📚 Mastering Experimental Methods: CCEA AS Psychology Practical Guide | 掌握实验方法：CCEA AS心理学实践考核指南
Research methods lie at the heart of CCEA AS Psychology, and the practical assessment demands more than just textbook knowledge. You are expected to design a sound experiment, recognise and control variables, handle ethical issues, select appropriate inferential statistics, and write a coherent scientific report. This guide distils the essential skills and common pitfalls to help you approach the experimental / practical component with confidence.
研究方法是CCEA AS心理学的核心，实践考核要求的不仅仅是课本知识。你需要设计严密的实验，识别并控制变量，处理伦理问题，选择合适的推断统计量，并撰写出条理清晰的科学报告。本指南浓缩了关键技能和常见误区，帮助你自信应对实验/实践考核。
1. Experimental Variables and Hypotheses | 实验变量与假设
In any experiment, you must identify the independent variable (IV) – the factor you manipulate – and the dependent variable (DV) – the outcome you measure. It is essential to operationalise both variables by providing clear, precise definitions of how they will be changed or measured. For example, if investigating the effect of caffeine on memory, the IV could be ‘amount of caffeine consumed (0 mg vs 200 mg)’, and the DV ‘number of words correctly recalled from a 20-word list’.
在任何实验中，必须明确自变量（IV）——你操控的因素，以及因变量（DV）——你测量的结果。关键是对两个变量进行操作化定义，清晰地说明如何改变或测量它们。例如，若研究咖啡因对记忆的影响，自变量可定义为“摄入咖啡因的量（0毫克 vs 200毫克）”，因变量为“正确回忆20个单词的数量”。
Hypotheses should be stated clearly. An experimental (alternative) hypothesis predicts a significant effect or relationship; it can be directional (one‑tailed) if previous research suggests a specific direction, or non‑directional (two‑tailed) if not. A null hypothesis always states there will be no significant difference or correlation, and inferential statistics test the probability of obtaining the data if the null is true.
假设必须清晰陈述。实验（备择）假设预测显著的效应或关系；如果先前研究提示明确方向，可使用定向假设（单尾），否则使用非定向假设（双尾）。零假设则总是声称不存在显著差异或相关，推断统计检验的正是零假设成立时得到该组数据的概率。
2. Experimental Designs: Independent, Repeated, Matched | 实验设计：独立组、重复测量、配对组
The three main experimental designs each have strengths and limitations. In an independent groups design, different participants are used in each condition, avoiding order effects but requiring more participants and risking participant variables. Random allocation to conditions is crucial here.
三种主要实验设计各有利弊。独立组设计中，每个条件使用不同的参与者，可避免顺序效应，但需要更多参与者，且有参与者变量干扰的风险。此时，随机分配至各条件是关键。
A repeated measures design exposes the same participants to all conditions, which controls participant variables and needs fewer people. However, it introduces order effects (practice, fatigue, boredom) that must be controlled by counterbalancing the sequence of conditions.
重复测量设计让同一组参与者接受所有条件，能控制参与者变量且所需人数较少。但会引入顺序效应（练习、疲劳、厌倦），必须通过平衡条件顺序来加以控制。
Matched pairs design is a compromise: different but similar participants are used, matched on key characteristics (e.g. IQ, memory span) before being randomly assigned to conditions. It reduces participant variables and avoids order effects, but matching is time‑consuming and imperfect.
配对组设计是一种折中方案：使用不同但相似的参与者，在关键特征（如智商、记忆广度）上匹配后随机分配到各条件。它可以减少参与者变量，避免顺序效应，但匹配过程耗时且难以完美实现。
3. Sampling Techniques and Participant Allocation | 抽样技术与参与者分配
The way you select participants affects both the internal and external validity of your experiment. Random sampling gives every member of the target population an equal chance, reducing bias, but is often impractical. Opportunity sampling uses whoever is available at the time, which is quick but likely to produce a biased, unrepresentative sample.
选择参与者的方式会影响实验的内部效度和外部效度。随机抽样使目标人群中每个人机会均等，减少了偏差，但常常难以实施。机会抽样选取当时方便找到的人，速度快，但很可能产生有偏差、不具代表性的样本。
Stratified sampling ensures subgroups (strata) are represented in proportion to their occurrence in the population, while volunteer (self‑selected) sampling relies on people responding to advertisements – it can yield motivated participants but may be biased towards certain personality types. Always consider the sample size: larger samples increase reliability and the power of statistical tests, while very small samples make it harder to detect real effects.
分层抽样确保各子群体按人口比例入选，而志愿者（自选）抽样依赖人们回应广告，可能得到动机较强的参与者，但会偏向特定性格类型。始终要考虑样本量：大样本能提高信度和统计检验力，过小的样本则难以探测真实效应。
4. Ethical Guidelines in Psychological Research | 心理研究中的伦理准则
Ethical practice is a non‑negotiable part of CCEA practical work. You must obtain informed consent from participants, meaning they know the nature of the study, what they will do, and that they can withdraw at any time without penalty. If some deception is necessary, it should be minimal, justified, and followed by a thorough debrief that explains the true purpose and offers support.
伦理实践是CCEA实验工作中不可妥协的部分。你必须获得参与者的知情同意，即他们了解研究的性质、将要做什么，以及可以随时无惩罚退出。若有必要使用欺骗，也应是轻微的、经得起辩护的，并在事后进行详尽的事后说明，解释真实目的并提供支持。
You must also protect participants from physical and psychological harm, guarantee confidentiality of data, and allow them the right to withdraw data after the study. When using observational methods, privacy must be respected – usually only observe in public settings unless consent is given. BPS (British Psychological Society) guidelines provide the framework you should follow.
你还必须保护参与者免受身心伤害，确保数据保密，并允许他们在研究结束后撤回数据。使用观察法时，必须尊重隐私——通常只在公共场合观察，除非获得同意。英国心理学会（BPS）的准则提供了应遵循的框架。
5. Controlling Extraneous Variables | 控制额外变量
An extraneous variable is any factor, other than the IV, that could affect the DV. If it varies systematically with the IV, it becomes a confounding variable and can invalidate your results. Participant variables (e.g. motivation, memory ability), situational variables (e.g. noise, time of day), and experimenter effects (e.g. unintentional bias) all need to be controlled.
额外变量是除自变量外任何可能影响因变量的因素。若它随自变量系统变化，就成为混淆变量，可能使结果无效。参与者变量（如动机、记忆能力）、情境变量（如噪音、一天中的时段）和实验者效应（如无意间的偏见）都需要被控制。
Standardisation of procedures (using the same instructions, timings, and environment for all participants), randomisation, and single‑ or double‑blind techniques are powerful controls. In a double‑blind trial, neither the participant nor the researcher knows which condition the participant is in, eliminating both demand characteristics and experimenter bias. Use of a control group or a placebo can also isolate the effect of the IV.
程序标准化（所有参与者遵循相同的指导语、时间安排和环境）、随机化，以及单盲或双盲技术，都是强有力的控制手段。在双盲试验中，参与者与研究者均不知道参与者处于何种条件，从而消除需求特征和实验者偏见。设立控制组或使用安慰剂也能分离出自变量的效应。
6. Data Collection: Self‑report and Observations | 数据收集：自我报告法与观察法
In AS practicals, you often gather data using questionnaires, interviews, or systematic observation. Questionnaires with closed questions yield quantitative data that is easy to analyse, whereas open questions produce qualitative data rich in detail. Interviews can be structured (fixed questions) or unstructured (free‑flowing), each with trade‑offs in reliability and depth.
在AS实践课中，你常使用问卷、访谈或系统观察收集数据。带有封闭式问题的问卷产生易于分析的量化数据，而开放式问题则获得细节丰富的质性数据。访谈可以是结构化的（固定问题）或非结构化的（自由交谈），两者在信度和深度上各有取舍。
Observational research requires clear behavioural categories that are observable, measurable, and mutually exclusive. You must decide between time sampling (noting behaviour at preset intervals) and event sampling (recording each occurrence of a target behaviour). Inter‑rater reliability – the extent to which two observers agree – should be checked to ensure consistency, often calculated using correlation or percentage agreement.
观察研究需要明确的行为类别，这些类别应可观察、可测量且互斥。你必须在时间取样（按预设间隔记录行为）和事件取样（记录每次目标行为的发生）之间做出选择。应检验评分者间信度，即两位观察者的一致性程度，通常通过相关或一致性百分比来计算。
7. Descriptive Statistics and Graphical Representation | 描述统计与图表呈现
Descriptive statistics summarise your data set. Measures of central tendency include the mean (arithmetic average), median (middle score), and mode (most frequent score). The mean is the most sensitive but can be distorted by outliers; the median is robust for skewed data. Measures of dispersion – range and standard deviation – indicate how spread out the scores are.
描述统计用于概括数据集。集中趋势的测量包括均值（算术平均数）、中位数（居中分数）和众数（出现频率最高的分数）。均值最为灵敏，但易受异常值影响；对于偏态分布，中位数更加稳健。离散度的测量——全距和标准差——则表明分数的离散程度。
Data should be presented clearly in tables with appropriate titles and units. Bar charts are used for discrete (nominal) data, histograms for continuous data, and scattergrams to display correlations between two co‑variables. In CCEA practicals, ensure all graphs have labelled axes and a clear title. A normal distribution is bell‑shaped and symmetrical, important for parametric tests, though at AS you will mainly use non‑parametric alternatives.
数据应在配有恰当标题和单位的表格中清晰呈现。条形图用于离散（定类）数据，直方图用于连续数据，散点图则显示两个协变量之间的相关。在CCEA实践报告中，确保所有图表都有带标签的坐标轴和清晰标题。正态分布呈钟形对称，对参数检验很重要，不过在AS阶段你主要使用非参数替代方法。
8. Choosing the Right Inferential Test | 选择合适的推断统计检验
Inferential statistics let you determine whether the observed effect is likely to have occurred by chance. The choice of test depends on the research design (independent groups, repeated measures, correlation) and the level of measurement (nominal, ordinal). CCEA AS candidates need to be familiar with the following non‑parametric tests: Sign test, Chi‑squared, Spearman’s rho, Mann‑Whitney U, and Wilcoxon signed ranks.
推断统计帮助你判断观察到的效应是否可能偶然产生。检验的选择取决于研究设计（独立组、重复测量、相关）和测量层次（定类、定序）。CCEA AS考生需要熟悉以下非参数检验：符号检验、卡方检验、斯皮尔曼等级相关系数、曼‑惠特尼U检验和威尔科克森符号秩检验。
A simple decision table helps avoid errors. For an independent groups design with nominal data, use Chi‑squared; for ordinal data, use Mann‑Whitney U. For repeated measures with nominal data, use the Sign test; for ordinal data, use Wilcoxon. For a correlation design with ordinal data, use Spearman’s rho. Always be aware whether your test is one‑tailed or two‑tailed, matching your hypothesis.
一个简单的决策表有助于避免错误。独立组设计，若为定类数据，用卡方检验；若为定序数据，用曼‑惠特尼U检验。重复测量设计，定类数据用符号检验，定序数据用威尔科克森检验。相关设计且数据为定序时，用斯皮尔曼等级相关系数。始终注意你的检验是单尾还是双尾，需与假设匹配。
9. Conducting and Reporting Inferential Statistics | 执行与报告推断统计
Once you have chosen a test, calculate the observed value using the appropriate formula and then compare it with a critical value from a statistical table. For the Sign test, the observed value S is simply the number of the less frequent sign (e.g. number of participants whose score decreased). For Chi‑squared:
χ² = Σ (O − E)² / E
where O = observed frequency, E = expected frequency. The calculated χ² is compared against the critical value at p ≤ 0.05 for the appropriate degrees of freedom.
选定检验后，使用恰当的公式计算观测值，再与统计表中的临界值作比较。对符号检验，观测值S就是较少出现的符号的频次（如分数下降的参与者人数）。对卡方：
χ² = Σ (O − E)² / E
其中O为观察频数，E为期望频数。将计算的χ²值与相应自由度下p ≤ 0.05的临界值比较。
If the observed value is equal to or more extreme than the critical value, the result is significant and the null hypothesis is rejected. Remember to report the significance level (p < 0.05 or p > 0.05) and note the possibility of a Type I error (false positive) or Type II error (false negative). Increasing sample size, improving controls, and setting an appropriate significance level all help reduce errors.
若观测值等于或大于（视检验而定）临界值，结果显著，拒绝零假设。记得报告显著性水平（p < 0.05 或 p > 0.05），并注意Ⅰ型错误（假阳性）和Ⅱ型错误（假阴性）的可能。增大样本量、改进控制措施、设定恰当的显著性水平都有助于减少错误。
10. Writing Up a Full Experimental Report | 撰写完整的实验报告
A standard psychological report follows the structure: Title, Abstract, Introduction, Method (Design, Participants, Apparatus/Materials, Procedure), Results, Discussion, References, and Appendices. The Abstract is a concise summary of all sections, usually around 150 words. In the Introduction, review relevant background theories and studies, and state your experimental and null hypotheses clearly.
标准的心理学报告结构依次为：标题、摘要、引言、方法（设计、参与者、器材/材料、程序）、结果、讨论、参考文献和附录。摘要是对所有部分的简洁概括，通常约150字。引言部分需回顾相关的背景理论与研究，并明确陈述你的实验假设和零假设。
The Method must be detailed enough to allow replication. Under Design, specify the experimental design, IV and DV with operationalisation, identification of controls, and ethical considerations. The Results section reports descriptive statistics (with tables or graphs) and the outcome of the inferential test, including observed/critical values and a statement of significance. The Discussion interprets findings, links them to the introduction, acknowledges limitations, and suggests improvements and future research.
方法部分必须足够详细以便复制。在“设计”中，说明实验设计类型、经操作化定义的自变量与因变量、控制措施的确认以及伦理考量。结果部分报告描述统计（附表格或图表）和推断统计结果，包括观测值/临界值及显著性陈述。讨论则解释结果、联系引言、承认局限，并提出改进建议与未来研究方向。
11. Reliability and Validity in Experimental Work | 实验中的信度与效度
Reliability refers to consistency. Internal reliability can be improved by standardising procedures and training observers. External reliability is about producing the same results on different occasions; test‑retest or inter‑rater checks help establish this. Validity concerns whether you are measuring what you intend to measure. Internal validity may be threatened by confounding variables, demand characteristics, or investigator effects; using control groups, standardisation, and blinding strengthens it.
信度指一致性。通过标准化程序、培训观察者可以提高内部信度。外部信度关乎在不同场合能否得出相同结果；重测信度或评分者间一致性检查有助于验证。效度关乎你是否测量到了想要测量的内容。混淆变量、需求特征或实验者效应会威胁内部效度；采用控制组、标准化和盲法则能增强内部效度。
Ecological validity (whether results generalise to real‑life settings) and population validity (generalising to other groups) are crucial for the usefulness of your study. A tightly controlled lab experiment may lack ecological validity, so consider how your task and setting resemble everyday life. Triangulation of methods or replicating across different samples can strengthen overall confidence in your conclusions.
生态效度（结果能否推广到真实生活情境）和人口效度（能否推广到其他群体）对研究的实用性至关重要。严格控制的实验室实验可能缺乏生态效度，因此要考量任务与环境在多大程度上与日常生活相似。方法多元互证或在不同样本中重复，可以增强你对结论的信心。
12. Common Pitfalls and Examiner Tips | 常见错误与考官提示
One common mistake is choosing an inappropriate inferential test – always check the design and data type before deciding. Another is vague operationalisation: simply stating you will measure ‘memory’ or ‘stress’ without precise definitions makes a study unreplicable. Failure to address ethical issues explicitly in the design, such as a missing debrief or consent form, can lose marks.
常见的错误是选择了不恰当的推断检验——决策之前一定要核对设计类型和数据类型。另一个是操作化定义含糊：只说要测量“记忆”或“压力”而没有精确定义，会使研究无法复制。设计中未能明确处理伦理问题，如缺少事后说明或同意书，也会导致丢分。
Many candidates present raw data without summarising it, or miss labelling graph axes. In the discussion, simply repeating results is not enough; you must interpret what the findings mean in light of the original hypothesis and psychological theories. Finally, time management in practical assessments is vital – practice writing reports under timed conditions, and always leave space for a strong, well‑evaluated conclusion.
许多考生只呈现原始数据而未加概括，或漏标了图表坐标轴。在讨论中，仅仅重复结果是不够的；你必须结合最初假设和心理学理论，解释这些发现意味着什么。最后，实践考核中的时间管理至关重要——在限时条件下练习撰写报告，并始终留出空间撰写一个有力且经过充分评价的结论。
Published by TutorHao | Psychology 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