NPS Provider Choice Tanzania

The dataset integrates four rounds of the Tanzania National Panel Survey (2008/09, 2010/11, 2012/13, and 2014/15) into a harmonized Uniform Panel Dataset (UPD) with 65,954 outpatient episodes reported by households between 2008 and 2015. Each record represents a healthcare visit within a four-week recall period, detailing facility type, cost, perceived quality, and socio-economic attributes. The data, collected by the National Bureau of Statistics in collaboration with the World Bank LSMS–ISA program, are nationally representative of Tanzania Mainland and Zanzibar. The study tests the hypothesis that healthcare provider choice in Tanzania is jointly determined by household resource allocation decisions and provider strategies shaped by fiscal and infrastructural constraints. Patients allocate scarce resources—money, time, and trust—between public and non-public providers, while providers adjust prices, quality, and capacity in response to demand and regulation. Using this dataset, a fixed-effects control function panel logit model estimates causal demand responses, which are then embedded in a dynamic Bayesian game framework to capture provider behaviour. The dataset contains variables on facility characteristics (ownership, distance, waiting time), costs (consultation, medication, transport), quality perceptions (satisfaction, drug availability, cleanliness), and household factors (income, education, insurance status, residence). Data were harmonized and cleaned to ensure consistency and representativeness across four survey waves, resulting in observations from 3,152 households across all administrative regions. This dataset is valuable for analyzing healthcare demand, equity, and provider competition in Tanzania and other LMICs. It can support econometric and game-theoretic modeling, policy simulations, and evaluation of health financing reforms. All variables are coded for standard software (Stata, R, Python) and accompanied by a detailed codebook. The data are ethically compliant and publicly available through the Tanzania NPS (LSMS–ISA).本数据集将坦桑尼亚国家面板调查（Tanzania National Panel Survey, TNPS）四轮调研数据（2008/09、2010/11、2012/13及2014/15）整合为标准化统一面板数据集（Uniform Panel Dataset, UPD），涵盖2008至2015年间家庭报告的65954例门诊就诊事件。每条记录对应一个四周回忆调查周期内的一次就医行为，详细记录了医疗机构类型、就诊成本、感知就医质量以及社会经济属性等信息。本数据集由坦桑尼亚国家统计局与世界银行生活水平测量研究-综合农业调查（Living Standards Measurement Study - Integrated Surveys on Agriculture, LSMS–ISA）项目合作采集，样本覆盖坦桑尼亚本土与桑给巴尔地区，具备全国代表性。本研究依托该数据集验证了一项假设：坦桑尼亚的医疗服务提供者选择，是由家庭资源分配决策与受财政及基础设施约束的服务提供方策略共同决定的。患者需在公立与非公立医疗服务机构之间分配自身有限的资源——金钱、时间与信任；而服务机构则会根据市场需求与监管要求，调整其定价、服务质量与接诊能力。研究人员借助本数据集，采用固定效应控制函数面板Logit模型（fixed-effects control function panel logit model）估算因果性就医需求响应，并将该结果嵌入动态贝叶斯博弈框架以刻画服务提供方的行为模式。本数据集包含以下维度的变量：医疗机构特征（所有权属性、距离、等待时长）、就诊成本（问诊费、药费、交通成本）、就医质量感知（满意度、药品可及性、环境卫生状况）以及家庭层面因素（收入、受教育水平、医保参保状态、居住地区）。为确保四轮调研数据的一致性与代表性，研究团队对数据进行了标准化整合与清洗处理，最终得到覆盖坦桑尼亚所有行政区域的3152个家庭的观测样本。本数据集对于分析坦桑尼亚及其他中低收入国家（Low- and Middle-Income Countries, LMICs）的医疗需求、医疗公平性与服务机构竞争格局具有重要价值。该数据集可支持计量经济学与博弈论建模、政策模拟以及医疗筹资改革评估等研究工作。所有变量均针对标准统计软件（Stata、R、Python）完成编码，并附带详细的变量编码手册。本数据集符合伦理规范，可通过坦桑尼亚国家面板调查（Tanzania National Panel Survey, TNPS, LSMS–ISA）项目公开获取。