Ask the experts: How science works

With stories on nutrition science hitting the media every week, it can be tricky deciphering which ones are legitimate. HFG editor Jenny de Montalk takes a look at how science works — and how to interpret science stories.

These days we’re bombarded with information about nutrition. It seems every five minutes there’s a new study published. Bloggers and other media are all vying for our attention with eye-catching headlines and promises of miraculous cure-alls.

With little time in our busy lives to properly scrutinise every health story we come across, making healthy choices can become confusing. But we’re here to help. Science is the backbone of all the content Healthy Food Guide produces. It’s our mission to provide nutrition-related information based on the body of established scientific evidence, and all our recipes are checked by our nutritionists.

We do this to save you time and provide peace of mind. But why is science so important and, for that matter, how does it work?

What is science?

Science is a system of finding out things about the world using observation and experimentation. And nutrition science, according to University of Auckland Head of Nutrition and Dietetics Clare Wall, looks at how the food we eat can affect how healthy we are. Science also describes the body of evidence so far revealed using scientific method.

How do we find our answers?

Researchers follow steps called the scientific method to come to conclusions about specific questions. According to the Science Media Centre’s Desk Guide for Covering Science, the steps look something like this:

How does nutrition science work?

In nutrition science, hypotheses are largely formed from population studies of diet and its association with a specific outcome, such as a disease, Dr Wall says. A hypothesis is like an educated guess that is the starting point for further investigation.

An example is the hypothesis: Higher intake of saturated fat increases the risk of cardiovascular disease (CVD). “It was observed in the 1970s that people who lived in southern Mediterranean countries had less CVD compared with those in Northern European countries,” she explains. There was also a big difference between the two populations in how much saturated fat they ate.

Researchers analysed the information on deaths from CVD and found there was a strong association with an increasing intake of saturated fats and CVD deaths. But, Dr Wall says, “This information can only demonstrate an association, not a definite link, because there can be many other differences between these two populations which could cause the observed difference in CVD.”

The hypothesis needs to be tested.

Putting a theory to the test

Testing can be done by conducting a randomised controlled trial. This is where you randomly assign people participating in a study into groups: those who receive an intervention, such as a medicine, and those in a control group who receive a placebo, which has no active ingredient.

When researchers and participants don’t know who is in the experimental group and who is receiving the placebo this is called a double-blind experiment. Double-blinding is the best way to reduce or eliminate bias by researchers looking for a particular result.

Challenges for nutrition science

Creating a control group for nutrition research can be difficult because people don’t just eat one type of food. “When you’re looking at altering components of the diet it becomes very complicated because you also have to control all other food components to make sure the diets are equally balanced. Easy to do in laboratory animals, but not in free living humans,” Dr Wall says.

Another challenge nutrition researchers face is that diseases associated with diet take a long time to occur, meaning studies need to be conducted over years to establish disease relationships. “Using biomarkers (such as nutrient levels in biological samples, e.g. blood and urine) can help overcome some of these problems but, currently, we do not have many reliable ones to choose from.”

What comes after the experiment?

When the trial is complete all the data are analysed and the researchers draw their conclusions and decide whether the experiment supports the hypothesis or not. These preliminary results are put up for review by the researchers’ peers.

University of Auckland Discipline of Nutrition and Dietetics professor Lynn Ferguson says most often a paper is sent off to three independent referees, within a given field, who are not collaborators in the trial.

Peer review is extremely important and it is a process that can last months, according to the Desk Guide for Covering Science. “The study is sent to scientists working in the same field, who are best positioned to decide whether the methods used were appropriate and the conclusions make sense. These ‘peer reviewers’ offer journal editors advice on the quality of the paper, whether or not it should be published and what changes should be made if it is to be published,” the guide says.

Where to next?

A study that has passed peer review and is published in a reputable journal continues to be evaluated. “Once published, a study may receive further critique from other scientists through letters to the editor of the journal, commentary articles or further research attempting to replicate the finding of the original study — science is an ongoing process,” the guide explains.

The importance of transparency

Scientists don’t just document the results of their experiments. They record every aspect of their methodology for complete transparency. “Sometimes we learn more from the actual research process than the outcomes,” Dr Wall says. “It also helps us to improve experimental design.”

Documenting methodology properly will help other scientists trying to replicate the results of an experiment. They will need to conduct the experiment in the same way the second time around to see if the results are the same.

Why scientists repeat research

Replication is important because the results of a study of one group of participants won’t automatically apply to all groups.

For example, if a study was done using 65-year-old men with heart disease, the results may not apply to women in their thirties.

How to pick a fact from a fad

Most people won’t hear about a study until it’s reported by the media. The Science Media Centre says the ideal time for journalists to write about research is after it has been peer reviewed and published. Be wary of stories reporting on results that haven’t yet been through this stage.

One of the first things to look for in a story about research is how big the study is. Were there just a few people studied, or were there thousands? If there were only 10 participants, the results might be interesting but further investigation is definitely needed.

Also, how long did the study last? As Dr Wall points out, a lot of nutrition research into dietary links to disease needs to happen over a long period of time.

Look for who is behind the study. Ideally, research will be carried out at a reputable institution. Sometimes big businesses fund health research. This doesn’t automatically mean there’s something wrong with the research, but those interests should be declared in the story so you know who paid for the study.

Another thing to consider is whether the study was done on humans or animals. Again, if the results are from an animal trial, they might be interesting but there needs to be more investigation done using humans.

Additionally, who were the people in the study? Was the population group similar to the one you’re part of? If you’re a young woman and the study’s participants were all men in their seventies, the results potentially don’t apply.

Finally, does the story say where the study fits in the wider body of scientific evidence? A good article should say whether the study’s results reinforce or conflict with previous studies. Headlines such as “Everything we thought about X is wrong” are rarely accurate.

The more we study nutrition, the better our scientific techniques become. This leads to more knowledge, which can appear to contradict previous findings. Also, our food and physical environment have changed significantly in the past 20 years, which changes how diet impacts on our health.

Science is an ongoing process. One study’s findings will add to a growing body of evidence but are rarely the answer on their own.

Good science story checklist

Whether you’re reading stories about science online or seeing them on TV, it pays to ask yourself a few key questions:

• Is the research being reported as a ‘breakthrough’, implying that everything that’s come before is wrong? Be wary; this is not how science typically works.
• Is the research being reported as ‘x causes y’? This might not be accurate. Look for words like ‘associated with’. This means there’s a correlation between two things, but the study doesn’t prove that one thing causes the other.
• Does the report tell you who conducted the research and how? (See How to pick a fact from a fad, above.) This is important to help understand how relevant the research might be to you.