Special education: Effectiveness and selection processes (SPEDU-EFFECT)

Currently 18% of teaching hours in grades 1-10 in Norwegian school are spent on special education. Does it work? Worldwide, research examining the effectiveness of special education as typically provided to all eligible students, is almost non-existing. B ecause students eligible to special education are entitled to get it, experiments are not possible. Therefore, observational studies have to be used. However, because there may be a range of unmeasured factors influencing who receivspecial education or not the results from such observational studies have been uncertain. Recently, statistical techniques have been developed which reduce these uncertainties. A few studies from the U.S. using such techniques have provided mixed results, but they may have little relevance to the Norwegian setting. One recently published study from Norway (Lekhal, 2018), however, adjusting for measured confounders and confounders common to students in the class or at the school (e.g. school climate, school specific teaching practices) suggested detrimental effects of special education on reading, but to no other effects. Of note, unmeasured student or family specific confounders (e.g. genetics) might still explain findings, or suppress actual associations. Using data from the Trondheim Early Secure Study (TESS), we will be able to provide answers to the question of effectiveness as this is long-term prospective study following students from 4 years of age to 5th grade. By using detailed information on a wide range of possible correlates of special education, we will be able to detect effects of special education on a range of outcomes including academic achievement and social/cognitive functioning. Almost 70% of special education recipients are boys, and variations according to school size and locales have also been noted. This raises the question "Are some students selected for special education more/less often due reasons outside the formal decision process?" (e.g. parental background, social skills, school factors, the student-teacher relationship). Because of the detailed information in TESS we will be able to address this question. With respect to gender and SES, our results show that the higher use of special education in 5th grade among boys and those with low SES is fully explained by the higher rates of ADHD, poor test results in math and reading, and increase perceived teacher helplessness when teaching the child, all in 3rd grade. Thus, there seem not be any cultural effects in this locale and in this age group. As regards the effect of special education, we examined the combined impact of teacher-reported special needs education and other special education efforts in grades1 and 3 on test performance in reading, writing, and math in grades 3 and 5, as well potential influence on task motivation for these three subjects, respectively. We did so by adjusting for all measured confounderes concerning selection into special education (above) as well as all unmeasured time-invariant confounders between performance, motivation and special education. The results revealed a small negative effect of special education in 3rd grade on match test scores in 5th grades. Otherwise, no effects of special education, for better or worse, were detected.

Siden hovedfunnene ble publisert i oktober 2018, forventes ingen effekter på de få ukene som har gått

Currently 18% of teaching hours in grades 1-10 are spent on special education. The question then arises: 'Does it work?' Research addressing this question (i.e. the effectiveness of special education as typically provided to all eligible students) is almo st non-existent. Because experimental techniques are not possible, observational studies have to be used. Recently, statistical techniques have been developed which allow for quasi-experimental analyses of observational data. A few studies from the U.S. u sing such techniques have provided mixed results, but they may have little relevance to the Norwegian setting. Using data from the Trondheim Early Secure Study, and extending it, we will evaluate effectiveness of special education using this long-term p rospective study incorporating bi-annual assessments of children from age 4 to 9th grade. By using detailed information on a wide range of possible correlates of special education (e.g. academic difficulties, test-scores, behavioral and attentional proble ms, cognitive and social functioning, and parental factors) we will be able to detect effects of special education on a range of important developmental outcomes. Further, we will investigate whether special education is effective for some individuals m ore than others due to student characteristics (e.g. type of problems), timing of intervention (e.g. early vs. later), and teacher, school or special education characteristics (e.g. delivered by teachers or assistants). Almost 70% of special education re cipients are boys, and variations in special education utilization related to school size and locale have also been noted. This raises the question 'Are some students selected for special education more/less often for reasons outside the formal decision p rocess?' (e.g. parental background, social skills, school factors, the student-teacher relationship). Because of the detailed information on these matters gathered over many years we will be able to address this