Designing an Experiment

             Experiment design is the process of systematically planning, executing, analyzing, and recording an experiment.  This process has many benefits ranging from making me much faster at running experiments to giving me the credibility I need on the job, in the laboratory, etc.  These benefits include things like helping me acquire accurate data (as apposed to theoretically driven math models), making me more confident in my data, giving me an easy way to openly share and explain my work to others, and providing my with the best way to learn engineering, i.e. using hands-on work to analyze and process real-world data. 

            How to actually “do” experiment design can be explained in several steps.  The first step is to find the “Why?” i.e. for what reason(s) am I doing this experiment.  This reason can range from validating a math model, to testing a design.  The second step is to use Knowledge Construction.  As its name indicates, Knowledge Construction is the process of acquiring information, specifically in looking for the best sources of knowledge available, understanding the knowledge, and applying it to my life.  Although not always used, step three is to make predictions on what will happen.  The fourth Step is the detailed process of figuring out exactly how I want to run my experiment and documenting my work.  When applying step four there are a few things I am looking for.  Initially, I want to set up and define the goals of my experiment (i.e. what am I measuring and how much data do I need?), then I need to generate ideas on how to measure my data, what can affect the results of my data, how will I calibrate my instruments, etc.  After I have a list of ideas, I need to select the best ones, make a step-by-step plan to execute them, and get my equipment ready for the experiment.  At this point we reach the fun part of experiment design.  In step five get to I execute the step-by-step plan created in the fourth step, process the data collected, and document the experiment in real-time.  Following this, I review my process and results and compare my findings with my original theory.  This is also a good time to go back and dig around some more in the literature.  Finally, step six is where I can loop back and run the process again, doing an iterative process to improve results, learning, and speed.

            A person can follow the correct experiment design procedure and still have skewed results simply because of their belief system, however.  A person who says he believes in science and data, but only uses what supports his beliefs and rejects the evidence that goes against them has an Alchemy belief system.  This is common simply because we like to be right.  In fact, it is a biological trait that, when we are challenged, our brains naturally want to kick into a “fight of flight” mode.  In contrast with Alchemy, people who hold the Scientific belief system believe that data and evidence will indicate what is most likely true.  Those who hold this latter view have to be skeptical of their own views, realizing that their interpretations of the data may not be correct.  This Scientific belief system is rare simply because of the natural tendency we have as human being to want to be right, whether or not it is supported by data.  Endorsing this rarer viewpoint, however, carries with it the benefits of better and quicker learning, and credibility.

            Here are some real-world examples of using this scientific method.  When learning about transistors and diodes if I can write down the fundamentals of how they work using semiconductors from memory or explain them to others I have clear evidence that I am grasping the concept.  If a thermometer I am designing using an Arduino and a thermocouple begins recording skewed temperatures I have data/evidence that my thermometer is not working correctly.  If a math model I am creating spits out values that are validated by experiments, I have data/evidence that my math model is most likely correct.