[目的/意义]全面解析数据问题的复杂表现及其背后的深层逻辑对于把握科研数据管理的核心内容、提 升数据管理成效具有重要的理论和现实意义。[方法/过程]运用布迪厄“场域-惯习”理论，突破传统二元对立的思 维局限，旨在深度揭示科研数据管理中的数据权力、主体行为与资源配置等现实问题。选取WOS和Retraction Watch同时收录的期刊撤稿论文作为分析对象，通过统计分析、比较分析和科学计量学方法，发现当前我国科研 数据管理问题呈现出增长过快、类型多样、层次复杂、原因交织、主体多元、分布广泛、处置薄弱的态势。[结果/结 论]数据问题映射出学术场域评价重压下的科研数据管理价值偏差、科研数据管理主体差序格局、不良数据管理 惯习、主体数据逐利行为与政策监管缺位等现实问题。因而，我国科研数据管理应标本兼治，既要有效优化学术 评价体系，健全科研诚信管理制度，也要平衡科研数据管理激励分配格局、提升数据素养、完善科研数据管理 体系。

[Purpose/significance] A comprehensive analysis of the complexity of data problems and the deep logic be⁃ hind them has important theoretical and practical guiding significance for grasping the core crux of scientific research data management and improving the scientific research data management performance. [Method/process] This paper uses Bourdieu's "field- habitus" theory to break through the limitation of traditional binary opposition, and aims to deeply reveal the practical problems of data power, subject behavior and resource allocation in scientific research data management.By selecting the retraction papers of periodicals included by WOS and Retraction Watch as the analysis objects, and using statistical analysis, comparative analysis and scientometrics methods, it is found that the current sci⁃ entific research data management problems in China were generally showing a trend of excessive growth, variable type, complexity hierarchy, intertwined reasons, plural subject, and wide distribution. [Result/conclusion] The results show that, data problem mapping scientific research data management value deviation, and the differential mode of associa⁃ tion in scientific research data management subject、the poor data management practices habitus、profit-seeking behav⁃ ior、absence of policy supervision. Therefore, the scientific research data management in China should address both symptoms and root causes. It is necessary not only to effectively optimize the academic evaluation system, improve the management system of scientific research integrity, but also to balance the incentive and distribution pattern of scientif⁃ ic research data management, improve data literacyand the scientific research data management system.