---
title: "Learning science library: 100+ papers, in plain words"
url: "https://duckyhelper.com/research/library/"
description: "100+ real learning-science papers behind DuckyHelper, sorted by principle, each with a citation, a link to the publisher and a one-line summary."
updated: "2026-10-09"
---

Library

# The learning-science library

These are the 100+ papers DuckyHelper's tutor is built on: tutoring, retrieval practice, spacing, worked examples, feedback, pictures and words, AI tutors and more. Each entry gives the real title, authors, year and journal, a link to the publisher or DOI, our one-line summary, and where we checked it, a short quote in the paper's own words. Browse by principle or topic.

Updated October 9, 2026 · 1 min read

## How we checked every entry

From the model film: real first pages of the papers in this library. Click to watch the whole film with sound.

1. Every paper is real: its title, authors, year and journal were matched against Crossref or OpenAlex, and we hold the full text of 129 of them.
2. A quote is only shown when we found it word for word: in the paper's PDF (103 papers), in a scanned PDF read with text recognition (5), or in the published abstract (11). One quote we could not check against the article is left out.
3. The one-line summary under each title is ours, in plain words. It is never shown as a quote.
4. We link to the publisher, the DOI or the authors' own page. We do not host copies of the papers.

[Download the annotated bibliography (PDF)](https://duckyhelper.com/ducky-pages/pages/papers/duckyhelper-annotated-bibliography.pdf): every paper in the library with its full citation and our note.

## Browse the papers

Pick a principle or a topic, or search by title, author, year or journal. Each title links to the publisher or the DOI.

Search the library

Principles

Topics

146 papers

1. [One-on-one tutoring](https://duckyhelper.com/research/principles/teach-one-on-one/) **The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring** Bloom, 1984. *Educational Researcher*. Tutored students scored about two standard deviations above students taught in class. “the average student under tutoring was about two standard deviations above the average of the control class”
2. [One-on-one tutoring](https://duckyhelper.com/research/principles/teach-one-on-one/) **Learning from human tutoring** Chi et al., 2001. *Cognitive Science*. Students learned just as well when tutors prompted instead of explaining. “students learned just as effectively even when tutors were suppressed from giving explanations and feedback”
3. [One-on-one tutoring](https://duckyhelper.com/research/principles/teach-one-on-one/) **The Wisdom of Practice: Lessons Learned from the Study of Highly Effective Tutors** Lepper & Woolverton, 2002. *Improving Academic Achievement (Academic Press, book chapter)*. The best math tutors tend motivation and emotion, not just cognition.
4. [One-on-one tutoring](https://duckyhelper.com/research/principles/teach-one-on-one/) **The Relative Effectiveness of Human Tutoring, Intelligent Tutoring Systems, and Other Tutoring Systems** VanLehn, 2011. *Educational Psychologist*. Step-based tutoring systems came nearly as close as human tutors. “the effect size of intelligent tutoring systems was 0.76, so they are nearly as effective as human tutoring”
5. [One-on-one tutoring](https://duckyhelper.com/research/principles/teach-one-on-one/) **The Promise of Tutoring for PreK-12 Learning: A Systematic Review and Meta-Analysis of the Experimental Evidence** Nickow, Oreopoulos & Quan, 2024. *American Educational Research Journal*. Tutoring programs show consistent, substantial gains, about 0.29 standard deviations. “tutoring programs yield consistently substantial positive impacts on learning”
6. [AI answers and learning](https://duckyhelper.com/research/principles/guide-dont-hand-over-answers/) **Generative AI without guardrails can harm learning: Evidence from high school mathematics** Bastani et al., 2025. *Proceedings of the National Academy of Sciences (PNAS)*. Unguarded GPT-4 help lifted practice grades, then cut exam grades 17%. “unfettered access to GPT-4 can harm educational outcomes”
7. [Retrieval practice](https://duckyhelper.com/research/principles/retrieval-practice/) **Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention** Roediger & Karpicke, 2006. *Psychological Science*. Taking recall tests beat restudying for remembering material days later. “Testing is a powerful means of improving learning, not just assessing it.”
8. [Retrieval practice](https://duckyhelper.com/research/principles/retrieval-practice/) **The Critical Importance of Retrieval for Learning** Karpicke & Roediger, 2008. *Science*. Repeated testing, not repeated studying, drove long-term vocabulary recall. “Repeated studying after learning had no effect on delayed recall, but repeated testing produced a large positive effect”
9. [Retrieval practice](https://duckyhelper.com/research/principles/retrieval-practice/) **Retrieval Practice Produces More Learning than Elaborative Studying with Concept Mapping** Karpicke & Blunt, 2011. *Science*. Practicing recall beat concept mapping, even on a concept-mapping test. “practicing retrieval produces greater gains in meaningful learning than elaborative studying with concept mapping”
10. [Retrieval practice](https://duckyhelper.com/research/principles/retrieval-practice/) **Improving Students’ Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology** Dunlosky, Rawson, Marsh, Nathan & Willingham, 2013. *Psychological Science in the Public Interest*. Of ten study techniques, practice testing and spacing were rated most useful. “Practice testing and distributed practice received high utility assessments”
11. [Spaced practice](https://duckyhelper.com/research/principles/spaced-practice/) **Über das Gedächtnis: Untersuchungen zur experimentellen Psychologie** Ebbinghaus, 1885. *Duncker & Humblot (book; English: Memory, 1913)*. First experiments on memory, charting the forgetting curve and effects of repetition.
12. [Spaced practice](https://duckyhelper.com/research/principles/spaced-practice/) **Distributed practice in verbal recall tasks: A review and quantitative synthesis** Cepeda et al., 2006. *Psychological Bulletin*. Across 317 experiments, spacing helps, and longer retention calls for longer gaps. “the ISI producing maximal retention increased as retention interval increased”
13. [Spaced practice](https://duckyhelper.com/research/principles/spaced-practice/) **Organizing Instruction and Study to Improve Student Learning** Pashler et al., 2007. *IES Practice Guide (NCER 2007-2004), U.S. Department of Education*. Seven research-based recommendations, including spacing, worked examples and quizzing.
14. [Worked examples](https://duckyhelper.com/research/principles/worked-examples/) **The Use of Worked Examples as a Substitute for Problem Solving in Learning Algebra** Sweller & Cooper, 1985. *Cognition and Instruction*. Studying worked examples beat problem solving for novices: faster, with fewer errors. “worked examples, as expected, require considerably less time to process than conventional problems”
15. [Worked examples](https://duckyhelper.com/research/principles/worked-examples/) **Cognitive Load During Problem Solving: Effects on Learning** Sweller, 1988. *Cognitive Science*. Conventional problem solving eats working memory needed to learn the underlying schemas. “conventional problem‐solving activity is not effective in schema acquisition”
16. [Worked examples](https://duckyhelper.com/research/principles/worked-examples/) **Learning from Examples: Instructional Principles from the Worked Examples Research** Atkinson, Derry, Renkl & Wortham, 2000. *Review of Educational Research*. Good worked examples are integrated and placed right next to matched practice. “examples should be presented in close proximity to matched practice problems”
17. [Worked examples](https://duckyhelper.com/research/principles/worked-examples/) **Toward an Instructionally Oriented Theory of Example-Based Learning** Renkl, 2014. *Cognitive Science*. Examples are a very effective way to start learning a new skill. “Learning from examples is a very effective means of initial cognitive skill acquisition”
18. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **The Role of Tutoring in Problem Solving** Wood, Bruner & Ross, 1976. *Journal of Child Psychology and Psychiatry*. Good tutors handle what is beyond the learner and leave the rest to them.
19. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **Mind in Society: The Development of Higher Psychological Processes** Vygotsky, 1978. *Harvard University Press (book)*. Learners can do more with guidance than alone: the zone of proximal development. “the distance between the actual developmental level as determined by independent problem solving”
20. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **The Expertise Reversal Effect** Kalyuga, Ayres, Chandler & Sweller, 2003. *Educational Psychologist*. Techniques that help novices can backfire with more experienced learners. “Instructional techniques that are highly effective with inexperienced learners can lose their effectiveness”
21. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **Why Minimal Guidance During Instruction Does Not Work: An Analysis of the Failure of Constructivist, Discovery, Problem-Based, Experiential, and Inquiry-Based Teaching** Kirschner, Sweller & Clark, 2006. *Educational Psychologist*. For novices, minimal guidance is less effective than strong, explicit guidance. “minimally guided instruction is less effective and less efficient than instructional approaches that place a strong emphasis on guidance”
22. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **Scaffolding vs. Hints in the Assistment System** Razzaq & Heffernan, 2006. *ITS*. Compares breaking a missed problem into scaffolding questions against giving hints on request in ASSISTments.
23. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **Experimental Evaluation of Automatic Hint Generation for a Logic Tutor** Stamper et al., 2011. *AIED*. Hints generated automatically from past students' data helped students persist in a logic course. “hints help students persist in deductive logic courses”
24. [Scaffolding and hints](https://duckyhelper.com/research/principles/scaffolding/) **Principles of Instruction: Research-Based Strategies That All Teachers Should Know** Rosenshine, 2012. *American Educator*. Ten research-based principles, like small steps and checking every student's understanding. “Present new material in small steps with student practice after each step”
25. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Why a Diagram is (Sometimes) Worth Ten Thousand Words** Larkin & Simon, 1987. *Cognitive Science*. Diagrams can make problem solving easier because they put related information in one place. “Diagrammatic representations are indexed by location in a plane”
26. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Dual coding theory and education** Clark & Paivio, 1991. *Educational Psychology Review*. Memory has verbal and imagery codes, and teaching can use both.
27. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Animation: can it facilitate?** Tversky, Morrison & Betrancourt, 2002. *International Journal of Human-Computer Studies*. A review that asks whether animation beats static graphics, and finds the benefits often come from other differences. “The assumption is that graphics can facilitate comprehension, learning, memory, communication and inference”
28. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Nine Ways to Reduce Cognitive Load in Multimedia Learning** Mayer & Moreno, 2003. *Educational Psychologist*. People learn by linking pictures and words, if neither channel is overloaded. “meaningful learning involves cognitive processing including building connections between pictorial and verbal representations”
29. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Activity and Imagined Activity Can Enhance Young Children's Reading Comprehension** Glenberg et al., 2004. *Journal of Educational Psychology*. Young readers who moved toys (or imagined moving them) to act out a story understood and remembered it better. “Both actual manipulation and imagined manipulation resulted in markedly better (compared with rereading) memory for and comprehension of the text material”
30. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **The Promise and Practice of Learner-Generated Drawing: Literature Review and Synthesis** Van Meter & Garner, 2005. *Educational Psychology Review*. A review of research on students making their own drawings to learn.
31. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **DeFT: A conceptual framework for considering learning with multiple representations** Ainsworth, 2006. *Learning and Instruction*. The DeFT framework for when learning with several representations (text, graphs, pictures) helps. “Multiple (external) representations can provide unique benefits when people are learning complex new ideas”
32. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Gesturing makes learning last** Cook, Mitchell & Goldin-Meadow, 2008. *Cognition*. Children who gestured while learning a new math idea kept what they learned better. “requiring children to gesture while learning the new concept helped them retain the knowledge they had gained during instruction”
33. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Gesturing Gives Children New Ideas About Math** Goldin-Meadow, Cook & Mitchell, 2009. *Psychological Science*. Children told to make correct gestures during a math lesson learned more than children making partial or no gestures. “children required to produce correct gestures learned more than children required to produce partially correct gestures, who learned more than children required to produce no gestures”
34. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Drawing to Learn in Science** Ainsworth, Prain & Tytler, 2011. *Science*. Argues that drawing should be a key part of science education. “Emerging research suggests drawing should be explicitly recognized as a key element in science education”
35. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **The Cognitive Science of Visual-Spatial Displays: Implications for Design** Hegarty, 2011. *Topics in Cognitive Science*. A review of the cognitive science of graphs, maps and diagrams, and what it means for designing them. “This paper reviews cognitive science perspectives on the design of visual-spatial displays”
36. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Gesture's Role in Speaking, Learning, and Creating Language** Goldin-Meadow & Alibali, 2013. *Annual Review of Psychology*. A review of how gesture helps people speak, think and learn. “the gestures speakers produce when they talk are integral to communication”
37. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Drawing pictures during learning from scientific text: testing the generative drawing effect and the prognostic drawing effect** Schmeck et al., 2014. *Contemporary Educational Psychology*. Tests whether drawing pictures while reading a science text improves learning, and whether drawing quality predicts it.
38. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Creating visual explanations improves learning** Bobek & Tversky, 2016. *Cognitive Research: Principles and Implications*. Students who created visual explanations of a bicycle pump and of chemical bonding learned more. “creating a visual explanation increased understanding particularly for participants of low spatial ability”
39. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **Effects of observing the instructor draw diagrams on learning from multimedia messages** Fiorella & Mayer, 2016. *Journal of Educational Psychology*. Watching an instructor draw diagrams while explaining helped low prior knowledge learners transfer more than seeing finished diagrams. “watching the instructor draw diagrams (by viewing the instructor's full body) resulted in significantly better transfer test performance than viewing already-drawn diagrams for learners with low prior knowledge”
40. [Pictures and words](https://duckyhelper.com/research/principles/draw-it-as-you-explain/) **The drawing effect: Evidence for reliable and robust memory benefits in free recall** Wammes, Meade & Fernandes, 2016. *Quarterly Journal of Experimental Psychology*. Drawing a word's meaning made it easier to recall than writing the word out, across seven experiments. “Drawn words were better recalled than written”
41. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Formative assessment and the design of instructional systems** Sadler, 1989. *Instructional Science*. A theory of formative assessment: learners need to know the goal, where they are, and how to close the gap. “Feedback is defined in a particular way to highlight its function in formative assessment”
42. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **The Instructional Effect of Feedback in Test-like Events** Bangert-Drowns et al., 1991. *Review of Educational Research*. A meta-analysis of 58 effect sizes on feedback after test-like questions. “The present meta-analysis reviewed 58 effect sizes from 40 reports”
43. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Cognitive Tutors: Lessons Learned** Anderson, Corbett, Koedinger & Pelletier, 1995. *Journal of the Learning Sciences*. Tutors giving immediate feedback matched classroom results in a third of the time. “the tutor provides immediate feedback, consisting of short and directed error messages”
44. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Feedback and Self-Regulated Learning: A Theoretical Synthesis** Butler & Winne, 1995. *Review of Educational Research*. A theory of how feedback works inside self-regulated learning, including the feedback learners give themselves.
45. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory** Kluger & DeNisi, 1996. *Psychological Bulletin*. Feedback helps on average, but over a third of interventions hurt performance. “FI effectiveness decreases as attention moves up the hierarchy closer to the self and away from the task”
46. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Assessment and Classroom Learning** Black & Wiliam, 1998. *Assessment in Education*. A landmark review of classroom formative assessment and the learning gains from frequent feedback. “innovations designed to strengthen the frequent feedback that students receive about their learning yield substantial learning gains”
47. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Help Seeking and Help Design in Interactive Learning Environments** Aleven et al., 2003. *Review of Educational Research*. A review of on-demand help in learning software: learners often do not use help well. “learners are not using help facilities effectively”
48. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Off-Task Behavior in the Cognitive Tutor Classroom: When Students "Game the System"** Baker et al., 2004. *CHI*. Students who abused hints and feedback to get answers learned much less than students who did not. “students who frequently game the system score substantially lower on a post-test than students who never game the system”
49. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Toward Meta-cognitive Tutoring: A Model of Help Seeking with a Cognitive Tutor** Aleven et al., 2006. *International Journal of AI in Education*. Extends the Cognitive Tutor with a model of good and bad help seeking so it can coach how students ask for help. “an intelligent tutoring system that provides guidance with respect to students' meta-cognitive abilities can help them to become better learners”
50. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Formative assessment and self-regulated learning: a model and seven principles of good feedback practice** Nicol & Macfarlane-Dick, 2006. *Studies in Higher Education*. Seven principles of good feedback practice that help students take control of their own learning. “seven principles of good feedback practice that support self-regulation”
51. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **The Power of Feedback** Hattie & Timperley, 2007. *Review of Educational Research*. Feedback strongly shapes learning, but its type and delivery decide the effect. “Feedback is one of the most powerful influences on learning and achievement”
52. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Focus on Formative Feedback** Shute, 2008. *Review of Educational Research*. Formative feedback works best when supportive, timely and specific. “formative feedback should be nonevaluative, supportive, timely, and specific”
53. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **Improving students’ help-seeking skills using metacognitive feedback in an intelligent tutoring system** Roll et al., 2011. *Learning and Instruction*. The Help Tutor gave students immediate feedback on how they used hints in a geometry tutor. “The present research investigated whether immediate metacognitive feedback on students' help-seeking errors can help students acquire better help-seeking skills”
54. [Feedback](https://duckyhelper.com/research/principles/immediate-feedback/) **The Power of Feedback Revisited: A Meta-Analysis of Educational Feedback Research** Wisniewski, Zierer & Hattie, 2020. *Frontiers in Psychology*. A meta-analysis of 435 studies on feedback; effects vary a lot with what information the feedback carries. “Overall results based on a random-effects model indicate a medium effect (d = 0.48) of feedback on student learning”
55. [Self-explanation](https://duckyhelper.com/research/principles/explain-it-back/) **Self-Explanations: How Students Study and Use Examples in Learning to Solve Problems** Chi et al., 1989. *Cognitive Science*. Students who explain worked examples to themselves learn with real understanding. ““Good” students learn with understanding: They generate many explanations”
56. [Self-explanation](https://duckyhelper.com/research/principles/explain-it-back/) **Eliciting Self-Explanations Improves Understanding** Chi et al., 1994. *Cognitive Science*. Prompting eighth graders to self-explain deepened their understanding of circulation. “Generating explanations to oneself (self‐explaining) facilitates that integration process”
57. [Interleaving](https://duckyhelper.com/research/principles/interleaving/) **Memory and Metamemory Considerations in the Training of Human Beings** Bjork, 1994. *Metacognition: Knowing about Knowing (MIT Press, book chapter)*. Conditions that make practice harder can make learning last longer.
58. [Interleaving](https://duckyhelper.com/research/principles/interleaving/) **The shuffling of mathematics problems improves learning** Rohrer & Taylor, 2007. *Instructional Science*. Mixing different problem types in practice improved later math test scores.
59. [Interleaving](https://duckyhelper.com/research/principles/interleaving/) **Learning Concepts and Categories: Is Spacing the “Enemy of Induction”?** Kornell & Bjork, 2008. *Psychological Science*. Interleaving artists' paintings improved learning each style, though massing felt better. “Surprisingly, induction profited from spacing, even though massing apparently created a sense of fluent learning”
60. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **The AI Teacher Test: Measuring the Pedagogical Ability of Blender and GPT-3 in Educational Dialogues** Tack & Piech, 2022. *EDM*. Compares Blender and GPT-3 replies with real teachers' replies in the same dialogues on speaking like a teacher, understanding a student and helping a student. “speak like a teacher, understand a student, help a student”
61. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **MathDial: A Dialogue Tutoring Dataset with Rich Pedagogical Properties Grounded in Math Reasoning Problems** Macina et al., 2023. *Findings of EMNLP 2023*. LLMs solve math well but reveal solutions too early when tutoring. “While models like GPT-3 are good problem solvers, they fail at tutoring”
62. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Opportunities and Challenges in Neural Dialog Tutoring** Macina et al., 2023. *EACL*. Tests generative language models on two language-learning tutoring datasets and maps where they still fall short of human tutors. “dialog tutoring has largely remained unaffected by these advances”
63. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Language Models as Science Tutors** Chevalier et al., 2024. *ICML*. TutorEval tests whether language models can answer questions about long textbook chapters the way a science tutor would, and TutorChat trains for it. “fine-tuning base models with existing dialogue datasets leads to poor performance on TutorEval”
64. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Stepwise Verification and Remediation of Student Reasoning Errors with Large Language Model Tutors** Daheim et al., 2024. *EMNLP*. A dataset of student math solutions with the first wrong step marked by teachers; tutors that first find the error give better feedback. “they struggle to precisely detect student's errors and tailor their feedback to these errors”
65. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Effective and Scalable Math Support: Experimental Evidence on the Impact of an AI-Math Tutor in Ghana** Henkel et al., 2024. *AIED*. A field experiment with about 500 students in Ghana using Rori, a WhatsApp math tutor, one hour a week during study hall. “an AI-powered math tutor accessible via WhatsApp”
66. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Towards Responsible Development of Generative AI for Education: An Evaluation-Driven Approach** Jurenka et al., 2024. *arXiv (Google DeepMind tech report)*. Educators and learners preferred a pedagogy-tuned Gemini tutor over prompting alone. “LearnLM-Tutor is consistently preferred over a prompt tuned Gemini by educators and learners”
67. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **LearnLM: Improving Gemini for Learning** LearnLM Team, Google, 2024. *arXiv (Google tech report)*. Training on teaching instructions made Gemini the experts' pick for learning. “tuned to present information by default, rather than engage users in service of learning”
68. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **SocraticLM: Exploring Socratic Personalized Teaching with Large Language Models** Liu et al., 2024. *NeurIPS 2024*. An LLM tuned on Socratic teaching dialogues out-taught GPT-4. “actively engaging students in the thought process required for genuine problem-solving mastery”
69. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **ChatGPT-generated help produces learning gains equivalent to human tutor-authored help on mathematics skills** Pardos & Bhandari, 2024. *PLOS ONE*. ChatGPT-written hints produced learning gains comparable to human tutor hints. “only the ChatGPT condition produces statistically significant learning gains compared to a no-help control”
70. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Ruffle&Riley: Insights from Designing and Evaluating a Large Language Model-Based Conversational Tutoring System** Schmucker et al., 2024. *AIED*. A conversational tutor where two LLM agents, a student and a professor, let the learner teach and be taught from a lesson text. “orchestration in a learning-by-teaching format via two LLM-based agents”
71. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Pedagogical Alignment of Large Language Models** Sonkar et al., 2024. *Findings of EMNLP*. Trains LLMs with preference learning to guide students through problems rather than reply with the answer at once. “often provide immediate answers rather than guiding students through the problemsolving process”
72. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Bridging the Novice-Expert Gap via Models of Decision-Making: A Case Study on Remediating Math Mistakes** Wang et al., 2024. *NAACL 2024*. Adding expert remediation decisions made LLM tutor replies far more preferred. “the expert's decision-making model is critical for LLMs to close the gap”
73. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise** Wang et al., 2024. *arXiv preprint (v1)*. AI guidance for human tutors raised student topic mastery by 4 points. “less likely to give away the answer to the student”
74. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **From Chalkboards to Chatbots: Evaluating the Impact of Generative AI on Learning Outcomes in Nigeria** De Simone et al., 2025. *World Bank Policy Research Working Paper*. A randomized trial in Nigeria where secondary students used Microsoft Copilot (GPT-4) for English as an after-school tutor for six weeks. “The effect on English, the main outcome of interest, was of 0.23 standard deviations”
75. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **From Problem-Solving to Teaching Problem-Solving: Aligning LLMs with Pedagogy using Reinforcement Learning** Dinucu-Jianu et al., 2025. *EMNLP*. Uses reinforcement learning with simulated students to turn an LLM into a tutor that guides instead of giving away answers. “effective pedagogy which requires strategically withholding answers”
76. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **AI tutoring outperforms in-class active learning: an RCT introducing a novel research-based design in an authentic educational setting** Kestin et al., 2025. *Scientific Reports*. Students learned more in less time with an AI tutor than in class. “students learn significantly more in less time when using the AI tutor”
77. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Findings of the BEA 2025 Shared Task on Pedagogical Ability Assessment of AI-powered Tutors** Kochmar et al., 2025. *BEA Workshop @ ACL*. A shared task where over 50 teams built systems to judge AI tutor replies on finding, locating and guiding a student's mistake. “mistake identification, precise location of the mistake, providing guidance, and feedback actionability”
78. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Evaluating Gemini in an arena for learning** LearnLM Team, Google, 2025. *arXiv (Google tech report)*. Educators ran blind, head-to-head tutoring comparisons of leading AI models. “experts preferred Gemini 2.5 Pro in 73.2% of these match-ups”
79. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **MathTutorBench: A Benchmark for Measuring Open-ended Pedagogical Capabilities of LLM Tutors** Macina et al., 2025. *EMNLP*. An open benchmark for how well LLMs teach math, with a reward model that tells expert from novice tutor replies; solving skill did not translate into good teaching. “subject expertise, indicated by solving ability, does not immediately translate to good teaching”
80. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Unifying AI Tutor Evaluation: An Evaluation Taxonomy for Pedagogical Ability Assessment of LLM-Powered AI Tutors** Maurya et al., 2025. *NAACL 2025*. An eight-dimension benchmark for judging how well LLMs actually tutor. “highlighting which LLMs are good tutors and which ones are more suitable as question-answering systems”
81. [AI and LLM tutors](https://duckyhelper.com/research/principles/ai-tutoring-evidence/) **Training LLM-Based Tutors to Improve Student Learning Outcomes in Dialogues** Scarlatos et al., 2025. *AIED*. Trains an LLM tutor to pick replies that a model predicts will help the student answer correctly, while keeping the replies pedagogically sound. “they are not trained to maximize student learning throughout the course of a dialogue”
82. How tutors talk **Collaborative dialogue patterns in naturalistic one-to-one tutoring** Graesser, Person & Magliano, 1995. *Applied Cognitive Psychology*. Everyday tutoring works through joint problem solving, questions and explanation with examples. “Naturalistic one‐to‐one tutoring is more effective than traditional classroom teaching methods”
83. How tutors talk **Understanding Tutor Learning: Knowledge-Building and Knowledge-Telling in Peer Tutors’ Explanations and Questions** Roscoe & Chi, 2007. *Review of Educational Research*. Peer tutors mostly tell knowledge instead of building it with students. “Peer tutors, even when trained, focus more on delivering knowledge rather than developing it”
84. How tutors talk **Deliberative Discourse Idealized and Realized: Accountable Talk in the Classroom and in Civic Life** Michaels, O’Connor & Resnick, 2008. *Studies in Philosophy and Education*. Describes Accountable Talk, classroom discussion practices that hold students to reasoning, evidence and each other's ideas. “Classroom discussion practices that can lead to reasoned participation by all students are presented and described by the authors”
85. How tutors talk **Dialogue Act Modeling in a Complex Task-Oriented Domain** Boyer et al., 2010. *SIGDIAL*. A classifier for dialogue acts in computer-programming tutoring that uses both the chat and the student's task actions. “The task-oriented domain involves tutoring in computer programming exercises”
86. How tutors talk **The Teacher-Student Chatroom Corpus** Caines et al., 2020. *NLP4CALL*. A corpus of one-to-one online chat lessons between English teachers and learners, with scaffolding and correction annotated. “the teacher was able to focus exclusively on the linguistic abilities and errors of the student”
87. How tutors talk **CIMA: A Large Open Access Dialogue Dataset for Tutoring** Stasaski, Kao & Hearst, 2020. *BEA Workshop, ACL 2020*. A tutoring dialogue dataset where tutors choose between hints and questions. “One-to-one tutoring is often an effective means to help students learn”
88. How tutors talk **Instructions and Guide for Diagnostic Questions: The NeurIPS 2020 Education Challenge** Wang et al., 2020. *arXiv (NeurIPS 2020 Competition)*. The NeurIPS 2020 challenge on Eedi diagnostic questions, where wrong answer choices reveal specific misconceptions. “the answers that the students give to these questions reveal key information about the specific nature of misconceptions”
89. How tutors talk **Measuring Conversational Uptake: A Case Study on Student-Teacher Interactions** Demszky et al., 2021. *ACL*. Measures how teachers build on what students say, and links that uptake to outcomes. “teachers' uptake of student contributions has been linked to higher student achievement”
90. How tutors talk **Is it a good move? Mining effective tutoring strategies from human-human tutorial dialogues** Lin et al., 2022. *Future Generation Computer Systems*. Mines human tutoring dialogues to find which tutor moves go with better learning.
91. How tutors talk **The TalkMoves Dataset: K-12 Mathematics Lesson Transcripts Annotated for Teacher and Student Discursive Moves** Suresh et al., 2022. *LREC*. 567 K-12 math lesson transcripts labelled for teacher and student talk moves from accountable talk theory. “sustained classroom discourse is a critical component of equitable, engaging, and rich learning environments”
92. How tutors talk **The NCTE Transcripts: A Dataset of Elementary Math Classroom Transcripts** Demszky & Hill, 2023. *BEA Workshop, ACL 2023*. Dialogic teacher moves track with better observation scores and learning outcomes. “these moves are correlated with better classroom observation scores and learning outcomes”
93. How tutors talk **Is ChatGPT a Good Teacher Coach? Measuring Zero-Shot Performance For Scoring and Providing Actionable Insights on Classroom Instruction** Wang & Demszky, 2023. *BEA Workshop @ ACL*. Tests whether ChatGPT can score classroom transcripts and give teachers useful coaching feedback. “We explore whether generative AI could become a cost-effective complement to expert feedback by serving as an automated teacher coach”
94. How tutors talk **Can Automated Feedback Improve Teachers’ Uptake of Student Ideas? Evidence From a Randomized Controlled Trial in a Large-Scale Online Course** Demszky et al., 2024. *Educational Evaluation and Policy Analysis*. A randomized trial with 1,136 instructors: automated feedback on uptake of student ideas made instructors build on students more. “We find that M-Powering Teachers improves instructors' uptake of student contributions by 13%”
95. How tutors talk **Using Large Language Models to Assess Tutors' Performance in Reacting to Students Making Math Errors** Kakarla et al., 2024. *AAAI Workshop on AI for Education*. Checks whether GPT-3.5 and GPT-4 can grade how real tutors respond when a student makes a math error. “Rather than drawing direct attention to the error, tutors should guide the students to identify and correct their mistakes on their own”
96. How tutors talk **LLM Based Math Tutoring: Challenges and Dataset** Miller & DiCerbo, 2024. *EdArXiv*. A dataset for checking whether LLM tutors get the math right while tutoring, plus a look at the kinds of student interactions. “evaluating their performance in real-time math tutoring scenarios presents distinct challenges”
97. How tutors talk **Edu-ConvoKit: An Open-Source Library for Education Conversation Data** Wang & Demszky, 2024. *NAACL (Demo)*. An open-source library for cleaning, annotating and analyzing conversation data from classrooms and tutoring. “an open-source library designed to handle pre-processing, annotation and analysis of conversation data in education”
98. Intelligent tutoring systems **Knowledge tracing: Modeling the acquisition of procedural knowledge** Corbett & Anderson, 1994. *User Modeling and User-Adapted Interaction*. Bayesian knowledge tracing: a tutor estimates, skill by skill, how likely the student is to know it, and updates after every answer.
99. Intelligent tutoring systems **Intelligent Tutoring Goes To School in the Big City** Koedinger, Anderson, Hadley & Mark, 1997. *International Journal of Artificial Intelligence in Education*. An algebra tutor in Pittsburgh city schools raised standardized test scores.
100. Intelligent tutoring systems **AutoTutor: A simulation of a human tutor** Graesser et al., 1999. *Cognitive Systems Research*. A computer tutor that holds natural-language dialogue like a human tutor.
101. Intelligent tutoring systems **Performance Factors Analysis - A New Alternative to Knowledge Tracing** Pavlik, Cen & Koedinger, 2009. *AIED*. Performance Factors Analysis, a simple alternative to knowledge tracing that counts a student's past successes and failures per skill.
102. Intelligent tutoring systems **The Knowledge-Learning-Instruction Framework: Bridging the Science-Practice Chasm to Enhance Robust Student Learning** Koedinger, Corbett & Perfetti, 2012. *Cognitive Science*. The Knowledge-Learning-Instruction framework: match the kind of instruction to the kind of knowledge and learning involved. “we describe the Knowledge-Learning-Instruction (KLI) framework”
103. Intelligent tutoring systems **The ASSISTments Ecosystem: Building a Platform that Brings Scientists and Teachers Together for Minimally Invasive Research on Human Learning and Teaching** Heffernan & Heffernan, 2014. *International Journal of AI in Education*. ASSISTments, a free platform where teachers assign problems with hints and students get immediate feedback, also used for research. “The system gives immediate feedback to students while they are working”
104. Intelligent tutoring systems **Intelligent tutoring systems and learning outcomes: A meta-analysis** Ma et al., 2014. *Journal of Educational Psychology*. A meta-analysis of 107 effect sizes comparing learning with intelligent tutoring systems to other kinds of instruction. “107 effect sizes involving 14,321 participants were extracted and analyzed”
105. Intelligent tutoring systems **AutoTutor and Family: A Review of 17 Years of Natural Language Tutoring** Nye, Graesser & Hu, 2014. *International Journal of Artificial Intelligence in Education*. Conversational tutoring produced learning gains in physics, computer literacy and more. “AutoTutor is a natural language tutoring system that has produced learning gains across multiple domains”
106. Intelligent tutoring systems **Deep Knowledge Tracing** Piech et al., 2015. *NeurIPS*. Uses recurrent neural networks to model what a student knows from their sequence of answers. “we explore the utility of using Recurrent Neural Networks (RNNs) to model student learning”
107. Intelligent tutoring systems **Effectiveness of Intelligent Tutoring Systems: A Meta-Analytic Review** Kulik & Fletcher, 2016. *Review of Educational Research*. A meta-analysis of 50 controlled evaluations of intelligent tutoring systems. “The median effect of intelligent tutoring in the 50 evaluations was to raise test scores 0.66 standard deviations over conventional levels”
108. Intelligent tutoring systems **Online Mathematics Homework Increases Student Achievement** Roschelle et al., 2016. *AERA Open*. A randomized trial with 2,850 seventh graders in Maine: online homework with timely feedback and hints raised math scores. “Results showed that the intervention significantly increased student scores on an end-of-the-year standardized mathematics assessment”
109. Intelligent tutoring systems **Dynamic Key-Value Memory Networks for Knowledge Tracing** Zhang et al., 2017. *WWW*. A memory network for knowledge tracing that outputs a mastery level for each concept. “directly output a student's mastery level of each concept”
110. Intelligent tutoring systems **A Self-Attentive model for Knowledge Tracing** Pandey & Karypis, 2019. *EDM*. A self-attention model for knowledge tracing that handles sparse student data better than earlier neural models. “Knowledge tracing is the task of modeling each student's mastery of knowledge concepts”
111. Intelligent tutoring systems **EdNet: A Large-Scale Hierarchical Dataset in Education** Choi et al., 2020. *AIED*. A large public dataset of student interactions from an English test-prep tutor, for knowledge tracing research. “there is no public large-scale benchmark dataset”
112. Intelligent tutoring systems **Context-Aware Attentive Knowledge Tracing** Ghosh, Heffernan & Lan, 2020. *KDD*. Attentive knowledge tracing with interpretable parts from cognitive and psychometric models, including decay of what was learned. “predicting future learner performance given their past performance in educational applications”
113. Explanations, analogies and misconceptions **Analogical problem solving** Gick & Holyoak, 1980. *Cognitive Psychology*. A solved story helps crack an analogous problem, mostly when learners are hinted.
114. Explanations, analogies and misconceptions **Accommodation of a scientific conception: Toward a theory of conceptual change** Posner et al., 1982. *Science Education*. A classic theory of conceptual change: learners give up an idea only when it stops working and a new one makes sense. “learning is the result of the interaction between what the student is taught and his current ideas or concepts”
115. Explanations, analogies and misconceptions **Structure-mapping: A theoretical framework for analogy** Gentner, 1983. *Cognitive Science*. Analogies work by mapping relations between domains, not surface features. “Relations between objects, rather than attributes of objects, are mapped from base to target”
116. Explanations, analogies and misconceptions **Force concept inventory** Hestenes, Wells & Swackhamer, 1992. *The Physics Teacher*. The Force Concept Inventory, a test that shows how common sense beliefs about force survive regular physics teaching. “Instruction that does not take them into account is almost totally ineffective, at least for the majority of students”
117. Explanations, analogies and misconceptions **Mental models of the earth: A study of conceptual change in childhood** Vosniadou & Brewer, 1992. *Cognitive Psychology*. Children's ideas about the shape of the Earth show consistent mental models that mix what they are told with what they see. “many children said that the earth is round but also stated that it has an end or edge from which people could fall”
118. Explanations, analogies and misconceptions **Toward an Epistemology of Physics** diSessa, 1993. *Cognition and Instruction*. Describes intuitive physics as many small pieces of knowledge (p-prims) that learning has to reorganize.
119. Explanations, analogies and misconceptions **From things to processes: A theory of conceptual change for learning science concepts** Chi, Slotta & De Leeuw, 1994. *Learning and Instruction*. Argues that some science ideas are hard because students treat processes, like heat or current, as if they were things.
120. Explanations, analogies and misconceptions **The misunderstood limits of folk science: an illusion of explanatory depth** Rozenblit & Keil, 2002. *Cognitive Science*. People think they understand how things work much better than they do, until they try to explain it. “People feel they understand complex phenomena with far greater precision, coherence, and depth than they really do”
121. Explanations, analogies and misconceptions **Learning and transfer: A general role for analogical encoding** Gentner, Loewenstein & Thompson, 2003. *Journal of Educational Psychology*. Comparing two cases side by side helped learners pull out the shared idea and use it later. “Experiment 2 showed a marked advantage for comparing two cases over studying the 2 cases separately”
122. Explanations, analogies and misconceptions **The structure and function of explanations** Lombrozo, 2006. *Trends in Cognitive Sciences*. A review of how explanations shape learning and generalization by fitting new facts into what we already believe. “explanations accommodate novel information in the context of prior beliefs, and do so in a way that fosters generalization”
123. Explanations, analogies and misconceptions **Cognitive Supports for Analogies in the Mathematics Classroom** Richland, Zur & Holyoak, 2007. *Science*. Compares how teachers in Hong Kong, Japan and the US support analogies in math lessons. “Variations in the effective use of analogies in math instruction across countries may contribute to performance differences in the TIMSS studies”
124. Explanations, analogies and misconceptions **Why Instructional Explanations Often Do Not Work: A Framework for Understanding the Effectiveness of Instructional Explanations** Wittwer & Renkl, 2008. *Educational Psychologist*. A framework for why instructional explanations often fail and how to design ones that work. “research shows that they often do not contribute to learning”
125. Explanations, analogies and misconceptions **Selective effects of explanation on learning during early childhood** Legare & Lombrozo, 2014. *Journal of Experimental Child Psychology*. Asking young children to explain a toy helped them learn how it works more than only watching it. “explanation promotes causal learning”
126. Memory, motivation and metacognition **The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information** Miller, 1956. *Psychological Review*. The classic paper on the limits of short-term memory and on chunking information to stretch them.
127. Memory, motivation and metacognition **Levels of processing: A framework for memory research** Craik & Lockhart, 1972. *Journal of Verbal Learning and Verbal Behavior*. Memory depends on how deeply information is processed, not on which store it sits in. “An alternative framework for human memory research is then outlined in terms of depth or levels of processing”
128. Memory, motivation and metacognition **Working Memory** Baddeley & Hitch, 1974. *Psychology of Learning and Motivation*. The working memory model: a small active workspace for thinking, not just a short-term store.
129. Memory, motivation and metacognition **Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry** Flavell, 1979. *American Psychologist*. The paper that named metacognition: knowing and monitoring your own thinking.
130. Memory, motivation and metacognition **Motivational and self-regulated learning components of classroom academic performance** Pintrich & De Groot, 1990. *Journal of Educational Psychology*. Self-efficacy, self-regulation and strategy use predicted classroom performance for 173 seventh graders. “Self-efficacy and intrinsic value were positively related to cognitive engagement and performance”
131. Memory, motivation and metacognition **Unskilled and unaware of it: How difficulties in recognizing one's own incompetence lead to inflated self-assessments** Kruger & Dunning, 1999. *Journal of Personality and Social Psychology*. People with the least skill tend to overrate their skill the most, because judging it needs the same skill. “participants scoring in the bottom quartile on tests of humor, grammar, and logic grossly overestimated their test performance and ability”
132. Memory, motivation and metacognition **Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being** Ryan & Deci, 2000. *American Psychologist*. Self-determination theory: people are most motivated when they feel competent, autonomous and related to others. “The findings have led to the postulate of three innate psychological needs”
133. Memory, motivation and metacognition **The magical number 4 in short-term memory: A reconsideration of mental storage capacity** Cowan, 2001. *Behavioral and Brain Sciences*. Working memory holds about three to five chunks, fewer than the famous seven. “it is only three to five chunks”
134. Memory, motivation and metacognition **Becoming a Self-Regulated Learner: An Overview** Zimmerman, 2002. *Theory Into Practice*. An overview of self-regulated learning: setting goals, using strategies, and checking your own progress.
135. Memory, motivation and metacognition **Implicit Theories of Intelligence Predict Achievement Across an Adolescent Transition: A Longitudinal Study and an Intervention** Blackwell, Trzesniewski & Dweck, 2007. *Child Development*. Seventh graders who believed intelligence can grow did better in math over junior high. “the belief that intelligence is malleable (incremental theory) predicted an upward trajectory in grades over the two years of junior high school”
136. Memory, motivation and metacognition **Better to be frustrated than bored: The incidence, persistence, and impact of learners’ cognitive-affective states during interactions with three different computer-based learning environments** Baker et al., 2010. *International Journal of Human-Computer Studies*. Tracks learners' feelings in three computer learning environments; boredom was more persistent and more harmful than frustration.
137. Memory, motivation and metacognition **Mind-Set Interventions Are a Scalable Treatment for Academic Underachievement** Paunesku et al., 2015. *Psychological Science*. Short online growth mindset and purpose lessons raised grades for students at risk of dropping out. “delivered brief growth-mind-set and sense-of-purpose interventions through online modules to 1,594 students in 13 geographically diverse high schools”
138. Memory, motivation and metacognition **To What Extent and Under Which Circumstances Are Growth Mind-Sets Important to Academic Achievement? Two Meta-Analyses** Sisk et al., 2018. *Psychological Science*. Two meta-analyses found weak overall links between growth mindset and achievement, stronger for some groups. “Overall effects were weak”
139. Memory, motivation and metacognition **A national experiment reveals where a growth mindset improves achievement** Yeager et al., 2019. *Nature*. A national randomized experiment: a short online growth mindset lesson raised grades for lower-achieving students. “improved grades among lower-achieving students and increased overall enrolment to advanced mathematics courses”
140. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **A Time For Telling** Schwartz & Bransford, 1998. *Cognition and Instruction*. Comparing contrasting cases first prepares students to learn from a later lecture.
141. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **How People Learn: Brain, Mind, Experience, and School: Expanded Edition** Bransford, Brown & Cocking (eds.), National Research Council, 2000. *National Academies Press (book)*. Synthesis of learning science: how experts differ from novices and how transfer works.
142. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **The Hidden Lives of Learners** Nuthall, 2007. *NZCER Press (book)*. Decades of classroom recordings show peers and private worlds shape what students learn.
143. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **Productive Failure** Kapur, 2008. *Cognition and Instruction*. Students who struggled first with hard problems later transferred knowledge better. “suggesting a latent productivity in what initially seemed to be failure”
144. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **The ICAP Framework: Linking Cognitive Engagement to Active Learning Outcomes** Chi & Wylie, 2014. *Educational Psychologist*. Learning rises from passive to active to constructive to interactive engagement. “from passive to active to constructive to interactive, their learning will increase”
145. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **Active learning increases student performance in science, engineering, and mathematics** Freeman et al., 2014. *Proceedings of the National Academy of Sciences (PNAS)*. Across 225 studies, active learning raised exam scores and cut failure rates. “student performance on examinations and concept inventories increased by 0.47 SDs under active learning”
146. [Active learning](https://duckyhelper.com/research/principles/active-learning/) **Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom** Deslauriers et al., 2019. *Proceedings of the National Academy of Sciences (PNAS)*. Students in active classes learned more but felt they learned less. “when students experience the increased cognitive effort associated with active learning, they initially take that effort to signify poorer learning”

Every paper here is real and was checked: title, authors, year, venue and link. A quote is only shown when quote_status starts with "verified" (found word for word in the paper PDF, its OCR, or its published abstract). "note" is our own one-line summary, never a quote.

## Frequently asked questions

### Are these all real papers?

Yes. Each one was matched by title, authors, year and venue against Crossref or OpenAlex, and most were read in full. Every card links to the publisher, the DOI or the authors' page.

### Can I download the papers here?

No. We link to the original publisher, the DOI or the authors' own page instead of hosting copies, and many of the papers can be read at that link.

### Why do some papers have no quote?

We only show a quote when we found it word for word in the paper or its abstract. Some older papers are scans or books where we could not check a line, so they show our summary only.

## Keep reading

### How people learn: 12 principles

Twelve learning-science principles behind DuckyHelper, from one-on-one tutoring to spacing, each with the research and the tutor behaviour it drives.

### A model built to teach

DuckyHelper's tutor is trained on 1,000,000+ tutoring examples and inspired by 100+ learning-science papers: a tutor, not an answer machine.

### What the evidence says about AI tutors

Randomized trials of AI tutors so far: Harvard physics, Ghana, Nigeria, Tutor CoPilot, and the PNAS study where AI help hurt. What they show, and do not.

### Research: a tutor built on learning science

How DuckyHelper's tutor is built to teach: the model, a library of 100+ learning-science papers, the principles it follows and our white papers.

### Papers and technical notes

DuckyHelper's white papers: how the tutor is designed from learning science, drawing on the learner's screen, and hints before answers. PDF and HTML.

## Ready to learn

DuckyHelper is ready on your Mac. Learn with Eddy on its blackboard, or with Ducky right on your screen.

[Download for Mac](https://duckyhelper.com/download/)

[See what it does](https://duckyhelper.com/features/)

The Mac app needs Apple silicon and macOS 14 Sonoma or later. There is no Chromebook, Windows or web version right now.