Performance of an Abbreviated Lubben Social Network Scale Among Three Ethnic Groups of Older Asian Americans

Background and Purpose: One of the most frequently used instruments to gauge social networks among older adult populations is the Lubben Social Network Scale (LSNS). With its brevity, easy administration, psychometric qualities, and ability to combine or discern social ties to family and friends, the LSNS−6 has been widely used in gerontological research and geriatric care settings. Translated into many languages, the LSNS−6 has been applied to older adults with diverse racial/ethnic backgrounds. However, simultaneous cross-group validation of the LSNS−6 has rarely been conducted. The present study examined the cross-validation of the LSNS−6 in measurement properties, factor structures, and correlations with selected social and health indicators in Chinese, Korean, and Vietnamese American samples aged 50 and above.

Methods: A total of 605 participants aged 50 or above (242 Chinese, 150 Koreans, and 213 Vietnamese) were selected from the 2015 Asian American Quality of Life (AAQoL) Survey, conducted with self-identified Asian Americans aged 18 or above living in Central Texas. Data on the LSNS−6 were analyzed on measurement qualities (internal consistency and corrected item-total correlation), dimensionality (exploratory and confirmatory factor analyses), and correlations with other indicators.

Results: The 6 items in the LSNS−6 showed high Cronbach’s alphas (0.82−0.87) in the three Asian American subgroups. High levels of internal consistency were also observed in the family and friend subscales, but the alpha values in the friend subscale (0.84−0.91) were consistently higher than those of the family subscale (0.71−0.80) across all ethnic groups. The results from the corrected item-total correlations also demonstrated homogeneity within the overall and sub-scales, with all coefficients exceeding the criterion of 0.30. A higher level of homogeneity was manifested in the friend subscale (0.65−0.84) than the family subscale (0.43−0.68). It is noteworthy that the first item in the family subscale (the number of family in frequent contacts) demonstrated substantially low coefficients in all ethnic groups. Confirming the underlying construct of the LSNS−6, the exploratory factor analysis yielded the two-factor solution, representing family and friends, in all three groups. Furthermore, the confirmatory factor analysis showed that the factor structure was invariant across the groups. Both subscale and total scores of the LSNS-6 were associated in expected directions with the social and health indicators considered.

Conclusions and Implications: The findings not only confirmed the measurement qualities of the LSNS-6 within each group but also provided support for measurement invariance across the groups. The finding on differences in family and friend networks requires cultural considerations and warrants further investigations.