The Scientific Method is Messy
Click Icon to View Within The Classroom Artifact
Click Icon to View Beyond The Classroom Artifact
Artifacts: Explained
​
The within the classroom artifact that I chose is a recreation of the notes that I took during BIOL 303, Cell and Molecular Biology Lab. The notes reflect my professor's teachings regarding the scientific method and the lack of linearities that are canonically taught. It demonstrates the moment when I realized the true chaos of the scientific method and the process of experimentation.
​
The beyond the classroom artifact that I chose is an imitation of the excel documents that I used during my research in Sutton Labs. The top row on the first page of the document shows the many criteria that I initially looked for in the articles that I read. It is clear, though, that these criteria were not always listed in the experimental sections of the document. The second page of the document highlights the revised list of descriptors and evidences the more effective data collection that I was able to do with this condensed list. The artifact as a whole documents the idea that the first method you try during research may not be the most effective, and that reevaluations of the methods used can increase the success of a project by yielding better and more informative results.
When learning something new, generalizations can be a useful tool to understand concepts. Sometimes it is easier for initial understanding to ignore exceptions or slight nuances to rules. Just like when learning language rules, scientific rules are often more complicated than they are explained to us in introductory chemistry courses. However, as we become more advanced in our knowledge and understanding, it is important to consider the more complicated ideas that were previously ignored.
​
In my introductory science courses I was taught the steps in the scientific method:
1. Make an observation
2. Pose a question
3. Form a hypothesis
4. Design an experiment
5. Collect data
6. Draw conclusions from the data.
This process seemed to be very simple, straightforward and linear in its progression. The steps were clear and seemingly manageable. However, in my BIOL 303L, Cell and Molecular Biology lab, the veiled simplicity of the scientific method was removed. On the very first day of lecture, the TA explained to the class that in this lab there would be no straightforward, correct answers like there had been for all of the science labs that we had taken in the past. In real life, the scientific method is messier than what we were taught. More often than not, an experimental design will be flawed and, after data collection, the experiment will be re-worked and run again. Or, sometimes the data that you collected will not reveal what was expected and you have to work backwards from the data to form a new hypothesis about what caused the results you got. Most commonly, though, the data will not support a clear conclusion, but instead it opens up a whole new set of questions.
​
Despite what we had been taught, the steps of the scientific method are not clear cut and straightforward. Scientists in the real world need to understand how to adapt those six simple steps to fit the real-life complexities of experimental analysis. Throughout the rest of that semester, the TA taught us to focus more on the big picture than a single data point. How does the experiment we just did fit into the larger series of experiments we performed throughout the semester? What were the flaws in the experiment and how could they be eliminated? If the data was not what you expected, what are some possible reasons why? All of these questions were prevalent in the learning outcomes for this lab and taught me the importance of open-mindedness in scientific exploration.
​
These questions were especially helpful as I began an undergraduate research project where I worked on data mining to build descriptors for a Python program to predict the formation of perovskite crystal structures. The project requires me to read over 200 articles that discuss the formation of different perovskite crystal compounds and to pull out the relevant synthesis information from each one. It seemed like a simple, straightforward task at first, but it proved to be overwhelming. As I began reading articles, I was looking for four descriptors: compound elements, reaction temperature, reaction time, pressure conditions. However, my experimental design was soon proven to be flawed. Each of the articles I read included different information regarding the synthesis reactions that successfully formed perovskiets. Some articles included three-step heating processes for various amounts of time; some articles said nothing about the pressure conditions for the reaction; some articles discussed the starting materials in the reaction and some didn’t.
​
Very quickly, my four selected descriptors were expanded into 10, in an attempt to accommodate the wide scope of information I was finding. Unfortunately, adding more descriptors just led to more holes in my data collection and decreased my ability to make the generalizations required to build a computer program for prediction of perovskite formation. It was at this point that I thought back to my Cell and Molecular Biology lab. I realized that I had been sticking too strictly to my initial experimental design. The data I was collection was not useful or effective, so, rather than collecting more useless data, I needed to go back and re-design my experiment. Using what I had learned from the first round of data collection I created more widely mentioned descriptors and eliminated the least-commonly used ones so that I would not have as many holes in my data.
​
The re-working of my experiment in its infancy has led to a much more efficient and effective data collection method that has yielded the necessary information to start building my program. From both of these experiences, I have learned that it is easy to get tunnel vision when working on an experiment. However, it is critical that scientists continually evaluate their experimental designs and data collections. This method of operation is a departure from the straightforward, linear scientific method that we learned in introductory science. However, the real-life scientific method is messy. It is not linear and straightfroward, and it is also not set in stone. As scientists, we need to understand these ideas to create effective experiments that yield informative data. If we fail to break free of the cannonized version of the scientific method, scientific exploration will halt and discovery will suffer.