"讨论部分的"五段式结构"。" 第一段：简要重述核心发现（2–3句）。不是重复结果，而是"用大白话"概括"我们发现了什么"。第二段：与前人研究对话。你的发现和Smith (2022)一致吗？和Jones (2023)矛盾吗？为什么？第三段：理论贡献。你的发现对XX理论意味着什么？需要修正？需要加边界条件？第四段：实践/政策启示。你的发现对教育实践/临床干预/组织管理有什么启示？第五段：局限性与未来方向。诚实地讨论你的研究有哪些不足，未来可以怎么改进。

"讨论的"对话感"营造。" 最好的讨论不是"独白"，而是"与前人对话"。营造方法：①引用对立观点："While Smith (2022) argued that X causes Y through mechanism A, our findings suggest an alternative pathway: X → B → Y." ②承认分歧："Our results contradict Lee's (2021) finding that X has no effect on Y. The discrepancy may be due to [样本差异/方法差异/情境差异]." ③整合分歧："Taken together, Smith's and our findings suggest that X → Y holds in [情境A] but not in [情境B]." 这种"对话感"让审稿人感受到你在"参与学术对话"，而非"自说自话"。

"讨论的"意义升级"技巧。" 从"小意义"升级到"大意义"：①从具体发现到理论意涵："This finding suggests that self-determination theory may need to incorporate a 'cultural moderation' component." ②从统计结果到现实影响："An effect size of d = 0.45, while modest, translates to a 12% improvement in writing scores—a meaningful gain in educational practice." ③从单一研究到领域地图："Our study, together with recent work by [作者], points to an emerging consensus that [新共识]." ④从学术到社会："These findings have implications beyond academia. Policymakers designing AI-in-education initiatives should consider [具体建议]."

"讨论中的"局限性"表述艺术。" 局限性不是"认输"，而是"展示学术诚信"的机会。表述艺术：①具体而非笼统："A limitation is that our sample was drawn from a single city (Beijing), limiting generalizability to rural populations."（好）→ "A limitation is that our sample may not be representative."（太笼统）②将局限转化为未来方向："Because we could not experimentally manipulate X, causal inference is limited. Future research using a randomized controlled trial design could address this." ③不要把致命伤当"局限性"轻描淡写：如果你的研究有根本性的方法缺陷，不要藏在"局限性"里一笔带过。④局限性的数量：3–5个为宜。太多→读者怀疑你的研究质量。太少（0个）→审稿人觉得你不自省。

"讨论的"结尾句"力量。" 讨论部分的最后一段话（有时就是最后一句），是读者合上论文前看到的最后内容。它应该：①回到引言中的"大问题"：你的研究对这个大问题意味着什么？②给出有力结论："This study provides the first empirical evidence that AI-generated feedback can meaningfully improve writing coherence—challenging the assumption that only human feedback can nurture writing skills." ③避免夸张："This study revolutionizes the field of education."（太夸张）→ "This study offers a modest but meaningful step toward understanding how AI can support writing development." ④留下余韵："The question is no longer whether AI can support learning, but how to design AI interactions that nurture—rather than replace—human thinking."

"讨论部分的"五段式结构"。" 第一段：简要重述核心发现（2–3句）。不是重复结果，而是"用大白话"概括"我们发现了什么"。第二段：与前人研究对话。你的发现和Smith (2022)一致吗？和Jones (2023)矛盾吗？为什么？第三段：理论贡献。你的发现对XX理论意味着什么？需要修正？需要加边界条件？第四段：实践/政策启示。你的发现对教育实践/临床干预/组织管理有什么启示？第五段：局限性与未来方向。诚实地讨论你的研究有哪些不足，未来可以怎么改进。

"讨论的"对话感"营造。" 最好的讨论不是"独白"，而是"与前人对话"。营造方法：①引用对立观点："While Smith (2022) argued that X causes Y through mechanism A, our findings suggest an alternative pathway: X → B → Y." ②承认分歧："Our results contradict Lee's (2021) finding that X has no effect on Y. The discrepancy may be due to [样本差异/方法差异/情境差异]." ③整合分歧："Taken together, Smith's and our findings suggest that X → Y holds in [情境A] but not in [情境B]." 这种"对话感"让审稿人感受到你在"参与学术对话"，而非"自说自话"。

"讨论的"意义升级"技巧。" 从"小意义"升级到"大意义"：①从具体发现到理论意涵："This finding suggests that self-determination theory may need to incorporate a 'cultural moderation' component." ②从统计结果到现实影响："An effect size of d = 0.45, while modest, translates to a 12% improvement in writing scores—a meaningful gain in educational practice." ③从单一研究到领域地图："Our study, together with recent work by [作者], points to an emerging consensus that [新共识]." ④从学术到社会："These findings have implications beyond academia. Policymakers designing AI-in-education initiatives should consider [具体建议]."

"讨论中的"局限性"表述艺术。" 局限性不是"认输"，而是"展示学术诚信"的机会。表述艺术：①具体而非笼统："A limitation is that our sample was drawn from a single city (Beijing), limiting generalizability to rural populations."（好）→ "A limitation is that our sample may not be representative."（太笼统）②将局限转化为未来方向："Because we could not experimentally manipulate X, causal inference is limited. Future research using a randomized controlled trial design could address this." ③不要把致命伤当"局限性"轻描淡写：如果你的研究有根本性的方法缺陷，不要藏在"局限性"里一笔带过。④局限性的数量：3–5个为宜。太多→读者怀疑你的研究质量。太少（0个）→审稿人觉得你不自省。

"讨论的"结尾句"力量。" 讨论部分的最后一段话（有时就是最后一句），是读者合上论文前看到的最后内容。它应该：①回到引言中的"大问题"：你的研究对这个大问题意味着什么？②给出有力结论："This study provides the first empirical evidence that AI-generated feedback can meaningfully improve writing coherence—challenging the assumption that only human feedback can nurture writing skills." ③避免夸张："This study revolutionizes the field of education."（太夸张）→ "This study offers a modest but meaningful step toward understanding how AI can support writing development." ④留下余韵："The question is no longer whether AI can support learning, but how to design AI interactions that nurture—rather than replace—human thinking."