Three rounds of collaborative lesson study with 15 teachers: how should changes in self-efficacy be interpreted?
A blended lesson-study approach based on USP in a technology-supported environment

English translation
A 2018 journal article documents a cross-school lesson-study process and teachers’ self-reported outcomes. It is useful for discussing how design research can improve an approach, and for distinguishing teachers’ perceptions from student learning and causal evidence. The original gives only the publication month and year; this site does not invent a day or present an older paper as recent news.
Sources and editorial checks
Editorial check:
- Source type
- Publication status unverified
- Scope
- The implementation and research setting described in the material is mainland China.
- Study design
- Design research with three iterative rounds across three semesters, combining teacher surveys and interviews; no independent control group.
- Population and context
- A cross-school Chinese language teaching community involving 10 primary schools in Liwan District, Guangzhou.
- Sample
- 15 Chinese language teachers from 10 primary schools in Liwan District, Guangzhou, and 1 teacher educator.
- Intervention and comparison
- Three rounds of design iteration within the same teaching community; no independent comparison group.
- Main findings
- The paper reports that teachers’ self-reported self-efficacy was higher in round three than in the first two rounds; the difference between rounds one and two was not significant. Student learning outcomes were not directly reported.
- Outcomes and uncertainty
- The main measure is teachers’ self-reported self-efficacy. Student learning outcomes are not reported, and causal effects are not identified.
- Funding and conflicts of interest
- The first page lists Huaqiao University’s research start-up funding project 17SKBS212; no separate conflict-of-interest statement was found.
- Resource version
- China Educational Technology, Issue 3 (No. 374), 2018, pp. 97–102; PDF in a university institutional archive. · Source date: Unspecified
Limitations
- The sample is small and limited to one region, subject and teaching community.
- The measure is teachers’ self-reported self-efficacy, not student learning outcomes.
- The activities, roles and experiences changed across rounds, so the causal effect of any single component or technology cannot be isolated.
- The original gives March 2018 only. sourceDate remains null; the version field retains the month and year, and the PDF upload path is not used to infer a publication date.
- The original questionnaire data have not been independently recalculated, and the specific peer-review process has not been verified.
Access: Free · Status: Available. Evidence from other regions should be considered alongside local curricula, language and implementation conditions.
Editorial analysis
Original analysis based on the sources listed. Research findings, policy goals, and editorial judgment are identified separately; original rights remain with their sources.
Three rounds of collaborative lesson study with 15 teachers: how should changes in self-efficacy be interpreted?
This is historical research published in Issue 3 of China Educational Technology in 2018, not a recent paper. A PDF in a Beijing Normal University institutional archive provides the complete six-page text. The publication information gives only the month and year, so this review retains the month and does not invent a specific date. Read the full paper.
The study followed one cross-school teaching community in Guangzhou: 15 Chinese language teachers from 10 primary schools and one teacher educator iterated through lesson study over three consecutive semesters. The paper reports higher teacher-reported self-efficacy in the third round than in the first two; the difference between the first and second rounds was not significant. This result merits discussion, but it cannot be translated directly into a claim that students learned more.
First ask: whose measure changed, and what did it measure?
A teacher’s sense of being able to carry out teaching tasks is useful information about professional learning. It may help explain whether teachers are willing to try something or feel supported, but it is at a different level from what students actually do in class. Describing the former as the latter skips a process that needs to be measured. Even if the two are related in some settings, one measure cannot stand in for the other.
When reading research like this, name the outcome in full: who answered, what they were asked, and when. Do not leave only a statement of statistical significance and summarize it broadly as “teaching effectiveness improved.” Self-reports may also be affected by familiarity with an activity, expectations or a change in the standard respondents use for comparison. These are possible explanations to investigate further, not biases already shown to have affected this study.
Three rounds of improvement show how an approach developed
A strength of design research is that it can identify difficulties in real practice, revise a process and leave evidence that others can discuss. It need not be dismissed as worthless because it lacks a randomized control group. But the questions it is best suited to answer differ from estimating the net effect of an intervention.
When the same community practises over time while activities, roles and forms of support also change, later differences may reflect accumulated experience, changing relationships among participants and revisions to the approach at the same time. Without a suitable comparison, the changes cannot all be attributed to one platform, script or step. A statistically significant difference between rounds does not by itself establish attribution. Statistical difference, practical educational importance and generalizability should be discussed separately.
Treat the process as a set of questions to test
For a school teaching team, this paper is better read as a list of questions than as a standard process to copy: What teaching problem are participants studying together? Can colleagues understand the observation records? Are lesson revisions based on student work, classroom observation or impressions? What should be retained or removed in the next round? These are questions this article suggests bringing to collaborative lesson study, not an approach already proven effective in other schools.
A small-scale trial can also record its costs: how much time teachers spend, which discussions actually change their judgments, and which steps merely add materials. To study student outcomes, researchers would need to define learning measures and comparison arrangements clearly and report cases where participants did not take part, withdrew or showed no improvement. The number of reflections written or resources uploaded cannot be treated as the ultimate learning gain.
Boundaries on using this historical material
The paper examines online-supported collaborative teacher development at the time; it does not test today’s generative AI. Using it to think about teacher collaboration is reasonable. Using it as evidence that a particular AI product works goes beyond what the study examined. It presents a small practice model from one region and one subject. Its relevance to other schools depends on curriculum, organizational conditions and teachers’ needs.
The first page discloses support from a Huaqiao University research start-up project. This review has not independently recalculated the survey or obtained records of the specific peer-review process. Readers can check the methods and discussion in the original, while continuing to distinguish teacher self-reports, student outcomes and causal claims.
Source: China Educational Technology; institutional archive at Beijing Normal University · aic-fe.bnu.edu.cn