In a single-case experimental design (SCED), it is recommended that interrupted time-series analysis (ITSA) be applied in the AB design. This paper demonstrates that sample size determination in ITSA can be conducted using the Fisher information matrix (FIM) to predict estimation precision and statistical power. The simulation results corroborate that the FIM accurately predicts the empirical statistical power. Because of its low computational load, researchers can check the statistical power of various parameters (effect sizes) and sample sizes using FIM. Consequently, FIM serves as a valuable tool for optimizing the experimental design of SCED. Furthermore, we developed a graphical user interface (GUI) application capable of calculating the predicted statistical power.
This paper aims to examine the effect of rating scale orientation and labels on the measurement of subjective social position. Literature on the effects of layout and labels on responses frequently reported the left-side selection bias. Among the theoretical hypotheses that may explain this bias, we focused on the pseudoneglect and the “Left and top means first” heuristic. To examine the prediction from each hypothesis, we conducted a web survey experiment (Study 1). The results, however, did not seem to support either of the two hypotheses. Therefore, we proposed another hypothesis that assumes the “Moving rightward means going up” heuristic and conducted a web survey experiment (Study 2). The results partially supported our hypothesis. Based on the results of our two experiments, we recommended that the measurement of subjective social position should use a vertically oriented rating scale with the verbal anchors “Upper—Lower” and the descending numerical labels.
Count data frequently arise in applied research but are often challenging to model due to overdispersion, underdispersion, excess zeros, and unobserved heterogeneity. This study analyzes the Fisman_Miguel dataset using seven models: Poisson, negative binomial (NB), zero-inflated Poisson (ZIP), zero-inflated negative binomial (ZINB), Poisson hurdle (PH), negative binomial hurdle (NBH), and Conway–Maxwell Poisson (COM-Poisson) to model diplomatic misconduct. Parameters are estimated via maximum likelihood estimation, while model fit is evaluated using information criteria (AIC and BIC) and log-likelihood values. The results indicate that models accounting for overdispersion and excess zeros, particularly the ZINB and NBH, provide a superior fit compared to standard models. Moreover, the inclusion of random intercepts reveals substantial cross-country heterogeneity, highlighting the hierarchical structure of the data. Overall, the findings emphasize the importance of adopting flexible count data models in policy-oriented analyses, where reliance on standard Poisson assumptions may lead to misleading conclusions.
This study investigated the effects of topic variation on the accuracy of authorship analysis using Japanese stylometric features including n-grams of characters, parts-of-speech (POS), postpositional particles, and function words. We classified 4,574 blog texts from 2,287 authors into six topic groups and further divided them into similar and dissimilar topic groups using Latent Dirichlet Allocation (LDA) and hierarchical cluster analysis (HCA). Bayesian modeling and effect size analysis revealed that n-grams of characters were highly susceptible to topic variations, with increasing n exacerbating topic effects. In contrast, postpositional particles remained robust across all n-gram levels, whereas n-grams of POS and function words exhibited smaller topic effects than n-grams of characters despite significant topic effects at higher n-gram levels. These findings provide useful insights into authorship analyses, underscoring the importance of non-content-based stylometric features.
As the world transitions to hybrid work, the Indian IT sector faces a dilemma involving digital innovation and emotional complexity. As organizations increasingly adopt hybrid work arrangements, emotional understanding and resilience have become important considerations in managing digitally mediated work. Grounded in the JD–R model and Upper Echelons Theory, this study adopts a two-stage analytical framework to examine and prioritize factors among IT professionals working under hybrid work arrangements, conceptualizing EI as a role-contingent and hierarchically differentiated resource. In the first stage, structural path analysis is applied on data collected from 592 Indian IT professionals to examine the influence of six psychological factors on four EI components across managerial levels. In the second stage, Spherical Fuzzy Analytic Hierarchy Process (SF-AHP) is applied to expert judgments obtained from 40 senior industry professionals and academicians to prioritize these EI factors across levels under uncertainty. The results reveal differentiated dominance patterns, with empathy emerging as most salient at middle and lower levels, while adaptability consistently ranks as dominant at lower and senior levels. Self-expression assumes a bridging role, gaining prominence in roles involving cross-level coordination. By integrating statistical validation with expert-driven prioritization, the study provides a practical framework for designing hierarchy-specific emotional intelligence development programs across hierarchical roles in hybrid work environments.
We study the arithmetic-harmonic inequality (AHI) index, a bounded and scale-invariant measure of dispersion for positive random variables. Explicit expressions are derived within the generalized inverse Gaussian (GIG) family, including the inverse Gaussian and gamma distributions as special cases. We investigate the associated estimator, establish its asymptotic properties, and derive first-order bias approximations. A Monte Carlo study evaluates its finite-sample performance, and an application to GDP per capita data from countries in the Americas illustrates the behavior of the AHI index within the Atkinson family. The results support the AHI index as a tractable and interpretable measure of economic dispersion.
In Malaysia, the rapid proliferation of shopping malls has heightened consumer expectations, driving malls to diversify their scale, formats, and characteristics. This study employs block clustering to simultaneously classify consumers (rows) alongside area brand asset factors in Subang Jaya and store image factors at Sunway Pyramid (columns). Compared to traditional methods like K-means, this machine learning technique more efficiently processes complex data interrelationships. The analysis revealed minimal overlap between area-specific and mall-specific factors within the expectation clusters, suggesting a relatively weak correlation between the two. Furthermore, the identified consumer clusters reflect distinct trends in age, transportation, number of shopping companions, and the prioritization of the area versus the mall. These findings offer strategic insights into leveraging area brand assets and store image factors to enhance regional appeal and shopping mall satisfaction.
This article discusses an exact inference approach in longitudinal experimental designs based on a discrete conditional distribution. It is derived from a mixed effects model for binary data (with logit link function) that is a generalization of the Rasch model. Exact hypothesis tests are suggested. Their application in scenarios of high practical relevance in empirical research is discussed. In a particular case, very common in clinical research, it is shown that uniformly most powerful and uniformly most powerful unbiased tests are applicable. For computational reasons, the exact distributions are approximated by Monte Carlo techniques. All suggested procedures are illustrated in a hypothetical example referring to a clinical research setting and a real-world data example from developmental psychology. Both examples show the practicability of the use of sampling algorithms to approximate the exact distributions of the test statistics. The computational burden is practically negligible.
Once the symmetry model is rejected by statistical tests, we aim to compare several contingency tables with respect to their deviations from symmetry, regardless of the sample size. In this study, we propose a measure to evaluate the degree of departure from symmetry using cosine similarity for square contingency tables that are represented by the same categories. This proposal is based on the Fisher-Rao distance, which allows asymmetry to be interpreted as the geodesic distance between two distributions. Unlike the divergence-based measures proposed in previous studies, the proposed measure provides a geometrically simple way to represent departures from symmetry, allowing intuitive interpretation through visualization on a two-dimensional plane. A simulation study demonstrates that, for square contingency tables under an asymmetry model, the proposed measure complements symmetry tests by providing an interpretable quantification of the degree of asymmetry after rejection, facilitating comparisons across tables.
The authors proposed a model to accurately estimate the rate of unplanned purchasing of purchased items for each product category. The model is a hierarchical item response theory model that considers consumer heterogeneity and product category hierarchy. Simulations were run using the proposed model for a small sample of in-store survey data that improved the estimation accuracy compared to the conventional method. Analysis was also conducted with the real data, and the rate of unplanned purchasing for each product category was estimated. By reducing the estimation error using the proposed model, retailers and other practitioners can reduce the possibility of errors when creating a sales floor and conducting sales promotions.
This study employs hierarchical Bayesian binomial logit models to examine the store image indicators that influence customer satisfaction at competing shopping malls in Malaysia, and whether area attractiveness and brand asset indicators affect shopping mall satisfaction. We construct proposed shopping mall satisfaction and area attractiveness ranking models for quantification with store image indicators and area attractiveness as explanatory variables, and the area attractiveness ranking model uses area brand asset indicators. The areas under assessment are Subang Jaya and Petaling Jaya. The shopping malls examined in this study are Sunway Pyramid, located in Subang Jaya, and 1Utama shopping center, located in Petaling Jaya. As the two areas are adjacent, the malls share a competitive relationship. We also estimate the two models simultaneously, providing measures to enhance shopping mall satisfaction and area attractiveness based on the estimation results of the models.
Knowledge Tracing (KT) estimates a learner’s knowledge state within a given domain based on their response data. Although approaches such as Bayesian modeling and neural-based models have achieved remarkable estimation accuracy, conventional models update the learner’s knowledge state only at the time response data are obtained, thereby preventing the consideration of real-time state changes between responses. In this study, we introduce the concept of continuous time and propose a knowledge tracing method grounded in mechanical dynamics. Experimental results demonstrate that incorporating “indeterminate” evaluations–cases where it cannot be determined whether constraints are considered–enhances the overall validity of state estimation. Despite accounting for such ambiguous states, the proposed model achieves accuracy comparable to Bayesian knowledge tracing methods in predicting learners’ correct response rates.
Herein, we propose an automated essay scoring (AES) model that utilizes the argumentative graph (AG) information of an essay. An AG is a logical structure that influences the holistic score and key aspects of organization in scoring. Previous AES models have considered the statistics of logical elements of the graph and the relationships among these elements. In this study, we employ entire sentence containing logical elements to estimate the relationships among the sentences. Then, the proposed method extracts the relevant features using a graph attention network; this network considers the graph structure and is combined with a conventional AES model. We conducted experiments on the Automated Student Assessment Prize (ASAP) and demonstrated that the proposed model exhibits improved accuracy compared to similar AG-based baseline methods. Furthermore, we validated AG and demonstrated that the proposed method achieves high accuracy when AG is random or no logical structure exists in an essay.
Kernel methods have long played a central role in machine learning by enabling nonlinear modeling in high dimensional feature spaces. Among them, the Gaussian Radial Basis Function (RBF) kernel is widely used due to its smoothness, universality, and excellent performance in supervised learning. In contrast, the Jaccard/Tanimoto coefficient, originally defined over sets and binary vectors, has achieved notable success in chemoinformatics and information retrieval. Recent studies have shown that the Jaccard/Tanimoto coefficient can be extended to real-valued vectors and interpreted as a valid kernel. In this paper, we propose a natural generalization of the Tanimoto kernel by introducing a bandwidth-like parameter, leading to the Bandwidth-Adjustable Tanimoto (BAT) kernel. We focus on the shared structure of the BAT and the Gaussian RBF kernels in terms of their infinite polynomial expansions and investigate the universality of the BAT kernels, clarifying their relationship to the universal Taylor kernels. In particular, we show that by incorporating a bias term, the BAT kernel attains the universal approximation property comparable to that of the Gaussian kernel. Through experiments on classification, regression, and clustering tasks using benchmark datasets, we demonstrate that the BAT kernel performs comparably to the Gaussian kernel, while even the original Tanimoto kernel achieves competitive results. In particular, the BAT kernel attains performance on par with the Gaussian kernel in classification tasks, and it slightly surpasses the Gaussian kernel in regression and clustering tasks. A comparison of their functional forms, together with experiments on both regression and classification tasks involving nonlinear structures, further shows that the BAT kernel yields smoother predictive surfaces than the RBF kernel. Overall, our results indicate that the BAT kernel offers universality and practical performance, making it an alternative to the Gaussian kernel.
This study introduces a data-driven cut score method that integrates Item Response Theory (IRT), cluster analysis, and Gaussian-Otsu thresholding to objectively determine cut scores. The proposed approach, abbreviated as DDCS, provides a fully repeatable and non-judgmental framework that eliminates the need for expert panels used in traditional methods such as Angoff, while maintaining the conceptual goal of identifying a borderline competency threshold. In this study, the performance of DDCS was compared with the traditional Angoff approach, both using simulated data generated under different sample sizes, test lengths, and distributions of ability, difficulty, and discrimination parameters, and with realistic data. The two methods were evaluated using metrics such as accuracy, sensitivity, specificity, bias, absolute bias, and RMSE. Simulation results showed that the DDCS method outperformed the traditional Angoff approach in 88.9
In daily life, humans selectively search for information about options to make decisions. The metalevel MDP framework, proposed to understand this information search process, has so far been evaluated only for its predictive performance regarding group differences in summary measures under unrealistic scenarios. This study aimed to examine whether the metalevel MDP can account for the sequential patterns of information search in more realistic decision-making situations. We conducted a chatbot-based experiment that enables diverse information search actions. We then cast the existing greedy metalevel MDP (GML-MDP) into a probabilistic form to estimate participant-level latent parameters and assess its fit to sequential action data. The experimental results showed that the model partially explains the information search process, and we discussed potential directions for model improvement.
This study proposes an item response theory (IRT) model for bounded continuous data: the censored normal response model (CNRM). This model has a structure similar to the Tobit models, which makes it possible to take into account the ceiling and floor effects on item responses. The CNRM is formulated as a special case of the generalized normal ogive framework, which unifies several existing models. A parameter estimation method using the EM algorithm is shown and is applied to simulated and real data. The results suggest that the CNRM provides a computationally efficient and highly interpretable alternative to Molenaar et al.’s (2022) zero-and-one inflated approach.
