Research Designs

Research designs can be broadly classified into two categories: quasi experimental research designs and experimental research designs.

The second category of research designs, the experimental studies, portray the skill of randomizing the treatments and the control groups. The randomization  consists of controlling the variables that are not included openly in the study. Because the groups used for comparison are not truly randomized, control groups in the other type of research designs are quasi experimental designs.

Thus, quasi experimental research design is sometimes called correlational designs.

An experimental research design is said to be experimental only if the area of focus is being randomly assigned to treatment groups or the control (or the comparison) groups. Cook and Campbell (1979) highlights ten different types of experimental research designs.

Statistics Solutions can assist you in deciding which research design is right for your study.

Classical experimental research designs involve randomization of the subjects into control and treatment groups. These experimental designs are divided into three categories: between subject classical experimental research designs, within subjects (repeated measures) classical experimental designs and matched pair classical experimental research designs.

Between subject classical experimental research designs involve the comparison between the subjects that encounter different treatments. In this type of classical experiment design, there are different subjects for each level of the independent variable(s). Examples of this include factorial designs and randomized block designs.

In within subject classical experimental research designs, the comparison is made on the same subjects. Hence it is termed “within subjects research designs.” In this type of classical experiment research design, the same subjects are used for each level of the independent variable. Thus the subject has their own controls.

Matched pair classical experimental research design is better than the other two types of classical experimental research designs. This research design avoids invalidity of the within subject research designs. This research designs controls the threat of subject fatigue only for the matched subject characteristics.

As the name suggests, lottery experimental research designs are used by the researcher when lotteries are expected.

Mandated control experimental research designs are used in the case of the military where the control levels are sufficiently high – so much so that the random task of randomizing the treatment and control conditions are accepted.

Waiting list experimental research designs are used when the demand of a particular thing outdoes the supply of that thing. In other words, as the name suggests, this type of experimental research design is involved in those cases where there exists a waiting list group or some waiting queue expecting their demands be fulfilled, etc.

Equivalent time series experimental research designs are used when the treatment cannot be given to all the groups in a simultaneous manner. One example of this is training sessions that are carried out to train different groups of people.  This is a type of equivalent time series experimental research design.

Spatial separation experimental research designs are used when the treatment groups are separated and there is no experiment-related-communication with each other.

Mandated change/unknown solution experimental research designs are required when there is a problem that needs to be changed, but there is no clear solution for the problem.

Tie breaking experimental research designs are used when the treatments that are to be received are based on merit.

The assumption that is made in experimental research designs is that the researcher working with the experimental research designs assumes that he has considered all types of threats that are associated with the experimental research designs.