop new friendships or to improve existing ones. The general decline in friendship networks in later life (Kalmijn, 2003; Van Tilburg, 1998) may be the main motivating factor for activity to enrich friendship, to compensate for this decline. A multifaceted, flexible intervention program is necessary to meet diverse needs of participants (Cattan et al., 2005).
Age did have an effect on success in friendship development, although not in the way we had predicted on the basis of socioemotional selectivity theory. Age had no effect on the likelihood of developing new friendships; it only affected the likelihood of improving friendships. This finding shows the potential for interventions that focus on the development of new relationships such as friendship in later life. It is important to note that although the age range of participants was 52 to 80 years, the largest group (59%) was in their sixties. Thus, the oldest old, for whom socioemotional selectivity may be operating more strongly, are underrepresented in this intervention study.
Limitations
One of the limitations of these studies is that participants in the program are self-selected; they have volunteered to participate in the Friendship Enrichment Program and the two studies. When the studies began, the intervention was already implemented in social service programs for older adults; we did not have a large sample of lonely people available to whom different experimental conditions could be randomly assigned. Furthermore, the Friendship Enrichment Program is based on the notion of empowerment, that is, supporting people to meet goals that they have set for themselves. Random assignment to the intervention or control condition would be contradictory to the goal of empowerment.
We did not have a baseline measurement for loneliness and availability of friendships in Study 1. However in the second study we found no significant differences in loneliness between T0 and T1; there were also very few developments in friendship between T0 and T1. Therefore, using T1 as a first measurement point for this article is defensible. Our assumption that it takes time to develop new friendships and reduce loneliness is supported by the data from Study 2 (Martina & Stevens, in press).
Conclusion
Despite the limitations of the studies involved, this article adds to available knowledge on the possibilities for interventions to enhance friendship and improve well-being in later life. The action-oriented approach can also be applied to other goals that are attractive to older adults, such as optimal health, autonomy, harmonious partnership, and family relations.
|
Footnotes |
|---|
1 Center for Psychogerontology, Radboud University, Nijmegen, the Netherlands.
Decision Editor: Linda S. Noelker, PhD
Received for publication December 15, 2005. Accepted for publication April 14, 2006.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. Street, S. Burge, J. Quadagno, and A. Barrett The Salience of Social Relationships for Resident Well-Being in Assisted Living J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2007; 62(2): S129 - S134. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Bode, D. T. D. de Ridder, R. G. Kuijer, and J. M. Bensing Effects of an Intervention Promoting Proactive Coping Competencies in Middle and Late Adulthood Gerontologist, February 1, 2007; 47(1): 42 - 51. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
