The design of sampling depends on the design of the research I am going to conduct. As I see it, there are two research designs that can be deployed in the context of bike safety. To analyze the risk factors and whether helmets can be regarded as an adequate measure of bike safety, a case-control study would be appropriate. To assess and explore the issue theoretically, a systematic review would suffice.
Generally, systematic reviews have the most theoretical value since they summarize the existing evidence and assess them critically. The approaches deployed when conducting a systematic review facilitate the reduction of possible publication bias and give strong evidence to this or that research issue (Melnyk & Fineout-Overholt, 2015). Some of the challenges that can be faced when conducting a systematic review are data retrieval, evaluating the quality of data, and choosing the most optimal method of analyzing it (DiCenso, Guyatt, & Ciliska, 2014). Despite these challenges, such a method appears relevant for our study. Indeed, there is a significant number of qualitative and quantitative studies on the issue of bike safety which calls for systemization and synthesis of the information to deeper explore the issue. Consequently, a systematic sample is needed to enlist the existing works on the subject of bike safety. On the other hand, a case-control study would contribute to the overall field of research and provide practical evidence for future analyses. In this case, stratified sampling would be needed. The population of adolescents and pre-adolescents registered with a bicycle-related injury would be divided by several strata. The strata would be based on the severity of the injuries, with helmet usage or non-usage taken as variables. In addition, the controls’ gender, age, and extra safety measures such as enhanced bicycle frames, smart lighting, louder horns, etc., can be regarded as variables as well.
Other sampling designs that can be used in a case-control study are random sampling and matching. At that, random sampling would include selecting a certain number of cases from a larger population by assigning them numbers and using a lottery method to include them into study. Its main advantage is that it is the simplest existing sampling method. On the other hand, the one serious flaw in random sampling is that it overlooks the characteristics of each individual case. To effectively assess the direct consequences of the controls being exposed to some factors – risk factors, in this case – matching seems more appropriate. It helps the controls to be conformed according to the variables, thus homogenizing the sample. If the data is unmatched, it is likely that the research will end up with a plethora of minor strata with not enough evidence for each stratum. Albeit the many advantages that matching has, there are several drawbacks to this method. For instance, matching can prove to be costly and take more time than random sampling. Also, matching increases the amount of exclusion criteria which can diminish the sample size (DiCenso et al., 2014).
Overall, on the current stage of the research as serious as that, it seems justifiable to rely on what is feasible at the moment. It appears that a systematic review of existing findings is more valuable than any observational studies, with its findings based entirely on systemized existing evidence, the possibility of applying critical synthesis and assessment skills, and significant value as a platform for future research.
References
DiCenso, A., Guyatt, G., & Ciliska, D. (2014). Evidence-Based Nursing: A Guide to Clinical Practice. St. Louis, MO: Elsevier Health Sciences.
Melnyk, B. M., & Fineout-Overholt, E. (2015). Evidence-based practice in nursing and healthcare: A guide to best practice (3rd ed.). Philadelphia, PA: Lippincott Williams & Wilkins.