While it is clear that workforce issues determine the safety of patients, the intensity of such a correlation remains vague. The reason for that is poor sampling utilized by researchers. A sample must be representative of the whole population because only in this case research findings may apply to it. The excerpt describes one example of bad sampling, which is a convenience sample. As a result, the design of the discussed research study may be considered biased. This means that the study results may favor a particular group of nurses from a certain level of experience or discipline.
However, if researchers had used random sampling, they would have avoided bias by giving all nurses an equal chance to be chosen. Perhaps the easiest way to make a random sample is to create a database of nurses, having given each of the identification numbers. Then, the desired number of participants needs to be randomly chosen by statistical software, such as STATA or SPSS. In such a case, researchers exclude the possibility of favoritism by the sampler and self-selection by nurses.
Nevertheless, in some cases, for example, when researchers aim only to determine if there is a link between workforce issues and patient safety, non-random sampling may be sufficient. However, it has to be taken into account that this will be regarded as a serious study limitation, and further research will be needed to gain more insights into the nature of the relationship. Moreover, bad sampling may lead to low generalizability of findings and the impossibility to apply them to various healthcare settings.