Validation of the 6-Item De Jong Gierveld Loneliness Scale in Korean Young Adults

Background and Purpose: Loneliness is generally conceptualized as the subjective perception of social isolation and feelings of lack of connection to others. Loneliness is closely linked to social isolation which represents the lack/limited contact with others and many deleterious outcomes of physical and mental health. This is of specific concern for young adults in South Korea, where increased suicide and lonely death rates are an increasing concern. Prior research suggests loneliness as a multidimensional construct. Specifically, the De Jong-Gierveld Loneliness Scale (DJGLS) is constructed to assess two latent factors, social and emotional loneliness. While the scale is one of the most popular instruments with the UCLA Loneliness scale, its psychometric properties have not been tested in Korean contexts. The current study aimed to test the qualities of the 6-item DJGLS with Korean young adults.

Methods: This study employed survey data collected in 2021 from a national sample of college students in South Korea (N=1,067). The study included a balanced ratio of the sample in sex (female vs. male) and geographic locations (Seoul vs. other cities). The sample was 21.8 years old on average. The short version of the DJGLS consisted of six items, three items respectively for emotional loneliness and social loneliness (1=none of the time to 5=all of the time). We tested (1) confirmatory factor analyses (CFA) to determine the factor structure of the DJGLS, (2) internal consistency reliability, and (3) correlations with self-rated general health (1=poor to 5=excellent) and psychological distress (Patient Health Questionnaire-4) to assess the concurrent validity.

Results: The mean scores were 2.56 for items in emotional loneliness and 2.66 for items in social loneliness. In agreement with the previous literature, we identified a two-factor CFA model with six items (emotional and social loneliness). The model fits were acceptable, except RMSEA (χ²(8) = 273.035, p < .001; RMSEA 90% CI: 0.158~0.193; CFI = .985; TLI = .973, SRMR = 0.048). We found one item in emotional loneliness (i.e., I miss having people around me) cross-loaded on the factor of social loneliness. We removed the item, following the literature guidelines and substantial interpretation. The revised model with five items has an excellent model fit: (χ²(4) = 9.465, p = .051; RMSEA 90% CI: 0.000~0.066; CFI = 1.000; TLI = 0.999; SRMR = 0.007). All factor loadings were over 0.73 and statistically significant (p < .001). Internal consistency reliability was acceptable (emotional loneliness: α = .66; social loneliness: α = .91). Overall loneliness and two subscales of emotional and social loneliness were negatively associated with health and psychological distress.

Implications and Conclusion: The short version of the DJGLS is overall a reliable and valid scale to measure loneliness in Korean young adults. The study suggests an adapted application of removing one item for better estimation in Korea. We will discuss cultural differences and considerations to best inform social work practice and research in diverse contexts.