Confounding Variables and Validity

Confounding Variables

A variable that is confounding is an external variable that has a statistical correlation with the independent variable. This also indicates that whenever there is a change in the independent variable the confounding variable changes along with it. If sometimes the researcher may miss taking the confounding variable into account, then   the conclusion derived may be false that there is a causal relationship between the dependant and independent variable.

Validity: When discussing the specifics of data analysis, it is not right to ignore the importance of validity in any study. Validity exists at each stage of the research process. One stage of outcome of not proving validity has been discussed above where  when there is a failure in operationalising the variables of interest the conclusion that  are drawn are inappropriate in context to the research question.  In general terminology, validity refers to the design of the study and the results it provides that would be appropriate for generalisation of the population in interest. A researcher, for a complete research, has to focus on three kinds of validity:

Internal Validity:  It is applicable for those studies which target to establish a causal relationship between the variables. It helps to identify the degree of the inferences that the study is able to make. The core concept of internal validity is that it confirms that the outcome of the study is due to the manipulation of independent variable and nothing more.

Construct Validity: this type of validity has a link with operationalization. It talks about the extent to which the researcher is able to claim the accurate inferences. Any study can claim to have construct validity if the researcher is able to show that all the variables in the study were operationalised in the right way.

External Validity: In order to generalise results  for a larger population the researcher needs to prove external validity which claims that the results drawn from the chosen sample are good enough to be generalised for the entire population.

Throughout the design of the study, it is important to keep the concept of validity applicable. Validity has to be preserved at each stage of the research  if it lacks validity then the conclusions that are drawn may be not applicable for the population as a whole and sometimes may show even dangerous outcomes.





Leave a Reply