The research team, led by Hongyu An of Langfang Normal University together with colleagues from institutions in Croatia, Kazakhstan, and China, framed the question around a distinction that earlier studies often blurred. Psychological well-being, in the tradition of Ryff's six-dimensional model, captures positive functioning across self-acceptance, positive relations, autonomy, environmental mastery, purpose in life, and personal growth. Loneliness, by contrast, measures a perceived gap between the social relationships a person wants and the relationships they actually have.

Treating low loneliness as simply the mirror image of good well-being, the authors argue, risks conflating two psychologically distinct states that may carry different weights in shaping how students experience university life. The theoretical scaffolding combines social cognitive theory and self-determination theory. Self-determination theory holds that satisfying three basic needs, autonomy, competence, and relatedness, fuels autonomous motivation and positive functioning.

Together, the two frameworks predict that socially skilled students should report higher well-being and lower loneliness, and that both states should, in turn, relate to how deeply students embed themselves in academic work, optional activities, and course commitment. To measure the outcome, the team used the Academic Life Assessment Scale, or ALAS, which captures perceived academic ability, involvement in non-compulsory activities, and commitment to the course. Social skills were assessed with a 23-item instrument covering affiliative skills, leadership, team integration, interpersonal skills, and social engagement.

Loneliness was measured with the 20-item UCLA Loneliness Scale, and psychological well-being with an 18-item adaptation of Ryff's scale. The questionnaire was translated into Chinese and back-translated to preserve conceptual equivalence, then administered electronically via WeChat during scheduled class sessions. Of 275 collected questionnaires, 24 were incomplete, leaving 251 complete responses with no missing-value imputation required.

The sample was heavily skewed in ways the authors flag explicitly. Women made up 76.9 percent of respondents, 92.0 percent were aged 18 to 21, and 97.2 percent were first-year students. Arts and humanities dominated the study areas at 57.8 percent, followed by engineering at 31.1 percent.

Most students reported strong grades, with 57.4 percent reporting a GPA between 3.4 and 4.0. Because the sample is a non-probability convenience draw from a single institution, the authors caution that generalization beyond this specific student group should be made with care. The primary structural model told a consistent story for most hypothesized links.

Social skills were negatively associated with loneliness, with a standardized coefficient of minus 0.349, and positively associated with psychological well-being at 0.621, both highly significant. Social skills also showed a direct positive association with academic life integration at 0.401, and psychological well-being was positively associated with the outcome at 0.284. The conditional path from loneliness to academic integration, however, fell short of significance at minus 0.081 with a p-value of 0.093, even though loneliness correlated negatively with the outcome at the construct level.

In other words, once social skills and well-being were held constant, loneliness added little unique explanatory weight. The mediation results sharpened the picture. The indirect association running from social skills through psychological well-being to academic life integration was significant, with a standardized estimate of 0.176 and a 95 percent bootstrap confidence interval from 0.099 to 0.271.

The parallel indirect association through loneliness was not, at 0.028 with a confidence interval spanning minus 0.006 to 0.060 and a p-value of 0.090. The total indirect effect was 0.205, and the total association between social skills and academic integration reached 0.606. The primary model explained 43.0 percent of the variance in academic life integration, 38.6 percent of the variance in well-being, but only 12.2 percent of the variance in loneliness, indicating that social skills account for just one slice of the factors shaping students' feelings of disconnection.

Because the six-dimensional well-being construct showed weak convergent validity, with an average variance extracted of only 0.444, a purpose-in-life loading that was effectively zero, and a weak positive-relations loading, the team ran sensitivity analyses. Dropping purpose in life produced a five-dimensional model, and dropping positive relations as well produced a four-dimensional model with stronger measurement properties, including an average variance extracted of 0.648. The well-being pathway remained significant across all three specifications, at 0.176, 0.174, and 0.145 respectively.

The loneliness pathway, however, turned significant only in the four-dimensional model, at 0.037, alongside a significant loneliness-to-integration path of minus 0.107. The authors interpret this as specification uncertainty rather than evidence for or against a loneliness mechanism, noting that the improved statistics concern the narrower retained construct and do not validate the full 18-item instrument. Interpretively, the researchers argue that the significant well-being pathway fits both theories: socially skilled students may draw on peer relationships, mentorship, and institutional support to build resilience, self-efficacy, and a sense of agency, which in turn aligns with stronger perceived academic ability and course commitment.

The direct social-skills effect that survived the inclusion of both mediators suggests that communication, cooperation, and help-seeking may also support participation in university life through channels not captured by the two mediators. The authors stress that these are theoretical interpretations rather than separately tested mechanisms, and that the cross-sectional, single-source design cannot establish causal direction, rule out reciprocal associations, or exclude common method bias, though procedural safeguards and full-collinearity variance inflation factors between 1.212 and 1.734 offered some reassurance. The study's limitations are stated plainly.

Global model fit was limited, with standardized root mean square residual values of 0.085 and 0.092 that exceeded bootstrap reference thresholds, and the very high reliability of the loneliness scale, with a Cronbach's alpha of 0.966, raises the possibility of item redundancy. The item screening and dimensional reductions require independent validation, and the shifting support for the loneliness path across specifications signals model-specification uncertainty. The authors call for longitudinal designs, independent measurement validation, and more diverse institutions and learning settings.

For practitioners, they suggest that structured peer tasks with explicit communication and help-seeking practices, combined with first-year mentoring and accessible well-being support, are worth piloting, provided such initiatives are evaluated with baseline and follow-up assessment rather than assuming that more social contact alone produces better academic integration.

Subject of Research: Associations among social skills, psychological well-being, loneliness, and academic life integration in university students

Article Title: How social skills relate to academic life integration: psychological well-being and loneliness

Article References: An, H., Obrenovic, B., Du, G., Dzhanegizova, A., Godinić, D., & Gu, X. (2026). How social skills relate to academic life integration: psychological well-being and loneliness. Frontiers in Psychology, 17, Article 1891012. https://doi.org/10.3389/fpsyg.2026.1891012

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1891012

Keywords: social skills, psychological well-being, loneliness, academic life integration, university students, PLS-SEM, cross-sectional study, self-determination theory, social cognitive theory, student engagement, higher education, mediation analysis