📚 GCSE WJEC Business: Guide to Experimental Research | GCSE WJEC 商务：实验操作指南
Experiments are a powerful tool in business research, allowing firms to test ideas under controlled conditions before committing large resources. Whether launching a new product, changing a price, or tweaking an advertisement, an experimental approach provides evidence rather than guesswork. This guide breaks down everything you need to know about designing, carrying out, and interpreting experiments for your WJEC GCSE Business studies, equipping you with the skills to evaluate real-world business decisions and excel in your exam.
实验是商业研究中一种强大的工具，能让企业在投入大量资源之前，在受控条件下测试想法。无论是推出新产品、调整价格还是修改广告，实验方法都能提供证据而不靠猜测。本指南详细解析了 WJEC GCSE 商务课程中你需要掌握的实验设计、实施和结果解读的全部内容，帮助你评估现实商业决策并在考试中脱颖而出。
1. What Is an Experiment in Business? | 什么是商务实验？
In business, an experiment is a research method where one or more factors (variables) are deliberately changed to observe the effect on another factor, while keeping all other conditions the same. Unlike surveys or interviews, experiments seek to establish cause-and-effect relationships. For example, a retailer might change the background music in one store but not in another, then compare sales figures to see if the music influenced customer spending.
在商务领域，实验是一种研究方法，即故意改变一个或多个因素（变量），观察其对另一个因素的影响，同时保持其他条件不变。与调查或访谈不同，实验旨在建立因果关系。比如，零售商可能在某一店铺改变背景音乐，而另一店铺不做改变，然后比较销售数据，判断音乐是否影响了顾客消费。
Experiments can be conducted in a laboratory setting, where researchers create an artificial environment, or in the field, where they use real-life settings such as shops, offices, or online platforms. Business students need to understand that experiments are not just for scientists in white coats — marketers, HR managers, and operations directors all use experimental thinking to test innovations.
实验可以在实验室环境中进行，研究者创造一个受控的人为环境；也可以在实地进行，利用真实的场景如商店、办公室或在线平台。商科学生需要明白，实验不仅仅是穿白大褂的科学家的专属——市场营销人员、人力资源经理和运营总监都会运用实验思维来检验创新。
2. Types of Experiments: Laboratory, Field, and Natural | 实验类型：实验室实验、实地实验与自然实验
Laboratory experiments take place in a controlled, artificial setting where the researcher can manipulate the independent variable precisely and eliminate confounding factors. In business, this might involve using computer simulations or mock shops. They offer high internal validity because extraneous variables can be held constant, but the artificial context may limit how well findings apply to the real world (low external validity).
实验室实验在受控的人为环境中进行，研究者可以精确操控自变量并消除混杂因素。在商务中，这可能涉及使用计算机模拟或模拟商店。它们具有较高的内部效度，因为可以保持无关变量不变；但人为环境可能限制研究结果在现实中的推广（外部效度低）。
Field experiments occur in natural, everyday surroundings. A company might test a new promotional offer in a few branches while leaving others unchanged. Participants often do not know they are in an experiment, so behaviour is more authentic. The downside is that it is harder to control all variables, meaning other factors (like a local festival) could affect the results.
实地实验发生在自然的日常环境中。公司可能在几家分店测试新促销方案，而其他分店保持不变。参与者通常不知道自己正在被实验，因此行为更为真实。缺点是难以控制所有变量，意味着其他因素（如当地节日）可能影响结果。
Natural experiments are not truly designed by the researcher; instead, they take advantage of changes that happen naturally, such as a government policy shift or a sudden supply shortage. The researcher compares data before and after the event. These can be very useful for business analysis, but the lack of deliberate manipulation means causality is harder to prove.
自然实验并非由研究者主动设计，而是利用自然发生的变化，如政府政策改变或突然的供应短缺。研究者比较事件前后的数据。这对商业分析非常有用，但由于缺乏刻意操控，因果关系较难证明。
3. Key Features of a Good Experiment | 成功实验的关键特征
Every well-designed experiment aims to isolate the impact of one factor. To achieve this, you must identify the independent variable (the one you change), the dependent variable (the one you measure), and any control variables (constants). For example, if testing whether a loyalty card increases customer visits, the loyalty card is the independent variable, the number of visits is the dependent variable, and factors like store location, opening hours, and competitor activity should be controlled.
每个精心设计的实验都旨在分离某一因素的影响。要做到这一点，必须确定自变量（你改变的变量）、因变量（你测量的变量）以及各类控制变量（保持不变的量）。例如，测试会员卡是否能增加顾客光顾次数时，会员卡是自变量，到店次数是因变量，而店铺位置、营业时间和竞争者活动等应被控制。
Reliability and validity are central. Reliability means that if you repeat the experiment under the same conditions, you would get similar results. Validity refers to whether the experiment truly measures what it intends to measure. A business experiment might be reliable but lack validity if, say, the loyalty card effect is actually caused by a discount introduced at the same time.
信度和效度至关重要。信度意味着如果在相同条件下重复实验，会得到相似的结果。效度则指实验是否真正测量了它想要测量的东西。一个商务实验可能信度高但缺乏效度，例如，会员卡的效果实际上是由同时推出的折扣引起的。
4. Hypothesis Formulation and Variables | 假设提出与变量设定
Before running an experiment, you must state a clear hypothesis — a testable prediction about the relationship between variables. It is usually written as ‘If… then…’ or as a comparative statement: ‘Customers who receive a personalised email will spend 10% more than those who receive a generic one.’ A null hypothesis (H₀) assumes no effect, while the alternative hypothesis (H₁) proposes there is an effect.
在开展实验前，你必须提出明确的假设——一个关于变量间关系的可检验的预测。通常表述为“如果……那么……”或比较性陈述：“收到个性化邮件的顾客将比收到通用邮件的顾客多消费10%。”零假设（H₀）假定没有效果，备择假设（H₁）则提出有某种效果。
You also need to operationalise your variables. Operationalisation means defining exactly how a variable will be measured or manipulated. For ‘customer satisfaction’, you might use a score from 1 to 10 from a survey. Without clear definitions, replication becomes impossible and results become ambiguous.
你还需要将变量操作化。操作化是指精确定义如何测量或操控一个变量。对于“顾客满意度”，你可能采用问卷中1到10分的评分。没有清晰的定义，就无法复制实验，结果也会模棱两可。
5. Designing an Experiment: Control Groups and Randomisation | 实验设计：对照组与随机化
An effective experiment compares an experimental group (which receives the treatment) against a control group (which does not). The control group provides a baseline, showing what would happen without the intervention. In business, a control group might be a set of customers who do not see a new advert, while the experimental group does.
一个有效的实验会设置实验组（接受处理）与对照组（不接受处理）进行比较。对照组提供基线，展示在没有干预的情况下会发生什么。在商务中，对照组可能是一组没有看到新广告的顾客，而实验组则看到了。
Randomisation means assigning participants to groups purely by chance. This reduces bias and helps ensure any differences observed are due to the independent variable, not pre-existing differences among people. For instance, if you let volunteers choose which group to join, you might end up with more tech-savvy customers in the experimental group, skewing results.
随机化意味着纯粹靠偶然将参与者分配到不同组。这减少了偏差，有助于确保观察到的任何差异都源于自变量，而非人们原本的差异。例如，如果让志愿者自己选择加入哪一组，可能会使实验组里技术娴熟的顾客偏多，从而扭曲结果。
6. Data Collection and Measurement in Experiments | 实验中的数据收集与测量
Data during an experiment can be collected through digital analytics, tills receipts, observation, questionnaires, or interviews. Before-and-after measurements strengthen the design. A pretest-posttest approach measures the dependent variable both before and after introducing the independent variable, clearly revealing any change. In a retail experiment, sales figures for the four weeks before and after a shelf display change would be compared.
实验期间的数据可以通过数字分析、收银小票、观察、问卷或访谈来收集。前后测量能加强实验设计。前测-后测方法分别在引入自变量前后测量因变量，清楚地显示任何变化。在零售实验中，可以比较货架陈列改变前后四周的销售额。
Ensure measurement tools are consistent. If different observers are used, train them to apply the same criteria. In WJEC assessments, examiners often ask about the appropriateness of data collection methods — linking back to validity and reliability.
确保测量工具一致。如果使用不同观察员，要培训他们采用相同的评判标准。在 WJEC 考试中，考官常会询问数据收集方法是否恰当——这又会联系到效度和信度。
7. Pilot Studies and Ethical Considerations | 试点研究与伦理考量
A pilot study is a small-scale trial run of the experiment design. It helps identify practical problems — such as confusing instructions, equipment failures, or unexpected costs — before the main experiment begins. For example, a supermarket testing a new layout might try it first in one small store to iron out issues.
试点研究是对实验设计进行的小规模试运行。它有助于在正式实验开始前发现实际问题，例如指令含混、设备故障或意想不到的成本。比如，一家超市测试新布局时，可以先在一家小店尝试以消除问题。
Ethics matter even in business experiments. Participants should give informed consent, their data must be kept confidential, and they must be allowed to withdraw at any time. Deception may be used only when essential and must be justified. For instance, a study effect of a loyalty scheme might involve observing customers without telling them upfront, but this raises privacy concerns that must be addressed.
即便是商业实验，伦理也至关重要。参与者应给予知情同意，数据必须保密，且他们可以在任何时候退出。欺骗仅在必要时才能使用，并且必须证明其合理性。例如，研究忠诚计划的效果可能涉及在不事先告知的情况下观察顾客，但这会引起必须解决的隐私问题。
8. Analysing Experimental Results: Mean, Mode, Median | 分析实验结果：均值、众数和中位数
Once data is collected, you need to calculate measures of central tendency. The mean (average) is likely the most common measure in business experiments:
Mean = Σx ÷ n
数据收集完成后，需要计算集中趋势量数。均值（平均数）很可能是商务实验中最常用的指标：
均值 = Σx ÷ n
The mode is the value that appears most often — useful for categorical data like the most popular product colour. The median is the middle value when data is ordered; it is less affected by extreme scores (outliers), which often appear in sales data if a few customers spend huge amounts.
众数是出现频率最高的值——适合类别数据，如最受欢迎的产品颜色。中位数是排序后处于中间位置的值；它受极端值（离群值）影响较小，而销售数据中常会有少数顾客花费极高的情况。
| Day | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
|---|---|---|---|---|---|---|---|---|---|---|
| Sales (£) | 220 | 215 | 230 | 218 | 500 | 225 | 210 | 228 | 222 | 235 |
Here, the mean is (220+215+230+218+500+225+210+228+222+235) ÷ 10 = 270.3, but the median (put in order: 210, 215, 218, 220, 222, 225, 228, 230, 235, 500) is 223.5. The outlier of £500 pulls the mean up. Business analysts therefore often report both mean and median.
这里，均值是 (220+215+230+218+500+225+210+228+222+235) ÷ 10 = 270.3，但中位数（排序后：210, 215, 218, 220, 222, 225, 228, 230, 235, 500）是 223.5。离群值 500 把均值拉高了。因此，商业分析师通常同时报告均值和中位数。
9. Practical Example: Testing a New Product | 实例：测试新产品
Imagine ‘BeaniGo’, a coffee shop chain, wants to launch a new oat-milk latte. Before a national rollout, it runs a field experiment in four stores: two experimental stores promote the latte with a ‘healthy choice’ banner, and two control stores simply add it to the menu without advertising. The independent variable is the promotional banner; the dependent variable is the number of oat-milk lattes sold per week.
设想连锁咖啡品牌“BeaniGo”想推出新款燕麦拿铁。在全面上市前，它选择四家店进行实地实验：两家实验店用“健康之选”横幅推广该拿铁，两家对照店只将其加入菜单而不做广告。自变量是促销横幅；因变量是每周销售的燕麦拿铁杯数。
After four weeks, the average weekly sales in experimental stores were 340 units, compared with 120 units in control stores. The company concluded the promotion significantly boosted interest. However, they also checked that store footfall, weather, and pricing were similar across groups, so the result was valid. This experiment saved the company from a costly nationwide campaign that might have failed without evidence.
四周后，实验店平均每周销售 340 杯，对照店为 120 杯。公司得出结论：促销大幅提高了兴趣。不过，他们还检查了各组的客流量、天气和定价是否相似，确保结果有效。这个实验让公司避免了一场可能因缺乏证据而失败的昂贵全国推广活动。
10. Advantages and Disadvantages of Experimental Research | 实验研究的优点与缺点
Experiments offer clear cause-and-effect evidence, which is invaluable for businesses wanting to optimise marketing spend, product design, or employee motivation schemes. They are replicable, so other firms can verify findings or adapt them. Additionally, many experiments can be conducted quickly online, enabling agile decision-making.
实验提供了清晰的因果证据，对于想优化营销支出、产品设计或员工激励方案的企业来说极其宝贵。它们具有可复制性，其他公司可以验证或调整研究结果。此外，许多实验可以在线上快速进行，实现敏捷决策。
On the downside, laboratory experiments lack realism, and field experiments can lose control over extraneous variables. Ethical constraints may limit manipulation — for example, you cannot deliberately worsen staff conditions just to see the effect on productivity. There is also the risk of the Hawthorne effect: when people know they are in an experiment, their behaviour changes, distorting results.
缺点是，实验室实验缺乏真实感，而实地实验又可能对无关变量失去控制。伦理限制也会制约操控——例如，不能为了观察对生产率的影响而故意恶化员工待遇。此外还存在霍桑效应的风险：当人们知道自己正被实验时，行为会改变，从而扭曲结果。
11. Applying Experiments in Real Business Decisions | 实验在实际商业决策中的应用
Businesses use experiments routinely. A/B testing on websites involves showing two versions of a page to different user segments to see which generates more clicks or sales. Streamers like Netflix test different thumbnail images to maximise viewer engagement. Retailers test store layouts, and manufacturers use trial markets to gauge demand before mass production.
企业经常使用实验。A/B 测试在网站中向不同用户群体展示两个版本的页面，看哪个带来更多点击或销售。像 Netflix 这样的流媒体公司测试不同的缩略图，以最大化观众参与度。零售商测试店面布局，制造商则用试销市场在大规模生产前评估需求。
For your GCSE Business, case studies may feature such experiments. When evaluating a business decision, you can discuss whether an experiment was the right method, whether variables were properly controlled, and what conclusions can be drawn confidently. Linking your analysis to reliability, sample size, and possible bias will score high marks.
在 GCSE 商务课程中，案例研究可能会涉及这类实验。在评估商业决策时，你可以讨论实验是否是正确的方法，变量是否被妥善控制，以及可以自信地得出什么结论。将分析与信度、样本量和可能的偏差联系起来会得到高分。
12. Common Pitfalls and How to Avoid Them | 常见陷阱及规避方法
Confounding variables are the greatest threat. For instance, if a store’s sales rise after a price cut experiment, but a competitor closed down at the same time, the competitor’s closure is a confounding variable. Always list potential confounders in the planning stage and try to control them, or at least acknowledge them in the evaluation.
混淆变量是最大的威胁。例如，如果降价实验后一家店的销售额上升，但同时有个竞争对手倒闭了，那么竞争对手的倒闭就是一个混淆变量。务必在规划阶段列出潜在的混淆因素，并设法控制，或至少在评估中承认它们的存在。
Sampling bias occurs when the experimental and control groups are not comparable. Random allocation helps, but with small samples — common in niche businesses — differences may persist. A low sample size also makes it harder to detect clear effects and reduces generalisability. Pilot studies can help estimate required sample sizes.
抽样偏差发生在实验组和对照组不可比的情况下。随机分配有所助益，但在小众企业的通常小样本中，差异仍可能持续。低样本量还会使明确效果的检测更加困难，并降低推广性。试点研究有助于估算所需的样本量。
Finally, beware of the research bias — the experimenter may unconsciously influence results or interpret data to support a pre‑existing belief. Keeping the data collection procedure standardised and, where possible, blinded (not knowing which group received treatment) prevents this.
最后，警惕研究偏差——实验者可能无意中影响结果，或为支持先入之见而解读数据。将数据收集程序标准化，并在可能情况下实施盲法（不知道哪组接受了处理）可以防止这一点。
Published by TutorHao | GCSE WJEC Business Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导