Replication Data for: "Digital Exclusion and ICT4Ag Adoption Among Smallholder Farmers in Puntland, Somalia: Structural Gatekeepers, Awareness–Usage Disparities, and Institutional Pathways for Inclusive Digital Agriculture"

Abstract & Data Package Overview:This repository contains the complete replication package, including the primary micro-dataset, comprehensive variable definitions register, technical analysis workflows, and multi-language analytical source codes (R, Python, and Stata) supporting the manuscript titled "Digital Exclusion and ICT4Ag Adoption Among Smallholder Farmers in Puntland, Somalia: Structural Gatekeepers, Awareness–Usage Disparities, and Institutional Pathways for Inclusive Digital Agriculture." Context & Geographic Focus:The data was generated through an empirical cross-sectional survey executed across three distinct agronomic and administrative zones in Puntland State, Somalia: Bosaso (urban/maritime trading hub), Garowe (administrative capital region), and Carmo (peri-urban/rural agricultural zone). The underlying targeted population captures N=110 smallholder farmers, cooperative leaders, and regional agronomic stakeholders surveyed between October and December. Empirical Scope & Variables Trapped:The primary records map comprehensive socioeconomic markers alongside digital divide variables, including:1. Demographic Indicators: Age distributions, gender balances, formal educational attainment brackets, and dominant agricultural production systems (crop, livestock, or mixed farming models).2. Infrastructure Access Metrics: Distinctions between basic voice/SMS feature-phone dependence and active smartphone hardware ownership, cellular data connectivity profiles, and primary stable electricity power configurations.3. The Awareness-Usage Disparity Ledger: Granular usage data tracking the difference between broad financial tech awareness (e.g., mobile money platforms) and functional deployment of specialized agtech software channels (e.g., micro-climate apps, digital crop marketplaces, and specialized farm management interfaces).4. Structural Barrier Severity Matrices: 5-point Likert scale arrays assessing specific resource and accessibility gatekeepers, ranked by severity from critical to moderate (device costs, internet charges, local training limitations, lack of Somali semantic localization, and signal instability).5. Institutional Governance Frameworks: Evaluations regarding the effectiveness of public and decentralized agricultural extension support networks. Folder Inventory and Replication Guidance:To ensure continuous computational transparency, this repository packages four functional operational files:- 'Puntland_Agri_Barriers_dataset.csv': The fully anonymized, de-identified primary data file comprising exactly 110 clean rows. All personally identifiable information (PII) and household GPS track data have been fully scrubbed out.- 'Data_Dictionary.txt': The structural register defining exact column names, attribute limits, categorical encoding levels, scale specifications, and sample properties.- 'Statistical_Analysis_Workflow.txt': A clean human-readable step-by-step procedural log breaking down dataset execution across six processing phases (Pre-processing, Descriptive Profiling, Reliability testing, Bivariate Chi-Square analysis, Multivariate Logistic Regression, and Graphic export).- Source Code Automation Scripts: 'Analysis_Script_R_ggplot2.R' (featuring integrated package checks and high-resolution 300 DPI visualization models), 'Analysis_Script_Python_Stats.py' (executing Cronbach's alpha internal consistency metrics and statsmodels multivariate logit modeling), and 'Analysis_Script_Stata.do' (Stata 17 layout utilizing automated log capturing and outreg2 outputs).摘要与数据包概览：本仓库包含完整的复现数据包，涵盖支撑题为《索马里邦特兰州小农的数字排斥与信息通信技术赋能农业（ICT4Ag）采纳：结构性守门人、认知-使用差距与包容性数字农业的制度路径》的学术论文所需的核心微观数据集、详尽的变量定义总表、技术分析流程以及多语言分析源代码（支持R、Python与Stata语言）。 研究背景与地理聚焦：本数据集源自针对索马里邦特兰州三个各具特色的农艺与行政区开展的实证横断面调查：博萨索（城市/海运贸易枢纽）、加罗韦（行政首都区域）以及卡尔莫（城郊/农村农业区）。本次调查于10月至12月间开展，共纳入110名目标受访者，包括小农、合作社负责人以及区域农艺利益相关者。 实证研究范畴与变量清单：核心数据记录涵盖了全面的社会经济标志物与数字鸿沟相关变量，具体包括： 1. 人口统计指标：年龄分布、性别结构、正规教育程度层级以及主导农业生产模式（种植业、畜牧业或混合经营模式）。 2. 基础设施接入指标：基础语音/短信功能机依赖与活跃智能手机持有情况的区分、蜂窝数据连接情况以及主要稳定电力供应配置。 3. 认知-使用差距台账：追踪广义金融科技认知（如移动货币平台）与专业农业科技（agtech）软件渠道功能性使用情况之间差异的精细化使用数据，例如微气候应用、数字农产品交易平台以及专用农场管理界面。 4. 结构性障碍严重程度矩阵：采用5级李克特量表对特定资源与可及性守门因素进行评估，按严重程度从高到低排序（设备成本、互联网资费、本地培训不足、索马里语语义本地化缺失以及信号不稳定）。 5. 制度治理框架：针对公共与去中心化农业推广支持网络的有效性开展的评估。 文件夹清单与复现指南：为确保计算过程全程透明，本仓库包含四类功能性操作文件： - 「Puntland_Agri_Barriers_dataset.csv」：完全匿名去标识化的核心原始数据文件，共包含110条干净记录。所有个人可识别信息（PII）与家庭GPS轨迹数据均已彻底清除。 - 「Data_Dictionary.txt」：结构性变量定义总表，明确标注了各列名称、属性限值、分类编码层级、量表规格以及样本特征。 - 「Statistical_Analysis_Workflow.txt」：清晰易读的分步程序日志，将数据集处理流程划分为六个阶段：预处理、描述性分析、信度检验、二元卡方分析、多元逻辑回归以及图形导出。 - 源代码自动化脚本：包括「Analysis_Script_R_ggplot2.R」（集成了包依赖检查与300 DPI高分辨率可视化模型）、「Analysis_Script_Python_Stats.py」（用于执行克朗巴哈α系数内部一致性指标检验以及statsmodels库的多元logit建模）以及「Analysis_Script_Stata.do」（采用Stata 17格式，支持自动日志捕获与outreg2输出）。