Artificial intelligence is increasingly embedded in U.S. educational institutions for tasks such as dropout prediction and student performance monitoring, yet these systems introduce intertwined legal, ethical, and fairness risks. This study develops and evaluates a regulatory-aligned AI pipeline that integrates fairness auditing, bias mitigation, and privacy preservation within an educational context. Using a privacy-safe synthetic dataset modeling realistic demographic, academic, and behavioral patterns, we benchmark five machine-learning models, Logistic Regression, Random Forest, XGBoost, MLP, and SVM, across baseline, fairness-aware, and privacy-enhanced conditions. Fairness audits conducted with the Fairlearn framework reveal notable disparities across academic-risk and access groups, particularly in selection-rate metrics. A manually implemented reweighing mechanism and adaptive thresholding substantially narrow these gaps with only marginal losses in predictive performance. Differential-privacy simulation through Gaussian noise injection demonstrates that privacy reinforcement entails a measurable but manageable accuracy reduction (~1–2%). A human-in-the-loop policy layer emulates U.S. regulatory requirements under the AI Bill of Rights and FERPA by designating high-risk predictions for human review rather than full automation. Collectively, results show that a governance-first machine-learning workflow can achieve strong predictive validity while satisfying emerging ethical and legal expectations for accountability, fairness, and privacy in educational AI deployment. This framework provides a replicable reference architecture for responsible AI adoption across academic institutions and education-technology providers.

Abstract Phosphor‐converted white light‐emitting diodes are transforming energy‐efficient lighting technologies. At the core of their performance are Ce 3+ and Eu 2+ activated phosphors, whose photoluminescent properties significantly impact the lighting devices’ efficiency and, equally important, govern the light quality. However, the discovery of new phosphors that meet the stringent requirements needed for commercialization has remained a challenge. Recent advances in data‐driven approaches, particularly machine learning and optimization algorithms, have begun to accelerate this process. These methods enable the accurate prediction of key photoluminescent properties and facilitate the exploration of vast compositional spaces of inorganic phosphors, guiding the targeted synthesis of new high‐performance phosphors. This review highlights the major progress in data‐driven discovery of Ce 3+ and Eu 2+ phosphors, emphasizing property prediction, materials screening, and experimental validation. It concludes with an outlook on future opportunities and challenges in the application of artificial intelligence to phosphor discovery.

The predictive validity of machine learning models depends on the training data. In some cases, training data contains historical, social, or demographic inequalities, which leads algorithms to reproduce unfair results. This paper proposes a fairness-constrained stacking meta-learning approach for reducing bias in classification by aggregating a set of classifiers through a constrained ensemble learning scheme. A set of base classifiers, including Decision Tree, Naive Bayes, Support Vector Machine (SVM), and LightGBM, are trained and evaluated on the Adult Census Income dataset using both predictive and fairness metrics. The final meta-model is constructed as an aggregation of only the fair-performing models, while models failing to meet the fairness threshold are excluded. Learned weights are then optimized to maximize the F1-score while maintaining fairness constraints. Experimental results demonstrate that the proposed method achieves predictive performance (Accuracy = 0.91, F1-score = 0.82) while substantially reducing disparity between demographic groups (EOD = 0.03 for sex and 0.04 for race). These findings indicate that fairness-aware stacking ensembles can provide a solution for mitigating algorithmic bias through an aggregation framework that balances accuracy and fairness.

The Structured Assessment of Protective Factors (SAPROF) is a measure of protective factors intended to augment violence risk assessment. While prior research supports the predictive validity of SAPROF ratings, factorial and convergent validity have been underexamined, each of which is required to ensure that the instrument measures intended targeted constructs and converges with test scores from established measures. We evaluated the structural and convergent properties of SAPROF ratings through examining its factor structure and convergence with measures of relevant constructs, as a function of ethnocultural heritage, in a treated sample of 461 men with sexual offense conviction histories. The SAPROF was rated from institutional files pre-and posttreatment. Results of exploratory structural equation modeling (ESEM) of pre and post SAPROF item ratings identified a temporally stable three-factor model that was invariant across ethnocultural groups; however, it departed from the developers' original subscale structure-the factors were termed Internal-Prosocial, Motivational-Lifestyle, and External-to reflect continuity with, yet departure from, the current subscale structure. SAPROF ratings were correlated in theoretically and clinically meaningful ways with scores on relevant risk-need-responsivity (RNR) measures. The results support the structural and convergent validity of SAPROF ratings and identified a slightly modified subscale structure in the present sample.

In an innovative simulation study, Perley-Robertson et al. found that two correctional risk assessment tools were robust to missing data, with summation, proration, and multiple imputation producing nearly identical relative predictive validity results. However, the uniform deletion of items across cases may have preserved their risk rankings and, consequently, relative predictive accuracy. We extend this research by applying identical missing data conditions (1%-50% of items deleted in 10% increments) to one third, two thirds, and three thirds of a high-risk intimate partner violence (IPV) sample assessed on the Ontario Domestic Assault Risk Assessment (ODARA) and Spousal Assault Risk Assessment-Version 2 (SARA-V2; N = 267). Neither missing data nor the handling method affected relative predictive accuracy, though summation underestimated absolute risk. These findings support proration or multiple imputation when IPV risk scale items are missing within a research sample, and underscore that proration is preferable to summed totals in practice.

Human–AI collaboration is redefining how cognitive assessment is conducted in modern workplaces, blending computational precision with human interpretive judgment. This paper investigates methodological advances that integrate artificial intelligence into psychometric testing and cognitive evaluation to enhance workplace productivity and talent management. Traditional assessments often suffer from evaluator bias, inconsistent scoring, and limited scalability, while AI systems offer adaptive testing, real-time analytics, and pattern recognition that improve reliability and objectivity. However, the absence of human contextual interpretation can limit AI’s effectiveness in capturing emotional and situational nuances. To address this, the study proposes a hybrid assessment framework where AI models assist human experts in evaluating cognitive flexibility, problem-solving, and emotional intelligence through multimodal data, including linguistic and behavioral cues. Using correlation analysis and performance-based validation, results show that human–AI collaboration significantly improves predictive validity of job performance indicators by 18–22% over traditional methods. The study emphasizes the importance of transparent algorithmic processes and ethical oversight to ensure fairness and inclusivity. Overall, this research advances methodological innovation in cognitive assessment, paving the way for data-driven, human-centered talent management systems that balance automation with empathy and contextual insight.

Mobile sports betting is an increasingly popular way to wager in the United States, but few studies have examined the social cognitive predictors of engaging in this specific form of gambling. The present studies were designed to do so with the application of the Reasoned Action Approach (RAA; Fishbein & Ajzen, 2010). Study 1 was a belief elicitation study that identified the most common behavioral, normative, and control beliefs regarding mobile sports betting. Study 1 also reported an initial test of the degree to which attitude, perceived normative pressure, and perceived behavioral control predict intentions to place mobile sports bets. Study 2 examined the predictive validity of the identified beliefs and tested whether future betting behavior can be predicted from the model. The RAA significantly predicted mobile sports gambling intentions and future behavior. The most important beliefs contributing to the prediction of mobile sports gambling were the degree that participants endorsed gambling is fun and easy, leads to addiction, and results in winning money. Implications for future research and interventions are discussed.

Dampening of positive affect (PA) constitutes a transdiagnostic risk and maintenance factor for affective dysregulation in various psychopathologies, including depression. However, the motives underlying this PA downregulation strategy remain unclear, even though they may be highly relevant for improving traditional psychological treatments. This study examined whether avoidance of negative emotional contrasts (NECs) and diminished preference for positive emotions were predictive of dampening. The latter was operationalised as low pro- and high contra-hedonic emotion regulation (ER) goal endorsement. An adult community sample (N = 159) completed an online survey, and multiple linear regressions were conducted to examine the predictive validity of both factors, after controlling for age, gender, and repetitive negative thinking (RNT). Higher levels of NEC avoidance and higher contra-hedonic ER goal endorsement were consistently found to uniquely predict concurrent dampening levels, above and beyond age, gender and RNT. Crucially, inclusion of both factors in the same regression model still yielded evidence for the unique predictive validity of NEC avoidance. Findings support the possibility that dampening is motivated by NEC avoidance rather than solely by emotional preferences. Study limitations are noted, and implications for future research and clinical practice are discussed.

Dietary and nutrient intake directly impact health, whereby adherence to certain dietary patterns is linked to positive outcomes. Traditional methods like the Food Frequency Questionnaire (FFQ) and 24-hour recall are subjective, highlighting the need for advanced techniques that incorporate phenotypic and metabolic data. This pilot exploratory study aimed to assess the feasibility of using machine-learning techniques that integrate routinely collected phenotypic and biochemical data to predict adherence to well-characterized dietary quality indices. A total of 138 participants were recruited in the Dietary Deal cross-sectional study to collect data on dietary intake (FFQ, 24-hour recall), biochemical markers, physical activity estimation, quality-of-life questionnaires, and anthropometric determinations. The Mediterranean Diet Adherence Screener (MEDAS 17p), the Alternative Healthy Eating Index (AHEI), the Dietary Approaches to Stop Hypertension (DASH), and a pro-vegetarian model were tested as quality indices. Biochemical and dietary data were integrated using adjusted logistic regressions through STATA (v. 18.0) statistical program to identify biochemical markers associated with food consumption to predict dietary quality. Subsequently, an algorithm based on machine-learning techniques was developed, and the predictive capacity of the obtained models was determined using receiver operating characteristic (ROC) curves and related metrics (area under the curve). A computational algorithm was created for probability classification, adjusted for age, sex, body mass index, physical activity, and SF-36. Key biochemical parameters included glucose, triglycerides, high-density lipoprotein cholesterol, homocysteine, and albumin. Homocysteine (p = 0.007 for AHEI, p = 0.040 for pro-vegetarian), folate (p = 0.039 for DASH, p = 0.019 for pro-vegetarian), and vitamin C (p < 0.001 for AHEI, p = 0.023 for DASH) emerged as significant variables across diet quality indices. The explanatory capacity of the fully adjusted model ranged from R2 = 22.07% to 35.76%, depending on the index. The model's accuracy ranged from 72.46% to 78.26%, with ROC values between 0.79 and 0.87, indicating moderate to good predictive validity of the training data on itself. This pilot exploratory analysis demonstrates the feasibility of integrating dietary and biochemical data to suitably predict adherence to validated dietary quality indices, Although not intended as a deployable prediction tool, the study provides preliminary evidence supporting the potential of routinely collected clinical data to inform personalized precision dietary advice through objective computational algorithms for precision nutrition implementation.

Objectives: This study investigated the temporal dynamics of category-based attentional orienting (CAO) under the influences of prediction (top-down) and perceptual load (bottom-up) across color and shape dimensions, combining behavioral and event-related potential (ERP) measures. Methods: Across two experiments, we manipulated predictive validity and perceptual load during a visual search for category-defined targets. Results: The results revealed a critical dimension-specific effect of prediction: invalid predictions elicited a larger N2pc component (indexing attentional selection) for shape-defined targets, but not color-defined targets, indicating that shape CAO relies more heavily on predictive information during early processing. At the behavioral level, a combined analysis of the two experiments revealed an interaction between prediction and perceptual load on accuracy, suggesting their integration can occur at later stages. Conclusions: These findings demonstrate that prediction and perceptual load exhibit distinct temporal profiles, primarily independently modulating early attentional orienting, with their interactive effects on behavior being more nuanced and dimension-dependent. This study elucidates the distinct temporal and dimensional mechanisms through which top-down and bottom-up sources of uncertainty shape attentional orienting to categories.

A frailty-intrinsic capacity index to predict disability in community-dwelling older Japanese adults

Healthcare discrimination is an important barrier to accessing services among Latino populations in the United States. However, few validated scales have been developed to systematically examine this issue. In this study, we evaluated the validity and reliability of a bilingual healthcare discrimination scale in a sample of churchgoing Latino adults in Los Angeles, California. The study sample included 336 participants (foreign-born: 250; US-born: 86) who attended 12 Catholic churches in Los Angeles. Psychometric testing of the 7-item healthcare discrimination (HCD) scale included internal consistency; split-half reliability; convergent, discriminant, and predictive validity; and confirmatory factor analyses. The HCD had relatively high internal consistency (full sample Cronbach’s α = 0.92; foreign-born: 0.91; US-born: 0.92) and showed good convergent and discriminant validity, as it was moderately correlated with the depression scale (full sample r = 0.28, p < 0.001) and weakly correlated with the acculturation scale (full sample r = 0.15, p = 0.008). Confirmatory factor analyses yielded further support for a one-factor solution. Our study finds that the HCD is a valid and reliable scale for use among churchgoing Latino adult populations in the United States. Future studies should examine the psychometric properties of the HCD among Latinos of diverse backgrounds, geographic locations, religious beliefs, and languages.

ABSTRACT The present study aimed to validate the Bergen Facebook Addiction Scale (BFAS). In study 1, personality dimensions were employed to establish criterion and predictive validity on a sample size of 120 young adults in the age group of 18–23 years who participated voluntarily. The EFA revealed a 12 items, three-factor solution (Time, Withdrawal and Mood Modification) with 65.86% of variance. To validate correlations were examined between total BFAS score, BFAS dimensions, total NEO-FFI score, NEO-FFI dimensions, BIS and BAS. Hierarchical Regression indicated that age and gender had predictive efficiency of 15% whereas in step 2 Age, Gender, Neuroticism, Extraversion, Openness, Agreeableness, Conscientiousness, BIS, BAS Drive, BAS Fun Seeking and BAS Reward Responsiveness predicted Facebook Addiction to 12.3%. In study 2, good factor solution was revealed thus, confirming the three-factor solution from study 1. The overall results of this studies indicated that a three-factor solution was more relevant for Indian adolescents.

Usefulness of 3D joint space width on weight-bearing CT in comparison with 2D joint space width on radiographs for predicting 24-month worsening of knee osteoarthritis pain and function in the MOST study.

This study reports a five-layer surface plasmon resonance biosensor architecture comprising a BK7 glass substrate, silver plasmonic film, monolayer graphene, black phosphorus dielectric, and analyte region, engineered for high-precision detection of low refractive index media. The graphene–black phosphorus heterostructure synergistically exploits the exceptionally high surface-to-volume ratio of graphene and the anisotropic optical response of black phosphorus, enabling pronounced electromagnetic field confinement at the sensor interface. In particular, the detection procedure is mainly dependent on the emergence of the angular surface plasmon resonance based on the optimum values of the different geometrical and structural parameters. Therefore, the electromagnetic optimization using COMSOL Multiphysics was performed by varying the silver thickness, graphene thickness and black phosphorus thickness over an analyte index range of 1.29–1.38 RIU. The optimized configuration achieved a maximum sensitivity of 300°/RIU at n = 1.35 RIU, with a figure of merit of 45.455 RIU–1 and a detection limit of 0.018 RIU, surpassing previously reported architectures. Furthermore, predictive validation employing K-nearest neighbours regression demonstrated excellent reliability, yielding R² values between 92–100% and mean absolute errors of 0.005–0.012 RIU.

Social health has been increasingly recognized as an important determinant of dementia progression and quality of life. The Social Network Index (SNI), developed by Cohen, is widely used to assess social networks as a proxy for social health. This study aimed to adapt the SNI cross-culturally and validate its psychometric properties for use among Indonesians with dementia. A cross-sectional study was conducted at the Memory Clinic of Dr. Sardjito General Hospital, Yogyakarta, Indonesia, involving 56 individuals with mild to moderate dementia. The original 12-item SNI was translated and culturally adapted according to WHO guidelines. Construct validity was examined using Confirmatory Factor Analysis (CFA). Composite Reliability (CR) and Average Variance Extracted (AVE) were used to assess internal consistency. The finalized modified Indonesian version of the SNI (m-SNI-INA), consisting of 10 items (2 excluded due to low factor loadings), demonstrated strong construct validity with factor loadings >0.5 and good model fit (GFI=0.958, AGFI=0.934, CFI=1.000, RMSEA=0.000). Reliability was high (CR=0.91; AVE=0.53). The m-SNI-INA is a valid, reliable, and culturally adapted tool for assessing social networks in people with dementia in Indonesia. Further studies should examine its predictive validity in larger populations.

This study develops and validates a multi-dimensional identity-based scale for measuring LGBTQ+-friendly destination images, grounded in the sexual orientation, gender identity, and gender expression dimensions. Through qualitative and empirical testing with 654 international tourists in Thailand, the scale demonstrates strong psychometric properties, including reliability, convergent validity, discriminant validity, and nomological validity. The overall LGBTQ+-friendly destination image significantly predicts tourists’ satisfaction, word-of-mouth, and revisit intentions, highlighting its predictive validity. Notably, LGBTQ+ friendliness has universal relevance; while friendliness toward diverse sexual orientations and gender expressions is a central determinant for all tourists, including non-LGBTQ+ travelers, friendliness toward diverse gender identities holds particular importance for LGBTQ+ travelers. The scale is designed for broad application in assessing LGBTQ+-friendly destination images in diverse tourism contexts. This validated tool offers destination marketing organizations a comprehensive framework to assess and enhance LGBTQ+ friendliness, benefiting both LGBTQ+ and non-LGBTQ+ travelers through its association with general safety and openness.

ObjectiveTo analyze the influencing factors of postoperative hypotension (POH) following video-assisted thoracoscopic lung resection (VATS) and evaluate a predictive model combining non-invasive hemodynamic parameters with the inferior vena cava collapsibility index (IVCCI).MethodsA prospective study enrolled 100 VATS patients (September 2024–March 2025). Patients were stratified into POH (n = 36) and Non-POH (n = 64) groups based on mean arterial pressure (MAP ≤ 65 mmHg or ≥ 30% reduction from baseline) within 24 hours postoperatively. Hemodynamic parameters (cardiac output [CO], systemic vascular resistance [SVR], stroke volume [SV], stroke volume variation [SVV], left ventricular stroke work [LVSW]) were monitored using the Non-invasive Continuous Arterial Blood Pressure And Cardiac Output Monitoring System. IVCCI was ultrasonographically measured post-extubation. Linear regression analyzed correlations between post-anesthetic emergence period hemodynamic parameters and POH, while multivariable logistic regression analysis was employed to identify predictive factors, leading to the development and validation of a clinical prediction model.ResultsThe incidence of early POH was 36%. Multivariate analysis demonstrated that a model combining post-anesthesia emergence period left ventricular stroke work (PA_LVSW, OR = 0.880, P < 0.01), IVCCI (OR = 1.095, P = 0.01), baseline MAP, and ASA achieved an AUC of 0.940 (95% CI: 0.895–0.985), with 83.3% sensitivity and 89.1% specificity, outperforming individual predictors (IVCCI: AUC = 0.65; PA_LVSW༚AUC = 0.84).ConclusionEarly POH after VATS is closely associated with cardiac function suppression and volume status imbalance. The multiparameter model integrating PA_LVSW, IVCCI, ASA physical status, and baseline MAP effectively predicts POH.Trial registrationChinese Clinical Trial Register, ChiCTR2500100275. Registered 7 April 2025 Retrospectively registered, https//www.chictr.org.cn/showprojEN.html? proj=259,898.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12871-025-03440-0.

Enhancing predictive validity in the english placement test: evidence and insights from a Thai public university

The Stress and Adversity Inventory for Adults (Adult STRAIN) systematically assesses the count, severity, timing, types of lifetime stressors, primary life domains, and core social-psychological characteristics. The study aimed to replicate findings from the original Adult STRAIN validation study with a sample of middle-aged and older African American adults. Participants from the Health among Older Adults Living in Detroit study [N = 200; M(SD) = 67.48 years old (8.53), range = 50-89; 74.50% female], completed two home visits, daily diaries, and questionnaires. Pearson correlations and regression models assessed concurrent, discriminant, predictive, and comparative predictive validity (vs. the Perceived Stress Scale-4 and Risky Family Questionnaire) of the Adult STRAIN to stress-related (e.g. subjective physical health) and expected unrelated outcomes (e.g. personality variables), with and without covariates. The present study provides evidence of the Adult STRAIN as a valid measure of cumulative lifetime stressor exposure in older African American adults.