| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
Arlene R. Gordon Research Institute, Lighthouse International, New York, New York.
Address correspondence to Kathrin Boerner, Lighthouse International, 111 East 59th Street, New York, NY 10022. E-mail: kboerner{at}lighthouse.org
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods.This multilevel analysis examined change in support provision over time, as well as the degree to which sociodemographics, health, functional disability, use of rehabilitation services, and the receipt of support predicted support provision. Participants, 449 older adults with a progressive, chronic visual impairment, were interviewed three times over an 18-month period.
Results.Affective and instrumental support provided to family and friends each showed a decrease over time. Age, gender, and education emerged as predictors of support provided at baseline. Receipt of support was positively related to support provision at all three time points, both within and across support types.
Discussion.Findings indicated that there may be feasible ways of providing support, even by older adults who are in need of support themselves.
Although social support and its impact are among the most widely studied topics in gerontology, the vast majority of this research focuses on received support. Little attention has been paid to the older person as support provider (Litwin, 1998
). Because older support recipients often experience chronic disability, it is uncommon to consider their potential to provide support. Research has shown, however, that older adults do reciprocate the support they receive (Peters & Kaiser, 1985; Wentowski, 1981), and, furthermore, that helping others can be beneficial to emotional states (Silverstein & Bengtson, 1994). Moreover, reciprocity in giving and receiving help has been found to be associated with life satisfaction, happiness, and self-esteem (Antonucci, Fuhrer, & Jackson, 1990; Liang, Krause, & Bennett, 2001). Overall, research has shown that support functions play an important role in the process of adaptation to chronic impairment (e.g., Penninx et al., 1999; Reinhardt, 1996). Thus, study of support provision in addition to support receipt in disabled older adults is needed.
Addressing the issue of support among older adults who are dealing with chronic vision loss, a common age-related impairment affecting approximately 20% of adults aged 65 and older (The Lighthouse, Inc., 1995), is particularly relevant because of its strong association with functional disability (Horowitz & Reinhardt, 2000). The increase in functional disability that typically accompanies a chronic impairment such as age-related vision loss makes conducting certain instrumental tasks (e.g., handling finances, or performing household tasks such as cooking) difficult or even impossible. As a result, the person not only needs to rely on support from others in order to accomplish specific tasks but also may have less of a capacity to provide help to others with these types of tasks. In support of this idea, a recent study on network and health changes demonstrated that decline in health and functional ability predicted a decrease in instrumental support provided (van Tilburg & Broese van Groenou, 2002). Thus, increased support needs may affect the ways in which a chronically impaired person can provide support. However, although some types of support provision may no longer be feasible as a result of a disability (help with household tasks), there may be increases in other types of support provision that can still be performed (e.g., affective support). Because of such dynamics, the challenge of dealing with chronic vision loss seems to provide a good sample case in which to study changes in patterns of support provision over time, as well as the determinants of support provision among older adults. Longitudinal research on support provision in those with age-related eye disease is especially important, because this disease is characterized by a gradual onset and progressive deterioration (Faye, 1984). Thus, the ability to provide various types of support may vary over time with increasing levels of impairment. This would also add to the general support literature, considering requests by other researchers (Antonucci & Akiyama, 1987b; Liang, et al., 2001) who have noted that most available support research is limited to cross-sectional design and that future work based on longitudinal data is needed. Thus, this study sought to examine support provided over time by older adults with age-related vision loss, a common, chronic disabling condition and to identify the factors that influence levels of support provision in this population.
Linking the Give and Take of Support
The changes in types and amount of support provision that are likely to occur when one becomes chronically impaired are related to the increase in levels of functional disability and the resulting need for more support from others. Therefore, the study of "giving while in need" should address the links between the "give and take" of support. Some prior studies have shown a strong, positive correlation between giving and receiving support (Liang et al., 2001; Litwin, 1998; van Tilburg & Broese van Groenou, 2002). This finding seems to be in line with the basic tenets of equity theory (Walster, Walster, & Berscheid, 1978), that is, people strive for a balance between give and take in relationships and feel uncomfortable with an imbalance.
The empirical evidence regarding equity theory has been inconsistent, with some authors finding little support in its favor (e.g., Davey & Eggebeen, 1998). Yet, others have pointed out that an interpretation of these mixed findings is complicated by methodological aspects, such as different ways of conceptualizing (Wentowksi, 1981) or measuring support exchange (Litwin, 1998). Overall, there is still substantial evidence in favor of equity theory, suggesting that balanced exchanges can enhance life satisfaction (Antonucci, Fuhrer, & Jackson, 1990) and that lack of reciprocation can have negative effects on one's morale (Stoller, 1985).
It is important to realize that even under conditions of balanced exchange, the actual type of support does not necessarily have to match. For example, in a cross-sectional examination of baseline data from the current study, Reinhardt (2001) found that the participants reported receiving more instrumental support than they provided. In terms of affective support, however, the opposite was perceived; that is, participants reported that they provided more affective support than they received. These findings suggest that older adults who deal with a chronic impairment may strive for balance in their support exchanges, in the sense that they may try to compensate for an increased need in instrumental support by providing more affective support.
Differences in the reciprocation of support type have also been examined over time. Van Tilburg (1998) found that the youngest of the older adults in his sample reported an increase in instrumental support provided and received over time. Yet, the older adults in his sample reported an increase in instrumental support received and emotional support provided. He concluded that, if younger adults need more instrumental support, they may compensate by giving more instrumental support, whereas older adults may try to maintain reciprocity by giving more emotional support. However, in an earlier longitudinal study, Klein Ikkink and van Tilburg (1998) found no evidence to support the idea that a lack of instrumental reciprocity could be "balanced out" by giving more emotional support. Thus, more research is needed to clarify the conditions under which the phenomenon of balancing out may be more or less likely to occur. For example, in each of the papers just cited, support exchange was examined at the network level, not by relationship type. The phenomenon of balancing out, however, may be more typical for certain relationship types (e.g., friendship) than for others (e.g., family relationships).
There is cross-sectional research suggesting that tolerance of an imbalance in give and take differs depending on relationship type. For example, Ingersoll-Dayton and Antonucci (1988) found that, during old age, maintaining balance in support exchanges was more important in friendships than family relationships and with spouses than children. Silverstein, Conroy, Wang, Giarrusso, and Bengtson (2002) suggested that the long time lag between investment and return may differentiate family from friendship support exchanges. The latter would require more immediate reciprocity. Thus, an understanding of the link between the receipt and provision of support across various relationship types seems to be key in understanding patterns and ways of support provision.
Predisposing Factors of Support Provision
With regard to the individual characteristics that may influence support provision, the literature points to links between support exchanges and age, gender, education, health, and functional disability. A consistent effect of age indicates that younger adults tend to report higher levels of support provision than older adults (e.g., Ingersoll-Dayton & Antonucci, 1988; Litwin, 1998; van Tilburg, 1998; van Tilburg & Broese van Groenou, 2002). In contrast, findings on the relationship between gender and social support have been inconclusive. Although the finding that women provide more support than men has been considered well documented (Antonucci & Akiyama, 1987b), Antonucci and Akiyama (1987a) found inconclusive patterns in their own research. Furthermore, van Tilburg and Broese van Groenou (2002) reported from their study on network and health changes that women provided lower levels of instrumental support than did men. Because of such inconsistencies, researchers have called for future work on gender differences in the context of social exchange (e.g., Liang et al., 2001).
Individual characteristics, such as education, health, and functional disability, have also been included in the analysis of social exchange. Evidence from different studies suggests that a higher educational level is related to higher levels of support provision (Liang, et al., 2001; Litwin, 1998, van Tilburg & Broese van Groenou, 2002), and poorer health and greater functional impairment are linked to lower levels of support provision (Klein Ikkink & van Tilburg, 1998; Litwin, 1998; van Tilburg & Broese van Groenou, 2002). Although these variables have failed to emerge as significant predictors in a few studies (e.g., Litwin, 1998; Penninx et al., 1999), there is some consensus that measures of health or functional impairment are important to include when the meaning and effects of support exchanges are examined (Liang et al., 2001; Penninx et al., 1999; Silverstein et al., 2002).
The ability to maintain certain levels or types of support provision even when one is disabled may also be influenced by using rehabilitation. Older adults who become visually impaired may experience initial social withdrawal but also dependence on those closest to them (Greig, West, & Overbury, 1986). Vision rehabilitation service goals can include helping people to regain some level of independence, learn how to accept help when needed, and maintain social interactions. The latter could facilitate support provision. This study extends prior work by examining the provision of support by disabled elders who are in need of support themselves for multiple support and relationship types over time. Furthermore, it adds to the literature by addressing the role of several predisposing factors, including the use of rehabilitation services as a factor of unique relevance for a disabled population, and by taking into consideration the links between support provided and received.
Hypotheses for Current Analyses
On the basis of the notion of balancing out a lack of instrumental reciprocity by providing more emotional support, we expect that the provision of instrumental support will decrease while the provision of affective support will either remain stable or increase over time. However, we also hypothesize significant individual variation in support provision at baseline as well as in change over time. We expect that this variation will be accounted for in the following ways: Participants who are younger, have higher levels of education, and report better health and lower functional disability will start out with significantly higher levels of support provision. It is noted that because of prior inconclusive evidence on the role of gender in support exchanges, gender effects on baseline levels of support will be explored. Further, we hypothesize that the use of rehabilitation services will be positively associated with the provision of instrumental support over time. In terms of the link between support provided and received, the latter is expected to show a positive association with support provision at each time point. Thus, in line with the basic tenets of equity theory, we expect that the disabled elderly people in our sample will maintain the give and take of support in their relationships, even though some types of support provision may become increasingly difficult to accomplish.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Participants who had data for at least two of the three time points were included in this study. Seventy-one percent of the baseline sample had Time 2 data (n = 406), 62% (n = 356) had Time 3 data, and55% (n = 313) of older adults had data for all three waves. Longitudinal study participants (n = 449; M = 79.8 and SD = 6.9) were significantly younger, F(1,568) = 6.48 and p <.05, than drop-outs (
; 
and 
). Furthermore, as would be expected in longitudinal research with older adults, those who remained in the study had lower levels of functional disability at baseline (
; 
).
The study sample (
) was half (52%) female and mostly White (84%), with 12% African American and 4% Hispanic. Education was variable, with 30% having less than high school, 28% having high school, and 42% having more than high school. Although most participants reported having macular degeneration (66.4%), significant portions also reported having cataracts (35.7%) and glaucoma (26.2%). Half of the sample indicated that they had been having vision problems for the past year, whereas another 22% had problems with vision loss for the past 2 years and the remainder (28%) for 3–5 years.
Measures
Social support
Social support measures were based on the Arizona Social Support Interview Schedule (Barrera, Sandler, & Ramsey, 1981). Older adults were first asked to list the family members (Time 1 
; range = 0–15) and friends (Time 1 
; range = 0–12) who are important to them (up to 15 people were listed). They were then asked for each of the persons in their network from whom they received and to whom they provided affective support (three items: intimate interaction, advice, and positive feedback) and instrumental support (three items: material aid, physical assistance, and checking in on or watching home). Support variables were assessed separately for family members and friends. Affective support variables were computed by counting the number of family members and friends for whom participants usually provide, and from whom they usually receive, affective support. The variables for instrumental support were computed in the same way. The potential range for each support variable was 0–45.
Health
Self-rated health was assessed with a single-item indicator of subjective health, rated on a 5-point Likert scale ranging from 1 = very poor to 5 = excellent.
Functional disability
This variable was assessed with the OARS Multidimensional Functional Assessment Questionnaire (Center for the Study of Aging and Human Development, 1975). Fourteen items were summed that assessed difficulty in the performance of instrumental (seven items) and personal (seven items) daily living tasks rated on a 3-point scale (0 = no difficulty to 2 = needs help or cannot do task; range = 0–28; high score = high functional disability). Cronbach alphas for the three time points ranged from.90 to.91.
Use of rehabilitation services
Participants were asked whether or not they received any services (e.g., low vision clinical service, or training in home management).
Education
Education was assessed on a 4-point scale ranging from less than high school (1) to college graduate level or professional degree (4).
Gender
The gender variable was dummy coded, with 1 representing female and 0 representing male gender.
Analysis Plan
Hypotheses were tested by use of hierarchical linear modeling (HLM), a multilevel analysis (MLA) of two-level data (MLA, Version 3.2; Busing, Meijer, & van der Leeden, 1997). With HLM, both the average rate of change over time and individual variability in change over time can be assessed. Moreover, the extent to which individual characteristics (e.g., age or gender) explain individual variation at baseline or in change over time can be determined. In addition, HLM can tolerate an incomplete longitudinal data set because it utilizes data from all participants with available data for at least two time points.
Parameter estimates were derived in a two-level iterative manner, utilizing full information maximum likelihood (FIML), the numerical method implemented in the MLA program (MLA, Version 3.2; Busing, Meijer, & van der Leeden, 1997). "Level 1" equations reflect the level of the individual (i.e., the support provision score for each individual) at a given time point; b0 and b1 represent the intercept and slope, and e symbolizes the residual. "Level 2" equations, then, serve to account for the Level 1 coefficients b0 and b1 in terms of differences between individuals; estimates are labeled as gamma (
) and u.
First, an unconditional model and a conditional model were applied to each support provision variable. The unconditional model examined the dependent variable at a given point in time as a function of baseline status, rate of change over time, and random error. Multiplying the coefficient for the time effect by 18 gave an estimate of the total units of change across the three time points (labeled as 0, 6, and 18) of the study. The conditional model predicted that the dependent variable at baseline was a function of age, gender, health, functional disability (activities of daily living), and education. The conditional model also predicted that the rate of change in support provision would depend on whether or not adults used rehabilitation services and that a person's score on the dependent variable was associated with the level of received support at a given time point. As a way to reduce problems associated with multicollinearity among independent predictors, support receipt variables were centered at their respective grand means.
In the next step, these models were compared to test whether or not the conditional model represented a significant improvement over the unconditional model in terms of its fit to the data. The significance of model change was tested by using the difference between deviance values (i.e., lack of correspondence between model and data) from the unconditional and conditional model as 
2 value and the number of parameters that were added for the conditional model as degrees of freedom (see van Tilburg, 1998). A significant 2 value indicates a model improvement of the conditional over the unconditional model.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
) and 3.6 (
), and 76.2% (
) of respondents used at least one type of rehabilitation service over the course of the study.
|
2 value suggested that the conditional model with the added predictor variables had significantly better model fit than the unconditional model (see footnotes to Tables 2 and 3). Finally, self-rated health, functional disability, and use of rehabilitation services did not indicate any of the expected effects, suggesting that support provision over time was not affected by these predictors.
|
|
The right side of Table 2 shows that participants started out with an average score of 4.03 for instrumental family support provision, and this level decreased at the rate of approximately one family member (.90 units) over the course of the study. Although it was noted that scores for family versus friendship support provision at baseline were higher, there also appeared to be more of a decrease over time in the provision of instrumental family compared with friendship support. Age and gender showed significant effects at baseline; being younger and of male gender predicted higher levels of instrumental support provided to family members. Similar to results for instrumental friendship support, the provision and receipt of instrumental family support were positively related. Thus, as predicted, those who received more family support at a given time point were also likely to provide more support to family members. Yet, the strength of the associations between instrumental and affective support received and instrumental support provided for family seemed to be approximately equal (.33 and.29, respectively), suggesting that the likelihood of providing more instrumental support to family members could be enhanced by receiving affective as much as instrumental family support.
Affective Support Provided to Friends and Family
The left side of Table 3 shows that the average level of affective friendship support provided was 2.99. Contrary to prediction, on average, participants reported a significant decrease in this variable (.72 units over the course of the study). As hypothesized, younger age and higher levels of education predicted higher levels of support provided to friends at baseline. This was also the case for female gender. As expected, both affective and instrumental support received from friends were positively related to the outcome, with a strong association between affective support received and provided (.78) and a moderate, but significant link between instrumental support received and affective support provided (.30).
The right side of Table 3 shows that participants started out with an average affective family support provision score of 6.56. Interestingly, scores for affective family support provided at baseline appeared to be larger than all other support provision variables. There was a decrease in this variable of approximately 1.08 units over the course of the study. Variance components indicated variability around both the intercept and time effect. As expected, participants who were younger and had a higher level of education reported higher levels of support provision to family members at baseline. Also as expected, both affective and instrumental family support received were positively related to affective family support provision, with a strong association between affective support provided and received (.83) and a weaker link between instrumental received and affective provided (.15). This pattern seemed to differ from the links between support received from family and instrumental support provided to family, which were both moderate and appeared to be similar in magnitude.
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
In terms of average scores, the highest level of provision at baseline was reported for affective family support, and the lowest reported level was for instrumental friendship support. Within relationship types, more affective than instrumental support was provided to both family and friends. For older adults with vision loss, the provision of affective support may be the more preferable or feasible way of reciprocating the support they receive, because they may no longer be able to give certain types of instrumental support as a result of their disability. Reported levels of support provision may also reflect some degree of social desirability (i.e., wanting to see oneself as giving to others). However, the average decrease found in all support provision outcomes seemed to indicate that participants' perception of their own giving was not just based on "what's socially desirable." If this were the case, more stability in reported support provision would have been found.
With regard to age effects, being younger predicted higher levels of support provision in all outcome variables, with the exception of instrumental friendship support. It seems that the latter was generally the least common type of provided support, independent from the support provider's age. The age effects in the other outcomes were consistent with the prior research just reviewed, indicating higher levels of support provision among younger adults. However, in contrast to van Tilburg (1998), we did not find a differential age effect for instrumental versus affective support. This inconsistency may be due to the use of different types of measures or differences in study population. Therefore, the question of what kinds of support provision are more or less likely based on age deserves further research attention.
In terms of the gender effects, results varied by relationship type and support component, which emphasizes the importance of separate examination of these factors. Men provided higher levels of instrumental family support compared with women, whereas women provided higher levels of affective friend support compared with men. This seems consistent with prior research highlighting the prominent role of family for older men and the equal importance of both family and friend relationships for women (e.g., McIlvane & Reinhardt, 2001). It is reasonable that the traditional roles of men as responsible for instrumental tasks and women as nurturers emerged in relationship types that are especially important for each gender. Future research may explore in more detail the specific kinds of strategies within support types and their association with gender. A consideration of possible interactions between gender and other relationship characteristics (e.g., history of shared activities or level of affection in earlier life phases; Silverstein et al., 2002) in their effect on support provision may also provide insight.
Education was a significant predictor for the provision of affective friendship and family support. It is noted that, in an examination of the baseline sample of the current study, education emerged as a significant predictor of more positive adaptation to vision loss (Reinhardt, 2001). Thus, one explanation for the present finding may be that the ability to give support is part of successful adaptation. Education also had a significant, positive impact on instrumental friendship support provided. Perhaps instrumental family support provision is expected even if there is a lack of resources on the part of the support provider. In contrast, giving instrumental friendship support may be more dependent on one's resources.
Health and functional disability were expected to play a critical role in the prediction of support provision based on prior research described herein. However, these variables appeared to have little impact. Alternatively, this finding may be related to the fact that elders who dropped out of the study after baseline were more functionally disabled than study participants. Yet another explanation is that instead of a direct link, functional disability and health may have an indirect effect on support provision, mediated by support received from others. The relationship between health variables and support provision may also be interactive in nature rather than unidirectional. Finally, it is possible that an increase over time in disability rather than its baseline level affect support provision. Direct, indirect, and interactive effects, as well as the effect of change in disability on levels of support provided, could best be addressed in a dynamic, complex path model. The latter would have required a larger sample size and a more complete longitudinal data set than was available in the present study. Thus, future research using this approach to further examine the interrelationships among these variables is needed.
Contrary to predictions, use of vision rehabilitation services did not have an effect on rate of change in provision of instrumental support. This is interesting, considering that a major goal of rehabilitation is to maintain or regain some degree of independent functioning, which led to the hypothesis that the receipt of these services may help participants with the ability to continue to provide this type of support. However, this finding may be due to the limited variability in the rehabilitation measure (a majority of participants received some type of rehabilitation services). A more detailed account of rehabilitation services used would probably yield more variability, resulting in a larger effect size. Furthermore, future research should include other variables to predict the rate of change that reflect the changes older adults may make in adjustment to vision loss. This could include compensatory strategies learned in rehabilitation programs (e.g., developing systems to find things or match clothing) or any other adaptations that they may make in dealing with their daily life. Finally, it may be of interest to look at possible interactions between change in disability and use of rehabilitation.
With regard to the link between support receipt and provision, received support was a strong predictor for provision at each time point. People who received more also tended to give more support, which may be interpreted as an indication of a balance in give and take. Although these findings may be seen as a sign of reciprocity, it should be noted that transactions or perceptions of reciprocity were not specifically assessed in the present study. Lacking direct evidence for reciprocity, conclusions in this direction have to be very tentative, and other explanations for the link between receipt and provision of support have to be considered. One such explanation may be related to the way in which support was operationalized in the present study. Measures of support were based on the number of persons in the network from whom participants receive and to whom they give support. It may be the case that a person who was listed in the network does not exchange any of the types of support assessed. One potential limitation here is that, technically, with more network members being listed, there is a chance of higher scores for support across the different support functions. Levels of reported support could also be affected by the sociability of a person; that is, some participants who think of themselves as more "social" may have reported more of both types of support, whereas others reported less of both types. However, participants could have a few people in their network who provide a lot of support, or have many people in their network who provide just a few types of support, and the support scores would be similar. Thus, the nature of the measure constitutes a potential but not a necessary limitation in this regard.
Another possible explanation for the link between receipt and provision is that there may have been a reporting bias, in the sense that participants like to think of themselves as being part of relationships that are reciprocal. However, even in case of such reporting bias, the finding of a positive association between support provided and support received seems to fit within the framework of equity theory in this sample of elders with disability, because it would still indicate that people feel the need to have balanced social exchanges. The link between receipt and provision was particularly interesting with regard to instrumental support, because this support type had been expected to play a minor role for older adults dealing with chronic impairment who may have limited capacity to provide instrumental support. It is noted, however, that the measure of instrumental support used in this study included different types of strategies (e.g., physical assistance, or checking in on or watching home). Material aid (e.g., financial assistance or giving gifts), one of several common types of assistance exchanged between older parents and adult children (Cantor & Brennan, 2000), was also included. This may explain why visually impaired elders were still able to provide instrumental support and why there was a positive link between the receipt and provision. If the definition of instrumental support included strategies that a person can provide within the limitation of a chronic impairment, the notion of balancing out, as drawn from equity theory, could be broadened to include the possibility of compensating for an increased need of certain kinds of instrumental support with increased provision of other kinds of instrumental support.
There seemed to be no clear evidence that participants actually tried to compensate for a greater need in instrumental support by providing more affective support. Rather, the link between support provision and receipt was mostly stronger within rather than across types of support. Interestingly, though, this was not the case for the link between provision of instrumental family support across both the receipt of instrumental as well as affective family support. The provision of instrumental support to family members seems to depend as much on the levels of instrumental support received as on the receipt of affective support from family members. Thus, our findings neither confirm nor exclude the possibility of "balancing out." However, they do indicate that there may be different ways of providing support that are feasible, even for older adults who deal with a chronic impairment. This may have implications for vision rehabilitation interventions. Individuals may benefit from understanding that having a chronic impairment and needing more support does not necessarily mean that they can no longer reciprocate or that they can only reciprocate with affective support. Helping older adults with vision loss to explore new ways of reciprocating support that take into account the limitations caused by the impairment could be a promising intervention strategy.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication June 21, 2002. Accepted for publication December 19, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. P. Reinhardt, K. Boerner, and A. Horowitz Good to have but not to use: Differential impact of perceived and received support on well-being Journal of Social and Personal Relationships, February 1, 2006; 23(1): 117 - 129. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
