Experimental Research An experiment is a form of quantitative research that tests causal relationships. The researcher manipulates and controls the conditions under which individuals are observed to behave. Experimental research starts with a hypothesis and then modifies something in a particular relationship. The researcher has control over the environment, variables and individuals under study. At the end of the experiment, the outcome is compared with the situation before the modification. An experiment consists of a number of components:

Treatment or independent variable

Dependent variable

Pre-test

Post-test

Experimental group

Control group

Random assignment

Classical Experimental, Pre-Experimental, Quasi-Experimental and the Solomon Four-Group designs all differ in how they treat these components, thus impacting the reliability and validity of the experiment.

Classical Experimental Design comprises random assignment of cases to groups, a pre-test and a post-test, an experimental group and a control group. Each group is exposed to different conditions or stimulus materials. Random assignment is used to increase the likelihood that each group will contain individual with equal measures of characteristics. The pre-test measures the dependent variable prior to the introduction of the experimental conditions and the post-test measures the...

The validity threats to the Classical Experimental Design have to do with its artificial nature; external and construct validities are the primary concerns. The threats to external validity include:
interaction between testing situation and the experiment stimulus interaction between the experiment stimulus and the subject interaction between the experiment stimulus and history.

Often, it isn't possible to adopt a classical experimental design because it may not be possible to assign subjects at random to different conditions or treatments. Some Pre-Experimental Design studies solve this issue by either omitting the use of control groups or pre-tests. For example, a one-group, post-test design has only one group, a treatment and a post-test. But, this design is flawed because it has no way of telling whether the treatment had any influence on the group as distinct from other untested factors. A one-group pre-test/post-test design uses one group, a pre-test, a treatment and a post-test. This design is better than the one-group, post-test design because respondents are compared against themselves. But, because it lacks a control group, it leaves open the possibility that changes in the dependent variable might be a consequence of something other than the treatment.

Quasi-Experimental Designs make the identification of causal relationships more valid…