Experimental method: how

First, a very brief overview of experimental design.

When writing about the nature of experimental research, there are some key traits of the method that you should be able to discuss.

• Hypothesis testing - experiments propose a hypothesis and then create a situation in which to test whether or not that hypothesis can be supported.
• Variables - An experiment manipulates an independent variable and measures the effect on a dependent variable. Some experiments have more than one IV and DV. Other variables, called extraneous or confounding variables, are controlled - or at least there is an attempt at controlling them so that they do not affect the outcome of the experiment. Control of variables raises internal validity and helps to establish a cause-and-effect relationship.
• A big question in experiments is "who manipulates the independent variable?"  When the researcher does it and participants are randomly allocated to groups, this is a "true" experiment.  If participants are not randomly allocated, this is a quasi-experiment. If the research does not manipulate the IV, this is a natural experiment.
• Random allocation - it is important that participants are randomly allocated to conditions. When done in a large enough sample, random allocation is an effective way to reduce bias.
• A representative sample should be obtained. This means that the sample should be large enough and consist of participants that reflect the nature of the population being studied.
• Experiments can be done in a laboratory or in a more natural environment. When done in the "real world", these are called field experiments. Obviously, field experiments have less control over variables than lab experiments.
• Experiments use designs - for example, a repeated measures or independent samples design. In a repeated measures design, to make sure that the order of the conditions is not influencing the results, counter-balancing is often used - that is, some participants are given condition 1 first and then condition 2; other participants are given condition 2 and then condition 1.

Experiments in the biological approach

• One assumption of the biological approach is that animals may be used to give us insight into human behaviour. Of the three approaches that we study, only biologists use animals in their research. This, of course, leads to the problem of generalizability.
• Biologists often use double-blind testing - that is, neither the participants nor the researchers know which participants belong to the control group or the test group. Only after all data have been recorded do the researchers learn which participants were which. Performing an experiment in double-blind fashion can greatly lessen biases that may distort the results.
• Placebos are often used in order to determine whether just getting a treatment was enough to change behaviour.
• Matched pairs are also often used. This is when participants are allocated to groups based on their characteristics so that all groups to be tested have a relatively equal distribution of a variable. In the Whitehall study, for example, participants were matched for smoking. So, smokers were compared at different levels of the hierarchy to make sure that smoking, rather than their place in the social hierarchy, was not the reason for their level of cardiovascular health.

Experimental method: why

There are several reasons why researchers use experiments. The most significant is the ability to establish a cause and effect relationship. Since the testing conditions are tightly controlled, a strong argument can be made that the change in the IV led to the change in the DV. Experiments are used because we want to eliminate extraneous variables and focus on the role of the IV.

Biologists often employ a reductionist approach. Because of this, experimental research that looks to isolate variables is a logical choice of method.

Experiments also are highly standardized and can be replicated. This makes it possible for other psychologists to replicate the study and hopefully support the findings by obtaining the same results under the same or similar conditions. When this happens, we can say that the research is reliable.

The method can also objective; there Is very little chance for the experimenter's bias and personal opinions to affect the results. The use of single and double-blind methods, as well as the use of technology to assist in measurement, help to guarantee high levels of objectivity.

Experiments yield quantitative data which can be analysed using inferential statistical tests. These tests permit statements to be made about how likely the results are to have occurred by chance.