Randomized Experiment Research Articles

An Individual Participant Data Meta-Analysis to Support Power Analyses for Randomized Intervention Studies in Preschool: Cognitive and Socio-Emotional Learning Outcomes

There is a need for robust evidence about which educational interventions work in preschool to foster children’s cognitive and socio-emotional learning (SEL) outcomes. Lab-based individually randomized experiments can develop and refine such interventions, and field-based randomized experiments (e.g., cluster randomized trials) evaluate their effectiveness in real-world daycare center settings. Applying reliable estimates of design parameters in the context of a priori power analyses is essential to ensure that the sample size of these studies is adequate to support strong statistical conclusions regarding the strength of the intervention effect. However, there is little knowledge on relevant design parameters with preschool children. We therefore utilized a systematic collection of individual participant data from four German probability samples (554 ≤ N ≤ 2928) with preschool children (aged two to six years) to estimate and meta-analyze design parameters. These parameters are relevant for planning single-level (e.g., in non-clustered lab-based settings), two-level (children nested in daycare centers), and three-level (children nested in groups, with groups nested in daycare centers) randomized intervention studies targeting cognitive and SEL outcomes assessed with three methods (standardized tests, parent ratings, and educator ratings). The design parameters depict between-group and -center differences as well as the proportion of variance in the outcomes explained by different covariate sets (socio-demographic characteristics, baseline measures, and their combination) at the child, group, and center level. In conclusion, this paper provides a rich source of design parameters, recommendations, and illustrations to support a priori power analyses for randomized intervention studies in early childhood education research.

Commentary: Prioritising Randomized Trials; David P. Farrington's Legacy for Criminology in the 21st Century: As Criminological Research Tilts Towards Ideology, Randomized Experiments Are Key for Reinforcing True Science.

Commentary: Prioritising Randomized Trials; David P. Farrington's Legacy for Criminology in the 21st Century: As Criminological Research Tilts Towards Ideology, Randomized Experiments Are Key for Reinforcing True Science.

Causal inference with cross-temporal design.

When many participants in a randomized trial do not comply with their assigned intervention, the randomized encouragement design is a possible solution. In this design, the causal effects of the intervention can be estimated among participants who would have experienced the intervention if encouraged. For many policy interventions, encouragements cannot be randomized and investigators need to rely on observational data. To address this, we propose a cross-temporal design, which uses time to mimic a randomized encouragement experiment. However, time may be confounded with temporal trends that influence the outcomes. To disentangle these trends from the intervention effects, we replace the commonly used exclusion restrictions with temporal assumptions. We develop Bayesian procedures to estimate the causal effects and compare it to instrumental variables and matching approaches in simulations. The Bayesian approach outperforms the other 2 approaches in terms of estimation accuracy, and it is relatively robust to various violations of the common trends assumption. Taking advantage of the expansion of the Medicare Advantage (MA) program between 2011 and 2017, we implement the proposed method to estimate the effects of MA enrollment on the risk of skilled nursing facility residents being re-hospitalized within 30days after discharge from the hospital.

Artificial Intelligence Policies in Higher Education: A Randomized Field Experiment

Generative artificial intelligence (genAI) tools like ChatGPT are becoming increasingly embedded in daily life. In the field of education in particular, concerns have emerged about how these tools might positively or negatively impact student learning. In the current study, I conduct a randomized field experiment exploring how policies on genAI impact undergraduate students’ beliefs, behaviors, and grades. Students enrolled in two sections of the same upper-level undergraduate research methods course were randomly assigned to a section that either permitted or prohibited genAI tools. At the end of the semester, students enrolled in the section permitting genAI tools had a more positive view of genAI in higher education. However, students enrolled in the section prohibiting these tools had a higher final grade. These results highlight the need for evidence-based guidelines on how to minimize the risks and maximize the benefits of genAI in higher education.

Assessing the causal effect of genetically predicted metabolites and metabolic pathways on vitiligo: Evidence from Mendelian randomization and animal experiments

Assessing the causal effect of genetically predicted metabolites and metabolic pathways on vitiligo: Evidence from Mendelian randomization and animal experiments

Multi-objective Bayesian optimization of fused filament fabrication parameters for enhanced specific fracture energy in PLA-carbon fiber composites

Short carbon fiber-reinforced polymers (SCFRPs) are used in automotive and aerospace applications because of their superior strength-to-weight ratio and resistance to fatigue and thermal stresses. Fused filament fabrication (FFF) is considered an efficient fabrication method for SCFRP components, yet optimizing process parameters for specific performance targets is non-trivial and resource consuming. Herein, a multi-objective Bayesian optimization (MOBO) framework for FFF-based SCFRP fabrication has been developed to fine-tune process parameters to optimize the trade-off between tensile strength and weight of the products. Initial random experiments were conducted on printed samples with different infill densities, printing speeds, and layer heights to evaluate their impact on specific tensile fracture energy (STFE) and component weight. These parameters were found to influence STFE and weight significantly. Data-driven sequential experimentation was then applied to identify a set of optimal values for these parameters. A surrogate-based optimization framework was utilized, and optimal points were determined through iterative refinement, with the process culminating when further iterations did not significantly enhance weight or STFE, indicating Pareto optimality. The efficacy of the methodology was demonstrated by identifying FFF parameters that achieve Pareto optimal STFE-to-weight ratios in nine iterations. This data-driven approach provides an efficient route to process optimization compared to heuristic and computational methods, streamlining the SCFRP design process via FFF and potentially extending to other composite systems and manufacturing processes.

Multi-direction vibration isolation via transverse isotropy control of metal rubber

Abstract Metal rubber (MR) is widely used for vibration isolation due to its high damping and designable stiffness, but the isolation efficiency is relatively low in the non-molding direction due to its transverse isotropy. In this paper, the transverse isotropy of MR is controlled by design of manufacturing parameters (wire diameter, spiral diameter and compression ratio), material parameter (relative density) and geometric structure, to improve the isolation efficiency in the non-molding direction. Firstly, the helix unit was adopted to establish the relation between the transverse isotropy of MR material and manufacturing parameters, and manufacturing parameters were designed to control the transverse isotropy of MR material. Next, based on parametric modeling and genetic algorithm, material parameters and geometric structure were optimized to control the transverse isotropy of MR isolator. Finally, the transverse isotropy of the designed MR isolator was tested by quasi-static experiments, and multi-direction vibration isolation performance was verified by random vibration experiments. It is demonstrated that the transverse isotropy is controlled efficiently, and the vibration isolation efficiencies of the proposed MR isolator are all up to 85% under 15-2000Hz broadband random vibration in X, Y and Z directions.

Nonlinear moderating effects of individual social engagement in freemium strategies on digital content platforms

PurposeThis study aims to investigate the moderating effects of individual social engagement on the effectiveness of freemium strategies in digital content platforms.Design/methodology/approachThis study involved conducting a randomized field experiment with 74,758 consumers on a prominent e-book platform in China, comparing the effects of offering the first 50 chapters for free against no free content. Additionally, a causal random forest machine learning algorithm was applied to analyze data and optimize strategies based on individual social engagement levels.FindingsThis study indicates freemium strategies on digital content platforms can increase consumer willingness to pay but may reduce social community participation. These effects are moderated by consumers' prior social engagement, with excessive interaction leading to diminishing returns.Practical implicationsThe study offers actionable insights for digital content managers, showing how tailored freemium strategies can effectively balance consumer engagement and revenue generation. The findings suggest that platforms can significantly enhance profitability by moderating free content offerings based on detailed analysis of consumer engagement histories.Originality/valueThis study enhances the understanding of freemium strategies by showcasing their dual impact on consumer willingness to pay and social engagement, and detailing the complex, non-linear effects of individual social engagement, which challenges the traditional linear assumptions in existing literature. Additionally, it provides insights for implementing mixed marketing strategies on digital platforms, where multiple strategies often interact, guiding the effective management of these complexities.

Fact-checking information from large language models can decrease headline discernment

Fact checking can be an effective strategy against misinformation, but its implementation at scale is impeded by the overwhelming volume of information online. Recent AI language models have shown impressive ability in fact-checking tasks, but how humans interact with fact-checking information provided by these models is unclear. Here, we investigate the impact of fact-checking information generated by a popular large language model (LLM) on belief in, and sharing intent of, political news headlines in a preregistered randomized control experiment. Although the LLM accurately identifies most false headlines (90%), we find that this information does not significantly improve participants' ability to discern headline accuracy or share accurate news. In contrast, viewing human-generated fact checks enhances discernment in both cases. Subsequent analysis reveals that the AI fact-checker is harmful in specific cases: It decreases beliefs in true headlines that it mislabels as false and increases beliefs in false headlines that it is unsure about. On the positive side, AI fact-checking information increases the sharing intent for correctly labeled true headlines. When participants are given the option to view LLM fact checks and choose to do so, they are significantly more likely to share both true and false news but only more likely to believe false headlines. Our findings highlight an important source of potential harm stemming from AI applications and underscore the critical need for policies to prevent or mitigate such unintended consequences.

A Survey on Scenario Theory, Complexity, and Compression-Based Learning and Generalization.

This work investigates formal generalization error bounds that apply to support vector machines (SVMs) in realizable and agnostic learning problems. We focus on recently observed parallels between probably approximately correct (PAC)-learning bounds, such as compression and complexity-based bounds, and novel error guarantees derived within scenario theory. Scenario theory provides nonasymptotic and distributional-free error bounds for models trained by solving data-driven decision-making problems. Relevant theorems and assumptions are reviewed and discussed. We propose a numerical comparison of the tightness and effectiveness of theoretical error bounds for support vector classifiers trained on several randomized experiments from 13 real-life problems. This analysis allows for a fair comparison of different approaches from both conceptual and experimental standpoints. Based on the numerical results, we argue that the error guarantees derived from scenario theory are often tighter for realizable problems and always yield informative results, i.e., probability bounds tighter than a vacuous [0,1] interval. This work promotes scenario theory as an alternative tool for model selection, structural-risk minimization, and generalization error analysis of SVMs. In this way, we hope to bring the communities of scenario and statistical learning theory closer, so that they can benefit from each other's insights.

Inevitability of knowing less than nothing

A colloquial interpretation of entropy is that it is the knowledge gained upon learning the outcome of a random experiment. Conditional entropy is then interpreted as the knowledge gained upon learning the outcome of one random experiment after learning the outcome of another, possibly statistically dependent, random experiment. In the classical world, entropy and conditional entropy take only non-negative values, consistent with the intuition that one has regarding the aforementioned interpretations. However, for certain entangled states, one obtains negative values when evaluating commonly accepted and information-theoretically justified formulas for the quantum conditional entropy, leading to the confounding conclusion that one can know less than nothing in the quantum world. Here, we introduce a physically motivated framework for defining quantum conditional entropy, based on two simple postulates inspired by the second law of thermodynamics (non-decrease of entropy) and extensivity of entropy, and we argue that all plausible definitions of quantum conditional entropy should respect these two postulates. We then prove that all plausible quantum conditional entropies take on negative values for certain entangled states, so that it is inevitable that one can know less than nothing in the quantum world. All of our arguments are based on constructions of physical processes that respect the first postulate, the one inspired by the second law of thermodynamics.

Evaluation of Bone Supported Ultra Lock Ezy Bar versus Erich Arch Bar for the Treatment of Mandibular Fractures

ABSTRACT Introduction: Maxillofacial trauma treatment involves maxillomandibular fixation. Despite reduced need for post-operative MMF with plating devices, temporary intraoperative MMF is still necessary for proper tooth positioning. The aim of this research is to evaluate the cost-effectiveness, impact on patient gingival health, glove perforation, and time required for utilizing the Ultra Lock Ezy Bar compared to Erich’s Arch Bar. Methods: Ten patients with mandibular fractures were randomly split into two groups for treatment: Group A (study) and Group B (control). Group A received treatment with an Ultra Lock Ezy Bar and screws, while Group B received treatment using an Erich’s Arch Bar and wires. Follow-ups were done for suture removal, wound healing, and evaluations on gingival index, glove perforations, and treatment times. Results: In application time, group A had mean of 44.00 ± 3.391 and group B had 90 ± 9.354, with no significant difference. Neither group showed significant mean differences in removal time. Group B had statistically significant glove perforation (4.80 ± 0.837) compared to 0 in group A. Pre-operative gingival index comparison showed significance. After four weeks, group B’s mean index was 1.720 ± 0.2387 and group A’s was 1.120 ± 0.2049 with no statistical distinctions. Conclusion: To sum up, the randomized control experiment produced convincing results when comparing Erich’s Arch Bar with Ultra Lock Ezy Bar for mandibular fixation in ten patients. With notably quicker application and removal times, no glove perforations, and comparable gingival health immediately following surgery, Ultra Lock Ezy Bar showed clear benefits.

Gut microbiotas, inflammatory factors, and mental-behavioral disorders: A mendelian randomization study

BackgroundThe Mendelian randomization approach has emerged as a powerful tool, leveraging genetic variations as natural random experiments to minimize confounding and infer causality with unique advantages. Previous research has highlighted the crucial roles of gut microbiotas and inflammatory factors in mental-behavioral disorders, albeit to varying degrees. However, the precise causal relationship between gut microbiotas and mental-behavioral disorders remains elusive, and the potential role of inflammatory factors as mediators in this process is unclear. MethodsTo investigate the associations between gut microbiotas, inflammatory factors, and mental-behavioral disorders, we pooled data from large-scale genome-wide association studies (GWAS). Our study screened 27 diseases, encompassing nine subtypes of mental-behavioral disorders: neurodevelopmental disorders, eating disorders, sleep disorders, schizophrenia spectrum disorders, stress- and trauma-related disorders, mood and affective disorders, cognitive and executive function disorders, personality and somatization disorders, and addiction disorders. Mendelian randomization(MR) was employed to assess causal relationships between gut microbiotas, inflammatory factors, and these mental-behavioral disorders, with inverse variance weighting (IVW) as the primary statistical method. Furthermore, we explored whether inflammatory factors mediate the relationship between gut microbiotas and mental-behavioral disorders. ResultsHaving investigated the intricate interplay among gut microbiota, inflammatory factors, and mental-behavioral disorders, we have identified nine pivotal inflammatory factors that intricately regulate the progression of eight distinct disease subtypes. Noteworthy among these findings, levels of CC motif chemokine ligand 28 (CCL28) and CC motif chemokine ligand 25 (CCL25) are associated with the progression of attention-deficit/hyperactivity disorder (ADHD), interleukin-18 (IL-18) levels are linked to anorexia, IL-12β levels are related to schizophrenia (SZ) progression, IL-8 levels are associated with manic episodes, and IL-10 and monocyte chemoattractant protein-2 (MCP-2) levels are closely related to enduring personality changes(EPC). Additionally, fibroblast growth factor 19 (FGF19) levels are associated with cognitive disorders, while C-X-C motif chemokine ligand 1 (CXCL1) levels are related to executive functioning. ConclusionGut microbiotas and mental-behavioral disorders have causal relationships, with inflammatory factors mediating the pathway from gut microbiotas to these disorders.

A novel method for failure probability prediction of plain weave composites considering loading randomness and dispersion of strength

A new method based on combined residual stiffness-strength degradation is developed to predict the failure probability of plain weave composites subjected to random fatigue loadings. All the parameters presented in the proposed analytical model are characterized using the outcomes from quasi-static and constant amplitude fatigue testing. The evolution of residual strength is obtained based on combined residual stiffness-strength degradation model, which can greatly reduce the cost of the experiments. The Weibull distribution with two parameters is used to account for the dispersion of residual strength. Combing with randomness statistics of the fatigue loadings and the interference criterion of stress-strength, the fatigue failure behavior and failure probability are obtained. The narrow-band random vibration experiments were conducted to generate the random loadings and validate the predicted results. The approach proposed in this paper takes full advantage of residual stiffness or residual strength method and has better accuracy.

How effective is AI augmentation in human–AI collaboration? Evidence from a field experiment

PurposeCompanies increasingly leverage artificial intelligence (AI) to enhance human performance, particularly in e-commerce. However, the effectiveness of AI augmentation remains controversial. This study investigates whether, how and why AI enhances human agents’ sales through a randomized field experiment.Design/methodology/approachThis study conducts a two-by-two factorial randomized field experiment (N = 1,090) to investigate the effects of AI augmentation on sales. The experiment compares sales outcomes handled solely by human agents with those augmented by AI, while also examining the moderating effect of agents’ experience levels and the underlying mechanisms behind agents’ responses.FindingsThe results reveal that AI augmentation leads to a significant 5.46% increase in sales. Notably, the impact of AI augmentation varies based on agents’ experience levels, with inexperienced agents benefiting nearly six times more than their experienced counterparts. Mediation analysis shows that AI augmentation improves response timeliness, accuracy and sentiment, thereby boosting sales.Originality/valueThis study highlights the role of AI augmentation in human–AI collaboration, demonstrates the varying impacts of AI augmentation based on agents’ experience levels and offers insights for organizations on how to regulate AI augmentation to enhance agent responses and drive sales.

Divided Attention Has Limited Effects on Speech Sensorimotor Control.

When vowel formants are externally perturbed, speakers change their production to oppose that perturbation both during the ongoing production (compensation) and in future productions (adaptation). To date, attempts to explain the large variability across individuals in these responses have focused on trait-based characteristics such as auditory acuity, but evidence from other motor domains suggests that attention may modulate the motor response to sensory perturbations. Here, we test the extent to which divided attention impacts sensorimotor control for supralaryngeal articulation. Neurobiologically healthy speakers were exposed to random (Experiment 1) or consistent (Experiment 2) real-time auditory perturbation of vowel formants to measure online compensation and trial-to-trial adaptation, respectively. In both experiments, participants completed two conditions: one with a simultaneous visual distractor task to divide attention and one without this secondary task. Divided visual attention slightly reduced online compensation, but only starting > 300 ms after vowel onset, well beyond the typical duration of vowels in speech. Divided attention had no effect on adaptation. The results from both experiments suggest that the use of sensory feedback in typical speech motor control is a largely automatic process unaffected by divided visual attention, suggesting that the source of cross-speaker variability in response to formant perturbations likely lies within the speech production system rather than in higher-level cognitive processes. Methodologically, these results suggest that compensation for formant perturbations should be measured prior to 300 ms after vowel onset to avoid any potential impact of attention or other higher-order cognitive factors.

Low-resource dynamic loading identification of nonlinear system using pretraining

To address the challenge of load identification in nonlinear systems, an audio neural network-based method called WaveNet is proposed that leverages its capability to capture long-term dependencies in mechanical systems, enabling accurate load identification. Unlike traditional dynamic load identification methods that often encounter difficulties with matrix solutions, this approach takes advantage of WaveNet’s capabilities, enhancing both accuracy and efficiency. We integrate pre-training and transfer learning techniques to address the data scarcity challenges often encountered in real-world engineering applications. By transferring features across distributed datasets, this method reduces the dependency on single-task data, thereby improving model robustness. The performance of the WaveNet method is rigorously evaluated against traditional benchmarks such as Multilayer Perceptron (MLP) and Convolutional Neural Network (CNN) benchmarks under random load conditions applied to a complex structural framework. The proposed method achieves a root mean squared error (RMSE) of 1.521 and a determination coefficient (R²) of 0.996 in the random load case, demonstrating superior accuracy compared to other approaches. Moreover, its applicability is verified through simulations of both impact load and harmonic load scenarios, showcasing the effectiveness of transfer learning in overcoming domain discrepancies. Finally, the method is tested in random experiments to validate its engineering applicability. The results highlight the significant accuracy improvement in low-resource tasks achieved through pre-training, showcasing the potential and value of the proposed method.

An Overview of the Basic Theory of Probability Theory and its Commercial Application

Abstract. Probability theory is a key field of mathematics that focus on developing theoretical models and mathematical frameworks to describe and analyze random events or processes. As a part of basic science, probability theory holds a central place in pure mathematics while also serving a crucial function across various domains, including natural science, social science and engineering technology. This paper summarizes the main concepts, basic theories and commercial applications of probability theory. Probability theory encompasses ideas such as random experiment, sample space, event, probability, conditional probability, independence, Bayes Theorem, random variable, probability distribution, expected value and variance, law of large numbers and central limit theorem. This paper provides an overview of key concepts and theorems from probability theory, along with their practical applications in everyday life, in order to make people deeply recognize the broad and profound contents and research fields of probability theory and its wide application in various fields as an academic tool, which has a great relation with people.

Quantum Security of a Compact Multi-Signature

With the rapid advances in quantum computing, quantum security is now an indispensable property for any cryptographic system. In this paper, we study how to prove the security of a complex cryptographic system in the quantum random oracle model. We first give a variant of Zhandry’s compressed random oracle (CStO), called a compressed quantum random oracle with adaptive special points (CStOs). Then, we extend the on-line extraction technique of Don et al. (EUROCRYPT’22) from CStO to CStOs. We also extend the random experiment technique of Liu and Zhandry (CRYPTO’19) for extracting the CStO query that witnesses the future adversarial output. With these preparations, a systematic security proof in the quantum random oracle model can start with a random CStO experiment (that extracts the witness for the future adversarial output) and then converts this game to one involving CStOs. Next, the online extraction technique for CStOs can be applied to extract the witness for any online commitment. With this strategy, we give a security proof of our recent compact multi-signature framework that is converted from any weakly secure linear ID scheme. We also prove the quantum security of our recent lattice realization of this linear ID scheme by iteratively applying the weakly collapsing protocol technique of Liu and Zhandry (CRYPTO 2019). Combining these two results, we obtain the first quantum security proof for a compact multi-signature.

A Unified Framework for Covariate Adjustment Under Stratified Randomisation

ABSTRACTRandomisation, as a key technique in clinical trials, can eliminate sources of bias and produce comparable treatment groups. In randomised experiments, the treatment effect is a parameter of general interest. Researchers have explored the validity of using linear models to estimate the treatment effect and perform covariate adjustment and thus improve the estimation efficiency. However, the relationship between covariates and outcomes is not necessarily linear, and is often intricate. Advances in statistical theory and related computer technology allow us to use nonparametric and machine learning methods to better estimate the relationship between covariates and outcomes and thus obtain further efficiency gains. However, theoretical studies on how to draw valid inferences when using nonparametric and machine learning methods under stratified randomisation are yet to be conducted. In this paper, we discuss a unified framework for covariate adjustment and corresponding statistical inference under stratified randomisation and present a detailed proof of the validity of using local linear kernel‐weighted least squares regression for covariate adjustment in treatment effect estimators as a special case. In the case of high‐dimensional data, we additionally propose an algorithm for statistical inference using machine learning methods under stratified randomisation, which makes use of sample splitting to alleviate the requirements on the asymptotic properties of machine learning methods. Finally, we compare the performances of treatment effect estimators using different machine learning methods by considering various data generation scenarios, to guide practical research.