Long-run results of mainstreaming students with special educational needs in lower secondary school

ABSTRACTPolicies aiming to promote the mainstreaming of students with special educational needs in regular classrooms have become a focal point of political discussions in many countries. However, there is a scarcity of quantitative evidence with robust empirical designs that can shed light on the long-term educational outcomes associated with mainstream education. This study utilizes extensive longitudinal register data and employs a propensity score matching framework to investigate the disparities in upper secondary education completion between students educated in regular and segregated learning environments. The findings reveal that the likelihood of attaining an upper secondary certificate is significantly higher (by 18 percentage points) among students in regular classrooms. Furthermore, tentative results indicate that higher passing rates at the lower secondary school leaving exam may play a mediating role.KEYWORDS: Mainstreamingspecial educationpropensity score matchingeducational attainmentSEN Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 Downloaded on January 18 2023 at https://www.udir.no/tall-og-forskning/statistikk/statistikk-grunnskole/tall-om-elever-og-skoler/2 Data on SEN students in regular classes in only available for the schoolyear 2011/12. The analysis focusses on the year 9 cohort of 2011/12 (rather than on years below year 9), because this cohort can be followed for the longest period in their post-school career. This is important as SEN students are often slower to complete upper secondary education than regular students because they often take one or several preparatory years before enrolling in youth education programs.3 Until 2018/19, the last year of data by the time of analysis.4 Children and their parents are linked using the variables named MOR_ID and FAR_ID in the Population register (BEF).5 Yet, matching methods typically estimate the propensity score between treatment and covariates using parametric methods.6 The matching estimations were conducted using the KMATCH procedure in STATA 15. 95% confidence bands are obtained by bootstrapping the entire estimation process. Note that only regular class students on the “common support” are included in the calculations of the estimate, meaning that regular class students whose propensity scores are outside the range of propensity scores among special class/school students are excluded from the analysis.7 In a related study by Hienonen et al. (Citation2021), students in special classes demonstrated higher GPAs in year 9 compared to SEN students in regular classes. Although these results may initially appear contradictory to the findings of the present study, it is important to consider the differences in country contexts and outcome measures. Hienonen et al. assessed GPA based on teacher evaluations, while the present study focused on the outcome measure of passing the national standardized school leaving examination.