Eye tracking has emerged as a powerful tool for advancing autism research, including diagnostics and interventions. However, studies on joint attention (JA) in autistic children have predominantly concentrated on the responding to joint attention (RJA) construct, with limited focus on the initiating joint attention (IJA) construct. Moreover, despite the interactive nature of JA, researchers have often relied on passive paradigms to study JA in this population. To address these gaps, we developed a gaze-contingent eye-tracking battery targeting three developmentally appropriate JA skills for young children: RJA, IJA to request, and IJA to comment or reference. The development process was multifaceted and iterative, involving a series of collaborative steps and the allocation of various resources. These steps included determining the motion format and stimulus type, designing and prototyping the stimuli, recruiting an actor to serve as a communication partner in the stimuli, recording and editing videos for the stimuli, and building and test-running the battery. We developed the Interactive Eye Tracking for Joint Attention (IET-JA) battery, which consists of 32 JA stimuli: 16 RJA, 8 IJA-Request, and 8 IJA-Comment/Reference. The stimuli are dynamic (i.e., videos) and feature a preprogrammed interactive human communication partner who is responsive to the participant's gaze. The IET-JA takes approximately 8 minutes to complete, and its duration is expected to vary based on the participant's level of engagement. Implications for advancing methodologies, fostering team science, and enhancing iterative processes are discussed.

When data have a hierarchical structure, such as students nested within classrooms, ignoring dependencies between observations can compromise the validity of imputation procedures. Standard (tree-based) imputation methods implicitly assume independence between observations, limiting their applicability in multilevel data settings. Although multivariate imputation by chained equations (MICE) is widely used for hierarchical data, it has limitations, including sensitivity to model specification and computational complexity. Alternative tree-based approaches have shown promise for individual-level data, but remain largely unexplored for hierarchical contexts. In this simulation study, we systematically evaluate the performance of novel tree-based methods-chained random forests (missRanger) and extreme gradient boosting (mixgb)-explicitly adapted for multilevel data by incorporating dummy variables indicating cluster membership. We compare these tree-based methods and their adapted versions with traditional MICE imputation in terms of coefficient estimation bias, type I error rates, and statistical power under different cluster sizes (25 and 50), missingness mechanisms (missing completely at random [MCAR], missing at random [MAR]), and missingness rates (10%, 30%, 50%), using both random intercept and random slope data generation models. The results show that MICE provides robust and accurate inference for level 2 variables, especially at low missingness rates (10%). However, the adapted boosting approach (mixgb with cluster dummies) consistently outperforms other methods for level 1 variables at higher missingness rates (30%, 50%). For level 2 variables, while MICE retains better power at moderate missingness (30%), adapted boosting becomes superior at high missingness (50%), regardless of the missingness mechanism or cluster size. These findings highlight the potential of appropriately adapted tree-based imputation methods as effective alternatives to conventional MICE in multilevel data analyses.

In the structural equation modeling framework, binary variable models are generally considered a special case of ordinal variable models, as both involve similar scale assignment processes. However, the scaling processes of the two model types differ, with these differences becoming increasingly pronounced in the context of latent growth models (LGMs). To define scale units, the two types of LGMs-specifically, one with ordinal variables and the other with binary variables-depend on different observed scale references, such as thresholds and standard deviations, which are derived from observed categorical variables. Applying distinct observed scale references to binary and ordinal LGMs results in systematic differences in the scale units of their corresponding latent response variables. Consequently, in binary LGMs, the transformed latent response variables used for model estimation may fail to accurately reflect the corresponding population information, and as a result, their parameter estimates are more likely to be systematically biased than those obtained from ordinal LGMs. This study investigates the impact of these differences on estimating ordinal and binary LGMs and underscores potential estimation concerns in binary LGMs from both theoretical and empirical perspectives.

Perceptual threshold estimation stands as a fundamental variable manipulation in psychology and neuroscience, fine-tuning sensory input intensity to align perceptual precision across individuals. Current methods, such as constant stimuli and staircase procedures, inaccurately assume that perceptual thresholds are influenced solely by sensitivity, overlooking the role of response biases and consequently leading to unreliable estimates. The underlying reason is that classical methods traditionally focus on hit rates-the correct detection of stimuli-while neglecting false alarms, where participants incorrectly report a stimulus when none is present. According to signal detection theory, hit rates can arise from genuine sensitivity or a liberal response criterion, making it crucial to account for both measures. To overcome this confound, we developed bias-free versions of these methods by including target-absent trials to factor in decisional bias during threshold evaluation. Across 74 participants, we robustly demonstrated that widely used classical procedures overestimate perceptual thresholds due to uncontrolled interindividual variability in participant criterion. Conversely, the bias-free staircase procedure achieved reliable sensitivity thresholds, effectively mitigating the influence of decisional bias. We strongly advocate for the adoption of this straightforward, bias-free approach. Its effectiveness, simplicity, and quick execution make it highly feasible for widespread use in enhancing the reliability of threshold estimation and reducing variability in future scientific investigations across various fields.

This study presents an audio stimulus corpus consisting of 444 Western popular music drum and bass patterns called the Lucerne Groove Library. Common requirements for stimuli in music psychological research, particularly studies on the experience of groove, are outlined, followed by a description of how these criteria are addressed in the presented corpus. For example, the corpus is designed to combine ecological validity with high manipulability, facilitating the use of the corpus in a variety of experimental settings. The methods section provides a detailed account of material selection, audio creation, and measures. Ground-truth behavioural data for the stimuli were obtained through a listening experiment (e.g., participant'sratings on the urge to move in response to the stimuli, and style assignments ). Several structural data (e.g., tempo and event density) and audio features (e.g., low-frequency sub-band flux) were measured for each stimulus, and an overview over all data is provided. Potential applications of the corpus in future studies are discussed, including variables along which stimuli can be selected or manipulated. The corpus is published in different formats and with a range of accompanying data, all of which can be found in the online repository.

Cognitive diagnostic assessment (CDA) enables direct exploration of participants' cognitive structures or psychological latent traits (referred to as attributes), offering unique advantages within psychological methodologies. The Q-matrix, which delineates the relationship between items and attributes in CDA, is crucial for accurate diagnosis. However, ensuring the accuracy of the Q-matrix in practical applications is often challenging. Constructing a Q-matrix typically requires extensive calibration efforts from both test developers and domain experts, and even then, issues of accuracy and subjectivity remain. Although various Q-matrix validation methods have been developed to improve its quality, their implementation often presents a steep technical barrier for typical psychological researchers. These challenges have limited the broader application of CDA in psychological research. This paper provides a systematic review of Q-matrix validation methods under saturated cognitive diagnosis models (CDMs) and introduces Qval, a user-friendly and powerful R package that offers a one-stop solution for implementing a wide range of state-of-the-art validation procedures, including parameter estimation, validation methods, iterative procedures, and search algorithms. The Qval package leverages C++ code and parallel computing to improve computational efficiency. Additionally, this paper provides detailed guidance on how to implement Q-matrix validation procedures effectively.

Translation ambiguity occurs when a word in one language has multiple translations into another language, resulting in slower and less accurate language processing and production. We collected number-of-translation norms for 562 English words into Japanese, of which 70% were translation-ambiguous. When compared with previous norming studies with the same English words, our sample had the highest proportion of ambiguity. We then had a separate group of participants back-translate the correct Japanese translations into English and found that 54% were translation-ambiguous. We further analyzed types of errors in our sample, finding omission errors (i.e., no translation given) to be the most common in both translation directions. Errors made during translation were related to proficiency, wherein higher L2 proficiency was significantly positively correlated with making more semantic-type errors (i.e., providing a semantic associate instead of the correct translation), specifically in the Japanese to English direction. These norms provide information on how translation-ambiguous and translation-unambiguous words are produced in both directions of translation and can be used in future research examining translation ambiguity.

The expectation-maximization (EM) algorithm is widely used for parameter estimation in item response theory (IRT) modeling. However, when applied to datasets with large numbers of individuals and items, the standard EM algorithm can be slow to converge, with computationally expensive E-steps. We propose a modified EM algorithm to accelerate estimation for unidimensional two-parameter logistic IRT models. The modified algorithm uses a two-stage structure with partial-step updating over data subsets to reduce convergence time, while maintaining comparable accuracy and precision. The first two simulation studies evaluated its performance relative to standard EM, focusing on convergence time, parameter recovery, and standard error estimation across varying subset sizes and item counts. The third study demonstrated its scalability and runtime advantage in a large-scale testing scenario involving one million respondents and 100 items. The fourth study evaluated robustness under departures from unidimensionality. The proposed algorithm showed time reductions under smaller subsets with 36 items, consistent reductions across all subset sizes with 54 items, and the largest reduction (60%) with 40-item forms constructed from 100 items, while maintaining comparable estimation performance. These results highlight the algorithm's potential for large-scale applications involving tens of thousands of respondents and moderate-to-large item pools, with modifications that can be integrated into existing EM routines. We conclude with directions for operational use and potential extensions to multidimensional IRT estimation.

The FAIR (Findable, Accessible, Interoperable, and Reusable) data principles form the foundation of the open data movement. However, while many current practices ensure data are findable and accessible, true interoperability and reusability remain limited. This paper introduces the Truth Effect Database (TED), a large-scale, trial-level, open database harmonizing data from illusory truth effect studies designed to enhance interoperability and reusability. TED currently integrates data from 59 studies in 29 publications, spanning 12,249 participants and 808,231 trials, accounting for a wide range of dispositional and contextual variables. To promote usability, TED focuses on user-friendly data submission using a custom entry website and data extraction using the R package acdcquery. These tools guide researchers through both data entry and retrieval, eliminating the need for direct interaction with the database's internal structure. We illustrated the utility of TED through Bayesian multilevel analyses, highlighting substantial variance in the illusory truth effect at the subject level, moderated by the delay between exposure and judgment phases in truth effect paradigms. Beyond this first demonstration, TED provides the foundation for a wide range of future research. These include (living) meta-analyses, simulation-based power analyses, rigorous replication and reanalysis of existing studies, and the validation and development of formal cognitive models. As an open and extensible infrastructure, TED serves as a blueprint for sustainable, community-driven database development in psychological science.

Scale abbreviation is a crucial task for researchers aiming to reduce response burden and optimize data collection when using self-report instruments such as online surveys and questionnaires. Among various data-driven strategies available for scale abbreviation, supervised machine learning (SML) algorithms have emerged as a prominent approach due to their accuracy in predicting total scores from the original instrument. However, previous studies offer limited insights into how SML-abbreviated scales can be evaluated using both SML and psychometric metrics across different feature selection techniques. To address this gap, the current study aims to evaluate the effectiveness of seven feature selection methods: item-total-correlation-based filters (ITC), Minimum-Redundancy-Maximum-Relevance (MRMR), Lasso, Sequential Forward Selection (SFS), Sequential Backward Selection (SBS), Genetic Algorithms (GA), and Non-dominated Sorting Genetic Algorithms-II (NSGA-II), all used in conjunction with SML. Additionally, the psychometric properties of these SML methods are compared with two non-SML approaches. Using simulated datasets varying in sample size, model error, and factorial correlations, the study examines predictive accuracy, reliability, and the ability to recover both inter-subscale correlations and external criterion correlations. The findings indicate that no single method consistently excels across all conditions, with specific feature selection techniques performing better under certain circumstances. Key insights are provided to guide researchers in selecting appropriate feature selection methods based on their specific dataset characteristics and goals.