Young Children Can Learn Scientific Reasoning with Teachable Agents
Open accessAbstract
A teachable agent (TA) is an instructional technology that capitalizes on the organizing metaphor of teaching another, in this case, a computer agent. As one instance, students teach their agents by creating concept maps that connect nodes with relational links. TAs use simple artificial intelligence to draw conclusions based on what they have been taught and to trace the path of their reasoning visually. TAs literally make thinking visible, with the goal of helping children learn to reason. TAs also provide interactive feedback and engender in students a sense of responsibility toward improving their agents' knowledge. We describe, in detail, a TA designed to teach hierarchical reasoning in science, and then present a 2-year research study using this TA with 153 fourth-grade children learning about organisms, taxonomies, and ecosystems. We show that the TA improved learning from the standard curriculum as measured by the curriculum's accompanying tests. The TA also helped children learn hierarchical reasoning, as measured by researcher-designed tests. The research indicates that, contrary to some theoretical positions, it is possible to help younger children learn scientific modes of reasoning, specifically by means of TA software.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Some algorithms were designed with the integration of role-play into the pedagogical design, where students are taught by an intelligent agent rather than being taught by the learning system.
“Some algorithms were designed with the integration of roleplay into the pedagogical design, where students are taught by an intelligent agent rather than being taught by the learning system [39].”
From A Review of Artificial Intelligence (AI) in Education from 2010 to 2020 (2021) · cites this paper as [39]
Section 2, “Three Design Principles for Teachable Agents,” states the opposite of the attributed teaching arrangement: students teach the computer character, rather than the character teaching them. The paper does use a teaching narrative, but it assigns the student the role of teacher. (LLM, conf=0.90)
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe recursive feedback may have the potential to foster learners' abilities to reason in specific ways because human-computer interaction is able to engender among students a sense of responsibility toward improving the construction of a knowledge repository.
“e recursive feedback may have the potential to foster learners' abilities to reason in specific ways because the human-computer interaction is able to engender among the students a sense of responsibility toward improving the construction of knowledge repository [39].”
From A Review of Artificial Intelligence (AI) in Education from 2010 to 2020 (2021) · cites this paper as [39]
Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Experimental — accuracy not yet measuredThe four studies reviewed focused on the utilization of modelling to support general reasoning, while the reasoning model should be largely domain-specific.
“e four studies reviewed focused on the utilization of modelling to support general reasoning, while the reasoning model should be largely domain-specific [24,39,40,42].”
From A Review of Artificial Intelligence (AI) in Education from 2010 to 2020 (2021) · cites this paper as [39]
Another research showed that students were able to learn the relevant facts and pairwise relations, while they may still fail to reason with them very well.
“Another research showed a similar conclusion: students were able to learn the relevant facts and pairwise relations, while they may still fail to reason with them very well [39].”
From A Review of Artificial Intelligence (AI) in Education from 2010 to 2020 (2021) · cites this paper as [39]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 40 · Cited by 54
Research works printed but not linked: 1
- [23]23. L. Martin and D.L. Schwartz, "Prospective Adaptation in the Use of Representational Tools," Cognition and Instruction, vol. 27, no. 4, pp. 1-31, 2009.
- Authors
- Chin, D. B., Dohmen, I. M., & Schwartz, D. L.
- Affiliations
- Stanford University (US)
- Venue
- IEEE Transactions on Learning Technologies
- Published
- 2013-07
- Volume
- 6
- Issue
- 3
- Pages
- 248-257
- ISSN
- 1939-1382 (ISSN-L)
- OpenAlex
- W2153695592
- Semantic Scholar
- ce2db7dfa30bce8b4c04c65811c92bb808ddae3d
- Cited by
- 54
- References
- 40
- Citations by year
- 2026: 8 · 2025: 6 · 2024: 6 · 2023: 4 · 2022: 3 · 2021: 4 · 2020: 3 · 2018: 2 · 2017: 5 · 2016: 5 · 2015: 3 · 2014: 3 · 2013: 2
- FWCI
- 4.79
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- IEEE Transactions on Learning Technologies (publishedVersion)
- Funders
- National Science Foundation (0634044, 0904324)
- Institute of Education Sciences (R305H060089)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
- Social Sciences › Education › Science Education and Pedagogy
- Social Sciences › Psychology › Developmental and Educational Psychology › Innovative Teaching and Learning Methods
- Keywords
- teachable agents, scientific reasoning, hierarchical reasoning, young children, fourth grade, concept maps, AI agents, taxonomy, ecosystems, interactive feedback, making thinking visible, science education
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
