Quasi Experimental Research Design In Health Care

Quality Medical Service in Quasi Experimental

Quasi-Experimental Research,Implementation of Quasi Experimental Research,Pre-Experimental Design,Non Experimental Control Group,Compulsion on the use of Control Group and Randomization.

Quasi-Experimental Research

     Under “Experimental Research” in this encyclopedia, T. Cook and Campell's (1979) definition that experiments are characterized by manipulation, control, and randomization was cited. However, when conducting research in field settings, it is not always possible to implement design that meets these three criteria. 

    Quasi-experimental research is similar to experimental research in that there is manipulation of an independent variable. It differs from experimental research because either there is no control group, no random selection, no random assignment, and/or no active manipulation. 

Implementation of Quasi Experimental Research

    Quasi-experimental research is a useful way to rest causality in settings when it is impossible or unethical to randomly assign subjects to treatment and control groups or to withhold treatment from some subjects. 

    The main disadvantage of quasi-experimental research is the increased threat to internal validity (see “Experimental Research” for a review of types of design validity). Within quasi-experimental designs, a distinction is made between pre-experimental, nonequivalent control group designs, and interrupted time series designs. 

    Note also that the boundaries between experimental and quasi-experimental research have blurred. Often investigators like to define their study as experimental when in fact it is quasi-experimental.

Pre-Experimental Design

    Pre-experimental designs are the weakest of the quasi-experimental designs. They may lack a control/comparison group, observation before the intervention (commonly known as pretests), or both. 

    Their use is strongly discouraged because they do not permit even remote inferences about the direction and dynamics of change and causality.

Non Experimental Control Group

    Nonequivalent control group designs refer to situations in which naturally occurring groups of subjects are used as control/comparison group or those in which it is impossible or unethical to withhold treatment from a given group. 

    In spite of the absence of randomization, nonequivalent control group designs can be considered relatively strong designs. The use of a control group and a pretest significantly increase the strength of non-equivalent control group designs. Good pretest data will enable the researcher to improve the level of analysis. 

    When subjects from different settings are used, a nonequivalent control group design may control some threats to internal validity, such as compensatory rivalry and demoralization of controls. When subjects in each group are naturally kept separate, it is less likely that they will have contact with each other, and it is often useful to minimize contact between treatment and control groups.

Compulsion on the use of Control Group and Randomization

    In time series designs the researcher does not always use a control group and does not use randomization. An interrupted time series study uses several observations of subjects over time with a treatment given at a specified point (or longitudinally over time, with start and end time points).     

    A time series study can be designed to study the same individuals at specified intervals or to study different individuals at some common point in time. When the researcher studies one group of subjects, the subjects act as their own controls, which provides the researcher with equivalent control groups.

     Time series designs are used when a control group population is not available. When only one group is available to the researcher, the time series design significantly increases the strength of the research.