Public trust in science has taken a beating during the pandemic, and experts argue helping students understand natural uncertainty in science could help restore it.
鈥淲hen scientific findings change, the perception by the public, understandably, is often that something went wrong鈥攚hen in fact that鈥檚 intrinsic to how science progresses,鈥� said Joshua Rosenberg, an assistant professor of STEM education and faculty fellow at the Center for Enhancing Education in Mathematics and Sciences at the University of Tennessee. 鈥淲e have evidence that鈥檚 inherently uncertain, and we weigh that evidence in light of what we already know, and we sort of update how confident we are over time. But that鈥檚 often not how science is communicated. It鈥檚 often not how science is learned.鈥�
For example, when the SARS-COV-2 virus first sparked a global pandemic in 2019, scientists scrambled to understand how the virus spread, mutated, and affected different groups of people. Research findings and public health recommendations based on them changed over time as studies looked a larger groups of people in different areas and situations.
Many grew frustrated and confused by what seemed to be conflicting findings. The share of U.S. adults who expressed some confidence in scientists fell from , according to the Pew Research Center. Only 29 percent reported having 鈥済reat confidence鈥� in the field.
鈥淲hen we teach science as a collection of facts, it鈥檚 easy to think about those facts as fixed,鈥� Rosenberg said. 鈥淲hereas, if we teach science as a way of figuring out how the world works, then it鈥檚 much easier for students to see that that generates things that we can count on, but that it鈥檚 also noting when we learn new things that change what we know.鈥�
In a , Rosenberg and Marcus Kubsch, a physics educator at the Leibniz Institute for Science and Mathematics Education in Kiel, Germany, argue that students need more exposure to concepts of subjective probability and uncertainty in earlier grades.
Alex Edwards, a 6-8 grade science teacher at the independent Tate鈥檚 School in Knoxville, Tenn., said mistrust in science among his students in recent years has become 鈥渞eally tough,鈥� and they often struggle to understand why findings should change over time or how confirmation bias can develop.
The way science curricula scaffold lessons can instill a mistrust in the subject if students don鈥檛 accept degrees of uncertainty, Edwards said. 鈥淲e teach little chunks at a time. We teach stuff that鈥檚 not necessarily exactly right but understandable, so that we can go back later and teach it more.鈥�
For example, students may learn in the early grades that Earth is a sphere rather than flat, and then later learn that the planet鈥檚 rotation makes it an oblate spheroid rather than a perfect ball. 鈥淭hat is a better explanation, but it鈥檚 a little harder to explain [to young students] than, 鈥榯he world is a sphere.鈥� The world as a sphere is than the world being flat,鈥� said Edwards. 鈥淏ut people can get the idea that if something is a little wrong, then it鈥檚 all wrong.鈥�
Instead, Rosenberg and his colleagues argue that science teachers need to help students understand variation, probability, and uncertainty as part of the normal process of science. While the Next Generation Science Standards developed in 2013 include these concepts, the researchers said students often only read or hear about them, but they have fewer opportunities to conduct experiments on their own and discuss how and why their findings may vary.
For example, Kubsch has started a program in which German preservice teachers learn, in three to four 90-minute sessions over the course of a school year, how to teach students to reason about uncertainty using a three-part strategy:
- Be open to new evidence as scientific knowledge changes, rather than holding findings as unchangeable;
- Evaluate new evidence in light of prior information; and;
- Always consider alternative explanations for a finding.
Kubsch also developed an app called the 鈥淐onfidence Updater鈥� that teachers can use to help students think through their own claims and certainty of their findings.
A little less confidence may help
Every year, Alex Edwards poses a deceptively simple question to 6th grade science students: Are 6th grade boys or girls taller?
This could be a very basic data collection task: measure yourself and your classmates, chart the data, compare averages and report back. But Edwards likes to push back. Students realize some classmates round heights to the nearest inch while others round to the nearest quarter inch. They go back to develop a uniform system and multiple measures for each student. They realize boys are more likely than girls to be either very tall or very short, and discuss how to deal with outliers. Around and around the class goes, until finally students come up with their final height charts.
鈥淭hat graph [of boys鈥� and girls鈥� average heights] will usually be almost neck-and-neck. And they鈥檒l just look at it and go, this one鈥檚 taller. ... so whoever guessed girls, they were right and whoever guessed boys, they were wrong,鈥� Edwards said. 鈥淎nd then I say, 鈥楬ey, was my question, are boys in this class taller than the girls in this class?鈥� They鈥檒l say, no, it was all the 6th grade boys and all the 6th grade girls in the world. So how do we know we鈥檙e right with this? And that鈥檚 where I start to put that little bit of that uncertainty to them.鈥�
Hee-Sun Lee, a senior research scientist at the Concord Consortium, a science and digital education research group, asked more than 6,000 students to analyze data from either scientists or computer models, then make a claim and explain both their reasoning for the claim based on the data, their level of certainty in their claim, and the potential reasons for uncertainty. Lee found students鈥� after going through the tasks that made them think explicitly about their sources of uncertainty.
Edwards agreed that it鈥檚 important to regularly remind students of how variation and uncertainty support science. He starts every science test鈥攆rom 6th through 8th grade鈥攚ith a cover sheet of the same set of questions which act as mental reminders that scientific models are not always correct and that science is a process and not just facts to be studied.
Students likewise know they鈥檒l get 10 points off a lab report for describing a hypothesis as 鈥渃orrect鈥� or 鈥減roved鈥� rather than 鈥渟upported.鈥�
鈥淰ocabulary matters, and the way [students] perceive it in their minds, if they鈥檙e just saying 鈥榳e鈥檙e right,鈥� then that鈥檚 such a definite thing. There鈥檚 no room for there to be something else going on,鈥� Edwards said. 鈥淏ut if they use terminology like 鈥榮upported,鈥� then hopefully they鈥檝e made the little connection back that data supports this鈥攄oesn鈥檛 necessarily confirm it, but it at least supports it. It doesn鈥檛 mean that there are not other explanations out there, but this is the one that we had the most evidence for.鈥�
