📚 Year 13 OCR Psychology: Experimental & Practical Assessment Essentials | OCR A-Level 心理学：实验与实践考核要点
Experimental and practical skills sit right at the heart of the OCR A-Level Psychology course. In Year 13, you are expected not only to understand classic studies but also to design, conduct, analyse, and critique your own investigations. This guide pulls together the key assessment points that frequently appear in Paper 1 (Research Methods) and the practical components across the two-year course, helping you turn methodical rigour into higher marks.
实验与实践技能是 OCR A-Level 心理学课程的核心。进入 Year 13 后，你不仅要理解经典研究，还需要自行设计、实施、分析和批判心理学研究。本文梳理了试卷一（研究方法）以及整个两年课程中实践考核的高频要点，帮助你用严谨的方法论换取更高分数。
1. The Role of Research Methods in OCR Psychology | 研究方法在 OCR 心理学中的地位
Research methods form the backbone of the OCR Psychology H567 specification. Every section of the course, from social influence to forensic psychology, requires you to evaluate how data were gathered and interpret the methodological choices researchers made. Your practical investigation portfolio, though not formally submitted for moderation, is the foundation for the skills-based questions in Paper 1 Section B.
研究方法是 OCR 心理学 H567 课程的主干。从社会影响到法医心理学的每一个单元，都要求你评价数据是如何收集的，并解释研究者为何做出特定的方法选择。你的实践调查档案虽然不作为正式样本提交审核，却是试卷一 B 部分技能型考题的直接基础。
- Examiners will ask you to spot weaknesses in hypothetical studies.
- 考官会要求你找出虚拟研究中的方法缺陷。
- You must be able to propose improvements in validity, reliability, and ethics.
- 你必须能提出关于效度、信度和伦理的改进方案。
- Knowledge of the research cycle – from aim to reporting – is assessed repeatedly.
- 从研究目标到报告撰写的整个研究循环会被反复考查。
2. Experimental Designs and Their Trade-offs | 实验设计及其权衡
Independent groups, repeated measures, and matched pairs are the three core experimental designs you must master. Each carries inherent advantages and specific threats to internal validity, and the OCR mark scheme rewards precise comparisons. You should be able to match the design to the research context and justify your choice using psychological terminology.
独立组设计、重复测量设计和配对组设计是你必须掌握的三种核心实验设计。每种设计都带有天然优势和特定的内部效度威胁，OCR 评分方案青睐精确的比较。你应当能够将设计与研究情境相匹配，并使用心理学术语说明理由。
| Design 设计 | Main Strength 主要优点 | Main Weakness 主要缺点 | Control Strategy 控制策略 |
|---|---|---|---|
| Independent Groups 独立组 | No order effects 无顺序效应 | Participant variables 被试变量 | Random allocation 随机分配 |
| Repeated Measures 重复测量 | Controls participant EVs 控制被试外扰变量 | Order/fatigue effects 顺序/疲劳效应 | Counterbalancing 抵消平衡 |
| Matched Pairs 配对组 | Reduces participant variables 减少被试变量 | Time-consuming and imperfect matching 耗时且匹配不完美 | Pre-test on key trait 关键特质前测 |
When identifying the design in a novel scenario, look for phrases such as "each participant took part in all conditions" (repeated measures) or "participants were split into two separate groups" (independent groups).
在识别新情境中的实验设计时，留意“每名被试参加了所有条件”（重复测量）或“被试被分成两个独立组”（独立组）等表述。
3. Variables, Operationalisation, and Hypothesis Writing | 变量、操作化与假设撰写
Clear operationalisation transforms a vague research idea into an empirical investigation. You will be asked to identify the independent variable (IV), dependent variable (DV), and at least one extraneous variable (EV) that should be controlled. In Year 13, you are expected to move beyond simple dichotomies and operationalise variables such as anxiety, memory, or helping behaviour with measurable indicators.
清晰的操作化能将模糊的研究构想转化为实证调查。你将被要求识别自变量（IV）、因变量（DV）以及至少一个应加以控制的额外变量（EV）。到了 Year 13，你需要超越简单的二分，用可量化的指标来操作化诸如焦虑、记忆或助人行为等变量。
- IV: "Time of day" → operationalised as "09:00 vs 15:00 testing session".
- 自变量：“每天时段” 可操作化为 “09:00 与 15:00 的测试时段”。
- DV: "Cognitive performance" → operationalised as "number of correct answers on a 20-item digit span test".
- 因变量：“认知表现” 可操作化为 “在 20 题数字广度测验中正确回答的数目”。
- Hypothesis: Must state both conditions, the population, and the predicted outcome. A directional (one-tailed) hypothesis is used when prior research strongly points one way; otherwise, keep it non-directional (two-tailed).
- 假设：必须陈述两种条件、目标人群和预测结果。当先前研究强烈指向某一方向时使用方向性（单尾）假设；否则保持非方向性（双尾）。
Directional hypothesis: Participants tested at 09:00 will recall significantly more digits than those tested at 15:00.
方向性假设：09:00 测试的被试将比 15:00 测试的被试显著回忆更多数字。
4. Sampling Methods and Ethical Safeguards | 抽样方法与伦理保障
Opportunity, volunteer, random, stratified, and systematic sampling each carry distinct implications for representativeness and generalisability. In the extended-response questions, you may need to recommend an alternative sampling method and explain how it would improve external validity.
机会抽样、志愿者抽样、随机抽样、分层抽样和系统抽样各自对代表性和推广性有着不同的含义。在拓展作答题目中，你可能需要推荐一种替代抽样方法，并解释它将如何增强外部效度。
| Method 方法 | Bias Risk 偏差风险 | Generalisability 推广性 |
|---|---|---|
| Opportunity 机会 | High – researcher selects available people 高——研究者选取便利人群 | Low 低 |
| Volunteer 志愿者 | Self-selection bias 自荐偏差 | Moderate 中等 |
| Random 随机 | Low – every member has equal chance 低——所有人机会均等 | High 高 |
| Stratified 分层 | Low – reflects subgroups proportionally 低——按比例反映亚群体 | High 高 |
Ethical guidelines from the British Psychological Society (BPS) are embedded throughout the OCR specification. When designing a study, always address informed consent, right to withdraw, confidentiality, protection from harm, and debriefing. Propose specific strategies: gaining presumptive consent, offering counselling referrals, or using anonymised codes.
英国心理学会（BPS）的伦理准则贯穿 OCR 课程始终。设计研究时，务必处理知情同意、退出权、保密、免受伤害以及事后解释。提出具体策略：获取推定同意、提供心理咨询转介或使用匿名编码。
5. Experimental Control and Extraneous Variables | 实验控制与外扰变量
Standardisation of procedure, instructions, and environment is the first line of defence against extraneous variables. Even a seemingly minor detail – such as the gender of the experimenter or the temperature of the room – can become a confounding variable if it systematically co-varies with the IV.
程序、指导语和环境的标准化是抵御外扰变量的第一道防线。即使是一个看似微小的细节——例如实验者性别或室温——只要它随自变量发生系统性共变，就可能成为混淆变量。
- Demand characteristics: Participants guess the aim and alter behaviour. Use single-blind procedures or deception followed by thorough debriefing.
- 需求特征：被试猜测目的并改变行为。采用单盲程序或欺骗后细致的事后解释。
- Experimenter effects: Researcher’s expectations influence outcomes. Use double-blind protocols where neither participants nor the data-gathering researcher know the condition allocation.
- 实验者效应：研究者的期望影响结果。采用双盲方案，使被试和收集数据的研究者均不知晓条件分配。
- Order effects in repeated measures: Counterbalancing (ABBA) is essential. For example, half the participants complete condition A then B, and half complete B then A.
- 重复测量中的顺序效应：抵消平衡（ABBA）至关重要。例如，一半被试先完成条件 A 再完成 B，另一半先完成 B 再完成条件 A。
6. Descriptive Statistics and Visual Representation | 描述统计与图表呈现
You must be able to calculate measures of central tendency (mean, median, mode) and dispersion (range, variance, standard deviation). In Year 13, you go further by evaluating which measure is most appropriate for a given data set: a skewed distribution inflates the mean, making the median a better summary; when extreme outliers exist, the interquartile range often tells a more honest story than the range.
你必须能够计算集中量数（平均数、中位数、众数）和离散量数（全距、方差、标准差）。在 Year 13，你还会进一步评价哪种量数最适合给定的数据集：偏态分布会拉高平均数，此时中位数是更佳概括；当存在极端异常值时，四分位距通常比全距更真实。
Standard deviation s = √[Σ(x − x̄)² ÷ (n − 1)]
标准差 s = √[Σ(x − x̄)² ÷ (n − 1)]
Choosing the right graph is a common mark point. Bar charts are for discrete categories and means comparison; histograms display continuous data distributions; scattergrams depict correlations. Always label both axes with the operationalised variable names and units, and provide a clear, descriptive title.
选择正确的图表是常见的得分要点。条形图适合离散类别和均值比较；直方图呈现连续数据分布；散点图表现相关关系。切记在两个坐标轴上标注操作化变量的名称和单位，并写明清晰、具有描述性的标题。
7. Probability, Levels of Significance, and Errors | 概率、置信水平与错误类型
quences of false positives are severe. You are expected to interpret an observed value against a critical value table, making a binary decision about the null hypothesis.
在 OCR 心理学中，大多数情境下常规置信阈值为 p ≤ 0.05，但当假阳性的后果严重时可能采用更严格水平（p ≤ 0.01）。你需要根据临界值表解释观测值，并对零假设做出二元判断。
- Null hypothesis (H₀): States there is no effect, no difference, or no correlation.
- 零假设 (H₀)：声明无效应、无差异或无相关。
- Alternative hypothesis (H₁): The prediction you actually test – already operationalised.
- 备择假设 (H₁)：你实际检验的预测——已经操作化。
- Type I error: Rejecting a true null (false positive). Likened to shouting 'fire' when there is none.
- I 类错误：拒斥了正确的零假设（假阳性）。好比无火情却高喊“着火了”。
- Type II error: Failing to reject a false null (false negative). Missing a genuine effect due to weak power or a tiny sample.
- II 类错误：未能拒斥错误的零假设（假阴性）。因统计效力弱或样本量太小而错过真实效应。
8. Inferential Statistics: Choosing and Justifying the Test | 推论统计：选择并论证统计检验
Selecting the appropriate inferential test is a skill that separates top-grade students from the rest. The decision tree depends on three questions: (1) Is the researcher testing a difference or a correlation? (2) What is the experimental design – independent groups, repeated measures, or matched pairs? (3) What is the level of measurement – nominal, ordinal, or interval/ratio?
选择合适的推论统计检验是拉开高分差距的关键技能。决策树取决于三个问题：(1) 研究者在检验差异还是相关？ (2) 实验设计是什么——独立组、重复测量还是配对组？ (3) 测量层次是什么——称名、顺序还是等距/比率？
| Test 检验 | Design/Relationship 设计/关系 | Data Level 数据层次 | Symbol 符号 |
|---|---|---|---|
| Chi-Square 卡方 | Independent groups, difference 独立组, 差异 | Nominal 称名 | χ² |
| Mann-Whitney U 曼-惠特尼 U | Independent groups, difference 独立组, 差异 | Ordinal/Interval (non-parametric) 顺序或非参数等距 | U |
| Wilcoxon Signed-Ranks 威尔科克森符号秩 | Repeated measures/Matched pairs 重复测量/配对组 | Ordinal/Interval (non-parametric) 顺序或非参数等距 | T |
| Spearman’s Rho 斯皮尔曼等级相关 | Correlation 相关 | Ordinal/Interval (non-parametric) 顺序或非参数等距 | rₛ |
For parametric tests like the related t-test or unrelated t-test, you must also justify that the data meet the assumptions: approximately normal distribution, interval data, and homogeneity of variance. OCR expects you to state these conditions explicitly.
对于参数检验，如相关 t 检验或无关 t 检验，你还需要论证数据满足假设：近似正态分布、等距数据和方差齐性。OCR 要求你明确陈述这些条件。
χ² = Σ((O − E)² ÷ E)
χ² = Σ((O − E)² ÷ E)
9. Writing the Practical Report in the Correct Format | 按正确格式撰写实践报告
The practical report structure is a near-certain feature of the exam. Every section – Abstract, Introduction, Method, Results, Discussion, and References – must follow the conventions of a psychological journal article. In Paper 1, you might be given a raw draft and asked to identify mistakes in formatting, sequencing, or content.
实践报告的结构几乎是考试的必考内容。每个部分——摘要、引言、方法、结果、讨论和参考文献——都必须遵循心理学期刊文章的规范。在试卷一，你可能会拿到一份原始草稿，并被要求找出格式、顺序或内容方面的错误。
- Abstract: Written last, summarises aim, hypothesis, method, results, and conclusion in about 150 words.
- 摘要：最后撰写，用约 150 字概括研究目的、假设、方法、结果和结论。
- Introduction: Funnel structure – broad background, critical evaluation of previous studies, justification for the current study, and statement of operationalised hypotheses.
- 引言：漏斗结构——广泛的背景，对先前研究的批判评价，本研究的意义陈述，以及操作化假设的表述。
- Method: Contains subsections on design, participants (with sampling), materials/apparatus, and procedure. Enough detail for replication.
- 方法：包含设计、被试（及抽样）、材料/仪器和程序等子部分。细节足供重复实验。
- Results: Descriptive statistics first, then inferential test with calculated and critical values, followed by a statement of significance. Do not interpret yet.
- 结果：先呈现描述统计，再呈现推论检验的计算值和临界值，随后进行显著性陈述。此处暂不解释。
- Discussion: Interpret findings, relate back to the hypothesis, compare with previous research, acknowledge limitations, and suggest improvements.
- 讨论：解释发现，与假设相联系，与先前研究进行比较，承认局限性，并提出改进建议。
10. Reliability, Validity, and Methodological Critique | 信度、效度与方法论批判
OCR examiners frequently ask you to assess 'reliability' and 'validity' with surgical precision. Reliability concerns consistency – can the measurement be repeated with similar outcomes? Validity concerns truth – does the tool measure what it claims to measure?
OCR 考官常要求你精准地评析“信度”和“效度”。信度关乎一致性——这一测量能否被重复且得到相似结果？效度关乎真实性——该工具是否测量了其声称要测量的东西？
- Internal reliability: Consistency within a test (e.g., split-half method).
- 内部信度：测验内部的一致性（如分半法）。
- External reliability: Consistency over time (test-retest) or across observers (inter-rater).
- 外部信度：跨时间的一致性（重测）或跨观察者的一致性（评分者间信度）。
- Internal validity: Whether the IV truly caused the change in the DV – threatened by demand characteristics, experimenter effects, and poor controls.
- 内部效度：IV 是否真正导致了 DV 的变化——受到需求特征、实验者效应和控制不佳的威胁。
- External validity: Generalisability – ecological validity (real-world setting), population validity (sample to target), and temporal validity (historical period).
- 外部效度：推广性——生态效度（真实环境）、人群效度（样本到目标人群）和时代效度（历史时期）。
When you propose improvements, link each suggestion to a specific concept: 'To raise inter-rater reliability, the observers could use a standardised coding sheet and conduct a pilot correlation check.'
当你提出改进建议时，将每一条建议与特定概念挂钩：“为提高评分者间信度，观察者可使用标准化编码表，并进行先导相关检查。”
11. Time Management and Command Words in the Exam | 考试时间管理与指令词
Practical-based questions range from short 2-mark identification items to extended 12-mark 'design an investigation' prompts. You must know what each command word demands: 'Identify' requires a brief term or phrase; 'Explain' asks for a point plus elaboration; 'Discuss' or 'Evaluate' requires multiple paragraphs weighing strengths and limitations.
实践类题目涵盖从 2 分简答题到 12 分的“设计一项调查”拓展题。你必须清楚每个指令词的要求：“识别”(Identify) 仅需简短术语或短语；“解释”(Explain) 要求要点加阐述；“讨论/评价”(Discuss/Evaluate) 要求多个段落权衡优势与不足。
- For 'Design a study' questions, follow a mental checklist: aim, hypothesis, design, IV & DV, sampling, procedure, ethics, data analysis, and one strength plus one limitation.
- 对于“设计一项研究”类问题，遵循思维清单：目的、假设、设计、IV 与 DV、抽样、程序、伦理、数据分析，以及一个优势和一个局限性。
- If you are shown a raw data set, calculate descriptive statistics before scanning for inferential test cues. Show all working – marks are awarded for process.
- 若给出原始数据集，先计算描述统计，再搜寻推论检验线索。展示所有演算过程——过程有分。
- Keep an eye on the clock: roughly one minute per mark is a safe guideline.
- 留意时间：大约每分钟 1 分是一条安全的原则。
12. Common Mistakes and Final Revision Tips | 常见错误与终极复习建议
A frequent error is describing an experimental design as 'correlational' when the researcher has merely measured two co-occurring variables without manipulation. A correlation is a relationship, not a design for establishing cause-and-effect. Another pitfall is writing generic ethical comments such as 'obtain consent' without specifying how – for instance, 'send an opt-out consent form to parents two weeks before the study and provide a child-friendly information sheet'.
一个常见错误是把实验设计说成“相关”，当研究者只是测量两个同时发生的变量而不加操纵时尤其如此。相关是一种关系，而非确立因果的设计。另一个陷阱是写出笼统的伦理评论，如“获取同意”，却未具体说明如何操作——例如，“在研究前两周向家长发送选择退出知情同意书，并提供适龄被试的信息告知单”。
- Revise using real past-paper scenarios: annotate the IV, DV, design, and one fatal flaw for each.
- 使用真实的历年真题情境复习：为每一道题标注 IV、DV、设计和一个致命缺陷。
- Create a table of inferential tests with their design conditions and learned mnemonics (Carrots Should Come Mashed With Swede Turnip: Chi-square, Mann-Whitney, Wilcoxon, Spearman’s Rho, Related t-test, Unrelated t-test).
- 制作一张推论检验表格，配上设计条件和记忆口诀。
- Practise writing full practical reports under timed conditions, so that the structure becomes automatic.
- 在限时条件下练习撰写完整的实践报告，让结构变成肌肉记忆。
- Use technical language deliberately: write 'the study lacked ecological validity because the task took place in an artificial laboratory setting' rather than 'it wasn’t real life'.
- 刻意使用专业术语：写“该研究缺乏生态效度，因为任务在人工实验室环境中进行”，而不是“它不是真实生活”。
Published by TutorHao | Psychology Revision Series |
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