← Back to Chapter 7

These are research notes and source trails used while drafting the manuscript. They are educational and not medical advice.

Purpose: evidence you can cite to justify (and strengthen) the claims in Chapter 7—especially around effort, iteration, self-trust, and why AI shortcuts can quietly steal development.

  1. Chapter 7 claims → research backing (quick map)

Craft builds durable confidence (not the performative kind).

Bandura’s self-efficacy theory emphasizes mastery experiences as a primary driver of efficacy beliefs (confidence that generalizes).

Over time, repeated mastery experiences predict persistence and resilience under challenge.

Key sources: 1

Praise that targets process/strategy supports resilience more than praise that targets ‘being smart.’

Experimental work shows intelligence praise can reduce resilience after setbacks, compared to effort/process praise.

Dweck’s work describes how process praise supports learning goals and persistence.

Key sources: 2 3

The ‘messy middle’ (drafts, corrections, iterations) is not a detour; it is the learning.

Bjork & Bjork’s ‘desirable difficulties’ explains why learning that feels harder can stick better and transfer more.

Kapur’s ‘productive failure’ shows that struggling first can improve later learning after instruction.

Key sources: 4 5

Good struggle is *productive*, not pointless suffering.

NCTM (Principles to Actions) defines productive struggle as deepening understanding rather than chasing correct answers; it recommends supporting it as a teaching practice.

Key sources: 6

Iteration + feedback is how expertise grows.

Ericsson et al. popularized deliberate practice: focused practice with feedback and correction is central to expertise.

Key sources: 7

Practice changes the brain—literally.

Neuroscience reviews describe activity-dependent myelination and white-matter plasticity as mechanisms linked to learning and experience-driven change.

Key sources: 8 9

AI can amplify craft *after* contact, but it can steal learning when it replaces the first attempt.

A 2025 arXiv preprint (Kosmyna et al.) reports lower brain connectivity and lower ownership in an LLM-assisted writing condition vs. no-tool writing (preprint caveat applies).

A 2024 review synthesizes many ChatGPT-in-education studies and reports mixed effects (engagement can rise or fall depending on use).

A 2025 meta-analysis finds mixed impacts on performance and higher-order thinking, reinforcing that implementation matters.

Key sources: 10 11 12

Policy bodies recommend a human-centered approach and guardrails.

UNESCO and OECD guidance emphasize human-centered, responsible use, including protecting learning goals, integrity, and student development.

Key sources: 13 14

  1. Insert-ready mini-paragraphs (fun voice, manuscript-ready)

Mastery → confidence

Psychologists call it self-efficacy: the belief that you can handle what’s in front of you. The fastest way kids build it isn’t pep talks — it’s mastery experiences: hard thing → effort → “I did it.”¹

Process praise

There’s a reason “You’re so smart!” can backfire. Research finds that when kids are praised for intelligence, they can become more afraid of mistakes; when they’re praised for strategy and effort, they’re more likely to persist.² ³

Desirable difficulty

Learning that feels a little harder can actually stick better. Researchers call these ‘desirable difficulties’ — the kind of struggle that makes learning last, not the kind that just makes kids miserable.⁴

Productive failure

Sometimes the best teaching sequence is: attempt first, teach second. ‘Productive failure’ research suggests that letting learners wrestle with a problem before instruction can produce deeper learning later.⁵

AI as coach not ghost

The boundary isn’t ‘no AI.’ It’s ‘no AI that replaces the rep.’ Early studies and broader reviews suggest AI can help as a tutor or feedback coach, but can reduce deep engagement when it does the core thinking for the student.¹⁰ ¹¹ ¹²

  1. Citable, short quotations (≤25 words each)

Mueller & Dweck (1998) (abstract):

““Praise for intelligence… had more negative consequences… than praise for effort.””

Source: 2

NCTM (Principles to Actions) (as summarized in ERIC):

“Productive struggle means students “delv[e] more deeply into understanding… instead of simply seeking correct solutions.””

Source: 6

  1. AI section caveats (useful for credibility)

Be explicit when evidence is preliminary. For example, Kosmyna et al. (2025) is an arXiv preprint (not peer-reviewed yet). That’s still citeable, but label it accurately.

Don’t claim ‘AI always harms learning.’ The literature is mixed: outcomes depend heavily on *how* AI is used (replacement vs. feedback/tutoring).

If you mention brain mechanisms (myelin/white matter), avoid oversimplified claims like “myelin makes signals 100x faster.” Stick to: practice is associated with plasticity in white matter/myelination, and that supports more efficient performance over time.

  1. Full reference list (URLs)

1. American Psychological Association. “Self-efficacy: The theory at the heart of human agency.” (overview of Bandura’s theory; highlights mastery experiences as key source).
https://www.apa.org/research-practice/conduct-research/self-efficacy-human-agency

2. Mueller, C. M., & Dweck, C. S. (1998). “Praise for intelligence can undermine children’s motivation and performance.” Journal of Personality and Social Psychology.
https://pubmed.ncbi.nlm.nih.gov/9686450/

3. Dweck, C. S. “The Perils and Promises of Praise.” (Process/effort praise vs. intelligence praise.)
https://teaching.temple.edu/sites/teaching/files/resource/pdf/Dweck-Perils%20%26%20Promises%20of%20Praise.pdf

4. Bjork, E. L., & Bjork, R. A. (2011). “Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning.”
https://bjorklab.psych.ucla.edu/wp-content/uploads/sites/13/2016/04/EBjork_RBjork_2011.pdf

5. Kapur, M. (2008). “Productive Failure.” Cognition and Instruction, 26(3).
https://arch.kuleuven.be/studeren/tall/artikels/productive-failure-kapur.pdf/@@download/file/Productive%20Failure%20Kapur.pdf

6. National Council of Teachers of Mathematics (NCTM). Principles to Actions: Ensuring Mathematical Success for All (2014). (Defines ‘productive struggle’ and recommends supporting it.)
https://www.nctm.org/uploadedFiles/Standards_and_Positions/Principles_to_Actions/Principles%20to%20Actions%20overview.pdf

7. Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). “The role of deliberate practice in the acquisition of expert performance.” Psychological Review, 100(3), 363–406.
https://doi.org/10.1037/0033-295X.100.3.363

8. Fields, R. D. (2015). “A new mechanism of nervous system plasticity: activity-dependent myelination.” Nature Reviews Neuroscience.
https://pubmed.ncbi.nlm.nih.gov/26585800/

9. Sampaio-Baptista, C., & Johansen-Berg, H. (2017). “White Matter Plasticity in the Adult Brain.” Neuron.
https://pmc.ncbi.nlm.nih.gov/articles/PMC5766826/

10. Kosmyna, N. et al. (2025). “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task.” arXiv preprint.
https://arxiv.org/abs/2506.08872

11. Lo, C. K. (2024). “The influence of ChatGPT on student engagement.” Computers & Education.
https://www.sciencedirect.com/science/article/pii/S0360131524001143

12. Wang, J. et al. (2025). “The effect of ChatGPT on students' learning performance, learning perception, and higher-order thinking.” Humanities and Social Sciences Communications.
https://www.nature.com/articles/s41599-025-04787-y

13. UNESCO (2023, updated 2025). Guidance for generative AI in education and research.
https://cdn.table.media/assets/wp-content/uploads/2023/09/386693eng.pdf

14. OECD (2023). Generative AI in the classroom: From hype to reality?
https://one.oecd.org/document/EDU/EDPC%282023%2911/en/pdf