Claims: L 2

A 1 · A 2 · A 3 · A 4 · A 5 · A 6 · A 7 · A 8 · B 1 · B 2 · C 1 · C 2 · C 3 · C 4 · C 5 · C 6 · D 1 · D 2 · D 3 · E 1 · E 2 · E 3 · E 4 · F 1 · F 2 · F 3 · G 1 · G 2 · H 1 · H 2 · I 1 · I 2 · I 3 · I 4 · I 5 · J · K · L 1 · L 2 · L 3 · M 1 · M 2 · M 3 · M 4 · N 1 · N 2 · O 1 · O 2 · P 1 · P 2 · P 3 · P 4 · P 5 · Q · R 1 · R 2 · R 3 · S 1 · S 2 · S 3 · S 4 · S 5 · S 6 · S 7 · S 8 · T 1 · T 2 · T 3 · T 4 · T 5 · T 6 · T 7 · T 8 · T 9 · T 10 · U · V · W 1 · W 2 · X · Y · Z · 0-9

• Learning Quality & Cognitive Dependency discourse is heavily faculty-driven (Fac% 82.3), concentrated in K-12 and professor communities - Learning Quality & Cognitive Dependency discourse is heavily faculty-driven (Fac% 82.3), concentrated in K-12 and professor communities
• Learning readiness is the WEC component most strongly associated with academic outcomes (0.1–0.15 SD) - Learning readiness is the WEC component most strongly associated with academic outcomes (0.1–0.15 SD)
• Learning relational categories becomes more difficult as the complexity of the knowledge type increases - Learning relational categories becomes more difficult as the complexity of the knowledge type increases
• Learning should be assessed with retention and transfer tests rather than current practice performance - Learning should be assessed with retention and transfer tests rather than current practice performance
• Learning Space Redesign Alone Insufficient
• Learning strategies instruction produced no significantly different gains in word reading subskills or oral reading fluency measures than typical adult education instruction - Learning strategies instruction produced no significantly different gains in word reading subskills or oral reading fluency measures than typical adult education instruction
• Learning strategies relate to test anxiety: persistence positively, environmental structuring negatively, time management not significantly, and women report higher test anxiety - Learning strategies relate to test anxiety: persistence positively, environmental structuring negatively, time management not significantly, and women report higher test anxiety
• Learning strategy outcome effects are driven by different differentiation groups than the performance effect - Learning strategy outcome effects are driven by different differentiation groups than the performance effect
• Learning Studio participation increased students' comfort with, and skills around, the Studio technologies - Learning Studio participation increased students' comfort with, and skills around, the Studio technologies
• Learning style (visual-verbal) shows no significant correlation with the effectiveness of pictorial elucidation for adolescent EFL learners - Learning style (visual-verbal) shows no significant correlation with the effectiveness of pictorial elucidation for adolescent EFL learners
• Learning styles have been used in education for classifying students by their preferences for taking in, processing, and demonstrating information - Learning styles have been used in education for classifying students by their preferences for taking in, processing, and demonstrating information
• Learning styles impact teachers, pedagogy, student engagement, and assessment in K-12 education - Learning styles impact teachers, pedagogy, student engagement, and assessment in K-12 education
• Learning Styles Matching Does Not Improve Learning
• Learning styles measured by Kolb's inventory showed no significant change over one year of varied instruction - Learning styles measured by Kolb's inventory showed no significant change over one year of varied instruction
• Learning techniques have much smaller effects on deep learning and far transfer than on surface learning and near transfer - Effects were d = 0.60 on surface outcomes (1,473 effects) against 0.26 on deep outcomes (109 effects), and 0.61 on near transfer against 0.39 on far transfer; most studies measured surface or factual
• Learning the CVS with self-generated, vicarious, combined, or no errors during PS-I does not differentially affect CVS posttest performance when controlling for prior knowledge - Learning the CVS with self-generated, vicarious, combined, or no errors during PS-I does not differentially affect CVS posttest performance when controlling for prior knowledge
• Learning through the experience of solving problems can build both content knowledge and thinking strategies - Learning through the experience of solving problems can build both content knowledge and thinking strategies
• Learning time and recall time are negatively correlated (r = -.380) in individual serial learning of words and figures - Learning time and recall time are negatively correlated (r = -.380) in individual serial learning of words and figures
• Learning timing produces Group X Time interactions in both memory systems: declarative consolidation time courses differ between afternoon and evening groups, and finger-tapping speed and accuracy show highly significant Group X Time interactions - Learning timing produces Group X Time interactions in both memory systems: declarative consolidation time courses differ between afternoon and evening groups, and finger-tapping speed and accuracy sho
• Learning varied tasks of the same type enables transfer to unencountered tasks of that type - Learning varied tasks of the same type enables transfer to unencountered tasks of that type
• Learning with a ChatGPT-based flipped learning guiding approach significantly improves students’ performance, self-efficacy, learning attitudes, intrinsic motivation and creative thinking in a courseware project - Learning with a ChatGPT-based flipped learning guiding approach significantly improves students’ performance, self-efficacy, learning attitudes, intrinsic motivation and creative thinking in a coursew
• Learning with collaboration scripts outdoors yields biology knowledge gains in both iterations, with joint regulation activities mediating script-component effects - Learning with collaboration scripts outdoors yields biology knowledge gains in both iterations, with joint regulation activities mediating script-component effects
• Learning with CSCL scripts yields a large positive effect on collaboration skills compared to unstructured CSCL - Learning with CSCL scripts yields a large positive effect on collaboration skills compared to unstructured CSCL
• Learning with erroneous examples often imposed higher cognitive load than correct examples or problem-solving - Learning with erroneous examples often imposed higher cognitive load than correct examples or problem-solving
• Learning with intelligent tutoring systems yields a moderate, statistically significant overall advantage in achievement over non-ITS modes of instruction - Learning with intelligent tutoring systems yields a moderate, statistically significant overall advantage in achievement over non-ITS modes of instruction
• Learning with more than two external representations in STEM has advantageous effects on students' cognitive load and performance compared to learning with two representations. - Learning with more than two external representations in STEM has advantageous effects on students' cognitive load and performance compared to learning with two representations.
• Learning with more than two external representations shows no notable differences in learning time compared to learning with two representations. - Learning with more than two external representations shows no notable differences in learning time compared to learning with two representations.
• Learning with multimedia pedagogical agents is more effective than learning without these agents, with a small overall effect (g+ = 0.20) - Learning with multimedia pedagogical agents is more effective than learning without these agents, with a small overall effect (g+ = 0.20)
• Learning with multiple representations imposes high demands and often fails to produce expected results because learners do not integrate the representations - Learning with multiple representations imposes high demands and often fails to produce expected results because learners do not integrate the representations
• Learning with tree diagrams increased experienced cognitive load across intrinsic, extraneous, germane, and overall measures - Learning with tree diagrams increased experienced cognitive load across intrinsic, extraneous, germane, and overall measures
• Learning-by-design shows medium- to large-sized positive effects on student achievement in K-12 STEM education, per meta-analysis cited by the review - Learning-by-design shows medium- to large-sized positive effects on student achievement in K-12 STEM education, per meta-analysis cited by the review
• Learning-by-teaching an LLM improved assignment and project performance versus historical cohorts, but not exam performance, in an undergraduate CS course - Learning-by-teaching an LLM improved assignment and project performance versus historical cohorts, but not exam performance, in an undergraduate CS course
• Learning-curve analysis of tutor logs revealed a flaw in the algebra cognitive model's handling of equations of the form ax = b when the coefficient is 1, and the tutor now provides targeted extra practice on that skill - Learning-curve analysis of tutor logs revealed a flaw in the algebra cognitive model's handling of equations of the form ax = b when the coefficient is 1, and the tutor now provides targeted extra pra
• Learning-curve plateaus were caused by learners' failure to overcome specific difficulties at critical stages, not by the order in which habits developed - Learning-curve plateaus were caused by learners' failure to overcome specific difficulties at critical stages, not by the order in which habits developed
• Learning-oriented engagement is recurrent but selective in everyday human–LLM conversations, with cognitive engagement in 31.9% and constructive engagement in 4.9% of user turns - Learning-oriented engagement is recurrent but selective in everyday human–LLM conversations, with cognitive engagement in 31.9% and constructive engagement in 4.9% of user turns
• Learning-oriented intentions did not significantly predict enacted in-chat behavior; only selecting final solutions predicted executive and answer-seeking requests - Learning-oriented intentions did not significantly predict enacted in-chat behavior; only selecting final solutions predicted executive and answer-seeking requests
• Learning-oriented video groups differ significantly from the productivity-oriented group on the primary ENA axis, with large effect sizes - Learning-oriented video groups differ significantly from the productivity-oriented group on the primary ENA axis, with large effect sizes
• Learning-outcome effects of open- versus closed-book generative tasks are mixed - Learning-outcome effects of open- versus closed-book generative tasks are mixed
• Learning-pattern differentiation and integration increase with educational progression within one type of education, while context change induces temporary diffuse patterns - Learning-pattern differentiation and integration increase with educational progression within one type of education, while context change induces temporary diffuse patterns
• Learning-process and psychological constructs correlate only weakly with aggregate final artistic performance - Learning-process and psychological constructs correlate only weakly with aggregate final artistic performance
• Learning-stories research indicates meaningful learning involves felt ownership of knowledge, emotional integration, generativity, real-life context, interaction with others' learning, and transformed self-perceptions - Learning-stories research indicates meaningful learning involves felt ownership of knowledge, emotional integration, generativity, real-life context, interaction with others' learning, and transformed
• Learning-technique effects do not shrink across test delays from the same day to six months - Effects were d = 0.58 under a day, 0.59 from a day to a week, 0.56 from a week to a month and 0.51 from one to six months, with no overall differences; 68% of effects were measured within a day.
• LearnLens reduces teacher feedback time, with half of teachers previously spending 10–30 minutes per assignment versus a median below 5 minutes with the system - LearnLens reduces teacher feedback time, with half of teachers previously spending 10–30 minutes per assignment versus a median below 5 minutes with the system
• LearnLens's modular pipeline outperforms direct-prompt baselines on scoring quality while matching latency and cost on 100 authentic student answers - LearnLens's modular pipeline outperforms direct-prompt baselines on scoring quality while matching latency and cost on 100 authentic student answers
• LearnLM receives the highest expert pedagogy-rubric marks across rubric categories, with the largest leads in active learning, metacognition, and curiosity - LearnLM receives the highest expert pedagogy-rubric marks across rubric categories, with the largest leads in active learning, metacognition, and curiosity
• LearnMate2 improved learning outcomes and user experience compared to a baseline combining an online learning platform and an LLM in a user study (n=16) - LearnMate2 improved learning outcomes and user experience compared to a baseline combining an online learning platform and an LLM in a user study (n=16)
• LEAs expressed broad interest in technical assistance, with the highest interest in professional learning opportunities - LEAs expressed broad interest in technical assistance, with the highest interest in professional learning opportunities
• LEAs most frequently used short-term behavioral health funds to expand internal staffing and direct services rather than partnerships and structural investments - LEAs most frequently used short-term behavioral health funds to expand internal staffing and direct services rather than partnerships and structural investments
• LEAs steadily expanded TK access, with 94% offering TK and 70% of multi-site LEAs offering TK at all elementary sites in 2024–25 - LEAs steadily expanded TK access, with 94% offering TK and 70% of multi-site LEAs offering TK at all elementary sites in 2024–25
• LEAs' most frequent sustainability strategy in SBHIP project plans was Medi-Cal billing, despite many LEAs lacking billing infrastructure - LEAs' most frequent sustainability strategy in SBHIP project plans was Medi-Cal billing, despite many LEAs lacking billing infrastructure
• LEAs' technical assistance needs for school-based behavioral health span care coordination, sustainability, infrastructure, and services for students - LEAs' technical assistance needs for school-based behavioral health span care coordination, sustainability, infrastructure, and services for students
• Least-to-most prompting and teaching wh-word meanings meet evidence-based criteria; explicit comprehension-strategy instruction is potentially evidence-based - Least-to-most prompting and teaching wh-word meanings meet evidence-based criteria; explicit comprehension-strategy instruction is potentially evidence-based
• Leave-one-out cross-validation over 36 students confirms significant accuracy and stability gains from eye-tracking in both self-explanation and exploration assessment - Leave-one-out cross-validation over 36 students confirms significant accuracy and stability gains from eye-tracking in both self-explanation and exploration assessment
• Leave-one-out sensitivity analysis confirms the pooled productivity effect is robust (g = 0.24–0.38, all p < 0.05), with two studies (P002, P004) exerting disproportionate influence. - Leave-one-out sensitivity analysis confirms the pooled productivity effect is robust (g = 0.24–0.38, all p < 0.05), with two studies (P002, P004) exerting disproportionate influence.
• Leavers reported lower end-of-semester interest, lower prior achievement, and lower biology grades than stayers, with no differences in FG status, gender, or baseline interest - Leavers reported lower end-of-semester interest, lower prior achievement, and lower biology grades than stayers, with no differences in FG status, gender, or baseline interest
• LeBlanc and Matson (1995) found a statistically significant positive effect on one social-emotional outcome but insufficient information prevented WWC verification or inclusion in the rating - LeBlanc and Matson (1995) found a statistically significant positive effect on one social-emotional outcome but insufficient information prevented WWC verification or inclusion in the rating
• LecturaAgents scores higher than Instructional Agents, GenMentor, and Learn Your Way on lecture content, personalization, and assessment quality - LecturaAgents scores higher than Instructional Agents, GenMentor, and Learn Your Way on lecture content, personalization, and assessment quality
• Lecture modality (written, audio, video, live) shows no consistent effect on memory for lecture content or on the note-taking effect - Lecture modality (written, audio, video, live) shows no consistent effect on memory for lecture content or on the note-taking effect
• Lecture-based and 4C/ID-based training produced comparable knowledge retention among emergency care professionals - Lecture-based and 4C/ID-based training produced comparable knowledge retention among emergency care professionals
• Lecture-based learning events show a clearer decline in EDA across within-event segments than flipped learning events - Lecture-based learning events show a clearer decline in EDA across within-event segments than flipped learning events
• Lecturing is the dominant teaching method in educational psychology, with well-equipped laboratories in about one-third of institutions - Lecturing is the dominant teaching method in educational psychology, with well-equipped laboratories in about one-third of institutions
• Lecturing persists for identifiable psychological and practical reasons: it is easy, exploits personal opinion, and satisfies pupils' craving for authority and ready-made knowledge - Lecturing persists for identifiable psychological and practical reasons: it is easy, exploits personal opinion, and satisfies pupils' craving for authority and ready-made knowledge
• Lee argues the brain and spinal cord show a high degree of resistance to fatigue, with no indisputable evidence of genuine nervous-system fatigue - Lee argues the brain and spinal cord show a high degree of resistance to fatigue, with no indisputable evidence of genuine nervous-system fatigue
• Left ventrolateral prefrontal cortex activation shows linear age-related increases during reappraisal - Left ventrolateral prefrontal cortex activation shows linear age-related increases during reappraisal
• Legal and Regulatory Knowledge is the least represented AIL dimension in existing conceptualizations yet is framed as a core competency given evolving regulation - Legal and Regulatory Knowledge is the least represented AIL dimension in existing conceptualizations yet is framed as a core competency given evolving regulation
• Legal professionals frequently use free NMT with high self-reported satisfaction yet perform little to no post-editing, so satisfaction is not correlated with quality - Legal professionals frequently use free NMT with high self-reported satisfaction yet perform little to no post-editing, so satisfaction is not correlated with quality
• Leisure digital reading habits show a small positive association with reading comprehension (r = .055), much smaller than for print reading habits - Leisure digital reading habits show a small positive association with reading comprehension (r = .055), much smaller than for print reading habits
• Leisure readers are far more likely than non-leisure readers to enjoy reading and to engage in mental imagery, in both educational tracks - Leisure readers are far more likely than non-leisure readers to enjoy reading and to engage in mental imagery, in both educational tracks
• LENA automated adult word counts are reliable for Spanish-language home environments - LENA automated adult word counts are reliable for Spanish-language home environments
• Lengthening consolidation time in the simultaneous condition reduces forgetting to a rate comparable to sequential presentation (Experiment 4) - Lengthening consolidation time in the simultaneous condition reduces forgetting to a rate comparable to sequential presentation (Experiment 4)
• Lengthening the interval between dictated digits markedly lowered children's memory span scores on one testing day - Lengthening the interval between dictated digits markedly lowered children's memory span scores on one testing day
• Les LLMs utilisés comme juges d'évaluation scientifique présentent douze types de biais qui désavantagent les productions s'écartant des canons dominants - Les LLMs utilisés comme juges d'évaluation scientifique présentent douze types de biais qui désavantagent les productions s'écartant des canons dominants
• Lesion evidence links written-language processes to language, working memory, and other cognitive skills - Lesion evidence links written-language processes to language, working memory, and other cognitive skills
• Less experienced teachers are more likely to belong to emotion regulation profiles characterized by avoidance strategies - Less experienced teachers are more likely to belong to emotion regulation profiles characterized by avoidance strategies
• Less positive school organizational conditions account for the higher minority teacher turnover rate; once conditions are controlled, the minority-nonminority turnover difference becomes statistically insignificant - Less positive school organizational conditions account for the higher minority teacher turnover rate; once conditions are controlled, the minority-nonminority turnover difference becomes statistically
• Less skilled comprehenders make fewer knowledge-based coherence and elaborative inferences than skilled comprehenders even when the relevant knowledge is equally available - Less skilled comprehenders make fewer knowledge-based coherence and elaborative inferences than skilled comprehenders even when the relevant knowledge is equally available
• Less skilled comprehenders' inference failures stem more from incorrect premise recall, whereas skilled comprehenders fail more at integrating recalled information - Less skilled comprehenders' inference failures stem more from incorrect premise recall, whereas skilled comprehenders fail more at integrating recalled information
• Less skilled readers are more likely to select high-density text under learner control - Less skilled readers are more likely to select high-density text under learner control
• Less sophisticated epistemic belief profiles are more prevalent in younger students and nonacademic tracks, while sophisticated profiles are more prevalent in academic tracks and higher grade levels - Less sophisticated epistemic belief profiles are more prevalent in younger students and nonacademic tracks, while sophisticated profiles are more prevalent in academic tracks and higher grade levels
• Less successful teachers tended to provide less support because they mistakenly thought prompting was not part of scaffolding - Less successful teachers tended to provide less support because they mistakenly thought prompting was not part of scaffolding
• Less than half of Chicago elementary schools keep pace with ITBS grade-level content expectations during the last three years of elementary schooling - Less than half of Chicago elementary schools keep pace with ITBS grade-level content expectations during the last three years of elementary schooling
• Lesson designers identify generic feedback and unclear title-page and instruction sections as key weaknesses of LLM-generated lessons - Lesson designers identify generic feedback and unclear title-page and instruction sections as key weaknesses of LLM-generated lessons
• Lesson designers report LLM-generated lessons save substantial time drafting scenarios and multiple-choice questions - Lesson designers report LLM-generated lessons save substantial time drafting scenarios and multiple-choice questions
• Lesson ratings showed no statistically significant differences between lessons observed earlier versus later in the initiative on any dimension - Lesson ratings showed no statistically significant differences between lessons observed earlier versus later in the initiative on any dimension
• Lesson variability benefits transfer only in combination with worked-example instruction, as shown by a variability-by-instruction interaction - Lesson variability benefits transfer only in combination with worked-example instruction, as shown by a variability-by-instruction interaction
• Lesson-plan quality grows steadily across three iterative design rounds, but no product reaches the excellent grade - Lesson-plan quality grows steadily across three iterative design rounds, but no product reaches the excellent grade
• Lesson-probe data paths showed clear separation favoring the peer tutoring condition for most probes across all eight participants - Lesson-probe data paths showed clear separation favoring the peer tutoring condition for most probes across all eight participants
• Lessons highly rated for both the mathematics and access to mathematics used varied participation structures, but none of the 14 was also highly rated on agency, authority, and identity - Lessons highly rated for both the mathematics and access to mathematics used varied participation structures, but none of the 14 was also highly rated on agency, authority, and identity
• Lessons in Character has potentially positive effects on academic achievement, with statistically significant effects on mathematics grades and attendance in one randomized controlled trial - Lessons in Character has potentially positive effects on academic achievement, with statistically significant effects on mathematics grades and attendance in one randomized controlled trial
• Lessons in Character shows no discernible effects on knowledge, attitudes, and values - Lessons in Character shows no discernible effects on knowledge, attitudes, and values
• Lessons in Character shows no discernible effects on student behavior - Lessons in Character shows no discernible effects on student behavior
• Lessons using contrasts between critical aspects produced greater long-term retention than the other lessons in each learning study cycle, though not significantly so in individual experiments - Lessons using contrasts between critical aspects produced greater long-term retention than the other lessons in each learning study cycle, though not significantly so in individual experiments
• Letter fluency shows no age-group difference in ANOVA, but older adults outperform young adults once processing speed and IQ are controlled - Letter fluency shows no age-group difference in ANOVA, but older adults outperform young adults once processing speed and IQ are controlled
• Letter size has less effect on legibility than spacing of letters or arrangement of words - Letter size has less effect on legibility than spacing of letters or arrangement of words
• Letter sound only instruction may interfere with letter name learning relative to incidental exposure - Letter sound only instruction may interfere with letter name learning relative to incidental exposure
• Letter type, letter order, textual frequency, number of strokes, and symmetry did not substantially or consistently contribute to letter writing - Letter type, letter order, textual frequency, number of strokes, and symmetry did not substantially or consistently contribute to letter writing
• Letter writing automaticity is marginally related to spelling (b = .11, p = .06) but not uniquely related to word reading (b = -.07, p = .28) in kindergartners - Letter writing automaticity is marginally related to spelling (b = .11, p = .06) but not uniquely related to word reading (b = -.07, p = .28) in kindergartners
• Letter-cancellation practice produced only small gains and some losses in parts-of-speech cancellation, while the unpracticed control group gained on all parts of speech - Letter-cancellation practice produced only small gains and some losses in parts-of-speech cancellation, while the unpracticed control group gained on all parts of speech
• Letter-name knowledge strongly predicts correct letter writing, especially for 4- and 5-year-olds - Letter-name knowledge strongly predicts correct letter writing, especially for 4- and 5-year-olds
• Letter-sound knowledge and phonological awareness are associated with letter writing, but much less strongly than letter-name knowledge - Letter-sound knowledge and phonological awareness are associated with letter writing, but much less strongly than letter-name knowledge
• Letter-writing difficulty varies widely across letters, with O easiest and K hardest, and knowing the letter name raises per-letter probabilities substantially - Letter-writing difficulty varies widely across letters, with O easiest and K hardest, and knowing the letter name raises per-letter probabilities substantially
• Letter-writing fluency has a small unique concurrent relation to kindergarten compositional fluency but no significant relation to first-grade writing outcomes - Letter-writing fluency has a small unique concurrent relation to kindergarten compositional fluency but no significant relation to first-grade writing outcomes
• Letting learners pace instruction helped on complex material in one experiment, while learner control in general showed almost no effect across 18 studies
• Level of implementation moderates Dialogic Reading's effects on children's oral language - Level of implementation moderates Dialogic Reading's effects on children's oral language
• Leveled Literacy Intervention has positive effects on general reading achievement, potentially positive effects on reading fluency, and no discernible effects on alphabetics for beginning readers in grades K–2 - Leveled Literacy Intervention has positive effects on general reading achievement, potentially positive effects on reading fluency, and no discernible effects on alphabetics for beginning readers in g
• Leveling up lessons preceding the source lesson improves Review Exercise accuracy, indicating transfer of learning benefits across lessons within a skill - Leveling up lessons preceding the source lesson improves Review Exercise accuracy, indicating transfer of learning benefits across lessons within a skill
• Levels of joint visual attention are correlated with collaboration quality across collaborative groups - Levels of joint visual attention are correlated with collaboration quality across collaborative groups
• Leveraging pre-existing networks or hubs may support greater potential for long-term sustainability of cross-sector reform - Leveraging pre-existing networks or hubs may support greater potential for long-term sustainability of cross-sector reform
• Lexia Early Reading outperforms no supplemental instruction on kindergarten alphabetics outcomes in Gale (2006), significant by author but not WWC calculations - Lexia Early Reading outperforms no supplemental instruction on kindergarten alphabetics outcomes in Gale (2006), significant by author but not WWC calculations
• Lexia Early Reading shows no significant advantage over Earobics on alphabetics, with several negative effects substantively important - Lexia Early Reading shows no significant advantage over Earobics on alphabetics, with several negative effects substantively important
• Lexia Reading has potentially positive effects on alphabetics for beginning readers - Lexia Reading has potentially positive effects on alphabetics for beginning readers
• Lexia Reading has potentially positive effects on comprehension - Lexia Reading has potentially positive effects on comprehension
• Lexia Reading shows no discernible effects on general reading achievement - Lexia Reading shows no discernible effects on general reading achievement
• Lexia Reading shows no discernible effects on reading fluency - Lexia Reading shows no discernible effects on reading fluency
• Lexical adoption shows no significant condition effect after controlling for input volume, but AI-first participants adopted more words at the second dialogue position - Lexical adoption shows no significant condition effect after controlling for input volume, but AI-first participants adopted more words at the second dialogue position
• Lexical LREs dominate grammar-related LREs in young learners' L2 writing noticing - Lexical LREs dominate grammar-related LREs in young learners' L2 writing noticing
• Lexical processing at 36 months did not predict later school readiness over and above concurrent vocabulary, but predicted BSRA better than vocabulary at 60 months did - Lexical processing at 36 months did not predict later school readiness over and above concurrent vocabulary, but predicted BSRA better than vocabulary at 60 months did
• Lexical processing latency at 36 months does not predict concurrent grammatical complexity (LDS average length of phrase) - Lexical processing latency at 36 months does not predict concurrent grammatical complexity (LDS average length of phrase)
• Lexical processing latency at 36 months predicts school readiness (BSRA) at 60 months for overall and noun-trial latency, but not verb-trial latency - Lexical processing latency at 36 months predicts school readiness (BSRA) at 60 months for overall and noun-trial latency, but not verb-trial latency
• Lexical processing speed at 36 months is associated with concurrent receptive vocabulary, but more weakly than at younger ages - Lexical processing speed at 36 months is associated with concurrent receptive vocabulary, but more weakly than at younger ages
• Lexical richness and fluency measures differentiate between low-level L2 learners' comprehensibility - Lexical richness and fluency measures differentiate between low-level L2 learners' comprehensibility
• Lexical simplification (high-frequency target words) substantially reduces word-level processing times in adult sentence reading - Lexical simplification (high-frequency target words) substantially reduces word-level processing times in adult sentence reading
• Lexical simplification benefits non-university adults more than university adults, though it helps both groups - Lexical simplification benefits non-university adults more than university adults, though it helps both groups
• Lexical units brought to the 'fully learned' stage by the tutor scored significantly lower on the posttest than units students already knew before introduction - Lexical units brought to the 'fully learned' stage by the tutor scored significantly lower on the posttest than units students already knew before introduction
• LFA discovered that backward application of the circle area formula is a distinct, harder skill with no transfer from forward application - LFA discovered that backward application of the circle area formula is a distinct, harder skill with no transfer from forward application
• LFA-discovered cognitive models outperformed both original in-use and best hand-generated models in all eleven DataShop datasets - LFA-discovered cognitive models outperformed both original in-use and best hand-generated models in all eleven DataShop datasets
• Liberal discourse grounds democratic education in knowledge of democratic institutions and critical thinking across the curriculum - Liberal discourse grounds democratic education in knowledge of democratic institutions and critical thinking across the curriculum
• Lied magnet students in grades 6-8 attempted and completed significantly more STEM credits than matched comparison students - Lied magnet students in grades 6-8 attempted and completed significantly more STEM credits than matched comparison students
• Lifetime exposure to narrative fiction correlates with ToM and empathy, but correlational designs preclude causal conclusions - Lifetime exposure to narrative fiction correlates with ToM and empathy, but correlational designs preclude causal conclusions
• Lifetime stress severity correlates with poorer learning performance and reduced putamen/ACC activation - Lifetime stress severity correlates with poorer learning performance and reduced putamen/ACC activation
• Light AI-assisted edits to human-written abstracts are flagged by commercial detectors at 64–80% (Pangram) and 38–49% (GPTZero) rates - Light AI-assisted edits to human-written abstracts are flagged by commercial detectors at 64–80% (Pangram) and 38–49% (GPTZero) rates
• Light and moderate physical activity and sedentary behavior show no significant relationship with any creativity component or overall creativity - Light and moderate physical activity and sedentary behavior show no significant relationship with any creativity component or overall creativity
• Light verbs drive construction-based sorting in intermediate-level but not beginner or advanced EFL learners - Light verbs drive construction-based sorting in intermediate-level but not beginner or advanced EFL learners
• Lighthouse institutions showed uneven baseline capacity for adaptive courseware implementation, lowest in course redesign capacity - Lighthouse institutions showed uneven baseline capacity for adaptive courseware implementation, lowest in course redesign capacity
• Limited contextual understanding of pragmatic, cultural, and discourse information is a major technical limitation, reported in 16 of 54 studies (30%) - Limited contextual understanding of pragmatic, cultural, and discourse information is a major technical limitation, reported in 16 of 54 studies (30%)
• Limited device access and a six-month engagement constrained some Board members' contributions, with one member reporting expectations were not met - Limited device access and a six-month engagement constrained some Board members' contributions, with one member reporting expectations were not met
• Limited English fluency and low native-language literacy among parents hinder school communication in ELL communities - Limited English fluency and low native-language literacy among parents hinder school communication in ELL communities
• Limited human capital is one of the biggest challenges in rural school turnaround, making SIG staff-replacement requirements often neither possible nor desirable - Limited human capital is one of the biggest challenges in rural school turnaround, making SIG staff-replacement requirements often neither possible nor desirable
• Limited instruction directed at complex structures also proved effective, provided the target structures are readily available in noninstructional input - Limited instruction directed at complex structures also proved effective, provided the target structures are readily available in noninstructional input
• Limited output accuracy and over-reliance are the most pressing challenges Chinese HSS students report in GenAI use, alongside strong ethical concern - Limited output accuracy and over-reliance are the most pressing challenges Chinese HSS students report in GenAI use, alongside strong ethical concern
• Limited school staff time is a crucial obstacle impeding cross-sector collaboration - Limited school staff time is a crucial obstacle impeding cross-sector collaboration
• Limited teacher knowledge of language development and curriculum procedures was the third key implementation barrier - Limited teacher knowledge of language development and curriculum procedures was the third key implementation barrier
• Limiting verbose Math Agent guidance sharply reduced answer giveaway but decreased cognitive engagement - Limiting verbose Math Agent guidance sharply reduced answer giveaway but decreased cognitive engagement
• LINCS surpassed its targets for program completion, exceeding its proposal goal of preparing students for supply chain management positions - LINCS surpassed its targets for program completion, exceeding its proposal goal of preparing students for supply chain management positions
• Linear path design in EERs tends to foster collaboration, while open-path structures enable positive social interdependence - Linear path design in EERs tends to foster collaboration, while open-path structures enable positive social interdependence
• Linear regression outperformed equipercentile and mean-sigma equating for predicting ARM scores and was selected as the final linking model - Linear regression outperformed equipercentile and mean-sigma equating for predicting ARM scores and was selected as the final linking model
• Linguist review of draft items revealed communication problems with terminology and imprecisely reasoned attitude statements - Linguist review of draft items revealed communication problems with terminology and imprecisely reasoned attitude statements
• Linguistic and psycholinguistic features such as verb ratio and aggressiveness are positively correlated with perceived ethnic bias - Linguistic and psycholinguistic features such as verb ratio and aggressiveness are positively correlated with perceived ethnic bias
• Linguistic coherence predicts productive collaboration, strengthened by joint visual attention data - Linguistic coherence predicts productive collaboration, strengthened by joint visual attention data
• Linguistic comprehension instruction shows a small immediate effect on generalized linguistic comprehension outcomes - Linguistic comprehension instruction shows a small immediate effect on generalized linguistic comprehension outcomes
• Linguistic comprehension instruction shows no clear immediate effect on generalized reading comprehension outcomes - Linguistic comprehension instruction shows no clear immediate effect on generalized reading comprehension outcomes
• Linguistic delay may affect verbal mathematical abilities without impairing conceptual counting development, as deaf children's one-to-one correspondence understanding was age-appropriate - Linguistic delay may affect verbal mathematical abilities without impairing conceptual counting development, as deaf children's one-to-one correspondence understanding was age-appropriate
• Linguistic group and perceptual modality show a significant disordinal interaction in metacognition: Arabic speakers monitor better in the auditory modality, Chinese speakers in the visual modality - Linguistic group and perceptual modality show a significant disordinal interaction in metacognition: Arabic speakers monitor better in the auditory modality, Chinese speakers in the visual modality
• Linguistic inferencing uniquely predicts all three components of children's scientific reasoning over and above visuospatial inferencing - Linguistic inferencing uniquely predicts all three components of children's scientific reasoning over and above visuospatial inferencing
• Linguistic intelligence dominated both series while natural intelligence was least frequent - Linguistic intelligence dominated both series while natural intelligence was least frequent
• Linguistic problem difficulty and lower linguistic abilities increase attention on situational text during the first solution phase of word-problem solving - Linguistic problem difficulty and lower linguistic abilities increase attention on situational text during the first solution phase of word-problem solving
• Linguistic selection reasons were uttered most regardless of the selection task, suggesting a default linguistic selection strategy - Linguistic selection reasons were uttered most regardless of the selection task, suggesting a default linguistic selection strategy
• Linguistic simplification of math word problems can improve performance, particularly among English language learners and students from low-socioeconomic backgrounds - Linguistic simplification of math word problems can improve performance, particularly among English language learners and students from low-socioeconomic backgrounds
• Linguistic-level discourse analysis of students restating animation descriptions in their own words documents collaborative behavior and informs task redesign - Linguistic-level discourse analysis of students restating animation descriptions in their own words documents collaborative behavior and informs task redesign
• Linguistic-probe late-N1 latency decreases with age, unlike the stable early-N1 latency for nonlinguistic probes - Linguistic-probe late-N1 latency decreases with age, unlike the stable early-N1 latency for nonlinguistic probes
• Linguistically varying an instructional video into easy, moderate, and difficult versions does not significantly affect students’ expectancies of success or intrinsic task value compared with the moderate version - Linguistically varying an instructional video into easy, moderate, and difficult versions does not significantly affect students’ expectancies of success or intrinsic task value compared with the mode
• Linkages between picture elements and interpretation statements are better recognized with an LDI integrated than an LDI separated audio structure - Linkages between picture elements and interpretation statements are better recognized with an LDI integrated than an LDI separated audio structure
• Linked Learning benefits occurred only in pathways independently certified as high-quality on seven implementation dimensions - Linked Learning benefits occurred only in pathways independently certified as high-quality on seven implementation dimensions
• Linked learning communities produced no statistically significant or substantively important effect on degree attainment - Linked learning communities produced no statistically significant or substantively important effect on degree attainment
• Linked learning communities produced no statistically significant or substantively important effects on academic achievement - Linked learning communities produced no statistically significant or substantively important effects on academic achievement
• Linked learning communities produced no statistically significant or substantively important effects on credit accumulation across six studies - Linked learning communities produced no statistically significant or substantively important effects on credit accumulation across six studies
• Linked learning communities show no discernible effects on any of five outcome domains for community college students in developmental education - Linked learning communities show no discernible effects on any of five outcome domains for community college students in developmental education
• Linked learning communities show no overall effect on progress in developmental education, with a statistically significant but non-substantive gender interaction at Houston - Linked learning communities show no overall effect on progress in developmental education, with a statistically significant but non-substantive gender interaction at Houston
• Linked learning communities show no overall enrollment effect, but statistically significant gender-by-condition interactions on registration at Houston and substantively important differences at Hillsborough - Linked learning communities show no overall enrollment effect, but statistically significant gender-by-condition interactions on registration at Houston and substantively important differences at Hill
• Linking CPS, ISBE, and ICCB records through the CDDA ID matched 97 percent of the 2012 CPS graduating class to ICCB data - Linking CPS, ISBE, and ICCB records through the CDDA ID matched 97 percent of the 2012 CPS graduating class to ICCB data
• Linking studies connect MAP Growth scores to state summative proficiency predictions and external measures - Linking studies connect MAP Growth scores to state summative proficiency predictions and external measures
• Linking studies predict proficiency on state summative assessments from MAP Growth scores - Linking studies predict proficiency on state summative assessments from MAP Growth scores
• LiPS has mixed effects on alphabetics for beginning readers: positive and statistically significant in one trial, negative and substantively important (but not significant) in another - LiPS has mixed effects on alphabetics for beginning readers: positive and statistically significant in one trial, negative and substantively important (but not significant) in another
• LiPS has no discernible effects on reading comprehension for students with learning disabilities - LiPS has no discernible effects on reading comprehension for students with learning disabilities
• LiPS has potentially negative effects on writing: comparison students significantly outperformed LiPS students on the KTEA Spelling subtest - LiPS has potentially negative effects on writing: comparison students significantly outperformed LiPS students on the KTEA Spelling subtest
• LiPS has potentially positive effects on alphabetics for students with learning disabilities, with statistically significant advantages on two of eight measures - LiPS has potentially positive effects on alphabetics for students with learning disabilities, with statistically significant advantages on two of eight measures
• LiPS has potentially positive effects on comprehension for beginning readers, based on one randomized trial showing a statistically significant positive effect on passage comprehension - LiPS has potentially positive effects on comprehension for beginning readers, based on one randomized trial showing a statistically significant positive effect on passage comprehension
• LiPS shows potentially positive effects on math: no statistically significant effect on the WJ–R Calculation subtest, but a positive effect size of 0.30 judged substantively important - LiPS shows potentially positive effects on math: no statistically significant effect on the WJ–R Calculation subtest, but a positive effect size of 0.30 judged substantively important
• LiPS shows potentially positive effects on reading fluency: authors found significant effects on GORT–III Reading Accuracy and Reading Rate, but WWC analyses did not confirm significance - LiPS shows potentially positive effects on reading fluency: authors found significant effects on GORT–III Reading Accuracy and Reading Rate, but WWC analyses did not confirm significance
• LIS users recognize Pepper's isolated signs far better than its full signed sentences in an exploratory evaluation - LIS users recognize Pepper's isolated signs far better than its full signed sentences in an exploratory evaluation
• Listen/understand negatively affected dyads' appropriateness, while Extend/revise's negative effect became insignificant once Inviting behavior was controlled - Listen/understand negatively affected dyads' appropriateness, while Extend/revise's negative effect became insignificant once Inviting behavior was controlled
• Listening to vocabulary lessons while physically active improves End-of-Unit exam performance over traditional intentional learning alone - Listening to vocabulary lessons while physically active improves End-of-Unit exam performance over traditional intentional learning alone
• Literacy benefits of dual language programs did not extend to the English learner subsample in the strongest study - Literacy benefits of dual language programs did not extend to the English learner subsample in the strongest study
• Literacy coaching responsibilities fall into two major areas: teacher mentoring and literacy program advocacy - Literacy coaching responsibilities fall into two major areas: teacher mentoring and literacy program advocacy
• Literacy Express has positive effects on oral language for preschool children, with significant effects in two of three studies - Literacy Express has positive effects on oral language for preschool children, with significant effects in two of three studies
• Literacy Express has positive effects on phonological processing for preschool children, with significant effects in two of three studies - Literacy Express has positive effects on phonological processing for preschool children, with significant effects in two of three studies
• Literacy Express has positive effects on print knowledge for preschool children, with significant effects in two of three studies - Literacy Express has positive effects on print knowledge for preschool children, with significant effects in two of three studies
• Literacy Express shows no discernible effects on cognition and math for preschool children - Literacy Express shows no discernible effects on cognition and math for preschool children
• Literacy First shows strong ESSA evidence of positive effects in phonics and related alphabetics, reading fluency, and reading comprehension for grades K–2 - Literacy First shows strong ESSA evidence of positive effects in phonics and related alphabetics, reading fluency, and reading comprehension for grades K–2
• Literacy is associated with functional status in older adults, but after full adjustment total literacy remains associated only with instrumental activities of daily living and mobility disability - Literacy is associated with functional status in older adults, but after full adjustment total literacy remains associated only with instrumental activities of daily living and mobility disability
• Literacy screener benchmarks predict which grade 3 students will not meet MCAS ELA expectations, but not necessarily who will meet them - Literacy screener benchmarks predict which grade 3 students will not meet MCAS ELA expectations, but not necessarily who will meet them
• Literary appreciation is hardly attainable as a by-product of formal language-training instruction - Literary appreciation is hardly attainable as a by-product of formal language-training instruction
• Literature and expert opinion converge on a limited number of key noncognitive attributes that matter for student success, with multiple available instruments - Literature and expert opinion converge on a limited number of key noncognitive attributes that matter for student success, with multiple available instruments
• Little attention has been paid to the undesirable consequences of indiscriminate AI application, including automation-driven socio-economic inequality - Little attention has been paid to the undesirable consequences of indiscriminate AI application, including automation-driven socio-economic inequality
• Little evidence exists of student achievement changes following Common Core implementation in other places, with the one positive study limited to wealthier districts - Little evidence exists of student achievement changes following Common Core implementation in other places, with the one positive study limited to wealthier districts
• Little evidence of heterogeneity in preschool effects by child characteristics, with suggestive larger behavioral drawbacks for lower-SES children - Little evidence of heterogeneity in preschool effects by child characteristics, with suggestive larger behavioral drawbacks for lower-SES children
• Little evidence that Philadelphia charter schools help or harm achievement of students in nearby district schools - Little evidence that Philadelphia charter schools help or harm achievement of students in nearby district schools
• Little is known about growth in reading for students with visual impairments as they progress through school - Little is known about growth in reading for students with visual impairments as they progress through school
• Little is known about how much variance in SEL constructs is stable over time versus specific to a given time or context - Little is known about how much variance in SEL constructs is stable over time versus specific to a given time or context
• Little is known about substitute teachers' work conditions, motivations, experiences, and support despite their ubiquity in the teaching workforce - Little is known about substitute teachers' work conditions, motivations, experiences, and support despite their ubiquity in the teaching workforce
• Little is known about teachers certified through alternative certification programs and their experiences after certification - Little is known about teachers certified through alternative certification programs and their experiences after certification
• Little of the existing IRT scaling and linking expertise from large-scale educational assessment is applied to scaling psychological and social-emotional survey constructs - Little of the existing IRT scaling and linking expertise from large-scale educational assessment is applied to scaling psychological and social-emotional survey constructs
• Little or no relation between absolute and relative accuracy in metacomprehension research - Little or no relation between absolute and relative accuracy in metacomprehension research
• Little was known about the circumstances under which the trifactor model performs well versus poorly prior to this study - Little was known about the circumstances under which the trifactor model performs well versus poorly prior to this study
• Little was known before this study about students who rapidly guess over the course of several tests - Little was known before this study about students who rapidly guess over the course of several tests
• LittleLearner's capabilities tightly track its controlled exposure: strong on in-scope tasks, degrading beyond the K-5 boundary - LittleLearner's capabilities tightly track its controlled exposure: strong on in-scope tasks, degrading beyond the K-5 boundary
• Live feedback shifts the router's strategy from the Feynman prompt favored in simulation to the scaffolding Coach prompt, with superior live performance - Live feedback shifts the router's strategy from the Feynman prompt favored in simulation to the scaffolding Coach prompt, with superior live performance
• Live simulations with real patients showed the greatest effects among simulation types (g = 2.27, medical education only) - Live simulations with real patients showed the greatest effects among simulation types (g = 2.27, medical education only)
• Living in a very safe neighborhood is associated with higher kindergarten curiosity, but only for lower- and middle-resourced children - Living in a very safe neighborhood is associated with higher kindergarten curiosity, but only for lower- and middle-resourced children
• Living in close geographic proximity to homicide is followed by lower attendance, test scores, GPA, and more disciplinary incidents for CPS students - Living in close geographic proximity to homicide is followed by lower attendance, test scores, GPA, and more disciplinary incidents for CPS students
• Llama 4 Maverick shows the strongest human-LLM alignment, while Claude Sonnet 4 performs weakest with poor calibration - Llama 4 Maverick shows the strongest human-LLM alignment, while Claude Sonnet 4 performs weakest with poor calibration
• Llama Guard models misclassify subtle, context-dependent unsafe education prompts (e.g., exam-answer and cheating requests) as safe - Llama Guard models misclassify subtle, context-dependent unsafe education prompts (e.g., exam-answer and cheating requests) as safe
• LLaMA produces a fivefold higher victimization-to-politics vocabulary ratio for the Roma persona (≈0.15) than for elite peers (≈0.03) - LLaMA produces a fivefold higher victimization-to-politics vocabulary ratio for the Roma persona (≈0.15) than for elite peers (≈0.03)
• LLaMA shows a persona-conditioned vocabulary collapse for the Roma minority persona, with TTR falling 0.018 points below its own cross-persona mean - LLaMA shows a persona-conditioned vocabulary collapse for the Roma minority persona, with TTR falling 0.018 points below its own cross-persona mean
• LLaMA's LoRA adaptation compresses its prediction range to roughly 2.0–4.0, while Gemma maintains a broader distribution closer to the ground truth - LLaMA's LoRA adaptation compresses its prediction range to roughly 2.0–4.0, while Gemma maintains a broader distribution closer to the ground truth
• LlamaLo added knowledge in almost every post, mostly correct, and most posts were internally coherent - LlamaLo added knowledge in almost every post, mostly correct, and most posts were internally coherent
• LlamaLo exhibited bias and presented contested assertions as facts, notably in the Israeli-Palestinian conflict discussion - LlamaLo exhibited bias and presented contested assertions as facts, notably in the Israeli-Palestinian conflict discussion
• LlamaLo occasionally employed an aggressive or sarcastic tone, sometimes without user provocation, though it changed tone after being commented on - LlamaLo occasionally employed an aggressive or sarcastic tone, sometimes without user provocation, though it changed tone after being commented on
• LlamaLo's responses to social posts were solely social and it did not push back to resume the discussion - LlamaLo's responses to social posts were solely social and it did not push back to resume the discussion
• LLC did not affect employment for all participants within the first 24 months of enrollment - LLC did not affect employment for all participants within the first 24 months of enrollment
• LLC increased postsecondary education enrollment: 31 percent of participants enrolled versus 22 percent of the control group - LLC increased postsecondary education enrollment: 31 percent of participants enrolled versus 22 percent of the control group
• LLC increased the likelihood of earnings among later enrollees by 11 percentage points - LLC increased the likelihood of earnings among later enrollees by 11 percentage points
• LLC produced large impacts on service use in the two years after enrollment, particularly services the program intended to offer - LLC produced large impacts on service use in the two years after enrollment, particularly services the program intended to offer
• LLI had a negative impact on secondary students' mastery of English language arts/literacy standards - LLI had a negative impact on secondary students' mastery of English language arts/literacy standards
• LLI had no impact on secondary students' reading comprehension in the Oakland RCT - LLI had no impact on secondary students' reading comprehension in the Oakland RCT
• LLM access boosts individual story creativity while significantly homogenising output across writers (review attribution) - LLM access boosts individual story creativity while significantly homogenising output across writers (review attribution)
• LLM accuracy drops from 96% on text-only to 74% on multimodal physics problems, with significant differences emerging among models - LLM accuracy drops from 96% on text-only to 74% on multimodal physics problems, with significant differences emerging among models
• LLM accuracy increases monotonically with prompt sophistication across semantically equivalent medical questions, with Low Literacy users scoring significantly below Prompt Engineer users - LLM accuracy increases monotonically with prompt sophistication across semantically equivalent medical questions, with Low Literacy users scoring significantly below Prompt Engineer users
• LLM accuracy on the CDPK pedagogy benchmark ranges from 28% to 89%, with closed-weight reasoning models dominating the top of the leaderboard - LLM accuracy on the CDPK pedagogy benchmark ranges from 28% to 89%, with closed-weight reasoning models dominating the top of the leaderboard
• LLM accuracy varies across mathematically equivalent problem representations even under deterministic decoding, in both interaction conditions - LLM accuracy varies across mathematically equivalent problem representations even under deterministic decoding, in both interaction conditions
• LLM adoption varies substantially across CS subfield assignments under a shared instructional intervention, with highest use in Algorithms (89.6%) and Web Development (80.6%) and lowest in Software Engineering (15.2%) - LLM adoption varies substantially across CS subfield assignments under a shared instructional intervention, with highest use in Algorithms (89.6%) and Web Development (80.6%) and lowest in Software En
• LLM agreement with human examiners is limited in fast mode and improves in reasoning or thinking modes - LLM agreement with human examiners is limited in fast mode and improves in reasoning or thinking modes
• LLM alignment with expert teaching ratings does not predict, and is often negatively associated with, alignment with student learning gains - LLM alignment with expert teaching ratings does not predict, and is often negatively associated with, alignment with student learning gains
• LLM and human written 7C analyses show no overall difference in behavioral alignment or evidence correspondence, but align less on Communication and Constructive dimensions - LLM and human written 7C analyses show no overall difference in behavioral alignment or evidence correspondence, but align less on Communication and Constructive dimensions
• LLM assessment conversations can be derailed off-topic by gaming attempts, and one successful prompt revision was not demonstrated to be replicable - LLM assessment conversations can be derailed off-topic by gaming attempts, and one successful prompt revision was not demonstrated to be replicable
• LLM assistance can enhance employees' creativity, especially for those with skills to reflect on AI use - LLM assistance can enhance employees' creativity, especially for those with skills to reflect on AI use
• LLM assistance shifts instructor effort from manual setup toward pedagogical expression - LLM assistance shifts instructor effort from manual setup toward pedagogical expression
• LLM assistance value in VR instruction varies with instructional intent type - LLM assistance value in VR instruction varies with instructional intent type
• LLM attack success rates increase consistently from single-turn to static multi-turn and dynamic multi-turn interactions in educational settings - LLM attack success rates increase consistently from single-turn to static multi-turn and dynamic multi-turn interactions in educational settings
• LLM capabilities enabled family education systems across conversational, embodied, and spatial interface forms to generate support from the context of an unfolding interaction - LLM capabilities enabled family education systems across conversational, embodied, and spatial interface forms to generate support from the context of an unfolding interaction
• LLM coders replicate many human discursive moves but lack questioning, adaptive responsiveness, and emotional expressiveness - LLM coders replicate many human discursive moves but lack questioning, adaptive responsiveness, and emotional expressiveness
• LLM configurations show high within-configuration reliability when re-coding the same chat log dataset, with ChatGPT4o/temperature=0 highest - LLM configurations show high within-configuration reliability when re-coding the same chat log dataset, with ChatGPT4o/temperature=0 highest
• LLM diagnostic performance drops sharply when student responses contain vague rather than evident evidence - LLM diagnostic performance drops sharply when student responses contain vague rather than evident evidence
• LLM difficulty labels track structural question features: stem length, option length, and digits in stems rise monotonically across Easy, Medium, and Hard labels - LLM difficulty labels track structural question features: stem length, option length, and digits in stems rise monotonically across Easy, Medium, and Hard labels
• LLM difficulty predictions for fraction items are less stable across repeated runs than for the overall item set - LLM difficulty predictions for fraction items are less stable across repeated runs than for the overall item set
• LLM digital shadows reproduce the human pattern of positive science perceptions anchored in negative mathematics associations - LLM digital shadows reproduce the human pattern of positive science perceptions anchored in negative mathematics associations
• LLM distractor-generation strategies differ by domain: error description and simulation dominate on math while conceptual linking and discriminability checks dominate on science - LLM distractor-generation strategies differ by domain: error description and simulation dominate on math while conceptual linking and discriminability checks dominate on science
• LLM error simulation rarely fails, but the model is less likely to reproduce a targeted distractor on problems with longer correct solutions - LLM error simulation rarely fails, but the model is less likely to reproduce a targeted distractor on problems with longer correct solutions
• LLM errors on multimodal physics problems fall into four categories, with visual processing errors most prevalent, and error type distribution does not differ significantly across models - LLM errors on multimodal physics problems fall into four categories, with visual processing errors most prevalent, and error type distribution does not differ significantly across models
• LLM factual errors during collaborative circuit debugging occurred almost exclusively after image inputs, attributed to limited 3D spatial reasoning from breadboard or PCB photos - LLM factual errors during collaborative circuit debugging occurred almost exclusively after image inputs, attributed to limited 3D spatial reasoning from breadboard or PCB photos
• LLM feedback adaptivity to student performance occurs only in frequently produced focus types and only in large or closed models - LLM feedback adaptivity to student performance occurs only in frequently produced focus types and only in large or closed models
• LLM feedback agents classify optimal logic proof steps near-ceiling but perform poorly on valid-alternative and incorrect solutions - LLM feedback agents classify optimal logic proof steps near-ceiling but perform poorly on valid-alternative and incorrect solutions
• LLM feedback agents show model-specific over-rejection of valid-alternative reasoning and over-validation of incorrect solutions - LLM feedback agents show model-specific over-rejection of valid-alternative reasoning and over-validation of incorrect solutions
• LLM feedback for students marked by race, language, disability, gender, or motivation shows positive feedback bias and feedback withholding bias — more praise and less substantive critique than for unmarked counterparts - LLM feedback for students marked by race, language, disability, gender, or motivation shows positive feedback bias and feedback withholding bias — more praise and less substantive critique than for un
• LLM feedback receives higher expert ratings on six of seven quality dimensions, with tone the exception - LLM feedback receives higher expert ratings on six of seven quality dimensions, with tone the exception
• LLM feedback sentiment tracks assigned scores: low scores come with more criticism, high scores with more praise - LLM feedback sentiment tracks assigned scores: low scores come with more criticism, high scores with more praise
• LLM graders show model-specific precision/recall trade-offs, so a single metric is insufficient to judge grading quality - LLM graders show model-specific precision/recall trade-offs, so a single metric is insufficient to judge grading quality
• LLM grading patterns cluster into six groups aligned with vendor families, with highest correlations within same-vendor model families - LLM grading patterns cluster into six groups aligned with vendor families, with highest correlations within same-vendor model families
• LLM grading performance relative to human graders is mixed: comparable in 12 reviewed cases, worse in 11, with the strongest alignment on short, straightforward tasks - LLM grading performance relative to human graders is mixed: comparable in 12 reviewed cases, worse in 11, with the strongest alignment on short, straightforward tasks
• LLM history tutors reduce marginalized students' access to contested geopolitical framing, with the coup theory appearing in 2.6% of Low Tier responses versus 7.9% for the baseline - LLM history tutors reduce marginalized students' access to contested geopolitical framing, with the coup theory appearing in 2.6% of Low Tier responses versus 7.9% for the baseline
• LLM learners are younger, more likely male, and score higher on technology affinity than non-learners - LLM learners are younger, more likely male, and score higher on technology affinity than non-learners
• LLM misconception analysis that ties student errors to instructional content reveals insights not apparent from performance data alone - LLM misconception analysis that ties student errors to instructional content reveals insights not apparent from performance data alone
• LLM observer ratings converge excellently with human expert judgments in applying the CAARS observer instrument to generated narratives - LLM observer ratings converge excellently with human expert judgments in applying the CAARS observer instrument to generated narratives
• LLM output can be misleading or erroneous because a stochastic element in text generation sometimes selects less likely continuations, so it requires continuous supervision and critical evaluation by domain experts. - LLM output can be misleading or erroneous because a stochastic element in text generation sometimes selects less likely continuations, so it requires continuous supervision and critical evaluation by
• LLM participant agents clarify stage goals, keep discussions on topic, and prompt deeper ethical reflection - LLM participant agents clarify stage goals, keep discussions on topic, and prompt deeper ethical reflection
• LLM performance drops from offline to real-time proxy evaluation when clues arrive incrementally - LLM performance drops from offline to real-time proxy evaluation when clues arrive incrementally
• LLM ratings of classroom transcripts are more correlated with each other than with expert human ratings, across the same and different tasks - LLM ratings of classroom transcripts are more correlated with each other than with expert human ratings, across the same and different tasks
• LLM responses that are incomplete or incorrect prompt students to question, verify and improve the information, fostering critical thinking through cycles of validation and correction - LLM responses that are incomplete or incorrect prompt students to question, verify and improve the information, fostering critical thinking through cycles of validation and correction
• LLM scaffolding level is associated with the student's next dialogue act, and the association persists within broad request categories - LLM scaffolding level is associated with the student's next dialogue act, and the association persists within broad request categories
• LLM scaffolding level varies systematically across student request types and across assignments, but high-assistance responses remain dominant - LLM scaffolding level varies systematically across student request types and across assignments, but high-assistance responses remain dominant
• LLM scoring performance is dimension-specific: Reasoning shows the strongest agreement and Terminology the weakest across all three strategies - LLM scoring performance is dimension-specific: Reasoning shows the strongest agreement and Terminology the weakest across all three strategies
• LLM self-detection performance is strongly task-dependent: reliable for programming tasks and longer reflective responses but poor for short-answer questions - LLM self-detection performance is strongly task-dependent: reliable for programming tasks and longer reflective responses but poor for short-answer questions
• LLM Short-Answer Graders Agree With One Human Rater More, and the Other Less, Than the Two Raters Agree With Each Other
• LLM student simulators exhibit near-zero Selective Flip Score, flipping at similarly high rates under targeted, misaligned, and generic feedback - LLM student simulators exhibit near-zero Selective Flip Score, flipping at similarly high rates under targeted, misaligned, and generic feedback
• LLM teaching effectiveness varies substantially across subject domains, strongest where knowledge points apply directly and weakest in physics and chemistry - LLM teaching effectiveness varies substantially across subject domains, strongest where knowledge points apply directly and weakest in physics and chemistry
• LLM usage varied across assignments and tasks, increasing for later assignments (especially discussion points), while Low-Reliance students concentrated use early and reduced it after Midterm 2 - LLM usage varied across assignments and tasks, increasing for later assignments (especially discussion points), while Low-Reliance students concentrated use early and reduced it after Midterm 2
• LLM use for code generation is significantly negatively correlated with student final grades - LLM use for code generation is significantly negatively correlated with student final grades
• LLM use for debugging is significantly negatively correlated with student final grades - LLM use for debugging is significantly negatively correlated with student final grades
• LLM use for seeking additional explanations does not significantly impact final grades - LLM use for seeking additional explanations does not significantly impact final grades
• LLM-annotated empathy dimensions correlate positively with intention to follow, self-efficacy, and chatbot usefulness despite participants not preferring empathetic chatbots - LLM-annotated empathy dimensions correlate positively with intention to follow, self-efficacy, and chatbot usefulness despite participants not preferring empathetic chatbots
• LLM-annotated EPITOME scores show the Non-Empathetic chatbot expresses significantly lower empathy on all three dimensions, while Standard and Empathetic chatbots are not significantly different - LLM-annotated EPITOME scores show the Non-Empathetic chatbot expresses significantly lower empathy on all three dimensions, while Standard and Empathetic chatbots are not significantly different
• LLM-as-a-judge ratings found only small, incremental differences between systems, with ceiling effects on several criteria - LLM-as-a-judge ratings found only small, incremental differences between systems, with ceiling effects on several criteria
• LLM-as-a-Judge scores correlate with blinded human expert judgments, more strongly for Medical Narrative Generation than Story Direction - LLM-as-a-Judge scores correlate with blinded human expert judgments, more strongly for Medical Narrative Generation than Story Direction
• LLM-as-judge automated evaluation can achieve human-level agreement when carefully validated - LLM-as-judge automated evaluation can achieve human-level agreement when carefully validated
• LLM-as-judge evaluation scored CyberAGENTS highest on domain grounding (4.46) and agent role fidelity (4.35) - LLM-as-judge evaluation scored CyberAGENTS highest on domain grounding (4.46) and agent role fidelity (4.35)
• LLM-assigned difficulty labels fail to predict empirical item difficulty in classroom retrieval practice (ρ=0.06), as do co-generated Bloom levels (ρ=0.02) - LLM-assigned difficulty labels fail to predict empirical item difficulty in classroom retrieval practice (ρ=0.06), as do co-generated Bloom levels (ρ=0.02)
• LLM-assisted hardware debugging requires consistent human feedback because the LLM's understanding of the circuit drifts - LLM-assisted hardware debugging requires consistent human feedback because the LLM's understanding of the circuit drifts
• LLM-assisted self-study yields higher post-intervention exam performance than traditional resources among early-career knowledge workers - LLM-assisted self-study yields higher post-intervention exam performance than traditional resources among early-career knowledge workers
• LLM-based assessment can replace similarity-based methods because it needs no ground-truth response and no correctness threshold - LLM-based assessment can replace similarity-based methods because it needs no ground-truth response and no correctness threshold
• LLM-based automated assessment of student code self-explanations outperforms semantic similarity methods on a balanced classification task - LLM-based automated assessment of student code self-explanations outperforms semantic similarity methods on a balanced classification task
• LLM-based essay grading is not fully reproducible: repeated ChatGPT sessions on the same submissions show only moderate consistency (Krippendorff's alpha 0.625) - LLM-based essay grading is not fully reproducible: repeated ChatGPT sessions on the same submissions show only moderate consistency (Krippendorff's alpha 0.625)
• LLM-based everyday learning is predominantly an individual activity, with 89% learning alone - LLM-based everyday learning is predominantly an individual activity, with 89% learning alone
• LLM-based extraction of three variables from 156 meta-analyses cost about $50 in computing fees, far less than estimated human-assistant labor costs - LLM-based extraction of three variables from 156 meta-analyses cost about $50 in computing fees, far less than estimated human-assistant labor costs
• LLM-based rubric scores agreed with expert ratings at the total-score level and on cognitive-process dimensions, but agreement was low for more subjective dimensions - LLM-based rubric scores agreed with expert ratings at the total-score level and on cognitive-process dimensions, but agreement was low for more subjective dimensions
• LLM-based scoring of Spanish argumentative writing achieves 60–80% per-item accuracy, slightly below human inter-rater agreement - LLM-based scoring of Spanish argumentative writing achieves 60–80% per-item accuracy, slightly below human inter-rater agreement
• LLM-based scoring produced fewer false negatives than Deep Tutor on a balanced test set, which the authors expect would spare students the frustration of correct explanations judged incorrect - LLM-based scoring produced fewer false negatives than Deep Tutor on a balanced test set, which the authors expect would spare students the frustration of correct explanations judged incorrect
• LLM-based simulated students surpass rule-based systems in natural language generation and situational flexibility, but concerns about bias and evaluation reliability persist - LLM-based simulated students surpass rule-based systems in natural language generation and situational flexibility, but concerns about bias and evaluation reliability persist
• LLM-based systems mediated family education through four recurring forms: producing materials, structuring dialogue, interpreting context, and connecting to external knowledge and values - LLM-based systems mediated family education through four recurring forms: producing materials, structuring dialogue, interpreting context, and connecting to external knowledge and values
• LLM-based tools may interfere with reading comprehension and retention despite learners finding them helpful - LLM-based tools may interfere with reading comprehension and retention despite learners finding them helpful
• LLM-extracted cognitive and linguistic features add predictive value beyond readily available item metadata (grade, domain, item type, word count) - LLM-extracted cognitive and linguistic features add predictive value beyond readily available item metadata (grade, domain, item type, word count)
• LLM-generated 7C collaboration assessment scores fall within the range of human expert variability across ten discussions - LLM-generated 7C collaboration assessment scores fall within the range of human expert variability across ten discussions
• LLM-generated assessment conversations may reveal more information than intended and present scoring challenges compared with pre-structured conversations - LLM-generated assessment conversations may reveal more information than intended and present scoring challenges compared with pre-structured conversations
• LLM-generated difficulty ratings show only moderate rank alignment with empirical student performance on basic arithmetic items - LLM-generated difficulty ratings show only moderate rank alignment with empirical student performance on basic arithmetic items
• LLM-generated distractors for math multiple-choice questions are judged mathematically valid more often than plausible to students, and rarely reproduce the full set of human-written distractors - On Eedi math MCQs, the best LLM method matched all three human-authored distractors on 10.95% of test questions, and teacher-raters scored LLM distractors below human ones, with a larger gap for plaus
• LLM-generated erroneous code aligns with authentic student errors at levels that depend on model and prompting strategy, with Claude Sonnet 4 most consistently aligned - LLM-generated erroneous code aligns with authentic student errors at levels that depend on model and prompting strategy, with Claude Sonnet 4 most consistently aligned
• LLM-generated errors received significantly higher plausibility ratings than authentic student errors, explained by a narrower, cleaner error-type distribution - LLM-generated errors received significantly higher plausibility ratings than authentic student errors, explained by a narrower, cleaner error-type distribution
• LLM-generated feedback is more comprehensive in volume than human feedback but is sometimes too generic, abstract, or misaligned with the assigned grade - LLM-generated feedback is more comprehensive in volume than human feedback but is sometimes too generic, abstract, or misaligned with the assigned grade
• LLM-generated forum replies are more neutral and less varied in sentiment than human replies - LLM-generated forum replies are more neutral and less varied in sentiment than human replies
• LLM-generated ideas diverge from lesson ideas in grounding, misconception substance, vocabulary, uncertainty markers, and analogy use - LLM-generated ideas diverge from lesson ideas in grounding, misconception substance, vocabulary, uncertainty markers, and analogy use
• LLM-generated lessons produced inauthentic references even when combined with RAG over verified academic papers - LLM-generated lessons produced inauthentic references even when combined with RAG over verified academic papers
• LLM-generated misconceptions grounded in quiz content, response patterns, and lecture transcripts are rated excellent by subject matter experts - LLM-generated misconceptions grounded in quiz content, response patterns, and lecture transcripts are rated excellent by subject matter experts
• LLM-generated news sentiment features did not improve ETF price-forecasting model performance in this project - LLM-generated news sentiment features did not improve ETF price-forecasting model performance in this project
• LLM-generated programming code was reliably detected, but inserting beginner bugs had an inconsistent effect across the two programming questions - LLM-generated programming code was reliably detected, but inserting beginner bugs had an inconsistent effect across the two programming questions
• LLM-generated responses were substantially longer than authentic student submissions, suggesting response length as a signal the detector may exploit - LLM-generated responses were substantially longer than authentic student submissions, suggesting response length as a signal the detector may exploit
• LLM-generated retrieval practice was practical to deploy and well received, but high correctness rates (about 70% of items at p=1.0) restricted score variance - LLM-generated retrieval practice was practical to deploy and well received, but high correctness rates (about 70% of items at p=1.0) restricted score variance
• LLM-generated robot control code succeeds on 100%, 94.4%, and 88.9% of simple, composite, and complex instructions respectively - LLM-generated robot control code succeeds on 100%, 94.4%, and 88.9% of simple, composite, and complex instructions respectively
• LLM-group participants reported low ownership of their essays and fell behind in their ability to quote from essays written minutes earlier - LLM-group participants reported low ownership of their essays and fell behind in their ability to quote from essays written minutes earlier
• LLM-human factor-structure similarity is consistently lower than human-human similarity on both assessments - LLM-human factor-structure similarity is consistently lower than human-human similarity on both assessments
• LLM-judge evaluation of tutoring sycophancy shows systematic self-judge blind spots and missed sycophancy even under judge consensus - LLM-judge evaluation of tutoring sycophancy shows systematic self-judge blind spots and missed sycophancy even under judge consensus
• LLM-rewritten Python error messages improved subjective evaluations but did not reliably improve objective debugging performance - LLM-rewritten Python error messages improved subjective evaluations but did not reliably improve objective debugging performance
• LLM-simulated student ideas moderately overlap with teacher- and student-generated lesson ideas (mean cosine similarity 0.52), with larger LLMs slightly outperforming Mistral-7B - LLM-simulated student ideas moderately overlap with teacher- and student-generated lesson ideas (mean cosine similarity 0.52), with larger LLMs slightly outperforming Mistral-7B
• LLM-simulated students' self-reported ADHD persona scores remain stable throughout extended multi-turn conversations - LLM-simulated students' self-reported ADHD persona scores remain stable throughout extended multi-turn conversations
• LLM-supported socio-scientific inquiry lowered cognitive load but produced lower-quality reasoning and argumentation than conventional search - LLM-supported socio-scientific inquiry lowered cognitive load but produced lower-quality reasoning and argumentation than conventional search
• LLMKT outperforms existing knowledge tracing methods at predicting student turn correctness in the CoMTA and MathDial tutoring dialogue datasets, and generally outperforms DKT-Sem. - LLMKT outperforms existing knowledge tracing methods at predicting student turn correctness in the CoMTA and MathDial tutoring dialogue datasets, and generally outperforms DKT-Sem.
• LLMKT's predicted knowledge change curves on CoMTA are mixed across the 15 most frequent KCs, though overall they mostly resemble the power law of practice when dialogues have sufficient turns. - LLMKT's predicted knowledge change curves on CoMTA are mixed across the 15 most frequent KCs, though overall they mostly resemble the power law of practice when dialogues have sufficient turns.
• LLMs across five models generate diverse erroneous Java code, with mean AST edit distances varying by model and prompting strategy - LLMs across five models generate diverse erroneous Java code, with mean AST edit distances varying by model and prompting strategy
• LLMs act as catalysts for collaboration by aiding idea generation, organization and peer feedback, and by democratizing access to knowledge in large-class settings - LLMs act as catalysts for collaboration by aiding idea generation, organization and peer feedback, and by democratizing access to knowledge in large-class settings
• LLMs agree strongly with their own consensus grades (most ICC > 0.8), far more than with human teachers' grades - LLMs agree strongly with their own consensus grades (most ICC > 0.8), far more than with human teachers' grades
• LLMs and instructors agree broadly on where feedback belongs but diverge on exact sentence spans, with span overlap strongly goal-dependent and LLMs highlighting more of each essay - LLMs and instructors agree broadly on where feedback belongs but diverge on exact sentence spans, with span overlap strongly goal-dependent and LLMs highlighting more of each essay
• LLMs are more intrusive and more consistent than humans, eager to view neutral discussions as opportunities for positive reinforcement - LLMs are more intrusive and more consistent than humans, eager to view neutral discussions as opportunities for positive reinforcement
• LLMs are more stable than supervised models across dataset shifts, though they do not surpass within-dataset fine-tuned performance - LLMs are more stable than supervised models across dataset shifts, though they do not surpass within-dataset fine-tuned performance
• LLMs can deliver scalable, personalized, interactive feedback almost instantly, reducing dependency on teacher availability and supporting self-regulated learning in large courses - LLMs can deliver scalable, personalized, interactive feedback almost instantly, reducing dependency on teacher availability and supporting self-regulated learning in large courses
• LLMs cover most teacher feedback focus types but fail to match teacher feedback focus distributions, with LearnLM closest - LLMs cover most teacher feedback focus types but fail to match teacher feedback focus distributions, with LearnLM closest
• LLMs diagnose Applying skills better than Knowing skills, but with higher evidence over-attribution and false-attribution - LLMs diagnose Applying skills better than Knowing skills, but with higher evidence over-attribution and false-attribution
• LLMs differ systematically in grading strictness, forming liberal, restrictive, and balanced grading styles on programming assignments - LLMs differ systematically in grading strictness, forming liberal, restrictive, and balanced grading styles on programming assignments
• LLMs exhibit an inherent abstraction bias, consistently generating questions at higher-than-intended cognitive levels regardless of prompting strategy - LLMs exhibit an inherent abstraction bias, consistently generating questions at higher-than-intended cognitive levels regardless of prompting strategy
• LLMs exhibit demographic sensitivity: systematic output differences attributable solely to demographic context when task inputs are held constant - LLMs exhibit demographic sensitivity: systematic output differences attributable solely to demographic context when task inputs are held constant
• LLMs expressed unjustified confidence in recommendations based on visual inputs, asserting incorrect pin-connection diagnoses with high stated certainty - LLMs expressed unjustified confidence in recommendations based on visual inputs, asserting incorrect pin-connection diagnoses with high stated certainty
• LLMs frequently over-attribute evidential strength and assert Evident evidence even for incorrect diagnoses, and the two tendencies are strongly correlated - LLMs frequently over-attribute evidential strength and assert Evident evidence even for incorrect diagnoses, and the two tendencies are strongly correlated
• LLMs help bridge the gap between theoretical curricula and real-world skills by creating simulated scenarios, cases and problem contexts that are hard to reproduce in traditional classrooms - LLMs help bridge the gap between theoretical curricula and real-world skills by creating simulated scenarios, cases and problem contexts that are hard to reproduce in traditional classrooms
• LLMs in educational applications hallucinate in 15–25% of responses and can amplify bias against non-native language inputs - LLMs in educational applications hallucinate in 15–25% of responses and can amplify bias against non-native language inputs
• LLMs overrate short underdeveloped essays and over-penalize minor language errors in well-developed essays - LLMs overrate short underdeveloped essays and over-penalize minor language errors in well-developed essays
• LLMs perform competently at comparing item difficulty but fall below random chance on comparing item discrimination - LLMs perform competently at comparing item difficulty but fall below random chance on comparing item discrimination
• LLMs perform worse on complex, open-ended, or subjective assignments than on short, clearly formulated answers - LLMs perform worse on complex, open-ended, or subjective assignments than on short, clearly formulated answers
• LLMs predict common distractors above chance but fall 15 percentage points below chance at predicting which errors students rarely select - LLMs predict common distractors above chance but fall 15 percentage points below chance at predicting which errors students rarely select
• LLMs reconstruct highly cited researchers' coauthor lists at approximately twice the rate of lower-cited peers, with the disparity growing with model size - LLMs reconstruct highly cited researchers' coauthor lists at approximately twice the rate of lower-cited peers, with the disparity growing with model size
• LLMs redistributed educational labor while family and institutional rules left parents and professionals responsible for interpreting outputs and deciding how they entered practice - LLMs redistributed educational labor while family and institutional rules left parents and professionals responsible for interpreting outputs and deciding how they entered practice
• LLMs show a directional asymmetry in educational control: they reliably increase cognitive demand but struggle to lower it - LLMs show a directional asymmetry in educational control: they reliably increase cognitive demand but struggle to lower it
• LLMs show a systematic optimism bias in knowledge tracing, predicting far more accurately for students who answer correctly than for those who answer incorrectly - LLMs show a systematic optimism bias in knowledge tracing, predicting far more accurately for students who answer correctly than for those who answer incorrectly
• LLMs show robust zero- and few-shot Bloom classification of OOD AI-assisted questions compared with ML and BERT models - LLMs show robust zero- and few-shot Bloom classification of OOD AI-assisted questions compared with ML and BERT models
• LLMs simulating survey respondents under AI-assistant instructions report higher mathematics self-efficacy and lower anxiety than under human-shadow conditioning - LLMs simulating survey respondents under AI-assistant instructions report higher mathematics self-efficacy and lower anxiety than under human-shadow conditioning
• LLMs struggle most with Chemistry riddles, with best EM accuracy at or below 75% versus above 90% in other subjects - LLMs struggle most with Chemistry riddles, with best EM accuracy at or below 75% versus above 90% in other subjects
• LLMs support the shift from memorization-focused to competency-based assessment by generating rubrics, varied assessment items and authentic scenario-based tasks - LLMs support the shift from memorization-focused to competency-based assessment by generating rubrics, varied assessment items and authentic scenario-based tasks
• LLMs systematically underestimate fraction items whose difficulty is driven by misconceptions (the "Easy Trap") - LLMs systematically underestimate fraction items whose difficulty is driven by misconceptions (the "Easy Trap")
• LLMs use fewer extreme grades than human examiners, with grade-dependent opposing biases compressing grades toward the middle - LLMs use fewer extreme grades than human examiners, with grade-dependent opposing biases compressing grades toward the middle
• LLMs' generation probabilities for multiple-choice distractors correlate moderately with how often students choose them - Across 3,202 four-option MCQs with real student response distributions, LLaMA and Qwen models (0.5B–72B) assigned distractor probabilities that correlated with student selection at Pearson r = 0.28–0.
• LMS activity contributes only modestly to explained variance in exam grades, and gender-by-LMS interactions add no explained variance in any course - LMS activity contributes only modestly to explained variance in exam grades, and gender-by-LMS interactions add no explained variance in any course
• LMS log analysis showed groups spent roughly 9 days to 2 weeks per module after M1, and completed implementation (M5) faster than expected - LMS log analysis showed groups spent roughly 9 days to 2 weeks per module after M1, and completed implementation (M5) faster than expected
• LMS process indicators documented adaptive engagement during the eight-week intervention, with most experimental students reaching mastery by unit end - LMS process indicators documented adaptive engagement during the eight-week intervention, with most experimental students reaching mastery by unit end
• LMS quality did not significantly influence academic performance and did not independently mediate the AI literacy–performance relationship - LMS quality did not significantly influence academic performance and did not independently mediate the AI literacy–performance relationship
• LMS quality predicts blended-learning satisfaction, which predicts academic performance (β = .604 and β = .535) - LMS quality predicts blended-learning satisfaction, which predicts academic performance (β = .604 and β = .535)
• LO assessment supported by CA tools prompted faculty toward more authentic, participatory, and reflective teaching practices - LO assessment supported by CA tools prompted faculty toward more authentic, participatory, and reflective teaching practices
• Loaded moments (de)escalate through moment-by-moment co-construction such as mutual trade-offs, turnings, or refusals - Loaded moments (de)escalate through moment-by-moment co-construction such as mutual trade-offs, turnings, or refusals
• Loaded moments emerge when instigating circumstances co-construct incompatibilities between teacher and students - Loaded moments emerge when instigating circumstances co-construct incompatibilities between teacher and students
• Loan conditionalities compel policy reform, with SAP-era requirements later evolving into subtler forms such as Poverty Reduction Strategy Papers and impact-evaluation demands - Loan conditionalities compel policy reform, with SAP-era requirements later evolving into subtler forms such as Poverty Reduction Strategy Papers and impact-evaluation demands
• Loans are the most common pathway of World Bank influence, and their scale and discretionary weight amplify their impact - Loans are the most common pathway of World Bank influence, and their scale and discretionary weight amplify their impact
• Local costs (switch implementation) showed no significant dependence on inhibition or working-memory performance at any of the preschool ages tested - Local costs (switch implementation) showed no significant dependence on inhibition or working-memory performance at any of the preschool ages tested
• Local participants learning a foreign language show the strongest attitude–proficiency association; immigrant participants learning a local language the weakest - Local participants learning a foreign language show the strongest attitude–proficiency association; immigrant participants learning a local language the weakest
• Local School Council decisions collectively produced racialized differences in SRO presence, with Black students more than twice as likely as other students to have an SRO in their school - Local School Council decisions collectively produced racialized differences in SRO presence, with Black students more than twice as likely as other students to have an SRO in their school
• Localizing word-problem character names improved MATHia performance only for specific subgroup-majority schools, with no improvement in sense of belonging overall - Localizing word-problem character names improved MATHia performance only for specific subgroup-majority schools, with no improvement in sense of belonging overall
• Lockdown measures are positively associated with expected political participation but show null or negative associations with expected electoral participation - Lockdown measures are positively associated with expected political participation but show null or negative associations with expected electoral participation
• Lockdown stringency and school closure duration are associated with greater declines in student trust in civic institutions - Lockdown stringency and school closure duration are associated with greater declines in student trust in civic institutions
• Lockdown training and drill participation improved students' knowledge and perceptions of preparedness without increasing anxiety, per the studies the brief reviews - Lockdown training and drill participation improved students' knowledge and perceptions of preparedness without increasing anxiety, per the studies the brief reviews
• Lockdown was significantly associated with higher disengagement (β = 0.12) but not with self-efficacy among university students - Lockdown was significantly associated with higher disengagement (β = 0.12) but not with self-efficacy among university students
• Locus of control is negatively correlated with and predicts subjective well-being, ranking second after self-esteem - Locus of control is negatively correlated with and predicts subjective well-being, ranking second after self-esteem
• Log analysis reveals novel post-acceptance behaviors (slow tab accept, slow accept then modify) that show suggestion acceptance is a multi-step process - Log analysis reveals novel post-acceptance behaviors (slow tab accept, slow accept then modify) that show suggestion acceptance is a multi-step process
• Log-based behaviours predicted problem-solving performance in the DI-PS group but hardly explained knowledge acquisition in either group - Log-based behaviours predicted problem-solving performance in the DI-PS group but hardly explained knowledge acquisition in either group
• Log-based evaluation showed chatbot alignment was stronger for responsiveness (88.9%) and persona (81.5%) than for rules (70.4%) and purpose (59.3%) - Log-based evaluation showed chatbot alignment was stronger for responsiveness (88.9%) and persona (81.5%) than for rules (70.4%) and purpose (59.3%)
• Log-likelihood comparisons indicate Pychometrik can produce more accurate item parameter estimates than the existing tool for items with RIT differences larger than 2 - Log-likelihood comparisons indicate Pychometrik can produce more accurate item parameter estimates than the existing tool for items with RIT differences larger than 2
• Logged time investment on an online mathematics preparation platform differed significantly across SRL intervention groups, rising from control through diary and training + diary to training + diary + peer feedback groups - Logged time investment on an online mathematics preparation platform differed significantly across SRL intervention groups, rising from control through diary and training + diary to training + diary +
• Logistic regression can identify the ELP score at which students have a probability of 0.5 of being proficient on the academic content assessment, as a candidate English-proficient standard - Logistic regression can identify the ELP score at which students have a probability of 0.5 of being proficient on the academic content assessment, as a candidate English-proficient standard
• Logistic regression performs poorly (mean accuracy 0.4243) and SVM moderately (0.7374) on physiological exam-outcome prediction - Logistic regression performs poorly (mean accuracy 0.4243) and SVM moderately (0.7374) on physiological exam-outcome prediction
• Logistic regression shows all traced KAM practices except fading are significant positive predictors of adoption, with maturation strongest (odds ratios up to 12.36) and model AUC = 0.79 - Logistic regression shows all traced KAM practices except fading are significant positive predictors of adoption, with maturation strongest (odds ratios up to 12.36) and model AUC = 0.79
• Loneliness functions as a biological social-feedback signal that motivates relationship cultivation, and AI companions may silence that signal - Loneliness functions as a biological social-feedback signal that motivates relationship cultivation, and AI companions may silence that signal
• Long Beach schools with assigned math coaches showed markedly and disproportionately rising CAASPP scores, though causation could not be tied to coaching - Long Beach schools with assigned math coaches showed markedly and disproportionately rising CAASPP scores, though causation could not be tied to coaching
• Long practice periods were not much less favorable than short ones in addition practice, with evidence favoring the longer periods - Long practice periods were not much less favorable than short ones in addition practice, with evidence favoring the longer periods
• Long school days or years are among the policies most consistently associated with positive charter-school impacts - Long school days or years are among the policies most consistently associated with positive charter-school impacts
• Long school days or years, comprehensive behavioral policies with rewards and sanctions, and an achievement-focused mission are most strongly associated with charter school success - Long school days or years, comprehensive behavioral policies with rewards and sanctions, and an achievement-focused mission are most strongly associated with charter school success
• Long video-based modelling examples did not improve diagnostic strategy knowledge, skills, self-efficacy, or cognitive load relative to independently practising in a computer simulation - Long video-based modelling examples did not improve diagnostic strategy knowledge, skills, self-efficacy, or cognitive load relative to independently practising in a computer simulation
• Long-term academic benefits of preschool are almost entirely mediated by children's earlier skill development - Long-term academic benefits of preschool are almost entirely mediated by children's earlier skill development
• Long-term and late-arriving English Learners were considerably less likely to enroll in four-year colleges than their peers - Long-term and late-arriving English Learners were considerably less likely to enroll in four-year colleges than their peers
• Long-term Buddhist meditation training induces gamma synchrony whose ratio to slow oscillations correlates with cumulative practice hours (r = 0.79) - Long-term Buddhist meditation training induces gamma synchrony whose ratio to slow oscillations correlates with cumulative practice hours (r = 0.79)
• Long-term effects of grading mismatch are smaller and more nuanced than those of persistent effects - Long-term effects of grading mismatch are smaller and more nuanced than those of persistent effects
• Long-term effects of systematic phonics are short-lived, reducing from d = 0.29 immediately after training to d = 0.07 after about a year - Long-term effects of systematic phonics are short-lived, reducing from d = 0.29 immediately after training to d = 0.07 after about a year
• Long-term English Learners with IEPs were on-track in ninth grade at near district-average rates but graduated at substantially lower rates, similar to non-English Learners with IEPs - Long-term English Learners with IEPs were on-track in ninth grade at near district-average rates but graduated at substantially lower rates, similar to non-English Learners with IEPs
• Long-term English Learners without IEPs had the lowest Freshman OnTrack rate and substantially lower graduation rates than district average - Long-term English Learners without IEPs had the lowest Freshman OnTrack rate and substantially lower graduation rates than district average
• Long-term Hispanic ELLs had the worst course performance and graduation rates, with only 52 percent graduating within four years - Long-term Hispanic ELLs had the worst course performance and graduation rates, with only 52 percent graduating within four years
• Long-term learning gains may be optimized in collaborative situations that display discouraging short-term process costs - Long-term learning gains may be optimized in collaborative situations that display discouraging short-term process costs
• Long-term memory is effectively unlimited and holds schemas that vary in their degree of automation - Long-term memory is effectively unlimited and holds schemas that vary in their degree of automation
• Long-term nature exposure shows inconsistent effects on processing speed, positive in the greened schoolyard study but null in the classroom green wall study - Long-term nature exposure shows inconsistent effects on processing speed, positive in the greened schoolyard study but null in the classroom green wall study