Introduction
Great technological advances that have equally been paralleled by increased availability of health related information both to health practitioners and the general public that is present in the form of statistics. This form of presentation is the one that easier to understand and comprehend for many people. As such, statistics play an important role in the field of health research.
As with any other research field, there is an important need for indivividuals to have good understanding of the concepts used in statistics so that they can read and comprehend research literature such as research reports and other forms of research publications. This is also an equally essential requirement for individuals to undertake credible research studies.
In public health research, some of the statistical concepts include statistical errors and their relationship to concepts like population and sample size. This paper will inform the importance of statistical errors in public health research by means of description and analysis of a contextual problem set as an example.
Types of errors
Two types of errors are commonly referred to in statistics. These are type 1 and type 2 errors. The type 1error occurs in a situation where rejection of the null hypothesis can be found. It specifically takes place when the null hypothesis is wrongly rejected.So, in case with the type 1 error, the hypothesis is often true but it turns out to be rejected by the researcher (Mitchell and Jolley, 2012, p.147). An example on how the type 1 error can occur is given by an illustration given by McKillup (2011, p.61). The description of a clinical trial is provided. It has been done to test the efficacy of a new drug compared to a different one in treating a particular illness, researcher ends up with a conclusion that there is a difference between the mean scores of the drugs under investigation in the research.
So, in comparison, the overall conclusion in this particular example is that the new drug being under investigation does not work (does not treat well), whereas the result will be opposite in practice and it will show that the drug does treat well. On the other hand, the situation is visa versa with the type 2 error.
Researcher retains the null hypothesis when he or she is actually supposed to reject the null hypothesis. In considering the same example given above and bringing it under perspective, if the type 2 error is made, the researcher may wrongly conclude that the new drug under testing for efficacy in the treatment of a particular illness works. The type 2 error shows that the drug is not efficacious in practice(McKillup, 2011, p.61).
It is generally opined by authors such as Easton and McColl, 1997 and Sullivan (2008, p.29) that the type 1 error is very serious. Consequently, owing to this important consideration, it is important for researchers to avoid the error compared to making the type 2 error. They shall minimize a type 2 error. There is a considered argument that no one need to spend time or commit momentary resources on a treatment that does not work.
Statistical Power Adjustments
As shown in the above examples and being aware of the impacts that either type of error will make on the research, it is logical to assume that adjustments on the statistical power are necessary. The minimizing the type 2 error in the above example would be achieved by increasing the sample size as it is a common practice within the most research studies (Easton and McColl, 1997).
Since the risk involved in controlling one type of an error increases the risk for the other one, it is better to gather enough information while carrying out the study and include it into the research(Smith, 2009).Moreover, it is fully on discretion on researcher’s discretion to determine which error they will give priority in order to minimize the risks.
References
Easton, V.J &McColl.J.H. (1997). Statistics Glossary. Web.
McKillup.S. (2011).Statistics Explained: An Introductory Guide for Life Scientists.(2nd Ed). Cambridge: Cambridge University Press.
Mitchell, M.L & Jolley.J.M. (2012). Research Design Explained. (8th Ed). Wadsworth: Cengage Learning.
Smith.S. (2009). Statistical errors (Type I, Type II, Power). Web.
Sullivan.L.M. (2008).Essentials of Biostatistics in Public Health. Sudbury: Jones & Bartlett Learning.