Executive Summary
The results of the yearly end grade examinations in the country cannot easily be translated into statements about students' performance in relation to the curriculum. Very little is known about issues such as: What and how much do children learn in schools? Are their learning needs met in schools? What kind of variations are there and, if so, what can these variations be explained by? Do the worldwide acknowledged determinants on students' performance in schools produce similar results in Afghan cultural context? Do the inputs produce expected learning outcomes? What lessons can be drawn from existing experiences, e.g. of SCA experiences of Model schools? Etc.
This study, though small in size, can bee seen as an attempt towards varied aspects of assessment of learning achievements and as a general picture of students performance in primary level of schools The study focuses on language and math at primary school level of schools i.e. end of grades 3 and 6. A test based on national curriculum and textbooks (of language and math) for the related grades has been elaborated and used. Additionally, a simple questionnaire was used to collect data about the students, the school and the teachers. Data was collected from cold climate schools at the end of the school year (beginning of November). Totally 1175 students (21 % from government schools) of grade 3 taught by 39 teachers and 837 (15 % from government schools) of grade 6, taught by 64 teachers participated in the study.
The much dispersed student test results (ranging from 0 to 100) and the low average scores (average scores 52 of grade 3 and 53 of grade 6) of students indicate a relatively low level of achievements in primary level, which might be caused by, among others, short school year. Children of grade 3 were found to be weaker in math than in language (average scores 57 and 48 respectively), while this difference was found to be smaller in grade 6. However, the result of SCA supported schools were found to be remarkably better in both grade 3 and 6 compared to government schools (average scores 57 and 55 compared to 36 and 43 respectively). The difference might be explained by combined effects of slightly better teachers' working conditions and regular supervision of schools. No statistically significant differences, though girls' scored slightly better, have been observed between boys and girls.
No particular correlation between quality inputs such as class size, in-service teacher training, availability of classrooms, etc and students' test results been observed. Similarly, no pattern of special magnitude has been traced between teachers' education background and students' scores. These somehow surprising findings might be explained by the fact that actual teaching in classrooms do not differ between these categories of teachers and learning conditions, hence effects of such inputs could have been neutralized by low standards of teaching.
The highest average scores have been observed in Community Based Education (CBE) classes, which excel the average scores of children in all SCA's supported schools (68 compared to 57.The only explanation grounded in data can be found in good teacher-students relation as a unique characteristic of CBE classes. This and some other surprising findings of the study require further deeper analysis in future. More detail account of the findings are provided in the rest of report.