Contribution of social support and recreation companionship to the life satisfaction of people with persistent mental illness
McCormick, Bryan P
Social relationships have been seen as a contributor to happiness and well being since the time of Durkheim’s classic study of suicide. For people with persistent mental illness, social relationships appear to be problematic. This study examined the contribution of social support and recreation companionship to life satisfaction among people with persistent mental illness. Clients from a community mental health program were surveyed on social support networks, recreation companionship networks, and life satisfaction. Findings indicated that minority group status, educational attainment, social support satisfaction, and recreation satisfaction were significantly related to life satisfaction. Neither the size of participants’ social support nor recreation companionship networks were significantly related to life satisfaction. Multivariate analysis revealed that recreation satisfaction significantly predicted life satisfaction.
KEY WORDS: Social Support, Recreation, Life Satisfaction, Mental Illness, Social Networks
Services for people with persistent mental illness (PMI) have undergone a dramatic shift in the past 30 years. Jacobson and Burchard (1992) characterized this shift as one from total institutional care, with little hope of discharge, to services that commonly emphasize goals of community integration and productivity. One outcome of community integration is the development of a supportive social network. Among the general population, social companionship appears to impact well being (Rook, 1987; Vaux, 1988); whereas the absence of such ties has been found to be related to premature death (Bloom, 1990). However, there has been some indication that community integration goals for people with PMI have been unsuccessful. For example, Horwitz and Reinhard (1995) stated that the social networks of people with (PMI) are smaller than the general population and “marked by social isolation” (p. 207).
Along with the shift in service delivery settings has come a shift in the underlying philosophy of services. Increasingly, the realization has been made among those serving clients with PMI that “comforting” may be a more realistic goal than “curing” (Oliver, Huxley, Bridges, & Mohammad, 1997). To this end, mental health services have begun to consider indicators of improved life quality as treatment outcomes. Oliver et at. stated that regardless of what professionals serving people with PMI desire to achieve, “we are certain that the public, including the patients and their families, expect us to be centrally concerned with fostering satisfaction and enhancing role performance” (p. 19). As a result, clients@ satisfaction with their lives has become an increasingly important indicator of successful care for people with PMI.
One area typically addressed through therapeutic recreation intervention is the area of community (re)integration (cf. Austin, 1997). Recreation and leisure interventions appear to offer possibilities for the development of social relationships, particularly given the social nature of leisure. For example, Scott and Godbey (1992) stated that “leisure is largely organized around people interacting with others as a result of mutual tastes and out of a sense of belonging” (p. 48). Thus, leisure may provide a context for social interaction, leading to social companionship and social integration. If therapeutic recreation professionals are to develop efficacious interventions in working with people with PMI, a better understanding of their community life and its contribution to their well being is required.
Review of Literature Well Being and Life Satisfaction
Diener (1994) stated that the study of happiness has occupied western thinkers since Aristotle. In recent decades, behavioral scientists have continued to be interested in the concept of happiness; however, it has been studied under various rubrics such as subjective well being, morale, positive affect, life satisfaction, and quality of life. The most global concept that incorporates the various terms noted in the more general study of happiness is that of subjective well being (Diener). He characterized subjective well being as a global, positive reaction to one’s life. Furthermore, it “includes all of the lower-order components such as life satisfaction and hedonic level. Life satisfaction refers to a conscious global judgement of one’s life. Hedonic level or balance refers to the pleasantness minus unpleasantness of one’s emotional life” (Diener, p. 108). In addition, the construct is considered to be subjective in nature, residing in experience of the individual. This characterization is somewhat in contrast to other global measures of well being such as “quality of life” (QOL). Lehman (1996) noted that “at a minimum, QOL covers persons’ sense of well being; often it also includes how they are doing (functional status) and what they have (access to resources and opportunities)” (p. 78). Thus, QOL covers at least the subjective component of well being; however, it may also be considered to be inclusive of objective factors such as functioning and resource access. Typically, the study of QOL has assessed well being through measurements of life satisfaction.
Life satisfaction represents a cognitive assessment of one’s subjective life circumstances. This cognitive component of well being is thought to be a relatively stable construct, unlike hedonic level which is considered to be more variable (Diener, 1994). Although there is some expected correlation between hedonic level and life satisfaction over time, they are also expected to diverge at times. This divergence is due to the fact that life satisfaction is a global measure and hedonic level is a result of ongoing reactions to events. Affective assessments may be based on reactions to immediate factors of short duration; whereas life satisfaction is based on a comparison of one’s standards to one’s expectations (Pavot & Diener, 1993). Thus, one may be “down in the dumps” and still be satisfied with life. As a result, the construct of life satisfaction provides a relatively stable, global indicator of well being.
Social Ties and Well Being
Among the general population, social ties appear to positively impact well being (House, Landis, & Umberson, 1988; Iso-Ahola & Park, 1996: Rook, 1987; Vaux, 1988). For example, House et al. noted that prospective studies of social relationships and mortality have found that persons low in social network integration were as much as two times more likely to die than those high on this variable. Thus, among the general population, social integration and social relationships appear to play a critical role in well being; yet, what is “happening” in these relationships that prevents negative outcomes continues to be a source of examination.
One outcome of social ties that has received research attention has been that of social support. Vaux (1988) characterized socially supportive behaviors as “specific acts generally recognized (by most members of a culture) as being intentional efforts to help a person, either spontaneously or upon request” (p. 29). As indicated by this characterization of social support, much of the research in this area has focused on the utilitarian function of social relationships (Rook, 1987). That is, social support implies intentional efforts to help another.
Although the above studies were based on general population samples, social relationships appear to have similar implications for people with PMI. Horowitz and Reinhard (1995) noted that “there is general recognition that the quality of life of persons with chronic [sic] mental illness in large part depends on their social relationships in the community” (p. 206). For example, Lehman (1983) found that social relations, measured as total social contacts, number of contacts in the home, and intimacy of contacts, were all positively related to well being among people with PMI.
The concept of social support has gained considerable attention in the past two decades, yet its development has been approached from a number of different perspectives (Sarason, Sarason, & Pierce, 1990; Starker, 1986; Winemiller, Mitchell, Sutliff, & Cline, 1993). One approach has been to examine the nature of social integration or connection as an indication of available social support. In this approach, measurement has focused on both the identification of a network of supportive others, as well as considering individuals’ perceptions of social support.
Social integration vs. perceived social support. In general, social support research has begun with an identification of a network of others to whom individuals can turn for support. In this sense, social support networks have been considered to be indicative of social integration. Those who have fewer people available to them are considered to be less socially integrated than those who have a greater number of people available to them. Measures of this sort have shown to be significant in prediction of mortality in prospective studies (cf. Bloom, 1990; House et al., 1988). However, there have been criticisms of using support networks as indicative of social support. Sarason, Sarason, et al. (1990) noted that asking people to identify those who they see frequently, or even those people who are important in their lives, may identify members of a network who are sources of conflict and negative feelings. For example, although someone may identify family members as someone they could turn to for certain types of support, the relationship may be highly conflicted.
A second limitation of the support networks approach to examining social support has been raised by Vaux ( 1988). He noted that a linear relationship is assumed in most examinations of support networks and well being. That is, the additive effect of an additional network member is assumed to be constant regardless of whether the network size increases from zero to one or from 20 to 21. Vaux has suggested that a “threshold” effect may be more conceptually useful; indicating that above some number of network members, additional members do not significantly affect well being. Finally, although measures of social integration have been tied to mortality in the general population, they are not as effective in predicting well being among smaller populations (Sarason, Sarason, et al., 1990).
Another conceptual approach to social support has been through an examination of perceptions or appraisals of social support. “Support appraisals are subjective, evaluative assessments of a person’s supportive relationships and the supportive behavior that occurs within them” (Vaux, 1988, p. 29). In this sense, the actual people or acts of support they provide are of less importance than how an individual feels about the support. This perceptual approach to social support is typically measured through people’s satisfaction with available support. Perceptions of support have been found to be most consistently associated with health outcomes (Doeglas, et al., 1996; Sarason, Pierce, & Sarason, 1990; Sarason, Sarason, et al., 1990). Sarason, Sarason, et al. stated that “support emanates from not so much what is done but from what that indicates to the recipient about the relationship” (p. 17).
The experience of mental illness appears to be marked by social isolation. Studies have consistently found that the social networks of people with mental illness are smaller than those of the general population (Biegel, Tracy, & Song, 1995), thus indicating poor social integration. Furthermore, at least one study has found evidence of a hierarchy of social isolation among people with PML Sokolovsky, Cohen, Berger, and Geiger (1978) found that people with thought disorders demonstrated smaller social networks than those with mood disorders. In addition, people with mental illness have been found to perceive their social networks as less supportive (Biegel et al.; Lyons, Perrotta, & Hancher-Kvam, 1988) and tend to be less satisfied with their social relationships than the general population (Rosenfield, 1992). Overall, people with mental illness are poorly socially integrated and are dissatisfied with the few social relationships which they have.
In spite of the conceptual problems in identifying social support networks and their relative weakness in predicting well being, their enumeration still has value. The reason for the inclusion of objective social networks is that they may be considered to be indicators of an opportunity structure through which people can meet their needs (Bigelow, Brodsky, Stewart, & Olson, 1982; Bigelow, McFarland, & Olson, 1991; Oliver et al., 1997). Although perceptions of social support may be a better indicator of well being, they are to some extent tied to an opportunity structure of available relationships.
Buffering vs. main effects hypotheses. In addition to examining different conceptualizations of social support, researchers have also examined its operation in well being. In this body of literature, two predominant themes have emerged: the effects of social support in time of stress or crisis, and the effects of social support on overall well being. In the social support literature, two themes have been characterized as that of a “main effect” and a “buffering effect” (Bloom, 1990; Penninx et al., 1997; Vaux, 1988). In the main effect approach, the argument is made that social support has a direct positive relationship to health and well being, regardless of other life circumstances. In contrast, the buffering hypothesis has noted that social support itself may not have a direct effect on health but may protect people in times of environmental stress, thus limiting (buffering) the effect of the stressors.
According to the buffering hypothesis, social support interacts with stressors such that social support makes a greater contribution to health and well being among those who are stressed than among those who are not experiencing stressors. Pennix et al. (1997) noted that “there is sufficient empirical evidence in favor of both approaches” (p. 393). However, some authors have come to question the utility of social support in explaining everyday well being due to its apparent bias towards measuring instrumental acts. Rook (1987) noted that:
Although social relationships are often desired for the aid and security they afford, they are also sought in and of themselves because they provide opportunities for pleasurable companionship and intimacy. From this perspective, social integration does not serve an extrinsic purpose but instead affords many intrinsic satisfactions, such as shared leisure and recreation or discussion of common interests. (p. 1133)
Thus, although social support may be accessed in times of stress, companionship may have intrinsic values and contribute to everyday well being.
Although much of the research on the benefits of social relationships on well being has focused on relatively instrumental aspects of social support, Vaux (1988) indicated that there might be more to social relationships than intentional efforts to help. He stated that:
At the extreme, one might argue that a large component of social support goes on in close relationships. What constitutes support in these relationships may be embedded in the private language that characterizes communication in close relationships-special gestures, private jokes, obscure allusions, private meanings, and shared associations. (pp. 16-17)
One way this non-instrumental function of social relationships has been examined is through the concept of companionship (Rook, 1987). Rook noted that “social support and companionship make equally important but complementary contributions to psychological well being. Support protects people from the debilitating effects of life stress, whereas companionship protects people from the emptiness and despair associated with loneliness” (p. 1133). It is interesting to note that many of the early attempts to create typologies of social support included categories related to social companionship (cf. Barrera & Ainlay, 1983; Vaux, Stewart, & Riedel, 1987; Wills, 1985).
In the few studies that have examined companionship, recreation and leisure play important roles. For example, Rook (1987) defined companionship as relationships formed around -shared leisure and other activities that are undertaken primarily for the intrinsic goal of enjoyment” (p. 1133). In an examination of the relative contributions of social support and companionship to psychological well being, Rook found that the two concepts make different contributions to well being. Social support appeared to buffer the negative effects of life stressors on psychological well being most notably among people experiencing greater life stress.
Rook’s (1987) study confirmed the buffering hypothesis of social support; however, companionship was found to make a significant positive contribution to psychological well being regardless of the level of life stress experienced. Rook interpreted these findings as indicating that social support and companionship may make different contributions to overall well being. Social support may be more influential in assisting people in coping with stress, whereas companionship may make greater contributions in combating emptiness and loneliness. Given the strong social component of recreation, it appears possible that it may play a role in the development of companionship and social support. Recent studies have examined the nature of companionship and social support as they relate to recreation and leisure.
Recreation, Social Support, and Companionship
One of the first developments in this area was by Coleman and Iso-Ahola (1993). They proposed a model for describing the relationship of leisure to health via self-determination and leisure-based social support. According to the authors, “leisure benefits health by buffering people against personal stress produced by life circumstances. Leisure-based social support and leisure generated self-determination are identified as two important mediators of the influence of leisure on the stress-health relationship” (Coleman & Iso-Ahola, p. 112). However, in a test of the above model, Coleman (1993) found leisure-based social support did not demonstrate a significant main or buffering effect on a “seriousness of illness” rating (health measure). In another examination of this model, Iso-Ahola and Park (1996) examined the effects of leisure friendship and leisure companionship on the stress-illness relationship among Taekwondo practitioners. They operationalized leisure companionship as an index of participation in social activities with co-participants. Leisure friendship was operationalized using a modified version of the Social Support Appraisals scale (Vaux et al., 1986) that appeared to indicate the degree to which respondents felt that co-participants in Taekwondo cared for and respected the respondent. In contrast to Coleman’s findings, Iso-Ahola and Park found that “leisure-generated friendship and companionship interact with life stress in a manner consistent with their being buffers against the adverse effects of life stress on physical and mental health” (pp. 182-183). Thus these findings supported a buffering effect of leisure-based relationships on health, but only among those reporting higher degrees of life stress.
Purpose of The Study
Overall, the study of social connections and well being has been an interest of social and behavioral researchers since the time of Durkheim. In the past three decades, the construct of social support has received a great deal of research attention (Rhodes & Lakey, 1999). However, some researchers have argued that the construct of companionship can be distinguished from that of social support (Rook, 1987). More recently recreation and leisure researchers have begun to look its role in developing social networks and the impact of these networks on health and well being (e.g., Coleman, 1993; Iso-Ahola & Park, 1996). Within therapeutic recreation practice, interventions targeting social networks typically fall under the broad area of “community reintegration.” If therapeutic recreation is to play a part in community integration of people with persistent mental illness, a greater understanding of the role of recreation companionship and social support in well being is needed.
This study examined the relative contributions of social support and recreation companionship to life satisfaction among people with PML Unlike previous studies of social support and leisure, this study did not seek to identify buffering relationships. Instead, a main effects (Vaux, 1988) approach was taken, in which effects were examined under the assumption that social support and companionship contribute to life satisfaction regardless of life stress. In addition, at least one previous study of leisure and social support found that negative life events, which has been the typical means for operationalizing life stress, are atypical occurrences in most people’s lives (McCormick, 1995). Specifically, this study sought to identify if (a) recreation companionship and satisfaction are related to life satisfaction and (b) whether this relationship exists if the effects of social support are controlled.
A convenience sample (N = 77) was drawn from three treatment programs offered by a community mental health center in a mediumsized midwestern city. Participants were living in independent or supervised living situations at the time of the survey. In general, participants were predominantly Caucasian (55%), male (61 %), and in their early 40s (M = 42.0; SD = 10.5). Most participants had not completed high school (54%), were never married (71 %), were unemployed at the time of the survey (86%), and lived in supervised living arrangements (66%). A summary of characteristics of the study sample can be found in
Instrumentation. Data were collected through the use of three principal scales and demographic questions. The first scale was the Satisfaction with Life Scale (SWLS; Diener, Emmons, Larsen, & Griffin, 1985) which is composed of five items on which participants identify the degree to which they agree or disagree with each statement on a 7-point Likert scale. Previous studies reported the intemal consistency of the SWLS as ranging from 0.79 to 0.89 (Pavot & Diener, 1993).
The second scale used was the short version of the Social Support Questionnaire– Short Form, Revised (SSQSR; Sarason, Sarason, Shearin, & Pierce, 1987). The SSQSR is composed of six items with each item in two parts. The first part asks respondents to list social support network members who perform different support functions. Participants list up to nine different individuals. In addition, respondents note the nature of the relationship by identifying the social identity (friend, sister, mother, cousin) of each network member. The second part of each item asks respondents how satisfied (6-point Likert-type scale; I = very dissatisfied, 6 = very satisfied) they are with the specified support function. A total of six different support functions are assessed. Examples of support functions assessed are: (a) “Whom can you really count on to be dependable when you need help?” and (b) “Whom can you really count on to help you feel better when you are felling generally down-inthe-dumps?” Two scores are calculated for the SSQSR. First, an average network size (SSQN) is calculated by summing the total number of people identified across all six support functions and dividing by six (number of support functions assessed). The second score is that of average satisfaction with social support (SSQS), which represents the average satisfaction score across the six support functions. Sarason et al. stated that the items in the SSQSR are “very general in nature and reflect the affective aspects of relationships” (p. 507). Alpha coefficients of internal consistency for both SSQS and SSQN have been reported to be between 0.90 to 0.93 (Sarason et al.). Based on procedures specified by Sarason et al., the SSQN and SSQR for this study were scored by averaging scores across the six support functions.
The third principal scale addressed recreation networks and recreation satisfaction. This instrument was designed for the purposes of this study and was developed so that the format was similar to that of the SSQSR. A series of six questions asked participants to first identify the companions with whom they engaged in the following categories of recreation activities: activities in the home, games, sports, outdoor or nature activities, crafts activities, and entertainment activities. Categories were modified from those used by Dattilo, Hoge, and Malley (1996) in the TRAIL leisure assessment battery. These categories were chosen because they represented a broad spectrum of recreation activities, and included both home-based and out-of-home pursuits. The same format as that of the SSQSR was used. Participants identified up to nine different individuals in each activity category and their relationship to this recreation network member (RECNET). Participants also could identify if they did not engage in the particular activity category. In addition, participants were asked to rate their satisfaction with each category of activity (RECSAT). As with the SSQSR, average ratings were calculated. The recreation instrument differed from the SSQSR in that not all categories of activities were necessarily valid responses. For example, participants could identify that they did not engage in certain activity categories. As a result, average network and satisfaction scores were calculated across only those activity categories in which respondents indicated participation. Table 2 shows descriptive statistics for variables used in this study.
Procedures. The scales described above were compiled into a survey that was presented to participants via group administrations in group homes (8 administrations) and community treatment centers (2 administrations). A total of 10 group administrations were conducted. The survey was administered as follows. All clients present were verbally given information on the nature and purpose of the study and invited to participate. Clients were informed that those who elected to participate would be offered a choice of either food or merchandise certificates worth $5. At this point clients who were not interested in participating were dismissed and those who expressed interest were then presented (both written and oral) with a statement of informed consent. Following acquisition of informed consent, clients were presented with the survey instrument. Given a relatively low rate of literacy skills within this population, members of the research team worked with clients individually or in pairs to complete the survey. The principal investigator and research assistants aided participants by either reading questions to the client, or assisting them in reading any difficult words. Research team members were instructed not to interpret questions or possible responses for clients.
At each group administration, counts were taken of all clients present, and then compared to completed surveys. Based on these figures, the response rate was 53% (77/145). This rate should be considered only an estimate, as it is possible that at group administrations all potential participants may not have been present. Thus, the reported response rate only represents the percentage of clients who were presented with information on the survey and chose to complete it.
Protocol for missing data. In examining raw data, it was noted that there were a number of missing cases on important variables. In order to maintain as many cases as possible, missing data were replaced. Tabachnik and Fidell (1989) noted that mean replacement represents the most conservative form of replacement of missing data. Due to the exploratory nature of this research and the sample used, a conservative approach was chosen. One of the assumptions of replacing missing data is that missing data are randomly distributed (Tabachnik & Fidell). This assumption is particularly an issue when the number of missing data points is large compared to the size of the sample. In order to determine if missing data represented systematic differences, tests of mean differences of the dependent variable (life satisfaction) based on grouping SSQN, SSQS, RECNET, and RECSAT variables into (a) valid responses and (b) missing responses were performed. No significant differences were found between respondents who offered responses and those who offered no response (missing data). As a result, the assumption of random distribution of missing data appeared to be met. As a result, 13 values were replaced with the mean for RECNET, 10 values were replaced for RECSAT, 4 values were replaced for SSQN, 8 values were replaced for SSQS, and 2 values were replaced on SWLS.
Data transformations. In addition to the replacement of missing values, variables were examined for normality of distribution. Both network variables (RECNET and SSQN) were found to be positively skewed, while both satisfaction variables (RECSAT and SSQS) were found to be negatively skewed. Life satisfaction and demographic variables demonstrated adequate normality. In order to improve the meeting of assumptions of the analytic processes, non-normally distributed variables were transformed. Positively skewed variables were transformed using square roots, while negatively skewed variables were transformed using the square root of the reflect (constant – value; Tabachnik & Fidell, 1989). It should be noted that the square root of the reflect procedure reversed the direction of correlations. This reversal was due to the fact that larger values on the satisfaction variables produced smaller reflect values; whereas smaller values on the satisfaction variables created larger reflect values. As a result of these procedures, lower scores on the transformed RECSAT and SSQS actually indicated higher satisfaction. Transformed means and standard deviations are presented in Table 2. Subsequent analyses were performed using transformed variables.
Regression analyses. Data were analyzed using multiple regression analysis. Since the research questions related to establishing relationships among social support and recreation network variables, and life satisfaction, correlational methods were considered appropriate. In addition, previous research has identified significant multivariate correlation between recreation network variables and health variables in a sample from the general population (McCormick, 1995). Data were analyzed first through examining Pearson product-moment correlation coefficients (Table 3). Subsequently, two backwards stepwise multiple regression (SPSS, 1988) procedures were performed. A backwards stepwise regression model was used in order to identify the most parsimonious model for the data. First, demographic variables of age, highest educational level attained (number of years of formal education; scale ranged from 4 to 19+), gender (1 = male; 0 = female), and racial minority status (1 = minority; 0 = majority) were regressed as a block on the dependent variable. Subsequent steps in the first procedure removed variables with the least significant prediction of the dependent variable. This procedure continued until only those variables that significantly predicted the dependent variable remained. In the second backwards stepwise regression model, SSQN, SSQS, RECNET, RECSAT, and the demographic variables of highest educational level and racial minority status were entered into the regression equation. Again, subsequent steps removed nonsignificant predictors until only those variables that were significant predictors of life satisfaction remained.
A common concern in multiple regression analysis has to do with the number of independent variables-to-cases ratio. Although a typical rule of thumb is 20-to-1, Tabachnik and Fidell (1989) have stated that a 5-to-1 ratio is the bare minimum. The present study had approximately a 10-to-I ratio, indicating at least adequacy for multiple regression analysis.
Based on the descriptive statistics in Table 2, a number of overall statements can be made about this sample. First, average network sizes were small. On average, respondents identified only one recreation companion across activity categories. It should be noted that, in fact, 20% of this sample identified no recreation companions whatsoever. In addition, social support networks also were relatively small with respondents identifying, on average, two persons to whom they could turn for support across the functions identified. In contrast to the recreation companion network, only 7% of the sample identified no one in their social support network. However, given the small network sizes, respondents were generally satisfied with their social support and recreation companions. Scores between 4.5 to 5.5 would indicate a response of “fairly satisfied” on the satisfaction scale. Thus, although they did not have many others in their networks, they still tended to be satisfied with these aspects of their lives. Finally, respondents in this sample were in the slightly dissatisfied to undecided range of life satisfaction.
In addition to the descriptive findings, there were a number of significant bivariate relationships found within this sample (Table 3). First, age demonstrated a significant correlation with life satisfaction; older respondents tended to be more satisfied with their lives than did younger respondents. In addition, members of racial minorities also tended to report significantly higher life satisfaction than Caucasians. Among the general population, racial group membership has been found to be related to life satisfaction.- however, the findings have been that Caucasians report higher life satisfaction than did African-Americans (London, Crandall, & Seals, 1977; Redmond, 1988; Stock, Okun, Haring, & Witter, 1985).
The independent variables in this study also demonstrated significant bivariate correlation with life satisfaction. Principally, social support satisfaction and recreation satisfaction were significantly correlated with life satisfaction. In general, those respondents who were more satisfied with their social support and recreation were more satisfied with life. In contrast, neither average social support nor recreation companionship network size were significantly correlated with life satisfaction. In addition, although the size of one’s social support network was significantly correlated with social support satisfaction, recreation companionship network was not significantly correlated with recreation satisfaction. Finally, there appeared to be some consistency in respondents’ network sizes and satisfaction. Social support network size was significantly correlated with recreation companionship network size and social support satisfaction was significantly correlated with recreation satisfaction.
Following examination of bivariate relationships, backwards stepwise multiple regression analyses were performed to identify significant multivariate predictors of life satisfaction. First, demographic variables were examined (Table 4). Results of the analysis indicated that only minority group status and highest level of education completed were significant multivariate predictors of life satisfaction. Multivariate analysis found that being a member of a racial minority predicted higher life satisfaction, whereas higher educational attainment predicted lower life satisfaction. Subsequently, significant multivariate demographic predictors of life satisfaction were entered into the subsequent regression equations.
As can be noted in step 1 of Table 5, all study variables and the two significant demographic variables were entered into the initial equation. The initial regression equation (step 1) accounted for 26% of the variance in life satisfaction; however, only the variable of recreation satisfaction demonstrated a significant Beta coefficient. In examining the first step, two aspects appear worthy of notation. First, the demographic variables, which were significant predictors of life satisfaction when only demographic variables were considered (see Table 4), did not appear to predict life satisfaction significantly over-and-above the effects of the other variables. The second notable effect was that the significant bivariate relationship (see Table 3) of social support satisfaction with life satisfaction was not reproduced in the multivariate case. In addition, the social support and recreation companionship network variables were not significant in explaining life satisfaction. Through subsequent steps in the analysis, variables were removed one at a time until only significant variables remained. As can be seen by step 6 in Table 5, satisfaction with one’s recreation was the only significant predictor of satisfaction with life. Neither social support variables, nor demographic variables, demonstrated significant contributions to predicting life satisfaction.
In general, this study found that people with PMI demonstrated scores that indicated slight dissatisfaction with life. This finding is consistent with other studies that have shown that people with disabilities and disabling conditions report lower life satisfaction compared to general population samples (Pavot & Diener, 1993). In addition, it was found that demographic variables were associated with life satisfaction. Older participants and those who identified themselves as members of racial minorities were higher in perceived life satisfaction. Gender was not associated with life satisfaction. This finding is contrary to that of Lehman, Rachuba, and Postrado (1995) who found that among people with PMI, males reported a significantly higher level of general life satisfaction than did females. However, in this study none of the demographic variables were significantly associated with life satisfaction in multivariate analyses. Thus, although age and minority group status were associated with life satisfaction, their effects were negligible in the presence of recreation satisfaction.
The findings also indicated that among people with PMI. satisfaction with recreation is significantly related to overall satisfaction with life. This finding may not be surprising given the high rate of unemployment in this sample. That is, the quality of one’s recreation may make a major impact on life quality when one’s life is characterized by large amounts of unoccupied time. The contribution of recreation satisfaction to life satisfaction in the present study is consistent with previous research that has examined quality of life among people with PMI (e.g., Lehman, 1983; Rosenfield, 1992). In general, this study supports the important role that recreation and leisure experience play in the lives of people with PMI.
Although recreation satisfaction was a significant predictor of life satisfaction, the number of recreation companions was not. People who had few recreation companions were just as satisfied (or dissatisfied) with their lives as those who had many recreation companions. In addition, the average size of one’s social support network was also unrelated to life satisfaction. Although it may not be surprising that the number of recreation companions or socially supportive individuals in one’s life are not significant predictors of life satisfaction, previous studies of social integration have used these variables as predictors of well being (cf. Lehman, 1996, 1983). One explanation for the limited utility of network variables in this study is that they demonstrated somewhat of a “floor effect.” It should be noted that on average, respondents recreated with one other person across activity domains and across social support dimensions they could turn to only two persons. Thus, the network variables demonstrated little variation. In addition, counts Of people may not adequately reference what is “happening” in the identified relationships. What can be noted about companion and social support networks from this sample is that this population reports very small networks. From this finding it could be argued that they have a very limited social opportunity structure. Furthermore, this may indicate a relatively “fragile” social network in the sense that people with PMI have few alternatives if a network member is unavailable.
In general, the findings do not support the role of recreation companions, measured as number of available companions, in enhancing the life quality of this population. In addition, the number of socially supportive others was not indicative of well being. Yet, the number of socially supportive others in respondents’ networks was found to correlate with social support satisfaction. In addition, the number of recreation companions was positively correlated with the number of socially supportive network members. Although these relationships existed only in the bivariate case, it may be that this pattern indicates an underlying process. This hypothesized process might be such that social support network members are derived from recreation companions, and that the number of supportive others is related to greater social support satisfaction. Such a process would be consistent with the opportunity structure approach to structural properties of social networks (Bigelow et al., 1982; Bigelow et al., 1991; Oliver et al., 1997).
Implications and Summary
Skalko, Van Andel, and Desalvatore (1991) identified the need for the development of a sound research base in order to develop sound practices in therapeutic recreation, Given Horwitz and Reinhard’s (1995) statement regarding the absence of knowledge of the social networks of people with PMI, the need to develop a knowledge base still appears relevant. This study adds to the knowledge of the social support and recreation network aspects of the lives of people with persistent mental illness. In addition, the findings from this study add to the understanding of life quality in this population. Diener (1994) noted that life satisfaction is one component of subjective well being, representing the subjective aspect of quality of life (QOL). Many human service organizations and professions have been structured around the improvement of QOL for service recipients. Indeed. QOL and improved well being have both been noted as bases for the profession of therapeutic recreation (Wenzel, n.d.). However, questions remain regarding how, or if, services provided by such professions contribute to life quality. To the extent that we have a greater understanding of the nature of well being and quality of life, our profession is better equipped to address such issues beyond that of mere platitudes. A continued examination of the life satisfaction, well being, and life quality of people with persistent mental illness is necessary if we, as a profession, are to make valuable contributions to improving the lives of clients with PMI.
Although network variables were not significant predictors of life satisfaction in this study, there still may be utility in examining social network structures in this population. The findings on social networks reinforced previous findings that this population appears relatively socially isolated (Delespaul & deVries, 1987; Horwitz & Reinhard, 1995; Sokolovsky et al., 1978). More importantly, social network data may have utility in assessment. For example, the objective nature of networks enables practitioners to establish baseline information about social connections. In addition, the objective nature of networks may facilitate comparisons over time, which can be used as a basis for evaluating social integration interventions. However, until network variables can be established as predictors of life quality, they should be used cautiously in treatment planning.
Finally, this study has implications for guiding further study. First, this study was based on a convenience sample of one program in one city. A larger sample, with a more representative sampling procedure, would permit more in-depth analyses of relationships. For example, in this study the effects of variables were examined simultaneously; however, it may be more accurate that the relationships are analyzed in a sequential manner. The model proposed by Coleman and Iso-Ahola (1993) implies a path model in which leisure facilitates the development of leisure-based social support, which in-turn mediates the effects of life stress on health and well being. As a result, path or causal models may be more appropriate forms for examining social networks.
In summary, this study sought to identify relationships among recreation companionship, recreation satisfaction, and life satisfaction among people with PMI. The findings did not support the number of available recreation companions as a significant predictor of life satisfaction. However, it appears that recreation satisfaction does make an important contribution to overall life satisfaction in this population. In addition, the relationship of recreation satisfaction to life satisfaction exists when the effects of social support are controlled. Further study should continue to examine the character and processes of recreation and support networks among this population, and seek to examine if these may be indirectly related to life quality.
Austin, D. R. (1997). Therapeutic recreation: Processes and techniques. Champaign, IL: Sagamore.
Barrera, M., Jr., & Ainlay, S. L. (1983). The structure of social support: A conceptual and empirical analysis. Journal of Community Psychology, 11, 133-143.
Biegel, D. E., Tracy, E. M., & Song, L. (1995). Barriers to social network interventions with persons with severe and persistent mental illness: A survey of case managers. Community Mental Health Journal, 31, 335-349.
Bigelow, D. A., Brodsky, G., Stewart, L., & Olson, M. (1982). The concept and measurement of quality of life as a dependent variable in evaluation of mental health services. In G. J. Stahler and W. R. Tash (Eds.), Innovative approaches to mental health evaluation (pp. 345-366). New York: Academic Press.
Bigelow, D. A., McFarland, B. H., & Olson, M. M. (1991). Quality of life of community mental health program clients: Validating a measure. Community Mental Health Journal, 27, 43-55.
Bloom, J. R. (1990). The relationship of social support and health. Social Science & Medicine, 30, 635-637.
Coleman, D. (1993). Leisure based social support, leisure dispositions and health. Journal of Leisure Research, 25, 350-361.
Coleman, D., & Iso-Ahola, S. E. (1993). Leisure and health: The role of social support and selfdetermination. Journal of Leisure Research, 25, 111-128.
Dattilo, J., Hoge, G., & Malley, S. M. (1996). Interviewing people with mental retardation: Valid
ity and reliability strategies. Therapeutic Recreation Journal, 30, 163-178.
Delespaul, P. A@ E, G., & deVries, M. W. (1987). The daily life of ambulatory chronic mental patients. Journal of Nervous and Mental Disease, 175, 537544.
Diener, E. (1994). Assessing subjective well being: Progress and opportunities. Social Indicators Research, 31, 103-157.
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The Satisfaction With Life Scale. Journal of Personality Assessment, 49, 71-75.
Doeglas, D., Suurmeijer, T., Brian@on, S., Mourn, T., Krol, B., Bjelle, A., Sanderman, R., & Van Den Heu vel, W. (1996). An international study on measuring social support: Interactions and satisfaction. Social Science and Medicine, 43, 13891397.
House, J. S., Landis, K. R., & Umberson, D. (1988, July). Social relationships and health. Science, 241, 540-545.
Horwitz, A. V., & Reinhard, S. C. (1995). Family and social network supports for the chronically mentally ill: Patient perceptions. Research in Communit, and Mental Health, 8. 205-232.
Iso-Ahola, S. E., & Park, C. J. (1996). Leisure– related social support and self-determination as buffers of stress-illness relationship. Journal of Leisure Research, 28, 169-187.
Jacobson, J. W., & Burchard, S. N. (1992). Overview of clinical services, social adjustment, and work life in community living. In J. W. Jacobson, S. N. Burchard, & P. J. Carling (Eds.), Community living jbr people with developmental and psychiatric disabilities (pp. 5-19). Baltimore: Johns Hopkins Press.
Lehman, A. F. (1993). The well being of chronic mental patients: Assessing their quality of life. Archives of General Psychiatry, 40, 369 -373.
Lehman, A. F. (1996). Measures of quality of life among persons with severe and persistent mental disorders. Social Psychiatry and Psychiatric Epidemiology, 31, 78-88.
Lehman, A. F., Rachuba, L. T., & Postrado, L. T. (1995). Demographic influences on quality of life among persons with chronic mental illnesses. Evaluation & Program Planning, 18, 155-164.
London, M., Crandall, R., & Seals, G. W. (1977). The contribution ofjob and leisure satisfaction to quality of life. Journal of Applied Psychology, 62, 328-334.
Lyons, J. S., Perrotta, P., & Hancher-Kvam, S. (1988). Perceived social support from family and friends: Measurement across disparate samples. Journal of Personality Assessment, 52, 42- 47.
McCormick, B. (1995). Recreation companionship networks, recreation satisfaction, and social support in predicting psychological health. Abstracts from the 1995 Symposium on Leisure Research. Alexandria VA: National Recreation and Park Association.
Oliver, J., Huxley, P., Bridges, K., & Mohammad, H. (1997). Quality of life and mental health services. London: Routledge.
Pavot, W., & Diener, E. (1993). Review of the life satisfaction scale, Psychological Assessment, 5, 164-172.
Pennix, B. W. J. H., Van Tilburg. T., Deeg, D. J. H., Kriegsman, D. M. W., Boeke, A. J. P., & Van Eijk, J. T. M. (1997). Direct and buffer effects of social support and personal coping resources in individuals with arthritis. Social Science and Medicine, 44, 393-402.
Redmond, S. (1988). An analysis of the general well being of blacks and whites: Results of a national study. Joumal of Sociology and Social Welfare, 15, 57-71.
Rhodes, G. L., & Lakey, B. (1999). Social support and psychological disorder: Insights from social psychology. In R. M. Kowalski & M. R. Leary (Eds.). The social psychology of emotional and behavioral problems (pp. 281-309). Washington, DC: American Psychological Association.
Rook, K. S. (1987). Social support versus companionship: Effects on life stress, loneliness, and evaluations by others. Journal of Personality and Social PsYchology, 6, 1132-1147.
Rosenfield, S. (1992), Factors contributing to the subjective quality of life of the chronic mentally ill. Journal of Health and Social Behavior, 33, 299315.
Sarason, B. R., Pierce, G. R., & Sarason, I. G. (1990). Social support: The sense of acceptance and the role of relationships. In 1. G. Sarason, B. R. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 97-128). New York: John Wiley.
Sarason, B. R., Sarason, 1. G., & Pierce, G. R. (1990). Traditional views of social support and their impact on assessment. In B. R. Sarason, 1. G., Sarason, & G. R. Pierce (Eds.). Social support.- An interactional view (pp. 9-25). New York: John Wiley & Sons.
Sarason, 1. G., Sarason, B. R., Shearin, E. N., & Pierce, G. R. (1987). A brief measure of social support: Practical and theoretical implications. Journal of Social and Personal Relationships, 4, 497510.
Scott, D., & Godbey, G. C. (1992). An analysis of adult play groups: Social versus serious participation in contract bridge. Leisure Sciences, 14, 47-67.
Skalko, T., Van Andel, G., & Desalvatore, G. (1991). The benefits of therapeutic recreation in psychiatry. In C. P. Coyle, W. B. Kinney, B. Riley, & J. W. Shank (Eds.), Benefits of therapeutic recreation: A consensus view (pp. 289-352). Ravensdale, WA: Idyll Arbor.
Sokolovsky, J., Cohen, C., Berger, D., & Geiger, J. (1978). Personal networks of ex-mental patients in a Manhattan SRO hotel. Human Organization, 37, 5-15.
SPSS (1988). SPSS-x user’s guide. Chicago: SPSS Inc.
Starker, 1. (1986). Methodological and conceptual issues in research on social support. Hospital & Community Psychiatry, 40, 11-36.
Stock, W., Okun, M., Haring, M., & Witter, (1985). Race and subjective well being in adulthood: A black-white research synthesis. Human Development, 51, 192-197.
Tabachnik, B. G., & Fidell, L. S. (1989). Using
multivariate statistics (2nd ed.). New York: Harper & Row.
Vaux, A. (1988). Social support: Theory, research and intervention. New York: Prager.
Vaux, A., Phillips, P., Holly, L., Thompson, B., Williams, D., & Stewart, D. (1986). The Social Support Appraisals (SS-A) Scale: Studies of reliability and validity. American Journal of Community Psychology, 14, 195-219,
Vaux, A., Stewart, D., & Riedel, S. (1987). Modes of social support: The Social Support Behaviors (SSB) Scale. American Journal of Community Psychology, 15, 209-237.
Wenzel, K. (n.d.). Therapeutic recreation specialists: The quality of life professionals. (Available from the National Therapeutic Recreation Society, 22377 Belmont Ridge Rd, Ashburn, VA 201484501).
Wills, T. A. (1985). Supportive functions of interpersonal relationships. In S. Cohen, & S. L. Syme (Eds.), Social support and health (pp. 61- 82). Orlando: Academic Press.
Winemiller, D. R., Mitchell, E., Sutliff, J., & Cline, D. J. (1993). Measurement strategies in social support: A descriptive review of the literature. Journal of Clinical Psychology, 49, 638 – 648.
Bryan McCormick is on faculty in the Department of Recreation and Park Administration, Indiana University. This study was funded in part by grants from the National Therapeutic Recreation Society and the Department of Recreation and Park Administration, Indiana University.
Copyright National Recreation and Park Association Fourth Quarter 1999
Provided by ProQuest Information and Learning Company. All rights Reserved