What Tutoring Can’t Teach Kids — And What Only Project Work Can

Engineering firms don’t hire kids who got As on problem sets. They hire kids who figured out how to make something work when it didn’t. No tutor can teach that.

That’s not a knock on tutoring. Good tutoring is genuinely useful — it fills gaps, reinforces concepts, and builds confidence on tested material. But there’s a ceiling. And when you watch a kid who aced every tutoring session freeze completely the first time they face a problem with no right answer, you’re seeing exactly where that ceiling is.

The skills employers and universities actually care about — problem framing, tolerance for ambiguity, iterative thinking, collaborative debugging — are not skills that one-on-one tutoring can structurally deliver. Not because tutors are bad at their jobs. Because the learning context itself prevents it.

The Skills That Predict Engineering Career Success

The National Academy of Engineering’s 2004 report The Engineer of 2020 named a specific list of competencies needed for 21st-century engineering practice: strong analytical skills, practical ingenuity, creativity, communication ability, business and management fluency, ethical grounding, and the capacity to lead. Conspicuously absent from that list: “scores well on tests of known material.”

Sheppard, Macatangay, Colby, and Sullivan’s landmark 2009 study Educating Engineers: Designing for the Future of the Field — a Carnegie Foundation for the Advancement of Teaching project examining engineering education across dozens of universities — found that even engineering schools themselves were weak on teaching students to frame problems rather than solve pre-framed ones. Most instruction, the researchers noted, was built around “well-structured problems” with known solutions. Real engineering almost never looks like that.

Dym, Agogino, Eris, Frey, and Leifer published a widely-cited 2005 paper in the Journal of Engineering Education on engineering design thinking. Their central argument: design thinking is fundamentally different from analytical problem-solving, and it develops through iteration, failure, and open-ended exploration — not through instruction in a structured setting.

LinkedIn’s 2024 “Future of Work” analysis of 700 million professional profiles found that the fastest-growing skill requirements in engineering and technology roles are creativity, complex problem-solving, and critical thinking — not any specific technical knowledge. Technical knowledge has a short shelf life. The capacity to figure things out when they don’t work has a very long one.

What Tutoring Teaches Well — and the Ceiling It Hits

Private tutoring is very good at specific things. It closes knowledge gaps. It provides personalized pacing on defined curriculum. It rescues a student before an exam. For a kid who understands the concept but makes careless arithmetic errors, a good tutor is exactly the right intervention. For a kid who missed a unit and is falling behind in class, tutoring is efficient remediation.

The structural problem is that tutoring is designed around known problems with known solutions. The tutor knows the right answer. The student is working toward a known destination. That’s appropriate for academic content, but it trains a specific cognitive posture: working toward confirmation rather than working through uncertainty.

When researchers at Stanford’s d.school studied how design-thinking novices approach problems, they found that the most common failure mode wasn’t lack of knowledge — it was the inability to stay with a problem when its boundaries were unclear. Students who had been trained in academic settings were, as the researchers put it, “uncomfortable with the feeling of not knowing where they were going.” That discomfort, left unaddressed, limits them more than any content gap ever could.

There’s also the social dimension. One-on-one tutoring, by definition, removes the collaborative element. A significant body of research on cooperative learning — including Johnson and Johnson’s decades of work at the University of Minnesota — shows that peer collaboration improves both learning outcomes and the ability to give and receive constructive feedback. Neither of those skills develops in a private tutoring session.

The Project-Work Skills That Tutoring Structurally Cannot Deliver

The table below compares what each context trains — not what tutors try to teach, but what the structure itself makes possible or impossible.

Skill	Private Tutoring	Project-Based Work	Why the Difference
Solving well-structured problems	Strong	Moderate	Tutoring is built for this by design
Framing an ill-defined problem	Weak	Strong	Projects start with open briefs; tutoring starts with a question
Tolerating ambiguity mid-process	Weak	Strong	Tutors resolve ambiguity to keep sessions moving
Iterating after failure	Limited	Strong	Tutoring corrects errors; projects require rework
Giving and receiving peer feedback	None	Strong	One-on-one format removes peers entirely
Building something from scratch	None	Central	Projects produce artifacts; tutoring produces understanding
Managing a process over weeks	Limited	Strong	Projects span time; tutoring sessions are contained units
Debugging a complex system	Weak	Strong	Complexity is reduced in tutoring to clarify concepts

The row on debugging is worth pausing on. Debugging — not software debugging specifically, but the general skill of isolating why something isn’t working in a complex system — is one of the highest-value engineering skills there is. And it requires a specific condition: the system has to be complex enough that the failure mode isn’t obvious. Tutoring simplifies systems to make concepts clear. That’s exactly the opposite of what debugging practice requires.

Real Engineering Thinking Requires Ambiguity — and Tutoring Eliminates It

The 2018 PISA creative problem-solving assessment tested 15-year-olds in 41 countries on problems that were deliberately open-ended, with multiple valid approaches and no single correct answer. The results showed large variance that wasn’t predicted by math and reading scores — students who performed well on traditional PISA content sometimes performed poorly when problem structure was removed.

The researchers, and subsequent analysts including the OECD’s Andreas Schleicher, noted that this gap was likely explained by exposure: students who had more experience with open-ended, ill-structured tasks performed better on them. Students who had trained almost exclusively in structured academic settings struggled.

Tutoring is, structurally, a very structured setting. The tutor has a plan. There’s a learning objective. Problems are chosen because they’re appropriate for the student’s level. Ambiguity — the kind that exists in a real engineering problem — is managed out of the session to protect the student’s confidence and keep the session on track. Those are good tutoring practices. But they have a cost.

As an engineer who spent fifteen years on consumer product development, I saw this gap clearly in how new graduates performed. The ones who adapted quickly weren’t always the ones with the best grades. They were the ones who had built things that didn’t work and figured out why. That experience — specific and personal — is not something you can acquire by getting the right answer on a tutor’s worksheet.

What Project-Based Programs Actually Develop vs. What They Claim

Not all “project-based learning” is equal. The phrase has been applied to everything from a poster project in a third-grade classroom to multi-week engineering design challenges with real stakes and feedback loops. The research on project-based learning is positive but uneven, and the difference usually traces back to design quality.

Krajcik and Shin’s 2014 review in Educational Psychologist identified the elements that make project-based learning effective: a driving question that’s genuinely ill-defined, sustained inquiry over multiple weeks, public artifact production, and structured reflection. Programs that have those elements show measurable gains in problem-solving, self-regulation, and content knowledge. Programs that use the “project” label for what is essentially a directed activity with predetermined outcomes show much weaker effects.

What to look for in a program: Does the project have a real possibility of failure? Does the kid make design decisions that matter? Does something get built, not just written about? Is there a feedback loop — does the kid learn why their design didn’t work, not just that it didn’t work? Those are the structural markers of a project context that actually develops engineering thinking.

For more on the research behind hands-on versus passive learning approaches, hands-on STEM learning and constructivism research at HIWVE covers the neuroscience in detail. And if you’re wondering whether intensive tutoring programs might actually be creating the wrong kind of dependency, STEM tutoring and dependent learners research makes the case more directly.

How to Balance Tutoring and Project Work in Your Kid’s Schedule

The honest answer here is that these two things are not in competition — they serve different purposes, and both have legitimate roles. The problem comes when tutoring crowds out project time completely, not when it exists.

A useful framing: tutoring is maintenance on a car; project work is learning to drive. You need both, but if you spend all your time maintaining the car and never actually drive it, you haven’t really learned anything useful.

Use tutoring for closed-ended deficits

If your kid is struggling with fractions, can’t parse a word problem, or is about to fail a test, tutoring is the right tool. It’s efficient and targeted. Use it for that.

Use project contexts for open-ended development

Block time in the week for a project that doesn’t have a predetermined outcome. This could be a formal program, a family maker project, a coding challenge, or a physical build. What matters is that the outcome is uncertain and your kid has to make real decisions to get there.

Watch the ratio over a school year

If your kid is spending eight hours a week in tutoring and zero in any unstructured project context, the ratio is probably worth examining. Eight hours of gap-filling with no application context means the knowledge has nowhere to land.

What NOT to do: don’t “project-ify” tutoring

Adding a “project” to a tutoring session — like having a tutor assign a science fair topic — doesn’t produce the benefits of real project-based learning. The context still has a tutor managing the ambiguity in real time. The project needs its own space, its own stakes, and its own failure modes.

What to Watch For Over the Next 3 Months

If you add structured project time alongside existing tutoring, here are observable signals:

• Week 4: Your kid can describe what they’re trying to build and what went wrong last time without being prompted. That’s self-directed problem framing — a genuine indicator.
• Month 2 red flag: The project is always going smoothly and your kid never expresses frustration. Real projects hit walls. If there are no walls, the task is too structured.
• Month 3 self-check: Ask your kid to explain a choice they made in the project — not what they did, but why they did it over an alternative. If they can do that, they’re developing engineering thinking. If they look blank, the project may not be giving them real decision-making authority.

Frequently Asked Questions

Can a good tutor teach problem-solving skills?

A good tutor can teach heuristics, model problem-solving approaches, and help a student practice on increasingly complex problems. But the tutor still manages the session — selecting problems, providing hints, and preventing the student from getting truly stuck. Real problem-solving under ambiguity requires exactly the experience of getting stuck without a tutor to rescue you.

My kid’s school does project-based learning. Isn’t that enough?

It depends heavily on implementation quality. Many school “projects” are heavily scaffolded, have predetermined outcomes, and don’t give students real design authority. Krajcik and Shin’s (2014) review found that the quality of project-based learning in schools varies enormously. Evaluate by asking: could the project fail? Does your kid make decisions that actually matter to the outcome?

At what age should kids start project-based work?

Research from Papert’s constructionism framework and subsequent early-childhood work suggests that genuine making and building can start as early as 5–6, even if the projects are simple. The key is that the outcome is uncertain and the child controls the design. Complexity scales with age; the structural principle is the same.

Is tutoring worth it if my goal is engineering?

For specific skill gaps in math, physics, or related content — yes, unambiguously. For developing the thinking style of an engineer — it’s the wrong tool. Engineering programs and employers consistently report that the students who arrive best prepared have done the most building, not the most structured studying.

How much project time per week actually makes a difference?

The research doesn’t give a clean prescription, but Krajcik and Shin note that meaningful project-based learning typically unfolds over multiple weeks of sustained work. An hour or two per week of genuine project engagement over several months is more valuable than an intensive weekend that doesn’t persist.

What if my kid hates unstructured projects and only wants tutoring?

That preference itself is a signal worth noting. Children who resist open-ended tasks often do so because they’ve learned to be afraid of being wrong — a posture that structured academic settings can accidentally reinforce. Start with smaller, lower-stakes making experiences and build tolerance gradually. Building vs. watching: how brain circuits develop in kids covers why that resistance is so common and what actually helps.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

Sources

1. National Academy of Engineering. (2004). The Engineer of 2020: Visions of Engineering in the New Century. National Academies Press. https://nap.nationalacademies.org/catalog/10999/the-engineer-of-2020-visions-of-engineering-in-the-new-century
2. Sheppard, S., Macatangay, K., Colby, A., & Sullivan, W. (2009). Educating Engineers: Designing for the Future of the Field. Carnegie Foundation for the Advancement of Teaching. https://eric.ed.gov/?id=ED508360
3. Dym, C. L., Agogino, A. M., Eris, O., Frey, D. D., & Leifer, L. J. (2005). Engineering design thinking, teaching, and learning. Journal of Engineering Education, 94(1), pp. 103–120. https://doi.org/10.1002/j.2168-9830.2005.tb00832.x
4. Krajcik, J. S., & Shin, N. (2014). Project-based learning. In R. K. Sawyer (Ed.), The Cambridge Handbook of the Learning Sciences (2nd ed., pp. 275–297). Cambridge University Press. https://doi.org/10.1017/CBO9781139519526.018
5. OECD. (2019). PISA 2018 Results (Volume I): What Students Know and Can Do. OECD Publishing. https://doi.org/10.1787/5f07c754-en
6. Johnson, D. W., & Johnson, R. T. (2009). An educational psychology success story: Social interdependence theory and cooperative learning. Educational Researcher, 38(5), pp. 365–379. https://doi.org/10.3102/0013189X09339057
7. LinkedIn. (2024). 2024 Future of Work Report: AI at Work. LinkedIn Talent Solutions. https://business.linkedin.com/talent-solutions/resources/future-of-recruiting
8. Papert, S. (1980). Mindstorms: Children, Computers, and Powerful Ideas. Basic Books.