Juyoung Lee, Thad Starner, Kai Kunze et al. (6 authors) β’ Semantic Scholar
Finding: Meta-analysis of 522 CHI papers (2006β2024) reveals inconsistent NASA-TLX usage, stronger inter-subscale correlations than originally found, weakened discriminability, and proposes a shortened 3-subscale version (Mental Demand, Physical Demand, Performance).
Why it matters: NASA-TLX is a core measurement tool in HCI/UX research for assessing cognitive load in interface studies, and this standardization work directly affects how researchers design and report user studies involving workloadβrelevant to HCI teaching and AI-human interaction research.
Method: Systematic review of 522 CHI papers combined with meta-analysis of subscale data from 683 tasks, using correlation, regression, and factor analyses to assess psychometric structure.
Jargon: *NASA-TLX*: a 6-subscale self-report measure of subjective workload (mental demand, physical demand, temporal demand, performance, effort, frustration); *communality*: how much variance in a subscale is shared with other subscales in factor analysis.
Alexandre Bluet, Shen Tian, David Schultner et al. (4 authors) β’ PsyArXiv
Finding: Language transmission (written advice) enables cumulative intergenerational improvement in task performance, while behavioral observation alone does not β because language can convey hidden mental models that underlie successful behavior.
Why it matters: Directly relevant to learning sciences and knowledge transfer; has implications for how teams share tacit knowledge, how expertise is documented and transmitted, and why explanation matters in skill acquisition and training design.
Method: Ten-generation chain paradigm (N=2,000) comparing observation vs. written advice transmission on a hidden-structure decision task β a well-controlled experimental design for studying cumulative cultural learning.
Jargon: *Cumulative culture* β the ratcheting process by which knowledge builds across generations beyond what any individual could discover alone; *mental models* β internal representations of how a system works, not just what actions to take.
β οΈ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Finding: Argues for moving behavioral science research from controlled lab settings to naturalistic "ecosystem" phenotyping using smartphones, wearables, and digital traces, using habit formation as a central example.
Why it matters: Directly relevant to behavior change design, habit formation, and just-in-time interventions (nudges, persuasive technology), and offers a methodological framework for studying these phenomena in real-world contexts.
Method: Conceptual/theoretical framework paper (mislabeled as empirical) proposing multi-level phenotyping ecosystems integrating passive sensing, self-report, and physiological data.
Jargon: *Ecological momentary assessment (EMA)* β repeated real-time self-reports in daily life; *just-in-time intervention* β delivering behavioral nudges at contextually optimal moments; *phenotyping* β characterizing individuals by measurable behavioral/biological profiles.
β οΈ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
M. Rosanensi, I. N. Y. Sumadewa, Widia Febriana β’ Semantic Scholar
Finding: This paper critiques the visual interface design of an AR app for nutritional information using HCI principles (Nielsen's heuristics, typography, color contrast, visual hierarchy) rather than the technical AR implementation.
Why it matters: Touches on HCI/UX design principles and AR as an information medium, but is a narrow applied design review of a single health app with limited generalizability for teaching or research.
Method: Descriptive qualitative analysis using the MDLC Design phase as a framework to evaluate UI elements against Nielsen's 10 heuristics.
Jargon: *MDLC (Multimedia Development Life Cycle)* β a structured process for developing multimedia applications; *Vuforia* β an AR development platform; *Unity* β a game/AR engine used for building interactive applications.
Aline Moore Lorusso, Ghislaine Lex van den Boogerd, Melissa Jansen et al. (5 authors) β’ PsyArXiv
Finding: Long-term digital voice assistant (DVA) ownership slightly reshapes household social dynamics and trust patterns, with trust in DVAs decreasing over time and DVAs primarily trusted for knowledge tasks rather than social ones.
Why it matters: Tangentially relevant to AI's impact on human relationships and HCI, but focuses narrowly on smart speakers in domestic settings rather than work, collaboration, creativity, or behavior change.
Method: Multi-method design combining behavioral tasks (Information Dilemma Task), self-report surveys, and qualitative interviews across three experiments (n > 250).
β Includes replication | β Longitudinal design β οΈ Self-reported data only | No control group mentioned | Limited statistical detail in abstract
Finding: Decision-making factors (cognitive biases, escalation of commitment, individual differences) correlate with how school leaders implement instructional change in Philippine private higher education institutions.
Why it matters: Tangentially relevant to behavioral science (cognitive biases, decision-making) and learning sciences, but the application is narrowly administrative/institutional in a specific regional context.
Method: Descriptive-correlational design with t-tests and z-tests across 41 school leaders and 184 teachers.