Science for Everybody

Table of Contents

• Introduction
• Types of Experiment: An Overview
• Laboratory Experiments
• Field Experiments
• Natural Experiments
• Quasi-Experiments
• Comparing the Four Types
• True Experiments and Causal Conclusions
• Evaluation
• Strengths
• Weaknesses
• Summary
• Quick Questions
• Key Terms

---

Introduction

Psychologists use different types of experiment depending on how much control they need and how naturally they want behaviour to occur. Laboratory, field, natural and quasi-experiments differ mainly in where the study happens, whether the researcher manipulates the independent variable, and how confidently cause and effect can be inferred.

Choosing the right type of experiment involves a trade-off between control, realism, ethics and practical feasibility. Understanding those trade-offs is central to evaluating psychological research.

---

Types of Experiment: An Overview

All experiments compare conditions and measure a dependent variable. The key distinction is who or what produces the independent variable. In laboratory and field experiments, the researcher directly manipulates the IV. In natural experiments, the IV changes because of an event or circumstance outside the researcher’s control. In quasi-experiments, the IV is a pre-existing difference between groups.

This distinction matters because direct manipulation generally makes causal conclusions stronger. However, greater control can make behaviour less natural, while naturally occurring conditions may improve realism but introduce more uncontrolled variables.

Laboratory Experiments

A laboratory experiment takes place in a highly controlled environment. The setting does not have to be a literal laboratory; it may be a classroom, testing room or other controlled space. The researcher manipulates the IV, measures the DV and tries to keep extraneous variables constant.

The main strength of a laboratory experiment is control. Standardised instructions, controlled materials and carefully managed conditions make it easier to isolate the effect of the IV. This usually gives laboratory experiments high internal validity and makes exact replication easier.

However, laboratory settings and tasks can feel artificial. Participants may realise they are being studied and respond to demand characteristics. As a result, findings may have lower ecological validity or mundane realism than behaviour observed in everyday settings.

Gilchrist and Nesberg (1952): Laboratory Experiment Example

Aim: To investigate whether food deprivation influenced how participants perceived food-related stimuli.

Procedure or method: Participants were deprived of food and water for four hours and then shown pictures of food under controlled conditions. Their ratings were compared with a control group that had not been food deprived.

Findings: Food-deprived participants rated the pictures of food as brighter than participants in the control group.

Conclusion: A motivational state such as hunger can influence perception. The controlled setting makes this a useful example of a laboratory experiment because the researcher deliberately created different conditions.

Field Experiments

A field experiment takes place in a natural, everyday setting. The researcher still manipulates the IV, so it remains a true experiment, but participants often carry out ordinary behaviour in a real-world context.

Field experiments usually have higher ecological validity than laboratory experiments because the setting and behaviour are more natural. Participants may also be less aware that they are being studied, which can reduce demand characteristics.

The disadvantage is reduced control. Extraneous variables are harder to manage, replication is more difficult and researchers may face ethical issues if participants do not know that they are taking part.

Piliavin et al. (1969): Field Experiment Example

Aim: To investigate helping behaviour in a real-world situation and whether characteristics of a person needing help influenced bystander responses.

Procedure or method: On a busy New York subway, a researcher pretended to collapse. Different versions of the situation were staged, including a victim carrying a walking stick and a victim who appeared to smell of alcohol. Passengers’ helping behaviour was recorded.

Findings: More people helped when the victim carried a walking stick than when the victim appeared to have been drinking alcohol.

Conclusion: Characteristics of the person needing help can influence prosocial behaviour. Because the IV was manipulated in a natural setting, the study illustrates a field experiment.

Natural Experiments

In a natural experiment, the researcher takes advantage of a naturally occurring change in the IV. The researcher does not create the conditions. For example, a new law, environmental change or naturally occurring event might affect one group or one time period but not another.

A natural experiment is defined by how the IV arises, not by the location of the study. Participants could still be tested in a controlled room or laboratory after the natural change has occurred.

Natural experiments are valuable when manipulation would be impossible or unethical. However, participants cannot usually be randomly allocated to conditions, so pre-existing differences and other uncontrolled factors may confound the results.

Williams (1986): Natural Experiment Example

Aim: To investigate whether the introduction of television affected aggression in children.

Procedure or method: Williams monitored the behaviour of 6- to 11-year-old children in a Canadian town before and after television was introduced for the first time. The researcher did not decide when or whether television would be introduced, so the change in the IV occurred naturally.

Findings: Levels of aggression increased significantly after the children gained access to television.

Conclusion: The findings suggested that television exposure may influence aggressive behaviour. However, because the researcher did not control the naturally occurring change, alternative explanations cannot be ruled out as confidently as in a laboratory experiment.

Quasi-Experiments

A quasi-experiment uses an IV that is based on a pre-existing difference between people. Examples include age, biological sex, diagnosis or another characteristic that already exists before the study begins. The researcher measures the DV but cannot manipulate the defining characteristic or randomly allocate participants to that characteristic.

Quasi-experiments are often conducted under controlled conditions, which can reduce some extraneous variables. However, because the groups may differ in many ways besides the quasi-IV, it is harder to conclude that the IV alone caused any difference in the DV.

Strictly speaking, quasi-experiments are not true experiments because the researcher does not manipulate the IV directly.

Baron-Cohen et al. (1986): Quasi-Experiment Example

Aim: To investigate whether children from different pre-existing groups differed in their ability to arrange comic-strip stories into the correct sequence.

Procedure or method: Children with Down’s syndrome, children with autism and children described as ‘normal’ in the original source material completed a comic-strip sequencing task. Group membership already existed and was not created by the researchers.

Findings: Children in the autism group performed significantly worse when arranging the comic-strip stories into the correct sequence.

Conclusion: Performance differed between pre-existing participant groups. Because the researchers could not manipulate or randomly allocate diagnostic status, the study illustrates the logic of a quasi-experiment.

Comparing the Four Types

Laboratory experiments provide the highest level of control, so they are usually strongest for internal validity and replication. Field experiments keep researcher manipulation of the IV but move the study into a more natural setting, increasing realism while reducing control.

Natural experiments sacrifice direct manipulation because the IV occurs independently of the researcher. Quasi-experiments also lack direct manipulation because the IV is a participant characteristic that already exists. Both methods can answer important questions that cannot be investigated ethically or practically through direct manipulation, but both are more vulnerable to confounding variables.

A useful way to compare the methods is to ask two questions: Did the researcher manipulate the IV? How natural was the context in which behaviour was measured? Reflection question: which is more valuable in psychology – tightly controlled evidence about cause and effect, or behaviour that closely reflects everyday life? The strongest answer often depends on the research question.

True Experiments and Causal Conclusions

A true experiment involves direct researcher control over the IV. Laboratory and field experiments can therefore be true experiments. Natural and quasi-experiments are not true experiments in this strict sense because the researcher does not create the IV conditions.

This does not make natural or quasi-experiments unscientific. Instead, it changes what can be concluded. Researchers should be more cautious about claiming cause and effect when random allocation and tight control are not possible.

---

Evaluation

The value of each experimental method depends on the balance between control, realism, ethics and the type of causal conclusion the researcher wants to make.

Strengths

Laboratory Control Supports Causal Conclusions

A major strength of laboratory experiments is the high level of control over extraneous variables. When researchers standardise instructions and keep conditions constant, any systematic change in the DV is more likely to be linked to the IV. Therefore, laboratory experiments often have high internal validity and can provide strong evidence for cause and effect. Replication is also easier because the procedure can be described precisely and repeated by other researchers.

Field Experiments Can Produce More Natural Behaviour

Field experiments often achieve better ecological validity because participants behave in familiar, real-world surroundings. For example, helping behaviour on a subway may reflect everyday social behaviour more closely than a staged helping task in a laboratory. Participants may also be less aware of the research aim, which can reduce demand characteristics. Therefore, field findings may generalise more effectively to similar real-life settings.

Natural and Quasi-Experiments Make Important Research Possible

Natural and quasi-experiments allow psychologists to investigate variables that cannot be manipulated ethically or practically. Researchers cannot randomly assign a child to have autism, nor should they deliberately expose communities to potentially harmful social changes. Using naturally occurring events or pre-existing groups therefore expands the range of questions psychology can study while avoiding unethical manipulation.

Different Experiment Types Allow Methodological Choice

The four experiment types provide researchers with different balances of control and realism. This is useful because no single method is best for every question. A researcher interested in precise causal mechanisms may favour a laboratory experiment, whereas a researcher interested in real-world helping behaviour may choose a field experiment. Matching the method to the research aim can improve the overall validity of psychological evidence.

Weaknesses

Laboratory Experiments May Lack External Validity

The control that makes laboratory experiments powerful can also make them artificial. Unusual tasks, unfamiliar surroundings and awareness of being tested may change participant behaviour. This can reduce mundane realism and ecological validity, so findings may not generalise well to everyday life. Demand characteristics can further weaken validity if participants guess the aim and alter their responses.

Field Experiments Reduce Control and Create Ethical Problems

Field experiments take place in complex environments where many variables change at once. This makes confounding variables more difficult to identify and weakens certainty that the IV caused the DV. Replication can also be difficult because the exact real-world conditions may never recur. In addition, participants may not know they are taking part, creating possible problems with informed consent, privacy and the right to withdraw.

Natural and Quasi-Experiments Are Vulnerable to Confounding

Natural and quasi-experiments usually cannot randomly allocate participants to conditions. Groups may therefore differ before the study begins, and naturally occurring events may coincide with other changes. This creates alternative explanations for the findings and reduces internal validity. Researchers can use matching, statistical controls or carefully selected comparison groups, but these strategies cannot guarantee that every confound has been removed.

Samples, Culture and Time Can Limit Generalisation

The usefulness of an experiment also depends on who is studied, where the study takes place and when it occurs. A field experiment on one transport system or a natural experiment in one community may reflect local social norms rather than universal behaviour. Quasi-experiments can also involve small or specialised groups, increasing sample bias. Finally, older studies may have lower temporal validity because social attitudes, technology and everyday environments change over time. Researchers should therefore avoid assuming that findings automatically generalise across cultures, populations or historical periods.

Quick Questions

What is the defining feature of a laboratory experiment?

The researcher manipulates the IV and measures the DV in a highly controlled environment.

Why can field experiments have higher ecological validity than laboratory experiments?

They study behaviour in a more natural, everyday setting.

What makes an experiment natural rather than field-based?

In a natural experiment, the IV changes naturally and is not manipulated by the researcher.

Why can participants not usually be randomly allocated in a quasi-experiment?

The IV is a pre-existing characteristic, such as age or diagnosis, that the researcher cannot assign.

Which two types can be described as true experiments?

Laboratory and field experiments, because the researcher directly manipulates the IV.

Key Terms

Experiment – A research method used to investigate whether changes in one variable are associated with changes in another.

Independent variable (IV) – The variable that is changed by the researcher, or the naturally occurring difference used to create conditions.

Dependent variable (DV) – The variable that is measured to see whether it changes as a result of the IV.

Laboratory experiment – An experiment conducted in a highly controlled environment in which the researcher manipulates the IV and measures the DV.

Field experiment – An experiment conducted in a natural setting in which the researcher manipulates the IV and measures the DV.

Natural experiment – An experiment in which the IV changes naturally rather than being manipulated by the researcher.

Quasi-experiment – A study in which the IV is a pre-existing difference between people, such as age or a diagnosis, rather than a variable created by the researcher.

Extraneous variable – Any variable other than the IV that could influence the DV.

Confounding variable – An uncontrolled variable that changes systematically with the IV and provides an alternative explanation for the findings.

Internal validity – The extent to which a study measures the effect of the IV on the DV without interference from other variables.

External validity – The extent to which findings can be generalised beyond the study, including to other people, settings and times.

Mundane realism – The extent to which the tasks and setting resemble experiences participants might encounter in everyday life.

Random allocation – A procedure that gives participants an equal chance of being placed in each experimental condition.