Hot hand experimental data

This dataset corresponds to the raw data from the research study “The Hot-Hand Belief and the Loss-Averse Investor.” The experimental protocol was approved by two Research Ethics Committees: the University of Brasília (UnB), under technical decision No. 2.558.540, and the Federal University of Tocantins (UFT), under technical decision No. 2.850.741. A total of 226 undergraduate students from various majors at UFT and UnB were recruited to participate in the study. After data cleaning, the final sample comprised 208 individuals.

The original data file, RawDatahh.rds, is stored in RDS format—a binary file structure native to the R programming language. This format enables efficient storage of complex R objects, such as lists and data frames, while preserving internal structures including labels, data types, and hierarchies. Specifically, the RDS file contains: (i) the global configuration parameters that define the rules and structure of the experimental game; (ii) detailed behavioral data on participant decisions across multiple periods; and (iii) the full set of simulated price series for the financial assets used in the experiment.

For ease of use in statistical analysis and replication, a cleaned and reformatted version of the raw data is provided in CSV format as subset_RawDatahh.csv. This subset retains key variables capturing participants’ choices, asset prices, the monetary values of purchased shares, and the alpha indices estimated in the study.

The dataset subset_RawDatahh.csv is structured as panel data, with observations for multiple participants across several time periods. It includes numeric variables on the quantities of assets purchased, corresponding prices, monetary amounts paid per purchase, and the indices $\theta_1^i$ and $\theta_2^i$ tested under Hypothesis 1 of the study.

This version of the dataset also includes two key supplementary materials to support transparency and usability. First, it provides the original experimental instructions that were delivered to participants during the laboratory sessions, offering full insight into the procedures, decision tasks, and game mechanics involved in the experiment. These instructions are available in PDF format and reflect the exact wording and layout used in the sessions. Second, it includes a comprehensive user guide that outlines the structure of the dataset, defines the core variables, and provides practical instructions for importing, managing, and analyzing the data in R.

本数据集对应研究论文《热手信念与损失厌恶投资者》（The Hot-Hand Belief and the Loss-Averse Investor）的原始实验数据。本研究的实验方案已获两所研究伦理委员会批准：巴西利亚大学（University of Brasília, UnB），技术决议编号2.558.540；以及托坎廷斯联邦大学（Federal University of Tocantins, UFT），技术决议编号2.850.741。研究共招募了来自UFT与UnB多个专业的226名本科生参与。经数据清洗后，最终有效样本量为208名被试。

原始数据文件"RawDatahh.rds"采用RDS格式存储——这是R语言原生的二进制文件结构，可高效存储列表、数据框等复杂R对象，并保留标签、数据类型与层级结构等内部属性。具体而言，该RDS文件包含三部分内容：(i) 定义实验博弈规则与结构的全局配置参数；(ii) 被试多回合决策的详细行为数据；(iii) 实验中使用的金融资产的全套模拟价格序列。

为便于统计分析与结果复现，本数据集同时提供经清洗与重格式化的逗号分隔值（Comma-Separated Values, CSV）格式版本"subset_RawDatahh.csv"。该数据集保留了核心变量，包括被试选择、资产价格、买入股份的货币价值，以及本研究中估算的alpha指数。

数据集"subset_RawDatahh.csv"为面板数据结构，包含多名被试跨多个时间段的观测值。其包含的数值型变量包括资产购买数量、对应交易价格、每笔购买的支付金额，以及本研究假说1中检验的$oldsymbol{ heta_1^i}$与$oldsymbol{ heta_2^i}$指数。

本数据集版本还包含两项关键辅助材料，以提升研究透明度与易用性。其一，提供实验实验室环节中向被试发放的原版实验指导手册，完整呈现实验流程、决策任务与博弈机制，该文件采用PDF格式，完全还原了实验环节中的措辞与排版。其二，包含一份详尽的用户指南，概述数据集结构、定义核心变量，并提供在R中导入、管理与分析该数据的实操指引。