A-Level AQA Psychology: Key Points for Experimental and Practical Assessments | A-Level AQA 心理学：实验与实务考核要点

📚 A-Level AQA Psychology: Key Points for Experimental and Practical Assessments | A-Level AQA 心理学：实验与实务考核要点

In AQA A-Level Psychology, the experimental and practical components are the beating heart of the research methods assessment. Success demands more than memorising textbook definitions; you must be able to design investigations, identify methodological flaws, handle data, and critically evaluate every stage of a psychological study. This article consolidates the essential assessment points that examiners expect to see in experimental and practical questions, ensuring you can apply your knowledge with precision and confidence.

在AQA A-Level心理学中，实验与实务部分是研究方法考核的制胜核心。想要拿高分，仅靠死记硬背定义远远不够；你必须能够设计研究、识别方法缺陷、处理数据并批判性地评估心理学研究的每一个阶段。本文整合了考官在实验与实践题目中最为看重的考核要点，帮助你精准且自信地应用所学知识。

---

1. The Scientific Method and the Experimental Approach | 科学方法与实验取向

Psychology embraces the scientific method, meaning knowledge is built upon empirical evidence gathered through systematic, objective and replicable procedures. The experimental approach is the gold standard because it allows researchers to draw cause-and-effect conclusions by deliberately manipulating an independent variable (IV) while measuring the effect on a dependent variable (DV). All other variables that might influence the DV must be carefully controlled.

心理学信奉科学方法，知识建立在系统、客观且可重复的实证证据之上。实验法之所以被视为黄金标准，是因为它通过有目的地操纵自变量（IV）并测量其对因变量（DV）的影响，使研究者能够得出因果结论。所有可能影响因变量的其他变量都必须加以严格的控制。

An experiment is distinguishable from observational or correlational methods by the active manipulation of the IV. Without this manipulation, any observed relationship could be due to confounding variables, so non-experimental methods can only identify associations, not causation. For instance, a laboratory experiment on sleep deprivation and memory enables researchers to directly control the amount of sleep, whereas a correlation between self-reported sleep and test scores cannot rule out third-factor explanations.

实验法区别于观察法或相关法的核心特征在于对自变量的主动操纵。如果没有这种操纵，任何观察到的关系都有可能源自混淆变量，因此非实验方法只能识别关联，无法确立因果关系。例如，一项关于睡眠剥夺与记忆的实验室实验能使研究者直接控制睡眠时长，而自我报告的睡眠与考试成绩之间的相关则无法排除第三变量解释。

Key features of a well-conducted experiment include a clearly defined IV and DV, operationalised variables, standardised procedures, random allocation of participants where possible, and robust controls against extraneous and confounding variables. The goal is to isolate the IV as the only systematic difference between conditions.

一个设计严谨的实验应具备以下关键特征：明确界定的自变量与因变量、操作化定义的变量、标准化的程序、在可能的情况下对参与者进行随机分配，以及针对无关变量和混淆变量的有力控制。最终目标是让自变量成为不同实验条件之间唯一的系统性差异。

---

2. Types of Experiments | 实验类型

Laboratory experiments take place in a controlled, artificial setting. They offer high internal validity because the researcher can tightly regulate extraneous variables, but this artificiality often limits ecological validity. Demand characteristics and investigator effects are also potential threats, as participants may alter their behaviour based on perceived expectations.

实验室实验在受控的人工环境中进行。由于研究者可以严密调控无关变量，此类实验的内部效度较高，但环境的人工性常常限制了生态效度。需求特征与研究者效应也是潜在威胁，因为参与者可能会根据他们对期望的感知而改变行为。

Field experiments are conducted in a natural, real-world environment. The IV is still deliberately manipulated, but the setting increases ecological validity and reduces demand characteristics. The trade-off is less control over extraneous variables, which can introduce confounding variables and lower internal validity. A classic example is Bickman’s uniform study, where passers-by were asked to obey a person dressed either in a guard’s uniform or ordinary clothes.

现场实验在真实的自然环境里进行。自变量仍然被有意操纵，但场所本身提高了生态效度，也降低了需求特征。代价在于对无关变量的控制减弱，这可能引入混淆变量、拉低内部效度。比克曼的制服研究便是经典实例：路人被要求服从分别身穿保安制服或普通服装的指令。

Natural experiments exploit a naturally occurring event or policy change as the IV; the researcher does not manipulate it. This design is invaluable when deliberate manipulation would be unethical or impractical, but casual inference is weaker because participants are not randomly allocated to conditions. Quasi-experiments, meanwhile, have an IV based on pre-existing characteristics such as gender or culture, again without random assignment. Both designs demand cautious interpretation and explicit acknowledgement of potential confounding variables.

自然实验利用自然发生的事件或政策变化作为自变量，研究者并不主动操纵它。当刻意操纵存在伦理问题或不切实际时，这一设计极具价值，但由于参与者并未被随机分配到各条件，因果推论较弱。准实验则是以既有的特征（如性别或文化）作为自变量，同样缺乏随机分配。这两种设计都要求谨慎解读，并明确承认潜在的混淆变量。

---

3. Experimental Designs | 实验设计

Independent groups design uses different participants in each condition. Its main advantage is the elimination of order effects; participants take part in only one condition so there is no practice or fatigue carryover. The drawback is participant variables—individual differences between groups can act as a confounding variable. Random allocation helps to distribute these differences evenly, provided the sample is large enough.

独立组设计在不同实验条件下使用不同的参与者。其主要优势是消除顺序效应；每位参与者只接受一种条件，因此不会产生练习或疲劳的遗留影响。其缺点是参与者变量——组间的个体差异可能成为混淆变量。只要样本足够大，随机分配有助于均衡分布这些差异。

Repeated measures design exposes the same participants to all conditions. This controls participant variables perfectly because individuals act as their own controls. However, order effects—such as practice, boredom or fatigue—become a serious threat. Counterbalancing, where half the participants experience one order and half the reverse order, is the primary control technique.

重复测量设计让同一批参与者接受所有实验条件。由于个体自成对照，这完美地控制了参与者变量。然而，顺序效应——如练习效应、厌倦或疲劳——变成了严重威胁。平衡对抗是主要的控制手段，即一半参与者经历一种顺序，另一半则经历相反顺序。

Matched pairs design assigns different but carefully matched participants to each condition. Participants are first assessed on a characteristic relevant to the DV and then paired up; one member of each pair is randomly allocated to each condition. This design reduces participant variables without introducing order effects, but matching is time-consuming and can never be perfect.

配对组设计为不同条件分配不同但经过精心匹配的参与者。首先根据与因变量相关的某一特征对参与者进行评估，然后两两配对；每对中的一人被随机分配到各个条件。这一设计在不引入顺序效应的同时减少了参与者变量，但匹配过程耗时，且永远无法做到完美。

---

4. Variables and Operationalisation | 变量与操作性定义

The independent variable (IV) is the factor the researcher manipulates or changes to observe its effect. The dependent variable (DV) is the measurable outcome. Any other variable that might affect the DV, if not controlled, becomes a confounding variable that ruins the internal validity of the experiment. Extraneous variables are any variables other than the IV that could influence the DV; through careful design they are kept constant or minimised.

自变量是研究者为观测效果而操纵或改变的因素。因变量是可测量的结果。任何可能影响因变量的其他因素，若未加控制，就会变成混淆变量，从而破坏实验的内部效度。无关变量指除自变量外所有可能影响因变量的变量；通过周密的设计，它们能够被保持恒定或最小化。

Operationalisation is the process of turning abstract psychological constructs into precise, measurable procedures. Without it, replication becomes impossible and validity plummets. For example, “aggression” could be operationalised as the number of verbal insults delivered in a laboratory paradigm or the intensity of electric shocks a participant chooses to administer. A well-operationalised hypothesis leaves no ambiguity about how each variable is manipulated or measured.

操作性定义是将抽象的心理学构念转化为精确、可测量程序的过程。没有这一步，复制便无从进行，效度将急剧下降。例如，“攻击性”可以被操作化定义为在实验室范式中发出的言语侮辱次数，或参与者选择施加的电击强度。一个操作性定义清晰的假设不会在任何变量如何操纵或测量的问题上留下含糊空间。

---

5. Hypotheses Formulation | 假设的提出

A directional (one-tailed) hypothesis predicts the specific direction of the effect. It states which condition will produce a higher or lower score: “Participants who consume 200 mg of caffeine will recall more words from a list than participants who consume a placebo.” This is appropriate only when prior research strongly indicates the expected direction.

方向性（单尾）假设对效应的具体方向作出预测。它指明哪种条件将产生更高或更低的得分：“摄入200毫克咖啡因的参与者将比服用安慰剂的参与者回忆更多单词。”只有当先前研究强烈支持某种预期方向时，才适合采用此类假设。

A non-directional (two-tailed) hypothesis simply predicts that a difference or relationship exists, without stating the direction: “There will be a difference in the number of words recalled between participants who consume caffeine and those who consume a placebo.” The null hypothesis always states that there is no difference, no effect, or no association, and any observed result is due to chance alone.

非方向性（双尾）假设只预测存在差异或关联，但不说明方向：“摄入咖啡因的参与者与服用安慰剂的参与者在回忆单词数量上存在差异。”零假设总是声称没有差异、没有效应或没有关联，任何观察到的结果都仅是偶然所致。

In the examination, you must be able to identify and write both experimental and null hypotheses, ensuring they are operationalised. A poorly written hypothesis like “Caffeine will affect memory” is too vague; the exact manipulation, the specific measure and the clear prediction must all be present.

在考试中，你必须能够识别并写出实验假设与零假设，并确保它们已被操作化。像“咖啡因会影响记忆”这样含糊的假设过于模糊；精确的操纵、具体的测量以及清晰的预测都必不可少。

---

6. Sampling Methods | 抽样方法

Random sampling gives every member of the target population an equal chance of being selected, often using a random number generator or drawing names. This method minimises researcher bias and produces a highly representative sample, but it is rarely feasible because a complete list of the population is hard to obtain, and selected individuals may not be available.

随机抽样使目标总体中的每一位成员都有同等的机会被选中，通常使用随机数字生成器或抽签方式。该方法能最大程度地减少研究者偏差，产生高度代表性的样本，但它往往不可行，因为很难获得完整的总体名单，且被选中的个体未必能联系得上。

Opportunity sampling selects participants who are readily available and willing, such as classmates or people passing by. It is quick, convenient and economical, but it suffers from severe sample bias and results are unlikely to generalise. Volunteer (self-selected) sampling relies on participants responding to advertisements or invitations; while it can reach a wide audience, it tends to attract a certain personality type and leads to volunteer bias.

机会抽样选取那些方便且愿意参与的个体，比如同学或路人。这种方法快捷、便利、经济，但存在严重的样本偏差，结果难以推广。自愿（自选）抽样依靠参与者对广告或邀请的响应；虽然它可以触及广泛的受众，但容易吸引特定性格类型，从而产生志愿者偏差。

Systematic sampling uses a fixed interval (e.g. every 10th person on a register) and is more representative than opportunity sampling but still requires a sampling frame. Stratified sampling divides the population into relevant strata and then randomly samples from each in proportion to their occurrence in the population, which guarantees excellent representation—as long as the stratification categories are appropriate and well defined.

系统抽样按照固定的间隔（如注册名单上每隔10人抽取一位），比机会抽样更具代表性，但仍需要一个抽样框。分层抽样先将总体划分为相关的层，然后按各层在总体中所占的比例从中随机抽取，只要分层类别合适且明确，就能保证极佳的代表性。

---

7. Ethical Issues and Guidelines | 伦理问题与准则

All psychological research must abide by the ethical principles set out by the British Psychological Society (BPS). Informed consent requires participants to be told the aims, procedures, and their right to withdraw before agreeing to take part. When deception is unavoidable, a full debrief must be provided afterwards, explaining the true purpose and allowing participants to ask questions and have their data withdrawn.

所有心理学研究都必须遵守英国心理学会（BPS）制定的伦理原则。知情同意要求参与者在同意参加前被告知研究目的、程序以及随时退出的权利。当欺骗不可避免时，事后必须提供完整的解释，说明研究的真实目的，并允许参与者提问及撤除其数据。

Protection from harm dictates that participants should not experience physical or psychological injury greater than they would encounter in ordinary life. Confidentiality must be maintained by anonymising data and not disclosing personal details. The right to withdraw must be highlighted again and again—participants can leave at any stage and still expect full respect and, if applicable, payment.

免受伤害原则要求参与者不应经受超过日常生活水平的身体或心理伤害。保密性必须通过匿名化数据和不对外透露个人详情来维护。退出的权利需要一再强调——参与者可以在任何阶段离开，并依然应获得充分尊重，以及适用情况下的酬劳。

In practical assessments, you may be asked to identify ethical breaches in a scenario and suggest improvements. Common solutions include offering a presumptive consent technique, using a confederate to maintain deception, or providing a thorough debrief. Always connect your answer to specific BPS guidelines.

在实务考核中，你可能会被要求识别某个情境中的伦理违规行为并提出改进建议。常见解决方案包括采用预设同意技术、使用合谋者维持欺骗，或提供详尽的释疑说明。永远记得将你的回答与具体的BPS准则挂钩。

---

8. Controlling Confounding Variables | 控制混淆变量

Demand characteristics arise when participants form an interpretation of the experiment’s purpose and unconsciously change their behaviour to match that interpretation. Single-blind procedures, where participants are unaware of which condition they are in, help control for this. Investigator effects occur when the researcher’s expectations influence the outcome, either through subtle cues or biased recording. A double-blind design, in which neither participants nor the conducting researcher know the condition allocation, tackles both demand characteristics and investigator effects simultaneously.

需求特征指的是参与者对实验目的形成某种解读，并下意识地改变自己的行为以迎合这种解读。单盲程序——参与者不知道自己所处的实验条件——有助于控制这种效应。研究者效应则是指研究者的期望通过微妙的暗示或有偏见的记录影响结果。双盲设计让参与者与执行研究的研究者均不知晓条件分配，一举同时对付需求特征与研究者效应。

Order effects threaten repeated measures designs, as practice can improve performance and fatigue can worsen it. Counterbalancing distributes order effects evenly across conditions. Randomisation of task order and standardisation of instructions equally reduce systematic bias. Furthermore, standardising all procedures—including instructions, timing, and environmental conditions—ensures that all participants are treated identically except for the manipulated IV.

顺序效应威胁着重复测量设计，练习可提高表现，疲劳可恶化表现。平衡对抗能将顺序效应均匀分配至各条件。任务次序的随机化和指导语的标准化同样可以减少系统性偏差。此外，标准化所有程序——包括指导语、时间和环境条件——能确保除了被操纵的自变量外，所有参与者得到同等对待。

---

9. Data Analysis and Presentation | 数据分析与呈现

Descriptive statistics summarise data. Measures of central tendency include the mean, median and mode; the mean is most informative but heavily affected by outliers. Measures of dispersion, such as the range and standard deviation, tell us about the spread of scores. In practical write-ups, showing a summary table with these values is common.

描述统计用于概括数据。集中趋势的度量包括平均数、中位数和众数；平均数信息最丰富，但极易受异常值影响。离散量数，如全距和标准差，告诉我们分数分布的广度。在实际写作中，展示含有这些数值的摘要表格是很常见的。

Graphs must be carefully chosen: bar charts are ideal for displaying means across different conditions, often with error bars representing standard deviation; scattergrams reveal the relationship between two co-variables. Never connect points on a non-continuous bar chart with lines. Labelling axes clearly and providing a title is essential.

图表必须谨慎选用：条形图适合显示不同条件下的平均数，通常用误差棒表示标准差；散点图则揭示两个共变因素之间的关系。绝对不要用线连接非连续性条形图中的数据点。清晰标注坐标轴并配上标题至关重要。

Inferential statistics determine whether the results are significant. The choice of test depends on the experimental design, the level of measurement, and the distribution. For instance, an independent groups design with interval data calls for an independent t-test; a repeated measures design with ordinal data might use the Wilcoxon signed-rank test; a nominal data test of association uses chi-squared. The conventional significance level is p < 0.05. A test statistic can be presented as:

推断统计用于判断结果是否显著。选择何种检验取决于实验设计、测量层次以及数据分布。例如，采用等距数据的独立组设计应使用独立t检验；采用顺序数据的重复测量设计可能用到威尔科克森符号秩检验；名义数据的关联检验则使用卡方检验。通常的显著性水平为 p < 0.05。检验统计量可表示为：

t = (M₁ − M₂) / SE

Always state whether the observed value exceeds the critical value and whether the null hypothesis can be rejected. Effect sizes, such as Cohen’s d, should be reported alongside significance to convey the magnitude of the effect.

务必说明观测值是否超过了临界值，以及是否可以拒绝零假设。效应量，如科恩的d，应与显著性一同报告，以传达效应的幅度。

---

10. Validity and Reliability in Experimental Research | 实验研究中的效度与信度

Internal validity refers to the degree to which the observed effect is genuinely due to the manipulation of the IV rather than confounding variables. It can be strengthened by tight experimental controls, random allocation, and standardised procedures. External validity concerns the extent to which findings can be generalised beyond the specific study, including across settings (ecological validity), times, and populations.

内部效度指观察到的效应在多大程度上确实源自对自变量的操纵，而非混淆变量。通过严格的实验控制、随机分配和标准化程序可以强化内部效度。外部效度则关注研究结果能在多大程度上推广到特定研究之外的环境（生态效度）、时间和人群。

Validity of measurement includes face validity (does a test appear to measure what it claims?) and concurrent validity (does the score correlate with an established measure?). To improve validity, researchers can run pilot studies, use multiple measures, and check that classification systems are applied consistently.

测量效度包括表面效度（一个测验看起来是否测量了它声称要测量的构念？）和同时效度（测验分数是否与已知有效的测量存在相关？）。为提高效度，研究者可开展试点研究、使用多种测量工具，并检查分类系统是否得到了一致的应用。

Reliability is about consistency. If an experiment is reliable, repeating it under the same conditions yields similar results. Internal reliability concerns the consistency within a single test, often checked with split-half methods. External reliability refers to consistency over time, such as test-retest reliability, and inter-rater reliability ensures that multiple observers record the same behaviour similarly.

信度关乎一致性。若一个实验是可信的，在相同条件下重复实验会得到相似的结果。内部信度关乎单一测验内部的一致性，常通过分半法检验。外部信度指跨时间的一致性，如重测信度；评分者信度则确保多位观察者以相似的方式记录同一行为。

---

11. Writing a Practical Report | 撰写实验报告

In AQA examinations, you may be asked to design a study and outline the sections of a psychological report. The abstract provides a concise summary of the study comprising the aim, method, results and conclusion. It is written last but placed first. The introduction reviews relevant literature, states the rationale, and culminates in the experimental hypothesis.

在AQA考试中，你可能会被要求设计一项研究并概述心理学报告的各部分。摘要提供研究的简要综述，包含目的、方法、结果

Published by TutorHao | A-Level Psychology Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）