Studies on the relationship between the numbers of citations and downloads of scientific publications is beneficial for understanding the mechanism of citation patterns and research evaluation. However, seldom studies have considered directionality issues between downloads and citations or adopted a case-by-case time lag length between the download and citation time series of each individual publication. In this paper, we introduce the Granger-causal inference strategy to study the directionality between downloads and citations and set up the length of time lag between the time series for each case. By researching the publications on the Lancet, we find that publications have various directionality patterns, but highly cited publications tend to feature greater possibilities to have Granger causality. We apply a step-by-step manner to introduce the Granger-causal inference method to information science as four steps, namely conducting stationarity tests, determining time lag between time series, establishing cointegration test, and implementing Granger-causality inference. We hope that this method can be applied by future information scientists in their own research contexts.

研究引文数量与科学出版物的下载量之间的关系有助于理解引文模式和研究评价的机制。但是，很少有研究考虑下载和引文之间的方向性问题，或采用每个案例的下载和引文时间序列之间的个案时滞长度。在本文中，我们介绍了格兰杰因果推理策略，以研究下载与引用之间的方向性，并为每种情况设置时间序列之间的时间间隔。通过研究柳叶刀上的出版物，我们发现出版物具有不同的方向性模式，但被高度引用的出版物往往具有更大的可能性来具有格兰杰因果关系。我们采用逐步的方式将格兰杰因果推理方法引入信息科学四个步骤，即进行平稳性测试，确定时间序列之间的时间差，建立协整测试以及实施格兰杰因果关系推理。我们希望未来的信息科学家可以在他们自己的研究环境中应用这种方法。