Measurement Research & Technical Reports

This review presents a structured framework for the development and validation of psychological, psychiatric, and behavioral measurement scales. The authors synthesize contemporary psychometric standards and applied experience into five sequential phases comprising eighteen steps, from construct definition and item generation to factor extraction, reliability assessment, and validation. Emphasis is placed on theoretical clarity, systematic literature review, expert consultation, and engagement with target populations. The article outlines best practices in exploratory and confirmatory factor analysis, exploratory graph analysis, item response theory, and measurement invariance testing. Reliability evaluation includes internal consistency, test–retest stability, agreement indices, and measurement precision. Multiple forms of validity are addressed, including content, criterion-related, convergent, and discriminant validity. The framework is intended to improve methodological rigor, transparency, and reproducibility in scale development across clinical and research contexts.

The APA Handbook of Research Methods in Psychology is a comprehensive three-volume reference that presents the full range of research methods used across psychological science and related disciplines. It provides detailed descriptions of methodological approaches, starting with foundational issues such as philosophical frameworks, ethical considerations, and research planning, and progressing to measurement techniques, research designs (quantitative, qualitative, neuropsychological, and biological), and data analysis strategies. Volume sections cover sampling, psychometrics, experimental designs, observational methods, and advanced analytic procedures, highlighting how to align research questions with appropriate methods. The handbook emphasizes practical considerations for conducting high-quality research, including literature searching, workflow and reproducibility, normative data collection, and navigating the publication process. Each chapter synthesizes evidence and offers guidance on selecting, implementing, and evaluating techniques across diverse research contexts. The resource is intended for researchers, advanced graduate students, and practitioners seeking authoritative guidance on rigorous methodological practices and the theoretical underpinnings of research decisions in psychology and allied fields. It underscores the interplay between research design, measurement, and inference, supporting transparency and robust scientific inquiry across domains.

This chapter provides a comprehensive methodological overview of electroencephalography (EEG) and event-related potentials (ERP) as tools for investigating cognitive and affective processes in the social and behavioral sciences. It outlines the biophysical foundations of EEG signal generation, including postsynaptic potentials and volume conduction, and explains how ERPs are derived through time-locked averaging procedures. The chapter details core components of experimental design, including stimulus control, trial structure, artifact management, and signal preprocessing. Emphasis is placed on measurement reliability, temporal resolution, and the interpretation of component amplitudes and latencies in relation to underlying neural processes. The author also discusses common analytic strategies, including component quantification, difference waves, and statistical considerations relevant to repeated-measures designs. Practical guidance is provided regarding electrode configuration, referencing schemes, filtering, and artifact rejection to enhance data quality. The chapter situates EEG and ERP methods within broader research methodology, highlighting their strengths in capturing rapid neural dynamics while acknowledging spatial limitations. Overall, it serves as a foundational reference for researchers seeking to integrate electrophysiological measures into experimental research.

This paper develops a cross-disciplinary methodological framework for evaluating research impact and clarifies key concepts used in impact assessment. Using an adapted grounded theory analysis of peer-reviewed and grey literature, the authors propose definitions of research impact and research impact evaluation, emphasizing that impacts may differ across stakeholder groups and contexts and may be positive or negative. The article introduces five principal impact evaluation designs: experimental and statistical methods, systems analysis approaches, textual, oral and arts-based methods, indicator-based approaches, and evidence synthesis methods. The framework distinguishes between formative and summative purposes and between claims of necessary and sufficient causation. Methodological limitations include attribution challenges, long causal pathways, disciplinary heterogeneity, and risks associated with metric-driven assessment. The framework is intended to improve rigor, transparency, and methodological fit in real-world research impact evaluations.

This systematic review synthesizes research on electroencephalographic (EEG) indices used to quantify human performance across cognitive, operational, and applied environments. The authors evaluate spectral power measures, event-related potentials, functional connectivity indices, entropy metrics, and time–frequency features as candidate biomarkers of workload, vigilance, fatigue, and task efficiency. The review assesses methodological heterogeneity across studies, including electrode configurations, preprocessing pipelines, artifact management strategies, feature extraction procedures, and validation approaches. Particular attention is given to the reliability, sensitivity, and construct validity of commonly reported metrics such as frontal theta, parietal alpha suppression, and P300 amplitude. Limitations identified include small sample sizes, inconsistent reporting standards, variability in task paradigms, and limited ecological validation. The article concludes with recommendations for standardized acquisition protocols, clearer operational definitions, and multimodal validation strategies to improve reproducibility and translational applicability of EEG-based performance measurement.

This narrative review synthesizes current research on automated pain recognition systems that utilize machine learning to detect and classify pain from multiple data modalities. The authors systematically examine studies involving facial expressions, physiological signals, audio cues, and pupil dynamics as input sources for models intended to provide objective pain assessment, especially in populations with limited verbal communication ability. The methodology involves a literature analysis of peer-reviewed publications and empirical results, with attention to performance metrics and the strengths and limitations of each modality. Results indicate that facial expression and physiological signal-based models demonstrate promising classification accuracy, but challenges persist related to individual variability, environmental conditions, dataset heterogeneity, and model generalizability. The review highlights the need for robust evaluation frameworks, standardized datasets, and multimodal integration to improve reliability and clinical applicability. The authors also discuss potential clinical benefits of automated approaches for continuous monitoring and complementary use alongside traditional subjective assessments. The review concludes with recommendations for future research, emphasizing enhanced model robustness and validation across diverse populations and real-world contexts. The review situates automated pain recognition within interdisciplinary research integrating medicine, psychology, and computer science.

This editorial examines the methodological strengths of electroencephalography (EEG) and event-related potentials (ERPs) in affective and cognitive neuroscience, with emphasis on their clinical and translational relevance. The author highlights the millisecond-level temporal resolution of EEG as a primary advantage for investigating rapid perceptual, attentional, and emotional processes that are not accessible through slower hemodynamic imaging methods. The discussion reviews commonly studied ERP components, including the P300, N170, and N400, as objective indices of attentional allocation, face processing, and semantic integration. Clinical applications are considered across affective disorders, trauma-related conditions, and cognitive dysfunction, illustrating how electrophysiological markers may serve as sensitive indicators of treatment response and functional change. Methodological considerations are addressed, including challenges related to spatial resolution, signal-to-noise ratio, preprocessing variability, and reproducibility. The editorial also underscores the importance of standardized pipelines, transparent reporting, and integration with computational and machine learning approaches to enhance reliability and interpretability. Overall, the article positions ERPs as scalable, non-invasive tools capable of contributing to objective outcome measurement and precision-oriented clinical neuroscience.

This preprint investigates the classification of acute pain intensity states using noninvasive electroencephalography (EEG) and supervised machine learning methods. Data were collected from 52 healthy adults exposed to laser-evoked nociceptive stimuli. Continuous EEG recordings were segmented into four-second epochs and transformed into high-dimensional feature vectors comprising spectral band power, band ratios, Hjorth parameters, entropy measures, coherence indices, wavelet energies, and peak-frequency metrics. Nine conventional machine learning algorithms were evaluated using leave-one-participant-out cross-validation to estimate out-of-sample generalizability. A radial basis function support vector machine demonstrated the highest classification performance, achieving balanced sensitivity and specificity with overall accuracy approaching 89%, while maintaining minimal inference latency compatible with real-time deployment. Feature-importance analyses identified physiologically interpretable markers, including contralateral alpha suppression, midline theta enhancement, and frontal gamma activity. A prototype streaming pipeline further demonstrated feasibility using a reduced-channel EEG headset. Limitations include binary pain categorization, modest sample size, and reliance on single-modality data. The findings support the technical feasibility of EEG-based decision support tools for objective acute pain monitoring in clinical environments.

This systematic review evaluates electroencephalography (EEG) approaches for measuring cognitive workload using event-related potentials (ERPs) within single-task paradigms. The authors applied a structured literature search to identify empirical studies involving auditory probe stimuli combined with ERP measures to assess cognitive workload in healthy adult participants. Nineteen studies meeting inclusion criteria were analyzed, highlighting changes in ERP component amplitudes associated with increasing task difficulty. The review underscores that ERP amplitude changes with increasing cognitive load are dependent on task and stimulus features, emphasizing that the selection of primary task and stimuli shapes the sensitivity of ERP measures. It synthesizes conceptual insights regarding how specific ERP components reflect workload-related neural processes and discusses methodological considerations important for designing workload assessment experiments. Key contributions include clarifying the conditions under which ERP measures can reliably index cognitive workload and outlining practical implications for future research, such as optimizing stimulus design, task difficulty levels, and component selection to enhance interpretability. This review provides a foundational overview for researchers interested in using ERPs as neurophysiological markers of cognitive workload.

This article examines the integration of electroencephalographic (EEG) metrics with cognitive and affective theoretical models to inform biomimetic and neuroadaptive system design. The authors review spectral power indices, frontal alpha asymmetry, event-related potentials, connectivity measures, and time–frequency features as candidate neural markers for attention, working memory, emotional valence, and arousal. Emphasis is placed on construct validity, interpretability, and alignment between neural indices and computational representations of human cognitive architecture. The paper discusses methodological constraints including signal nonstationarity, inter-individual variability, preprocessing dependencies, and risks of reverse inference when mapping neural signals to psychological constructs. Design implications for adaptive human–machine interfaces and bio-inspired artificial systems are considered, with attention to ecological validity and real-time feasibility. Limitations include heterogeneity in task paradigms and lack of standardized validation protocols. The work contributes a translational perspective linking neuroscientific measurement to biomimetic engineering applications.

This review synthesizes commonly used electroencephalography (EEG) analysis methods and outlines applied principles for their selection in neuroscience and clinical neurology. EEG signals are categorized into time-invariant, accurate event-related, and random event-related types, with analytical strategies aligned to signal characteristics. Five principal methodological domains are described: power spectrum analysis, time–frequency analysis, connectivity analysis, source localization, and machine learning approaches. The review compares sub-methods including FFT, Welch, autoregressive modeling, wavelet transforms, coherence metrics, and spatial filtering techniques such as common spatial pattern. Strengths and limitations are discussed with attention to noise sensitivity, frequency resolution, model assumptions, and suitability for short versus long segments. The authors emphasize matching analytic technique to experimental design and signal properties to reduce bias and misinterpretation. The article provides a structured decision framework intended to improve methodological rigor, reproducibility, and translational validity in cognitive and clinical EEG research.

This paper presents a computer-vision method for detecting and localizing abnormal crowd behavior in surveillance video using a social force modeling framework. Instead of tracking individuals directly, the approach overlays a grid of particles on the image and advects them using a spatiotemporal average of optical flow, treating particle motion as a proxy for pedestrian dynamics in dense crowds. Interaction forces are estimated via a social force model and mapped back into the image plane to yield a force-flow representation for each pixel over time. Normal crowd behavior is modeled from randomly sampled spatiotemporal volumes of force flow using a topic-modeling approach, enabling likelihood-based identification of deviations. The method is evaluated on public crowd datasets with scenarios involving sudden dispersal and escape-like movement patterns, and it reports improved abnormality detection relative to baselines based primarily on optical-flow magnitude features. The work is frequently cited as an early example of combining motion-field estimation with interaction-based modeling to support automated monitoring of collective behavior in real-world, high-density scenes.

This longitudinal cohort study examined the extent to which passively sensed smartphone features predict depression and generalized anxiety symptom severity across varying temporal intervals. A community sample of 1,013 Android users installed a sensing application that continuously collected behavioral indicators, including GPS-derived mobility metrics, device and app usage, and communication patterns over a 16-week period. Repeated assessments of depression and anxiety were obtained using the PHQ-8 and GAD-7. Hierarchical linear regression models were used to evaluate associations between two-week aggregated sensing windows and symptom severity at distal, medial, and proximal lags. Results demonstrated that increased time spent at home relative to individual baseline levels prospectively predicted subsequent depression severity across multiple intervals, suggesting potential early warning utility. In contrast, certain circadian rhythm features showed stronger concurrent than prospective associations. Communication-related features exhibited differential relationships for depression versus anxiety, indicating symptom-specific behavioral signatures. Findings underscore the importance of temporal framing and person-centered deviations in digital phenotyping models and contribute to the methodological refinement of passive sensing approaches for mental health monitoring.

This study describes methods and initial results from a large-scale population assessment of objectively measured physical activity using wrist-worn accelerometers in the UK Biobank cohort. Over 100,000 participants aged 40–69 completed seven days of continuous wrist-worn accelerometer monitoring. Triaxial raw acceleration data were processed after calibration and noise removal to derive summary activity metrics. The analysis reported patterns of overall physical activity and variation by age, sex, time of day, day of the week, and season. Vector magnitude, a proxy for overall activity level, decreased with advancing age, and notable sex differences were observed across age bands. Differences between weekdays and weekend days and between seasons were generally small. The authors demonstrated the feasibility of collecting and processing large-scale objective physical activity data and established foundational activity metrics for use in future epidemiological analyses linking objectively measured behavior to health outcomes. The paper provides a methodological reference for researchers aiming to quantify physical activity objectively in large cohorts and highlights the potential for accelerometer measures to improve exposure assessment in population health research.

This systematic literature review examines digital phenotyping as a data analytics-driven method for assessing mental health using ubiquitous data from smartphones and wearable devices. The review included 5,422 articles published up to September 2024 and identified 74 primary studies that met inclusion criteria, reflecting a broad spectrum of research efforts in data collection, feature extraction, and analytical techniques. The authors categorize the types of data sources employed, including passive sensor data such as movement, phone usage, and physiological signals, and active data requiring user interaction. Methodological trends, such as the predominance of traditional machine learning approaches for behavior classification and prediction, are discussed, along with challenges related to dataset heterogeneity, reliance on self-report measures, and generalizability of models. The review also maps application domains and mental health conditions targeted in the literature, highlighting areas of promising progress as well as persistent obstacles that must be addressed to enable clinical translation. The authors detail implications for future research, underscoring the need for standardized study designs, robust evaluation frameworks, and techniques that integrate multimodal data to enhance predictive validity. The article situates digital phenotyping within the broader context of real-world behavioral monitoring and mental health research, emphasizing its potential to complement traditional assessments while noting ethical and practical considerations.

This article reports a large-scale experimental test of emotional contagion in an online social network context. The authors manipulated exposure to positive or negative emotional content in the News Feed of Facebook users (N ≈ 689,003) to assess whether observed emotional expressions influenced subsequent emotional expressions by users themselves. Participants were randomly assigned to conditions in which the prevalence of emotional content was reduced for a one-week period for either positive or negative posts. Emotional expression in participants’ own status updates was quantified as the proportion of positive or negative words used during the experimental interval. Results indicated that reducing exposure to positive emotional content led to fewer positive expressions and more negative expressions, while reducing exposure to negative content produced the opposite pattern. These findings provided evidence that emotions expressed by others through a social network platform can influence user behavior even in the absence of direct interaction or non-verbal cues. The study provides an empirical demonstration of large-scale emotional contagion mediated by algorithmically filtered digital communication channels but has generated subsequent discussion regarding ethical considerations and consent procedures in big data research.

This field-based study examined whether groups exhibit a measurable collective intelligence factor analogous to individual general intelligence. Teams completed a diverse battery of tasks requiring problem solving, coordination, and planning. Interaction dynamics were quantified using sociometric sensing technologies that captured conversational turn-taking, participation equality, and nonverbal engagement patterns over time. Statistical modeling demonstrated that group performance across tasks was explained by a single latent factor, distinct from average or maximum individual intelligence. Collective intelligence was predicted by social sensitivity, equality of conversational distribution, and the proportion of female group members. The findings indicate that group-level behavioral dynamics, rather than individual cognitive ability alone, shape collective performance. The study provides an empirical framework for quantifying interaction patterns in naturalistic team environments and modeling emergent group-level cognitive properties using automated behavioral measurement.

This study analyzed pedestrian movement in real public environments using automated video tracking to quantify emergent crowd behavior. Motion trajectories were extracted from surveillance recordings to model density fluctuations, lane formation, and oscillatory movement patterns at bottlenecks. Dynamical systems modeling demonstrated how self-organization arises from local interactions among individuals without centralized coordination. The analysis linked motion instability and density surges to safety risks, providing measurable indicators of hazardous crowd conditions. By applying computational tracking to large groups over time, the work established a quantitative foundation for understanding collective human motion in naturalistic settings. The findings illustrate how blob-based trajectory extraction and temporal modeling can reveal macro-level behavioral structure from micro-level movement data.

This large-scale field study quantified dynamic face-to-face interaction networks using wearable proximity sensors deployed in schools and conferences. Continuous data collection enabled high-resolution mapping of social contact structure across entire populations over multiple days. Temporal network analysis revealed heterogeneous contact durations, recurrent interaction patterns, and community clustering effects. The study demonstrated how automated sensing in natural environments can generate longitudinal behavioral datasets suitable for modeling information diffusion and disease transmission. The methodological framework emphasizes reproducible network reconstruction and statistical characterization of group interaction structure in ecologically valid settings.

This longitudinal observational study examined associations among mental health symptoms, physical activity, and environmental contexts using passively collected smartphone GPS data. Participants were recruited from a university community and provided Android smartphones configured to record GPS coordinates continuously for up to four weeks, allowing fine-grained assessment of movement patterns, time spent at locations of interest, and changes in activity spaces. Self-reported assessments of depression and stress were collected at baseline and during the monitoring period. Analytical methods included calculation of daily mobility features such as location entropy, radius of gyration, and percent time at home, which were related to symptom measures using mixed-effects models adjusting for age and gender. Results demonstrated that reduced mobility and increased time spent near home were associated with higher levels of depressive symptoms, consistent with theories linking restricted activity space to poorer mental health outcomes. The findings highlight the utility of passive GPS tracking to quantify contextual behavioral markers relevant to mental health research and illustrate methodological considerations for integrating high-resolution locational data with subjective symptom measures in naturalistic settings.

This updated AACN consensus statement provides evidence-based recommendations for the assessment of effort, response bias, and malingering in clinical and forensic neuropsychology. The document synthesizes empirical findings on performance validity tests (PVTs) and symptom validity tests (SVTs), emphasizing routine, multivariate validity assessment across cognitive and self-report domains. Guidance is provided on test selection, interpretation thresholds, base rate considerations, and management of false-positive risk. The statement clarifies distinctions among non-credible performance, psychiatric symptom exaggeration, and genuine cognitive impairment. It recommends the use of multiple independent indicators and cautions against reliance on single cut scores. Methodological considerations include demographic factors, psychiatric comorbidity, and neurological conditions that may influence validity outcomes. Limitations reflect evolving psychometric evidence and the need for culturally appropriate norms. The consensus framework supports defensible medicolegal opinions grounded in transparent and empirically supported validity assessment practices.

In *Daubert v. Merrell Dow Pharmaceuticals, Inc.*, the U.S. Supreme Court established the standard for admitting expert scientific testimony in federal court under the Federal Rules of Evidence. The case arose from claims that prenatal exposure to the drug Bendectin caused birth defects, leading to contested expert testimony. The Court held that the Federal Rules of Evidence govern admissibility of scientific evidence and rejected the long-standing Frye ‘‘general acceptance’’ test. Under Rule 702, the trial judge acts as a gatekeeper, ensuring that expert testimony is both relevant and reliable before it is presented to the jury. The Court outlined several non-exclusive factors for evaluating reliability, including testability, peer review and publication, known or potential error rates, and the existence of standards controlling the technique’s operation. The decision shifted emphasis from consensus in the scientific community to principles of scientific validity and relevance to the facts at issue, influencing how courts assess expert evidence across diverse disciplines. This ruling significantly impacts forensic and scientific expert testimony by requiring methodological rigor and judicial scrutiny prior to admissibility in legal proceedings.

This article examines the problem of group-to-individual (G2i) inference in scientific expert testimony, focusing on the logical and methodological challenges involved in applying population-based research findings to specific individuals in legal contexts. The authors distinguish between nomothetic scientific evidence derived from group data and idiographic conclusions about a particular litigant. They analyze how courts evaluate probabilistic and statistical evidence under evidentiary standards, including relevance and reliability thresholds. The article emphasizes base rates, conditional probability, and the inferential limits of epidemiological and psychological data when used to support individual causation or mental state claims. Methodological concerns include ecological fallacy, overgeneralization, and misinterpretation of statistical association as proof of individual effect. The authors propose structured reasoning frameworks to clarify the inferential bridge between group findings and case-specific conclusions. The analysis has direct implications for forensic neuropsychology and other behavioral sciences where expert opinions depend on translating research evidence to individual determinations.

In Kumho Tire Co. v. Carmichael, the United States Supreme Court clarified the scope of the trial judge’s gatekeeping role under Federal Rule of Evidence 702. The Court held that the reliability principles articulated in Daubert v. Merrell Dow Pharmaceuticals apply not only to scientific testimony but also to technical and other specialized expert knowledge. The decision emphasized that trial courts possess broad discretion to evaluate reliability, including consideration of factors such as testability, peer review, error rates, and general acceptance, where appropriate. However, these factors are flexible rather than mandatory. Applying this framework, the Court concluded that the district court did not abuse its discretion in excluding engineering testimony regarding tire failure causation, given concerns about methodological rigor and application to the specific facts. The ruling reinforced judicial responsibility to ensure expert testimony reflects reliable principles and methods before admission in civil litigation.

This National Research Council report provides a comprehensive evaluation of the scientific foundations of forensic disciplines in the United States. It identifies systemic deficiencies in standardization, validation research, accreditation, and oversight across multiple forensic domains, excluding nuclear DNA analysis, which was recognized as scientifically robust. The report emphasizes the need for empirical validation studies, quantification of error rates, development of standardized protocols, and transparent reporting practices. It critiques the historical reliance on subjective pattern-comparison methods without sufficient peer-reviewed research or statistical underpinning. Structural recommendations include the establishment of an independent national entity to oversee forensic science research and laboratory accreditation. The report highlights issues related to cognitive bias, training variability, and inconsistent laboratory practices. It concludes that strengthening forensic science requires sustained federal investment, interdisciplinary collaboration, and rigorous application of scientific methodology to ensure reliability, validity, and courtroom integrity.

This article examines the prosecutor’s fallacy within a contemporary forensic science framework grounded in likelihood ratio (LR) reporting. The author analyzes the conceptual distinction between likelihood ratios, prior odds, and posterior probabilities using Bayes’ theorem, demonstrating how conflation of these quantities leads to inferential error in courtroom settings. The paper critiques legacy reporting practices that encouraged statements about posterior probability of guilt and instead supports the use of LRs as measures of evidentiary strength. A modified real-case example illustrates how misinterpretation of statistical evidence can arise when conditional probabilities are transposed. The discussion addresses communication challenges, including lay misunderstanding of probabilistic reasoning, and explores methodological issues related to estimation of relevant probabilities. Limitations include dependency on assumptions about background population frequencies and model specification. The article provides structured guidance for experts and attorneys to avoid probabilistic fallacies while maintaining scientifically defensible testimony.