📚 Research Methods in Social Psychology: Experimental and Correlational Approaches | 社会心理学主要研究方法：实验与相关
Social psychology seeks to understand how individuals’ thoughts, feelings, and behaviours are influenced by the actual, imagined, or implied presence of others. To investigate these phenomena systematically, researchers rely on two principal methodological frameworks: experimental methods and correlational methods. Each approach offers distinct advantages and limitations, and understanding their differences is essential for critically evaluating psychological research.
社会心理学旨在理解个体的思想、情感和行为如何受到他人真实、想象或暗示的存在所影响。为了系统研究这些现象，研究者主要依赖两种方法论框架：实验方法和相关方法。每种方法各有其独特的优势和局限，理解二者的差异对于批判性评价心理学研究至关重要。
1. The Scientific Foundation of Social Psychology | 社会心理学的科学基础
Social psychology emerged as a distinct discipline in the early twentieth century, with researchers such as Norman Triplett (1898) and Kurt Lewin (1935) laying the groundwork for systematic empirical investigation. The field operates on the assumption that social behaviour can be studied objectively and quantitatively, in the same way that natural scientists study physical phenomena.
社会心理学在二十世纪初成为一个独立学科，诺曼·特里普莱特（1898年）和库尔特·勒温（1935年）等研究者为系统的实证研究奠定了基础。该领域的基本假设是，社会行为可以像自然科学家研究物理现象一样被客观、定量地研究。
At the core of scientific inquiry lies the hypothetico-deductive model, in which researchers formulate specific hypotheses derived from broader theories, then collect empirical data to test those predictions. Regardless of whether the chosen approach is experimental or correlational, all social psychological research must meet the standards of reliability, validity, and replicability.
科学探究的核心是假设-演绎模型，研究者从更广泛的理论中提出具体的假设，然后收集经验数据来检验这些预测。无论选择实验方法还是相关方法，所有社会心理学研究必须满足信度、效度和可复制性的标准。
2. The Experimental Method: Establishing Causality | 实验方法：确立因果关系
The experimental method is considered the ‘gold standard’ in social psychology because it is the only method that can definitively establish cause-and-effect relationships. In a true experiment, the researcher manipulates one or more independent variables (IV) and measures their effect on one or more dependent variables (DV), while controlling extraneous variables that might confound the results.
实验方法被认为是社会心理学中的“金标准”，因为它是唯一能够明确建立因果关系的方法。在真正的实验中，研究者操纵一个或多个自变量（IV），并测量它们对一个或多个因变量（DV）的影响，同时控制可能混淆结果的额外变量。
Consider a classic example: Milgram’s (1963) obedience study. The independent variable was the proximity of the experimenter or the victim, which was systematically varied across conditions. The dependent variable was the maximum shock level the participant was willing to administer. Because Milgram randomly assigned participants to different conditions, he could confidently attribute differences in obedience levels to the manipulated variables rather than to pre-existing differences among participants.
考虑一个经典例子：米尔格拉姆（1963年）的服从研究。自变量是实验者或受害者的接近程度，这一变量在不同条件下被系统变化。因变量是参与者愿意施加的最大电击水平。由于米尔格拉姆将参与者随机分配到不同条件，他可以有信心地将服从水平的差异归因于被操纵的变量，而非参与者间的既有差异。
Key features of a true experiment include: (a) random assignment of participants to conditions; (b) active manipulation of the independent variable; (c) control over extraneous variables; and (d) measurement of the dependent variable under standardised conditions. These features enable researchers to make strong causal inferences.
真正实验的关键特征包括：（a）参与者被随机分配到不同条件；（b）主动操纵自变量；（c）控制额外变量；（d）在标准化条件下测量因变量。这些特征使研究者能够做出强有力的因果推断。
3. Types of Experimental Designs | 实验设计的类型
Social psychologists employ several variations of experimental design, each with its own logic and applications. The most common distinction is between independent measures (between-subjects) designs and repeated measures (within-subjects) designs.
社会心理学家采用多种实验设计变体，每种设计都有其自身的逻辑和应用。最常见的区分是独立测量（被试间）设计和重复测量（被试内）设计。
In an independent measures design, different participants are assigned to different experimental conditions. This approach avoids order and carry-over effects because each participant experiences only one condition. However, it requires larger sample sizes and risks problems from individual differences between groups. A powerful variant is the matched-pairs design, in which participants are matched on relevant characteristics (e.g., age, gender, personality) before being assigned to different conditions.
在独立测量设计中，不同的参与者被分配到不同的实验条件。这种方法避免了顺序效应和延续效应，因为每个参与者只经历一种条件。然而，它需要更大的样本量，并可能因组间个体差异而产生问题。一种有力的变体是匹配设计，参与者先在相关特征（如年龄、性别、人格）上进行匹配，然后被分配到不同条件。
In a repeated measures design, the same participants take part in all conditions of the experiment. This design controls for individual differences perfectly and increases statistical power with smaller samples. However, it is susceptible to order effects, such as practice effects and fatigue effects, which are typically addressed through counterbalancing—systematically varying the order of conditions across participants.
在重复测量设计中，同一组参与者参加实验的所有条件。这种设计完美地控制了个体差异，并以较小的样本提高了统计效力。然而，它容易受到顺序效应的影响，如练习效应和疲劳效应，通常通过平衡设计——系统改变参与者接受条件的顺序——来解决这些问题。
Independent Measures | 独立测量: different participants per condition | 不同参与者参与不同条件
Repeated Measures | 重复测量: same participants in all conditions | 相同参与者参与所有条件
4. Variables: Independent, Dependent, and Confounding | 变量：自变量、因变量和混淆变量
The independent variable (IV) is the factor that the researcher deliberately manipulates in an experiment. It is typically a social situation or context variable, such as group size, the presence of authority figures, or the framing of a persuasive message. The dependent variable (DV) is the measured outcome that is expected to change as a result of the IV manipulation—for example, helping behaviour, conformity rates, or attitude scores.
自变量（IV）是研究者在实验中刻意操纵的因素。它通常是一个社会情境或背景变量，如群体规模、权威人物的在场、或说服信息的框架。因变量（DV）是预期会因自变量操纵而改变的测量结果——例如，助人行为、从众率或态度分数。
Confounding variables are extraneous factors that vary systematically with the IV and provide alternative explanations for the results. For example, if participants in one condition happen to be tested in the morning and participants in another condition in the afternoon, time of day could become a confound. Methodological controls—such as standardised instructions, automated procedures, and double-blind protocols—are essential to minimise confounds.
混淆变量是与自变量协同变动并能为结果提供替代解释的额外因素。例如，如果某个条件下的参与者在早上接受测试，而另一条件下的参与者在下午接受测试，那么一天中的时分就可能成为一个混淆因素。方法学控制——如标准化指导语、自动化程序和双盲协议——对于最小化混淆至关重要。
Additionally, researchers must consider the distinction between an extraneous variable (any variable other than the IV that may affect the DV) and a confounding variable (an extraneous variable that actually varies with the IV). A well-controlled experiment ensures that extraneous variables are held constant or balanced across conditions, preventing them from becoming confounds.
此外，研究者必须区分额外变量（除自变量外可能影响因变量的任何变量）与混淆变量（确实与自变量协同变化的额外变量）之间的区别。设计良好的实验确保额外变量在不同条件下保持恒定或平衡，从而防止它们成为混淆变量。
5. Experimental Control and Validity | 实验控制与效度
Control is the essential feature that distinguishes experiments from other research methods. Social psychologists seek control at multiple levels: controlling the research environment through laboratory settings, controlling participant expectations through deception or cover stories, and controlling extraneous variables through randomisation and standardisation.
控制是区分实验与其他研究方法的基本特征。社会心理学家在多个层面寻求控制：通过实验室环境控制研究环境，通过欺骗或掩饰故事控制参与者的预期，通过随机化和标准化控制额外变量。
Internal validity refers to the extent to which the research design allows confident causal conclusions about the effect of the IV on the DV. Threats to internal validity include history (unexpected events during the study), maturation (participants changing over time), testing effects (prior exposure to dependent measures), and selection bias. Random assignment and rigorous controls mitigate these threats.
内部效度指研究设计允许对自变量对因变量的影响做出有信心的因果结论的程度。对内部效度的威胁包括历史效应（研究期间的意外事件）、成熟效应（参与者随时间变化）、测试效应（先前接触过因变量测量）和选择偏差。随机分配和严格的控制可减轻这些威胁。
External validity, by contrast, concerns the generalisability of findings to other populations, settings, and times. A study conducted exclusively on undergraduate students in a university laboratory may have high internal validity but limited external validity. This tension between internal and external validity is a central methodological debate in social psychology. Techniques such as field experiments and replication across diverse samples help enhance external validity.
与此相对，外部效度涉及研究结果对其他人群、情境和时间点的可推广性。一项仅针对大学实验室中的本科生进行的研究可能具有较高的内部效度，但外部效度有限。内部效度与外部效度之间的这种张力是社会心理学的核心方法论辩论。现场实验和跨多样化样本的重复研究等技术有助于提高外部效度。
6. The Correlational Method: Measuring Associations | 相关方法：测量关联
The correlational method examines whether and to what extent two or more variables are related, without manipulating any variables. Researchers measure the naturally occurring co-variation between variables, expressed as a correlation coefficient (r) that ranges from -1.00 to +1.00. The sign indicates direction, while the absolute value indicates strength.
相关方法考察两个或多个变量之间是否存在关联以及关联的程度如何，不操纵任何变量。研究者测量变量之间自然发生的共变性，用相关系数（r）表示，范围从-1.00到+1.00。正负号表示方向，绝对值表示强度。
A positive correlation (e.g., r = +0.60) indicates that as one variable increases, the other tends to increase as well. A negative correlation (e.g., r = -0.45) indicates that as one variable increases, the other tends to decrease. A correlation near zero (e.g., r = +0.08) indicates little or no systematic relationship. The correlation coefficient also provides a measure of effect size: r² represents the proportion of variance shared between the two variables.
正相关（如 r = +0.60）表示当一个变量增加时，另一个变量往往也增加。负相关（如 r = -0.45）表示当一个变量增加时，另一个变量往往减少。接近零的相关（如 r = +0.08）表示几乎没有系统性的关系。相关系数还提供了效应量的度量：r² 代表两个变量共享的方差比例。
Common correlational techniques in social psychology include surveys and questionnaires, observational studies, and archival or secondary data analysis. For example, researchers studying the relationship between social media use and self-esteem might administer validated scales to a large, diverse sample and analyse the resulting correlation coefficient.
社会心理学中常见的相关技术包括问卷调查、观察研究和档案或二手数据分析。例如，研究社交媒体使用与自尊关系的研究者可能会向一个大型、多样化的样本施测经过验证的量表，并分析所得相关系数。
7. Correlation Is Not Causation | 相关不等于因果
The most fundamental limitation of the correlational method is that it cannot establish causality. The phrase ‘correlation does not imply causation’ is a cornerstone of research methodology. Three possible explanations exist for any observed correlation: X causes Y, Y causes X, or a third variable Z causes both X and Y. This third possibility is known as the third-variable problem or the confounding-variable problem.
相关方法最基本的局限是它无法确立因果关系。“相关不等于因果”这一说法是研究方法论的基石。任何观察到的相关都存在三种可能的解释：X导致Y、Y导致X、或者第三个变量Z同时导致X和Y。这第三种可能性被称为第三方变量问题或混淆变量问题。
Consider a well-known social psychological example: researchers observe a negative correlation between violent media consumption and prosocial behaviour in adolescents. Does violent media reduce prosocial behaviour? Or do adolescents who are less prosocial choose to consume more violent media? Or does a third variable, such as poor parental supervision, independently increase both violent media use and decreased prosocial behaviour?
考虑一个著名的社会心理学例子：研究者观察到青少年暴力媒体消费与亲社会行为之间存在负相关。是暴力媒体减少了亲社会行为？还是亲社会行为较少的青少年选择消费更多暴力媒体？或者某个第三方变量，如父母监管不力，独立地同时增加了暴力媒体使用并减少了亲社会行为？
The directionality problem is equally significant: the observed correlation cannot tell us which variable precedes the other in time. Longitudinal designs, in which the same individuals are measured at multiple time points, can provide some evidence about temporal precedence and help rule out the directionality problem. However, even longitudinal correlational studies cannot definitively control for all third variables—only experimental manipulation can.
方向性问题同样重要：观察到的相关无法告诉我们哪个变量在时间上先发生。纵向设计——对同一组个体在多个时间点进行测量——可以提供关于时间先后顺序的某些证据，并有助于排除方向性问题。然而，即使纵向相关研究也无法最终控制所有第三方变量——只有实验操纵才能做到。
Three possible explanations for r ≠ 0 | r ≠ 0 的三种可能解释
X → Y | X 导致 Y
Y → X | Y 导致 X
Z → X and Z → Y | Z 同时导致 X 和 Y
8. Experimental vs. Correlational Designs: A Comprehensive Comparison | 实验设计与相关设计：综合比较
Choosing between experimental and correlational methods requires thoughtful consideration of the research question, ethical constraints, practical feasibility, and intended level of causal inference. The table below summarises the key distinctions.
在实验方法和相关方法之间进行选择需要深思熟虑研究问题、伦理约束、实际可行性和预期的因果推断水平。下表概述了关键区别。
| Feature | 特征 | Experimental | 实验 | Correlational | 相关 |
| Primary goal | 主要目标 | Establish causation | 确立因果 | Describe associations | 描述关联 |
| IV manipulation | 自变量的操纵 | Active manipulation | 主动操纵 | None (measured only) | 无（仅测量） |
| Random assignment | 随机分配 | Yes | 是 | No | 否 |
| Primary strength | 主要优势 | High internal validity | 高内部效度 | High external validity, can study unethical topics | 高外部效度，可研究不合伦理的议题 |
| Primary limitation | 主要局限 | May lack generalisability, ethical limits | 可能缺乏推广性，伦理有限 | Cannot establish causality, third-variable problem | 无法确立因果，存在第三变量问题 |
| Typical design | 典型设计 | Laboratory or field experiments | 实验室或现场实验 | Surveys, observations, archival research | 调查、观察、档案研究 |
It is important to note that experiments and correlational studies are complementary rather than antagonistic. A rigorous research programme might begin with correlational studies to identify interesting relationships, then proceed to experiments to test causality, and finally return to correlational research to assess generalisability in real-world settings.
值得注意的是，实验和相关研究是互补而非对立的。一个严谨的研究项目可能从相关研究开始，以识别有趣的关联，然后通过实验来检验因果关系，最后再回到相关研究以评估在现实世界环境中的可推广性。
9. Beyond Laboratory Experiments: Field Experiments and Quasi-Experiments | 超越实验室实验：现场实验与准实验
While the laboratory experiment is the most common experimental design in social psychology, field experiments and quasi-experiments extend the logic of experimentation to real-world and naturally occurring social contexts.
尽管实验室实验是社会心理学中最常见的实验设计，但现场实验和准实验将实验的逻辑扩展到现实世界和自然发生的社会情境中。
In a field experiment, the researcher manipulates an IV in a natural setting where participants are unaware they are being studied. For example, Latané and Darley’s (1970) investigation of bystander intervention involved staged emergencies in public places. The advantage of field experiments is high ecological validity; the trade-off is reduced control over extraneous variables and ethical concerns about deception without informed consent.
在现场实验中，研究者在自然环境中操纵自变量，参与者不知道自己在参与研究。例如，拉塔内和达利（1970年）对旁观者干预的研究涉及在公共场所设计的紧急事件。现场实验的优势是高生态效度；代价是对额外变量的控制力减弱，以及在没有知情同意的情况下使用欺骗的伦理问题。
Quasi-experiments resemble true experiments but lack random assignment to conditions. For example, researchers might compare the altruistic behaviour of older and younger participants, drawing participants from naturally occurring age groups. Quasi-experiments provide useful insights but are vulnerable to selection effects because pre-existing groups may differ in multiple ways beyond the variable of interest.
准实验类似于真正实验，但缺乏随机分配到条件的过程。例如，研究者可能比较老年和年轻参与者的利他行为，从自然存在的年龄组中招募参与者。准实验提供了有用的见解，但容易受到选择效应的影响，因为既有群体可能在目标变量之外的其他多方面存在差异。
10. Ethical Considerations Across Methods | 不同方法中的伦理考量
Social psychology often studies sensitive topics—conformity, obedience, aggression, prejudice—that can involve deception, psychological stress, and invasion of privacy. Both experimental and correlational methods raise distinctive ethical questions.
社会心理学常常研究敏感议题——从众、服从、攻击性、偏见——这些议题可能涉及欺骗、心理压力和对隐私的侵犯。实验方法和相关方法都会引发独特的伦理问题。
In experimental research, the use of deception is common and historically controversial. Milgram’s obedience studies involved extensive deception about the true purpose of the research. Modern ethical guidelines from professional bodies such as the American Psychological Association (APA) and the British Psychological Society (BPS) require that deception be justified by significant scientific value, that less deceptive alternatives be considered first, and that participants be fully debriefed after the study.
在实验研究中，欺骗的使用很普遍，且在历史上一直备受争议。米尔格拉姆的服从研究就大量使用欺骗手段来掩盖研究的真实目的。来自美国心理学会（APA）和英国心理学会（BPS）等专业机构的现代伦理准则要求，欺骗必须由重大的科学价值来证明其合理性，必须首先考虑欺骗较少替代方案，并且研究结束后必须对参与者进行充分的汇报。
In correlational research, ethical concerns centre on informed consent, confidentiality, and the potential misuse of data. Survey studies may collect highly personal information, and researchers have an obligation to store data securely, protect participant anonymity, and ensure that participants can withdraw at any time without penalty. Additionally, the interpretation of correlational findings must be communicated responsibly to the public to avoid implying causal relationships that the data do not support.
在相关研究中，伦理问题围绕知情同意、保密性以及数据被滥用的潜在可能。调查研究可能收集高度个人信息，研究者有义务安全存储数据、保护参与者匿名性，并确保参与者可以在任何时候无惩罚地退出。此外，相关研究结果的解释必须以负责任的方式传达给公众，以避免暗示数据并不支持的因果关系。
11. Choosing the Appropriate Method for the Research Question | 根据研究问题选择合适的方法
The choice between experimental and correlational approaches must be driven primarily by the nature of the research question. Questions about causation—’Does discrimination cause decreased self-esteem?’—demand experimental methods. Questions about naturally occurring relationships—’Is self-esteem associated with academic achievement?’—lend themselves to correlational approaches.
在实验方法和相关方法之间的选择必须主要由研究问题的性质决定。关于因果关系的问题——“歧视是否导致自尊下降？”——需要实验方法。关于自然发生关系的问题——“自尊是否与学业成就相关？”——则适合相关方法。
Ethical and practical constraints may force a correlational approach even when a causal question is of interest. For example, it would be deeply unethical to experimentally manipulate childhood trauma or to randomly assign participants to experience prejudice. In such cases, well-designed longitudinal studies and cross-sectional correlational research represent the best available evidence.
即使在研究因果问题时，伦理和实际约束也可能迫使研究者采用相关方法。例如，实验性地操纵童年创伤或随机分配参与者去经历偏见是极不道德的。在这种情况下，设计良好的纵向研究和横断面相关研究代表了我们可获得的最佳证据。
Researchers should also be aware that no single method is universally superior. The overarching goal in social psychology is triangulation—converging evidence from multiple methodological approaches. When experimental, correlational, field, and longitudinal studies all point to the same conclusion, confidence in the validity of that conclusion increases substantially.
研究者还应认识到，没有任何单一方法是普遍优越的。社会心理学的总体目标是三角验证——来自多种方法学途径的汇聚证据。当实验、相关、现场和纵向研究都指向相同的结论时，我们对该结论效度的信心就会大幅提升。
12. Conclusion: The Complementary Roles of Experimental and Correlational Methods | 结论：实验方法与相关方法的互补作用
Social psychology’s rich body of knowledge about human social behaviour has been built upon a methodological foundation that includes both experimental and correlational methods. Experiments provide unmatched power to establish cause-and-effect relationships while maintaining strict control over extraneous variables. Correlational methods, though unable to prove causation, allow researchers to investigate meaningful relationships among variables in natural settings, including topics that are ethically or practically impossible to manipulate experimentally.
社会心理学关于人类社会行为的丰富知识体系建立在包含实验方法和相关方法的方法论基础之上。实验在控制额外变量方面提供了无与伦比的因果关系确立能力。相关方法虽然无法证明因果关系，但允许研究者在自然环境中考察变量之间有意义的关联，包括那些在伦理上或实践上无法通过实验操纵的议题。
As an A-Level psychology student, understanding the strengths and limitations of each approach enables you to think critically about the research evidence presented in textbooks and scientific journals. When evaluating a study, always consider the following questions: Was an appropriate method used for the research question? Were threats to validity adequately controlled? Have the authors acknowledged the limitations of their chosen design?
作为A-Level心理学的学生，理解每种方法的优势和局限使你能够批判性地思考教科书和科学期刊上呈现的研究证据。在评估研究时，始终考虑以下问题：是否针对研究问题使用了合适的方法？对效度的威胁是否得到充分控制？作者是否承认了其选择设计上的局限？
Ultimately, social psychologists do not choose between experiments and correlational studies as if they were mutually exclusive alternatives. Instead, skilled researchers strategically use both approaches—and often combine them in a single programme of research—to build a comprehensive understanding of complex social phenomena that is both causally informative and ecologically meaningful.
最终，社会心理学家不会像对待互斥选项一样在实验和相关研究之间二选一。相反，技能娴熟的研究者会策略性地使用这两种方法——并常常在单一研究计划中将它们结合——以建立对复杂社会现象的全面理解，这种理解既包含因果信息又具有生态学意义。
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Cancel reply