Natural and quasi-experiments | AQA A-Level Psychology Revision
- Revision Notes
- Aug 3
- 25 min read
Updated: Aug 23
For 7182 specification, first teach in September 2025
AQA A-Level Psychology | Free Revision Notes
Estimated study time: 45 minutes
These Natural and quasi-experiments A-Level Psychology revision notes explain how psychologists investigate variables they cannot or should not manipulate directly. You will distinguish naturally occurring changes from pre-existing participant differences and compare both methods with researcher-manipulated laboratory and field studies. You will also consider control, causation, ecological validity, ethics and replicability when evaluating each type of experiment.
Learning Objectives 🎯
By the end of this revision page, you should be able to:
Define a natural experiment.
Explain how the independent variable arises in a natural experiment.
Define a quasi-experiment.
Explain how pre-existing participant differences form the independent variable in a quasi-experiment.
Distinguish natural and quasi-experiments.
Compare natural and quasi-experiments with laboratory and field experiments.
Evaluate the strengths and limitations of natural and quasi-experiments.
Identify the experimental type used in an unfamiliar investigation.
Revision Notes 📚
Natural and quasi-experiments A-Level Psychology revision overview
Experiments investigate whether differences in an independent variable are associated with differences in a dependent variable.
In laboratory and field experiments, the researcher manipulates the independent variable.
In natural and quasi-experiments, the researcher does not directly manipulate it.
The central distinction is:
Type of experiment | How the independent variable is created |
Laboratory experiment | Manipulated by the researcher in a controlled setting |
Field experiment | Manipulated by the researcher in a natural setting |
Natural experiment | Produced by an event or circumstance that would occur without the researcher |
Quasi-experiment | Based on a pre-existing difference between participants |
The AQA specification requires knowledge of all four types and awareness of their strengths and limitations.
The experimental method
The experimental method investigates whether one variable may affect another.
An experiment normally includes:
An independent variable.
At least two conditions or groups.
A dependent variable.
A comparison of outcomes.
Consideration of extraneous variables.
For example, a psychologist may investigate whether school start time affects students’ alertness.
The researcher needs:
Different start-time conditions.
A measure of alertness.
A comparison between the conditions.
What distinguishes the four experimental types is how the independent variable arises and, for laboratory and field experiments, the setting in which it is manipulated.
The independent variable
The independent variable, abbreviated to IV, is the variable whose different levels or conditions are compared.
In a true laboratory or field experiment, the researcher actively manipulates the IV.
For example:
Quiet room compared with a noisy room.
Praise compared with no praise.
Phone visible compared with phone absent.
In natural and quasi-experiments, the researcher identifies an IV that already exists or has arisen independently.
The researcher still measures its possible effect on a DV, but does not create the difference directly.
The dependent variable
The dependent variable, abbreviated to DV, is the outcome measured by the researcher.
Examples include:
Number of words recalled.
Score on an attention test.
Time taken to complete a task.
Number of helping behaviours.
Rating of stress.
Examination performance.
The DV should be measured consistently in every condition or group.
The precise definition of variables is developed in identifying and measuring variables.
Natural experiments
What is a natural experiment?
A natural experiment is an experiment in which the change or difference in the independent variable occurs naturally rather than being manipulated by the researcher.
The event or circumstance would have happened even if the psychologist had not conducted the investigation.
The researcher:
Identifies a naturally occurring change or situation.
Identifies the conditions created by that change.
Measures a dependent variable.
Compares the outcomes.
The basic sequence is:
Naturally occurring event or difference → naturally created conditions → researcher measures the DV
What does “natural” mean?
In a natural experiment, the word natural describes how the independent variable arose.
It does not necessarily describe the physical setting.
A natural experiment may involve data collected:
In a school.
In a workplace.
In a hospital.
Through an online assessment.
In a controlled testing room.
From existing records.
The defining feature is that the researcher did not create or manipulate the IV.
This is different from a field experiment, where the setting is natural but the researcher does manipulate the IV.
Example of a natural experiment
Suppose a school changes its starting time from 8.30 am to 9.15 am for administrative reasons.
A psychologist investigates whether the change affects students’ alertness.
Naturally occurring IV: school start time.
Condition 1: alertness before the change.
Condition 2: alertness after the change.
DV: score on a standardised alertness task.
This is a natural experiment because the school, rather than the psychologist, changed the start time.
The psychologist uses the change as an opportunity to investigate its possible effect.
Another natural experiment
Consider the following investigation:
A local authority introduces additional street lighting in one neighbourhood but not another. A psychologist measures residents’ feelings of safety before and after the change.
The IV is the introduction of additional lighting.
The psychologist did not manipulate this variable. It arose from a local authority decision.
The DV might be:
A rating of perceived safety.
The number of residents using the area after dark.
The frequency of avoidance behaviour.
Natural events
Natural experiments may make use of events such as:
A change in school policy.
A change in workplace arrangements.
A new law or public policy.
A transport disruption.
A naturally occurring environmental event.
A difference in access to a service.
A major social event.
The psychologist does not cause the event.
Instead, the psychologist investigates its possible psychological or behavioural effects.
Conditions in a natural experiment
The naturally occurring IV must create at least two conditions or levels that can be compared.
These could be:
Before and after an event
The same people are assessed before and after a change.
For example:
Stress before an office relocation.
Stress after the relocation.
Affected and unaffected groups
One group experiences the event and another does not.
For example:
A school that changes its timetable.
A similar school that keeps its original timetable.
Different levels of naturally occurring exposure
Participants experience different levels of an event or circumstance.
For example:
Residents living close to a new source of environmental noise.
Residents living further away.
Natural experiments and experimental design
A natural experiment may use different experimental designs.
Repeated measures
The same participants are assessed before and after the naturally occurring change.
Independent groups
Different participants experience the different naturally occurring conditions.
Matched pairs
Participants in the conditions are matched on relevant characteristics where possible.
The type of experiment and the experimental design are separate decisions.
The designs are explored in Experimental designs.
The researcher’s role
Although the researcher does not manipulate the IV, they still make important decisions.
They decide:
What event to investigate.
How to define the conditions.
Which DV to measure.
When measurements will be taken.
Which participants will be included.
How potential confounding variables will be addressed.
How the results will be analysed.
A natural experiment is therefore not an entirely uncontrolled observation.
The researcher designs the investigation around an event they did not create.
Natural experiment or correlation?
Natural experiments and correlations can appear similar because neither necessarily involves direct manipulation by the researcher.
The distinction concerns how the variables are organised.
Natural experiment
A naturally occurring IV creates separate conditions.
The researcher compares the DV between those conditions.
The research question is framed in terms of a possible effect.
Correlation
Two co-variables are measured.
No variable is treated as a manipulated or naturally created condition.
The researcher examines the strength and direction of a relationship.
For example:
Comparing alertness before and after a naturally imposed timetable change may be a natural experiment.
Measuring the relationship between usual bedtime and alertness scores may be a correlation.
The distinction is developed in relationships between co-variables.
Strength of natural experiments: variables that cannot be manipulated
Natural experiments allow psychologists to investigate variables that would be impossible or unethical to manipulate deliberately.
A researcher could not ethically:
Cause a natural disaster.
Force a person to lose employment.
Remove access to an essential service.
Create serious environmental disruption.
Expose participants to dangerous conditions.
If such an event occurs independently, researchers may be able to investigate its psychological effects without causing it.
This extends the range of questions psychology can study.
Strength: real-world relevance
Naturally occurring events take place within genuine social and environmental contexts.
Participants may experience:
Real changes.
Meaningful consequences.
Ordinary pressures.
Authentic decisions.
The findings may therefore have practical value.
For example, research into a school timetable change could inform future educational decisions.
Strength: ecological validity
Natural experiments may have high ecological validity because participants experience real situations rather than artificial laboratory manipulations.
Their responses may reflect how they behave in everyday life.
For example:
Stress caused by a genuine workplace change may be more meaningful than stress produced by a short artificial task.
Behaviour following an actual policy change may be more realistic than responses to a hypothetical description.
However, ecological validity depends on the particular study.
A naturally occurring IV could still be measured using an artificial laboratory task.
Strength: reduced demand characteristics
Participants may be less likely to identify the research hypothesis because the researcher did not introduce the event.
They may see the change as part of ordinary life rather than as an experimental manipulation.
This could reduce demand characteristics, particularly when participants do not know which aspect of the event is being studied.
However, demand characteristics are still possible if:
Participants know they are being assessed.
They guess the purpose of the study.
The DV makes the aim obvious.
Repeated testing draws attention to the event.
Strength: opportunities for practical application
Natural experiments can evaluate the effects of real policies or environmental changes.
Their findings may help organisations decide whether to:
Keep a new timetable.
Change working arrangements.
Alter public information.
Expand a service.
Introduce an intervention elsewhere.
Because the event occurred in a real context, the findings may be directly relevant to similar settings.
Limitation of natural experiments: lack of control
A major limitation is that naturally occurring events are not controlled by the researcher.
Many factors may change at the same time.
For example, if a school introduces a later start time:
Lesson length may also change.
Transport arrangements may change.
Teachers may change their routines.
Students may sleep for different amounts of time.
The assessment may occur at a different point in the academic year.
Any of these factors could affect alertness.
Confounding variables
A confounding variable is an uncontrolled variable that changes systematically with the IV and could explain the findings.
Suppose students perform better after a timetable change.
This might be caused by:
The later start time.
Greater familiarity with the test.
Different teaching.
A less stressful point in the term.
Changes in the student sample.
Another policy introduced at the same time.
Because the researcher did not control the naturally occurring change, it may be difficult to isolate the effect of the IV.
Limitation: weaker causal conclusions
Natural experiments may investigate possible cause and effect, but causal conclusions are often weaker than in a well-controlled laboratory experiment.
The researcher cannot be certain that:
The IV was the only important difference.
Participants were equivalent across conditions.
The event occurred in the same way for everyone.
Another variable did not produce the outcome.
A careful conclusion might state:
The naturally occurring timetable change was associated with higher alertness scores.
A stronger causal statement would require evidence that alternative explanations had been controlled.
Limitation: no random allocation
Participants are usually not randomly allocated to the naturally occurring conditions.
The event determines who belongs in each condition.
For example:
Residents already live in the affected or unaffected neighbourhood.
Pupils already attend the school that changes its timetable.
Employees already work in a particular office.
The conditions may therefore contain participants who differ in important ways.
These participant differences may influence the DV.
Limitation: difficulty replicating the study
Naturally occurring events may be:
Unique.
Unpredictable.
Difficult to recreate.
Different each time they occur.
Accompanied by different environmental conditions.
Another researcher may be unable to reproduce the same event with the same participants and context.
This reduces exact replicability.
Replication and consistency are considered further in Reliability.
Limitation: timing and baseline measurements
A naturally occurring event may happen unexpectedly.
Researchers might not have collected measurements before it occurred.
Without a baseline, they may not know:
What participants were like before the event.
Whether a difference already existed.
How much change occurred.
Whether the groups were initially comparable.
Retrospective reports of earlier behaviour may be inaccurate.
A planned policy change is easier to investigate because researchers may be able to collect data before and after it.
Limitation: ethical issues
Although the researcher does not cause the event, ethical issues may still arise.
These include:
Informed consent.
Privacy.
Confidentiality.
Psychological distress.
Sensitive personal information.
The right to withdraw.
Repeated discussion of a difficult event.
Researchers must not assume that a naturally occurring situation can be studied without ethical safeguards.
These issues are developed in Ethics in psychological research.
Quasi-experiments
What is a quasi-experiment?
A quasi-experiment is an experiment in which the independent variable is a pre-existing difference between participants.
The researcher does not manipulate the IV and cannot randomly allocate participants to its conditions.
Examples of pre-existing variables include:
Age group.
Handedness.
Whether a person is multilingual.
Whether a person has a particular diagnosis.
Whether a person has experienced a specific life event.
Level of an existing ability.
Membership of an existing group.
The researcher selects participants who already belong to the relevant conditions and measures a DV.
What does “quasi” mean?
The word quasi means resembling or appearing similar to something without possessing every defining feature.
A quasi-experiment resembles a true experiment because:
It includes an IV.
It contains different conditions or groups.
A DV is measured.
The groups are compared.
However, it lacks direct manipulation of the IV and random allocation to its levels.
Example of a quasi-experiment
Suppose a psychologist compares younger and older adults on a memory test.
IV: age group.
Condition 1: adults aged 18 to 30.
Condition 2: adults aged 65 to 75.
DV: number of words recalled.
This is a quasi-experiment because age is a pre-existing participant characteristic.
The researcher cannot manipulate a participant’s age or randomly allocate someone to be a younger or older adult.
Another quasi-experiment
Consider this investigation:
A psychologist compares left-handed and right-handed participants on a spatial task.
IV: handedness.
Level 1: left-handed participants.
Level 2: right-handed participants.
DV: spatial-task score.
Handedness existed before the investigation and was not created by the researcher.
The study is therefore a quasi-experiment.
Pre-existing participant variables
A quasi-experimental IV is sometimes called a participant variable because it describes a characteristic of the participants.
The groups exist before the investigation begins.
The researcher:
Identifies the characteristic.
Recruits people from the relevant groups.
gives them the same task or assessment.
Measures the DV.
Compares the groups.
Unlike a laboratory experiment, participants cannot be allocated randomly to the levels of the IV.
Quasi-experiments and setting
A quasi-experiment may take place in:
A controlled laboratory.
A classroom.
A workplace.
A hospital.
A community setting.
An online environment.
The location does not determine whether it is quasi-experimental.
For example:
Younger and older adults attend the same controlled room and complete an identical memory task.
The setting is laboratory-like, but the IV is pre-existing.
The investigation is therefore a quasi-experiment rather than a laboratory experiment in which the IV was manipulated.
Standardisation in a quasi-experiment
Although the IV cannot be controlled, the procedure used to measure the DV may be highly standardised.
Researchers can keep constant:
Instructions.
Materials.
Time limits.
Testing room.
Equipment.
Order of stimuli.
Scoring procedures.
This helps ensure that differences in the DV are not caused by inconsistent testing.
However, standardisation cannot remove all pre-existing differences between the groups.
Participant variables in quasi-experiments
A central difficulty is that the IV may be connected with many other characteristics.
For example, younger and older adults may differ in:
Education.
Health.
Familiarity with technology.
Sleep patterns.
Motivation.
Life experience.
Medication use.
If the groups differ in memory scores, age itself may not be the only explanation.
Other participant variables may be confounded with the IV.
Quasi-experiment or natural experiment?
Both methods involve an IV that is not manipulated by the researcher.
The difference concerns the source of the IV.
Natural experiment | Quasi-experiment |
IV results from an event or circumstance | IV is an existing participant difference |
Conditions arise because something happens | Groups exist before the investigation |
Example: school introduces a later start time | Example: comparing younger and older pupils |
Researcher studies the effect of a naturally occurring change | Researcher studies differences between existing groups |
A useful memory rule is:
Natural experiment: something naturally happens.
Quasi-experiment: participants already differ.
An important borderline issue
Some investigations may contain features of both natural and quasi-experiments.
For example, groups might differ because one experienced a naturally occurring event and the other did not.
To classify the study, focus on how the IV is described:
If the IV is the occurrence of an external event, it is best described as a natural experiment.
If the IV is a stable or pre-existing participant characteristic, it is best described as a quasi-experiment.
In an examination, use the details provided in the scenario and explain your classification.
Quasi-experiment or correlation?
A quasi-experiment compares separate pre-existing groups on a DV.
A correlation measures two co-variables for each participant and examines their relationship.
For example:
Quasi-experiment
Compare participants with high and low existing levels of musical training on a memory test.
Correlation
Measure each participant’s years of musical training and memory score, then calculate the relationship between them.
The quasi-experiment treats musical-training group as an IV with conditions.
The correlation treats both measures as co-variables.
Strength of quasi-experiments: studying variables that cannot be manipulated
Quasi-experiments allow psychologists to investigate meaningful participant differences that cannot be manipulated.
A researcher cannot randomly assign participants to:
A particular age.
A developmental stage.
A naturally existing diagnosis.
A first language.
A history of a previous experience.
A stable biological characteristic.
Quasi-experiments therefore make important psychological questions accessible.
Strength: avoiding unethical manipulation
Some participant variables could be studied only unethically if manipulation were required.
For example, a researcher should not create:
A serious psychological disorder.
A damaging life experience.
A neurological injury.
Long-term deprivation.
A quasi-experiment allows comparisons involving existing groups without creating the characteristic.
The researcher must still protect participants and handle sensitive information carefully.
Strength: controlled measurement is possible
A quasi-experiment can be conducted using a standardised procedure.
For example, both age groups could:
Complete the same memory test.
Receive the same instructions.
Use the same equipment.
Work in the same room.
Receive the same time allowance.
This high level of procedural control may improve reliability.
It also reduces some alternative explanations relating to the immediate testing environment.
Strength: objective comparisons
Researchers may collect objective quantitative data, such as:
Test scores.
Response times.
Accuracy.
Number of errors.
Behavioural frequencies.
This makes differences between the groups easier to compare and analyse.
However, objective measurement does not remove the possibility that the groups differed in uncontrolled ways before the investigation.
Strength: practical and theoretical value
Quasi-experiments can help psychologists investigate:
Developmental differences.
Differences associated with brain injury.
Differences linked with existing experiences.
Variation between educational or social groups.
Characteristics associated with a diagnosis.
Findings may contribute to:
Psychological explanations.
Assessment procedures.
Educational support.
Future experimental research.
New hypotheses.
Limitation of quasi-experiments: no manipulation of the IV
Because the IV is not manipulated, the researcher cannot control how the groups came to differ.
For example, participants with extensive musical training may also differ from those without training in:
Family background.
Education.
Motivation.
Income.
Practice habits.
Other cognitive activities.
A difference in the DV may be associated with musical training without being caused by it.
Limitation: no random allocation
Participants cannot be randomly allocated to pre-existing groups.
A participant cannot be randomly assigned to:
A different age.
A different handedness.
A past life experience.
An existing diagnosis.
Random allocation normally helps distribute participant variables more evenly between conditions.
Without it, systematic group differences are more likely.
Limitation: participant variables
Participant variables are individual characteristics that may affect behaviour.
In a quasi-experiment, the groups may differ on many participant variables in addition to the IV.
Suppose older adults perform less well than younger adults on a computerised memory task.
Possible explanations include:
Age-related memory differences.
Less experience using computers.
Slower motor responses.
Different levels of anxiety.
Differences in eyesight.
Differences in educational experience.
The researcher must consider whether the DV measures the intended process fairly across both groups.
Limitation: weaker causal conclusions
Quasi-experiments are limited in their ability to establish cause and effect.
If two pre-existing groups differ on a DV, the researcher can conclude that:
The group characteristic is associated with the difference.
The researcher usually cannot conclude confidently that:
The characteristic directly caused the difference.
The groups may differ in many ways, creating alternative explanations.
Limitation: matching is difficult
Researchers may attempt to match the groups on relevant variables.
For example, younger and older participants might be matched on:
Education.
General health.
Language ability.
Familiarity with testing.
However, perfect matching is difficult because:
Not every relevant variable is known.
Some characteristics are difficult to measure.
Matching many variables reduces the available sample.
Participants may still differ in unmeasured ways.
Matching can improve the comparison but cannot create the same control as random allocation.
Limitation: labelling and social sensitivity
Quasi-experiments sometimes compare groups defined by personal characteristics or diagnoses.
This may create risks of:
Stigmatisation.
Stereotyping.
Misinterpretation of average group differences.
Treating group membership as a complete explanation of behaviour.
Ignoring differences within each group.
Researchers should report findings carefully.
A mean difference between groups does not mean that every member of one group differs from every member of the other.
Limitation: generalisation
Quasi-experimental groups may not represent everyone with the relevant characteristic.
For example, participants recruited through a specialist organisation may differ from other people in the same broad group.
The sample may also be:
Small.
Self-selecting.
Limited to one setting.
Unequal in size.
Unrepresentative of the target population.
Sampling and generalisation are examined in Populations and samples.
Comparing all four types of experiment
Main comparison
Feature | Laboratory | Field | Natural | Quasi |
Who creates the IV? | Researcher | Researcher | Event or circumstance | Pre-existing participant difference |
Is the IV manipulated? | Yes | Yes | No | No |
Typical setting | Controlled | Natural | Any setting | Any setting |
Random allocation possible? | Often | Sometimes | Usually not | Not to the IV |
Control over extraneous variables | Usually high | Usually lower | Usually lower | Procedure may be controlled, but group differences remain |
Ecological validity | May be lower | May be higher | Often high where a real event is studied | Depends on task and setting |
Causal conclusions | Potentially strong | Possible but reduced by lower control | Usually cautious | Usually cautious |
Replication | Usually easier | Often more difficult | Often difficult | Procedure may be replicable |
Main strength | Control | Natural behaviour | Studies real events that cannot be manipulated | Studies pre-existing differences |
Main limitation | Artificiality | Uncontrolled environment | Confounding variables | Participant variables |
Laboratory versus natural experiment
Both may compare conditions and measure a DV.
The main difference is manipulation.
Laboratory experiment
The researcher creates the levels of the IV.
The setting is highly controlled.
Random allocation may be used.
Causal conclusions may be stronger.
Natural experiment
The IV arises independently of the researcher.
Random allocation is usually impossible.
The event may have high real-world relevance.
Confounding variables are harder to control.
A concise comparison could state:
In a laboratory experiment, the researcher manipulates the IV and usually controls the testing environment. In a natural experiment, the IV results from an event that would have occurred without the researcher, so control and random allocation are usually reduced.
Field versus natural experiment
Field and natural experiments are frequently confused.
Both may take place in an everyday setting, but the source of the IV differs.
Field experiment
The researcher manipulates the IV in a natural environment.
Example:
A researcher changes the type of music played in a real café.
Natural experiment
A change occurs independently, and the researcher investigates it.
Example:
A café changes its opening hours for business reasons, and a psychologist investigates employee stress before and after the change.
The simplest distinction is:
Field describes the setting. Natural describes the source of the IV.
Laboratory versus quasi-experiment
A quasi-experiment may be conducted in a controlled setting, but that does not make it a laboratory experiment.
Laboratory experiment
Researcher manipulates the IV.
Participants may be randomly allocated to conditions.
Quasi-experiment
IV is a pre-existing participant characteristic.
Participants cannot be randomly allocated to the IV.
The measurement procedure may still be highly controlled.
For example:
Younger and older adults complete the same memory task in a laboratory.
This is a quasi-experiment because age group was not manipulated.
Field versus quasi-experiment
A quasi-experiment may take place in a natural setting.
For example:
A psychologist compares existing groups of experienced and inexperienced teachers during their normal lessons.
The study is quasi-experimental because teaching experience is a pre-existing participant difference.
Its classroom location does not make it a field experiment because the researcher did not manipulate the IV.
Natural versus quasi-experiment
The methods share several features.
Both:
Lack direct researcher manipulation of the IV.
Usually lack random allocation to the IV.
May study variables that cannot be manipulated.
Are vulnerable to confounding variables.
Require cautious causal conclusions.
May have practical value.
The difference is:
Natural experiments use naturally occurring external events or circumstances.
Quasi-experiments use pre-existing participant characteristics.
Control and causation
The four methods can be considered along a continuum of control.
Laboratory experiments
Usually provide the greatest control and strongest basis for causal conclusions.
Field experiments
Contain manipulation but less environmental control.
Natural experiments
Use naturally occurring conditions and usually have limited control over the IV.
Quasi-experiments
May control the immediate procedure, but cannot control the pre-existing differences that formed the groups.
This does not mean that every laboratory experiment is better than every natural or quasi-experiment.
A poorly designed laboratory study may have lower validity than a carefully designed natural study.
Ecological validity
Ecological validity depends on how well the investigation represents everyday behaviour.
Field experiments often study natural behaviour in ordinary settings.
Natural experiments often involve meaningful real-world events.
Laboratory experiments may use more artificial tasks.
Quasi-experiments may be either naturalistic or artificial depending on the procedure.
The experiment’s label alone does not determine its ecological validity.
The actual setting, task and participant experience must be considered.
Demand characteristics
Demand characteristics may vary between methods.
Type | Likely influence of demand characteristics |
Laboratory | Often higher because participants know they are being studied |
Field | May be lower if participants are unaware |
Natural | May be lower because the researcher did not introduce the event |
Quasi | Depends on whether participants recognise the group comparison and aim |
Even in natural and quasi-experiments, participants may change their behaviour if they understand the purpose of the measurement.
Replicability
Laboratory experiment
Usually easiest to replicate because the researcher controls and documents the procedure.
Field experiment
Harder to replicate because natural settings change.
Natural experiment
Often difficult to replicate because the event may be unique.
Quasi-experiment
The testing procedure may be replicable, although the exact participant groups may be difficult to recreate.
A method may therefore have:
A highly standardised measurement procedure.
Low replicability of the original IV or sample.
Ethical considerations
Each experimental type may create different ethical issues.
Laboratory experiment
Participants usually know that they are taking part, making consent and withdrawal easier to manage.
Field experiment
Participants may be unaware, creating issues involving consent, privacy and withdrawal.
Natural experiment
The event may involve distressing or sensitive circumstances, even though the researcher did not cause it.
Quasi-experiment
Group labels and personal characteristics may be sensitive, and findings could contribute to stereotypes.
Ethical evaluation should be applied to the specific procedure rather than attached automatically to the experimental label.
Choosing an appropriate method
A laboratory or field experiment may be suitable when the researcher can ethically manipulate the IV.
A natural experiment may be appropriate when:
A real-world change is already occurring.
Manipulation would be impossible.
Manipulation would be unethical.
The event has important practical consequences.
A quasi-experiment may be appropriate when:
The research question concerns an existing participant characteristic.
Participants cannot be allocated randomly to the IV.
A standardised comparison between groups is possible.
Classification decision tree
Use these questions to identify the experimental type.
Question 1: Did the researcher manipulate the IV?
Yes
Move to Question 2.
No
Move to Question 3.
Question 2: Where did the investigation take place?
Controlled or artificial setting: laboratory experiment
Natural everyday setting: field experiment
Question 3: What created the levels of the IV?
An event or external circumstance: natural experiment
A pre-existing participant difference: quasi-experiment
Worked classification examples
Example 1
A researcher asks participants to learn words in silence or while music is played in a controlled room.
Classification: Laboratory experiment.
Reason: The researcher manipulated the music condition in a controlled environment.
Example 2
A researcher changes the message displayed beside a lift and records stair use in a real office building.
Classification: Field experiment.
Reason: The researcher manipulated the message in a natural setting.
Example 3
A company independently introduces a four-day working week. A psychologist measures employee stress before and after the change.
Classification: Natural experiment.
Reason: The workplace change was not introduced by the psychologist.
Example 4
A psychologist compares people who speak one language with people who speak two languages on a problem-solving task.
Classification: Quasi-experiment.
Reason: Language background is a pre-existing participant difference.
Applying evaluation to a natural experiment
Consider this investigation:
One school changes to a later starting time. Another school keeps its usual timetable. Researchers compare students’ scores on an attention task.
Possible strengths include:
The timetable change is genuine.
The findings may have practical educational value.
Students experience real school conditions.
The researcher does not need to impose the timetable.
Possible limitations include:
The schools may differ in teaching quality.
Students were not randomly allocated.
The schools may have different facilities or populations.
Other timetable changes may have occurred.
One school may complete the assessment at a different time of year.
A conclusion that the later start time caused improved attention would therefore require caution.
Applying evaluation to a quasi-experiment
Consider another investigation:
A psychologist compares left-handed and right-handed participants on a computerised spatial task.
Possible strengths include:
Handedness can be studied without manipulation.
All participants can complete the same standardised task.
Instructions, equipment and timing can be controlled.
The results can be measured objectively.
Possible limitations include:
The groups may differ in age, experience or education.
Participants cannot be randomly allocated to handedness.
The task may not represent everyday spatial ability.
A difference would not establish that handedness caused performance.
The samples may not represent all left-handed or right-handed people.
A method for evaluating natural experiments
Use the following structure.
1. Identify the naturally occurring event
State what changed independently of the researcher.
2. Identify a benefit of studying it
Explain why direct manipulation would be impossible, unethical or unrealistic.
3. Identify an uncontrolled variable
Name a specific factor that may differ between conditions.
4. Explain the consequence
State how it weakens a causal conclusion.
5. Consider validity and application
Explain whether the real-world context increases ecological validity or practical value.
A method for evaluating quasi-experiments
Use the following structure.
1. Identify the pre-existing IV
State the participant characteristic forming the groups.
2. Explain why it cannot be manipulated
Link this to practicality or ethics.
3. Identify another group difference
Give a participant variable that may affect the DV.
4. Explain why random allocation is impossible
State how this limits causal inference.
5. Consider procedural control
Explain whether standardisation improves reliability.
Writing an effective comparison
A strong comparison might state:
Laboratory and field experiments involve direct manipulation of an independent variable, whereas natural and quasi-experiments use conditions that already exist or arise independently of the researcher. In a natural experiment, the IV results from an external event or circumstance. In a quasi-experiment, the IV is a pre-existing participant difference. Natural and quasi-experiments allow psychologists to study variables that may be impossible or unethical to manipulate, but the lack of random allocation and reduced control make causal conclusions less certain.
This answer:
Identifies the central similarity.
Gives the key distinction.
Compares manipulation.
Includes a strength.
Includes a limitation.
Links the methods with causal inference.
Writing an effective natural-experiment evaluation paragraph
One strength of a natural experiment is that it allows psychologists to investigate real events that could not be manipulated ethically. For example, researchers could examine the psychological effects of a genuine change in working arrangements without imposing it on employees. This may produce behaviour with high ecological validity. However, the researcher has limited control over other changes occurring at the same time, so any difference in stress might be caused by workload, staffing or seasonal factors rather than the new working arrangement.
Writing an effective quasi-experiment evaluation paragraph
One strength of a quasi-experiment is that it allows researchers to compare pre-existing groups, such as younger and older adults, without manipulating characteristics that cannot be changed. The testing procedure can still be standardised, improving reliability. However, the groups cannot be formed through random allocation and may differ in education, health or task experience. These participant variables make it difficult to conclude that age itself caused any difference in performance.
Key Words 🔑
Key word | Student-friendly definition | How it may be used in an exam |
Natural experiment | An experiment in which the IV results from an event or circumstance not created by the researcher. | Identify a study based on a naturally occurring change. |
Quasi-experiment | An experiment in which the IV is a pre-existing participant difference. | Identify a comparison between existing groups. |
Independent variable | The variable whose conditions or levels are compared. | Explain what differs between groups or situations. |
Dependent variable | The outcome measured by the researcher. | Identify the data collected in the investigation. |
Naturally occurring variable | A variable that changes without direct researcher manipulation. | Explain the defining feature of a natural experiment. |
Pre-existing difference | A participant characteristic that already exists before research begins. | Explain the defining feature of a quasi-experiment. |
Laboratory experiment | An experiment in which the researcher manipulates the IV in a controlled setting. | Compare researcher manipulation and control. |
Field experiment | An experiment in which the researcher manipulates the IV in a natural setting. | Distinguish a natural setting from a naturally occurring IV. |
Experimental condition | One level or version of the independent variable. | Identify the groups or situations being compared. |
Random allocation | Assigning participants to conditions using chance. | Explain why group differences are harder to control in natural and quasi-experiments. |
Participant variable | An individual characteristic that may affect performance or behaviour. | Evaluate pre-existing groups in a quasi-experiment. |
Extraneous variable | Any variable other than the IV that could affect the DV. | Identify an alternative influence on the results. |
Confounding variable | An uncontrolled variable that varies with the IV and could explain the findings. | Evaluate the internal validity of an investigation. |
Cause and effect | A relationship in which changing one factor produces a change in another. | Explain limits on conclusions from natural and quasi-experiments. |
Internal validity | The extent to which the IV rather than another variable caused the result. | Evaluate control and confounding variables. |
Ecological validity | The extent to which findings represent behaviour in everyday settings. | Evaluate real-world natural events or tasks. |
Demand characteristics | Cues that lead participants to guess the research aim and alter their behaviour. | Evaluate awareness of the study. |
Standardisation | Keeping the measurement procedure consistent across participants. | Explain how quasi-experimental reliability can be improved. |
Replication | Repeating an investigation using the same or similar procedures. | Evaluate the uniqueness of natural events. |
Correlation | A method examining the relationship between two co-variables. | Distinguish group comparisons from correlational analysis. |
Generalisation | Applying findings beyond the people and setting studied. | Evaluate small, unusual or unrepresentative groups. |
Common Mistakes ⚠️
Mistake: Saying that a natural experiment must take place outdoors or in a natural environment.
Why this is incorrect:The term describes how the IV arose, not the physical setting.
How to improve:State that the IV results from an event or circumstance not manipulated by the researcher.
Mistake: Calling every study conducted in a school or workplace a natural experiment.
Why this is incorrect:A researcher may manipulate an IV in a natural setting, making it a field experiment.
How to improve:Ask who created the difference between conditions.
Mistake: Confusing a field experiment with a natural experiment.
Why this is incorrect:A field experiment has a researcher-manipulated IV in a natural setting. A natural experiment has a naturally occurring IV.
How to improve:Remember: field refers to setting, natural refers to the origin of the IV.
Mistake: Saying that a quasi-experiment has no independent variable.
Why this is incorrect:A quasi-experiment has an IV, but it is a pre-existing participant difference.
How to improve:Identify the characteristic that divides participants into groups.
Mistake: Calling a comparison of age groups a natural experiment.
Why this is incorrect:Age is a pre-existing participant characteristic rather than an external event.
How to improve:Classify comparisons based on age as quasi-experiments.
Mistake: Saying the researcher manipulates the IV in a quasi-experiment.
Why this is incorrect:Participants already possess the characteristic forming the IV.
How to improve:Explain that the researcher selects or compares existing groups.
Mistake: Assuming a controlled room automatically makes a study a laboratory experiment.
Why this is incorrect:A quasi-experiment may use a controlled room while comparing a pre-existing IV.
How to improve:Classify the study according to how the IV was created before considering the setting.
Mistake: Saying natural and quasi-experiments can establish causation as confidently as laboratory experiments.
Why this is incorrect:They usually lack random allocation and have less control over alternative explanations.
How to improve:Use cautious language such as “associated with” or “may have affected”.
Mistake: Treating all extraneous variables as confounding variables.
Why this is incorrect:An extraneous variable becomes confounding only when it differs systematically between conditions and could explain the result.
How to improve:State exactly how the variable changes with the IV.
Mistake: Saying that natural experiments have no ethical issues because the researcher did not cause the event.
Why this is incorrect:Research may still involve distress, sensitive information, privacy and informed consent.
How to improve:Evaluate how participants are recruited, assessed and protected.
Mistake: Saying a natural experiment can always be replicated.
Why this is incorrect:Naturally occurring events may be unique or unpredictable.
How to improve:Distinguish repeating the measurement procedure from recreating the original event.
Mistake: Ignoring participant variables in a quasi-experiment.
Why this is incorrect:Pre-existing groups may differ in several ways besides the IV.
How to improve:Identify a specific group difference that could affect the DV.
Mistake: Describing a correlation as a quasi-experiment simply because the variables were not manipulated.
Why this is incorrect:A correlation measures two co-variables, whereas a quasi-experiment compares pre-existing groups on a DV.
How to improve:Identify whether the study uses group conditions or measures a relationship.
Mistake: Writing separate descriptions when asked to compare experimental types.
Why this is incorrect:A comparison requires direct similarities and differences.
How to improve:Use terms such as “whereas”, “both”, “unlike” and “in contrast”.
Exam-Style Questions ✍️
Question 1
Which one of the following best describes a natural experiment?
A. The researcher manipulates an IV in a controlled environment
C. The IV arises through an event not created by the researcher
D. The researcher measures two co-variables
[1 mark]
Question 2
Define a quasi-experiment.
[2 marks]
Question 3
A company introduces flexible working hours for business reasons. A psychologist compares employee stress before and after the change.
Explain why this is a natural experiment.
[3 marks]
Question 4
A psychologist compares younger and older adults on a test of memory.
Explain why this is a quasi-experiment.
[3 marks]
Question 5
Explain one difference between a natural experiment and a field experiment.
[4 marks]
Question 6
Explain one difference between a quasi-experiment and a laboratory experiment.
[4 marks]
Question 7
A local authority introduces additional lighting in one neighbourhood. A psychologist compares residents’ feelings of safety with those of residents in another neighbourhood.
Identify one possible confounding variable and explain how it could affect the findings.
[4 marks]
Question 8
A psychologist compares multilingual and monolingual students on a problem-solving task completed under standardised laboratory conditions.
Explain one strength and one limitation of using a quasi-experiment for this investigation.
[6 marks]
Question 9
A school changes its starting time from 8.30 am to 9.15 am. Researchers assess students’ alertness before and after the change.
Explain one strength and one limitation of using a natural experiment to investigate the effect of the timetable change.
[6 marks]
Question 10
Compare natural and quasi-experiments with laboratory and field experiments.
Refer to manipulation, control, validity and causal conclusions in your answer.
[8 marks]
Comments