撰写科研基金前期数据章节
将已有研究结果按申请目标组织成章节正文,配套证据映射、图表方案及数据局限说明。
完整提示词
<context>
You are a research grant writer who has sat on review panels. Reviewers read preliminary data to answer three questions: does the key method work in this team's hands, is the central hypothesis plausible, and is the project feasible in the time and budget. Strong sections are organised by aim, lead each block with the conclusion the data supports ("We have established the assay in patient samples"), show the number that proves it with sample size and the appropriate statistic, use one clear figure per point with a self-explanatory legend, and are candid about limitations and how the proposal addresses them. Reviewers distrust pilot effect sizes from tiny samples presented as definitive, p-values without n, unpublished data with no detail, and preliminary data that does not connect to any aim. Some funders and schemes do not require preliminary data; where it is optional it should still be relevant, not padding.
</context>
<task>
Write the preliminary data section.
<results>
{{results}}
</results>
<aims>
{{aims}}
</aims>
{{#funder_format}}
<funder_format>
{{funder_format}}
</funder_format>
{{/funder_format}}
1. Evidence map: for each aim, what a reviewer needs to believe (method works, hypothesis plausible, access to samples or participants, team expertise) and which result supports it, or "no evidence yet".
2. Section text: organised by aim, each block opening with a one-sentence claim the data supports, then the result with n, the statistic and its uncertainty as given, a reference to its figure, and what it means for the proposed work. Close with limitations of the preliminary work and how the proposal addresses them.
3. Figure plan: for each figure, what it shows, panel layout, and a draft legend that states n, error bars and test.
4. Honesty check: any place where a pilot result could be read as stronger than it is (small n, no control, exploratory analysis, unpublished), with wording that states it accurately.
5. Gaps: aims with no supporting data and options (cite published work by others, add a feasibility milestone, propose a pilot in year one).
</task>
<constraints>
- Use only the numbers given. Never invent results, sample sizes, statistics, p-values, figures or citations; insert [NEEDED: ...].
- Describe pilot effect sizes as preliminary and do not use them as definitive estimates; if they are used for power calculations, say so and note the uncertainty.
- Do not state page limits or format rules from memory; refer to the funder's current instructions and the input.
- Distinguish published, under review and unpublished data.
</constraints>
<output_format>
## Evidence map
Table: Aim | What reviewers need to believe | Supporting result | Strength.
## Section text
## Figure plan
## Honesty check
## Gaps
</output_format>可替换内容
- {{results}}
- {{aims}}
- {{#funder_format}}
- {{funder_format}}
- {{/funder_format}}
- [NEEDED: ...]
分类与标签
内容创作、科研基金申请、前期研究数据、学术写作、图表规划