Feedback to students on how they progress in school has been shown to affect further learning (Hattie, 2009), and very much so if the feedback information is used formatively (Hattie & Timperley, 2007). Less empirical work has been done on the effect of summative assessments, such as grades, on students’ learning. Grades have several functions and one is that they are assumed to motivate students to learn and achieve in school. However, the assumed positive effects of grading is far from clear and international research shows disparate results (Harlen & Deakin Crick, 2002; Klapp et al., 2014; Azmat & Iriberri, 2009). This is primarily due to lack of empirical studies within the field.

Previous research has identified a negative grading effect for low-achieving 7-9th grade students (13-15 years old) whereby individuals who previously received grades (in 6th grade) showed poorer grade development in subsequent grades compared to similarly capable peers who did not previously receive grades (Klapp, 2015; Klapp, 2018; Klapp, Cliffordson, & Gustafsson, 2016). A gender effect was found: graded girls tended to perform better than ungraded girls, and both graded and ungraded boys. However, these findings were obtained using data from 1980 following a survey of students born in 1967 birth cohort. As there have been several reforms and changes to the Swedish education system since then, it is important to re-evaluate whether these findings are still relevant to the current education system. The change from norm-relative to criterion-referenced assessment, the strengthened eligibility requirements as well as the revised grading scale could affect the impact that grades have on subsequent achievement and how underlying factors like self-perception, competition, and perception of grades influence performance.

In the beginning of the 1990s the norm-relative grading system was replaced by a criterion-referenced grading system with national goals, performance standards and a grading scale with a fail-step (F). In 2011, the criterion-referenced grading system was reformed, and a stricter grading regime was introduced. National tests were introduced in 3^rd Grade, grades in all school subjects were introduced in 6^th Grade and in all subsequent Grades, the performance standards had to be fully met and no compensatory grading was allowed. Besides, in order to continue from compulsory school to upper secondary school, students need to achieve at least eight passing grades to enter a vocational program and at least 12 passing grades to enter a theoretical program at upper secondary education (before the 2011 reform, passing grades were required in three subjects Swedish, English and mathematics). Thus, the present criterion-referenced system has a stricter grading regime and higher demands for passing grades, compared to the previous norm-relative grading system, which makes the failing more severe for students in the present system. Therefore, it is of high relevance to investigate the effect of grading on later achievement in the criterion-referenced grading system. The cohorts born before the reform (born 1998 and earlier) were not subject to the stricter grading regime and did not receive national tests and grades until 8^th Grade, while the cohorts born 1999 and onwards received the stricter grading regime. This circumstance makes it possible to apply a quasi-experimental design in order to investigate how grading affects students´ later achievement. 

Purpose

By comparing cohorts of students before the reform with students after the reform it is possible to investigate effects of grading on later achievement. The following research questions are in focus:

How does previously receiving grades affect subsequent performance in the criterion-referenced grading system?

Are subgroups differentially affected by being graded or not graded?

How does previously receiving grades affect subsequent performance in different school subjects?

Data from the Evaluation Through Follow-up (UGU) infrastructure before and after the 2011 reform were used to compare the grade progression of individuals who have previously received grades to those who have not previously received grades. The cohorts relevant to the present study are the cohorts born in 1998 (N=9176) and 2004 (N=9437), which were in Grade 6 in the academic years 2010/11 and 2016/17 respectively. Birth cohort 1998 was not previously graded prior to Grade 8 whereas birth cohort 2004 previously received grades in Grade 6. These cohorts can therefore be compared to investigate whether prior receipt of grades has any discernible impact on subsequent academic performance. These cohorts are also well-placed on either side of the academic year 2012/13 in which the reforms were first implemented and allows for the fluctuating effects of novelty and uncertainty of the new system to stabilized. Initially, descriptive statistics were estimated. Then, several multivariate multiple regression models were estimated in several steps. First, a basic model was estimated using independent variables graded/not graded status, cognitive ability measures, gender, and parental education level as a measure of socio-economic status (SES) and 9th Grade GPA as dependent variable. Then cross product terms were included to investigate interaction effects between the variables, thus fitting a saturated model. Finally, a model with only the significant interaction effects was estimated. Further, the modelling continued by estimating the same models but with separate school subjects used as dependent variable in order to investigate the grading effect on students´ performances in different school subjects. Missing information in the variables was handled by using the missing data modeling in the Mplus statistics program (Muthén, Kaplan & Hollis, 1987). To take account of possible clustering of students in schools (school level), the complex option offered by the Mplus program was used. Cohen´s d was computed as a measure of effect size (Ellis, 2010). The analyses were conducted in the SPSS program, version 28 and in the Mplus program, version 8.5 (Muthén & Muthén, 1998-2019).

The analysis suggests that low-ability students, and boys in particular, seem to be especially affected by grading. The effect size for the difference between ungraded and graded students on GPA 9 were .22, stronger for students receiving grades in 6th Grade. Analyses of the effect sizes are underway concerning differences between low-ability and high-ability students in the ungraded and graded cohorts and their performance on GPA 9. The basic, saturated, and final models all yielded highly significant results (all RMSEA < .001, all CFI/TLI = 1.000, all SRMR < .001). When the cohort differences are examined by gender, little difference is observed between graded and ungraded girls, but boys show a noticeable interaction between GPA and graded status. The relationship between cognitive ability and 9th Grade GPA is stronger for the graded cohort than for the ungraded cohort. When the genders are compared by grading status, the data show that the relationship between boys’ cognitive ability and 9th Grade GPA is much more consistent with that of girls in the graded cohort, but with a noticeable divergence between the genders in the ungraded cohort. Further analyses on the effect of grading on students´ performances in the different school subject are currently underway during spring 2022. Differences may exist concerning the relations between low- and high ability students in the ungraded and graded cohorts and the performances in different school subjects such as mathematics and arts. The study is a part of the research project “Student self-concept and school achievement: bidirectional relations and effects of social comparisons and grading” (2019-04531), funded by the Swedish Research Council.

Azmat, G., & Iriberri, N. (2009). The importance of relative performance feedback information: Evidence from a natural experiment using high school students. CEP Discussion Paper No. 915, Centre for Economic Performance, London School of Economics and Political Science. Ellis, P. D. (2010). The essential guide to effect sizes. New York, NY: Cambrideg University Press. Harlen, W & Deakin Crick, R. (2002). A systematic review of the impact of summative assessment and tests on students´ motivation for learning (EPPI-Centre Review, version1.1*). Research Evidence in Educational Library. Issue 1. London: EPPI-Centre, Social Science Research Unit, Institute of Education. Hattie, J.A.C. (2009). Visible learning: A synthesis of 800+ meta-analyses on achievement. London: Routledge. Hattie, J. A. C., & Timperley, H. (2007). The power of feed-back. Review of Educational Research, 77, 81-112. Klapp A. (2018). Does academic and social self-concept and motivation explain the effect of grading on students´ achievement? European Journal of Psychology of Education, 33(2), 355-376. DOI: 10.1007/s10212-017-0331-3 Klapp, A. (2015). Does grading affect educational attainment? A longitudinal study. Assessment in Education: Principles, Policy and Practice, 22(3), 302-323. Klapp, A., Cliffordson, C., & Gustafsson, J-E. (2014). The effect of being graded on later achievement: evidence from 13-year olds in Swedish compulsory school. Educational Psychology: An International Journal of Experimental Educational Psychology, 36(10), 1771-1789. DOI: 10.1080/01443410.2014.933176. Muthén, B., Kaplan, D., & Hollis, M. (1987). On structural equation modelling with data that are not missing completely at random. Psychometrica, 52, 431-462. Muthén B., & Muthén, L. (1998-2019). Mplus user´s guide (8th ed.) Los Angeles, CA: Author.