arXiv subjects

René F. Kizilcec

Publications and source records attributed to René F. Kizilcec.

18 recordsLinked to original sources

arXiv2026

Redact or Keep? A Fully Local AI Cascade for Educational Dialogue De-Identification

Educational dialogue is a valuable but sensitive resource for research: the same transcripts that capture authentic learning often capture personally identifiable information (PII) entangled with curricular content, where "Riemann" may refer to a real student or to a mathematical concept. Existing approaches force a tradeoff between governance and accuracy. Commercial Large Language Models (LLMs) can handle this ambiguity but require sending student data to third parties, while local named entity recognition (NER) systems preserve governance but over-redact curricular terms. We propose a fully local cascade framework that reframes de-identification from open-ended entity recognition to constrained privacy triage. A recall-first union proposer combines two lightweight encoders with deterministic rules to over-generate candidate spans; a context-aware reviewer then makes a binary Redact/Keep decision for each candidate using surrounding dialogue and speaker role. We evaluate three reviewer configurations against same-family LLM-only baselines and a commercial API on math tutoring transcripts from two large platforms. The strongest local configuration reaches 0.958 macro F1, compared with 0.767 for a same-family LLM-only baseline and 0.706 for the commercial API, while running entirely on a single laptop. On a targeted challenge set of curricular-personal name ambiguity, the same configuration degrades by only 0.03 F1 versus 0.19 to 0.25 for smaller reviewers. These results suggest that for educational de-identification, problem formulation matters more than model scale.

cs.CL↗

arXiv2026

A mechanistic model of trust based on neural information processing

Trust is central to human social interactions, manifesting as a critical information processing step in taking actions that make one vulnerable to another. We argue that trust depends on the decision-making processes that arise in neural systems. Building on advances in the cognitive neuroscience of decision making, we propose a mechanistic model of trust arising differently in multiple parallel systems that perform distinct, complementary information processing. Because each system learns via different computational mechanisms, they will interact with the environment differently, and trust can be created (or destroyed) in multiple ways. This systems- level taxonomy of information representations provides a principled basis for differentiating forms of trust, linking them to specific learning processes, and generating testable predictions about their expression in behavior. Furthermore, because these different computational processes are implemented by different neural circuits, our theory makes testable predictions about the different neural circuits underlying different kinds of trust. By situating trust within a broader theory of neural decision systems, our account unifies diverse findings across psychology, neuroscience, and the social sciences, and offers a foundation for explaining how humans develop, maintain, lose, and repair trust in a complex social world.

econ.GN↗

arXiv2026

Utility-Preserving De-Identification for Math Tutoring: Investigating Numeric Ambiguity in the MathEd-PII Benchmark Dataset

Large-scale sharing of dialogue data is key to advancing the science of teaching and learning, yet rigorous de-identification remains a major barrier. In mathematics tutoring transcripts, numeric expressions frequently resemble structured identifiers (e.g., dates or IDs), leading generic Personally Identifiable Information (PII) detection systems to over-redact core instructional content and reduce data utility. This work asks how to detect PII while preserving educational utility, focusing on this "numeric ambiguity" problem. We introduce MathEd-PII, the first benchmark dataset for PII detection in math tutoring dialogues, built with human-in-the-loop LLM annotation. Using density-based segmentation, we show that false PII redactions cluster in math-dense regions, confirming numeric ambiguity as a key failure mode. We then compare four detection strategies: a Presidio baseline and three LLM-based approaches with basic, math-aware, and segment-aware prompting. Domain-aware prompting, including both math-aware (F1: 0.802) and segment-aware versions (F1: 0.821), substantially outperforms the baseline (F1: 0.379) while reducing numeric false positives, demonstrating that de-identification must incorporate domain context to preserve analytic utility. This work provides a new benchmark and evidence that utility-preserving de-identification for tutoring data requires domain-aware modeling.

cs.CL↗

arXiv2026

Does Algorithmic Uncertainty Sway Human Experts? Evidence from a Field Experiment in Selective College Admissions

Algorithmic predictions are inherently uncertain: even models with similar aggregate accuracy can produce different predictions for the same individual, raising concerns that high-stakes decisions may become sensitive to arbitrary modeling choices. In this paper, we define \emph{algorithmic sensitivity} as the extent to which arbitrary modeling choices propagate into human decisions: how much a decision outcome shifts when a more favorable versus less favorable algorithmic prediction is presented to the decision-maker for the same individual. We estimate this in a randomized field experiment ($n=19{,}545$) embedded in a selective U.S. college admissions cycle, in which admissions officers reviewed each application alongside an algorithmic score while we randomly varied whether the score came from one of two similarly accurate prediction models. Although the two models performed similarly in aggregate, they frequently assigned different scores to the same applicant, creating exogenous variation in the score shown. Surprisingly, we find little evidence of algorithmic sensitivity: presenting a more favorable score does not meaningfully increase an applicant's probability of admission on average, even when the models disagree substantially. These findings suggest that, in this expert, high-stakes setting, human decision-making is largely invariant to arbitrary variation in algorithmic predictions, underscoring the role of professional discretion and institutional context in mediating the downstream effects of algorithmic uncertainty.

cs.HC↗

arXiv2026

Who Decides in AI-Mediated Learning? The Agency Allocation Framework

As AI-mediated learning systems increasingly shape how learners plan, make decisions, and progress through education, learner agency is becoming both more consequential and harder to conceptualize at scale. Existing research often treats agency as a proxy for engagement and self-regulation, leaving unclear who actually holds decision-making authority in large-scale, automated learning environments. This paper reframes learner agency as the allocation of decision authority across learners, educators, institutions, and AI systems. We introduce the Agency Allocation Framework (AAF) for analyzing how decisions are distributed, how choices are architected, what evidence supports them, and over what time horizons their consequences unfold. Drawing on a focused review of Learning at Scale literature and an illustrative tutoring-system example, we identify four recurring challenges for studying learner agency at scale: (1) conceptual ambiguity, (2) reliance on behavioral proxies, (3) trade-offs between efficiency and learner control, and (4) the redistribution of agency through AI-mediated systems. Rather than advocating more or less automation, the AAF supports systematic analysis of when AI scaffolds learners' capacity to act and when it substitutes for it. By making decision authority explicit, the framework provides researchers and designers with analytic tools for studying, comparing, and evaluating agency-preserving learning systems in increasingly automated educational contexts.

cs.CY↗

arXiv2026

Teachers' Perceived Benefits and Risks of AI Across Fifty-Five Countries: An Audit of LLM Alignment and Steerability

Teachers' trust in artificial intelligence (AI) in education depends on how they balance its perceived benefits and risks. Yet global discussions about scaling AI in education rely on fragmented evidence, as most studies of teachers' perceptions focus on single countries or small samples. This lack of representative cross-national evidence limits both theory building and policy development. At the same time, large language models (LLMs) are increasingly used in research, policy, and teachers' professional workflows, despite limited validation in education. To address these gaps, we conduct a large-scale audit of LLM alignment with teachers' perceptions of AI by combining representative international survey data with systematic model evaluation. Using OECD TALIS data from 55 countries and territories, we measure cross-national variation in teachers' perceived benefits and risks of AI. We then benchmark responses from eight state-of-the-art LLMs across four providers under both general and country-specific prompting, comparing higher- and lower-reasoning models. Results reveal substantial cross-national variation in teacher perceptions that is not reliably reflected in LLM outputs. Models compress country differences, overestimate both benefits and risks, and show limited gains from identity prompting or enhanced reasoning. This misalignment matters because LLM-generated guidance and professional discourse increasingly shape how teachers learn about and discuss AI, potentially influencing trust and future adoption decisions. Our findings caution against treating LLM outputs as substitutes for direct engagement with teachers when informing global AI-in-education initiatives. At the same time, some models (e.g., Gemini 3 Fast) partially capture cross-national ranking patterns, suggesting a complementary role in hypothesis generation and exploratory comparative analysis.

cs.CY↗

arXiv2026

Does the TalkMoves Codebook Generalize to One-on-One Tutoring and Multimodal Interaction?

Accountable Talk theory has been widely adopted to analyze classroom discourse and is increasingly used to annotate tutoring interactions. In particular, the TalkMoves codebook, grounded in Accountable Talk theory, is commonly used to label tutoring data and train models of effective instructional support. However, Accountable Talk was originally developed to characterize collaborative, whole-classroom oral discourse, not to identify talk moves in one-on-one tutoring environments using multimodal data (e.g., video, audio, chat). As tutoring platforms expand in scale and modality, questions remain about whether Accountable Talk-based codebooks generalize reliably beyond their original classroom context and data representation. This study examines whether the human-developed TalkMoves codebook generalizes in reliability, utility, and interpretability when applied to one-on-one tutoring across audio, chat, and multimodal data. We compare TalkMoves with a hybrid AI-human developed codebook using a workflow established in prior research. Two expert annotators with over 20 years of teaching experience applied both codebooks to six tutoring sessions spanning three modalities: chat-based, audio-only, and multimodal interactions. Results show that while Talk-Moves achieved higher overall inter-rater reliability than the AI-human codebook (k = 0.74 vs. 0.64), the AI-human codebook demonstrated broader empirical coverage and higher perceived usability across modalities. Both codebooks undercaptured tutoring-relevant moves and introduced ambiguity when identifying actions expressed through nonverbal and multimodal artifacts. Together, these findings highlight the uneven generalizability of TalkMoves to tutoring contexts and motivate the development of modality-aware, tutoring-grounded codebooks.

cs.HC↗

arXiv2026

Modernizing Ground Truth: Four Shifts Toward Improving Reliability and Validity in AI in Education

Generative Artificial Intelligence (GenAI) is now widespread in education, yet the efficacy of GenAI systems remains constrained by the quality and interpretation of the labeled data used to train and evaluate them. Studies commonly report inter-rater reliability (IRR), often summarized by a single coefficient such as Cohen's kappa (k), as a gatekeeper to ``ground truth.'' We argue that many educational assessment and practice support settings include challenges, such as high-inference constructs, skewed label distributions, and temporally segmented multimodal data, which yield potential misapplication or misinterpretation of threshold-based heuristics for IRR. The growing use of large language models as annotators and judges introduces risks such as automation bias and circular validation. We propose four practical shifts for establishing ground truth: (1) treat IRR as a diagnostic signal to localize disagreement and refine constructs rather than a mechanical acceptance threshold (e.g., k > 0.8); (2) require transparent reporting of rater expertise, codebook development, reconciliation procedures, and segmentation rules; (3) mitigate risks in LLM annotation through bias audits and verification workflows; and (4) complement agreement statistics with validity and effectiveness evidence for the intended use, including uncertainty-aware labeling (e.g., assigning different labels to the same item to capture nuance), criterion-related checks (e.g., predictive tests to check if labels forecast the intended outcome), and close-the-loop evaluations of whether systems trained on these labels improve learning beyond a reasonable control. We illustrate these shifts through case studies of multimodal tutoring data and provide actionable recommendations toward strengthening the evidence base of labeled AIED datasets.

cs.CY↗

arXiv2025

Fundamental Mechanisms of Human Learning: Implications for Research and Practice in Education

Learning underlies nearly all human behavior and is central to education and education reform. Although recent advances in neuroscience have revealed the fundamental structure of learning processes, these insights have yet to be integrated into research and practice. Specifically, neuroscience has found that decision-making is governed by a structured process of perception, action-selection, and execution, supported by multiple neural systems with distinct memory stores and learning mechanisms. These systems extract different types of information (categorical, predictive, structural, and sequential) challenging canonical models of memory used in learning and behavioral science research by providing a mechanistic account of how humans acquire and use knowledge. Because each system learns differently, effective teaching requires alignment with system-specific processes. We propose a unified model that integrates these neuroscientific insights, bridging basic mechanisms with outcomes in education, identity, belonging, and wellbeing. By translating first principles of neural information processing into a generalizable framework, this work advances theories of skill acquisition and transfer while establishing a foundation for interdisciplinary research to refine how learning is understood and supported across domains of human behavior.

cs.IT↗

arXiv2025

Inspection-Guided Randomization: A Flexible and Transparent Restricted Randomization Framework for Better Experimental Design

Randomized experiments are considered the gold standard for estimating causal effects. However, out of the set of possible randomized assignments, some may be likely to produce poor effect estimates and misleading conclusions. Restricted randomization is an experimental design strategy that filters out undesirable treatment assignments, but its application has primarily been limited to ensuring covariate balance in two-arm studies where the target estimand is the average treatment effect. Other experimental settings with different design desiderata and target effect estimands could also stand to benefit from a restricted randomization approach. We introduce Inspection-Guided Randomization (IGR), a transparent and flexible framework for restricted randomization that filters out undesirable treatment assignments by inspecting assignments against analyst-specified, domain-informed design desiderata. In IGR, the acceptable treatment assignments are locked in ex ante and pre-registered in the trial protocol, thus safeguarding against $p$-hacking and promoting reproducibility. Through illustrative simulation studies motivated by education and behavioral health interventions, we demonstrate how IGR can be used to improve effect estimates compared to benchmark designs in group formation experiments and experiments with interference.

stat.ME↗

arXiv2024

The Life Cycle of Large Language Models: A Review of Biases in Education

Large Language Models (LLMs) are increasingly adopted in educational contexts to provide personalized support to students and teachers. The unprecedented capacity of LLM-based applications to understand and generate natural language can potentially improve instructional effectiveness and learning outcomes, but the integration of LLMs in education technology has renewed concerns over algorithmic bias which may exacerbate educational inequities. In this review, building on prior work on mapping the traditional machine learning life cycle, we provide a holistic map of the LLM life cycle from the initial development of LLMs to customizing pre-trained models for various applications in educational settings. We explain each step in the LLM life cycle and identify potential sources of bias that may arise in the context of education. We discuss why current measures of bias from traditional machine learning fail to transfer to LLM-generated content in education, such as tutoring conversations because the text is high-dimensional, there can be multiple correct responses, and tailoring responses may be pedagogically desirable rather than unfair. This review aims to clarify the complex nature of bias in LLM applications and provide practical guidance for their evaluation to promote educational equity.

cs.CY↗

arXiv2024

Discursive objection strategies in online comments: Developing a classification schema and validating its training

Most Americans agree that misinformation, hate speech and harassment are harmful and inadequately curbed on social media through current moderation practices. In this paper, we aim to understand the discursive strategies employed by people in response to harmful speech in news comments. We conducted a content analysis of more than 6500 comment replies to trending news videos on YouTube and Twitter and identified seven distinct discursive objection strategies (Study 1). We examined the frequency of each strategy's occurrence from the 6500 comment replies, as well as from a second sample of 2004 replies (Study 2). Together, these studies show that people deploy a diversity of discursive strategies when objecting to speech, and reputational attacks are the most common. The resulting classification scheme accounts for different theoretical approaches for expressing objections and offers a comprehensive perspective on grassroots efforts aimed at stopping offensive or problematic speech on campus.

cs.CL↗

arXiv2024

Cultural Differences in Students' Privacy Concerns in Learning Analytics across Germany, South Korea, Spain, Sweden, and the United States

Applications of learning analytics (LA) can raise concerns from students about their privacy in higher education contexts. Developing effective privacy-enhancing practices requires a systematic understanding of students' privacy concerns and how they vary across national and cultural dimensions. We conducted a survey study with established instruments to measure privacy concerns and cultural values for university students in five countries (Germany, South Korea, Spain, Sweden, and the United States; N = 762). The results show that students generally trusted institutions with their data and disclosed information as they perceived the risks to be manageable even though they felt somewhat limited in their ability to control their privacy. Across the five countries, German and Swedish students stood out as the most trusting and least concerned, especially compared to US students who reported greater perceived risk and less control. Students in South Korea and Spain responded similarly on all five privacy dimensions (perceived privacy risk, perceived privacy control, privacy concerns, trusting beliefs, and non-self-disclosure behavior), despite their significant cultural differences. Culture measured at the individual level affected the antecedents and outcomes of privacy concerns more than country-level culture. Perceived privacy risk and privacy control increase with power distance. Trusting beliefs increase with a desire for uncertainty avoidance and lower masculinity. Non-self-disclosure behaviors rise with power distance and masculinity, and decrease with more uncertainty avoidance. Thus, cultural values related to trust in institutions, social equality and risk-taking should be considered when developing privacy-enhancing practices and policies in higher education.

cs.CY↗

arXiv2024

What Explains Teachers' Trust of AI in Education across Six Countries?

With growing expectations to use AI-based educational technology (AI-EdTech) to improve students' learning outcomes and enrich teaching practice, teachers play a central role in the adoption of AI-EdTech in classrooms. Teachers' willingness to accept vulnerability by integrating technology into their everyday teaching practice, that is, their trust in AI-EdTech, will depend on how much they expect it to benefit them versus how many concerns it raises for them. In this study, we surveyed 508 K-12 teachers across six countries on four continents to understand which teacher characteristics shape teachers' trust in AI-EdTech, and its proposed antecedents, perceived benefits and concerns about AI-EdTech. We examined a comprehensive set of characteristics including demographic and professional characteristics (age, gender, subject, years of experience, etc.), cultural values (Hofstede's cultural dimensions), geographic locations (Brazil, Israel, Japan, Norway, Sweden, USA), and psychological factors (self-efficacy and understanding). Using multiple regression analysis, we found that teachers with higher AI-EdTech self-efficacy and AI understanding perceive more benefits, fewer concerns, and report more trust in AI-EdTech. We also found geographic and cultural differences in teachers' trust in AI-EdTech, but no demographic differences emerged based on their age, gender, or level of education. The findings provide a comprehensive, international account of factors associated with teachers' trust in AI-EdTech. Efforts to raise teachers' understanding of, and trust in AI-EdTech, while considering their cultural values are encouraged to support its adoption in K-12 education.

cs.CY↗

arXiv2023

Evaluating a Learned Admission-Prediction Model as a Replacement for Standardized Tests in College Admissions

A growing number of college applications has presented an annual challenge for college admissions in the United States. Admission offices have historically relied on standardized test scores to organize large applicant pools into viable subsets for review. However, this approach may be subject to bias in test scores and selection bias in test-taking with recent trends toward test-optional admission. We explore a machine learning-based approach to replace the role of standardized tests in subset generation while taking into account a wide range of factors extracted from student applications to support a more holistic review. We evaluate the approach on data from an undergraduate admission office at a selective US institution (13,248 applications). We find that a prediction model trained on past admission data outperforms an SAT-based heuristic and matches the demographic composition of the last admitted class. We discuss the risks and opportunities for how such a learned model could be leveraged to support human decision-making in college admissions.

cs.CY↗

arXiv2021

Should College Dropout Prediction Models Include Protected Attributes?

Early identification of college dropouts can provide tremendous value for improving student success and institutional effectiveness, and predictive analytics are increasingly used for this purpose. However, ethical concerns have emerged about whether including protected attributes in the prediction models discriminates against underrepresented student groups and exacerbates existing inequities. We examine this issue in the context of a large U.S. research university with both residential and fully online degree-seeking students. Based on comprehensive institutional records for this entire student population across multiple years, we build machine learning models to predict student dropout after one academic year of study, and compare the overall performance and fairness of model predictions with or without four protected attributes (gender, URM, first-generation student, and high financial need). We find that including protected attributes does not impact the overall prediction performance and it only marginally improves algorithmic fairness of predictions. While these findings suggest that including protected attributes is preferred, our analysis also offers guidance on how to evaluate the impact in a local context, where institutional stakeholders seek to leverage predictive analytics to support student success.

cs.CY↗

arXiv2021

Algorithmic Fairness in Education

Data-driven predictive models are increasingly used in education to support students, instructors, and administrators. However, there are concerns about the fairness of the predictions and uses of these algorithmic systems. In this introduction to algorithmic fairness in education, we draw parallels to prior literature on educational access, bias, and discrimination, and we examine core components of algorithmic systems (measurement, model learning, and action) to identify sources of bias and discrimination in the process of developing and deploying these systems. Statistical, similarity-based, and causal notions of fairness are reviewed and contrasted in the way they apply in educational contexts. Recommendations for policy makers and developers of educational technology offer guidance for how to promote algorithmic fairness in education.

cs.CY↗

arXiv2020

Evaluation of Fairness Trade-offs in Predicting Student Success

Predictive models for identifying at-risk students early can help teaching staff direct resources to better support them, but there is a growing concern about the fairness of algorithmic systems in education. Predictive models may inadvertently introduce bias in who receives support and thereby exacerbate existing inequities. We examine this issue by building a predictive model of student success based on university administrative records. We find that the model exhibits gender and racial bias in two out of three fairness measures considered. We then apply post-hoc adjustments to improve model fairness to highlight trade-offs between the three fairness measures.

cs.CY↗