📚 Psychology Research Methods and Experimental Design | 心理学研究方法与实验设计
Psychology is an empirical science. To answer questions about behaviour and mental processes, psychologists rely on systematic research methods and carefully planned experimental designs. Understanding these methods is a core exam requirement for A-level, IB, and introductory university psychology courses.
心理学是一门实证科学。为了回答有关行为和心理过程的问题，心理学家依赖系统性的研究方法和精心设计的实验方案。理解这些方法是 A-level、IB 及大学入门心理学课程的核心考点。
1. The Scientific Method in Psychology | 心理学中的科学方法
The scientific method involves formulating a hypothesis, operationalising variables, collecting data under controlled conditions, analysing results, and drawing evidence-based conclusions. It aims to minimise bias and allow replication.
科学方法包括提出假设、对变量进行操作化定义、在受控条件下收集数据、分析结果并得出基于证据的结论。其目的是尽量减少偏差并允许重复验证。
Hypothesis: a falsifiable, testable prediction about the relationship between variables.
假设：关于变量之间关系的可证伪、可检验的预测。
Operationalisation: defining variables in measurable terms, e.g. 'aggression' as number of loud responses in 15 minutes.
操作化：以可测量的方式定义变量，如将“攻击行为”定义为 15 分钟内大声回应的次数。
Replication: repeating a study to check reliability of findings.
重复验证：重复一项研究以检查结果的信度。
2. Experimental Design: Independent Groups | 实验设计：独立组设计
In an independent groups design, different participants are assigned to the experimental condition and the control condition. For example, one group takes a memory test in silence, another group takes the same test with loud noise.
在独立组设计中，不同的被试被分配到实验条件组和控制条件组。例如，一组在安静环境中进行记忆测试，另一组在嘈杂环境中进行同样的测试。
Advantages: no order/practice effects; participants are less likely to guess the aim.
优点：不存在顺序或练习效应；被试不太可能猜测研究目的。
Disadvantages: individual differences can act as a confounding variable; requires more participants.
缺点：个体差异可能成为混淆变量；需要更多被试。
3. Repeated Measures Design | 重复测量设计
In a repeated measures design, the same participants take part in all conditions. This eliminates participant variables and requires fewer participants.
在重复测量设计中，同一组被试参与所有实验条件。这消除了被试变量并节省被试数量。
Limitations: order effects such as practice, fatigue, or boredom may occur. To control these, researchers use counterbalancing, e.g. ABBA order.
局限：可能出现练习、疲劳或厌倦等顺序效应。为控制这些效应，研究者采用平衡法，如 ABBA 顺序。
Counterbalancing example: Group 1 → Condition A then B; Group 2 → Condition B then A
平衡设计示例：组 1 → 先条件 A 后条件 B；组 2 → 先条件 B 后条件 A
4. Matched Pairs Design | 匹配组设计
Matched pairs involve different participants who are paired according to relevant characteristics such as age, IQ, or gender. One member of each pair goes into each condition.
匹配组设计使用不同的被试，但依据年龄、智商或性别等相关特征将每两名被试配对，每对中的一人进入一种实验条件。
This reduces participant differences while avoiding order effects. However, matching is difficult and time-consuming, and pairs can never be perfectly equal.
这种设计既减少了个体差异又避免了顺序效应。然而，匹配过程困难且耗时，而且配对永远不可能完全等同。
5. Variables: IV, DV, Confounding, and Control | 变量：自变量、因变量、混淆变量与控制变量
Every experiment manipulates at least one independent variable (IV) to measure its effect on the dependent variable (DV). Confounding variables are unwanted influences that may systematically vary with the IV, threatening internal validity.
每个实验至少操纵一个自变量（IV），以测量其对因变量（DV）的影响。混淆变量是可能随自变量系统性变化的额外干扰因素，威胁内部效度。
| Type of Variable | Example in a memory study |
| Independent variable | Noise level (quiet vs loud) |
| Dependent variable | Number of words recalled |
| Confounding variable | Time of day of testing |
| Control variable | Same list length, same instructions |
Control variables are kept constant across conditions to prevent them from becoming confounding variables.
控制变量在各条件中保持恒定，以防止它们变成混淆变量。
6. Sampling Methods | 抽样方法
Psychologists rarely study whole populations. Instead, they select a sample. The sampling method affects how representative the sample is and hence the generalisability of findings.
心理学家很少研究整个总体。相反，他们会选取样本。抽样方法影响样本的代表性，从而影响研究结果的推广性。
Random sampling: every member of the population has an equal chance of being chosen. High representativeness but hard to achieve.
随机抽样：总体中每个成员被选中的机会均等。代表性强但难以实现。
Opportunity sampling: selecting whoever is available, e.g. university students on campus. Easy but biased.
方便抽样：选择可获得的被试，如校园里的大学生。容易但有偏差。
Stratified sampling: population divided into subgroups, then proportional numbers are randomly selected from each.
分层抽样：将总体分为若干子群，然后从每个子群中按比例随机选取。
Volunteer sampling: participants respond to an advert. Ethical but may attract biased types.
自愿抽样：参与者对广告作出回应。符合伦理但可能吸引特定类型的人。
7. Reliability and Validity | 信度与效度
Reliability refers to consistency. A study is reliable if the same results are obtained when repeated under identical conditions. Test-retest, inter-observer, and split-half methods are used to measure reliability.
信度指一致性。如果在相同条件下重复研究能得到相同结果，那么该研究是可信的。测量信度的方法包括重测信度、评分者间信度及折半信度。
Validity refers to whether the study actually measures what it claims to measure. Internal validity concerns causal conclusions; external validity concerns generalising to other people, times, and settings.
效度指研究是否真正测量了它声称要测量的内容。内部效度关系到因果结论；外部效度关系到能否推广到其他人群、时间和情境。
Reliability ≠ Validity: a reliable measure can be consistently wrong
信度 ≠ 效度：一个可信的测量也可能持续出错
8. Ethical Considerations in Research | 研究中的伦理考量
Psychologists must follow ethical guidelines set by bodies such as the British Psychological Society (BPS) or the American Psychological Association (APA). Key principles include informed consent, right to withdraw, debriefing, confidentiality, and protection from harm.
心理学家必须遵守英国心理学会（BPS）或美国心理学会（APA）等机构制定的伦理准则。关键原则包括知情同意、退出权、事后说明、保密性以及避免伤害。
Deception should be avoided unless justified and revealed during debriefing.
应避免欺骗；除非有充分理由且事后说明中告知被试。
Data must be anonymised to protect participant identity.
数据必须匿名化以保护被试身份。
Participants have the right to withdraw at any time, including after the study ends.
被试有权在任何时候退出研究，包括研究结束后。
9. Correlational Research vs Experiments | 相关研究 vs 实验研究
Correlational research examines the strength and direction of association between two or more variables, using the correlation coefficient r. A positive r indicates that as one variable increases, the other also increases; a negative r indicates an inverse relationship.
相关研究检验两个或多个变量之间关联的强度与方向，使用相关系数 r。正的 r 表示一个变量增加时另一个也增加；负的 r 表示反向关系。
r = +1.0 → 完全正相关；r = 0 → 无相关；r = -1.0 → 完全负相关
Critical point: correlation does not imply causation. Only experiments, with manipulation of the IV and random assignment, can establish cause-and-effect relationships.
关键点：相关不等于因果。只有通过操纵自变量和随机分配的实验才能确立因果关系。
10. Descriptive and Inferential Statistics | 描述统计与推断统计
Descriptive statistics summarise data: measures of central tendency (mean, median, mode) and measures of dispersion (range, standard deviation). Inferential statistics tell us whether results are likely to be due to chance.
描述统计用来概括数据：集中趋势量数（平均数、中位数、众数）和离散量数（全距、标准差）。推断统计则告诉我们结果是否可能由偶然因素导致。
Significance testing uses a probability value p. In psychology, the conventional threshold is p ≤ 0.05, meaning there is less than a 5% probability that the result occurred by chance if the null hypothesis is true.
显著性检验使用概率值 p。在心理学中，常规临界值为 p ≤ 0.05，意味着如果零假设成立，结果由偶然产生的概率小于 5%。
11. Choosing the Right Experimental Design | 选择合适的实验设计
There is no single best design. The choice depends on the research question, the availability of participants, the risk of order effects, and ethical constraints.
没有唯一最佳的设计。选择取决于研究问题、被试可得性、顺序效应的风险以及伦理限制。
| Design | Best used when |
| Independent groups | Avoiding order effects is more important than controlling participant variables |
| Repeated measures | Participant numbers are limited and order effects can be counterbalanced |
| Matched pairs | Important individual differences exist but repeated measures are impractical |
In exam questions, always justify your choice by linking the design to possible confounding variables, order effects, and sampling bias.
在考试题目中，务必通过联系可能的混淆变量、顺序效应和抽样偏差来论证你的设计选择。
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