Guide to Conducting Economics Experiments | 经济学实验操作指南

📚 Guide to Conducting Economics Experiments | 经济学实验操作指南

Economics experiments offer a powerful way to test theories about human decision-making, market behaviour, and policy effects. Whether you are designing a classroom activity for IB Economics or collecting primary data for a WJEC coursework investigation, understanding the fundamentals of experimental method is essential. This guide walks you through every stage – from framing a testable hypothesis to analysing results and reporting findings.

经济学实验是检验人类决策、市场行为和政策效果理论的强有力工具。无论你是在为 IB 经济学设计课堂活动，还是为 WJEC 课程作业收集一手数据，掌握实验方法的基本原理都至关重要。本指南将带你走过每一个阶段——从构建可检验的假设，到分析结果和撰写报告。

1. What is an Economics Experiment? | 什么是经济学实验？

An economics experiment is a controlled setting in which participants make real choices, often involving monetary incentives, to reveal causal relationships. Unlike natural observation, experiments allow researchers to manipulate one variable (independent) while measuring its effect on another (dependent), holding other factors constant.

经济学实验是一种受控环境，参与者在其中做出真实选择，通常涉及金钱激励，以揭示因果关系。与自然观察不同，实验允许研究人员操纵一个变量（自变量），同时测量其对另一个变量（因变量）的影响，并保持其他因素不变。

Key features include random assignment to treatment and control groups, standardised instructions, and performance-based payments. This methodology has been used to study topics from auction behaviour to the impact of minimum wages.

关键特征包括随机分配到处理组和对照组、标准化指令以及基于表现的报酬。这种方法已被用于研究从拍卖行为到最低工资影响的各种课题。

2. Designing a Research Question | 设计研究问题

Start with a clear, focused question that can be answered through experimental data. For example, ‘Does a subsidy on healthy food increase consumption among students?’ Good questions are specific, measurable, and linked to economic theory.

从一个清晰、聚焦的问题开始，这个问题应能通过实验数据回答。例如，“对健康食品提供补贴是否会增加学生的消费量？”好的问题应具体、可量化，并与经济理论相联系。

Base your question on a behavioural anomaly or a policy gap. In IB commentaries or WJEC investigations, grounding the experiment in real-world context strengthens analysis. A hypothesis should predict the direction of change, such as ‘A 20% subsidy will raise demand for fruit by at least 15% compared to the control group.’

将问题建立在行为异常或政策缺口的基础上。在 IB 评论或 WJEC 调查中，将实验置于现实世界背景中会加强分析。假设应预测变化的方向，例如“20% 的补贴将使水果需求比对照组至少提高 15%”。

3. Identifying Variables | 识别变量

In experimental economics, the independent variable is the treatment you manipulate, such as price, information, or incentive structure. The dependent variable is the outcome you measure, such as quantity purchased, cooperation rate, or efficiency.

在实验经济学中，自变量是你操纵的处理因素，如价格、信息或激励结构。因变量是你测量的结果，如购买数量、合作率或效率。

Control variables are factors that must be held constant across sessions, including timing, payment method, and participant demographic profile. Confounding variables, like news events or weather, must be minimised by randomisation and consistent procedures.

控制变量是必须在各场次中保持恒定的因素，包括时间安排、支付方式和参与者人口学特征。混杂变量，如新闻事件或天气，必须通过随机化和一致的程序最小化。

4. Laboratory vs. Field Experiments | 实验室实验与实地实验

Laboratory experiments take place in a controlled environment, often a university lab, where participants interact via computers. They offer high internal validity because you can isolate causal mechanisms, but they may lack realism.

实验室实验在受控环境中进行，通常是在大学实验室，参与者通过计算机互动。它们具有很高的内部效度，因为你可以分离因果机制，但可能缺乏现实性。

Field experiments occur in natural settings, such as a supermarket or a classroom, where subjects may not even know they are in an experiment. They strengthen external validity and policy relevance, but controlling for extraneous factors is harder.

实地实验在自然环境中进行，如超市或教室，受试者甚至可能不知道自己正在参与实验。它们增强了外部效度和政策相关性，但控制外部因素更难。

For IB and WJEC students, a classroom experiment with a small incentive is a practical compromise: it mimics a lab setting but uses peer groups as participants, making it feasible and ethically straightforward.

对于 IB 和 WJEC 学生，一个带有小额激励的课堂实验是实用的折中方案：它模拟了实验室设置，但使用同伴群体作为参与者，使其可行且在伦理上简单明了。

5. Participant Recruitment and Incentives | 参与者招募与激励

Recruit a sample large enough to detect an effect; a minimum of 30 participants per treatment arm is often recommended. Use random assignment to treatment and control groups to avoid selection bias.

招募足够大的样本以检测效应；通常建议每个处理组至少 30 名参与者。使用随机分配至处理组和对照组以避免选择偏差。

Incentives are central to economic experiments. Pay should be linked to decisions to elicit true preferences—for instance, participants keep the goods they choose or receive a cash bonus based on performance. Even small real payments outperform hypothetical questions.

激励是经济学实验的核心。报酬应与决策挂钩，以引出真实偏好——例如，参与者可以保留他们选择的商品，或根据表现获得现金奖励。即使是小额的真实报酬也优于假设性问题。

In classroom settings, course credits or small gifts can serve as incentives, provided they are meaningful for effort. Always obtain informed consent and debrief participants after the session.

在课堂环境中，课程学分或小礼品可作为激励，只要它们对努力有意义。务必获得知情同意，并在实验结束后向参与者进行情况通报。

6. Writing Experimental Instructions | 编写实验指令

Instructions must be clear, neutral, and identical for all participants. Use simple language, avoid leading terms, and include examples. A comprehension quiz before the main task helps ensure understanding.

指令必须清晰、中立，并对所有参与者一致。使用简单的语言，避免引导性术语，并附上示例。在主任务前进行一次理解测验有助于确保理解。

For an experiment on consumer behaviour, you might write: ‘You have an endowment of £5. You will see a series of products at different prices. Please indicate how many units you wish to buy at each price. You will actually purchase one randomly selected decision.’ This makes choices consequential.

对于消费者行为实验，你可以这样写：“你有 5 英镑的初始禀赋。你将看到一系列不同价格的产品。请指明在每个价格下你希望购买多少单位。你将实际购买一个随机选定的决策。”这使得选择具有后果。

Instructions should explain how payoffs are calculated, the randomisation procedure, and privacy protections. Script everything to avoid experimenter demand effects.

指令应解释收益如何计算、随机化程序以及隐私保护。将所有内容脚本化，以避免实验者需求效应。

7. Running the Experiment Session | 进行实验环节

Arrange the physical or online setting to minimise distractions. If using paper and pencil, distribute decision sheets face down. Digital platforms (e.g., oTree, Qualtrics) standardise presentation and timing.

安排好实体或线上环境，以最大限度减少干扰。若使用纸笔，将决策表朝下分发。数字平台（如 oTree、Qualtrics）可标准化呈现和时间控制。

Start with a practice round, giving participants a risk-free opportunity to understand the interface. During the actual rounds, monitor for technical glitches and do not intervene unless absolutely necessary—every session should run under identical conditions.

从一个练习轮次开始，让参与者有无风险的机会理解界面。在实际轮次中，监控技术故障，除非绝对必要，不要干预——每一场次都应在相同条件下运行。

Keep a log of any unexpected events (e.g., a fire alarm, a participant asking a leading question) because these can affect data interpretation. This is crucial for IB essays and WJEC evaluation marks.

记录任何意外事件（例如，火警、参与者提出引导性问题），因为这些可能会影响数据解读。这对于 IB 论文和 WJEC 评估分数至关重要。

8. Data Collection and Ethical Practice | 数据收集与伦理实践

Collect data on choices, reaction times, and demographic characteristics where relevant. Anonymise all records immediately by assigning ID codes. Store data securely in password-protected files.

收集有关选择、反应时间和相关人口学特征的数据。立即通过分配 ID 代码对所有记录进行匿名化处理。将数据安全地存储在密码保护的文件中。

Ethical principles require that participants are not deceived unless methodologically essential and approved. In IB and WJEC work, full disclosure and the right to withdraw without penalty are mandatory. Debriefing should explain the purpose of the experiment and how data will be used.

伦理原则要求除非在方法论上至关重要且获得批准，否则不欺骗参与者。在 IB 和 WJEC 作业中，完全公开和不受惩罚的退出权是强制性的。情况通报应解释实验目的以及数据将如何使用。

If you are studying sensitive topics like discrimination or risk-taking, additional safeguards—such as counselling referral or pre‑screening questionnaires—may be needed. Your teacher can guide you through the institutional review process.

如果你研究的是歧视或冒险行为等敏感话题，可能需要额外的保护措施——例如咨询转介或预筛选问卷。你的老师可以指导你完成机构审查流程。

9. Analysing Experimental Data | 分析实验数据

Start with descriptive statistics: means, medians, standard deviations for each group. Visualise data with bar charts or box plots to spot patterns. For many classroom experiments, a simple comparison of group means suffices.

从描述性统计开始：各组的平均数、中位数、标准差。使用条形图或箱线图可视化数据以发现规律。对于许多课堂实验来说，简单的组间均值比较就足够了。

To draw causal conclusions, apply inferential tests. A common approach is the two‑sample t‑test comparing treatment and control averages. For non-normal data or small samples, use a Mann–Whitney U test. Report p‑values and effect sizes in plain language: ‘The treatment increased average contributions by 2.3 units (p = 0.03, Cohen’s d = 0.55).’

要得出因果结论，应用推断检验。常见的方法是比较处理组和对照组均值的双样本 t 检验。对于非正态数据或小样本，使用 Mann–Whitney U 检验。用简明的语言报告 p 值和效应量：“处理使平均贡献增加了 2.3 个单位（p = 0.03，Cohen’s d = 0.55）。”

Always check that assumptions of the chosen test are met. If participants’ decisions are interdependent (e.g., in a public-goods game), use cluster‑robust standard errors or session-level analysis to avoid inflated significance.

务必检查所选检验的假设是否得到满足。如果参与者的决策相互依赖（例如，在公共品博弈中），使用聚类稳健标准误或场次层面分析，以避免显著性被夸大。

10. Common Biases and How to Control Them | 常见偏差及其控制方法

Experimenter demand effects occur when participants infer the hypothesis and alter behaviour. Mitigate this by using neutral language, double‑blind procedures, and by not revealing the exact hypothesis until debriefing.

实验者需求效应发生在参与者推断出假设并改变行为时。通过使用中性语言、双盲程序以及不在通报前披露确切假设来减轻这种效应。

Selection bias arises if people who volunteer are systematically different. True random assignment from a homogeneous pool reduces this risk. Attrition bias, when certain types drop out, requires tracking dropout patterns and using intention‑to‑treat analysis.

如果自愿参与的人具有系统性差异，就会产生选择偏差。从同质群体中进行真正的随机分配可以降低这一风险。损耗偏差，即某些类型的人退出，需要跟踪退出模式并使用意向性治疗分析。

Order effects (learning or fatigue) can be balanced by counterbalancing treatments across sessions. Where feasible, replicate the experiment in a second block of time to confirm reliability. Mention these controls in your evaluation to gain higher marks in IB and WJEC.

顺序效应（学习或疲劳）可以通过在场次间平衡处理来抵消。在可行的情况下，在第二个时间段复制实验以确认可靠性。在评估中提及这些控制措施，以在 IB 和 WJEC 中获得更高分数。

11. Application in IB and WJEC Economics | 在 IB 和 WJEC 经济学中的应用

For IB Economics, experiments can enrich your internal assessment (IA) commentaries. Although IAs are primarily based on news articles, you may conduct a supplementary mini‑experiment to test a concept like price elasticity or nudge effectiveness. This shows initiative and strengthens evaluation.

对于 IB 经济学，实验可以丰富你的内部评估（IA）评论。尽管 IA 主要基于新闻文章，但你可以进行一个补充性微型实验，以检验价格弹性或助推有效性等概念。这展现了主动性，并加强了评估。

WJEC Eduqas AS/A‑Level Economics includes non‑examined assessment (NEA) components where students design and carry out an independent investigation. An experiment—field or lab—provides original quantitative data, which is highly valued. Ensure your investigation question stems from the specification, such as behavioural economics biases or labour market signalling.

WJEC Eduqas AS/A‑Level 经济学包括非考试评估（NEA）部分，学生需设计并开展独立调查。一个实验——无论是实地还是实验室——可以提供原始的定量数据，这非常受重视。确保你的调查问题源自课程大纲，例如行为经济学偏差或劳动力市场信号。

In both programmes, link your experimental findings to economic theory: compare outcomes with model predictions, and discuss policy implications. For instance, if your experiment shows that higher information transparency reduces price dispersion, relate this to market efficiency and consumer welfare.

在这两个项目中，将你的实验结果与经济理论联系起来：将结果与模型预测进行比较，并讨论政策含义。例如，如果你的实验显示更高的信息透明度减少了价格离散，请将其与市场效率和消费者福利联系起来。

12. Writing the Experimental Report | 撰写实验报告

Structure your report with standard sections: Introduction (research question and background), Method (design, participants, procedure), Results (descriptive and inferential statistics), Discussion (interpretation, limitations, implications), and References.

按照标准章节构建你的报告：引言（研究问题和背景）、方法（设计、参与者、程序）、结果（描述性和推断性统计）、讨论（解读、局限性、含义）和参考文献。

In the Discussion, explicitly address internal and external validity. Could the results be driven by artificial lab conditions? How generalisable are they? For WJEC NEA, the evaluation criteria reward critical reflection on methodology. For IB, the evaluation section of your IA can incorporate these reflections if you use experimental evidence to support your commentary.

在讨论中，明确说明内部效度和外部效度。结果是否可能由人为的实验室条件造成？它们有多大程度的可推广性？对于 WJEC NEA，评价标准奖励对方法论进行批判性反思。对于 IB，如果你使用实验证据支持你的评论，你的 IA 评价部分可以纳入这些反思。

Use tables to summarise key findings. For example:

使用表格总结关键结果。例如：

Group 组别	Mean Contribution (units) 平均贡献（单位）	Std Dev 标准差	N 样本量
Treatment 处理组	14.2	3.8	35
Control 对照组	10.5	4.1	35

t = 2.61, p = 0.011, Cohen’s d = 0.88

Conclude by restating the main result, its economic meaning, and suggestions for future research. A well‑executed experiment not only boosts your grade but also builds analytical skills valued in university economics courses.

在结尾处重申主要结果、其经济意义和对未来研究的建议。一个执行良好的实验不仅能提高你的成绩，还能培养大学经济学课程所看重的分析技能。

Published by TutorHao | Economics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）