AS CAIE Psychology: High-Frequency Topics and Common Error Analysis | AS CAIE 心理学：高频考点与易错题分析

📚 AS CAIE Psychology: High-Frequency Topics and Common Error Analysis | AS CAIE 心理学：高频考点与易错题分析

In the Cambridge International AS Level Psychology (9990) examination, certain themes and research methods appear time and again. At the same time, many candidates lose marks because of persistent misunderstandings rather than lack of knowledge. This article brings together the topics that are examined most frequently and the mistakes that examiners regularly flag in reports. Each point is presented first in English and then in Chinese, so you can study both the content and the precise wording required for top-band answers.

在剑桥 AS 心理学 (9990) 考试中，某些主题和研究方法反复出现。与此同时，许多考生不是因为知识欠缺，而是因为持续存在的误解而丢分。本文将高频考点与考官报告中反复指出的错误汇集一处。每一个要点都先以英文呈现，再以中文解释，帮助你同时掌握内容和高分答案所需的准确表述。

1. Independent vs. Dependent Variables | 自变量与因变量

The independent variable (IV) is the factor the researcher deliberately changes or manipulates. In an experiment it must have at least two levels, often an experimental condition and a control condition. The dependent variable (DV) is what is measured; it should be operationalised so that it produces numerical data or clearly definable categories.

自变量 (IV) 是研究者有意改变或操纵的因素。在实验中它必须至少有两个水平，通常是一个实验条件和一个控制条件。因变量 (DV) 是被测量的内容；它应当被操作化，以便产生数值数据或可清晰定义的类别。

A common error is labelling the participant’s behaviour as the IV. For example, ‘aggression’ is usually the DV, while the type of model observed (aggressive or non-aggressive) is the IV. Examiners also report that students often fail to operationalise variables precisely; instead of ‘memory was better’ they should write ‘number of words recalled correctly from a list of 20’.

一个常见错误是把参与者的行为标示为自变量。例如，“攻击性”通常是因变量，而观察到的榜样类型（攻击性或非攻击性）才是自变量。考官还指出，学生往往未能精确地操作化变量；他们应当写“从 20 个词中正确回忆出的单词数量”，而不是笼统地写“记忆更好”。

2. Experimental Designs: Independent Groups, Repeated Measures, Matched Pairs | 实验设计：独立组、重复测量与匹配对

Three experimental designs appear in almost every examination series. In an independent groups design different participants are allocated to each condition of the IV. The main strength is that there are no order effects, but a weakness is that participant variables may differ between groups, reducing internal validity.

三种实验设计几乎在每次考试中都会出现。在独立组设计中，不同的参与者被分配到自变量的各个条件。主要优点是没有顺序效应，但缺点是参与者变量可能在组间存在差异，从而降低内部效度。

The repeated measures design uses the same participants in all conditions. This controls participant variables perfectly and requires fewer participants. However, order effects (practice, fatigue) can distort results, so counterbalancing is needed. A repeated mistake is thinking that counterbalancing eliminates order effects entirely; it merely spreads them evenly across conditions.

重复测量设计在所有条件下使用相同的参与者。这可以完美控制参与者变量，且所需参与者更少。然而，顺序效应（练习效应、疲劳效应）可能歪曲结果，因此需要平衡法。一个重复的错误是认为平衡法可以完全消除顺序效应；它只不过是将顺序效应均匀地分散到各个条件中。

Matched pairs design involves pairing participants on a relevant variable (e.g., pre-existing aggression level) and then randomly assigning each member of a pair to a different condition. It reduces participant variables without introducing order effects, but matching is time-consuming and can never be perfect. Many AS scripts incorrectly describe Bandura’s 1961 study as independent groups, whereas it used matched pairs based on prior aggression ratings.

匹配对设计涉及在相关变量（例如已有的攻击性水平）上对参与者进行配对，然后将每对中的成员随机分配至不同条件。它减少了参与者变量且不带来顺序效应，但匹配过程耗时且永远无法完美。许多 AS 答卷错误地将 Bandura 1961 年的研究描述为独立组设计，而该研究实际上是根据先前的攻击性评级采用匹配对设计。

3. Hypotheses: Directional, Non‑directional and Null | 假设：定向、非定向与零假设

A directional (one‑tailed) hypothesis states the expected direction of the effect, e.g. ‘Participants who doodle while listening to a telephone message will recall more names than those who do not doodle.’ A non‑directional (two‑tailed) hypothesis predicts a difference but not its direction, e.g. ‘There will be a difference in the number of names recalled between doodling and control groups.’ The null hypothesis states there will be no difference or relationship.

定向（单尾）假设陈述了预期的效应方向，例如：“在听电话留言时涂鸦的参与者会比不涂鸦的参与者回忆起更多地名。”非定向（双尾）假设预测存在差异但不指明方向，例如：“涂鸦组与控制组在回忆起的名字数量上将存在差异。”零假设则陈述将不存在差异或关系。

A very common mistake is using a directional hypothesis when the previous research is inconsistent, making a non‑directional hypothesis more appropriate. Examiners often deduct marks because students write a directional hypothesis for a study that can only justify a non‑directional prediction. Also, the hypothesis must be operationalised; ‘memory will improve’ is too vague.

一个非常常见的错误是在先前研究不一致的情况下使用定向假设，此时采用非定向假设更为恰当。考官经常因为学生为只能支撑非定向预测的研究写了定向假设而扣分。此外，假设必须被操作化；“记忆会提高”过于模糊。

4. Sampling Methods and Biases | 抽样方法与偏差

Opportunity sampling involves selecting participants who are available and willing at the time of the study. It is quick and convenient, but the sample is often unrepresentative and prone to researcher bias. Random sampling gives every member of the target population an equal chance of being selected, which reduces bias but is not truly achievable in many school‑based investigations.

机会抽样选取研究当时可用且愿意参加的参与者。它快速便捷，但样本常常不具代表性且容易出现研究者偏差。随机抽样使目标人群中的每个成员都有同等被选中的机会，这可以减少偏差，但在许多基于学校的研究中并不真正可行。

Volunteer (self‑selected) sampling relies on individuals responding to advertisements. It can reach a wide audience but often attracts a particular type of participant, limiting generalisability. A frequent error is confusing random sampling with random allocation. Random sampling refers to how participants are selected from the population, whereas random allocation refers to how they are assigned to conditions.

志愿者（自我选择）抽样依赖于个人对广告作出回应。它可以接触到广泛的群体，但常常吸引特定类型的参与者，从而限制了可推广性。一个频繁的错误是混淆随机抽样与随机分配。随机抽样指的是如何从人群中选取参与者，而随机分配指的是如何将参与者分配到各实验条件中。

5. Ethical Issues: Consent, Deception, Debriefing | 伦理问题：知情同意、欺骗与事后说明

Before a study, participants should give informed consent. They must know the aims and procedures, and that they can withdraw at any time. When deception is involved, the researcher must justify that the study would be impossible without it and that no distress will be caused. After the study, participants must be fully debriefed, including revealing any deception and offering the right to withdraw their data.

在研究开始前，参与者应给出知情同意。他们必须了解研究目的和程序，并知道自己可以随时退出。当研究涉及欺骗时，研究者必须证明若无欺骗研究将无法进行，且不会造成任何痛苦。研究结束后，必须对参与者进行充分的事后说明，包括揭示任何欺骗，并提供撤回数据的权利。

Confidentiality is another core principle; data should be reported so that individuals cannot be identified. A typical exam mistake is stating that researchers ‘must’ obtain fully informed consent even when deception is necessary, without explaining the compromise of presumptive consent or prior general consent. Candidates also sometimes use the term ‘anonymity’ when they mean ‘confidentiality’.

保密性是另一个核心原则；数据的报告应使个人无法被识别。一个典型的考试错误是声称研究者“必须”获得完全知情同意，即使在需要欺骗的情况下也不解释推定同意或预先一般同意的折衷办法。考生有时在使用“匿名性”一词时，实际指的是“保密性”。

6. Validity and Reliability: Types and Threats | 效度与信度：类型与威胁

Internal validity concerns whether the IV truly caused the change in the DV, free from confounding variables. External validity refers to the extent findings can be generalised to other settings, populations, and times. Ecological validity is a subset of external validity, asking whether the task and setting reflect real life.

内部效度关注的是自变量是否真的引起了因变量的变化，同时没有混杂变量的干扰。外部效度指的是研究结果能在多大程度上推广到其他情境、人群和时间。生态效度是外部效度的一个子集，询问任务和情境是否反映了真实生活。

Reliability means consistency. If a study is replicated and yields similar results, it is reliable. Internal reliability refers to consistency within a test itself (e.g., split‑half method), while external reliability is stability over time (test‑retest). A widespread confusion is treating validity and reliability as interchangeable. A study can be highly reliable (consistent results) but still lack validity if it measures something irrelevant.

信度指的是一致性。如果一项研究被重复并得出相似结果，它就是信度高的。内部信度指测试本身的一致性（例如分半法），外部信度则是跨时间的稳定性（重测信度）。一个普遍的混淆是将效度和信度视为可以互换。一项研究可以是信度很高（结果一致）的，但如果它测量的是无关内容，则仍然缺乏效度。

A further common error is describing mundane realism as ecological validity. Mundane realism refers to whether the setting physically resembles the real world, whereas ecological validity is about whether the findings can be applied to real‑life behaviour. Bandura’s lab, for instance, had low mundane realism but may still have some ecological validity.

另一个常见错误是将表面真实性描述为生态效度。表面真实性指实验设置是否在物理上与现实世界相似，而生态效度则是关于研究结果能否应用于真实行为。例如，Bandura 的实验室表面真实性低，但仍可能具有一定生态效度。

7. Choosing the Right Inferential Statistical Test | 选择合适的推断统计检验

Selecting the correct inferential test depends on the level of measurement, the experimental design, and the hypothesis. The table below summarises the tests most commonly required in AS Psychology.

选择正确的推断检验取决于测量水平、实验设计和假设。下表总结了 AS 心理学中最常要求的检验。

Data Type	Design	Test
Nominal	Independent groups / association	Chi‑squared
Nominal	Repeated measures	Sign test (or binomial)
Ordinal / interval (non‑parametric)	Independent groups	Mann‑Whitney U
Ordinal / interval (non‑parametric)	Repeated measures	Wilcoxon signed‑ranks

A persistent error is pairing a test with the wrong design. For instance, using Wilcoxon for independent groups will lose all marks for that part. Another error is using the sign test when the data are not nominal; the sign test requires a simple yes/no or +/- categorisation. Always check whether a parametric test (e.g., t‑test) is justified by interval data and normal distribution assumptions, though at AS non‑parametric tests are more commonly expected unless specified.

一个持续存在的错误是把检验与错误的设计配对。例如对独立组使用 Wilcoxon 检验将会失去该部分的全部分数。另一个错误是在数据并非名义数据时使用符号检验；符号检验需要简单的“是/否”或“+/-”分类。务必检查参数检验（如 t 检验）是否有区间数据和正态分布假设的支撑，不过在 AS 课程中，除非特别说明，通常更期望使用非参数检验。

Students also frequently forget to compare the observed value with the critical value correctly. In Chi‑squared and Mann‑Whitney, the observed value must be equal to or greater than the critical value at p ≤ 0.05 for significance; in the sign test and Wilcoxon, the observed value must be equal to or less than the critical value. Mixing these rules is a common examination pitfall.

学生还经常忘记正确比较观察值与临界值。卡方检验和曼‑惠特尼 U 检验中，观察值必须等于或大于 p ≤ 0.05 水平下的临界值才显著；而在符号检验和 Wilcoxon 检验中，观察值必须等于或小于临界值。混淆这些规则是考试中常见的陷阱。

8. Observations and Self‑reports: Strengths and Weaknesses | 观察法与自我报告法：优缺点

Naturalistic observations take place in the participants’ own environment. They often have high ecological validity but low control and are subject to observer bias. Controlled observations, like those in Bandura’s study, allow precise operationalisation of variables and reliable coding, but may lack mundane realism.

自然观察在参与者自己的环境中进行。它们通常具有高生态效度但控制度低，且容易出现观察者偏差。控制性观察，像 Bandura 研究中的那样，允许对变量进行精确操作化和可靠编码，但可能缺乏表面真实性。

Covert observations (participants unaware) reduce demand characteristics but raise ethical concerns because of the lack of consent. Overt observations are ethically clearer but may cause social desirability bias. When evaluating questionnaires, candidates often mention ‘social desirability bias’ correctly but forget to explain that fixed‑choice questions can also lead to response acquiescence, reducing validity.

隐蔽观察（参与者不知情）减少需求特征，但因缺乏同意而引发伦理问题。公开观察在伦理上更清晰，但可能引起社会赞许偏差。在评价问卷时，考生经常正确地提到“社会赞许偏差”，但忘记解释固定选择题也可能导致默认作答偏差，从而降低效度。

Another frequent error is confusing open and closed questions in self‑reports. Open questions generate qualitative data rich in detail but difficult to analyse; closed questions produce quantitative data that are easy to compare but may limit the range of responses. Neither is inherently better; the choice depends on the research aim.

另一个常见错误是混淆自我报告中的开放式和封闭式问题。开放式问题产生细节丰富的定性数据，但难以分析；封闭式问题产生易于比较的定量数据，但可能限制回答的范围。两者并没有天生的优劣之分；选择取决于研究目的。

9. Classic Studies: Bandura et al. (1961) and Andrade (2010) Key Points | 经典研究：Bandura 等人 (1961) 与 Andrade (2010) 要点

In Bandura’s study on aggression, 72 children aged 3‑6 were put into three groups based on pre‑rated aggression. They observed an adult model behaving either aggressively or non‑aggressively towards a Bobo doll, or had no model. The children were then taken into a room with aggressive and non‑aggressive toys, and their behaviour was observed through a one‑way mirror. Boys imitated more physical aggression, especially in the aggressive‑model condition. A common mistake is saying all groups saw a model; the control group did not.

在 Bandura 关于攻击性的研究中，72 名 3 至 6 岁的儿童根据预先评定的攻击性水平被分为三组。他们观察到一位成人榜样对波波玩偶表现出攻击性或非攻击性行为，或者没有榜样。随后儿童被带进一个放有攻击性和非攻击性玩具的房间，通过单向镜观察其行为。男孩模仿了更多的身体攻击，特别是在攻击性榜样条件下。一个常见错误是说所有组都看到了榜样；控制组并没有。

The matched pairs design is a heavily examined feature. Children were matched on aggressiveness and then each member of the pair was assigned to a different condition. This design did not eliminate all participant variables but reduced the main ones. In evaluations, students sometimes write that the study lacked ecological validity because hitting a doll is not real aggression; however, Bandura argued that the study showed learning of specific actions, not necessarily real‑world violence, so the question is about the generalisation of learning, not aggression per se.

匹配对设计是高频考查特征。儿童根据攻击性进行配对，然后每对中的成员被分配至不同条件。这一设计并未消除所有参与者变量，但减少了主要变量。在评价中，学生有时写道该研究缺乏生态效度，因为击打玩偶并非真实的攻击；然而 Bandura 认为该研究展示的是具体行为的学习，不一定是现实世界的暴力，因此问题在于学习的可推广性，而非攻击性本身。

In Andrade’s (2010) study, 40 participants from a participant panel listened to a monotonous telephone message and wrote down names of party‑goers. Half were randomly allocated to a doodling condition (shading shapes) and half to a control. The DV was the number of names correctly recalled plus the number of false alarms. The doodling group recalled more names and made fewer false alarms. This was an independent groups design; a misassumption is that it used repeated measures simply because the task was continuous.

在 Andrade (2010) 的研究中，40 名来自参与者库的参与者聆听了一段单调的电话留言，并写下参加聚会者的名字。一半人被随机分配到涂鸦条件（涂画形状），另一半为控制条件。因变量是正确回忆的名字数量加上错误记忆的数量。涂鸦组回忆了更多名字且错误记忆更少。这是独立组设计；一个错误的假设是认为因为任务是连续的，所以采用了重复测量设计。

The study is often used to illustrate how a controlled lab experiment can still have high mundane realism, as people often doodle during dull tasks. However, the task itself was artificial, so some may argue ecological validity is limited. Candidates sometimes confusingly evaluate Andrade’s study as a field experiment, which it is not; it was conducted in a lab setting with standardised procedures.

该研究常被用来说明受控实验室实验仍能具有较高表面真实性，因为人们常在乏味任务过程中涂鸦。然而，任务本身是人为的，因此有人可能认为生态效度有限。考生有时将其评价为现场实验，这是混淆；该研究是在实验室环境中以标准化程序进行的。

10. Common Exam Pitfalls in AS Psychology | AS 心理学考试常见陷阱

Beyond individual topic errors, several patterns of mistakes appear year after year. One is providing generic evaluation points such as ‘the study had low ecological validity’ without linking the comment to any specific feature of the study. Each evaluation point must be contextualised.

除了单个主题的错误，一些错误模式年复一年地出现。其中之一是提供笼统的评价要点，如“这项研究生态效度低”，却没有将该评论与研究的具体特征联系起来。每个评价要点都必须置于具体情境中。

Another is failing to use psychological terminology precisely. For instance, writing ‘demand characteristics’ when revealing cues actually came from the researcher’s bias, or using ‘reliability’ when the description refers to validity. Examiners look for accurate use of terms such as operationalisation, standardisation, counterbalancing, random allocation, and inter‑rater reliability.

另一个错误是未能准确使用心理学术语。例如，在揭示线索实际上来自研究者偏差时使用“需求特征”，或在描述指向效度时使用“信度”。考官期望准确使用诸如操作化、标准化、平衡法、随机分配和评分者间信度等术语。

Finally, many candidates lose marks by writing too much about a topic that is not assessed, or by failing to address the command word directly. If the question asks ‘Explain’, do not launch into an evaluation full of strengths and weaknesses; if it says ‘Evaluate’, a balance of strengths and weaknesses supported by evidence is essential. Careful reading of the question often prevents the most avoidable drops in marks.

最后，许多考生因写了过多不评分的主题内容，或未能直接回应指令词而丢分。如果问题要求“解释”，就不要展开充满优缺点的评价；如果问题说“评价”，则需要有证据支持的优缺点平衡。仔细阅读问题常常可以防止最可避免的失分。

Published by TutorHao | Psychology Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）