4.2.3 - Experimental method and types of experiment

How experiments test cause and effect

Psychologists use experiments when they want to test whether one factor can influence another. The researcher identifies a factor that varies, then measures whether an outcome changes as a result. In AQA research methods, the essential vocabulary is independent variable, dependent variable, control and validity.

Manipulation is the key distinction here: in laboratory and field experiments the researcher creates the comparison, whereas natural and quasi-experiments use a difference the researcher did not create. All four are named in AQA's specification. A manipulated IV makes a causal test possible, but the label alone does not establish causation. We must also ask how participants entered the conditions, what else differed and whether the DV measured the intended outcome.

That distinction is the best first question to ask in an exam scenario:

To reason from a result to a cause, ask what the comparison rules out. Using the same learning time rules out extra study time; randomly allocating participants reduces systematic differences in their starting memory ability. Neither step guarantees perfect control. If the IV already exists, important differences can still be measured, but other pre-existing differences are harder to rule out.

Suppose one group tries a new revision app while another studies the same material for the same time without the app. The second group is a control group: an improvement in the app group alone might otherwise reflect ordinary revision. An active control, such as another app with the same contact time, can help separate the new app's particular features from attention, expectations or simply spending time revising. A control group is a comparison condition, not the same thing as keeping room temperature constant. It is also not automatically equivalent to the treatment group; allocation and other controls still matter.

Laboratory experiments

Laboratory experiments bring the situation under deliberate researcher control.

A laboratory experiment does not have to take place in a science laboratory. It could be in a classroom, testing room, computer suite or any setting arranged so the researcher has high control. The defining features are researcher manipulation of the IV and control over the situation.

The main strength of laboratory experiments is the opportunity for high internal validity. If the researcher controls noise, timing, instructions and materials, then differences in the DV are less likely to be caused by those alternative influences. Control must be demonstrated, not assumed from the word laboratory. Standardised procedures also make replication easier, allowing another researcher to check whether the finding is reliable.

A controlled word-learning experiment illustrates this logic. A researcher gives participants the same word list and learning time, assigns them to silence or music, and records the number of words recalled. Keeping the list and timing constant helps rule out easier words or longer study time as explanations. Random allocation, when used, also reduces systematic participant differences between groups; it does not guarantee that the groups are identical.

The classic limitation is lower ecological validity. The environment or task may be unlike everyday life, so participants may behave differently from how they would in normal situations. Laboratory experiments can also produce demand characteristics: participants may guess the aim or respond to the situation of being studied. A further issue is low mundane realism if the task itself is artificial, such as recalling disconnected word lists.

AO2 example: Dr Mensah tests whether sleep affects attention by asking participants to sleep either four hours or eight hours before completing a computer reaction-time task in a controlled room. This is a laboratory experiment because Dr Mensah manipulates the IV and measures the DV under controlled conditions. The control helps internal validity, but the computer task may not fully represent attention in real life.

Field experiments

Field experiments test a manipulated variable in an everyday setting.

In a field experiment, the researcher changes something in the participants' normal environment rather than bringing participants into a controlled setting. The IV is still manipulated by the researcher. The everyday setting distinguishes it from a laboratory experiment; the exact degree of control depends on the procedure.

For example, a researcher could change whether a café sign says “Please clear your tray” or “Most customers clear their trays”, then count trays cleared during fixed observation periods. The wording is deliberately changed, while clearing trays is measured in the setting where it normally happens. That makes it a field experiment. If one sign is always used on quiet mornings and the other at busy lunchtimes, the method label does not rule out time of day as an alternative explanation.

The key AO3 strength is ecological validity. Participants are often in a real-life situation and may not know they are in a study, so their behaviour may be more natural. Field experiments can also have high mundane realism because the task or social situation is part of everyday life.

The trade-off is lower control. In a subway, school corridor or workplace, many extraneous variables may affect behaviour: time of day, crowd size, participant mood, weather, social norms or unexpected events. This weakens internal validity because it is harder to isolate the effect of the IV. Field experiments may also raise ethical issues when participants are unaware they are being studied and therefore cannot give fully informed consent. Replication is harder because the same real-world conditions cannot be recreated exactly.