The issue of validity applies to any research or study and requires the researchers to develop methods and designs that would help to control the variables and achieve internal validity. Internal validity is “the extent to which [the experiment’s] design and the data it yields allow the researcher to draw accurate conclusions about cause-and-effect and other relationships within the data” (Leedy & Ormrod, 2015, p. 103). The main threat to internal validity of the experiment is represented by the confounding variables, such as outside events, change in participants’ qualities, testing, instrumentation, selection bias, and so on (Leedy & Ormrod, 2015, p. 199). These factors can significantly undermine the soundness of the researcher’s conclusions. Despite the fact that these confounding variables are sometimes out of the researcher’s control, there are some methods that might help to decrease their influence on the results, providing some degree of control over the outcome of the study.
specifically for you
for only $16.05 $11/page
First of all, Leedy and Ormrod (2015) stress the need for stabilizing the environment of the experiment in order to control the variables that can be controlled (p. 198). This can be achieved by conducting the experiment in a laboratory setting with controlled humidity, temperature, noise levels, and other factors that could influence the final result in one way or another. Secondly, the use of control group is strongly recommended (Leedy & Ormrod, 2015, p. 198) to provide a comparative measure that helps to ensure that any changes in the experimental group were provoked by the researcher’s intervention. Other strategies include random sampling, using pretests, and exposing all the participants to all of the studied treatments (Leedy & Ormrod, 2015). These methods are only applicable at the stage of designing the experiment. In order for the researcher to exclude confounding variables from the existing results, statistical analysis method can be used (Pourhoseingholi, Baghestani, & Vahedi, 2012, p. 79). It involves measuring the effects of confounding variables on the obtained results by collecting information on them (Pourhoseingholi et al., 2012, p. 80)
In my study of the impact of early childhood caries on academic performance and school absence in children from ethnic minorities and low socioeconomic standing, there are many confounding factors that could affect the results and conclusions regarding the correlation between the variables. For instance, school performance can be highly affected by outside events, which would make it difficult to confirm the correlation between caries and academic experiences. Leedy and Ormrod (2015) discuss an experiment where a psychological study attempted to investigate the effect of music on work performance of employees.
To decrease the influence of the history factor, i.e. the outside events, on work performance, Leedy and Ormrod (2015) recommend the use of a control group (p. 198). Apparently, my research could also benefit from the use of this method. However, it would be impossible to completely eliminate the confounding variables by engaging a control group, which is why statistical analysis method could be used to filter the end results. For example, the Analysis of Covariance (ANCOVA) could help to predict the influence of confounding factors and thus to remove those that could potentially affect the outcome: “ANCOVA tests whether certain factors have an effect on the outcome variable after removing the variance for which quantitative covariates (confounders) account” (Pourhoseingholi et al., 2012, p. 81). This analysis could significantly increase the reliability of results, providing for a well-grounded conclusion.
Leedy, P. D., & Ormrod, J. E. (2015). Practical research: Planning and design (11th ed.). Upper Saddle River, NJ: Pearson.
Pourhoseingholi, M. A., Baghestani, A. R., & Vahedi, M. (2012). How to control confounding effects by statistical analysis. Gastroenterol Hepatol Bed Bench 5(2): 79-83.