Great discussion! Nice explanations of Type I and Type II errors, understanding sensitivity and specificity when testing the null and alternative hypotheses. I would add that we cannot have 100% confidence in our findings and choosing a 99% confidence interval has other risks such as wasting resources because the sample size has to be so much larger. So yes, we cannot draw conclusions from our research, on conclude that we can make inferences regarding the findings. This is why you see researchers continue to repeat studies to validate their findings. What would you recommend to help researchers achieve the correct sample size with adequate power? this was my discussion!
type I errors occur when the researcher incorrectly concludes that the hypothesis is baseless. This type of error’s probability is represented by an alpha symbol (Brereton, 2020). The alpha symbol is the value that the researcher can choose, and it indicates that 5% of type I errors are acceptable. Therefore, the lower the alpha value, the less likely it is that the researcher will commit this type of error. Type I error is caused by the use of haphazard and inappropriate research methods, which leads to the researcher drawing incorrect conclusions. Type II errors, on the other hand, occur when the researcher incorrectly asserts that the null hypothesis is true. Beta represents the probability of committing a type II error (Brereton, 2020). This type of error can be reduced by ensuring that the sample size and the number of test subjects are sufficient to detect the actual differences. We can therefore conclude that type I errors are worse than type II errors.
These two errors are impossible for the researchers to commit. To avoid type I errors, researchers should only include their findings once they have achieved a high level of confidence (Kim, 2015). A level of confidence of 95% is regarded as optimal. Consequently, this is the rate that researchers should strive to accomplish. In addition, they must perform their tests for a lengthier duration to ensure that a sufficient sample size has been tested, thereby enhancing the credibility of the test results (Kim, 2015). Errors of type II can be reduced by increasing the sample size and decreasing the number of variants. In addition, researchers can reduce the statistical significance threshold to increase statistical power and reduce the likelihood of Type II errors.
References
Kim, H. Y. (2015). Statistical notes for clinical researchers: Type I and type II errors in the statistical decision. Restorative Dentistry & Endodontics, 40(3), 249-252.
Brereton, R. G. (2020). Alpha, beta, type 1 and 2 errors, Ergon Pearson and Jerzy Neyman.