[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-315027-105":59,"doc-detail-315027-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","what-exam-scores-can-and-cannot-prove-about-unauthorized-ai-assistance-evidence-from-a-highly-public-classroom-episode","What exam scores can and cannot prove about unauthorized AI assistance: Evidence from a highly public classroom episode","","Spring 2026 data from an economics course at Brown University evaluates what released exam scores can and cannot establish about unauthorized generative AI use. Among 59 students, average midterm scores dropped from 95.7/100 to 48.8 after the final became proctored. Competing accounts include anxiety, harder exams, withdrawals, and regression to the mean. Correlation between midterm and final scores is 0.06, and permutation tests show no detectable association, consistent with a two-component mixture where midterm scores carry no information about final performance.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":20,"@type":76,"position":81},"https://docshare.wps.com/document/exam/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/what-exam-scores-can-and-cannot-prove-about-unauthorized-ai-assistance-evidence-from-a-highly-public-classroom-episode/315027/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/what-exam-scores-can-and-cannot-prove-about-unauthorized-ai-assistance-evidence-from-a-highly-public-classroom-episode/315027.png","ImageObject",300,407,{"name":92,"@type":93},"wps_ap_test_251126_0180","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-01","2026-09-20",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",11,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What event motivated the study of exam scores?","Question",{"text":113,"@type":114},"After the professor announced the final would be proctored and would void the midterm if distributions did not match, 27 students dropped or skipped. The 59 remaining students averaged about 49 on the final, far below reported final-exam means, which the professor interpreted as evidence of unauthorized AI use.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"What do the released scores show about the relationship between midterm and final performance?",{"text":118,"@type":114},"The correlation between a student’s midterm and final scores is only 0.06. Individual changes range widely, from small gains to severe losses.",{"name":120,"@type":111,"acceptedAnswer":121},"How do the authors evaluate whether the data require unauthorized assistance explanations?",{"text":122,"@type":114},"They run tests including a permutation test for any association and then fit a two-component mixture model contrasting coupled (same underlying proficiency) versus decoupled (independent scores) pairs. Models that treat midterm information as irrelevant describe the data as well as models permitting an association.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},315027,1789902654,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":19,"category_name":20,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":145,"preview_page_count":4},8796095027276,"https://avatar.qwps.com/avatar/d3BzX2FwX3Rlc3RfMjUxMTI2XzAxODA=","arXiv :2607 .27978v1 [ stat .AP] 30 Jul 2026  \nWhat exam scores can and cannot prove about unauthorized AI assistance: Evidence from a highly public classroom episode  \nChad M. Topaz 1,2,3,* and Utsav Bahl4  \n1Williams College, Williamstown, MA 01267, USA  \n2 QSIDE Institute, Williamstown, MA 01267, USA  \n3 University of Colorado, Boulder, CO 80309, USA  \n4 University of Cambridge, Cambridge CB2 1TN, UK  \n* Corresponding author: [cmt6@williams.edu](cmt6@williams.edu)  \nJuly 31, 2026  \nAbstract  \nIn spring 2026, an economics professor at Brown University gave a take-home midterm and, after unusually high scores, made the final exam proctored. Among the 59 students who completed the course, average scores fell from 95.7 out of 100 to 48.8 . The instructor attributed the drop to unauthorized use of generative AI on the midterm; others proposed test anxiety, a harder final, student withdrawals, and regression to the mean. The publicly released scores, one midterm–final pair per student, show two striking patterns: the correlation between a student’s two scores is only 0.06, and individual changes range from a 4-point gain to a 100-point loss. If both exams measured the same proficiency, students who scored higher on the midterm should generally have scored higher on the final, and most scores should have fallen by similar amounts. Permutation tests find no statistically detectable association between students’ midterm and final scores, yet would detect association of the strength the alternative explanations predict more than 90 percent of the time. Once model complexity is accounted for, a model in which a student’s midterm carries no information about that student’s final describes the data as well as any model that permits an association. Neither result rules out a weak association, but together they show that the data do not require any. In simulated classes where each student’s two scores remain linked through that student’s own proficiency, as the alternative explanations imply, the two patterns almost never appear together; they appear together regularly only when that link is nearly severed. More than one mechanism could have severed it: midterm answers that were not the students’ own work would have done so, and so would a final that tested substantially different knowledge and skills, with no assistance involved. The scores cannot distinguish these possibilities or identify which students, if any, used unauthorized assistance. Our study offers a framework for evaluating statistical evidence in future disputes over generative AI use in academic assessment.  \nKeywords: Academic integrity, Generative artificial intelligence, Academic misconduct, Educational measurement, Exam scores, Mixture model, Permutation test, Statistical evidence  \n1 Introduction  \nIn the spring of 2026, Roberto Serrano, a professor of economics at Brown University, administered a take-home, closed-book midterm to his class of 86 students and received back a set of papers averaging 96 out of 100, with roughly 40 perfect scores (Serrano, 2026; Gioino, 2026) . His previous midterms, which he describes as easier, had averaged between 65 and 80 (Serrano, 2026) . He then announced that the final examination would instead be proctored in person, and that he would void the midterm if the two grade distributions failed to match. Following the announcement, 27 of the 86 students either dropped the course or skipped the final, and the 59 students who remained averaged approximately 49, below any final-exam mean he reports having recorded in the course (Inside Higher Ed, 2026a; Goldstein, 2026) . He concluded that at least 50 students had made unauthorized use of generative artificial intelligence (AI) on the midterm, and said so publicly (Serrano, 2026; Goldstein, 2026) .  \nThe episode became national news, and public reaction divided between two interpretations. On one, the 47-point collapse was decisive evidence of mass cheating. On the other, advanced mainly","cbCaioH4ZN8FqxnN","https://ap.wps.com/l/cbCaioH4ZN8FqxnN","pdf",758360,48,"English","# Abstract\n# 1 Introduction","[{\"question\":\"What event motivated the study of exam scores?\",\"answer\":\"After the professor announced the final would be proctored and would void the midterm if distributions did not match, 27 students dropped or skipped. The 59 remaining students averaged about 49 on the final, far below reported final-exam means, which the professor interpreted as evidence of unauthorized AI use.\"},{\"question\":\"What do the released scores show about the relationship between midterm and final performance?\",\"answer\":\"The correlation between a student’s midterm and final scores is only 0.06. Individual changes range widely, from small gains to severe losses.\"},{\"question\":\"How do the authors evaluate whether the data require unauthorized assistance explanations?\",\"answer\":\"They run tests including a permutation test for any association and then fit a two-component mixture model contrasting coupled (same underlying proficiency) versus decoupled (independent scores) pairs. Models that treat midterm information as irrelevant describe the data as well as models permitting an association.\"}]","What exam scores can and cannot prove about unauthorized AI assistance: Evidence from a highly public classroom episode | PDF",121]