Causality
Understanding cause and effect relationships in psychology
Definition
Causation is the relationship between cause and effect, where changes in one variable (cause) directly result in changes in another (effect). Causality in psychology is about understanding how and why one variable influences another, while asking how confident we can be that the relationship is truly causal. This depends on whether researchers have isolated variables, controlled extraneous factors, and ruled out alternative explanations in the face of complex, interacting, and often hidden variables.
"A fundamental challenge for psychologists is understanding the relationship between cause and effect. Psychologists must determine whether one variable directly causes changes in another, or whether the relationship is merely correlational. Establishing causality requires careful experimental design, including the manipulation of variables, control groups, and random assignment."
Source: IBO (2023). Psychology guide. International Baccalaureate Organization, p. 21. ibo.org
Typical Exam Question Types
"Discuss how well psychologists can establish causality in research."
"Discuss the challenges in establishing causal relationships in psychology."
Conditions for Establishing Causation
Before claiming that A causes B, three logical conditions must all be satisfied. Missing even one means you can only claim a correlation, not causation. Examiners expect you to use these conditions when evaluating whether a study can support a causal claim.
| Condition | Description |
|---|---|
| Covariation | The cause and effect must vary together. When the cause is present, the effect is more likely to occur; when absent, the effect is less likely. |
| Temporal Precedence | The cause must occur before the effect. Establishing the correct time sequence is essential to rule out reverse causality. |
| No Plausible Alternative Explanations | Other variables must be ruled out as explanations for the observed relationship. This requires controlling for confounding variables and considering third-variable problems. |
Establishing Causality Through Experimental Manipulation
The true experiment is the only research design that can directly establish causality. Each of the following design features works to eliminate alternative explanations — the more of these a study uses, the stronger its causal claim. When evaluating a study, ask which of these features are present and which are missing.
| Method | Description |
|---|---|
| Well-Constructed Operationalisation | Both independent and dependent variables need to be clearly measurable and capture the essence of the construct. Variables need to be clearly isolated. |
| Control Groups | A group of participants that does not receive the experimental treatment, allowing researchers to compare outcomes with the experimental group. |
| Random Assignment | Process of allocating participants to different groups entirely by chance. Ensures each participant has equal likelihood of being in any group, reducing bias and controlling for individual differences. |
| Standardised Procedures | Consistent methods and instructions used across all participants. Ensures everyone experiences the experiment the same way, increasing reliability and reducing extraneous variables. |
| Double-Blind Design | Neither participants nor researchers directly interacting with them know which condition participants are in. Reduces both participant bias (social desirability) and researcher bias (expectancy effects). |
Experimental Controls
Control conditions allow researchers to isolate the effect of the independent variable by providing a baseline for comparison. Without a proper control condition, any observed change in the dependent variable could be due to factors other than the treatment — making causal inference impossible.
| Control Type | Description |
|---|---|
| Placebo | A treatment with no therapeutic effect given to participants to assess effects of active treatment by comparing it to non-active placebo. |
| Wait-Listing | Participants placed on waiting list to receive treatment later. Allows comparison between those receiving intervention and those who have not yet received it. |
Validity in Causal Research
Internal validity asks: did the IV actually cause the change in the DV? External validity asks: do the findings hold beyond the lab? Both are essential — a study can be internally valid but completely artificial, or ecologically realistic but confounded. Strong causal claims require both.
| Type | Description |
|---|---|
| Internal Validity | The extent that a measurement conducted within a study accurately measures or assesses what it claims to measure. |
| External Validity | The extent to which findings from a study can be generalized to a different setting or population. |
| Mundane Realism | The degree to which a research study, experimental materials, procedures and setting resemble real-life situations and experiences. Impacts external validity. |
| Replication | Findings should be consistent across different studies, samples, and contexts to establish robust causality. |
Complexity in Establishing Causation
Human behaviour is rarely caused by a single factor. Mediators explain the mechanism (how A causes B), while moderators explain the conditions (when or for whom A causes B). Third variables, bidirectional ambiguity, and ethical constraints all make causal inference in psychology far more challenging than in controlled laboratory sciences.
| Factor | Description |
|---|---|
| Complexity | Human behavior is influenced by many variables, often with Mediators (factors that explain how or why one variable affects another — the mechanism in the relationship) and Moderators (factors that influence the strength or direction of that relationship — the condition that changes the relationship, explaining context dependency). |
| Influence (Main effects, Mediation) | Describes the independent causal impact of a single factor on an outcome. Mediation unpacks the mechanism behind the influence. |
| Interaction (Moderation) | Occurs when the causal influence of one factor depends on another factor. Moderation explains the conditions under which an effect holds. |
| Bidirectional ambiguity and feedback loops | When the direction of cause and effect between two variables is unclear. Challenging to determine which variable influences the other or if there is a mutual interaction where cause and effect reinforce one another. |
| Third variable problem | A statistically observed relationship between two variables (A and B) that are not directly or causally linked, but appear connected due to a third, unobserved variable (C) or simple coincidence. |
| Reductionism and alternative explanations | Reductionism seeks to find the most basic possible explanation, valuable for establishing cause-and-effect but risks overlooking broader social and cultural context. Holism emphasises that behaviour is more than the sum of its parts. |
| Non-experimental research | Only TRUE experiments can conclude a causal relationship. Other methods cannot rule out alternative explanations. |
| Ethical limitations | Ethics ensure research protects participants, but also mean psychologists must rely on creative, indirect methods like natural experiments, longitudinal studies, or statistical modelling instead of experimental testing. |
| Agency | An individual's capacity to act independently and make their own free choices. Determinism argues events are predetermined. Most psychologists adopt a compatibilist perspective — agency and causality are not mutually exclusive. |
| Motivation | The impetus that gives purpose or direction to behaviour. Without motivation, even strong external cues may have no causal effect. |
| Significance | Statistical significance does not prove causality on its own (correlation ≠ causation), but it is a necessary first step. |
| Effect Size | Causality in psychology matters most when it has real-world impact. A causal factor that only produces trivial effects is of little practical value. |
Agency and Motivation in Causality
A unique challenge in psychology is that humans are not passive — they have agency (the capacity to choose) and motivation (internal drives that direct behaviour). This means the same external cause can produce different effects in different people, complicating causal claims. Most psychologists adopt a compatibilist view: agency and causality coexist.
| Concept | Description |
|---|---|
| Agency | An individual's capacity to act independently and make their own free choices. Often seen as opposite of being caused by external forces. Most psychologists adopt a compatibilist perspective, arguing that agency and causality are not mutually exclusive. |
| Determinism | The argument that events, including human actions and choices, are predetermined and inevitable. Explaining causal mechanisms behind decision-making explains how agency works. |
| Motivation | The impetus that gives purpose or direction to behaviour, operating at conscious or unconscious levels. Explains the internal mechanism through which external factors translate into behavior. |
Why is Causality Important? — CAUSALITY Mnemonic
Use this framework to evaluate causal claims in any study.
| Mnemonic | Validity Check | Why This Matters |
|---|---|---|
| C – Controls | Were extraneous variables controlled? Are there any control groups? | Having controls increase the internal validity and reduce possible effects of bias (history, maturation, demand characteristics). |
| A – Assignment | Was random assignment used to balance participant differences? Were there counterbalancing? Are the participant and/or researcher blinded? | Prevents selection bias. |
| U – Understanding variables | Were IV and DV operationalized clearly? | Affects measurement validity which affects the internal validity and credibility of results. |
| S – Statistical significance | Are the tests appropriate? Are there any type 1 or type 2 error? Is the sample big enough for normative statistical testing? | How likely are results due to chance? Significance supports causal inference but must be interpreted cautiously. |
| A – Alternative explanations | Are variables just correlated? Were rival hypotheses considered (third variables, bidirectional ambiguity)? | Looking at possible alternative explanations demonstrates critical thinking. If no other alternative explanations are possible, confidence of causal claim is boosted. |
| L – Limitations | What weaknesses (sample, ecological validity) restrict causal confidence? | Acknowledging limits shows critical thinking. Limitations may affect the confidence of causal inferences. |
| I – Internal validity | Was the design rigorous enough that DV changes can only be attributed to IV? | Internal validity is the core of causality. Without internal validity, causal claims collapse. |
| T – Triangulation / type of study | Was causality supported by multiple methods (e.g., RCT + longitudinal)? | Triangulation strengthens trust. Increases robustness; causal claims are not method-dependent. |
| Y – Why it matters | Does the causal link have practical or theoretical importance? | Causality is meaningful only if it advances knowledge or improves lives. Causal claims must connect to real-world or theory. |
Step-by-Step Answer Strategy
- 1. Restate the claim
- 2. State challenges and limitations in establishing causality
- 3. Use examples of research methods (experiments, correlational studies, longitudinal designs)
- 4. Analyse strengths/limitations (Experiments establish causality but may lack external validity; correlational studies lack internal validity)
- 5. Bring in own knowledge (Confounding variables, third-variable problem, bidirectional causality, moderators and mediators)
- 6. Balance the argument (Multiple methods combined provide stronger evidence; causality is complex and context-dependent)
- 7. Conclude (Psychologists can establish causality through rigorous experimental design, but must acknowledge limitations and complexity)