📚 Mastering Experimental Methods for AQA GCSE Psychology: Practical Assessment Essentials | AQA GCSE心理学实验方法精通：实践考核要点
The experimental method lies at the heart of psychological inquiry. In the AQA GCSE specification, practical knowledge is tested not only through standalone research-method questions but also via scenario-based items that demand a deep understanding of how investigations are designed, conducted, and evaluated. This article unpacks every essential element—from formulating hypotheses and controlling variables to handling data ethically—so you can approach your practical assessment with clarity and confidence.
实验方法是心理学探究的核心。在AQA GCSE考试规范中，实践知识不仅通过独立的研究方法题目进行考察，还会融入情境题，要求你深入理解研究如何设计、实施和评估。本文逐一拆解每一个关键要素——从提出假设、控制变量到合乎伦理地处理数据——助你清晰自信地应对实践考核。
1. Variables and Hypotheses | 变量与假设
A well‑designed experiment starts by identifying the independent variable (IV)—the factor the researcher deliberately changes or manipulates—and the dependent variable (DV)—the factor that is measured to see the effect of the IV. All other variables that could influence the DV are called extraneous variables; if they are not controlled, they become confounding variables that ruin the study's internal validity.
一项设计精良的实验始于明确自变量（IV）和因变量（DV）。自变量是研究者有意改变或操纵的因素，因变量是为观察自变量效果而测量的因素。所有其他可能影响因变量的因素都叫作额外变量；如果不加控制，它们就会变成混淆变量，损害研究的内部效度。
Operationalisation means stating precisely how the IV and DV will be manipulated and measured. For example, instead of saying 'we will test memory', you must specify: 'participants will recall the number of words correctly from a list of 20 after a 2‑minute delay'. A hypothesis must be testable and directional (one‑tailed) if previous research points to a specific outcome, or non‑directional (two‑tailed) if not.
操作化是指精确说明如何操纵自变量和测量因变量。例如，不能说“我们将测试记忆力”，而必须明确：“参与者在延迟2分钟后，从20个单词的列表中正确回忆出的单词数量”。假设必须可检验，若先前研究指向特定结果，便采用方向性（单尾）假设，否则采用非方向性（双尾）假设。
- Null hypothesis: 'There will be no significant difference in the number of words recalled between participants who learn with music and those who learn in silence.' / 零假设：“在有无音乐环境下学习的参与者，其回忆的单词数量没有显著差异。”
- Alternative hypothesis (directional): 'Participants who learn in silence will recall significantly more words than those who learn with music.' / 备择假设（方向性）：“在安静环境下学习的参与者回忆的单词数量将显著多于在有音乐环境下学习的参与者。”
2. Types of Experiments | 实验类型
AQA GCSE Psychology requires you to distinguish among four main types of experiments. Laboratory experiments take place in a controlled, artificial setting where the researcher manipulates the IV and strictly controls extraneous variables. They have high internal validity and are replicable, but may lack ecological validity because the environment feels unnatural to participants.
AQA GCSE心理学要求你区分四种主要的实验类型。实验室实验在受控的人工环境中进行，研究者操纵自变量并严格控制额外变量。这类实验内部效度高且可重复，但由于环境对参与者而言不自然，可能缺乏生态效度。
Field experiments are conducted in the participants' natural surroundings (e.g. a school or shopping centre). The IV is still deliberately manipulated, but some control over extraneous variables is inevitably lost. They offer greater ecological validity, yet ethical issues such as lack of informed consent often arise because participants may not know they are being studied.
现场实验在参与者的自然环境中进行（如学校或购物中心）。自变量仍被有意操纵，但对额外变量的控制难免有所丧失。现场实验的生态效度更高，但常因参与者可能不知情而引发知情同意等伦理问题。
Natural experiments exploit a naturally occurring event or change (e.g. a new school policy) as the IV. The researcher does not directly manipulate the IV but simply measures its effect on the DV. They are useful when it would be unethical or impossible to manipulate the IV, yet causal conclusions are weaker because participants cannot be randomly allocated to conditions.
自然实验利用自然发生的事件或变化（如一项新的学校政策）作为自变量。研究者不直接操纵自变量，只是测量其对因变量的影响。当操纵自变量不合伦理或不可能时，自然实验很有用，但由于参与者无法被随机分配，因果推论较弱。
Quasi‑experiments also lack random allocation; the IV is a pre‑existing characteristic of the participants (e.g. gender, age, phobia status). Comparisons are made between groups that differ on this characteristic. Like natural experiments, they sacrifice some control for practical or ethical reasons.
准实验同样缺乏随机分配；自变量是参与者原有的特征（如性别、年龄、恐惧症状况），并对具备不同特征的群体进行比较。与自然实验一样，出于现实或伦理考量，这类实验牺牲了一定的控制力。
3. Experimental Designs | 实验设计
Once the type of experiment is chosen, you must decide how to allocate participants to conditions. The three classic designs have distinct strengths and weaknesses that exam questions frequently target.
选定实验类型之后，你必须决定如何将参与者分配到不同的实验条件。三种经典设计各有优劣，是考试常考的重点。
- Independent groups: Different participants are placed in each condition. This avoids order effects (e.g. practice or fatigue), but participant variables (individual differences) can confound the results. Random allocation to conditions is essential. / 独立组设计：不同参与者进入不同条件。这避免了顺序效应（如练习或疲劳），但参与者变量（个体差异）可能混淆结果。随机分配至各组至关重要。
- Repeated measures: The same participants take part in every condition. This eliminates participant variables and requires fewer participants; however, order effects become a threat. Counterbalancing (half the participants do condition A then B, the other half do B then A) is used to control them. / 重复测量设计：同一批参与者参加所有条件。这消除了参与者变量且所需人数较少，但顺序效应成为威胁。采用平衡法（一半参与者先做条件A后做条件B，另一半先B后A）加以控制。
- Matched pairs: Participants are paired on key relevant traits (e.g. memory span) and then one member of each pair is randomly allocated to each condition. This reduces participant variables without introducing order effects, but matching is time‑consuming and can never be perfect. / 配对组设计：参与者按关键相关特质（如记忆广度）配对，每对成员再被随机分配到不同条件。这既减少了参与者变量，又不会引入顺序效应，但配对过程耗时且不可能完美。
| Design / 设计 | Key threat / 主要威胁 | Control method / 控制方法 | Participant numbers / 参与者数量 |
|---|---|---|---|
| Independent groups / 独立组 | Participant variables / 参与者变量 | Random allocation / 随机分配 | Higher / 较多 |
| Repeated measures / 重复测量 | Order effects / 顺序效应 | Counterbalancing / 平衡法 | Lower / 较少 |
| Matched pairs / 配对组 | Imperfect matching / 不完美匹配 | Pre‑test and careful pairing / 前测与精细配对 | Moderate / 中等 |
4. Control of Extraneous Variables and Standardisation | 额外变量的控制与标准化
To establish a cause‑and‑effect relationship, you must minimise the influence of extraneous variables. Standardised procedures mean that every participant is tested under exactly the same conditions, with identical instructions, timings, stimuli, and environment. This ensures that any differences in the DV can be attributed to the IV rather than to unintended variations.
要确立因果关系，你必须将额外变量的影响降至最低。标准化程序意味着每位参与者均在完全相同的条件下接受测试，使用相同的指导语、时间安排、刺激材料和环境。这能确保因变量的任何差异都可归因于自变量，而非无意间的变化。
Common extraneous variables include: situational variables (e.g. noise, lighting, time of day) and participant variables (e.g. mood, motivation, prior knowledge). Randomisation—using chance to assign participants to conditions or to determine the order of tasks—helps spread the effects of these variables evenly across groups. For example, generating a random sequence of word lists prevents order bias.
常见的额外变量包括：情境变量（如噪音、照明、时段）和参与者变量（如情绪、动机、先前知识）。随机化——利用机率分配参与者或决定任务顺序——有助于将这些变量的效应均衡分散在各组之间。例如，生成随机单词列表序列可防止顺序偏差。
When a variable cannot be eliminated, researchers may use a control group that does not receive the IV (or receives a placebo) to establish a baseline comparison. This is particularly common in drug trials or intervention studies.
当变量无法消除时，研究者可使用控制组，这组不接受自变量处理（或接受安慰剂），以建立基线比较。这在药物试验或干预研究中尤为常见。
5. Sampling Methods | 抽样方法
The sample is the group of participants who actually take part in the study, drawn from a target population. A representative sample allows generalisation of findings back to that population. AQA expects you to evaluate four main sampling techniques.
样本是实际参与研究的一组参与者，从目标总体中抽取。一个有代表性的样本能将研究结果推广回该总体。AQA要求你评估四种主要抽样技术。
- Random sampling: Every member of the target population has an equal chance of being selected (e.g. names drawn from a hat). It is unbiased but can be impractical and may still produce an unrepresentative sample by chance. / 随机抽样：目标总体中每个成员被选中的几率相等（如从帽中抽名字）。它无偏，但可能不切实际，且仍可能因偶然产生不具代表性的样本。
- Opportunity sampling: Recruiting whoever is available and willing at the time (e.g. students in the corridor). It is quick and easy, yet highly prone to researcher bias and likely to produce a biased, unrepresentative sample. / 便利抽样：招募当时恰好有空且愿意的人（如走廊上的学生）。它快捷简便，但极易受研究者偏见影响，且很可能产生有偏、不具代表性的样本。
- Volunteer (self‑selected) sampling: Participants respond to an advertisement or sign‑up sheet. They tend to be highly motivated, but volunteer bias means the sample may be unusually keen or have specific traits that differ from the wider population. / 志愿者（自选）抽样：参与者响应广告或报名表。他们通常积极性很高，但志愿者偏差意味着样本可能异常热衷或具备不同于普通人群的特定特质。
- Systematic sampling: Selecting every nth person from a list of the population (e.g. every 5th name in the register). It is more objective than opportunity sampling but can introduce bias if the list has an underlying pattern. / 系统抽样：从总体名单中每隔n个人选一个（如从注册名单每隔5个名字选一人）。它比便利抽样更客观，但若名单存在潜在规律则可能引入偏差。
6. Ethical Considerations | 伦理考量
All psychological research with human participants must adhere to the ethical guidelines set by the British Psychological Society (BPS). In the AQA GCSE, you must not only list these principles but also apply them to scenarios, suggesting how ethical issues could be addressed.
所有涉及人类参与者的心理学研究都必须遵守英国心理学会（BPS）制定的伦理准则。在AQA GCSE考试中，你不仅要列出这些原则，还要将其应用于情境，建议如何处理伦理问题。
- Informed consent: Participants should be told the aims, procedures, and their right to withdraw before agreeing to take part. If full disclosure would alter behaviour, researchers may use presumptive consent (asking a similar group) or debrief afterward. / 知情同意：参与者在同意参加前应被告知研究目的、程序及退出权利。若完全披露会改变行为，研究者可采用假设同意（询问类似群体）或事后进行事后解释。
- Deception: Withholding information or misleading participants should be avoided unless scientifically justified. A cost‑benefit analysis must show that the knowledge gained outweighs the distress, and a thorough debrief must be provided. / 欺骗：避免隐瞒信息或误导参与者，除非有科学上的正当理由。必须进行成本‑收益分析，表明获得的知识大于困扰，并提供详尽的事后解释。
- Protection from harm: Researchers must anticipate and minimise physical and psychological harm. Participants should not leave the study in a worse state than when they entered. Offering counselling or the right to withdraw data is essential. / 免受伤害：研究者必须预见并尽量减少身心伤害。参与者不应在离开研究时陷入比参与前更糟的状态。提供咨询服务或允许撤回数据是必要的。
- Confidentiality: Participants' identities and personal data must be protected, usually by using code numbers instead of names. Data should be stored securely and not shared without permission. / 保密：参与者的身份与个人数据必须受到保护，通常用代码编号替代姓名。数据应安全储存，未经许可不得分享。
- Right to withdraw: Participants must be told they can leave the study at any time without penalty, and they may request that their data be removed. This right continues even after the study has finished. / 退出权利：参与者须被告知他们可随时退出研究而不受惩罚，并可要求删除其数据。即使在研究结束后，该权利依然有效。
7. Pilot Studies and Refining Procedures | 预实验与程序优化
A pilot study is a small‑scale trial run of the investigation before the main data collection. It helps identify any flaws in the design, ambiguous instructions, or confusing materials. For example, if participants in a pilot misunderstand the rating scale, the researcher can reword it to ensure clarity. Piloting is not optional—it is a hallmark of rigorous psychological research and can save time, money, and ethical problems.
预实验是在正式收据收集之前进行的小规模试运行。它有助于识别设计中的缺陷、模糊的指导语或令人困惑的材料。例如，若预实验中的参与者误解了评分量表，研究者可以重新措辞以确保清晰。预实验不是可有可无的——它是严谨心理学研究的标志，能节省时间、金钱并避免伦理问题。
Feedback from pilot participants often leads to standardisation improvements, such as fixing timing, adjusting the difficulty of stimuli, or rewording the debriefing script. Pilot studies also allow you to check that equipment works and that the DV produces a meaningful range of scores (avoiding floor or ceiling effects).
来自预实验参与者的反馈往往带来标准化改进，例如修正时间安排、调整刺激材料的难度或改写事后解释词。预实验还能让你检查设备是否运作正常，以及因变量能否产生有意义的分数范围（避免地板效应或天花板效应）。
8. Data Handling: Descriptive Statistics | 数据处理：描述统计
After collecting data, researchers summarise it using measures of central tendency (mean, median, mode) and measures of dispersion (range, and sometimes standard deviation although you will primarily use the range at GCSE). The choice of which measure to report depends on the level of measurement and the presence of outliers.
收集数据后，研究者使用集中趋势量数（平均数、中位数、众数）和离散量数（全距，GCSE阶段主要使用全距）进行概括。选用哪种量数取决于测量层次及是否存在异常值。
- Mean: The arithmetic average. It uses all scores, making it sensitive and representative, but it is easily distorted by outliers. / 平均数：算术平均值。它利用了所有分数，灵敏且有代表性，但易被异常值扭曲。
- Median: The middle value when scores are ordered. It is unaffected by extreme scores; ideal for ordinal data or when outliers exist. / 中位数：将分数排序后位于中间的数值。它不受极端分数的影响，是顺序数据或存在异常值时的理想选择。
- Mode: The most frequently occurring score. It can be used with any type of data but may not be representative if there are multiple modes or if it misses the overall pattern. / 众数：出现频率最高的分数。适用于任何类型的数据，但若出现多个众数或无法反映总体模式，其代表性可能不足。
- Range: The difference between the highest and lowest scores. It is quick to calculate but gives no information about how scores are distributed around the centre. / 全距：最高分与最低分之差。计算快捷，但不提供分数围绕中心如何分布的信息。
ticipant scored 45 due to cheating, the mean would leap to 19.6 while the median would remain 11, demonstrating the median’s robustness.
例如，在一项记忆测试中，分数为 7, 9, 11, 14, 15，平均数为 11.2，中位数为 11，无众数（或各出现一次），全距 = 15 − 7 = 8。若一名参与者因作弊得分为 45，平均数将跃升至 19.6，而中位数仍为 11，这展示了中位数的稳健性。
9. Data Presentation: Tables, Graphs and Charts | 数据呈现：表格、图表与图形
Raw data should be organised into summary tables showing conditions and the corresponding measures of central tendency and dispersion. A well‑constructed table is clearly labelled, with units stated, and rows/columns handled logically.
原始数据应组织成汇总表，显示实验条件及相应的集中趋势和离散量数。一个结构良好的表格应清晰标注，明确单位，行列逻辑清晰。
Graphical displays make patterns visible at a glance. For experiments comparing conditions, bar charts are standard: separate bars for each condition’s mean, with the DV on the y‑axis and conditions on the x‑axis. Remember that bar charts are used for discrete categories; the bars do not touch.
图形展示能让模式一目了然。比较不同实验条件时，条形图是标准呈现方式：每个条件用独立的条形表示其平均数，因变量置于纵轴，条件置于横轴。记住条形图用于离散类别，条形之间不相连。
- Line graphs are used to show continuous change over time or across levels of an IV that has a natural order (e.g. age groups). / 折线图用于显示随时间连续变化或按具有自然顺序的自变量水平（如年龄段）的变化。
- Scattergrams display the relationship between two co‑variables in a correlation; you may be asked to interpret the direction and strength of a correlation from a scattergram. / 散点图展示相关研究中两个共变量之间的关系；你可能会被要求根据散点图解读相关的方向和强度。
- Frequency distribution tables and histograms (for continuous data, with touching bars) can also appear in exam questions. / 频数分布表和直方图（用于连续数据，条形相连）也可能出现在考题中。
10. Interpretation of Findings and Drawing Conclusions | 结果解读与结论得出
Once the data is analysed, you must return to the hypotheses. If the descriptive statistics show a difference in the predicted direction (e.g. the mean of the silence condition is noticeably higher than the music condition), the researcher may reject the null hypothesis and accept the alternative hypothesis. However, at GCSE level you are not required to compute inferential statistics; you rely on a sensible comparison of descriptive measures.
数据分析完成后，你必须回到假设本身。若描述统计显示差异朝向预测方向（例如安静条件下的平均数明显高于音乐条件），研究者可以拒绝零假设并接受备择假设。然而，在GCSE阶段你不必计算推断统计；你需要依赖于对描述性量数的合理比较。
A sound conclusion states whether the results support the hypothesis, refers back to the original aim, and acknowledges any limitations that might prevent a firm causal claim. For eye‑witness testimony studies, for instance, you might conclude that misleading post‑event information did reduce accuracy, but that the lack of a control group means we cannot be certain the IV was the sole cause.
合理的结论应指出结果是否支持假设，回溯最初目的，并承认可能妨碍坚固因果推论的任何局限。例如，在目击证词研究中，你可能会得出结论：误导性事后信息确实降低了准确性，但由于缺乏控制组，我们不能确定自变量是唯一的原因。
Use of percentage change or ratios can strengthen your interpretation. For example, 'the silence group recalled 25% more words on average than the music group' is more compelling than just quoting two means.
使用百分比变化或比值能增强你的解读。例如，“安静组平均回忆的单词比音乐组多25%”比仅列出两个平均数更有说服力。
11. Evaluation: Reliability and Validity | 评估：信度与效度
Every practical investigation must be assessed for reliability (consistency) and validity (accuracy). At GCSE, you need to define each term and suggest ways to improve them.
每项实践研究都必须评估其信度（一致性）与效度（准确性）。在GCSE阶段，你需要定义每个术语并提出改进方法。
- Internal reliability: The extent to which the procedure is consistent within itself. For example, using standardised instructions and the same material for every participant improves internal reliability. / 内部信度：程序本身的一致程度。例如，使用标准化指导语和对每位参与者使用相同材料可提高内部信度。
- External reliability: The extent to which the study can be replicated by other researchers and yield similar results. Replicability depends on the detail of the standardised procedure report. / 外部信度：研究可被其他研究者重复并得出相似结果的程度。可重复性取决于标准化程序报告的详尽程度。
- Internal validity: Whether the study truly measures what it was intended to measure, with no confounding variables. For instance, if a memory test is actually measuring reading speed, internal validity is low. Face validity (do the items appear to measure the construct?) is a quick check. / 内部效度：研究是否真正测量了它想要测量的东西，且没有混淆变量。例如，若记忆测试实际上在测量阅读速度，内部效度就低。表面效度（项目是否看起来在测量该构念？）是一种快速检查手段。
- External validity: The degree to which findings can be generalised. Ecological validity (can results apply to real‑life settings?) and population validity (can results apply to people beyond the sample?) fall under this heading. Many laboratory experiments struggle with ecological validity because the artificial environment may not reflect everyday behaviour. / 外部效度：研究结果可概括化的程度。生态效度（结果能否适用于真实生活环境？）和总体效度（结果能否推广至样本以外的人群？）均属于这一范畴。许多实验室实验因人工环境无法反映日常行为而在生态效度上遇到困难。
To improve reliability, replicate the study with a larger sample; to improve validity, add a control group, use double‑blind procedures, or conduct the experiment in a natural setting (field experiment). Always link your suggestion to the specific threat identified.
要提高信度，可扩大样本重复研究；要提高效度，可增加控制组、使用双盲程序或在自然环境中进行实验（现场实验）。你的改进建议务必与所识别的具体威胁挂钩。
12. Practical Assessment Tips and Common Pitfalls | 实践考核技巧与常见错误
Exam questions on practical methods often provide a scenario and ask you to identify the IV, DV, design, or ethical issue, or to suggest an improvement. Read the stem carefully and highlight keywords. When writing a hypothesis, always include both conditions of the IV and the DV in operationalised terms. Avoid vague phrases like 'memory will be better'; state exactly what 'better' means and the units of measurement.
实践方法考题通常会提供一个情境，要求你识别自变量、因变量、设计或伦理问题，或提出改进建议。仔细阅读题干并划出关键词。撰写假设时，务必同时包含自变量的两种条件及操作化的因变量。避免使用“记忆会更好”这类模糊措辞；明确说明“更好”的含义和度量单位。
Common pitfalls include: confusing the experimental design with the type of experiment; stating a directional hypothesis without justification; forgetting to mention counterbalancing when evaluating a repeated measures design; describing sample size as 'too small' without explaining why that matters (lack of representativeness, low population validity); and quoting a mean without acknowledging an extreme range that skews the interpretation.
常见错误包括：混淆实验设计与实验类型；未给出理由便使用方向性假设；在评估重复测量设计时忘记提及平衡法；仅说样本量“太小”而不解释为何重要（缺乏代表性，总体效度低）；以及引用平均数却不承认存在扭曲解读的极端全距。
For design‑your‑own‑study questions, structure your answer logically: aim and hypothesis, IV & DV (operationalised), design with justification, sampling method and its weakness, ethical considerations, procedure steps (standardised), and a brief evaluation point. Use psychological vocabulary precisely—'random allocation' not just 'random', 'confounding variable' not 'extraneous variable' when it has actually distorted the DV.
对于自主设计研究的题目，按逻辑组织答案：目的与假设、自变量与因变量（操作化）、实验设计及理由、抽样方法及其弱点、伦理考量、程序步骤（标准化）以及简要的评估点。准确使用心理学词汇——是“随机分配”而不仅仅是“随机”，当变量已切实扭曲因变量时使用“混淆变量”而非“额外变量”。
Finally, always link back to the real world. If the experiment is about learning in silence vs. music, comment on what the findings might mean for revision strategies in a school context. Demonstrating applied understanding can lift your answer to the highest mark band.
最后，始终联系实际。若实验是关于安静学习与音乐环境下的学习，不妨评论研究结果对学校复习策略可能有何意义。展现应用性理解能把你的答案推向最高分数段。
Published by TutorHao | Psychology Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图