Experimental Methods and Types of Experiment

An experiment is a research method in which the researcher investigates whether one variable has an effect on another. To do this, the independent variable (IV) is manipulated (or is allowed to vary) so its effect on the dependent variable (DV) can be measured.

• The independent variable (IV) is the thing that changes between conditions — the presumed cause.
• The dependent variable (DV) is what you measure — the presumed effect.
• Extraneous variables are any other variables that could affect the DV; the researcher tries to control them so they do not obscure the effect of the IV.

Suppose you want to know whether caffeine improves memory. You give one group caffeine and another group none (the IV), then measure how many words each group recalls from a list (the DV). Everything else — the word list, the room, the time allowed — is held constant so that any difference in recall can be attributed to the caffeine.

The defining feature of an experiment is that the IV is deliberately changed while other variables are controlled, so that its effect on the DV can be isolated.

Because the IV is manipulated and confounding factors are controlled, an experiment is the only method that can establish cause and effect. Methods such as correlations, observations and self-report can show that variables are linked, but they cannot show that one causes the other.

AQA requires you to know four types of experiment. The key that separates them is two questions:

1. Does the researcher manipulate the IV? (Or does it vary on its own?)
2. Where does the study take place? (A controlled setting or a natural one?)

Type	Who or what sets the IV?	Setting
Laboratory experiment	Researcher manipulates the IV	Controlled environment
Field experiment	Researcher manipulates the IV	Natural, everyday setting
Natural experiment	IV varies naturally (an event or condition), not manipulated	Often a real-world setting
Quasi-experiment	IV is a pre-existing feature of the participants, not manipulated	Often a controlled setting

The first split is control of the IV. In laboratory and field experiments, the researcher actively changes the IV. In natural and quasi experiments, the researcher does not: they take advantage of an IV that already exists or that arises on its own, and simply record the DV.

Get the manipulation question right first. If the researcher does not manipulate the IV, the study can only be a natural or quasi-experiment — never a lab or field experiment.

The next slides define each type with an example, then set out its strengths and limitations.

In a laboratory experiment, the researcher manipulates the IV inside a controlled environment and holds extraneous variables constant. It does not have to be a room full of equipment; "laboratory" here means a setting the researcher controls. Example: Loftus and Palmer tested whether the wording of a question ("smashed" vs "hit") changed the speed participants estimated for a car crash they had watched on film — a controlled setting with a manipulated IV.

In a field experiment, the researcher still manipulates the IV, but does so in a natural, everyday setting where participants normally behave. Example: a researcher varies whether a confederate who collapses on a train appears drunk or ill (the IV) and records how many passengers help (the DV).

	Laboratory experiment	Field experiment
IV	Manipulated by researcher	Manipulated by researcher
Setting	Controlled	Natural / everyday
Strengths	High control of extraneous variables; highly replicable; can infer causation	Higher ecological validity; behaviour is natural, so fewer demand characteristics
Limitations	Artificial setting can lower ecological validity; demand characteristics more likely	Less control of extraneous variables; ethical issues (often no informed consent); harder to replicate

Demand characteristics are cues that lead participants to guess the study's aim and change their behaviour. Ecological validity is the extent to which findings generalise to real-life settings.

The trade-off between these two types is control versus realism. A lab buys control at the cost of realism; a field study buys realism at the cost of control.

In a natural experiment, the researcher does not manipulate the IV. The IV varies naturally — through a real-world event or a naturally occurring condition — and the researcher records its effect on the DV. The researcher has no control over who ends up in which condition; they simply measure what happens.

Example: when a remote community first gains access to television, researchers can compare children's aggression before and after. No one manipulated the arrival of television — it happened naturally — but its effect on behaviour (the DV) can be measured.

Strengths	Limitations
Allows study of variables that could not be manipulated for practical or ethical reasons (you cannot ethically inflict a disaster on people)	Naturally occurring events are rare, so studies are hard to replicate
High external validity because the IV is a real event, not an artificial one	Participants are not randomly allocated to conditions, so causation is harder to establish
	Possible confounding variables the researcher cannot control

The IV in a natural experiment is genuinely an event or condition in the environment: a flood, a change in the law, a school closure, the introduction of a new technology.

Because the researcher does not control who experiences the IV, a natural experiment can show that a real event was associated with a change in the DV, but it cannot safely prove the event caused it. Confounding variables may offer an alternative explanation.

In a quasi-experiment, the IV is a pre-existing difference between people — a fixed characteristic they already have — rather than something the researcher manipulates. Common examples of such an IV are age, gender, or having a particular diagnosis (for example comparing people with and without OCD).

Example: comparing the memory scores of a group of 20-year-olds and a group of 70-year-olds. Age is the IV, but the researcher did not (and could not) assign people to an age; participants arrived already belonging to one group.

Strengths	Limitations
Allows comparisons that would otherwise be impossible, since the characteristic cannot be manipulated	No random allocation, so confounding participant variables may explain the results
Often carried out under controlled conditions, giving some of the rigour of a lab study	The researcher cannot conclude that the IV caused any difference between the groups

Note the subtlety: a quasi-experiment can still be run in a controlled setting with a well-standardised procedure. What makes it quasi is not the setting but the fact that the IV is a fixed feature of the participants, so people cannot be randomly allocated to conditions.

Without random allocation, the two groups may differ in ways beyond the IV (their participant variables). A difference in the DV might be due to one of those differences, not the IV itself — which is why causation cannot be safely claimed.

The most commonly confused pair is natural and quasi-experiments, because in both the researcher does not manipulate the IV. Students who stop there get them mixed up. The difference is what the un-manipulated IV actually is.

	Natural experiment	Quasi-experiment
Does the researcher manipulate the IV?	No	No
What is the IV?	An event or condition in the environment	A fixed characteristic of the participants themselves
Example of the IV	A flood, a new law, TV arriving in a town	Age, gender, having a diagnosis
Could the IV, in principle, be applied to anyone?	Yes — the event could have happened to a different group	No — it is who the participant already is

A quick test: ask whether the IV is something that happened to the participants (natural) or something the participants are (quasi). A hurricane is something that happened — natural. Being 70 years old, or having a phobia, is something the participant is — quasi.

Both natural and quasi-experiments lack a manipulated IV and lack random allocation, so both are limited in establishing cause and effect. They differ only in the source of the IV: an environmental event versus a pre-existing personal characteristic.

1. Saying the researcher manipulates the IV in a natural experiment

They do not. In a natural experiment the IV varies on its own through a real-world event or condition, and the researcher only records the DV. If the researcher had manipulated the IV in a natural setting, it would be a field experiment, not a natural one.

2. Confusing natural and quasi-experiments

In both, the IV is not manipulated. The distinction is the source of the IV: an event or condition in the environment (natural) versus a fixed characteristic of the participant such as age or a diagnosis (quasi). Name which one the IV is, and justify it.

3. Thinking a field experiment has no control at all

A field experiment still has a manipulated IV and some control over the procedure — it simply has less control than a lab study because it takes place in a natural setting. "Less control" is not the same as "no control".

4. Claiming a quasi-experiment can prove causation

Because the IV is a pre-existing difference and there is no random allocation, confounding participant variables may explain the results. A quasi-experiment can reveal a difference between groups, but it cannot safely conclude the IV caused it.

5. Mixing up the IV and the DV

The IV is what changes or varies between conditions (the presumed cause); the DV is what you measure (the presumed effect). Always state which is which, and remember you operationalise both — define exactly how each is set or measured.

6. Assuming "laboratory" always means a room with equipment

A laboratory experiment is defined by a controlled environment and a manipulated IV, not by lab coats or apparatus. A classroom or office can host a lab experiment if the researcher controls the conditions and manipulates the IV.