Categorization studies, in which stimuli vary along a category continuum, are becoming increasingly popular in psychological science. These studies demonstrate the effect of category ambiguity on various behavioral and neural measures. In such studies, researchers manipulate objective category levels by varying the physical properties of the stimuli, and then use these levels as predictors of behavior-assuming they map directly onto participants' perceived locations along the category continuum. This approach might not be optimal, considering the variability in participants' category boundary locations (their point of subjective equality, or PSE). In this tutorial, we propose addressing this issue by estimating participants' individual points of subjective equality, adjusting category levels relative to these points, and conducting statistical analyses on the subjective category levels. Implementing this method significantly improves the statistical power of the analysis in both experimental and simulated data. Adjusting stimulus levels by the points of subjective equality is highly suited for social categorization studies, in which points of subjective equality vary significantly. On a broader scale, it can be applied to a variety of categorization, discrimination, and decision-making studies.

Accurately quantifying belief strength in heuristics-and-biases tasks is crucial yet methodologically challenging. In this paper, we introduce an automated method leveraging large language models (LLMs) to systematically measure and manipulate belief strength. We specifically tested this method in the widely used "lawyer-engineer" base-rate neglect task, in which stereotypical descriptions (e.g., someone enjoying mathematical puzzles) conflict with normative base-rate information (e.g., engineers represent a very small percentage of the sample). Using this approach, we created an open-access database containing over 100,000 unique items systematically varying in stereotype-driven belief strength. Validation studies demonstrate that our LLM-derived belief strength measure correlates strongly with human typicality ratings and robustly predicts human choices in a base-rate neglect task. Additionally, our method revealed substantial and previously unnoticed variability in stereotype-driven belief strength in popular base-rate items from existing research, underlining the need to control for this in future studies. We further highlight methodological improvements achievable by refining the LLM prompt, as well as ways to enhance cross-cultural validity. The database presented here serves as a powerful resource for researchers, facilitating rigorous, replicable, and theoretically precise experimental designs, as well as enabling advancements in cognitive and computational modeling of reasoning. To support its use, we provide the R package baserater, which allows researchers to access the database to apply or adapt the method to their own research.

Psycholinguistic research has become increasingly reliant on online experimentation, making it an attractive approach for studying speech production. However, concerns remain about data quality and participant engagement in online settings. In this preregistered study, we used two tasks-picture naming and reading aloud-to test whether the lexical frequency effect (low-frequency words having shorter speech onset times than high-frequency words) could be reliably detected in the online environment (run at home), both with and without experimenter supervision. Participants completed the same two tasks at home and in the lab. Half of the participants performed both tasks with supervision and the other half unsupervised. In the naming task, all conditions yielded consistent frequency effects (~27-41 ms), comparable to previous online and lab findings. In the reading aloud task, lexical frequency effect emerged in all conditions except for the home-supervised, where the effect was in the expected direction but nonsignificant (~12 ms). Notably, participants were overall faster at home than in the lab (~10 ms), and unsupervised settings yielded the largest effect sizes. This suggests that experimenter presence may inadvertently dampen subtle effects, possibly due to increased self-monitoring or reduced comfort. Such findings indicate the reliability of online platforms for speech production research in psycholinguistics and highlight the nuanced influence of supervision on speech outcomes.

Singapore English, shaped by unique historical and sociocultural factors, diverges significantly from American and British English varieties. Despite its centrality in Singapore’s communication landscape, comprehensive word frequency estimates for this dialect have been lacking, limiting psycholinguistic research that accurately reflects local language processing. Dialectal variation fundamentally influences cognitive mechanisms, and neglecting these differences can distort theoretical models of language processing. This study addresses this gap by constructing word frequency (WF) and contextual diversity (CD) estimates from the National Speech Corpus (NSC), a large-scale speech corpus of Singapore English. We analyzed three NSC subparts comprising over 10,000 transcripts and 30 + million words to generate frequency estimates for 50,000 words, including both standard and uniquely Singaporean terms. To validate these measures, we used lexical decision data from the Auditory English Lexicon Project, featuring Singaporean participants who responded to stimuli across different accent conditions. Linear and binomial mixed-effect models compared NSC-based estimates against SUBTLEX-US and SUBTLEX-UK in predicting reaction times and accuracy. NSC estimates produced the lowest Akaike information criterion (AIC) values across both British and Singaporean accents, indicating superior fit to local language processing patterns. These results demonstrate that WF and CD effects are shaped by specific language experiences and highlight the necessity of dialect-appropriate psycholinguistic tools. This database advances cognitive science by enabling more accurate modeling of dialect-sensitive language processing, representing one of the largest psycholinguistic resources available for a less-studied English variety, and underscores the importance of developing similar resources for other English varieties worldwide.Supplementary InformationThe online version contains supplementary material available at 10.3758/s13428-026-03012-1.

Meta-analyses often include multiple dependent effect sizes, yet current methods typically neglect the resulting within-study dependencies or fail to address model uncertainty and publication bias adequately. We extend robust Bayesian meta-analysis (RoBMA) to a multilevel framework, simultaneously handling within-study dependencies, model uncertainty, heterogeneity, moderators, and publication bias. Specifically, the three-level RoBMA integrates approximate Bayesian selection models with PET-PEESE adjustments within a hierarchical Bayesian setting. We illustrate the methodology through empirical examples and demonstrate its performance via simulations. The approach is implemented in the RoBMA R package and JASP.

The planning of sample size for research studies often focuses on obtaining a significant result given a specified level of power, significance, and an anticipated effect size. This planning requires prior knowledge of the study design and a statistical analysis to calculate the proposed sample size. However, there may not be one specific testable analysis from which to derive power (Silberzahn et al., Advances in Methods and Practices in Psychological Science, 1(3), 337356, 2018) or a hypothesis to test for the project (e.g., creation of a stimuli database). Modern power and sample size planning suggestions include accuracy in parameter estimation (AIPE, Kelley, Behavior Research Methods, 39(4), 755–766, 2007; Maxell et al., Annual Review of Psychology, 59, 537–563, 2008) and simulation of proposed analyses (Chalmers & Adkins, The Quantitative Methods for Psychology, 16(4), 248–280, 2020). These toolkits offer flexibility in traditional power analyses that focus on the if-this, then-that approach. However, both AIPE and simulation require either a specific parameter (e.g., mean, effect size, etc.) or a statistical test for planning sample size. In this tutorial, we explore how AIPE and simulation approaches can be combined to accommodate studies that may not have a specific hypothesis test or wish to account for the potential of a multiverse of analyses. Specifically, we focus on studies that use multiple items and suggest that sample sizes can be planned to measure those items adequately and precisely, regardless of the statistical test. This tutorial also provides multiple code vignettes and package functionality that researchers can adapt and apply to their own measures.

The present study examined individual differences in 24 measures of cognitive ability in a sample of young adults (N = 255). Each measure was completed twice, separated by a period of 2 weeks, to assess test–retest reliability and retesting (i.e., practice) effects. Latent variable modeling was used to assess the convergent and discriminant validity of the measures, as they were selected to measure seven different cognitive constructs (attention control, processing speed, working memory, primary memory, secondary memory, fluid intelligence, and spatial ability). The measures showed adequate to high intrasession and intersession reliability. Construct-level estimates were highly reliable, and the measurement structure was invariant across the two testing occasions. In several instances, correlations among latent variables warranted further testing to ensure adequate discriminability. Finally, latent state-trait modeling indicated that the majority of systematic variance in cognitive measures is due to latent traits, rather than state-specific or task-specific factors. We discuss the practical and theoretical implications of these findings.

Eye-tracking corpora have advanced our understanding of reading processes by providing large-scale datasets of naturalistic reading behavior. However, existing corpora have almost exclusively sampled from typically hearing readers of spoken languages. Here, we present the Signers’ Eye-movements in English Reading (SEER) Corpus, a dataset of eye-movement behaviors from 41 skilled deaf adult readers who are early signers of American Sign Language (ASL), as well as a comparative group of 101 typically hearing monolingual English readers. Participants read 200 English sentences presented one at a time. In addition to eye-tracking data, the corpus includes detailed participant information: a standardized measure of reading proficiency, spelling recognition, and nonverbal intelligence for all participants. Information for the deaf participants include ASL comprehension scores, age of ASL acquisition, and phonological awareness scores (for a subset of participants). We report comparative analyses of reading behaviors at both the word level and sentence level. We also examine group differences in the effects of word length, frequency, and surprisal on local measures. The results indicate stronger effects of length and surprisal, but equivalent frequency effects (on content words) for deaf compared to hearing readers. The SEER Corpus offers researchers the opportunity to test hypotheses about reading development and efficiency in bimodal bilinguals who are first language users of ASL and skilled readers of English, supporting broader investigations of visual language processing. The corpus is preregistered and publicly available (https://doi.org/10.17605/OSF.IO/7P4F2) to facilitate replication, cross-study comparisons, and exploration of preliminary hypotheses in this understudied population.

Ceiling or floor effects pose analytic challenges in behavioral and psychological research. In this study, we developed novel Tobit modeling approaches, estimated using maximum likelihood (ML) or Bayesian methods, to address these effects for widely used statistical analyses, including the dependent-sample t-test and moderated regression. Simulation studies were conducted to compare the performance of the proposed modeling approaches to the conventional approach where ceiling or floor data are treated as if true values. The conventional approach was found to yield biased estimates, inflated Type I error rates, and poor confidence interval coverage, even with as little as 10% ceiling data. In contrast, the proposed approaches with either ML or Bayesian estimation provided accurate estimates and inference results across most studied conditions (e.g., with 30% ceiling data). Real data examples further illustrated the impact of modeling choices. To facilitate implementations of the proposed Tobit modeling approaches, we provide simulated datasets along with R and Mplus scripts online. Implications of the findings and future research directions were discussed.Supplementary InformationThe online version contains supplementary material available at 10.3758/s13428-025-02904-y.

A major challenge in studying unconscious processing is to effectively suppress the critical stimulus while allowing maximal signal strength for adequate sensitivity to detect an effect, if it exists. A possible way to do this is to calibrate stimulus strength. While calibrating stimulus strength is common in psychophysics, current calibration methods are not designed to find the maximal intensity in which the stimulus can still be rendered unconscious (i.e., find the upper subliminal threshold for each participant). Here, we demonstrate how calibration can be utilized to estimate, for each observer, this targeted threshold. We present a novel calibration procedure: the Subliminal Threshold Estimation Procedure (STEP), specifically designed for estimating the upper subliminal threshold for each individual. Using simulations, we showed that STEP outperforms existing calibration methods, which yielded strikingly low accuracy. We then further validated STEP using three empirical experiments. Together, these results establish STEP as highly beneficial for the study of unconscious processing.Supplementary InformationThe online version contains supplementary material available at 10.3758/s13428-025-02872-3.