Science, Observation, and Experiment
01/10/99
The pages that follow attempt to explain a course, Environmental Biology, and a way of knowing, science. They will demand some attention, and they are deliberately not designed to attract or even retain your attention. The surf is down here. It's a quiet backwater where you will have to expend some energy of your own to travel. Anyone can click a mouse, not everyone can think. By the way, you probably don't want to try this all in one chunk, either. Take it a bit at a time. Think about what I'm saying. You don't have to agree with me; like a lot of things on the internet, what I am putting into words here may be trash or treasure. The knowledgeable people of the future - your future - will be the ones who can find and analyze information quickly. They will also have to be able to distinguish the wheat from the chaff.
I will try to use the hypertext capabilities of the web sparingly, because much of what I have to say is, in fact, linear.
Sequences are how we attempt to deliver "general education" at Marietta. We, the faculty, believe that a liberally educated person has to have some faculty (and note that I use the word in two of its related meanings here) in the various areas of the liberal arts. At Marietta, that means the humanities, the social sciences, the fine arts, and the physical sciences. We use sequential courses because we think it's important to pursue these studies in some depth. With two unconnected courses, much time is spent merely acquainting you with the ways of thinking in the field. In a sequence course, that way of thinking can be built in throughout the year. And, believe me, there are profound differences. Let me give you an example.
To most people, science, technology, and engineering are the same thing, so it might surprise you to learn, for instance, that scientists and engineers often have very different ways of approaching a problem. Now, first of all, realize that I'm going to use some stereotypes here, and put them on with a broad brush, and exaggerate the differences. Human minds can't handle continua - where one thing merges imperceptibly into another - as well as we can discrete phenomena, and that's where stereotypes are useful. They can be used to teach general concepts, as long as we're careful to remember that things are not always so simple. To my story:
I am married to an engineer. She has had a lot of experience with many of the really exciting things that have happened in space since Apollo. She's worked at JPL with the teams that brought us Voyager, Magellan, and Galileo. She's worked with the engineers that built Apollo on improvements to the space shuttle. And, she's currently working on the space station. So, she's a competent, high-tech sort of engineer. Her dad is an engineer. Her grandfather was an engineer. Her brother's an engineer. Her mom's a biochemist. Go figure. Anyway, she has her unique, engineer way of doing things, and I have my scientific way of doing things. Actually, the two are very close together in that both involve observation and experimentation; what really differs is the ritualized form in which an engineer works. A perfect example occurred several years ago. We were visiting her father's office at Purdue University, and her father wanted to hook up a computer monitor to his new portable computer. Father and daughter examined the connections carefully to determine the exact type and size of screwdriver needed, then spent the next 15 minutes ransacking his colleagues' offices looking for the appropriate screwdriver (being department chair has its advantages). Meanwhile, I found a small paper clip about the right size, and made the connection. I was playing Solitaire by the time they got back. Now, my way wasn't the right way, necessarily. Trial and error means, well, error. Engineer's work with other people's money, materials, and lives. They can't afford to be wrong very often, and they go to great lengths to do things right the first time. Doctors are the same way. Academic scientists, on the other hand, have a lot more freedom to simply try new approaches. Of course, neither side is an absolute. When I am collecting insects, I want to be damn sure that I have a real purpose and that the specimens are going to serve some useful purpose. Otherwise I'm just killing bugs, and that bothers me, morally. So I try to do things right the first time. I don't get too upset if I'm wrong, but I try to do things right. And, my wife experiments as well. Right now they're using computers to predict what will happen with electrical levels once they start adding on to the space station. Better to experiment with electrons on silicon than to find you have a problem with billions of dollars (and rubles) worth of hardware in space - and maybe a few lives as well. We don't want to get too complacent with our way of thinking, however, lest we make mistakes because of it. My father-in-law and my wife made a funny, but time-consuming mistake when they failed to realize that a tool other than the one designed for the purpose would suffice for a simple task. Many of my experiments failed when I failed to take into account variables (which I then didn't control for) because I rushed into the experiment without thinking. If engineers and scientists have different ways of approaching problems, is it any wonder that poets and scientists have trouble communicating?
Back to sequences. In the general biology sequence - or the environmental biology sequence - one of our goals, perhaps the most important one, is to get you to be able to think like a scientist. Not that you have to spend the rest of your life doing so, but just so you can if you want to. Of course, if you go on to a career in science, this will be of great utility, but, if not, it will be part of your well-rounded education and contribute to your ability to understand science, use it in your daily life, and communicate with scientists.
What does it mean, to think like a scientist? Science is one way to attempt to understand the universe. Scientists, of course, mostly think that it is the best way to understand the universe, but people with other ways of looking at the universe - painters, poets, priests, etc. - may differ. Science begins by making observations of the universe, then attempting to explain these observations. Most every system of looking at the universe does this as well. The difference is in the next step. Scientists then check their explanations by making sure they work for all such observations. If they don't - and this is the important part - they chuck out the old explanation and come up with a new, better explanation that explains all of the observations. These explanations, by the way, have various names. Untested, we call them hypotheses. Once verified, we call them theories, but we always hold open the possibility that future observations will not fit the theory. At that point, we either modify and re-test the theory or we throw it out and start over. A good example is our understanding of the universe. Newton, who worked in the 1600's, had a perfectly good model of the universe. He saw it essentially as a huge billiard table, and demonstrated how the positions of the planets could be determined mathematically. When one can use mathematics to predict future events, it is a good sign that your theory, or model, is accurate. Newton's model did not work, however, when one examined the universe at the molecular level. When scientists began to do that - when they looked at the atomic level for how light is produced or investigated the nature of electricity - Newton's model no longer worked. Einstein and other physicists came up with new theories to explain the atomic world, and, as it turned out, their model was more accurate than Newton's when applied to larger situations as well. One model of the universe replaced another.
You're probably wondering about experiment. Actually, you can be a scientist without ever doing an experiment. Experiments are formalized ways of making observations. In an experiment, the scientist manipulates the situation to create the best possible conditions for making the most important observations.
Let's consider a possible experiment. You want to see if acid rain might have an effect on tomato plants. You start with a premise, or hypothesis, that acid will slow or stop the growth of the plants. You and a friend each plant a few tomato plants in your backyards, and you measure them daily. You both add acid to the water you use to water the plants. After a few weeks, you notice that the plants in your yard are much smaller. What can you conclude? That acid had an effect?
A scientist would have many problems with your experiment - and I hope you spotted them too. First, you can make no conclusion about the effect of acid. A good scientist would note that both sets of plants were treated with acid, so the fact that yours were stunted means nothing in terms of the acid. A good explanation of the observation (that your plants were shorter) would have to turn to differences in the ways the plants were treated at your house vs. your friends. Maybe one yard had more sun. Better soil. Maybe you watered in the afternoon and your friend in the morning. Did you use seeds from the same source? How much of a difference was there, anyway? Was it significant (scientists have a precise way of defining significance based on statistics) or was it just due to a random chance? Finding out the reason for the differences may lead you into lots of exciting hypotheses, but it will not address the question you originally asked about acid. That was the one thing that both plants had in common, so any difference in growth could not have been due to acid. One final thing. You treated all the plants with acid. What plants were you planning to compare your results to? All of the things that can affect the outcome of the experiment are called variables. A good experiment eliminates all but one variable, and that variable is the thing that we are testing. Usually, a good experiment will have one situation where the variable is different from normal, and one situation where the variable has not been altered from normal. This latter situation is known as the control, and serves as a basis for comparison.
Let's set up the experiment again, eliminating all variables except the one we want to test, acidity. To eliminate variables associated with different weather at various places, we would want to do the experiment in one place. Ideally it would be carried out by only a few people, each of whom was using a written plan or protocol to eliminate differences in watering, weeding, etc. Ideally, the experiment might be done in a greenhouse, where other problems (variables) such as insects, neighborhood dogs, hailstorms, etc. can be controlled to a greater extent. Obviously, all of the plants would be in the same potting soil, in similar containers, etc. It is even a good idea to alternate the control and test plants on the greenhouse bench to ensure that differences in position on the bench don't affect the final results. We would also want to be sure that all the seeds came from the same bag, and we would take steps to randomly assign seeds from the bags to either of the two treatments (control or experimental). We would use as many plants as practical. This, in the long run, ensures that the results are not too adversely affected by chance. Suppose you use only two plants, and a bug bites off the top of one while it was growing. It would end up shorter, and, if you didn't see the damage, you might assume the difference was due to the acid. If you had many plants, the chances of this happening - and happening only in one of the two treatments - would be greatly reduced.
O.K. how are you going to apply the acid? This is the only thing you want to vary between the plants, so you have to think about this. If you add the acid in a dry form how will that affect the results? If you add the acid in water, will the results you see be from the acid or the extra water the one set of plants receive when you add the acid water? How might you balance this? Where are you going to add the acid? Should it be applied to the whole plant or just one part?
Think about this before reading on.
Some of these questions can be answered by considering what you are trying to prove. I would imagine that students trying this experiment are interested in acid rain. Therefore, they might want to apply the acid in a way that simulates natural rain, perhaps by using a misting bottle. To control for the effects of the water in the mist, the control plants could be misted with plain water, and the experimental plants misted with water into which acid had been introduced. A very effective experiment could be carried out by varying the amount of acid. This would involve setting up several experimental groups; each one would get a higher dose of the acid than the preceding group.
Now a word about bias. When you set up an experiment to test a hypothesis, you have some idea of the results in advance - or, at least you think you do. Everyone is subject to subconscious desires, and, usually, these desires want our hypothesis to be right. If you are not very careful, you might end up affecting the results in some subtle way. You might, for instance, not stretch the tape measure as tightly on the control plants, making them seem slightly bigger. Or, you might round down the measurements on the experimental plants and round up the numbers on the controls. You probably wouldn't even notice yourself doing this. This is called bias, and it must be controlled in an experiment as well.
Scientists must be scrupulously honest. It isn't enough to simply avoid lying; a scientist must go further and be sure she isn't lying to herself, even unconsciously. One way to prevent such bias from entering into an experiment is to conduct a blind experiment. In such an experiment, the person doing the measuring does not know which group is the experimental and which is the control. In our experiment, it would work like this. One of the friends would label all the plants. Maybe he would number the plants from 1 to 50. He would spray the plants each day with a mister, using a mister with acid for the odd plants and a mister with plain water for the even ones. The other partner would be unaware of this scheme. She would measure the plants daily and record the results. Unless the acid had an odor or left a residue, the person doing the measuring would be unaware of which plants were being treated and would thus be unable to affect the results unconsciously (conscious deception is another matter - it's called fraud). After the final measurements are made, both partners can analyze the results. In some experiments, the blind is even kept on until the results are analyzed; sometimes the results are sent to an independent person for this analysis. In medical experiments with humans, often both the researcher and the patient are kept in the dark about whether the patient is receiving the experimental treatment or the control placebo. This is then known as a double-blind experiment. Why would such an elaborate scheme be needed?
A final lesson about science and experiments. Suppose the experiment showed that there was no difference between the plants treated with acid and those which were not treated. Was the experiment a waste? Certainly not. It suggests (we don't usually say a single experiment proves anything) that acid may not have a direct effect on the plants. That is useful knowledge. It also raises a number of interesting questions. Might acid have effects in other situations? In conjunction with other variables such as soil type? Only outdoors in natural sunlight? Only for certain plants? If so (for any of these) why? What is the mechanism? If not, how do plants protect themselves against the effects of acids? A good experiment often leaves us with more questions than we started with.
In the final analysis, if an experiment is properly designed and carried out, there is no such thing as "bad results". Even results that disprove our hypotheses move us in new directions. It is the willingness to pursue leads in all directions, the ability to abandon long-held beliefs in the light of new, contradictory evidence, and the constant pursuit of verification of our ideas that helps to separate science from other ways of looking at the universe.
As an assignment, I want you to design an experiment to test the hypothesis that increased ultraviolet light decreases the hatching rate of bullfrog eggs. Write a short description of your proposed experiment, including provisions for replication and controls. Email your proposal to your instructor. Some instructors will give bonus points to anyone who designs a satisfactory experiment and extra points to the best (yet still practical) experiment. Check with your instructor for details.
Environmental Biology Top Page About Sequences About Ecosystems
Return to McShaffrey Home Page