How do personal networks evolve as individuals age? To what degree do social disadvantage and contextual factors matter for network dynamics in later life? This paper answers these two questions based on egocentric network data of older adults over a ten-year period. Specifically, I use longitudinal and nationally representative data on 1,168 older adults from the National Social Life, Health, and Aging Project. I use between-within models to separate the within- and between-individual effects of sociodemographic characteristics and contextual factors on three aspects of social connectedness in later life: network size, frequency of contact, and proportion of kin. Patterns of network change vary among people of different races and ethnicities as well as educational levels. Black and Hispanic respondents have a significantly smaller network size and a higher average frequency of contact with confidants. Moreover, Hispanic respondents have a higher proportion of kin in the network, compared to White respondents. Similarly, older adults with less education have a smaller network size, higher frequency of contact and higher proportion of kin in their confidant networks compared to those who attended college. Older adults who have better mental health are more likely to have a higher frequency of contact and higher proportion of kin. When an older adult starts to work for pay, their frequency of contact with confidants tends to increase. Older adults living in neighborhoods with stronger social ties are more likely to have a larger network size, higher frequency of contact, and lower proportion of kin in their confidant network. The above results show that disadvantaged backgrounds and contextual factors are associated with certain less favorable network characteristics, which helps to explain the concentration of social disadvantage on certain populations.

个人网络如何随着个人年龄的增长而发展？社会劣势和背景因素在多大程度上对晚年网络动态有影响？本文基于十年间老年人的以自我为中心的网络数据回答了这两个问题。具体来说，我使用了来自国家社会生活、健康和老龄化项目的 1,168 名老年人的纵向和全国代表性数据。我使用组间模型来区分社会人口特征和背景因素对晚年社会联系三个方面的个体内部和个体间影响：网络规模、联系频率和亲属比例。网络变化的模式因不同种族和民族以及教育水平的人而异。黑人和西班牙裔受访者的网络规模要小得多，但与知己接触的平均频率更高。此外，与白人受访者相比，西班牙裔受访者在网络中的亲属比例更高。同样，与上过大学的老年人相比，受教育程度较低的老年人网络规模较小，联系频率较高，其知己网络中的亲属比例较高。心理健康状况较好的老年人更有可能有更高的接触频率和更高的亲属比例。当一个年长的成年人开始为薪水工作时，他们与知己接触的频率往往会增加。生活在社会关系更强的社区的老年人更有可能拥有更大的网络规模、更高的接触频率、在他们的知己网络中，亲属的比例较低。上述结果表明，弱势背景和情境因素与某些不太有利的网络特征相关，这有助于解释社会弱势集中在某些人群身上。