Oh, dear, we have one afternoon a week for about 10 weeks to inspire a small group of three undergraduates to devote their lives to research of some kind. What should we do?
I think back on my own earliest days of independent research at University of Michigan’s Biological Station, near Pellston, Michigan, in my third year of college. My main project made all the mistakes I hope these students do not make. You see, I studied a temporary pond as it thawed from a Michigan winter, then followed it through the summer. I catalogued the species that were in the pond, and described the physical environment, water temperature, turbidity, oxygen content, that kind of thing. I tried to measure both number and kind of species. The study involved daily bike rides through the forest to the pond. It also involved a lot of keying out of insect larvae and other invertebrates. The part I was least good at was begging the technician to run my water samples for micronutrients. My advisor for this project, John Gannon, came down on me hard for my failure to get enough of these samples run.
So, the project was a success in inspiring me to be a scientist. It was a success in a very thorough description of a temporary pond and the stages it went through. This description was used by subsequent generations of Bug Camp students. But as science it was a failure, a kind of failure I hope our students do not have.
You see, I did not have controls. You may wonder what a control for a pond would be and you would be right to wonder. Controls come from a question and my project did not have much of a question. How the pond will change over the spring is not much of a question. Even if I said something specific, like with warmer temperatures, there will be more species diversity, that would not be much of a hypothesis. This is because it would be hard to control for all the other springtime things that come with warming. Also, I would have to have a scientific reason that I thought it.
Not only did I not have controls, I did not have replicates. I could have studied two or three woodland temporary ponds, and seen how they were similar and how they were different. Then I might have had a chance of coming up with a real hypothesis. For example, I could have hypothesized that ponds with big predators, say diving beetle larvae, would have greater species diversity because the beetles kept the common prey low in number, letting rarer things have a chance. But I not only did not have a hypothesis, or replicates. I did not even know that I should have had them. I was young, and either was not well advised at that stage, or did not listen to the advice.
So, we will want these students that will only work 40 hours in total to have a hypothesis, and to have replicates, and to have controls. Isn’t that a lot to ask in such a short time? Don’t we also want them to experience the joy of independent research, of discovering things for themselves? Yes, we want all that and we can get it.
Let’s go back to my pond example to see how we could do it. At some point late in my study, I did become aware of its flaws. I did a little with some other woodland ponds. I found a pond of about the same size as my pond that was in the open, full of sedges, and water plants. The hypothesis I then considered was that the sunlit pond with plants in it would have more insects and invertebrates because there was more food. I could test this by comparing several sunlit, plant-filled ponds with several dark woodland ponds. It would not be a perfect test, because there were other differences, like water temperature. But those problems could be mentioned in the discussion. As an undergraduate project, it would have been much better. If, that is, it also fulfilled the first goal of exciting me about science.
Another thing I could have done is found a lot of ponds of different sizes, but otherwise similar, and hypothesized that larger ponds would have more species diversity. There I would have looked for a relationship between pond size, the independent variable, and species diversity, the response variable. The independent variable gets plotted on the x axis, and the dependent variable gets plotted on the y axis. Another good thing about this kind of hypothesis is I could have made a plot of what I predicted, to compare to what I found. What would the control be here? Well, it would not have a control, exactly. That is because it is a comparative study. In this kind of study you try to keep everything the same except for the variable of interest, not by doing an experiment, but by depending on natural experiments. In this case it would be ponds that were naturally of different sizes. Try to have everything else the same, including the kind of trees overhead, levels of light penetrance, distance from forest edge, even depth of the pond.
Many environmental projects are not experiments, but instead rely on comparison for their power. This is what we hope our students will do. We also hope they will come up with their own questions, from having read a little about our wonderful social amoebae on our lab web page. They will need to have a main variable, with replicates in each condition. For example, they could hypothesize that gram for gram, they will find more social amoebae in samples of dung than in samples of soil because there are more bacteria in dung, and social amoebae eat bacteria. Then they would collect several samples of dung, and several samples of soil, culture out the social amoebae, and compare them.
Just one last note on hypotheses. They are not guesses. They are arrived at after you know something about the system. You think they are going to be true, but await a real test. They have to have a predicted relationship, and a reason for it. Just the prediction is not enough. I’ll have other blogs on this.