Caldwell University

Identifying component skills necessary for the emergence of intraverbal tacts, or verbal responses under control of both a verbal and nonverbal antecedent stimulus, is important because the occasion for this skill often occurs in a child's everyday life. Previous research has begun to identify a sequence of component skills that may lead to the emergence of multiply controlled intraverbals. However, it remains unclear which component skills are necessary versus sufficient. The purpose of this study was to evaluate the effects of teaching a subset of component skills, element tact and intraverbal categorization, to identify the skills sufficient for emergence of intraverbal tacts. A multiple-probe design was used to assess intraverbal-tact emergence for five participants diagnosed with autism spectrum disorder during pre-and post-element-tact and intraverbal-categorization teaching sessions. Emergence of intraverbal tacts was also assessed during recombinative-generalization probes. Results indicated that intraverbal tacts emerged for all participants following acquisition of element tacts and intraverbal categorizations. As no other component skills were taught, these data suggest that these component skills may be sufficient for intraverbal tact emergence. Implications for identifying necessary component skills and directions for future research are discussed.

ABSTRACT There is increased demand for supervisors to develop staff training procedures that are efficient, such as those that promote generalized responding. Computer‐based instruction (CBI) is effective and efficient and can be enhanced by general‐case procedures. Therefore, experiment one employed a non‐concurrent multiple baseline design with adults to evaluate the generalized effects of general‐case analysis to inform CBI. We found this combination moderately effective in facilitating generalization of staff training. Experiment two used a non‐concurrent multiple baseline to evaluate general case analysis and modified general case programming to inform CBI to specifically increase recombinative generalization. We then identified the extent of recombinative generalization along a recombinative generalization gradient. Most participants demonstrated recombinative generalization and that fewer combinations of procedures were more likely to occur than all of them combined. Implications for staff trainers and researchers include pre‐training of skill acquisition trial arrangements and further evaluations of general‐case procedures.

Social validity in applied behavior analysis remains critically important. However, assessing social validity and responding to its results pose challenges for the field. Social validity assessments are often treated as a one-time measure conducted after intervention. Researchers have urged consideration of social validity as an ongoing, dynamic, and interactive process (i.e., interventions are adjusted based on social validity assessment results). Yet, there has been limited discussion on the interactive process of social validity and the variables that may influence it. This article (a) reviews definitions of social validity and the purpose of its assessment; (b) examines how studies published in the Journal of Applied Behavior Analysis from 2010 to 2020 have assessed social validity as an interactive process; and (c) offers considerations for researchers and practitioners to enhance opportunities for an interactive process of social validity, including issues related to the measurement rigor and data interpretability of social validity assessments.

Lung cancer is the leading cause of cancer-related mortality, with approximately 2.5 million new cases and 1.8 million deaths annually, making reliable diagnosis a clinical priority. Although deep learning models have achieved strong performance in lung cancer classification, evaluation has largely focused on predictive accuracy, leaving their decision-making processes insufficiently examined. This study compares three architecturally distinct models: a Convolutional Neural Network (CNN), a pretrained ResNet50, and a Vision Transformer (ViT), trained on the IQ-OTH/NCCD lung cancer CT dataset. Local Interpretable Model-Agnostic Explanations (LIME) were applied to investigate model reasoning. In addition to standard performance metrics, a dual-correlation framework was introduced to measure both prediction agreement and explanation agreement across model pairs. All three models achieved strong classification performance, with ResNet50 attaining 98.61% accuracy, CNN 97.91%, and ViT 93.75%, while all achieved ROC-AUC scores of 0.99. Prediction correlations exceeded 0.99 across all model pairs, indicating highly consistent outputs. However, LIME explanation correlations remained below 0.26, revealing substantial differences in the image regions used to reach those predictions. Analysis of misclassified samples further identified a consistent spatial pattern: incorrect predictions were associated with attention outside the lung parenchyma, whereas correct predictions focused primarily within lung regions. These findings demonstrate that prediction agreement is a poor proxy for reasoning consistency, and that interpretability evaluation must be treated as an independent validation criterion alongside predictive performance in clinical AI systems.

It is widely recognized that individuals with autism spectrum disorder (ASD) often require individualized educational programming. However, only a few studies have evaluated using an assessment to identify the most efficient and effective instructional components for individual learners. We replicated and extended Seaver and Bourret (2014) by conducting prompt-type and prompt-fading assessments with two adults with ASD to identify the most efficient prompt type (model, vocal-gestural, or full physical) and prompt-fading procedure (progressive prompt delay, most-to-least, or least-to-most). Following these assessments, we compared the most and least efficient procedures while evaluating whether the assessment results generalized to additional skill types. We also compared the assessment results to those of a decision-making tool, the Systematic Worksheet for the Evaluation of Effective Prompting Strategies (SWEEPS; Cowan et al., 2023). The most efficient prompt type was learner-specific, whereas both participants learned most efficiently using a progressive prompt delay. These outcomes were replicated across additional functional skills and showed clear differences between the most and least efficient procedures. The assessment results were also consistent with the SWEEPS.