Inside Ethiopian classrooms, an autonomy gap between policy and practice

In classrooms across Ethiopia’s Gurage Zone, students sit in rows, copy notes from the blackboard, and recite memorized answers while teachers control nearly every decision about what is learned, how it is learned, and how it is assessed. A new study published in Frontiers in Education offers the most detailed picture yet of just how rare opportunities for learner autonomy are in Ethiopian secondary school English classrooms, and how much those opportunities matter when they do appear. Drawing on 136 observed lessons, surveys of 412 students, and interviews and focus groups with 34 teachers, the research documents what the authors call an autonomy gap: a wide distance between national policy that champions self-directed learning and classroom realities shaped by large classes, teacher-centered instruction, and an exam-driven curriculum.

The stakes extend beyond one country. Learner autonomy, broadly defined as taking responsibility for one’s own learning, has been associated in decades of research with stronger engagement, higher motivation, and better retention of language knowledge. In English as a Foreign Language settings, where students rarely encounter English outside the classroom, autonomy is increasingly viewed as necessary rather than optional. Ethiopia’s 2021 General Education Curriculum Framework explicitly states that students should become self-directed learners who can plan, monitor, and evaluate their own learning, and the national EFL syllabus for grades 9 through 12 lists developing autonomous learning skills as a cross-cutting competency. Yet previous studies have consistently described Ethiopian EFL classrooms as teacher-dominated, and no large-scale mixed-methods study had ever measured the link between teachers’ observed autonomy-supportive behaviors and students’ actual autonomous actions.

The research team, led by Esayas Teshome Taddese of INTI International University in Malaysia, together with colleagues at Wolkite University in Ethiopia and Ajman University in the United Arab Emirates, used an explanatory sequential mixed-methods design with quantitative priority. From four of the eight districts in Gurage Zone, they randomly selected two government secondary schools each, then recruited 34 of 36 eligible EFL teachers and 412 Grade 12 students selected systematically from class lists. Each teacher was observed during four 40-minute lessons using a 30-item schedule called the Teacher Autonomy-Supportive Practices Observation Schedule, or TAPOS, adapted from an established protocol and contextualized for Ethiopian classrooms. The tool rated five dimensions of autonomy support on five-point frequency scales: providing choices, providing rationales, acknowledging negative feelings, avoiding controlling language, and encouraging self-initiation.

Students completed two further instruments in a single 50-minute session. The Student Perception of Autonomy Support Scale, a 24-item questionnaire adapted from the Learning Climate Questionnaire and translated into Amharic through back-translation, measured how students interpreted their teacher’s behavior, with two distractor items about general teaching quality excluded from analysis. The Learner Autonomy Inventory, a 40-item scale with subscales for goal-setting and planning, self-monitoring, strategy use, and resource-seeking, captured students’ self-reported autonomous behaviors. Both instruments showed good internal consistency in pilot testing. Because students were nested within teachers, the researchers treated teacher-level and student-level variables separately throughout, a distinction they maintain carefully: observed autonomy-supportive practices, measured by TAPOS at the teacher level, are conceptually distinct from perceived autonomy support, measured by SPASS at the student level.

The observational results were stark. Across 136 lessons, the mean TAPOS score was just 2.1 out of 5. The most common dimension was providing rationales, observed in 68 percent of lessons, but even that fell below the scale midpoint, and when rationales appeared they were typically generic, framed around exam preparation rather than genuine usefulness. Providing choices occurred in only 39 percent of lessons, most often as a choice between two homework exercises. Acknowledging negative feelings appeared in just 12 percent of lessons. Avoiding controlling language was rare in the opposite sense: teachers used phrases like you must finish by tomorrow in 74 percent of lessons. Most striking of all, encouraging self-initiation, in which teachers invite students to suggest topics, pose questions, or make decisions about their learning, occurred in only 4 percent of lessons. Teachers with master’s degrees scored slightly higher than those with bachelor’s degrees, but urban and rural schools showed no significant differences.

Student data mirrored the classroom picture. The mean SPASS score for the 22 autonomy-specific items was 2.7 out of 5, closely tracking the observed practices and suggesting that students’ perceptions largely reflect what teachers actually do. The lowest-rated statements concerned asking for student opinions about what to learn and allowing students to choose their own homework. Self-reported learner autonomy on the LAI averaged 2.4 out of 4, with goal-setting the weakest subscale: the least popular autonomous activity was setting weekly English goals, and only 12 percent of students said they read English texts outside class often or always. Interestingly, female students reported significantly higher autonomous behaviors than males, a difference consistent with findings from Korea and Spain, though the authors caution that some self-monitoring items may reflect compliance with school expectations rather than genuine autonomy.

The statistical core of the study lies in its multilevel modeling. An unconditional two-level model showed that approximately 17.5 percent of the variance in learner autonomy was attributable to differences between teacher clusters, justifying the hierarchical approach. In the full two-level linear mixed-effects model, perceived autonomy support was positively associated with learner autonomy, as was prior English achievement, while teacher-level TAPOS scores were also positively associated with learner autonomy after accounting for perceptions. Class size showed a negative association, indicating that larger classes constrain opportunities for individualized autonomy support. Teacher experience, initially included, was dropped from the final model because it was not significant and correlated substantially with teacher education level.

The mediation analysis provided the study’s most theoretically interesting result. Using a 2-1-1 multilevel mediation framework, in which the teacher-level TAPOS score served as the predictor, student perceptions as the mediator, and learner autonomy as the outcome, the researchers estimated a significant indirect effect of 0.087 with a 95 percent Monte Carlo confidence interval that excluded zero. Roughly 23 percent of the total association between teacher practices and learner autonomy operated through students’ perceptions of autonomy support. This pathway is consistent with Self-Determination Theory, which holds that autonomy-supportive teaching fosters intrinsic motivation and self-regulation by satisfying basic psychological needs. But the authors are careful to note that the cross-sectional design cannot establish temporal ordering or causation, and the small number of teacher-level units limits precision.

The qualitative phase added crucial nuance. Students expressed ambivalence: one high-autonomy student wished her teacher would let her choose essay topics, saying she was tired of always writing about My Village, while a low-autonomy student admitted that if asked what he wanted to learn, he would not know, because all he knew was what was in the textbook. Teachers, meanwhile, overwhelmingly invoked culture, with 30 of 34 suggesting that autonomy conflicts with Ethiopian norms in which the teacher decides like a parent. One teacher described a failed attempt at student choice that ended in twenty minutes of unresolved argument. Yet four teachers reported success with gradual, structured strategies, such as offering small binary options before expanding student decision-making, or using weekly self-assessment checklists that prompted students to honestly identify what they did not understand. Both teachers and students identified the national examination, with its 100 multiple-choice questions, as the primary structural barrier, rendering student opinion irrelevant to assessment.

The study’s implications are layered. Theoretically, it offers preliminary evidence that Self-Determination Theory’s mechanisms operate in a non-Western, resource-poor setting, and that autonomy can take culturally congruent forms through structured rather than open-ended choices, consistent with the Culturally Appropriate Autonomy Model. Practically, the authors recommend that teachers begin with structured options, provide explicit and meaningful rationales, and introduce simple self-assessment routines. They call for teacher education programs to include modules on fostering learner autonomy, noting that none of the six programs in Gurage Zone currently does. At the policy level, they suggest that even modest changes to national examinations, such as items requiring essay planning or self-evaluation, could signal the value of autonomy within the assessment system. The findings echo patterns documented in Saudi Arabia and Jordan, where exam-driven curricula similarly suppress learner independence. What emerges is neither a story of cultural impossibility nor a simple prescription, but evidence that the autonomy gap in Ethiopian classrooms is real, measurable, and potentially closable through gradual, structured approaches that respect both cultural context and systemic constraints.

Subject of Research: Learner autonomy and autonomy-supportive teaching in Ethiopian secondary EFL classrooms

Article Title: Learner autonomy in Ethiopian secondary EFL classrooms: observed teacher practices and student perceptions

Article References: Taddese, E. T., Yassin, H. A., Dinsa, M. T., & Shorouk, A. (2026). Learner autonomy in Ethiopian secondary EFL classrooms: observed teacher practices and student perceptions. Frontiers in Education, 11, Article 1940198. https://doi.org/10.3389/feduc.2026.1940198

Image Credits: AI Generated

DOI: 10.3389/feduc.2026.1940198

Keywords: learner autonomy, EFL, Ethiopia, autonomy-supportive teaching, Self-Determination Theory, multilevel modeling, mediation analysis, classroom observation, teacher education, assessment policy, secondary education, mixed methods