Multidimensional policy citation features: Insights into policymakers' policy adoption decision-making

Abstract

Scholarly citation has been extensively utilized to explore scholars' citation behaviors and elucidate their citation decision-making during academic writing. Likewise, policy citation serves as an invaluable instrument for scrutinizing policymakers' policy adoption decisions during policy formulation. This study offers a fine-grained portrayal and measurement of policy citation features, providing insights into policy adoption decisions from a novel perspective of citation choice. Specifically, each policy citation is conceptualized as a multidimensional feature collection consisting of 26 interpretable citation features (comprising 51 measurements) pertinent to the adoption of specific policies. These features are classified into three main categories: citation authority, citation proximity, and citation continuity. Utilizing large-scale information and communications technology (ICT) policies in China as empirical data, we conduct a series of logistic regressions and Random Forest-based classification experiments to quantitatively evaluate the importance of each constructed citation feature on policymakers' adoption decisions. Our empirical results reveal that policymakers' adoption of specific policies is predominantly influenced by citation proximity, followed by citation authority and citation continuity. Notably, central and local policymakers exhibit distinct adoption patterns; the former tends to prioritize policy continuity in their decision-making, whereas the latter lean towards adopting high-impact policies and learning from policy adoption pioneers. Moreover, the impact of policies is intricately entwined with their citation patterns, with high-cited policies often spearheading policy innovations, while low-cited policies tend to follow and imitate.