| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Arlene R. Gordon Research Institute, Lighthouse International, New York.
2 Department of Psychology, Fordham University, New York.
Address correspondence to Kathrin Boerner, Lighthouse International, 111 E. 59 St., New York, NY 10022. E-mail: kboerner{at}lighthouse.org
 |
Abstract |
|---|
 TOP Abstract MethodResultsDiscussionReferences |
|---|
Methods. Latent growth curve methodology was used to examine change over time in social negativity and instrumental family support in relation to age, gender, and family network size at baseline. Participants, 570 older adults with chronic visual impairment, were interviewed three times over an 18-month period.
Results. Social negativity received showed a decrease over time, whereas levels of social negativity initiated remained more stable. Links with instrumental support were positive but stronger for received compared with initiated social negativity.
Discussion. The differential pattern of stability and change over time in received versus initiated social negativity and their links to instrumental support suggest different origins for the initiation versus receipt of social negativity.
After decades of research on the beneficial effects of social relationships, there has been an increasing interest in the nature and implications of negative social exchange. Definitions of negative exchange, also referred to as "social negativity" (Finch, Okun, Pool, & Ruehlman, 1999
), include social behaviors such as displaying negative affect (e.g., being angry, upset, or hostile), being overly critical and demanding, showing insensitivity, and making goal attainment difficult (i.e., social interference/hindrance; Rook, 1984; Ruehlman & Karoly, 1991). The potential cost of such social exchange for older adults' well-being (e.g., Reinhardt, 2001) and physical health outcome (Kiecolt-Glaser et al., 1993) has been repeatedly demonstrated. In fact, some authors have suggested that aspects of well-being may be more strongly affected by negative than by positive exchanges (Morgan, Neal, & Carder, 1997; Rook, 1997). This notion was supported in a recent study that found positive exchange to be unrelated and negative exchange to be a significant predictor of affect over time (Newsom, Nishishiba, Morgan, & Rook, 2003).
There is also evidence that the impact of negative exchange can be intensified under stressful life circumstances. For example, Ingersoll-Dayton, Morgan, and Antonucci (1997) found that negative exchanges had a greater impact on well-being among more stressed individuals, whereas the effect of positive and negative exchanges was comparable among less stressed persons. Thus, being confronted with negative exchanges may be particularly problematic when a person is dealing with other life stressors (Rook, 2003). Dealing with a chronic health impairment is one common stressor occurring in later life that has been shown to predict negativity in social interactions involving assistance provided to disabled older adults (Newsom & Schulz, 1998). Negative reactions to being helped with daily life tasks have also emerged as predictors of depression in this population (Clark & Stephens, 1996; Newsom & Schulz, 1998).
Because social negativity yields potentially harmful consequences for well-being, questions about the chronicity of these types of interactions become particularly critical. Some studies have shown a reduction in negative exchanges with age (e.g., Akiyama, Antonucci, Takahashi & Langfahl, 2003; Smith & Goodnow, 1999), suggesting an overall decrease over the life course. Yet, this research has primarily been cross-sectional in nature. Evidence from longitudinal research, on the other hand, points to a pattern of stability in negative exchanges over time (Krause & Rook, 2003; Pagel, Erdly, & Becker, 1987; Wheaton, 1990). On the basis of these findings, Krause and Rook (2003) concluded that negative exchanges may be fairly chronic. However, this research did not address the predictive role of life circumstances such as living with a chronic disability. Furthermore, studies that address "the other side" of a social exchange, which is the initiation of negativity by older adults, have been scarce. Thus, important questions remain in regard to how stable negative exchanges may be when a person is dealing with other life stressors and whether such a pattern may be different for received compared with initiated negativity. The current study extends prior work by examining stability and change in both initiated and received social negativity over time among older adults with chronic disability. It should be noted that the term "social negativity" in the current article is not meant to imply a trait-like attribute. Rather, it is used as a synonym for other terms that are used in the support literature (e.g., negative interactions, negative exchange).
Patterns of Social Negativity Over Time
Findings from several longitudinal studies suggest that negative exchanges may be a rather stable phenomenon. That is, some people are more likely to report social interactions that are characterized by negativity than others. In a study on spousal caregivers, Pagel and colleagues (1987) found that baseline indicators of negative exchange were strongly linked to their follow-up measures 12 months later. Wheaton (1990) reported similar strong links between negative exchange measures over a 2-year period in a study on stress and well-being. In a recent study, Krause and Rook (2003) examined the stability of conflict among older adults by relationship type. They found not only that negative exchanges were quite stable over a 6-year period but also that those older adults who reported interpersonal conflicts in one relationship type (e.g., spouse) tended to report them in other relationships (e.g., child). The authors concluded that negative exchanges are fairly chronic and occur across relationships and that this stability may be due to the influence of person factors (e.g., limited social skills, personality). This is consistent with prior research demonstrating a critical role of personality variables in the prediction of negative social exchange (Finch et al., 1999; Newsom & Schulz, 1998).
The notion of stability in negative social exchange shows an interesting contrast to longitudinal findings on other social network and support outcomes, indicating varying patterns of change in these variables over time, depending on support and relationship type assessed (Martire, Schulz, Mittelmark, & Newsom, 1999; Reinhardt, Boerner, & Benn, 2003; van Tilburg, 1998). Although it is possible that these findings from the general support literature do not apply to the phenomenon of social negativity, there are also conceptual reasons to expect change rather than stability in negative exchange over time. According to socioemotional selectivity theory (Carstensen, Isaacowitz, & Charles, 1999), people become more selective in their choice of social partners as they grow older, in the sense that they tend to reduce contact with network members who provide fewer emotional rewards. Regarding social negativity among older adults, this means that the expected pattern over time would involve a decrease rather than stability. Theoretical formulations from the literature on dealing with life stress make similar proposals. For example, the "mobilization–minimization hypothesis" holds that negative events trigger strong psychological and social responses that mobilize the organism to actively seek out ways to minimize or undo the negative impact of the event (Taylor, 1991). Thus, one's response to the experience of social negativity would likely include efforts to minimize its occurrence.
If the occurrence of social negativity in one's relationships is indeed a matter of limited social skills or personality traits, this would mean that older adults may not only be on the receiving end of social negativity but may also initiate these exchanges. A study by Vinokur and Vinokur-Kaplan (1990) even found that older adults reported more initiation than receipt of social negativity. This finding underscores the importance of including measures of both initiation and receipt of negativity in studies of social exchange. Taking a more indirect approach, Krause and Rook (2003) interpreted their finding of older adults reporting similar levels of negative exchanges across relationship type as evidence that negative exchanges originate with the older person rather than others and that this explains stability over time and across relationships. This interpretation could be substantiated by future research that directly assesses older adults' initiation of negativity in their social relationships.
Chronic Impairment and Social Negativity
The experience of chronic impairment and associated functional disability provides an important context to study social support as it places people in a position of needing help. One common late-life impairment reported by about 20% of adults aged 65 and above is age-related vision loss (Lighthouse Research Institute, 1995). The loss of vision late in life can be devastating because of its strong relationship with functional disability and the image of helplessness and dependency associated with blindness (Horowitz, 1994). Research has identified the receipt of social support as an important resource for more positive adaptation to vision loss in later life both initially when elders become impaired and over time (Jacobs, 1984; Reinhardt, 1996, 2001).
Support receipt, however, may affect levels of social negativity in several ways. Stroke patients in Clark and Stephens' study (1996) reported over- or underestimation of the amount of help needed and being helped incorrectly as reasons for negative reactions to assistance in daily life activities. Moreover, researchers have suggested that support enactment may emphasize a person's inability to complete tasks independently (Oxman, Freeman, Manheimer, & Stukel, 1994). Potential feelings of dependency that may result when one needs help to accomplish everyday tasks that used to be conducted independently can set the stage for negative interactions (Newsom, 1999). Thus, in a population of older adults who are in need of instrumental support, social negativity is, at least in part, likely to be related to the receipt of instrumental support.
If this were indeed the case, then this link between social negativity and instrumental support should also exist over time. In a previous analysis of change in social support receipt, based on the same dataset used for the current article, instrumental support receipt showed a decrease over time (Reinhardt, Boerner, & Benn, 2003). We suggested that this decrease may have been a result of higher support needed in the beginning stages of adaptation to vision loss concomitant with the general tendency among older adults to reduce their social network to its core members (Carstensen, Isaacowitz, & Charles, 1999). If instrumental support receipt decreases and if social negativity is indeed closely related to this support receipt over time, social negativity should decrease along with the receipt of support, rather than showing stability.
When studying social negativity, family relationships, in contrast to friend or formal relationships, are of particular importance for both conceptual and empirical reasons. From a conceptual standpoint, focusing on social negativity in families is more relevant because one is less likely to accept negativity in relationships that are more optional (i.e., friendships) and may even be inclined to end such negative relationships while focusing on network members who play a positive role in one's life. Empirical evidence suggests that the level of social negativity tends to be higher for family relationships compared with friendships (Morgan, 1989) and that social negativity in family but not friendship relationships is associated with higher depressive symptomatology (Reinhardt, 2001).
Thus, the current study seeks to understand the origins of negative exchange in later life by looking at the links of both initiated and received social negativity with the receipt of instrumental family support over time in a population of older adults who are, by definition of their vision status, likely to be in need of support. As correlates, we included family network size, gender, and age so that their association with our key study variables could be evaluated in the context of the analyses of change in negativity and support variables. Although initiated and received social negativity variables are expected to show a strong positive relationship, addressing them separately will allow us to determine if these variables indeed show a similar pattern of change over time. Further, this will permit a direct examination of the initiation and receipt of social negativity regarding their links to instrumental support and the correlates.
Hypotheses
We predict that social negativity will be positively related to instrumental family support both initially and over time. Further, in contrast to prior research suggesting stability in social negativity over time (Krause & Rook, 2003), we predict a decrease in social negativity. Stability would imply that those who report high levels of negativity initially tend to also report high levels over time. However, we expect that those who start out with the highest level of negativity will be the ones who show the most decrease over time. This prediction is based on the notion that, over time, most people adjust to the situation of dealing with a chronic impairment to some extent and may therefore need less instrumental support. Both of these circumstances are likely to reduce the potential for negativity in social interactions.
With respect to the correlates, we predict a negative association between age and baseline status of social negativity and a positive association between network size and negativity at baseline. This means that being older would be related to lower levels of social negativity, whereas reporting larger networks would be related to more social negativity. These hypotheses were based on prior evidence suggesting that higher levels of social negativity may be a matter of exposure to more network members (Smith & Goodnow, 1999) and that social negativity tends to decrease with age (Akiyama et al., 2003; Smith & Goodnow, 1999). Furthermore, findings with regard to the links between gender and support variables cross-sectionally (Antonucci & Akiyama, 1987; Martire et al., 1999) as well as longitudinal findings for gender and age effects on subsequent support assessments (Martire et al., 1999; van Tilburg, 1998) have been mixed both across studies and within studies depending on type of network and support variables. Therefore, associations of gender with baseline status of social negativity and of all correlates with change in negativity over time are being explored.
|
METHOD |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
). Income adequacy also varied, with a small portion reporting not having enough money for necessities (9%), just over one third (40%) having just enough money, and 50% reporting having enough for necessities and extras.
Measures
Family network size, support, and negativity
Measures were based on the Arizona Social Support Interview Schedule (Barrera, Sandler, & Ramsey, 1981). To assess family network size, older adults first listed family members who were important to them (up to 15 people). They were then asked from which of these persons (a count) they received instrumental support (i.e., material aid, physical assistance, checking in on/watching home). Scores indicate number of instrumental support types across persons (e.g., if a participant received all three types of instrumental support from one provider, the score would be 3). A single item assessed social negativity received (number of family members who sometimes make you angry or upset), and another single item was used to capture social negativity initiated by the participant (number of family members who you sometimes make angry or upset).
Age and gender
These variables were assessed with single item indicators. The gender variable was dummy coded, with (1) representing female and (0) representing male gender.
Analysis Plan
The research objectives of this article were addressed with latent growth curve methodology (LGCM; Duncan, Duncan, Striker, Li, & Alport, 1999; Raykov & Marcoulides, 2000), using the latest version of LISREL (8.54; du Toit & du Toit, 2001). This methodology was chosen because it allows the examination of change over time in two repeatedly followed variables within one model (social negativity and instrumental family support) and the assessment of their interrelationships at baseline as well as over time. In addition, it permits relating these patterns to correlates of interest (age, gender, and family network size). Yet a further benefit of the LGCM framework in repeated measure settings is that it provides an optimal way to deal with the fact that a considerable part of the sample participated in less than a complete number of repeated observations. This can be accomplished by applying the full information maximum likelihood (FIML) method for purposes of fitting latent variable models of change that are conceptualized within this framework (du Toit & du Toit, 2001). With this method, models of interest are fitted using all available data from all participants. The pertinent likelihood function is constructed by accumulating across the sample information bearing on the estimated parameters. Thus, the contribution of each person is fully taken into account, regardless of (a) the number of assessment occasions on which they have been followed and (b) which other variables related to the repeated measures that they have missing data on (Collins, Schafer, & Kam, 2001; du Toit & du Toit, 2001). Unlike imputation methods for handling missing data (Schafer, 1997), with FIML, one is not assigning (imputing) any value to a subject on a variable on which he or she may not have provided data. Instead, all available data in the sample are used to obtain the likelihood function, through whose maximization (as in the complete data case) one obtains estimates and standard errors for the parameters of the fitted model in addition to measures of its fit.
To respond to the research questions of interest, within the LGCM we utilized the level-and-shape model that has been widely popularized by McArdle and colleagues during the last decade (Duncan et al., 1999; McArdle, 1988; Raykov & Marcoulides, 2000). The model is built upon two latent variables: baseline true status of the repeatedly observed measure (i.e., social negativity received, social negativity initiated, instrumental family support) and change along this latent dimension across the entire study period. The loadings of the first latent variable (baseline true status) on all repeated assessments were fixed at 1, the loading of the second latent variable (latent change) on the first assessment was fixed at 0, its loading on the second assessment was left free, and that on the last assessment was fixed at 1 (McArdle, 1988). Two specific models were fitted that have identical specification (Model 1; Figure 1), with the only exception being that in Model 2, the repeated measure of received social negativity is replaced with the repeated measure of initiated social negativity (Figure 2).
|
|
|
RESULTS |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
|
When fitted to the data, using the above-mentioned FIML method, Models 1 and 2 were found to be acceptable means of data description and explanation. Specifically, for Model 1, the fit indexes were as follows: FIML 
2 = 29.74, df = 14, p =.008, root mean square error of approximation (RMSEA) =.044 with a 90% confidence interval (.022,.067), and p value for test of close fit =.63. The fit indexes for Model 2 were as follows: FIML 2 = 23.47, df = 14, p =.053, RMSEA =.034 (.0,.058), and p value for test of close fit =.85.
Social Negativity Received (Model 1)
As predicted, there was a significant average overall decrease in received social negativity (mean estimate = –.23, SE =.06, asymptotic z value = –3.64; see Figure 1). Most of this decrease appeared to occur between assessment occasions 1 and 2, with very little change between measurement points 2 and 3 (factor loading for change between first and second assessment occasions estimated at 1.07, SE =.08, z = 13.49; McArdle, 1988). However, there was also significant interindividual variability in this overall decrease (variance estimate = 1.31, SE =.18, z = 7.18), indicating that participants did not all change uniformly. As expected, baseline status in received social negativity was significantly negatively related to change in this dimension (covariance estimate = –1.61, SE =.14, z = –11.27, correlation estimate = –.90). Thus, there was a strong tendency for elderly persons starting high on received social negativity to be among those who dropped most on this dimension by the end of the study. We note that this is a finding on the latent dimension itself, that is, after accounting for measurement error and hence not explainable with the classic regression to the mean phenomenon. In line with prior analyses (Reinhardt, Boerner, & Benn, 2003), there was some drop in instrumental family support that approached significance (mean estimate = –.35, SE =.18, z = –1.94). There was also significant interindividual variability in this change: variance estimate = 4.62, SE = 1.31, z = 16.92. Associations of instrumental family support with social negativity received were partially consistent with predictions. Baseline status in instrumental family support was significantly and positively related to baseline status in received social negativity (covariance estimate = 1.26, SE =.27, z = 4.67, correlation estimate =.23), suggesting that at the beginning of the study, persons high/low on the former dimension tended to be among those high/low on the latter dimension as well. However, baseline status on instrumental family support was significantly negatively related to change in received social negativity (covariance estimate = –.53, SE =.25, z = –2.12, correlation estimate = –.14), indicating that persons high initially on instrumental family support tended to be among those dropping more than average on received social negativity. Finally, change in instrumental family support was only weakly related to change in received social negativity (covariance estimate =.33, SE =.25, z = 1.30, correlation estimate =.13). Although this association did not reach significance, it showed the predicted positive direction.
With regard to the baseline correlates, age was only weakly nonsignificantly related to baseline status of and change in received social negativity (covariance z values being.92 and –.22, respectively). Gender was significantly and negatively related to baseline status of social negativity received (covariance estimate = –.21, SE =.04, z = –4.78, correlation estimate = –.19). Women tended to be among those lower than average on social negativity received. Conversely, gender was significantly and positively related to change in received social negativity (covariance estimate =.11, SE =.04, z = 2.63, correlation estimate =.14), suggesting a tendency for women to be among those dropping less than average on social negativity received with time. As expected, family network size at baseline was significantly and positively related to baseline status on received social negativity (covariance estimate =.81, SE =.19, z = 4.17, correlation estimate =.17), suggesting a tendency for persons with larger family networks to be among those starting higher than average on received social negativity. Further, family network size was significantly and negatively related to overall change in received social negativity (covariance estimate = –.40, SE =.19, z = –2.17, correlation estimate = –.12). Since there was on average a considerable drop in received social negativity over time, this suggests that people with large networks tended to be among those who dropped more than average on this variable.
Social Negativity Initiated (Model 2)
Contrary to prediction, there was no significant average decrease in social negativity initiated (mean estimate = –.10, SE =.06, z = –.171). This means that compared with receipt of social negativity, participants showed more stability in reporting on negativity that they had initiated themselves. However, there was interindividual variation in this latent change variable (variance estimate = 1.38, SE =.21, t = 6.69). As was the case for received social negativity, baseline status in social negativity initiated was significantly and negatively related to change in this dimension (covariance estimate = –1.48, SE =.15, t = –10.09, correlation estimate = –.83). Thus, there was a moderate to strong negative relationship between starting position and change on the underlying latent dimension of initiated social negativity. Since there was on average no significant drop in this dimension, but there was individual variability in change over time, elderly persons high on baseline status of social negativity initiated tended to be among those dropping most on this dimension. As expected and also found in Model 1, baseline status on instrumental family support was significantly positively related to baseline status on initiated social negativity (covariance estimate = 1.11, SE =.26, z = 4.27, correlation estimate =.21) but, unlike in Model 1, not related to change in the latter dimension (covariance estimate = –.18, SE =.24, z = –.74, correlation estimate =.07).
As was the case for received social negativity, age was only weakly (nonsignificantly) related to baseline status of and change in social negativity initiated (covariance asymptotic z values = 1.29 and 1.72, respectively, corresponding correlation estimates =.05 and –.09), and gender was significantly negatively related to baseline status in social negativity initiated (covariance estimate = –.14, SE =.04, z = –3.15, correlation estimate = –.13), suggesting that women tended to be among those lower on initiated social negativity. However, unlike in Model 1, gender was only weakly (nonsignificantly) related to change in this dimension (covariance estimate =.01, SE =.04, z =.35, correlation estimate =.02). As in Model 1, family network size was significantly positively related to baseline status on social negativity initiated (covariance estimate =.52, SE =.20, z = 2.64, correlation estimate =.11), indicating that elderly persons with larger family networks tended to be among those who also initiated higher than average social negativity themselves. However, unlike in Model 1, family network size was not related to change in social negativity initiated (covariance estimate = –.14, SE =.18, z = –.77, correlation estimate = –.04).
|
DISCUSSION |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
that stability in social negativity is due to elderly persons' initiation. However, this explanation remains tentative because there was interindividual variation in initiated social negativity, which implies change among some participants. Those who reported the highest initial levels of social negativity were also likely to show the most decrease in both received and initiated social negativity. This suggests that stability in initiated negativity was not the case for all participants. The impact of person factors on social interactions may override the situation factor (i.e., dealing with a chronic impairment) in some cases, but not in others. Future research that teases apart the importance of these factors for certain individuals under certain conditions may help clarify this issue.
The decrease in social negativity received could be due to an active effort of maximizing the benefits gained from the social network and minimizing negative experiences, as suggested by socioemotional selectivity theory (Carstensen et al., 1999). This means that we are likely to focus on network members who have a positive role in our lives rather than on those who repeatedly upset us. Although this may be more difficult in family than in friend relationships, there may be some room for such changes even within the family. Another possibility is that network members may withdraw if they feel under constant attack, which can result in less receipt of negativity with no associated change in initiated negativity.
Hypotheses regarding associations of social negativity and instrumental family support were partially consistent with predictions. As expected, participants with higher initial levels of instrumental support received also reported higher baseline levels of received and initiated social negativity. This supports the notion of instrumental support need as having the potential for conflict around issues of dependency. Also, those with higher initial levels of instrumental support were likely to drop more in social negativity received. Perhaps being in need of instrumental support may cause strain in family interactions, which can improve over time as the person adapts to the situation. Or, those who begin with the most social negativity may lose more support providers over time owing to the withdrawal of network members, thus reducing possible sources of negativity. Research assessing reasons for change in network composition could help clarify the more viable explanation. Contrary to prediction, change in social negativity was not significantly related to change in instrumental support, and change in social negativity initiated was not significantly associated with instrumental support at baseline. Initiating social negativity may indeed have more to do with one's personality and therefore be less likely to change as support need situations change.
Regarding the role of the correlates, we did not find a significant association between age and social negativity. Perhaps age turns out to be less important when variables such as family network and instrumental family support are taken into consideration. As expected, those who had larger family networks generally started out with more social negativity. However, they also had more of a decrease in social negativity received over time. This seems reasonable as those who start out with fewer members in their family network may have less room to regulate their networks. They may have to retain the few relationships they have, even if they have negative components. A possible explanation for the finding that family network size did not predict change in social negativity initiated may be, again, that it is easier to turn away from people who upset us than to change our own reaction to others.
Women compared with men reported receiving and initiating less social negativity at baseline. Women may be better able to minimize negative interactions owing to their key role in the family context and greater experience in handling difficult interaction situations. However, women also tended to show less of a drop in social negativity received, which may mean that because of their role in the family, regulating one's family network by turning away from those who upset us may be less of an option for women who may feel more responsibility for maintaining relationships than men. The lack of a significant link of gender with initiated social negativity is consistent with the suggestion that there is less room for regulation with this variable overall.
The findings presented herein add to the literature by showing a differential pattern of stability and change over time in received versus initiated social negativity and demonstrate the interrelatedness between social negativity and the receipt of instrumental support in a population of older adults who are, over time, trying to adjust to the consequences of chronic disability. However, we acknowledge a few important limitations that should be taken into consideration when interpreting the findings. First, the repeatedly presented measures of social negativity as well as instrumental family support exhibited some lack of normality. The method we employed for model-fitting purposes that utilizes all available data and optimally handles missing data, FIML, is formally based on the assumption of normality. Yet, this method has been found in recent studies to be fairly robust against violations to normality (Collins et al., 2001; Schafer & Graham, 2002). Thus, our findings are to be treated with some degree of caution, and the need for replication studies remains. Second, the social negativity measures used in the current study were unidimensional, in that they reflected only the affective component (e.g., being angry or upset) of social negativity. Although negative affect appears to be a key feature of this construct, it is possible that the inclusion of other components (e.g., criticism or hindrance) would have yielded a slightly different picture.
Third, while examining both social negativity initiated and received, we were unable to determine whether or not these were exchanges among the same people in the family network. To capture the interactive dynamic of social negativity, however, data are needed that can show what happens over time within particular social dyads. This type of dyad data would be ideal to examine the interrelations of negativity initiated and received concurrently and over time to further inform the complex nature of negative social exchange. Future research that uses dyadic data to focus on the interactive component of negative exchange could directly address the important question of whether social negativity develops within particular situations or relationships or whether it is a characteristic that lies within the person. In the first case, one would expect a strong association between initiated and received negativity within a social dyad, whereas in the second case, a weak to moderate link may be more likely.
The current study was also limited to the perspective of the person in the family network dealing with the chronic impairment and resulting support needs. Future work that assesses the degree of concordance in family members' perception of social negativity and their explanation for this negativity could deepen our understanding of patterns of social negativity over time in the context of dealing with chronic disability in later life. Finally, further research investigating the impact of other life stressors on social negativity is needed to see if differential effects emerge depending on the nature of the stressor.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication October 23, 2003. Accepted for publication April 5, 2004.
|
References |
|---|
|
TOPAbstractMethodResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
