Internal and External Validity
Validity is a fundamental consideration when designing a study. Internal validity and external validity are the major types of validity in research design. For a study to be effective, both internal validity and external validity are taken into account. However, certain factors may jeopardize internal validity, as well as external validity. These factors, commonly referred to as threats, have a major effect to internal and external validity. This paper analyses the different types of threats. The paper also delves into the importance of considering these types of threats when designing a study.
Last but not least, the paper will make clear the trade off between internal and external validity. Internal validity refers to the dependability of an experiment’s results to the aims of the experiment, whereas external validity is the generalizability of an experiment’s results to environment and groups outside of the research setting. As Campbell and Stanley (1963) presented them, there are eight factors that threaten internal validity, namely: history, maturation, testing, instrumentation, biases, statistical regression, experimental mortality and selection maturation interaction.
Similarly, four factors threaten the external validity of experiments: effects of testing, effects of selection, effects of experimental arrangements and multiple treatment interference. (Kaifang Yang & Gerald J. Miler, Pg. 112) Each of these threats is important to consider when designing a study due to various reasons. For instance, internal validity only has to do with descriptive or explanatory experiments, where a researcher’s interest is on whether a variable or an occasion resulted to another.
If the researcher mistakenly concludes that there is a causal relationship between both occasions without knowing that a third occasion may have caused the result, such a design is ineffective as it has failed to deal with the history threat, since such an unanticipated event has occurred while the experiment was in progress. This example of a threat is also called the teacher effect or most commonly, confounding effect. A study can give inappropriate results if the researcher failed to consider other threats such as maturation, as aforementioned.
Maturation due to physical development or age is an important consideration as it can alter the study’s results. There is also need for a researcher to consider threats of external validity. It is important for anyone designing a study to consider effects of selection. Selecting a sample that is not representative of the population of interest is a major threat to external validity because many researchers may fail to follow the guidelines needed in representative sampling. It is important to note that small, random samples are never representative as this can lead to big errors in determining results.
This may result if the general rule of “sampling with replacement” is not adhered to. It is important that one considers all threats to avoid making errors. Since external validity is about generalization, it is vital for researchers to understand that errors are prone to be present in for example, a census conducted before a migration period would give ineffective results to generalize the population of such a nation before and after the migration. The trade off between internal and external validity states that internal validity is inversely proportional to external validity.
That is, increasing validity of an experiment’s causal inference in a particular time, place and with particular subjects may result to a decrease in the generalizable causal inference to other places, at another time and subjects. This relationship is especially evident in randomized experiments. The language of internal and external validity should be understood by anyone conducting a design study so as to obtain precise, unbiased inferences. Reference: Yang K. , Miller G. J. 2007. Handbook of Research Methods in Public Administration, CRC Press, 1998
Sample Essay of AssignmentExpert.com