Much of what we know about endurance sports originates as scientific research. This research is performed by professors at universities and published in journals. When a journal article publishes, one of the keywords or the title may pique the interest of a journalist. If the journalist has a technical background, he or she will summarize the finding in a news article in the popular press. And we’ve seen many of these headlines: “Coffee is shown to reduce dementia, study shows” or “Ultraendurance sports may not be good for your heart, scientists say”
On more than one occasion, I’ve read the news article and then looked up the original publication to read the results for myself. When you read the original research, there are often nuances or subtleties not considered in the news article. Or the main point of the research was much different than the journalists report (the result may have been an ancillary result).
But first, where does one get access to scientific publications. Most scientific publications can be found on the PubMed site. Here you can get the basics of the journal article: how the study was conducted (method), what results were seen, and any conclusions drawn from the results. Depending on the publisher and the how long ago the paper was published, you may be able to get the article for free. Otherwise, you have to pay, usually a ridiculous amount. If you live near a university, often you can access these articles from their library (for those how believe all information should be free, you may be able to get the pirated article from a Russian site, Sci-Hub.) .
A scientific research paper is broken down into several sections. The Introduction usually reviews the previously published research in this area, a great way to get the lay of the land. This section usually ends with how the current research either fills a gap in the current knowledge or extends previous knowledge. This is usually framed in terms of a hypothesis: “We hypothesize that running three times a week will lower blood pressure compared to no running”. The Methods section of the paper describes the subjects used to conduct the study (“non-smoking females between 19 and 45) and the methods to conduct the study (“Subjects ran on a treadmill for 30 minutes at 60% of VO2max during each session. There were three sessions a week: Monday, Thursday, and Saturday”). The Results section shows the results in table or graphical form and any statistically significant difference between the groups (“The average systolic pressure for the running group dropped 20% compared to the sedentary group”). The Discussion section (sometimes combined with the Results section) discusses the relevance of the results and attempts to explain the results relative to previously published research (“The reduction in blood pressure could be attributed to changes in vascular enlargement and flexibility”).
In looking at scientific research, either at a high level or digging into the actual paper, a number of questions should be asked:
What is being compared? The crux of scientific research in the area of human physiology is the comparison of one or more groups. As a minimum, there is the experimental group, the group that does something (takes the vitamin, does an exercise, is exposed to heat, etc.) and the control group that doesn’t do something (receives a starch pill vs the vitamin, is sedentary, exposed to normal temperature). Often there may be two experimental groups that are compared (high mileage runners vs low mileage runners) or a control and several experimental groups (starch pill, low dose of vitamin, high dose of vitamin). But at the end of the day, a comparison is being made to either prove or disprove the hypothesis (e.g. does running have an effect on blood pressure)*.
Who are the subjects? Each of the groups being compared are composed of individual subjects. These are either humans or laboratory animals. While dogs or rats might be better controlled than “free-range” humans, the conclusions from the research may be difficult to apply to humans. For human subjects, it’s important to know the nature of the subjects. Sedentary smokers are going to have a different response to diet or exercise compared to elite marathoners. Unfortunately, many exercise science research groups are located in universities with access to young people who are willing to and have time to volunteer for studies. Thus, not as many studies are done with mid- to late-life adults or children.
How many subjects? The number of subjects (the sample size) and how those subjects are divided among the groups (experimental and control) is important to be able to show statistically significant comparisons between groups. A small sample size may not be able to show the effect or the effect shown may not be applicable to the larger population.
Retrospective or prospective study? Often in science, time is a variable being measured. Did running barefoot for 12 weeks have an effect on injury rates? In a retrospective study, scientists look back at data and draw conclusions. This often involves examining data from another study to draw conclusions. A prospective study selects subject for the study and follows the subjects (the cohort) over a period of time. In general, prospective is considered superior because of less bias and confounding variables.
Original research or meta-analysis? Although most research is done as an original research study, occasionally a publication will analyze the results from a number of similar studies. This meta-analysis can be useful to increase the number of subjects being studied to show an affect. The challenge with a meta-analysis is making sure the experimental procedures were similar enough to allow pooling and comparison of data from the different studies.
Causation vs correlation. When making comparisons between two groups, scientists are looking first for correlation. For example, is there a correlation between exercise and blood pressure? The research is designed to show the correlation (e.g. increasing exercise is correlated with a lowering of blood pressure). Where things get tricky is extending the correlation to a causation (e.g. exercise causes a decrease in blood pressure). In many cases, correlation implies causation but not always. Sometimes there is a confounding or third variable that may contribute to the correlation. For example, the sales of ice cream correlates with drowning deaths so therefore ice cream causes drowning? The third variable here is warm weather; both ice cream sales and drowning deaths are correlated with warm weather.
Date of publication. Science is built on science so the most recent publication usually means it is building on the previous knowledge. What’s too old? It depends on how fast the field is moving but usually anything over ten years is certainly becoming outdated in the field of exercise physiology and sports medicine.
*On a technical level, these are called independent and dependent variables. The independent variable is the one being set or varied by the study (e.g. sedentary vs running) whereas the dependent variable results from or depends on the independent variable (e.g. blood pressure)