China’s sweeping attempt to reduce academic pressure has won support for its goals, but the teachers responsible for putting the policy into practice are far less convinced by how it is working.
A new peer-reviewed study of 12,524 primary and secondary school teachers in Sichuan Province found generally negative evaluations of the implementation effectiveness of China’s “Double Reduction” education policy. Yet the picture was not one of simple rejection. Teachers simultaneously recognised potential benefits for student development, worried about heavier demands on their own work, and expressed substantial interest in refining rather than abandoning the reform.
The distinction matters because large education reforms do not operate only through legislation or ministry directives. They depend on what happens in schools, classrooms and staff rooms after a policy reaches the people expected to deliver it. The new findings therefore offer a window into the gap that can emerge between a policy’s objectives and the experience of implementation on the ground.
A reform designed to reduce pressure
China introduced the Double Reduction policy as a major intervention aimed at reducing excessive homework and curbing the burden associated with off-campus academic tutoring. The reform is consequently about more than the amount of work students complete. It changes expectations around schools, teachers, families and the wider educational system.
That makes teachers especially important to understanding whether the policy is functioning as intended. A reform can be formally adopted while still producing uneven results if frontline implementers face incompatible workloads, limited resources, regional differences or uncertainty about how broad policy goals should translate into everyday teaching practice.
Researchers She Wanbin and Wu Yan examined those implementation dynamics through policy diffusion theory. Rather than treating teachers merely as recipients of a central policy, this perspective places attention on how implementers interpret a reform, assess its consequences and decide whether its practices are workable in their own settings.
More than 12,000 teachers across four regions
The researchers used a descriptive survey design and stratified random sampling to recruit 12,524 teachers from four regions of Sichuan with different levels of socioeconomic development. That regional structure was important because the same education policy can encounter very different conditions depending on local resources, staffing, school context and economic development.
Teachers completed a purpose-built Double Reduction Education Policy Implementation Survey. The instrument examined three broad dimensions: evaluations of implementation effectiveness, perceptions of the policy’s impacts, and intentions concerning how the policy should be improved.
The researchers reported a Cronbach’s alpha of 0.757 for the questionnaire, indicating acceptable internal consistency for the study’s measures. The Kaiser-Meyer-Olkin statistic was 0.811, with p < 0.001, supporting the suitability of the data structure for the analyses used by the researchers.
These measurement details are important because a very large sample does not automatically make a survey informative. The questions still need to capture the intended constructs with reasonable consistency. Here, the reported diagnostics provided evidence that the instrument was sufficiently coherent for the authors’ analysis, although it remained a self-developed questionnaire rather than a long-established international scale.
Teachers were not simply for or against the policy
The central finding was a tension. Teachers generally evaluated implementation effectiveness negatively, but their views of the policy’s consequences were more mixed.
On one side, respondents acknowledged that the reform could promote student development. Reducing excessive academic burdens is, after all, a goal that can align with broader concerns about student well-being and the quality of learning. On the other side, teachers reported concerns about the workload created by implementation.
This is a useful reminder that a policy can produce benefits for one part of an education system while transferring costs to another. If homework is redesigned, after-school provision changes, or schools take on responsibilities previously handled elsewhere, the work required to make those changes does not disappear. It can instead shift toward teachers and schools.
The study therefore does not support a simple conclusion that teachers opposed reducing student pressure. Their responses instead suggest that support for a policy objective can coexist with dissatisfaction about implementation. That difference is crucial for policymakers because the appropriate response to opposition to a goal is not necessarily the same as the response to concern about how that goal is being delivered.
Impact perceptions dominated the statistical model
The researchers used linear regression to examine how different dimensions related to teachers’ overall policy attitudes. Together, effectiveness evaluations, perceived impacts, improvement intentions, teaching experience and regional disparities explained 73.4% of the variance in policy attitudes, with R² = 0.734 and p < 0.001.
The strongest coefficient by a wide margin was teachers’ perception of policy impacts, at β = 0.873. Effectiveness evaluations had a smaller positive coefficient of β = 0.081. Improvement intentions were negatively associated with the overall attitude measure at β = -0.077, while teaching experience had a coefficient of β = -0.138 and regional disparities β = -0.019.
Those coefficients should not be read as if the study had identified five independent causes of teachers’ attitudes. This was a survey, not an experiment, and the measures were collected within the same research design. What the model does show is that teachers’ perceptions of the policy’s actual consequences were closely bound up with their broader assessment of it.
That is more informative than a headline measure of approval alone. It suggests that the practical effects teachers believe they are seeing may matter more to their overall response than abstract support for reform. For policy diffusion, that is a consequential distinction: implementers are more likely to sustain a reform when the lived results make sense within their professional environment.
Experience and geography also mattered
The negative coefficient for teaching experience indicates that attitudes differed with time in the profession, even after the other variables in the model were considered. The study does not justify reducing this to a claim that experienced teachers simply resist change. Longer-serving teachers may have accumulated more experience of previous reforms, different workload expectations or a stronger basis for comparing policy promises with classroom realities.
Regional disparities were statistically incorporated as well. Although their coefficient was comparatively small, the inclusion of teachers from four socioeconomic regions reinforces a central problem for national policy: implementation capacity is rarely distributed evenly.
A uniform national instruction can therefore generate non-uniform local experiences. A school with stronger staffing, professional support and local resources may absorb new responsibilities differently from a school already operating under tighter constraints. The researchers consequently argue against assuming that a single implementation strategy will work equally well everywhere.
Teachers wanted refinement, not merely retreat
One of the most policy-relevant findings was that teachers actively supported improving the reform, particularly through differentiated strategies. This moves the discussion beyond a binary choice between keeping the policy unchanged and reversing it.
The authors highlight several areas for more targeted implementation: teacher workload, alignment between teachers’ competencies and new demands, incentive mechanisms, differences between academic subjects, and regional levels of educational development.
These proposals share a common logic. A reform aimed at reducing student burden may be difficult to sustain if it creates unmanaged burden elsewhere. Likewise, a policy that expects different subjects and schools to adapt in identical ways can overlook meaningful differences in teaching practice and local capacity.
Differentiation does not necessarily mean weakening a national objective. It can mean preserving the objective while allowing the path toward it to reflect the conditions in which schools actually operate.
What this tells us about policy implementation
The broader lesson reaches beyond China’s education system. Governments often evaluate reforms using visible outputs: new rules, programme participation, workload limits or compliance measures. Frontline workers experience a different layer of policy. They encounter the additional planning, coordination, judgement and adaptation required to make a reform function.
When those implementation costs are ignored, a policy can appear successful administratively while becoming difficult to sustain professionally. Conversely, negative feedback from implementers does not automatically mean that the policy’s underlying objective is misguided. It may identify where the delivery model needs adjustment.
The Sichuan findings fit that more complicated picture. Teachers could see value in promoting student development while still questioning implementation effectiveness and reporting workload concerns. Their support for refinement suggests that the most productive policy response may be to treat frontline criticism as information about implementation rather than as simple resistance.
Important limits to what the study can show
The size of the sample is a major strength, but it should not be confused with causal evidence. The study was descriptive and survey-based. It can identify patterns in teachers’ reported attitudes and statistical relationships among measured variables, but it cannot establish that the Double Reduction policy caused particular teacher attitudes or that changing one survey dimension would necessarily change another.
The research was also conducted in Sichuan Province. Its four-region design captures meaningful socioeconomic variation within the province, but China’s education system spans many provinces and local contexts. The findings should therefore not automatically be treated as nationally representative of every Chinese teacher.
Self-reported data introduce another limitation. Teachers were reporting their own perceptions of effectiveness, impact and improvement needs. Those perceptions are exactly what the study sought to understand, but they are not the same as independently measured student outcomes, teacher working hours or administrative performance.
Finally, the questionnaire was developed for this study. Its reported reliability and sampling adequacy were acceptable, but further validation across different populations would strengthen confidence in how the instrument performs beyond this sample.
A reform is also the experience of the people implementing it
The Double Reduction policy was created to change the experience of students, but this study shows why education reform must also be evaluated through the experience of teachers.
Among more than 12,000 educators, the policy was not viewed through a single lens. Teachers questioned implementation effectiveness, recognised potential developmental benefits, worried about their own workload and wanted more differentiated implementation. Their impact perceptions were also the dominant component in the regression model explaining overall policy attitudes.
For policymakers, that combination is more actionable than either blanket praise or blanket criticism. It points toward a governance problem that can potentially be improved: how to preserve the aim of reducing excessive student pressure while designing implementation around the capacity, incentives and working conditions of the people expected to make the reform real.
Source Information
Study: Teachers’ perspectives on ‘Double Reduction’ education policy: a survey of 12,524 educators in China
Authors: She Wanbin and Wu Yan
Journal: Humanities and Social Sciences Communications
Published: 26 September 2026
DOI: 10.1057/s41599-026-09243-z
Study design: Descriptive survey using stratified random sampling and linear regression analysis
Sample: 12,524 primary and secondary school teachers across four socioeconomic regions of Sichuan Province, China
