Many scientists, including me, struggle to ask interesting questions. In part, I think it’s because we are not taught an explicit framework for asking good questions. For example, a simple web search of the “scientific method” turns up dozens of various techniques, all of which are aimed squarely at elementary school students or laypeople.
How does one go from “making observations” to “asking a question?” And what is a good question anyway? What is a hypothesis and how does it differ from a prediction? Do you even need a prediction if it’s not part of the standard scientific method?
Enter the QHPM Scientific Framework….. I came up with this idea as a Biology 101 lab instructor. It came from the observation that critical thinking is a learned skill that can be practiced and improved. I saw that my students were terrible at observing a phenomena and asking questions about it. So, I invented the “QHPM”—a weekly exercise in which students had to ask a Question, generate Hypotheses, form Predictions, and develop a Methodology to examine it.
At first, students were pretty against the idea. Although it was usually under half a page of writing, it was way beyond the scope of anything they’d ever done. How dare I ask them to think for themselves! It was too hard, they said. They weren’t scientists, they said. However, by the end of the semester, most of the students wrote that the QHPM was an important element in their progress as young scientists and critical thinkers.
A casual QHPM requires very little background knowledge or previous observation. However, it’s worth pointing out that the best science requires one to operate at the edge of what’s known. Therefore, you really must be an expert observer to ask the best questions. This requires a lot of reading, thoughtful observation, and careful synthesis. Fortunately, this isn’t necessary for a QHPM. Anyone can do this—even my sixth graders could generate excellent QHPMs worthy of scientific examination!
A QHPM can be informal and short. You can create one in your head while wandering through the woods. Here’s how it works:
Step 1: Ask a Question.
A good question is interesting regardless of the potential outcome. It should be specific and precise. I generally like to ask mechanistic “how” or “why” questions. Broad questions can generally be made more specific by exploiting gradients (temp, elevation, porosity, size, etc.), but it’s not required. For example: How does water temperature during development influence body size of common cuttlefish at 1-week post-hatch?
Here, I ask a very specific question that should be interesting regardless of the outcome. Changing the temperature could lead to smaller or larger body sizes, or potentially result in no change! Understanding the relationship between temperature and development can have implications for climate change, cephalopod husbandry, and ecomorphological evolutionary tradeoffs. It exploits a temperature gradient that will provide good signal. The question targets one species of cuttlefish and examines only two variables (temp and body size at 1 week).
Step 2: Form a Hypothesis.
A hypothesis is a “because” statement grounded in prior knowledge or logic. A hypothesis describes the mechanism by which a phenomenon occurs. A good hypothesis is mutually exclusive. This means that if it accurately describes the phenomenon, then nothing else will. It is possible to have multiple hypotheses, but the best hypotheses do not overlap in terms of their mechanistic explanation. Examples:
H1: Elevated water temperature during development will increase body size of cuttlefish at 1-week post-hatch because higher temps are known to increase metabolism in ectotherms. This one is based on logic. Makes sense, but I think it could be improved by looking at the literature….
H2: Elevated water temperature during development will influence body size at 1-week post-hatch because Palmegiano and D’Apote (1983) found temperature only influenced incubation time in common cuttlefish. This one is based on prior knowledge. Previous research found temperature didn’t make a difference in body size. Is that true? Nothing wrong with re-testing an old finding!
Unlike what is typically taught, I posit here that a hypothesis is not necessarily testable. For example, the big bang hypothesis was untestable when it was developed in the early 1900s. It was, however, based on careful observations of galaxies emitting redshifted light—a telltale sign that they are moving away from the observer—and also on a set of equations (the Friedmann equations) that were based on logic. These equations, by the way, contradicted Einstein’s own theory that the universe was static. It wasn’t until 1964 that the Big Bang hypothesis was supported. Arno Penzias and Robert Wilson discovered cosmic background radiation which demonstrated the entire radiation spectrum was redshifted.
Moral of the story: A good hypothesis does not need to be testable (indeed, the coolest hypotheses likely can’t be tested without advancements in measurement technology). It only needs to be 1) grounded in previous knowledge or 2) really solid logic. And, it should (as much as possible) answer your question.
Step 3: Form Predictions.
Predictions are If/Then statements that must be testable. Predictions provide evidence that supports or refutes your hypothesis. At their simplest level, predictions are graphs. If you change X, then Y will happen. X is your independent variable and Y is your dependent variable. It may be necessary to form multiple predictions to test every aspect of your hypothesis or to provide extra evidence, especially if your hypothesis isn’t specific enough. Embedded in every prediction is the idea that you will keep all of the other variables constant.
Here’s are some examples: If we increase the temperature, then the cuttlefish will be larger at 1-week post-hatch. (But, you may have some others that need to be considered. For example: If we raise the temperature too high, then the cuttlefish will die in the egg.)
For the visual thinkers out there, predictions can be graphed:
Step 4: Methods.
If you intend to do the experiment, how will you test your predictions? What tools/measurement devices will you need? What sampling rates and resolutions and sample sizes will you need? What statistics are necessary to answer your question based on your experimental design? Can you simplify your statistics or improve your statistical power by changing your design? What variables must be controlled?
Generally, there are multiple ways to answer every question. I encourage folks to answer questions in the simplest way possible while using modern technology to their advantage. For example, instead of using a ruler to measure something, take a photo and measure it digitally with far more accuracy and repeatability.
Another important aspect of good science is anticipating the sample size necessary for a robust test that minimizes Type 1 (false-positive) and Type 2 (false-negative) errors. Sometimes it’s difficult to know until you’ve made a few measurements, but it’s always worth thinking about before you finish the experiment.
The QHPM Scientific Framework provides a very powerful yet simple approach to asking and answering scientific questions. It can help young scientists crystallize their thoughts and help seasoned scientists brainstorm quickly. It can help very young students “think like a scientist.” Please consider including it in your science courses at all levels.
Q – Question: A good question is interesting regardless of the potential outcome. It should be specific and precise.
H – Hypothesis: A hypothesis is a “because” statement grounded in prior knowledge or logic. A hypothesis describes the mechanism by which a phenomenon occurs. It is not necessarily testable.
P – Prediction: Predictions are If/Then statements that must be testable. Predictions provide evidence that supports or refutes your hypothesis.
M – Methods: What techniques, tools, statistics, and experimental design will you use to test your predictions?