Background Large language models (LLMs) are increasingly used in personal finance education, yet the role of instructional prompt design in shaping explanatory quality and ethical responsiveness remains underexplored. This matters because trustworthy financial guidance and instruction depend on clarity, accuracy, and appropriately cautionary framing. Method We conducted a Design-Based Research (DBR) Phase 1 artefact-validation study using content analysis of 120 GPT-4-turbo (May 2025) generations. Ninety diagnostic prompts varied linguistic framing, cognitive level, and ethical intent; thirty Phase 1b prompts instantiated a derived taxonomy. Two expert reviewers scored outputs (0–3) for financial accuracy, conceptual clarity, and ethical tone (inter-rater reliability: <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\kappa =0.82\)</EquationSource> </InlineEquation>). Results Prompt form was a pedagogical determinant of content quality. Procedural prompts outperformed interrogatives on clarity (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(U=418\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(p<.001\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(r=.42\)</EquationSource> </InlineEquation>) and accuracy (<InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(U=457\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(p<.01\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(r=.31\)</EquationSource> </InlineEquation>). Ethical anchoring raised ethical tone (median 2.78 vs. 2.31; Cliff’s <InlineEquation ID="IEq8"> <EquationSource Format="TEX">\(\delta =.46\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq9"> <EquationSource Format="TEX">\(p<.001\)</EquationSource> </InlineEquation>) with a modest accuracy trade-off (<InlineEquation ID="IEq10"> <EquationSource Format="TEX">\(\delta =-.18\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq11"> <EquationSource Format="TEX">\(p=.04\)</EquationSource> </InlineEquation>). Taxonomy-guided prompting increased medians for clarity (<InlineEquation ID="IEq12"> <EquationSource Format="TEX">\(+0.46\)</EquationSource> </InlineEquation> to 2.92), accuracy (<InlineEquation ID="IEq13"> <EquationSource Format="TEX">\(+0.41\)</EquationSource> </InlineEquation> to 2.81), and ethical tone (<InlineEquation ID="IEq14"> <EquationSource Format="TEX">\(+0.37\)</EquationSource> </InlineEquation> to 2.71), and reduced low-quality outliers from <InlineEquation ID="IEq15"> <EquationSource Format="TEX">\(18\%\)</EquationSource> </InlineEquation> to <InlineEquation ID="IEq16"> <EquationSource Format="TEX">\(4\%\)</EquationSource> </InlineEquation> (<InlineEquation ID="IEq17"> <EquationSource Format="TEX">\(\chi ^2=7.91\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq18"> <EquationSource Format="TEX">\(p=.005\)</EquationSource> </InlineEquation>), evidencing higher instructional reliability at the content level. Conclusion The prompt taxonomy, validated here at the content level, operationalises three design principles - Guided Explanation, Contextualised Inquiry, and Comparative Reasoning. These links prompt forms to instructional functions via structural scaffolding and situated relevance. Treating prompt engineering as instructional design can systematically improve AI-mediated educational content in high-stakes finance. DBR Phase 2 will test impacts on learner outcomes (decision quality, self-regulation, calibrated trust) and robustness across models and retrieval-augmented settings.

Prompt engineering as cognitive scaffolding for ethical and explanatory quality in AI-mediated financial learning

期刊论文
出版模式：

开放获取
发表日期：

2026年1月26日

查看全文

Kehinde Aruleba,
Ebenezer Esenogho,
Cameron Modisane

摘要

Background

Large language models (LLMs) are increasingly used in personal finance education, yet the role of instructional prompt design in shaping explanatory quality and ethical responsiveness remains underexplored. This matters because trustworthy financial guidance and instruction depend on clarity, accuracy, and appropriately cautionary framing.

Method

We conducted a Design-Based Research (DBR) Phase 1 artefact-validation study using content analysis of 120 GPT-4-turbo (May 2025) generations. Ninety diagnostic prompts varied linguistic framing, cognitive level, and ethical intent; thirty Phase 1b prompts instantiated a derived taxonomy. Two expert reviewers scored outputs (0–3) for financial accuracy, conceptual clarity, and ethical tone (inter-rater reliability: \(\kappa =0.82\) ).

Results

Prompt form was a pedagogical determinant of content quality. Procedural prompts outperformed interrogatives on clarity ( \(U=418\) , \(p<.001\) , \(r=.42\) ) and accuracy ( \(U=457\) , \(p<.01\) , \(r=.31\) ). Ethical anchoring raised ethical tone (median 2.78 vs. 2.31; Cliff’s \(\delta =.46\) , \(p<.001\) ) with a modest accuracy trade-off ( \(\delta =-.18\) , \(p=.04\) ). Taxonomy-guided prompting increased medians for clarity ( \(+0.46\) to 2.92), accuracy ( \(+0.41\) to 2.81), and ethical tone ( \(+0.37\) to 2.71), and reduced low-quality outliers from \(18\%\) to \(4\%\) ( \(\chi ^2=7.91\) , \(p=.005\) ), evidencing higher instructional reliability at the content level.

Conclusion

The prompt taxonomy, validated here at the content level, operationalises three design principles - Guided Explanation, Contextualised Inquiry, and Comparative Reasoning. These links prompt forms to instructional functions via structural scaffolding and situated relevance. Treating prompt engineering as instructional design can systematically improve AI-mediated educational content in high-stakes finance. DBR Phase 2 will test impacts on learner outcomes (decision quality, self-regulation, calibrated trust) and robustness across models and retrieval-augmented settings.