Zhicheng Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhicheng Lin has a total of 394 co-authors. Here are the top co-authors by number of shared publications:
Scott O. Murray
Department of Psychology, University of Washington
5 shared publicationsSheng He
Department of Psychology, University of Minnesota, Twin Cities, MN, USA
4 shared publications- Yang Zhang
Department of Psychology, Soochow University
4 shared publications Zhong-Lin Lu
Department of Psychology, The Ohio State University, Columbus, USA
3 shared publications- Balazs Aczel
Institute of Psychology, Eötvös Loránd University, Budapest, Hungary
2 shared publications - Don van Ravenzwaaij
Faculty of Behavioural and Social Sciences, University of Groningen, Groningen, The Netherlands
2 shared publications
Latest Publications
- Open Access
- 16 October 2026
- Published by SAGE Publications in Advances in Methods and Practices in Psychological Science
- Vol. 9 (4)
Restructuring Scientific Papers for Human and AI Readers
Scientific communication faces a dual crisis: Exponential publication growth—now accelerated by writing assisted by artificial intelligence (AI)—overwhelms human readers and reviewers, while fragmented evidence impedes automated synthesis. This fragmentation is especially acute in the behavioral and social sciences, where poorly standardized stimulus sets and jingle-jangle measurement fallacies (identical labels for distinct constructs; different labels for the same construct) hinder comparison, while underreported context conceals effect heterogeneity. Current AI tools can summarize papers but cannot reliably synthesize findings across incommensurable studies; they also risk perpetuating biases when trained on unstructured, unverified text. I propose restructuring scientific papers for two audiences: front-loaded narratives for time-pressed human readers, paired with research-object packages containing executable environments, semantic annotations, and tidy trial-level data. This design makes papers queryable: Readers can interrogate data and probe analytic choices, while the packages ground automated verification and AI-assisted manuscript evaluation—before submission or during peer review—in inspectable data and executable analyses rather than narrative claims alone. Such papers could become nodes in continuously updated evidence networks, automatically contributing effect sizes to living, versioned meta-analyses and allowing corrections and retractions to propagate through dependent syntheses. Widespread adoption will require institutions to credit research-object packages as scholarly contributions and invest in the open standards, infrastructure, training, and governance needed for their creation and long-term stewardship.
- 16 October 2026
- Published by SAGE Publications in Advances in Methods and Practices in Psychological Science
- Vol. 9 (4)
Open AccessRestructuring Scientific Papers for Human and AI Readers
Scientific communication faces a dual crisis: Exponential publication growth—now accelerated by writing assisted by artificial intelligence (AI)—overwhelms human readers and reviewers, while fragmented evidence impedes automated synthesis. This fragmentation is especially acute in the behavioral and social sciences, where poorly standardized stimulus sets and jingle-jangle measurement fallacies (identical labels for distinct constructs; different labels for the same construct) hinder comparison, while underreported context conceals effect heterogeneity. Current AI tools can summarize papers but cannot reliably synthesize findings across incommensurable studies; they also risk perpetuating biases when trained on unstructured, unverified text. I propose restructuring scientific papers for two audiences: front-loaded narratives for time-pressed human readers, paired with research-object packages containing executable environments, semantic annotations, and tidy trial-level data. This design makes papers queryable: Readers can interrogate data and probe analytic choices, while the packages ground automated verification and AI-assisted manuscript evaluation—before submission or during peer review—in inspectable data and executable analyses rather than narrative claims alone. Such papers could become nodes in continuously updated evidence networks, automatically contributing effect sizes to living, versioned meta-analyses and allowing corrections and retractions to propagate through dependent syntheses. Widespread adoption will require institutions to credit research-object packages as scholarly contributions and invest in the open standards, infrastructure, training, and governance needed for their creation and long-term stewardship.
- Open Access
- 15 October 2026
- Published by Center for Open Science in PsyArXiv
Registered Reports as the default pathway for confirmatory publication
By rewarding speed, volume, and coherent positive narratives, traditional publishing incentivizes questionable scientific practices and analytic flexibility. The latter especially contributes to blurring the distinction between exploration and confirmation, which inflates the apparent strength of confirmatory claims. This systemic issue has gone unchecked for decades, leaving the scientific community with a body of scholarly literature that is unreliable and lacking validity. Registered Reports (RRs) address problematic publishing practices by requiring a priori registration of confirmatory tests, which is peer-reviewed before the research begins (stage 1). The editorial acceptance of a stage 1 plan establishes a commitment to publication regardless of research outcomes. Publication remains, nonetheless, contingent on a second peer review and editorial scrutiny (stage 2) to verify that the approved plan was followed, all deviations were transparently disclosed and justified, confirmatory and exploratory claims are clearly delineated, and the integrity of the design was maintained. Meta-scientific comparisons indicate that RRs reduce outcome-based publication bias and improve methodological rigor without compromising perceived novelty. By making accepted RR protocols and their review history citable research outputs, RRs also strengthen recognition for peer-review work. In this perspective, we argue that RRs should become the primary mode of confirmatory publication moving forward.
- 15 October 2026
- Published by Center for Open Science in PsyArXiv
Open AccessRegistered Reports as the default pathway for confirmatory publication
By rewarding speed, volume, and coherent positive narratives, traditional publishing incentivizes questionable scientific practices and analytic flexibility. The latter especially contributes to blurring the distinction between exploration and confirmation, which inflates the apparent strength of confirmatory claims. This systemic issue has gone unchecked for decades, leaving the scientific community with a body of scholarly literature that is unreliable and lacking validity. Registered Reports (RRs) address problematic publishing practices by requiring a priori registration of confirmatory tests, which is peer-reviewed before the research begins (stage 1). The editorial acceptance of a stage 1 plan establishes a commitment to publication regardless of research outcomes. Publication remains, nonetheless, contingent on a second peer review and editorial scrutiny (stage 2) to verify that the approved plan was followed, all deviations were transparently disclosed and justified, confirmatory and exploratory claims are clearly delineated, and the integrity of the design was maintained. Meta-scientific comparisons indicate that RRs reduce outcome-based publication bias and improve methodological rigor without compromising perceived novelty. By making accepted RR protocols and their review history citable research outputs, RRs also strengthen recognition for peer-review work. In this perspective, we argue that RRs should become the primary mode of confirmatory publication moving forward.
- Open Access
- 8 October 2026
- Published by Springer Nature in npj Digital Medicine
Governing cognitive labor delegated to AI in scholarly writing and medical education
Across scholarly communication and medical education, AI-assisted writing redistributes cognitive labor. Mapping publishing policies from 2023 through 11 August 2026 finds governance centered on permitted uses, human authorship and accountability, disclosure, verification, and confidentiality. Five sociotechnical domains—agency and development, epistemic quality, fairness, rights and governance, and economy and culture—inform tiered oversight of editing, drafting, surrogate reading, and methodological use, with safeguards for learning and assessment.
- 8 October 2026
- Published by Springer Nature in npj Digital Medicine
Open AccessGoverning cognitive labor delegated to AI in scholarly writing and medical education
Across scholarly communication and medical education, AI-assisted writing redistributes cognitive labor. Mapping publishing policies from 2023 through 11 August 2026 finds governance centered on permitted uses, human authorship and accountability, disclosure, verification, and confidentiality. Five sociotechnical domains—agency and development, epistemic quality, fairness, rights and governance, and economy and culture—inform tiered oversight of editing, drafting, surrogate reading, and methodological use, with safeguards for learning and assessment.
- 29 September 2026
- Published by Springer Nature in Nature
- Vol. 658 (8134) , 32-33
AI can widen science — but only if institutions stop rewarding the already measurable
- 29 September 2026
- Published by Springer Nature in Nature
- Vol. 658 (8134) , 32-33
AI can widen science — but only if institutions stop rewarding the already measurable
- 10 August 2026
- Published by Annual Reviews in Annual Review of Psychology
From Prompts to Constructs: A Dual-Validity Framework for Large Language Model Research in Psychology
Large language models (LLMs) are entering psychological research both as tools and as objects of inquiry. Yet many studies apply human instruments to LLMs without establishing that the outputs are reliable or interpretable, raising the risk of measurement phantoms—statistical regularities mistaken for genuine psychological phenomena. This review argues that robust AI psychological research requires integrating two methodological traditions: psychometric validation of what a score means and causal inference standards for what the results warrant. It develops a dual-validity framework in which evidentiary demands scale with scientific ambition: from tool use through behavioral characterization and human simulation to cognitive modeling. Classifying text may require only accuracy and reliability; claiming that an LLM simulates anxiety or illuminates cognitive mechanisms requires additional evidence, including construct validity evidence and experimental controls. Progress depends on developing computational analogs of psychological constructs rather than assuming human measures automatically apply to language models.
- 10 August 2026
- Published by Annual Reviews in Annual Review of Psychology
From Prompts to Constructs: A Dual-Validity Framework for Large Language Model Research in Psychology
Large language models (LLMs) are entering psychological research both as tools and as objects of inquiry. Yet many studies apply human instruments to LLMs without establishing that the outputs are reliable or interpretable, raising the risk of measurement phantoms—statistical regularities mistaken for genuine psychological phenomena. This review argues that robust AI psychological research requires integrating two methodological traditions: psychometric validation of what a score means and causal inference standards for what the results warrant. It develops a dual-validity framework in which evidentiary demands scale with scientific ambition: from tool use through behavioral characterization and human simulation to cognitive modeling. Classifying text may require only accuracy and reliability; claiming that an LLM simulates anxiety or illuminates cognitive mechanisms requires additional evidence, including construct validity evidence and experimental controls. Progress depends on developing computational analogs of psychological constructs rather than assuming human measures automatically apply to language models.
- Open Access
- 9 July 2026
- Published by SAGE Publications in Advances in Methods and Practices in Psychological Science
- Vol. 9 (3)
Artificial-Intelligence-Mediated Contamination in Online Research: Taxonomy, Risk Gradient, and Recommendations
For 2 decades, online research has relied on a quality heuristic: Careful, coherent responding is good data. That heuristic is no longer reliable. Autonomous artificial-intelligence (AI) agents can now pass nearly all conventional quality checks, and in text-rich crowd-work tasks, reported use of large language models approaches one third. When such consultation shapes the response process itself—not just its surface expression—the resulting data appear human-generated while embedding systematic, model-shaped distortions. I synthesize emerging evidence on how AI-mediated contamination varies across research settings in prevalence, mechanism, and inferential consequence; and distinguish three contamination pathways (full delegation, partial mediation, and spillover) and three vulnerability zones (text-rich tasks at highest risk, browser-based cognitive paradigms as an emerging vulnerability, and supervised or identity-vetted settings at lower risk). Even modest contamination can shift estimated public opinion, compress attitudinal extremes, and, over time, feed back into the training data for future models. Current platform countermeasures may raise the cost of contamination but have not been independently validated under adversarial conditions. I argue for a shift from ad hoc detection to infrastructure redesign: contamination-aware sensitivity analyses, explicit stratification of data collection by evidential role, transparency norms that balance open science with adversarial robustness, and a minimum reporting checklist for online studies in vulnerable settings. I close by asking when AI mediation should be treated not as contamination but as part of the ecological baseline of human responding—a question that requires the field to specify the target cognitive system in any given study.
- 9 July 2026
- Published by SAGE Publications in Advances in Methods and Practices in Psychological Science
- Vol. 9 (3)
Open AccessArtificial-Intelligence-Mediated Contamination in Online Research: Taxonomy, Risk Gradient, and Recommendations
For 2 decades, online research has relied on a quality heuristic: Careful, coherent responding is good data. That heuristic is no longer reliable. Autonomous artificial-intelligence (AI) agents can now pass nearly all conventional quality checks, and in text-rich crowd-work tasks, reported use of large language models approaches one third. When such consultation shapes the response process itself—not just its surface expression—the resulting data appear human-generated while embedding systematic, model-shaped distortions. I synthesize emerging evidence on how AI-mediated contamination varies across research settings in prevalence, mechanism, and inferential consequence; and distinguish three contamination pathways (full delegation, partial mediation, and spillover) and three vulnerability zones (text-rich tasks at highest risk, browser-based cognitive paradigms as an emerging vulnerability, and supervised or identity-vetted settings at lower risk). Even modest contamination can shift estimated public opinion, compress attitudinal extremes, and, over time, feed back into the training data for future models. Current platform countermeasures may raise the cost of contamination but have not been independently validated under adversarial conditions. I argue for a shift from ad hoc detection to infrastructure redesign: contamination-aware sensitivity analyses, explicit stratification of data collection by evidential role, transparency norms that balance open science with adversarial robustness, and a minimum reporting checklist for online studies in vulnerable settings. I close by asking when AI mediation should be treated not as contamination but as part of the ecological baseline of human responding—a question that requires the field to specify the target cognitive system in any given study.
- 1 July 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (8) , 216
A validity-guided workflow for robust large language model research in psychology
Large language models (LLMs) are rapidly being integrated into psychological and behavioral research as research tools, evaluation targets, human simulator
- 1 July 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (8) , 216
A validity-guided workflow for robust large language model research in psychology
Large language models (LLMs) are rapidly being integrated into psychological and behavioral research as research tools, evaluation targets, human simulator
- Open Access
- 29 June 2026
- Published by Oxford University Press (OUP) in PNAS Nexus
- Vol. 5 (7) , pgag236
Six misconceptions about large language models: A minimal model and diagnostic taxonomy
Large language models (LLMs) are now embedded in scientific, educational, and governance workflows, with debates centering on their capabilities, mechanisms, and impacts. Yet these debates remain structured by persistent folk theories—intuitive, informal explanatory models that guide attitudes and actions. Deflationary slogans (“just autocomplete,” “stochastic parrots,” and “average of the internet”) and anthropomorphic framings (“emergent agents” and “proto-minds”) each capture genuine features of current systems but mistake those features for the whole. This Perspective proposes a minimal working model of LLM-based systems centered on four distinctions: between pretraining and deployed systems; between the learned distribution and particular samples; among parametric, contextual, and external memory; and between task competence and agency. The model is used to diagnose six misconceptions about LLMs: next-token prediction, regression to the mean, training-data regurgitation, model memory, alignment, and understanding. For each, the analysis identifies what the misconception gets right, which distinctions it conflates, and what follows for capability evaluation, system design, and governance. Applied to publisher AI policies as governance case studies, the framework shows both how policy language can conflate these distinctions and how such errors can be corrected. The model thereby avoids the parrot–mind binary by treating LLMs as simulators of discourse and task performance, offering a diagnostic toolkit for locating and correcting the errors these folk theories perpetuate.
- 29 June 2026
- Published by Oxford University Press (OUP) in PNAS Nexus
- Vol. 5 (7) , pgag236
Open AccessSix misconceptions about large language models: A minimal model and diagnostic taxonomy
Large language models (LLMs) are now embedded in scientific, educational, and governance workflows, with debates centering on their capabilities, mechanisms, and impacts. Yet these debates remain structured by persistent folk theories—intuitive, informal explanatory models that guide attitudes and actions. Deflationary slogans (“just autocomplete,” “stochastic parrots,” and “average of the internet”) and anthropomorphic framings (“emergent agents” and “proto-minds”) each capture genuine features of current systems but mistake those features for the whole. This Perspective proposes a minimal working model of LLM-based systems centered on four distinctions: between pretraining and deployed systems; between the learned distribution and particular samples; among parametric, contextual, and external memory; and between task competence and agency. The model is used to diagnose six misconceptions about LLMs: next-token prediction, regression to the mean, training-data regurgitation, model memory, alignment, and understanding. For each, the analysis identifies what the misconception gets right, which distinctions it conflates, and what follows for capability evaluation, system design, and governance. Applied to publisher AI policies as governance case studies, the framework shows both how policy language can conflate these distinctions and how such errors can be corrected. The model thereby avoids the parrot–mind binary by treating LLMs as simulators of discourse and task performance, offering a diagnostic toolkit for locating and correcting the errors these folk theories perpetuate.
- Open Access
- 25 June 2026
- Published by Association for Computing Machinery (ACM) in Communications of the ACM
Hidden Prompts in Manuscripts Exploit AI-Assisted Peer Review
Highlighting the need for controlled AI integration in peer review and harmonized policies governing AI use in academic evaluation.
- 25 June 2026
- Published by Association for Computing Machinery (ACM) in Communications of the ACM
Open AccessHidden Prompts in Manuscripts Exploit AI-Assisted Peer Review
Highlighting the need for controlled AI integration in peer review and harmonized policies governing AI use in academic evaluation.
- 18 June 2026
- Published by Springer Nature in Nature Reviews Psychology
- Vol. 5 (8) , 511-512
Metacognition can mitigate AI-driven homogenization of ideas
- 18 June 2026
- Published by Springer Nature in Nature Reviews Psychology
- Vol. 5 (8) , 511-512
Metacognition can mitigate AI-driven homogenization of ideas