SOCIAL AND EMOTIONAL LONELINESS AS PREDICTORS OF COGNITION IN LATER LIFE

Abstract To explore the viability of a model illustrating the potential relationship between both social and emotional loneliness and both psychometric and everyday cognition in later life, 575 older adults (M = 73.49) completed measures of crystallized (Gc) and fluid (Gf) ability as well as indicators of self-rated participation in 84 everyday cognitive activities, self-rated stimulatory value of each activity, attitudinal predisposition toward an engaged lifestyle and everyday cognitive failures. Measures of social support, caregiving stress, needs for cognition and cognitive self-efficacy were treated as mediators of the loneliness-psychometric/everyday cognition relationship, controlling for age, gender, health, and education. Hierarchical regression analyses suggested that social loneliness predicted (p < .04) Gc as mediated by social support, cognitive self-efficacy and need for cognition, whereas emotional loneliness similarly predicted (p < .04) Gf. Parallel analyses indicated that social loneliness predicted (p < .04) everyday cognitive failures and that both social and emotional loneliness predicted (p < .03) engaged lifestyle attitudes. In each case, the overall model was statistically significant (p < .01). For everyday cognitive activities and the stimulation values of such, neither type of loneliness was predictive, though lifestyle attitudes and lifestyle activity were moderately interrelated. These findings support a model incorporating distinct dimensions of loneliness as a predictor of diverse aspects of both psychometric and everyday cognition. This suggests that not only are the antecedents of cognition in late adulthood social/emotional in nature, but also that interventions targeting the prevention of loneliness may enhance cognitive functioning in later life.

life; making social support vital to prevent social frailty. The municipalities and universities have been hosting interventions, e.g. club activities, to make social connection but the effects remain unclear. In this study, the relationship between activity participation and the size of social support was explored to develop an effective method to increase social connections. A cross-sectional study was conducted in October 2018 and 200 questionnaires were distributed at a lunch event at the housing complex. The participants were solitary and independent seniors over the age of 70 (M:F=14.4:67.5). The frequency of going out to participate in club activity, see friends, and the geriatric social support scale were used for analysis. As a result, compared to seniors who went out to see their friends, the size of social support was smaller in groups that remained isolated (N = 161, odds ratio = 0.26, 95% CI=0.11-0.60). On the contrary, seniors who joined club activities had similar size of social support despite the frequency of participation. Clubs in Japan are often closed groups with limited membership, which may have restricted the addition of new social connection. Future intervention researches should focus on modifying the membership system of clubs and opening of a public space where seniors can casually access and talk to acquaintances.

AGE DIFFERENCES IN COGNITIVE PERSONAL NETWORKS
Christopher S. Marcum, 1 Jielu Lin, 2 and Laura M. Koehly 3 , 1. National Institutes of Health,Bethesda,Maryland,United States,2. Northern Arizona University,Flagstaff,Arizona,United States,3. National Human Genome Research Institute,National Institutes of Health,Bethesda,Maryland,United States Previous research has found a negative association between network size and age, suggesting that people experience greater isolation with advancing age. In this paper, we evaluate age differences in how individuals perceive their social worlds to be structured, rather than focusing solely on network size. A nationally represented sample of respondents (n=1,824) reported on their own ties to their close personal network members (i.e., ego-alter ties) as well as their perceptions of acquaintanceship between those members (i.e., alter-alter ties). We used social network analysis to assess how the structure of these relationships vary by respondent age. We find a positive association between respondent age and personal network size and a negative association between network members' ages and the number of ties respondents' perceive their members to have to each other. This effect significantly weakens as respondent age increases. Moreover, we find evidence of age-homophily, intergenerational contact spanning three generations in both ego-alter and alter-alter ties, and age differences in ego network composition. Our results suggest that the evolution of our social worlds across the life course shifts in terms of size and structure. While contemporary close personal networks may grow slightly with age, perceived social ties among one's network members become less cohesive and less diverse with age. We discuss these results in the context of recent findings that suggest aging uniformly insulates individuals from social contact from both structural and symbolic perspectives.

DOES MOVING LATER IN LIFE INFLUENCE IN-PERSON CONTACT WITH CHILDREN, FAMILY, AND FRIENDS?
Joonyoung Cho, 1 and Jacqui Smith 1 , 1.

University of Michigan, Ann Arbor, Michigan, United States
Relocation is common in later life and older adults differ the distance they move. One possible consequence of relocation is that in-person contact frequency with social network members changes. To date, relatively little is known about how older adults' in-person contact frequency with their children, family members, and friends is influenced by the distance they move and if this differs by age group (50-64, 65-74, 75+). To examine this, we used information from the Health and Retirement Study about geographic mobility and social network contact frequency. The sample was restricted to respondents over age 50 in the 2006 wave with data on contact frequency with children, family members, and friends in 2006 and 2014 (N=5159). Distance moved from 2006 to 2014 was categorized as: stayer, <5 miles, 5-49.9 miles, and ≥50 miles. Linear regressions with covariates controlled revealed that moving ≥50 miles was significantly associated with less frequent in-person contact with family members In contrast, the 75+ group who moved ≥50 miles were significantly less likely to have in-person frequency only with their friends (β=-0.58, 95%CI [-1.04, -0.14], p≤0.05). These findings are discussed in relation to theories about social support and emotional well-being in old age.

SOCIAL AND EMOTIONAL LONELINESS AS PREDICTORS OF COGNITION IN LATER LIFE Bert Hayslip 1 , 1. university of north texas, Denton, Texas, United States
To explore the viability of a model illustrating the potential relationship between both social and emotional loneliness and both psychometric and everyday cognition in later life, 575 older adults (M = 73.49) completed measures of crystallized (Gc) and fluid (Gf) ability as well as indicators of self-rated participation in 84 everyday cognitive activities, self-rated stimulatory value of each activity, attitudinal predisposition toward an engaged lifestyle and everyday cognitive failures. Measures of social support, caregiving stress, needs for cognition and cognitive self-efficacy were treated as mediators of the loneliness-psychometric/everyday cognition relationship, controlling for age, gender, health, and education. Hierarchical regression analyses suggested that social loneliness predicted (p < .04) Gc as mediated by social support, cognitive self-efficacy and need for cognition, whereas emotional loneliness similarly predicted (p < .04) Gf. Parallel analyses indicated that social loneliness predicted (p < .04) everyday cognitive failures and that both social and emotional loneliness predicted (p < .03) engaged lifestyle attitudes. In each case, the overall model was statistically significant (p < .01). For everyday cognitive activities and the stimulation values of such, neither type of loneliness was predictive, though lifestyle attitudes and lifestyle activity were moderately interrelated. These findings support a model incorporating distinct dimensions of loneliness as a predictor of diverse aspects of both psychometric and everyday cognition. This suggests that not only are the antecedents of cognition in late adulthood social/emotional in nature, but also that interventions targeting the prevention of loneliness may enhance cognitive functioning in later life. Socioemotional Selectivity Theory poses that as we age our motivations transition from knowledge focused to emotionally focused (Carstensen, 1995). This shift to emotional motivation increases the relevance of relationships and social interactions for older adults. We examined different aspects of social support: frequency of positive/negative social interactions, satisfaction with positive social interactions, and bothered by social interactions --to investigate these effects on one's global well-being. Negative and positive social exchanges are linked to psychological health (Newsom et al., 2005), however one's perceptions of those social interactions are important to consider as well --being satisfied or bothered by social interactions shows a better perspective of the individual's experience than simply recording frequency. The results of our multi-group path analysis show that there are different effects of social supports on global well-being contingent on age, consistent with socioemotional selectivity theory. For adults under 30 years old (Mage = 24.0, range 18 to 29) social support did not significantly relate to well-being. For adults over 30 and under 50 (Mage = 38.9) frequency of positive social interactions is significantly related to well-being (B = .201). For adults over 50 (Mage = 58.8, range 50 to 87) the perception of social exchanges, not their frequency, are what influence well-being: both satisfaction with positive social interactions (B = .402) and being bothered by negative social interactions predict well-being (B = -.193). It is important to know that older adult's perceptions of social exchanges effect their well-being, future directions are discussed.

FEELING AND LIVING WELL: NETWORK COMPOSITION AND PREDICTING SELF-RATED HEALTH FOR RACIAL AND ETHNIC MINORITIES Eli X. Ornelas 1 , 1. University of Nebraska-Lincoln, Lincoln, Nebraska, United States
Racial and ethnic inequities in health among older adults in the U.S. are well documented. A substantial amount of the health literature focuses on physical and mental disparities; however, less research has examined racial and ethnic differentials in subjective, self-rated health (SRH). Prior research has documented racial and ethnic inequities in SRH, though mechanisms by which these disparities occur are still largely unknown. One potential mechanism by which these disparities may arise is through unequal access to psychosocial resources through variability in social networks. Utilizing data from the 2006 Health and Retirement Study (HRS) along with the 2006 HRS Psychosocial and Lifestyle Questionnaire, the current study seeks to explore racial and ethnic differentials in SRH from a social network perspective. Ordinal logistic regression is used to predict SRH by race and Hispanic ethnicity potentially mediated by possession and number of friends, frequency of contact with friends, and psychosocial and subjective well-being. Results indicate that older black adults are less likely to rate their health in a higher category of SRH than older white adults, and this relationship is not significant for other racial groups and Hispanics. Additionally, psychosocial and subjective well-being and frequent written-communication with friends are found to significantly predict better SRH. The results suggest that feeling well mentally is crucial for SRH, but equally important is regular contact with friends. Further research should employ more robust measures of social networks to elucidate the role that network composition plays in predicting SRH for racial and ethnic minorities.

VOLUNTEER ACTIVITY CHANGES AFTER MIDDLE YEARS OLD AND THE INFLUENCES OF VOLUNTEERING ON WELL-BEING
Hyunkee Lee 1 , 1. Korean Gerontological Society, Suwon, South Korea, Korea, Republic of The objectives of this paper are to examine volunteer activity changes over the life courses and the long-term influences of volunteer activity on the well-being. The paper analyzes the KLoSA data from wave 1 to wave 5, and selects 10,254 persons over 45 years old as study samples in the base year. For the statistical analysis, the two-way connected line plot and transition probability analysis techniques are employed and the GEE method is used for the multivariate regression estimation of the coefficients. The results show that the volunteer activities increase highest in one's late 40s and decrease abruptly from the age, and slow down 50s to one's late 70s, showing a repeating rise and fall pattern. Entering one's 80s, the volunteer activities come to be stable relatively than the previous years. And, the probability of continuing the volunteer status after 2 years is about 31.2 percent and the probability of stopping volunteer activity comes to be around 68.9 percent. As a last, the regression analysis results show that there is a positive relationship between volunteerism and well-being variables. Expecially the interaction term is statistically significant, showing a negative sign of the coefficient. This implies that the volunteerism contributes to increasing the well-being but decreasing by engaging excessively. The results support the previous study outcomes that excessive volunteer activities are not good for the health.

TYPES OF SOCIAL NETWORKS AND THEIR ASSOCIATION WITH MOBILITY AND DISABILITY IN LATE LIFE
Talha Ali, 1 Michael Elliott, 1 Toni C. Antonucci, 1 Belinda Needham, 1 Jonathan Zelner, 1 Carlos Mendes de Leon 1 , 1. University of Michigan, Ann Arbor, Michigan, United States Social networks are critical in maintaining late-life functional health, but, previous studies have focused on isolated dimensions of social networks. We examined whether network types, representing multiple interrelated network characteristics, are associated with mobility and disability