You have probably learned the scientific method at some point during your education. It describes the ideal way that science should be done. It goes something like this:
You observe a problem or phenomenon you would like to further understand. These observations lead to a testable hypothesis to gain insights into ‘how’ or ‘why’ or ‘what if’ about your topic. You design an experiment to test your hypothesis and your results clearly tell you whether your hypothesis is correct or not. If not, then you could come up with a new hypothesis to test. Alternatively, based on the new knowledge your previous experiment has given you, you can come up with a new testable hypothesis to take you further down the rabbit hole of scientific discovery and final conclusions. Also, ideally, you present your findings clearly with everyone else so that all of humanity benefits and pushes forward in understanding.
It all seems so rational, so clean, so ripe to produce progress.
However, in the real world, things are messier. Here are the issues at each step that make real-life science less than ideal.
The problem: There is a large, interesting problem, but you are only working on a very small slice of that problem. For any given research topic (like curing cancer or solving the world’s energy crisis), there are many smaller pieces of the puzzle that must be worked out, usually by independent groups. On a daily basis, you feel many steps removed from actually helping, but this is not how you describe your work at parties around non-scientists.
Observations: Your observations are based upon previous research which may or may not be correct for various reasons. Past observations may be incomplete, flawed or, in some unfortunate cases, completely fabricated. This isn’t too insurmountable because the rest of the steps are designed to be self-correcting.
Hypothesis: Sometimes these aren’t as clear as the “If I do this, then I predict this will happen” statements we were taught. Many times they are closer to “If I do this, then the results will tell me more.” Even though you are proceeding logically from past data and observations, your conclusions will turn into your hypotheses by the time it is all written up in a manuscript.
Experiment: You don’t really have the tools to test your hypothesis as directly as you would like. You either have to come up with another hypothesis, develop the tools you need or do the experiment that you can do even though it isn’t the absolute best way of answering your question. Eventually you do an experiment, and get results. Experiments don’t always work, so you spend a lot of time in this phase.
Results: Your results aren’t always what you expected. They probably make you realize that you should have added other controls in your experiment. Controls keep scientists honest by helping to avoid making wrong conclusions on technically flawed experiments.
Conclusions: Finally, you have arrived at the Holy Grail of new understanding. Most of the time is more anti-climactic than you would think. You also have to be careful about your claims because you may not absolutely know “one new thing.” You may have only eliminated one possibility. There could still be a couple of different interpretations of your results. You either have to go back to the beginning or talk about multiple possibilities in your presentation.
Presentation: Volumes can and have been written about this subject alone, and there will definitely be more to come here on this subject. This is the short version: Your science is worthless if you don’t communicate it. This is a moral truth about science as well as a literal one. No really, there’s a memo. “Publish or perish.” Publications are the predominant way scientists are judged. If you don’t publish, your results are worthless. So, you write up a manuscript describing what you did, what your results were and the new truth gleaned from all your hard work. The journal to which you submitted your manuscript sends it out to two or three independent reviewers in your field to evaluate its integrity. In about two to three weeks, your manuscript is returned to you in rejection, filled with reviewer’s suggestions for clearer writing, additional experiments and inclusion of their interpretations in the conclusions section. The severity of their comments determines how far back you have to go in the scientific method before final publication. Eventually, your revised and re-submitted manuscript gets accepted for publication. As uncomfortable as it may have been, the peer-review process did make your article a better one. After momentary celebration, you start back at the beginning of the method again.
I know sounds like I’m just complaining, but that’s not the case. I just want non-scientists to know what is really going on in the background. The human element plays a huge role in the research enterprise, and we can never completely eliminate it or control for it. I’m not sure we’d really want to.
Other caveats: This post deals primarily with basic research like that done at universities, non-profit and governmental labs. There are a whole slew of other issues related to research done in ‘industry’ where companies have a bottom line and actual, useful products to generate.
Stay tuned for the next installment of “Science in real life”… Funding because doing science costs money.