When you’re looking at research, how do you know you can trust the quality of scientific evidence on show? Here James gives a quick explainer on the different types of scientific study and what information they give us.
Like most people who rely on the weight of evidence to offer advice and recommendations, we at Elevated spend much time looking at the results of scientific studies and trying to work out what conclusions the boffins have come up with when it comes to their results.
Not all studies are created equal though… some may have smaller sample sizes, others may be prone to author bias, others may have vested interests from those that have funded the study. Here James lists the different types of scientific studies in ascending order (ie lowest first) of based on the strength of the evidence they provide.
Anecdotal evidence and expert opinions
Anecdotal evidence is a person’s own personal experience or view and not necessarily one that is representative of typical experiences. Expert opinion is a stand-alone piece of evidence as stated in, say, a court case or a newspaper article. Both are pretty weak unless they are backed up by appropriate scientific studies.
Experimental animal and cell studies
Animal research can be useful and can predict the effects seen in humans. However, observed effects can differ so human trials are required before an effect can be seen in humans. Tests on isolated human cells may produce different results when then expanded to a section of or the whole human body.
These observational studies are written records on particular subjects. They are low when it comes to hierarchy of evidence but can aid detection of new diseases or side effects of particular treatments.
Case control studies
Case control studies are also observational. The look at two groups of subjects, one with a condition or symptom and one without. They then look back to see what could have caused the condition. These studies may show correlation (a link between A and B) but not causation (A causes B).
Another observational study, similar to the case control study above. In cohort studies, researchers look at a group of people sharing a characteristic or treatment and compares them over time to another group that does not share the characteristic. Any differences in the two groups are noted.
Randomised control trials (RCTs)
These experimental studies randomly assign participants into test groups that receive a treatment and a control group usually receives a placebo. In a ‘blind’ trial participants don’t know what group they are in, in ‘double blind’ trials those conducting the experiment don’t know either! Blinding trials helps remove bias.
Systematic reviews and meta analysis
At the top of the evidence pyramid lie these studies that look at groups of RCTs that have tried to answer the same or similar research questions, critiques the studies for their quality and tries to draw conclusions based on the weight of evidence across the trials. Reviews help mitigate for bias and give us a more complete picture. A meta-analysis is a statistical analysis that combines the results of multiple scientific studies and looks at them from a statistical point of view. It is a cherry on top of a systematic review.