The complexity of farming and rural communities poses challenges to research and extension initiatives seeking industry-wide change. The effectiveness of these initiatives depends on factors ranging from individual psychology to the science of effects. Tools like intervention logic models, which visualise the causal chain of an intervention through to its impact, are useful for programme planning but are of limited usefulness for comparing the relative benefits of initiatives. Benefit-cost analysis can quantify relative benefits, but applications often oversimplify the causal chain and leave key impacts unquantified. This paper aims to develop a benefit-cost analysis modelling framework that both captures the important causal logic from intervention to impact as well as quantifying and monetising key benefit categories. We apply the framework to value the impacts of ‘Hill Country Futures’, a programme of research and extension designed to assist in future-proofing the environmental sustainability, profitability, and well-being of New Zealand's hill country farmers, their farm systems, and communities. The QUICK (Quantifying and Understanding the Impact of Capability and Knowledge) model uses an agent-based simulation to represent processes of skill development and practice improvement that result from both extension activities for management and research tools as well as the use of those tools. We calibrate the model using a combination of an external predictive adoption model, an expert workshop, and researcher judgment. We use a choice experiment to value simulated changes in financial, environmental, community, and well-being outcomes. We explore how optimising the programme of extension interventions through targeting could increase impact. Our results suggest that the benefits of the ‘Hill Country Futures’ programme outweigh the costs by a factor of 13.5. We find that targeting extension efforts towards building awareness could be slightly more beneficial than targeting both awareness and skill development. We find no evidence that targeting extension efforts towards the best resource would be beneficial. This paper combines desirable features from intervention logic models, benefit-cost analysis, choice modelling, and agent-based models to value the benefits of a large programme of research and extension. It highlights that such modelling can be useful for evaluating both planned and complete programmes.

中文翻译：

---

基于代理的方法快速评估农业研究和推广的效益

农业和农村社区的复杂性给寻求全行业变革的研究和推广计划带来了挑战。这些举措的有效性取决于从个人心理学到效果科学等各种因素。诸如干预逻辑模型之类的工具可以将干预的因果链及其影响可视化，对于计划规划很有用，但对于比较举措的相对效益而言作用有限。效益成本分析可以量化相对效益，但应用程序往往过于简化因果链，导致关键影响无法量化。本文旨在开发一个效益成本分析建模框架，该框架既捕获从干预到影响的重要因果逻辑，又对关键效益类别进行量化和货币化。我们应用该框架来评估“山地未来”的影响，这是一项研究和推广计划，旨在帮助新西兰山地农民及其农场系统的环境可持续性、盈利能力和福祉面向未来。社区。 QUICK（量化和理解能力和知识的影响）模型使用基于代理的模拟来表示技能开发和实践改进的过程，这些过程是由管理和研究工具的扩展活动以及这些工具的使用产生的。我们结合外部预测采用模型、专家研讨会和研究人员的判断来校准模型。我们使用选择实验来评估财务、环境、社区和福祉结果的模拟变化。我们探讨如何通过目标优化扩展干预计划来增加影响力。我们的结果表明，“山地未来”计划的收益超过成本 13.5 倍。我们发现，以提高意识为目标的推广工作可能比以意识和技能发展为目标的推广工作稍微有利一些。我们没有发现任何证据表明针对最佳资源的推广工作是有益的。本文结合了干预逻辑模型、效益成本分析、选择建模和基于代理的模型的理想特征，以评估大型研究和推广项目的效益。它强调这种建模对于评估计划的和完整的计划很有用。