[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-1-en-105":3,"doc-seo-255691-105":53,"doc-detail-255691-en":126},{"code":4,"msg":5,"data":6},0,"success",[7,14,19,24,29,34,39,44,49],{"id":8,"doc_module":9,"doc_module_name":10,"category_name":11,"show_sort_weight":12,"slug":13},11,1,"Template","Presentations",90,"presentations",{"id":15,"doc_module":9,"doc_module_name":10,"category_name":16,"show_sort_weight":17,"slug":18},12,"Resumes",80,"resumes",{"id":20,"doc_module":9,"doc_module_name":10,"category_name":21,"show_sort_weight":22,"slug":23},14,"Invoices",70,"invoices",{"id":25,"doc_module":9,"doc_module_name":10,"category_name":26,"show_sort_weight":27,"slug":28},15,"Posters",60,"posters",{"id":30,"doc_module":9,"doc_module_name":10,"category_name":31,"show_sort_weight":32,"slug":33},16,"Social Media",50,"social-media",{"id":35,"doc_module":9,"doc_module_name":10,"category_name":36,"show_sort_weight":37,"slug":38},17,"Forms",40,"forms",{"id":40,"doc_module":9,"doc_module_name":10,"category_name":41,"show_sort_weight":42,"slug":43},18,"Letters",30,"letters",{"id":45,"doc_module":9,"doc_module_name":10,"category_name":46,"show_sort_weight":47,"slug":48},21,"Paper Templates",5,"papers-templates",{"id":50,"doc_module":9,"doc_module_name":10,"category_name":51,"show_sort_weight":4,"slug":52},158,"General","general-158",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":119,"head_meta":121,"extra_data":123,"updated_unix":125},105,"en","how-students-really-use-chatgpt-uncovering-experiences-among-undergraduate-students","How students (really) use ChatGPT - Uncovering experiences among undergraduate students","","We examine how undergraduate students integrate ChatGPT into everyday self-directed learning. Naturalistic traces from 36 students (10,536 messages) are analyzed using a sequential mixed-methods pipeline with iterative qualitative coding and zero-shot language-model annotation validated against human labels. The study derives a five-category interaction taxonomy and connects time-lagged engagement patterns to specific interaction behaviors through regression and survival models.",{"@graph":63,"@context":118},[64,80,101],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,71,74,77],{"item":68,"name":69,"@type":70,"position":9},"https://docshare.wps.com","Home","ListItem",{"item":72,"name":10,"@type":70,"position":73},"https://docshare.wps.com/template/",2,{"item":75,"name":51,"@type":70,"position":76},"https://docshare.wps.com/template/general/",3,{"item":78,"name":59,"@type":70,"position":79},"https://docshare.wps.com/template/how-students-really-use-chatgpt-uncovering-experiences-among-undergraduate-students/255691/",4,{"url":78,"name":59,"@type":81,"image":82,"author":87,"headline":59,"publisher":90,"fileFormat":93,"inLanguage":57,"description":61,"dateModified":94,"datePublished":95,"encodingFormat":93,"isAccessibleForFree":96,"interactionStatistic":97},"DigitalDocument",{"url":83,"@type":84,"width":85,"height":86},"https://docshare.wps.com/thumbnails/how-students-really-use-chatgpt-uncovering-experiences-among-undergraduate-students/255691.png","ImageObject",442,249,{"name":88,"@type":89},"Theodora","Person",{"url":68,"name":91,"@type":92},"DocShare","Organization","application/pdf","2026-09-20","2026-09-13",true,{"@type":98,"interactionType":99,"userInteractionCount":73},"InteractionCounter",{"@type":100},"ViewAction",{"@type":102,"mainEntity":103},"FAQPage",[104,110,114],{"name":105,"@type":106,"acceptedAnswer":107},"How does the study analyze students’ real ChatGPT use?","Question",{"text":108,"@type":109},"It uses naturalistic logs from 36 undergraduates (10,536 messages) over Dec 2022–Jan 2024, combining iterative qualitative coding with zero-shot language-model annotation validated against human labels.","Answer",{"name":111,"@type":106,"acceptedAnswer":112},"What interaction categories does the paper identify?",{"text":113,"@type":109},"It produces a five-category taxonomy: Information Seeking, Content Generation, Language Use, Student–ChatGPT Interaction, and ChatGPT Response Behavior.",{"name":115,"@type":106,"acceptedAnswer":116},"Which system behavior predicts increased ongoing engagement?",{"text":117,"@type":109},"System-issued apologies are identified as the strongest positive predictor of returning the next week, leading to the mechanism called repair gratification.","https://schema.org",{"og:url":78,"og:type":120,"og:title":59,"og:site_name":91,"og:description":61},"article",{"robots":122,"canonical":78},"index,follow",{"doc_id":124,"site_id":56},255691,1789291975,{"code":4,"msg":5,"data":127},{"doc_id":124,"user_id":128,"nickname":88,"user_avatar":129,"doc_module":9,"category_id":50,"category_name":51,"doc_title":59,"doc_description":61,"doc_content":130,"file_id":131,"file_url":132,"file_type":133,"file_size":134,"view_count":73,"is_deleted":4,"is_public":9,"is_downloadable":9,"audit_status":9,"page_count":135,"language":136,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":137,"faqs":138,"seo_title":139,"seo_description":61,"update_tm":125,"read_time":20},687197207919,"https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552","Highlights  \nHow students (really) use ChatGPT: Uncovering experiences among undergraduate students  \nTawfiq Ammari, Meilun Chen, S M Mehedi Zaman, Kiran Garimella  \n• Naturalistic logs from 36 undergraduates (10,536 messages; Dec 2022–Jan 2024) .  \n• Five-category interaction taxonomy, reliably coded by humans and a language model.  \n• System apologies are the strongest positive predictor of returning the next week.  \n• Repair gratification: reward from a breakdown repaired, not a task completed.  \n• Every predictor of use and non-use maps onto expectancy violations.  \narXiv :2505 .24126v5 [ cs .HC] 7 Aug 2026  \nHow students (really) use ChatGPT: Uncovering experiences among undergraduate students  \nTawfiq Ammaria,∗ , Meilun Chena , S M Mehedi Zamana and Kiran Garimellaa  \na Rutgers University School of Communication and Information, 4 Huntington Street, New Brunswick, NJ, 08901, USA  \nARTICLE INFO  \nKeywords:  \nGenerative AI  \nSelf-directed learning  \nUses and Gratifications Theory Human–AI interaction Conversational repair Expectancy Violations Theory  \nAB STRACT  \nWe examine how undergraduate students integrate ChatGPT into everyday self-directed learning. We analyze 10,536 naturalistic messages donated by 36 students over more than a year. A sequential mixed-methods pipeline pairs iterative team-based qualitative coding with zero-shot language-model annotation validated against human labels (􀀔 = 0 .75–0.91) . It yields a five-category interaction taxonomy of Information Seeking, Content Generation, Language Use, Student–ChatGPT Interaction, and ChatGPT Response Behavior. Time-lagged linear regression and Cox proportional-hazards models link these categories to sustained engagement. Three findings stand out. First, structured, goal-driven tasks such as theory application, code writing, job-application content, and multiplechoice questions predict continued use, and ChatGPT becomes incorporated into students’ academic rhythms when gratifications are reliably fulfilled. Second, system-issued apologies are the single strongest positive predictor of increased engagement, outweighing every task-completion predictor. We name this mechanism repair gratification, the reward of a breakdown acknowledged and repaired rather than a task simply completed. Third, behaviors marking interactional strain, including prompt revision, frustration, and follow-up clarification, predict disengagement. When the cost of managing the system falls on the user without system-side accountability, students abandon the tool. We interpret these results through a triangulation of Self-Directed Learning, Uses and Gratifications Theory, and Expectancy Violations Theory, mapping every predictor onto the four conditions of positive violation, positive confirmation, negative violation, and negative confirmation. We close with design recommendations for graduated repair patterns, mode-aware interaction, and verification affordances, and we outline a participatory AI-literacy agenda for higher education.  \n1. Introduction  \nLarge language models and AI-driven chatbots such as ChatGPT are reshaping how learners search for information, plan assignments, and solicit feedback. We still lack a clear picture of how students weave the tool into academic routines spanning formal coursework and the informal, logistical practices through which they manage learning. Most of what we know comes from self-reports, single-course case studies, or public prompt repositories. What students actually ask ChatGPT to do, and how that use evolves across a semester, remain open questions (Bewersdorff, Hornberger, Nerdel, & Schiff, 2025; Sawalha, Taj, & Shoufan, 2024; Skjuve, Brandtzaeg, & Følstad, 2024) .  \nThis gap matters because institutions are already writing policies and building infrastructure around generative AI (Jin, Yan, Echeverria, Gašević, & Martinez-Maldonado, 2025) . OpenAI’s May 2024 release of ChatGPT Edu, developed with several U.S. universities, has intensified both ","cbCaivXEecRjUvx6","https://ap.wps.com/l/cbCaivXEecRjUvx6","pdf",1286485,41,"English","# Introduction\n## Interaction patterns and research gap\n## Methodological challenges","[{\"question\":\"How does the study analyze students’ real ChatGPT use?\",\"answer\":\"It uses naturalistic logs from 36 undergraduates (10,536 messages) over Dec 2022–Jan 2024, combining iterative qualitative coding with zero-shot language-model annotation validated against human labels.\"},{\"question\":\"What interaction categories does the paper identify?\",\"answer\":\"It produces a five-category taxonomy: Information Seeking, Content Generation, Language Use, Student–ChatGPT Interaction, and ChatGPT Response Behavior.\"},{\"question\":\"Which system behavior predicts increased ongoing engagement?\",\"answer\":\"System-issued apologies are identified as the strongest positive predictor of returning the next week, leading to the mechanism called repair gratification.\"}]","How students (really) use ChatGPT - Uncovering experiences among undergraduate students | PDF"]