What often receives less attention is how the research was actually done.
That part matters. A result from a laboratory experiment means something different from a result seen in a human clinical trial. An animal study can provide researchers useful information, but it does not tell us exactly what will happen in people. Even two human studies looking at the same question can come to different conclusions if their methods are different.
Research methods are the tools researchers use to investigate a question. Understanding these methods can help you interpret a study’s findings more accurately, rather than simply accepting the headline result.
Research in the laboratory
Many areas of biomedical research begin in the laboratory.
Researchers might work with individual proteins, cells, tissues or other biological material. These experiments allow them to look closely at what happens under controlled conditions.
For example, a peptide could be introduced to a group of cells to see whether it changes a particular cellular process. Researchers might measure changes in gene activity, cell signalling, inflammation or other biological markers.
This type of work can be useful for finding out whether something is worth investigating further.
There is a limit, though. A group of cells in a laboratory dish is not a human body. The cells do not handle digestion, blood circulation, different organs, or the many other systems that can affect how a substance behaves.
So if a laboratory study reports an intriguing effect, it is better viewed as an early finding rather than evidence that the same thing will happen in people.
Animal research
Animal studies are occasionally employed to advance research.
A living animal provides researchers a more complete biological system to work with. They can study things such as how a compound is absorbed, how long it remains in the body and how different organs respond to it.
Researchers can also use animal studies to identify potential safety issues before investigating a substance more widely in humans.
But there is no perfect animal model for human biology. Mice, rats and other animals have their physiology, metabolism and immune responses. Results can therefore change when research moves from animals to people.
The details of the experiment matter as well. Researchers need to consider the dose used, how the substance was given, how long the animals were followed and what was actually measured.
An animal study can add an important piece to the evidence, but it is still only one piece.
Studies involving people
Human research provides us information that laboratory and animal studies cannot provide on their own.
Clinical studies may look at safety, how the body responds to a substance, or whether it produces a particular change in people with a specific condition. Some studies are tiny and are mainly designed to gather early information. Others involve hundreds or thousands of participants.
It is tempting to assume that a larger study is always better. While size is important, it is not the only factor to consider.
Researchers must also consider participant selection, study duration, comparison group provisions, and assessed outcomes.
A study involving a small group might provide useful early evidence, but the findings may not apply to a much wider population. A larger study may provide a broader picture, particularly when the participants reflect the people who would eventually use the treatment.
Randomised controlled trials
You will often see the term “randomised controlled trial,” or RCT, when reading about clinical research.
In a typical RCT, participants are placed into different groups using a random process. One group may receive the treatment being studied, while another receives a placebo or another form of comparison.
Randomisation is important because people are different. Age, lifestyle, existing health conditions and many other factors can influence an outcome. Random assignment helps reduce the chance that these differences will unfairly affect the results.
Some trials are also blinded. This means that participants, researchers, or both may be unaware of which treatment an individual has received.
RCTs are widely used because this type of design can provide strong evidence about whether an intervention is responsible for an observed effect.
That does not mean every RCT is flawless. Poor study design, a small sample, short follow-up or problems with the way results are analysed can still affect the findings.
Observational studies
Sometimes researchers cannot, or do not need to, assign people to different treatments.
Instead, they observe what happens in the real world.
An observational study might look at whether people with a particular exposure, behaviour or characteristic have different outcomes from those without it. These studies can be useful for spotting patterns and developing new research questions.
There is an important distinction here: a relationship between two things does not necessarily mean one caused the other.
For instance, if two characteristics appear together in a study, another factor may be responsible for the connection. Researchers therefore have to consider possible confounding factors when interpreting observational results.
These studies can still be valuable. They simply answer different types of questions from controlled trials.
Reviews of existing research
Researchers do not always need to start with a new experiment. Occasionally the useful question is what the existing research says when it is considered as a whole.
A systematic review collects relevant studies using a planned method. The researchers then assess the studies and bring the findings together.
Some reviews also include a meta-analysis. The meta-analysis uses statistical methods to combine numerical results from several studies.
Reviews can be beneficial since a single study seldom provides a complete picture. Looking across several studies may show whether findings are broadly consistent or whether there are major differences between them.
There is a catch, though. A review cannot magically fix weak research. If the studies included in it have serious limitations, those limitations still matter.
How to read a research paper
You do not need to be a statistician to ask useful questions about a study.
Start with the basics.
How many participants were involved? Was there a control group? What exactly were researchers measuring? How long were participants followed? Was the study carried out in humans, animals or cells?
It is also worth looking at the actual outcome. A study may report a statistically significant change, but that does not automatically mean the change was large enough to have a meaningful effect in real life.
The limitations section is useful too. Researchers normally discuss weaknesses in their own work, and these can sometimes tell you more than the headline result.
Perhaps most importantly, look beyond one paper when possible. If several well-designed studies have reached similar findings, that presents a different picture from a single small experiment.
Why the method matters
Research usually involves detours from an idea to a definite answer.
A laboratory experiment may suggest that a peptide affects a particular pathway. Animal research may provide additional information about how it behaves in a living system. Human studies can then show whether the findings translate to people.
Occasionally they do. Occasionally, they fail to do so.
That is a normal part of science.
Understanding research methods helps put individual findings into context. It also simplifies the task of distinguishing between an early scientific observation and thoroughly tested evidence.
The aim of this section is not simply to collect research papers. It explains how the papers were made, what they found, and where the evidence is lacking.