Picture a physics worksheet from almost any decade of the last fifty years. Every value a student needs sits in a neat box at the top of the page: the mass of the cart, the angle of the ramp, the starting speed and the friction coefficient. The student picks an equation, plugs in the numbers and checks the answer against the key in the back of the book. A few minutes of work and a solid grade.
Then that student grows up, becomes an engineer, and never sees a box of givens again.
Real engineering problems arrive messy. A customer wants a lighter bracket and forgets to mention that it sits two inches from an exhaust port. A test report lists the load and leaves out the temperature. Some of the information is wrong, a fair amount of it is missing, and the one person who knows the answer is out until Tuesday. Before anyone solves anything, somebody has to notice the holes.
Most classroom STEM materials skip that step. They’re written to be finished, which makes sense for a 50-minute class period and a lot less sense for a career. The better ones leave things out on purpose, and they do it carefully.
A well-built incomplete problem gives students a goal and a few real constraints, then stops talking. Students have to work out what else they need to know. They make a list of unknowns. They argue about which ones matter (this part gets loud), and then they go measure something or run a quick test to find out.
That list of unknowns is the lesson. Most of it, anyway. The formulas still show up later, and they land harder because students already know why they need them.
Learning scientists call it problem-solving before instruction. A stricter version goes by a blunter name: productive failure. Students tackle a problem before anyone shows them the method, and most of them don’t solve it cleanly. That’s expected. The struggle shows each student exactly where their knowledge runs out, so when the teacher finally explains the concept, there’s a hook ready to hang it on. A study published on April 21, 2025, in the journal Instructional Science put the idea to work somewhere unexpected: a first-year linear algebra course for mechanical engineering students at ETH Zurich, one of Europe’s best-known technical universities. Researchers including Vera Baumgartner and Manu Kapur wrote ten preparatory exercise sheets that students completed before the matching lecture. The sheets introduced no new definitions. They pointed students at math they already knew and nudged them to notice where it stopped working.
The added work came to roughly ten extra hours over the whole year, or less than an hour a month. In the program’s second year, 20 percent more students passed the final exam than in the baseline years, and average exam performance climbed 13.2 percent.
The World Economic Forum’s Future of Jobs Report 2025, released in January 2025 and built on input from more than 1,000 employers, ranks analytical thinking as the top core skill for the third edition in a row. Seven out of 10 companies called it essential.
Analytical thinking is a slippery phrase. Day to day, a lot of it comes down to one habit: looking at a problem and figuring out what’s missing before grabbing a formula. A student who practices that at fourteen won’t have to learn it from scratch at twenty-four with a deadline breathing down their neck.
Teachers worry about this approach, and the worry makes sense. A room full of seventh graders staring at a problem with missing pieces can turn into a room full of seventh graders staring at the clock.
The research is fairly specific about what keeps it on track. The problem has to be challenging without tipping into frustrating, and it should leave room for more than one way in. Just as important, the teacher has to come back afterward, set the students’ rough attempts next to the standard method and talk through the differences. Skip that last step and the struggle doesn’t add up to much. Well, it adds up to a frustrated class, which is the exact outcome everyone was worried about in the first place.
New Hampshire has a good example running right now. In August 2026,
New Hampshire Public Radio reported on a University of New Hampshire tutoring program that pairs undergraduates with middle school students and their teachers. One Manchester class designed a fictional zoo. To make it work, students had to calculate how much food the animals would need and research each species’ biology and habitat requirements. None of those numbers came printed on a worksheet. A new $10 million federal grant will let UNH grow the program from its current partner districts in Manchester and Nashua to 40 middle schools, mostly in rural parts of the state.
Mentis Sciences has run its own version of this through its high school internship program. For the Mentis Bridge Building Competition intern teams got limited instructions and very little guidance on structural design. They were encouraged to ask questions and research design techniques on their own.
What happened next is a tidy little engineering lesson. The teams ran coupon tests (small sample tests on the materials before building anything) to find the best bond length for their joints, because the hot glue they had been given was very compliant. Too little bonding area and the joints became the weak points of the whole structure.
Joints are the part of a student bridge everyone ignores until test day. Composites engineers will grumble, to anyone within earshot, that joints get ignored in full-size structures too, usually right up until the first test. The interns learned it in high school, which puts them a few years ahead of schedule.
The strongest STEM materials share a handful of habits. They state the goal clearly and leave the path open. They include real constraints along with a few deliberate gaps, and they build in time for students to compare different attempts before anyone reveals the standard answer. The teacher’s guide spends as much effort on connecting messy student work to the formal concept as it does on the concept itself. None of this requires expensive equipment. Leaving a number off a page costs nothing, and it might be the cheapest upgrade available to any science teacher in the building.
→ State a clear goal and leave the path open
→ Leave out key information on purpose, so students have to go find it
→ Give students time to compare different attempts
→ Connect rough attempts back to the formal concept afterward
The most valuable thing a STEM worksheet can do is leave something out and let students go find it. Engineers spend their whole careers doing exactly that.
Mentis Sciences builds hands-on STEM tools with the same thinking behind its engineering work, where problems rarely arrive with every number filled in. The
Mentis Sciences Engineering Toolkit (MSET) scales down tests used in a working manufacturing facility into one integrated kit configured to complete 40 STEM tests, from buckling and three point flexure to buoyancy and pendulum dynamics.