Recent studies have focused on the negative effects of loneliness on health and quality of life in older adults. The De Jong Gierveld Loneliness Scale (DJGLS) has been widely used and has proven to be a valid and reliable instrument for loneliness assessment. However, research on this topic and on the validation of measurement scales among the older population is still incipient. The objective of this study was to examine the psychometric properties of the Spanish version of the 11-item DJGLS in Mexican older adults. Data from a representative sample of cognitively intact older adults aged 60 years and over (mean, standard deviation [SD]) age = 72.0 years (SD 8.1) from two Mexican cities (n = 1913), interviewed face to face at their homes during 2018-2019, were analyzed. The psychometric properties of the DJGLS were examined, including (1) construct validity, examined by Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), discriminate validity and convergent validity, (2) reliability, calculated using Cronbach's alpha. The overall data quality was high, and the scaling assumptions were generally met with few exceptions. Using EFA and CFA, the findings showed that the DJGLS presents a two-factor structure (Social Loneliness and Emotional Loneliness), with 11 items that explain 67.2% of the total variance. Reliability is adequate at the full-scale level (Cronbach´s α = 0.899), also for the two subscales Social and Emotional Loneliness (Cronbach´s α = 0.892 and 0.776, respectively). These results highlight that most participants with a low score for depressive symptoms and or with a high social support score belonged to the "No loneliness" group. The results showed that the Spanish version of the 11-item DJGLS is adequate for use in Mexican older adults and should be used not only for loneliness screening but also for social and emotional loneliness assessment.
Keywords: Mexico; loneliness; older adults; psychometric properties; scale.