The radish experiment doesn't do much good if nobody finds out about it. Obviously, if the radishes were successful in reducing tooth decay, a lot of people might want to eat more radishes so they could benefit from the effect. Perhaps even more important is that other researchers would like to better understand why radishes prevent the decay, so they might start doing studies on the effects of various radish ingredients. This might result in their finding some pill or food additive that would be even more effective in reducing decay, and would be acceptable even to people who dislike radishes. Perhaps it would eventually lead to some better understanding of biological processes that would be useful for treating other health problems as well.
Even if the results did not turn out to show a significant benefit from radishes, it is still useful to have the results published, especially if there has been some prior discussion or controversy about how radishes affect teeth. It might suggest to other researchers that some food other than radishes would be more appropriate to investigate, or it might cast doubt on prior results that implied radishes would be beneficial.
Most research reports start with an abstract, which is a summary of the report, so potential readers can quickly decide if it is of interest to them. The body of the report starts with an introduction that gives background information and discusses previous research in the area.
In experiments involving people or animals, there is usually a short description of the subjects and how they were selected. This is important because the choice of subjects can affect the outcome of the experiment. It is possible that radishes reduce tooth decay for twenty-year-old college students, but not for fifty-year-old construction workers. Further research might be needed to see if and why results apply to certain other groups.
Experimental reports also have a section describing the method. Here the researcher should explain everything about the experiment that is believed to be relevant to achieving the result. We have described the radish experiment fairly simply, but there might be a lot of details that could be important, like the type of radishes used, the average quantity eaten each day, and perhaps the time they were eaten relative to other meals. If some important factor is left out, others who try to repeat the experiment might do something differently which prevents them from getting the same result. The basic knowledge that we hope to get by performing the experiment is of the form "if we perform these actions we expect to get this result". Since the method tells what "these actions" are, it is crucial to understanding the meaning of the experiment.
The next section is the results. This is a summary of the data collected. The raw data in the case of the radish experiment might be the actual number of cavities for each person in each group. The published results would contain things like the average number of cavities per subject in each group, or the number who had no cavities in each group. There would also be reports of problems, such as people who dropped out of the experiment for one reason or another.
Finally the conclusions are reported. These are the interpretations of the experimenters about the meaning of the results. While the conclusions are worth considering, if the subject is a serious concern to us, we should consider the results for ourselves and depend as little as possible on the experimenter's opinions. The results section is where the real facts are.
In a study which made news headlines in 1989, aspirin taken regularly was shown to substantially reduce heart attacks. It was some time later that I ran into an article that mentioned an interesting fact that might have shown up in the "results" section of this report but not the conclusions: the number of people who died in the group who took the aspirin was almost exactly the same as the number in the group who did not! Although the aspirin was effective in reducing heart attacks it apparently wasn't effective in reducing the death rate! It may have had bad effects that slightly increased mortality from other sources. Most of the news stories said doctors didn't recommend taking the aspirin unless people had already had heart trouble. This would make sense since these people would have a larger than average chance of being helped by the aspirin but probably only an average chance of being hurt by it.
Experimental reports are normally published in peer-reviewed technical journals. Being peer-reviewed means the articles are sent to experts in the field to be reviewed, and only the articles receiving sufficiently favorable reviews are published. Once published, others around the world with interest in the results have access to the knowledge provided by the experiment.