Science can be defined as the practice of observing the natural world, and trying to make objective sense of it by uncovering facts or cause-and-effect relationships.
The gadgets in this book detect substances and conditions in the atmosphere that otherwise would be invisible to your senses. (Essentially, the gadgets are technological extensions of your senses.) Building them will help hone your skills with DIY electronics and Arduino programming. These are fun, interesting, and practical things to do—but doing them by themselves is not doing science.
Suppose you’d like to learn more from what you uncover. Maybe you’d like to measure atmospheric conditions over days or weeks, and then interpret those readings; or monitor the atmosphere in different parts of your neighborhood, county, or state, and compare that data usefully; or perhaps even organize people around the world to build gas sensors or photometers, and compare findings from these different places in meaningful ways. To do these things, you’ll need to apply some intellectual elbow grease to how you use your gadget. You’ll have to do some science.
The scientific method is the foundation of how most of the serious science in the world gets done. It’s a systematic process of investigation that tests ideas about how cause and effect operate in the natural world, helps to reduce or eliminate bias, and allows the meaningful comparison of information from different sources.
The scientific method is appealingly linear in the abstract: an observation leads to a question, which leads to a hypothesis, which leads to an experiment, which leads to a result, which (if you’re lucky) can lead to another question, and so the process begins again.
Testing hypotheses by gathering evidence is a core concern of science. What most scientists will tell you, though, is that their work tends not to progress as tidily as the scientific method looks on paper. More often they move back and forth between these steps, because science is an iterative process: a repeating process in which the end result is used as a starting point for the next run. Researchers often repeat the same steps over and over in order to test new ideas and tools, to deepen their questions about what they’re studying, and to figure out how to do their research more effectively and accurately.
Researchers also test each other’s hypotheses, because modern science demands that a result be replicable: that different people conducting identical experiments can come up with very similar, even identical results. If an experiment’s results cannot be duplicated independently of the original researcher or team, then those results are cast into doubt.
Still, the scientific method is featured in the early chapters of many a Science 101 textbook because it’s a good jumping-off point for learning how to set up an experiment and collect data.
At their most basic, the steps in the scientific method go like this:
- Observe something in the world.
- Ask a question about it.
- Formulate a potential answer (a hypothesis) for it.
- Conduct an experiment that tests the hypothesis.
Compare the predicted result to the actual result:
- Result supports the hypothesis.
- Result doesn’t support the hypothesis.
- Result partially supports the hypothesis.
- Consider the result.
- Ask another question and begin again.
Let’s look more closely at each step.
Observation and exploration of what’s going on in the environment is essential to figuring out what questions to ask. Asking a question really means "asking an answerable question," one that you can then test with an experiment. Testable questions begin with how, what, when, who, or which ("why" is impossible to answer).
Devising a good experiment question can itself involve several steps. Is your question uninteresting or interesting? Can it be narrowed down to look at a single thing, to collect data on one variable only (an observational question), or to change one variable and learn what results (a manipulative question)? It’s important to test only one variable—a factor that exists at different levels or amounts—at a time in your experiment, in order to be reasonably sure your test and the conclusions you draw from it are valid. If you test more than one variable at a time, then cause and effect relationships are much less clear.
An example of a testable question that would work with one of the gadgets in this book is, "When are atmospheric hydrocarbon levels outside my window at their highest concentration?"
Formulating a hypothesis involves using what you already know to come up with a potential answer to your question: an explanation for what you’ve observed. It’s not merely an educated guess; it is your formal statement of what you’re going to test (the variable) and a prediction of what the results will be.
For example, working off the previous question, you could form your hypothesis statement as, "If heavier car traffic increases atmospheric hydrocarbons outside my window, and I measure those levels from 4 to 6 pm as well as from 4 to 6 am, then I should detect higher levels from 4 to 6 pm, which is afternoon rush hour." The IF statement is your hypothesis; the AND statement is the design of your experiment; the THEN statement is your prediction of what you’ll learn from the experiment. Since what you’re measuring will be atmospheric hydrocarbons, the variable in this experiment is time.
On the face of it, this experiment sounds like a no-brainer: rush hour traffic means more car exhaust means higher levels of hydrocarbons, right? If your results support your hypothesis, then this experiment may be over.
But what if levels are low during both time periods? Those results fail to support your hypothesis. It wouldn’t hurt to check your build and programming, to be sure the gadget is working correctly. Assuming it is, you may need to ask a new question, or broaden the scope of your experiment—such as taking measurements more often during the day, or on different days of the week, or during different types of weather.
And what if you get high hydrocarbon levels from 4 to 6 pm, and also from 4 to 6 am? That’s a partial confirmation of your hypothesis that may lead you to…
Maybe there’s something going on in the world around you that you didn’t know about, like a delivery truck idling on the street early in the morning (we get that a lot on our street!). Do deliveries happen often enough to affect hydrocarbon levels most of the time, or just some of the time? To learn more, you’ll need to figure out the next interesting, testable question, reformulate your hypothesis, and restructure your experiment.
By testing enough variables (time of day, day of week, month, weather conditions, etc.) you should be able to build up a very accurate profile of the hydrocarbon pollution outside your window. In fact, you should be able to predict future events: for example, if tomorrow is a delivery day, you could confidently predict there will be more pollution than usual. Or if tomorrow is going to be rainy, there will be less pollution, as the raindrops wash the pollution out of the atmosphere.
When you feel you’ve gotten the pollution profile for your neighborhood down pat, your work has just begun—now it’s time to measure a different neighborhood! Science never stops!