Communication of the likelihood that a grant proposal is funded is important for individual researchers and academic institutions. For research project grants (RPGs) funded by the US National Institutes of Health (NIH), this is measured by the success rate, the fraction of grants reviewed in a given fiscal year that are funded. In 2016, I posted a relatively simple computational model that allowed estimation of these success rates from historical data of NIH appropriations and associated data, based largely on the fact that commitments to fund NIH grants are typically made over four fiscal years rather than coming from a single fiscal year's appropriations. Data for appropriations and success rates for fiscal years 2016-2020 are now available, allowing this model to be tested without no adjustable parameters. Over this period, the NIH appropriation increased by 29.7% to $39.313 billion, yet the RPG success rate increased from 0.183 to 0.201, a relative increase of only 9.8%. This difference is accurately reproduced by the model, indicating that the modest changes in success rate despite large appropriation increases reflect the consequences of the funding grants from multiple fiscal years and the response of the biomedical research community in submitting an increased number of proposals. The model can also be extended to estimate the success rates for individual institutes with reasonable accuracy.

中文翻译：

---

根据美国国立卫生研究院拨款历史对研究项目拨款的成功率进行建模：延长至2020年

对于个人研究人员和学术机构来说，传达资助提案的可能性的沟通很重要。对于由美国国立卫生研究院（NIH）资助的研究项目赠款（RPG），用成功率来衡量，该成功率是给定财政年度中审查的赠款中获资助的部分。2016年，我发布了一个相对简单的计算模型，该模型可以从NIH拨款的历史数据和相关数据中估算出这些成功率，主要基于这样一个事实，即资助NIH赠款的承诺通常是在四个财年内做出的，而不是来自于单个财政年度的拨款。现在可以获取2016-2020财年拨款和成功率的数据，从而可以在没有可调参数的情况下对该模型进行测试。在这期间 美国国立卫生研究院的拨款增加了29.7％，达到393.13亿美元，但RPG的成功率从0.183增加到了0.201，相对增加了9.8％。该模型准确地再现了这种差异，表明尽管拨款大量增加，但成功率的适度变化反映了来自多个会计年度的拨款赠款的影响，以及生物医学研究界对提交越来越多的提案的反应。该模型还可以扩展为以合理的准确性估算单个机构的成功率。表明尽管拨款大量增加，但成功率的适度变化反映了来自多个财政年度的拨款赠款的后果，以及生物医学研究界对提交越来越多的提案的反应。该模型还可以扩展为以合理的准确性估算单个机构的成功率。表明尽管拨款大量增加，但成功率的适度变化反映了来自多个财政年度的拨款赠款的后果，以及生物医学研究界对提交越来越多的提案的反应。该模型还可以扩展为以合理的准确性估算单个机构的成功率。