Levels of hierarchies can be useful for assessing the quality of evidence. In health sciences, these are portrayed as a pyramid with levels for the different types of study design.
Understanding study designs will help you judge the limitations of what can be concluded from a particular study.
While in health-related research this hierarchy of study types can be used as a guideline, you cannot rely on the hierarchy to substitute for critical appraisal. A strong cohort study would be more useful than a flawed systematic review. There are times when evidence is better sorted by its usefulness for your own research question than by type of study design.
In food research, there is no consensus around the hierarchy of evidence.* Best practice is to decide, when you plan your search, on the type of research you are looking for and what study designs would be appropriate for it.
For example, if you are interested in qualitative studies of people’s behaviour towards nutrition, or in animal studies, you are unlikely to find large RCTs as it is not ethical or possible to run randomised studies in certain areas, so they may not exist. You are more likely to find qualitative studies that may include survey and interview data, write-ups from focus groups, or other types of studies that exist involving crops or animals (often observational or non-randomised studies). Similarly, you might decide that challenge studies are the most appropriate way to research packaging and shelf life.
A systematic review is a study of studies, where researchers follow a predetermined and published protocol to find all the primary research studies done on a question, weigh the reliability of each one, and, if possible, extract the data from the studies in order to draw a conclusion from the combined evidence.
Randomised control trials (RCTs) randomly allocate participants to either an intervention or a control group, so that conclusions can be drawn about the efficacy of an intervention.
Cohort studies are observational, longitudinal studies that look at a group of people with a shared experience or characteristic to see how they fare over time in regards to a particular factor.
Case-controlled studies are observational studies in which a group of cases (i.e. people with a condition or disease) is compared to an analogous group (similar to the case group except that they don't have the condition) to see if a causal attribute can be found for the case group.
Cross-sectional studies are observational studies that describe a population at a certain point in time. These studies can locate correlations but not causal relationships.
*Reported in Research Synthesis Methods and EFSA Journal.