The possibility and probability degree of errors are common issues in both research design and research assessment. Independently of the project topic, scientists of all fields remain vigilant of the potential variations that may misinterpretations compromise their research results. Certain margins of errors, however, are factored within the statistical analysis itself, while others are a sign of design and sampling errors. The former is unavoidable and inconsequential for the scientific value of the research, operating as a pre-established part of the narrative. The latter has the capacity to compromise an entire project by virtue of affecting all of the collected data by its mere presence. This paper attempts to conceptualize standard errors and sampling errors in research and discuss them within the context of a real-world application.
The standard error is a natural data analysis variation that occurs when the standard deviation is applied to a chosen data sample. It is a mathematical consequence of the analysis and does not compromise the perceived statistical accuracy of the result. Sampling error, however, affects the representative value and the statistical accuracy of the result, indicating the sampling process has been conducted incorrectly. Sampling errors generally occur when a quantitative survey operates as the main research instrument. They are the result of the sub-set of a population having been incorrectly chosen to represent the whole population. Sampling errors are dangerous, as they do not account for other factors of influence on the final research result that is, due to the incorrect sampling, distributed abnormally within the group.
To article COVID-19 pandemic and mental health consequences: Systematic review of the current evidence was chosen to illustrate the concepts of standard and sampling errors. It is related to the largest problem encountered by health sciences within the last century and is therefore directly relevant to many recent and ongoing research projects. This particular article examines the long-term consequences of the COVID-19 virus on the mental health of those who have contracted it. The research question posed asked which, if any, were the long-term effects on the mental health of the COVID-19 patients (Vinegaard & Benros, 2020). An example of a sampling error considering the context could have been a not randomly selected research group, where the number of respondents with a mental disorder is higher than on average. Such a sample would have been a poor representative group, both due to the abnormal variable distribution and the way this abnormal distribution would inevitably affect the results. To elaborate, the COVID-19 patients with pre-existing mental health conditions are likely to experience disorders such as ongoing depression, anxiety, and, in rarer cases, PTSD and OCD independently.
In this case, like in the majority of others, sampling error affects the perception of causation and correlation between factors within the research procedure. Particularly when a research question attempts to study a relationship or a dynamic of influence between variables, the sampling error must be avoided altogether. By allowing it to affect the results of topical medical research, a scientist would not only compromise their professional integrity but publish work with false premises and claims for the general public to interpret. In health sciences, in particular, the proper representative value of the sample must be achieved and maintained, as the research is likely to be entirely discredited otherwise.
References
Vindegaard, N., & Benros, M. (2020). COVID-19 pandemic and mental health consequences: Systematic review of the current evidence.Brain, Behavior, And Immunity, 89, 531-542.