UNDERGRADUATE MENTORING… how to write a hypothesis.

​At this point in the trajectory, I am envisioning (although nothing ever goes according to plan) that we have taught our students how to read the literature and they have been introduced to the amount of field work they will do (if any) and what it looks like. This shows my biases as a field researcher, so for your group it might be more of an introduction to the lab theory (I’ll tackle this later) or modeling theory. Now it’s time for your students to develop their own hypothesis so that they can start testing it! Honestly, this is my favorite part. And to be even more honest, my most embarrassing moment in science is when I couldn’t verbalize the hypothesis guiding my masters work when my advisor asked me. My hope for the world is that all undergraduate researchers get the skills to verbalize hypotheses on the spot no matter what they go on to do. (Photo taken by Daphne Kuta)
 
Step 1: Review of the scientific method (but a realistic review). 
 
Yes, I know…. your students have learned the scientific method in every science class that they have had since elementary school. But, let’s face it – the ideal scientific method as they have learned it (i.e. very linear) is not how science works. The scientific method looks a lot more like a cycle or decision chart that leads you back to the “hypothesis drawing board” over and over again. Consequently, I like to begin by having an honest conversation with my students on how many times your hypothesis is proved incorrect and you go back to the drawing board. You’ll be amazed at how much students will cling to their original hypothesis even if the data doesn’t support it, so it’s best to preempt that by having a conversation about what it’s like to be proven wrong.
 
If you find it helpful, and your students are prepared to do this, you can also employ a discussion on multiple hypotheses (link here). Although I see the absolute merit in this, I think it might be hard to try this with undergraduates and so never have. If you do, let me know. 
 
Step 2: Going from broad to realistic.
 
In my experience, students’ “first try” at a hypothesis is way too broad. They want to solve the world, specifically climate change, and I don’t blame them. That is, after all, how we pitch what we do to others. “I work on how mountains are built.” “In order to better understand anthropogenic climate change, we must do …” This is how we write grants, the introductions to scientific reports, and how we teach classes. So, it’s not surprising that students immediately start here -- but that is now actually the scale of a hypothesis that the students can answer in a single undergraduate thesis. 
 
For this step, I usually take out a grant or paper that I have written and have them read it. I ask them what my hypothesis is. Usually, they point to a large-scale question. They are probably not wrong about what they have identified (if you think like someone who has ever learned science as a linear process that tackles big questions). However, this starts a conversation about scaling down a big-picture problem into testable hypotheses.  In my work, I point to all of the hypothesis that must be proven correct for a geochemical signal to be a proxy. For example (and this would be something I would model), I hypothesize that the laws governing radioactivity are at play to age date this stalagmite, and to make that true, I assume that the stalagmite is a closed system and has a high concentration of U in it. To test this hypothesis, I do U-Th chemistry on stalagmite samples. If the dates come out to be unusable, then the hypothesis is proven wrong. This is a hypothesis of a very small scale, but it is actually testable for an undergraduate. 
 
Step 3: The undergraduate’s practice! 
 
After we have the above conversation, I have students read a paper from the literature that they are already familiar with and try to identify the hypotheses that were tested. In some (well-written) papers, the hypothesis will be very clear. In others, it might have to be gleaned through many iterations of reading the discussion. In either way, the students are asked to articulate the hypotheses and to identify what is a broad-scale hypothesis and what are smaller testable hypotheses the authors were after. We have a class discussion on this work, and I have the students write some of the hypotheses on the board. It truly depends on the students if you need to do this more than once, but practice is always good. 
 
At this point the students should have a since of what scale hypothesis I am trying to get them to formulate. 
 
Step 4: Everyone has a hypothesis!
 
Finally, I return the students to the literature that they have been working on reading. I ask them to formulate a question (based on what they have been introduced to) that they would like to answer. Hopefully, this is something we actually have funding and techniques to answer, but regardless we iterate until it meets the boundaries of the project and their desires. After there is a solid question, I ask them to hypothesize an answer. I am always amazed at how well they do at articulating their hypotheses. The hypothesis in question might need some wordsmithing, but regularly they have a great start. At this point, you would want to have a conversation on whether that hypothesis is something they can test in the time they have or with the equipment/method they have determined they need, but that is all details. Your students have come up with their own hypothesis!

---