鑑於國際閱讀評比與國內閱讀力調查報告屢屢提及我國閱讀教育表現呈現城鄉差距的情況,本研究從教育治理的理論觀點出發,分析臺灣地方政府推動閱讀教育之策略與治理特性,進而探究改善閱讀教育城鄉差異的可能途徑。研究採取多重個案研究法,依據極端案例抽樣、 最大變異抽樣與理論為本抽樣等三原則,篩選出四個個案縣市,並針對2010~2020年中央與地方政府的政策與官方文件,挑選符合重要政治性抽樣、準則抽樣和資料適足性等三個原則之政策文本,進行內容分析。本研究的主要結論為:一、縣市閱讀教育治理的城鄉特性:(一)閱讀表現優異的個案縣市展現法制治理特性;(二)閱讀表現優異的都會個案傾向反映市場治理特性;(三)偏鄉地方政府政治能力會影響閱讀教育治理策略的多元化。二、閱讀教育城鄉差異的改善途徑:(一)偏鄉地方可藉由政治能力填補政府技術能力的不足;(二)偏鄉地方可透過網絡治理彌補政府技術能力的不足。本研究亦根據研究結果對於當前教育政策實務與未來研究提出建議。 Research Background and Purposes International studies and domestic education surveys have identified considerable gaps in reading performance between the rural and urban regions of Taiwan. In contrast to extant studies, the present study, from a macroscopic perspective, explored possible explanations for the urban-rural discrepancies and focused on government intervention for improving adult reading ability and student reading performance. Based on theories of governance and public policy, the developed an analytical framework to understand the strategies adopted by local governments in Taiwan to promote reading education, identify the characteristics of local reading education governance in Taiwan, and suggest potential approaches to resolving rural-urban differences in reading education. Theoretical Perspectives and Framework This study considered governance as a dynamic process of power exercise and decision-making. In a democracy, governance refers to an interactive process through which the government, stakeholders, and civil society determine educational policies and affairs. Specifically, the study defined education governance as the collective behavior of the government and relevant actors in the formation and implementation of educational policies. If stakeholders from the public, the private, and the third sectors become progressively involved in developing education policies, governments may no longer be the sole and dominant policy actors in the education domain. Accordingly, this study considered reading education governance in Taiwan to be the collective interaction process between the government and relevant stakeholders on reading education policy development. Because policy instruments or tools are perceived as a governmental approach to achieving policy goals and addressing policy problems, scholars have viewed education governance from the perspective of public policy as involving the process of selecting and implementing policy instruments. This study developed an analytical framework for education governance, which comprises two key components, namely governance models and governance strategies. Both components were determined by Christopher Hood’s NATO (nodality-authority-treasure-organization) model, which is regarded as the most representative classification structure in the literature on public administration and policy. The NATO model covers almost all aspects of government actions that may affect the people through four approaches and accordingly divides policy instruments into four categories: (1) nodality, (2) authority, (3) treasure, (4) and organization. Furthermore, the present study divided governance models into four categories depending on which types of policy instruments are selected and implemented by the government. The main categories of education governance in the classification framework are outlined as follows: (1) legal governance, which relies on the legitimacy of law orders to generate control and obedience, reflects the administrative logic of the traditional bureaucratic system; (2) corporatist governance, in which the government plans and negotiates education policies under organizational and national systems, reflects the characteristics of traditional public administration; (3) market governance, in which the government prefers to adopt market mechanisms to generate competition and incentives, reflects the logic of new public management that emphasizes flexible principles and efficient methods to achieve policy goals; and (4) network governance, in which the government prefers operations not related to bureaucratic or hierarchical principles, involves the collaborative interaction of multiple actors in the educational domain. Research Methods Because of the lack of literature on how local governments govern reading education, this study adopted an exploratory research orientation to obtain a comprehensive understanding of this topic. A comparative case study approach was applied to provide a more thorough description of the change in governance and policies regarding reading education. Specifically, the study adopted focused case comparisons and followed the principle of multiple-case holistic analysis. To expand the depth and breadth of research data, this study applied a two-stage intentional sampling method. The first stage involved extreme or deviant case sampling to extract useful information from successful case experiences. This stage entailed searching for local counties and cities with excellent reading performance. The second stage involved selecting cases that offer valuable insights into the understanding of governance characteristics and strategies related to local government promotion of reading education. The cases were selected from those that stood out in the first stage based on the principles of maximum variation sampling. In addition, the principle of theory-based sampling was adopted for case selection in the second stage. Finally, the study selected four cases that exhibited considerable differences in their local governments in terms of governance capacity. This study perceives governance capability as a collective term that refers to a government’s political capability and technical capability and is a crucial factor affecting the choice and effectiveness of governance strategies. In accordance with the principles of sampling politically important cases, criterion sampling, and data sufficiency, this study collected prominent policy and official documents published during 2010-2020 by four individual local governments and then conducted a comparative content analysis. Research Results and Implications The study yielded five major findings. First, all cases with high performance reflected the features of legal governance when promoting reading education. Second, reading education in urban counties with high performance was inclined to exhibit strategies relating to market governance. Third, local governments’ political capacity was influential in the diversification of governance strategies for promoting reading education in rural counties. Fourth, rural counties could compensate for their insufficient technical capacity with political capacity. Fifth, rural counties could compensate for their insufficient political capacity through network governance. The study findings can contribute to the understanding of differences in local governments’ governance strategies and characteristics related to the promotion of reading education. This study argues that when local governments prefer to use bidding and tax incentives as policy instruments for achieving their policy goals and attach great importance to the effective use of resources and the control of cost efficiency, their local communities would be more vulnerable to problems of sacrificing social justice in the governance process. In addition, local governments in rural regions with insufficient financial resources and funds for education still produce effective reading education governance if the local government and council heads are affiliated with the same political party; this is because affiliation with the same political party engenders a considerable degree of political capacity for local governments. Finally, the introduction of civil assistance and resources would contribute to narrowing the rural-urban gap in reading education and performance through the implementation of specific governance strategies (such as setting up interdisciplinary teams, promoting private sector sponsorship and cooperation, and implementing other organizational policy tools that demonstrate the characteristics of network governance).