True Experiments

• True experiments are considered the most accurate form of experimental research. They are commonly used in the physical sciences, such as chemistry, geology and physics, to prove or disprove a hypothesis, or theory, but are more difficult to apply to social sciences, such as biology or psychology. The characteristic feature of a true experiment is that subjects are randomly assigned. This means that any difference between them is purely due to chance. For simplicity, just one variable is manipulated and tested.
  
  

Quasi Experiments

• Quasi experiments, also known as natural or in-situ experiments, are experiments in which subjects are beyond the control of the experimenter and are not randomly assigned. The fact that quasi experiments use intact groups means that they are typically less powerful research tools than true experiments because of the possibility of introducing systematic errors. Quasi experiments are not typically performed in the laboratory, but some variables in laboratory experiments, such as gender, are not randomly assigned. Quasi experiments are often used to investigate subjects other than people, such as plots of land in an agricultural or environmental study.
  
  

Bias

• In a true experiment, randomization means that it is unlikely that one subject is significantly superior or inferior to another. This means that variables associated with individuals are not sources of constant or systematic error, although there still may an element of bias in the overall design of the experiment. In a quasi experiment, however, nonrandom assignment of subjects may be a source of constant, but not always obvious, error. Subjects in a quasi experiment must be tested, or analyzed, prior to the start of the experiment.
  
  

Assignment

• In true experiments, randomization can be achieved by tossing a coin, rolling a die or using a computerized random number generator to assign treatments to subjects. In quasi experiments, nonrandom assignment -- based on first-come, first-served, perceived ability or some other criterion -- can significantly influence the outcome of the experiment, particularly if differences in immeasurable characteristics, such as intelligence, exist among subjects.