Longitudinal relationships among loneliness, social anxiety, self-esteem and self-concept in rural left-behind children: a cross-lagged study

Background: Loneliness is one of the most common mental health problems among Rural Left-behind Children (RLBC) in China. Social anxiety, self-esteem and self-concept are primary correlates that are linked with loneliness, yet there is limited evidence of causal or temporal associations between these variables in RLBC. Therefore, this study explores the gender differences in RLBCs loneliness, social anxiety, self-esteem and self-concept, and aims to clarify the longitudinal cross-lagged relationship among the four. Method: Using Children’s Loneliness Scale (CLS), Social Anxiety Scale for Children (SASC), Rosenberg’s Self-Esteem Scale (SES) and Piers-Harris Child Self-concept Scale (PHCSS) to assess the 812 RLBC. All assessments of CLS, SASC, SES, and PHCSS were measured in identical conditions at schools at the baseline (T1) and the end of the 1 year (T2). Descriptive statistics and Mann-Whitney U test were used for gender differences in variables. Partial correlation analysis of controlling gender was used to investigate the internal relationship of variables. A cross-lagged analysis were used to examine the longitudinal relationship between these variables. Result: The result show that gender differences between CLS and SES were not statistically significant ( P >0.05), but the gender difference between SASC and PHCSS was statistically significant ( P <0.05). Additionally, both simultaneous and sequential correlations between CLS, SASC, SES and PHCSS were significant (all P <0.001). Lastly, CLS ( β= 0.491, P< 0.001), SASC ( β= 0.229, P< 0.01), SES ( β= -0.140, P< 0.05) and PHCSS ( β= 0.121, P< 0.001) at T1 significantly predicted CLS at T2. Conclusion: There were gender differences in social anxiety and self-concept of RLBC. Meanwhile, social anxiety, self-esteem and self-concept have unidirectional forecasting effect on loneliness.