Background: Inclusive education plays a critical role in meeting the learning needs of all students, and preventing social exclusion. However, in many developing regions, its implementation is still insufficient in both width and depth. One of the critical barriers to the implementation is the parental and public worry that including students with special educational needs (SEN) would infringe upon the development of students without SEN. Although studies in developed regions observed neutral or even beneficial inclusion effects on students without SEN, these findings were less persuasive to raise social support in the developing regions, where poor materials and insufficient trained teachers are provided for inclusion. A further worry is that the neutral results can be attributed to the positive effects for some students and negative effects for others. However, few studies explore the distinctive inclusion effects due to individual features.

Objective: This study aimed to examine the association of inclusive education with academic performance (math, reading, arts) and social-emotional skills (15 skills from OECD 2021) of students without SEN. It further aimed to use person-oriented approaches to explore the distinctive inclusion effects due to types of SEN students in class, as well as gender, socioeconomic status and development level of students without SEN.

Method: 1155 10-year-old Chinese students without SEN from the OECD “Survey on Social-Emotional Skills” were selected using the Propensity Score Matching method (579 from non-inclusive class, 576 from inclusive class). 2-level Hierarchical Linear Modeling (student nested in school) was employed to accommodate the structure of the SSES database. The model was estimated by Bayes estimator, default priors, and probit link. Quantile regression was used to analyze the inclusion effects in students’ different development levels.

Result: The analysis controlled students’ gender and socioeconomic status at the individual level, and average class size, rate of disadvantaged students and students with special needs at the school level. Results illustrated that (1) inclusive education had insignificant relations with academic performance, and had relatively slight but significant positive associations with students’ optimism, tolerance and sociability (β=0.07-0.08); (2) socioeconomic status positively moderated the associations of inclusive education with trust, tolerance and assertiveness, in which inclusion might have greater social-emotional benefits to those from higher socioeconomic status families; (3) inclusive education had a tighter linkage with optimism for students having a lower level of optimism (last 25-50%).

Conclusion: In line with studies in developed regions, this study indicated that in developing regions, inclusive education might not hinder academic achievement, and slightly benefit some social-emotional skills of students without SEN, regardless of inadequate experience, trained teachers and support. This result is stimulating for implementing inclusive education, despite the small effect size. Interacting with SEN students helps students without SEN better comprehend and accept the differences while reduce their prejudices. As a consequence, their tolerance and sociability are enhanced. Furthermore, they tend to feel better about themselves and develop stronger optimism through the comparison with SEN students in class.

However, some benefits of inclusive education are more targeted at the students from higher socioeconomic status. Their parents are more knowledgeable about inclusive education and show greater acceptance and support, deeply influencing their attitudes and behaviors toward SEN students. Additionally, the inclusion effect on social-emotional skills may differ due to the development level of specific skills, which needs further investigation.

Findings help holistically comprehend the role of inclusive education in 15 specific social-emotional skills, and its individual differences, and imply the potentially crucial individual factors for further studies. Findings also contribute to relieving parental and public worries, while raising more support for implementing inclusive education.

Sample The sample came from the Survey on Social-Emotional Skills (SSES) database from OECD (2021). 10-year-old students from Chinese Suzhou were chosen and the 1:1 nearest neighbor and 0.01 caliper of Propensity Score Matching (PSM) without replacement was employed to reduce the selection bias and difference in sample size (Rosenbaum & Rubin, 1985). After controlling all 5 covariates, 579 students (49% boys and 51% girls) from non-inclusive classes and 576 students (43% boys and 57% girls) from inclusive classes were selected (see Appendix-A). Measurement Academic Outcomes compromised math, reading/language, and arts grades from the school registry exams. Social-Emotional Skills were assessed as 15 facets of the Big Five and reported by students (see details in Appendix-B). All facets were measured by 3 items with 5-point Likert scales, and showed good reliability (OECD, 2021). Inclusive Education was based on whether there were one or more students with special needs in class (0=non-inclusive class, 1=inclusive class). Special needs, reported by parents, were divided into 3 types of difficulties: (1) Physical and Sensory (i.e., hearing, vision, mobility), (2) Learning, (3) Behavior, (4) Social-Emotional. Socioeconomic Status (SES) was composed of home possessions, parents’ occupational status and educational level, similar to PISA (OECD, 2019). Gender was reported by students (0=boy, 1=girl). Average Class Size was reported by school principals, ranging from “1 = 15 or fewer” to “9 = more than 50”. The Rates of Disadvantaged Students and Students with Special Needs were reported by school principals (from 1 = less than 5% to 5 = more than 50%). Analysis Two-level hierarchical linear modeling (i.e., student-school) with Bayes estimator was conducted by Mplus 8.3 after the PSM sample selection. Three models were estimated: null model, covariates-only model, and random intercept model. The first two models were to estimate the variances of academic and social-emotional outcomes at the individual level, and the variances explained by covariates (R2). The last model estimated the effect and △R2 of inclusive education on academic and social-emotional outcomes, and explored the interaction of “inclusive×SES” and “inclusive×gender”. Continuous variables were grand mean centered, and binary variables were centered by effect coding (e.g., -0.5=boy, 0.5=girl). All models were converged and showed good model fits (PPP=0.42-0.52, the symmetry axis of Δχ2 95% PPI was close to 0; Gelman, 2014). In addition, quantile regression was finally employed to examine the inclusion effect in students’ different development levels.

Through the results (see Appendix-C), we found slight but significant positive effects of inclusive education on optimism, tolerance and sociability (i.e., β=0.07-0.08, △R²≤0.01; Cohen, 1992). We also observed insignificant academic differences between inclusive and non-inclusive classes. Despite the limitation of the small effect, this discovery is stimulating for implementing inclusive education. It was consistent with studies in developed countries (e.g., Ruijs et al., 2010; Szumski et al., 2017), and further confirmed the benefits of inclusion to students without SEN from the regions lacking inclusion experience, supports and trained teachers. Two mechanisms can explain the social-emotional benefits of inclusion. Interacting with SEN students, students without SEN can better comprehend and accept interpersonal differences, and develop fewer prejudices while more friendliness (Keith et al., 2015). Thus, their tolerance and sociability are enhanced. Furthermore, comparing themselves with SEN students, those without SEN tend to feel better about themselves (Ruijs et al., 2010; Marsh et al., 2019), fueling their optimism, and some students’ assertiveness. Apart from general neutral or positive associations, some social-emotional benefits were more targeted at higher SES students (i.e., trust, tolerance and assertiveness). With better understanding of inclusion and its potential benefits, higher SES parents show greater acceptance, and encourage their students to actively interact with SEN students (de Boer et al., 2010; Leyse & Kirk, 2004). Additionally, a stronger linkage between inclusion and optimism was observed for those having a lower level of optimism (last 25-50%), which requires further investigation. In conclusion, this study holistically examined the role of inclusive education in 15 social-emotional skills of students without SEN and explored critical individual differences, which helps further understand the inclusion impacts and implies future research direction. Additionally, findings could relieve parental and public worries, and raise more supports for inclusive education in China and other developing regions.

Cohen, J. (1992). Quantitative methods in psychology: A power primer. Psychological Bulletin, 112, 115-159. de Boer, A., Pijl, S. J., & Minnaert, A. (2012). Students’ attitudes towards peers with disabilities: A review of the literature. International Journal of Disability, Development and Education, 59(4), 379-392. https://doi.org/10.1080/1034912X.2012.723944 Gelman, A., Carlin, J. B., Stern, H. S., & Rubin, D. B. (2014). Bayesian data analysis (3rd edition). Chapman and Hall/CRC. Keith, J. M., Bennetto, L., & Rogge, R. D. (2015). The relationship between contact and attitudes: Reducing prejudice toward individuals with intellectual and developmental disabilities. Research in Developmental Disabilities, 47, 14-26. https://doi.org/10.1016/j.ridd.2015.07.032 Leyser, Y., & Kirk, R. (2004). Evaluating inclusion: An examination of parent views and factors influencing their perspectives. International Journal of Disability, Development and Education, 51(3), 271-285. https://doi.org/10.1080/1034912042000259233 Marsh, H. W., Parker, P. D., & Pekrun, R. (2019). Three paradoxical effects on academic self-concept across countries, schools, and students. European Psychologist, 24(3), 231-242. https://doi.org/10.1027/1016-9040/a000332 OECD (2019). PISA 2018 Assessment and Analytical Framework. Paris: PISA,OECD Publishing. OECD. (2021). OECD Survey on Social and Emotional Skills Technical Report. https://www.oecd.org/education/ceri/social-emotional-skills-study/sses-technical-report.pdf Rosenbaum, P. R., & Rubin, D. B. (1985). Constructing a control group using multivariate matched sampling methods that “incorporate the propensity score”. The American Statistician, 39, 33-38. https://doi.org/10.1080/00031305.1985.10479383 Ruijs, N. M., Van der Veen, I., & Peetsma, T. T. (2010). Inclusive education and students without special educational needs. Educational Research, 52(4), 351-390. https://doi.org/10.1080/00131881.2010.524749 Szumski, G., Smogorzewska, J., & Karwowski, M. (2017). Academic achievement of students without special educational needs in inclusive classrooms: A meta-analysis. Educational Research Review, 21, 33-54. https://doi.org/10.1016/j.edurev.2017.02.004