When choosing a research methodology, one word should always come to mind, feasible. The feasibility of what you can and cannot do given time, money, resources, and collaborators must be taken into consideration before conducting a study. Researchers that have limited amounts of the aforementioned may be better served by retrospective observational designs where data on predictors and outcomes already exists. Prospective and experimental designs require much more time and effort to conduct. A significantly larger amount of empirical complexity and experience is needed to conduct these types of designs. There must also have to be sufficient time to follow-up on the outcomes of interest.
The PICO (population, intervention, comparator, and outcome) mnemonic is an excellent tool for defining important parts of a research methodology. The population should be defined in regards to inclusion and exclusion criteria. In order for studies and experiments to be replicated, the intervention or treatment must be explicitly described. If the goal of a research study is to show evidence of a treatment effect, then a comparison, control, or comparator group is necessary. Comparator participants should possess similar demographic and clinical characteristics to treatment participants to truly understand any associations or effects. Finally, the primary outcome should be measured at the current "gold standard" level to increase the precision and accuracy of research findings. The "gold standard" outcome is also more generalizable and understood by clinicians because it is part of their lexicon and cognitive schema.