Bayesian Analysis Of Variance Research Articles

Population Performance and Individual Agreement of Coronary Artery Disease Polygenic Risk Scores.

Polygenic risk scores (PRSs) for coronary artery disease (CAD) are a growing clinical and commercial reality. Whether existing scores provide similar individual-level assessments of disease liability is a critical consideration for clinical implementation that remains uncharacterized. Characterize the reliability of CAD PRSs that perform equivalently at the population level at predicting individual-level risk. Cross-sectional Study. All of Us Research Program (AOU), Penn Medicine Biobank (PMBB), and UCLA ATLAS Precision Health Biobank. Volunteers of diverse genetic backgrounds enrolled in AOU, PMBB, and UCLA with available electronic health record and genotyping data. Polygenic risk for CAD from previously published PRSs and new PRSs developed separately from the testing cohorts. Sets of CAD PRSs that perform population prediction equivalently were identified by comparing calibration and discrimination (Brier score and AUROC) of generalized linear models of prevalent CAD using Bayesian analysis of variance. Among equivalently performing scores, individual-level agreement between risk estimates was tested with intraclass correlation (ICC) and Light's Kappa, measures of inter-rater reliability. 50 PRSs were calculated for 171,095 AOU participants. When included in a model of prevalent CAD, 48 scores had practically equivalent Brier scores and AUROCs (region of practical equivalence = 0.02). Across these scores, 84% of participants had at least one score in both the top and bottom risk quintile. Continuous agreement of individual risk predictions from the 48 scores was poor, with an ICC of 0.351 (95% CI; 0.349, 0.352). Agreement between two statistically equivalent scores was moderate, with an ICC of 0.649 (95% CI; 0.646, 0.652). Light's Kappa, used to evaluate consistency of assignment to high-risk thresholds, did not exceed 0.56 (interpreted as 'fair') across statistically and practically equivalent scores. Repeating the analysis among 41,193 PMBB and 50,748 UCLA participants yielded different sets of statistically and practically equivalent scores which also lacked strong individual agreement. Across three diverse biobanks, CAD PRSs that performed equivalently at the population level produced unreliable individual risk estimates. Approaches to clinical implementation of CAD PRSs must consider the potential for discordant individual risk estimates from otherwise indistinguishable scores.

Comments on the measurement of effect sizes for indirect effects in Bayesian analysis of variance.

Bayesian analysis of variance (BANOVA), implemented through R packages, offers a Bayesian approach to analyzing experimental data. A tutorial in Psychological Methods extensively documents BANOVA. This note critically examines a method for evaluating mediation using partial eta-squared as an effect size measure within the BANOVA framework. We first identify an error in the formula for partial eta-squared and propose a corrected version. Subsequently, we discuss limitations in the interpretability of this effect size measure, drawing on previous research, and argue for its potential unsuitability in assessing indirect effects in mediation analysis. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Consistency of Bayes factor estimates in Bayesian analysis of variance.

Factorial designs lend themselves to a variety of analyses with Bayesian methodology. The de facto standard is Bayesian analysis of variance (ANOVA) with Monte Carlo integration. Alternative, and readily available methods, are Bayesian ANOVA with Laplace approximation as well as Bayesian t tests for individual effects. This simulation study compared the three approaches regarding ordinal and metric agreement of the resulting Bayes factors for a 2 × 2 mixed design. Simulation results indicate remarkable disagreement of the three methods in certain cases, particularly when effect sizes are small and studies include small sample sizes. Findings further replicate and extend previous observations of substantial variability of ANOVAs with Monte Carlo integration across different runs of one and the same analysis. These observations showcase important limitations of current implementations of Bayesian ANOVA. Researchers should be mindful of these limitations when interpreting corresponding analyses, ideally applying multiple approaches to establish converging results. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Religion, personality, or none of them? Exploratory evidence on their correlations with economic preference parameters.

Previous empirical research in the social sciences suggests sizable differences across religious denominations for various outcomes of interest, such as educational attainment, marital stability, wealth, or fertility. A small body of previous experimental literature has investigated possible differences in economic preference parameters (including time preference and risk attitude) between religious denominations that might explain those differences. This research adds to the extant literature on religion and preferences by including information on subjects' Big Five personality traits and analyzing potential correlations with loss aversion. It combines experimental data from incentivized choices with information on religious affiliation during high school and Big Five personality traits to test for possible correlations of religious denomination with risk attitude, time preference, and loss aversion, using Bayesian analysis of variance (ANOVA) and Bayesian regression analysis. Bayesian ANOVA results suggest no preference differences between the religions analyzed in this research. When controlling for Big Five personality traits and a host of other background variables, Bayesian regression results suggest no effects of either religious affiliation or Big Five personality traits measures on the three economic preference parameters analyzed here. These findings highlight the complexity of the relationship between religion, personality traits, and economic preference parameters, suggesting that previously observed differences may be influenced by the preference measures used or other unobserved factors.

The Relationships Between External and Internal Training Loads in Mixed Martial Arts.

As a multidisciplined combat sport, relationships between external and internal training loads and intensities of mixed martial arts (MMA) have not been described. The aim of this study was to determine the external loads and intensities of MMA training categories and their relationship to internal loads and intensities. Twenty MMA athletes (age = 23.3 [5.3] y, mass = 72.1 [7.2]kg, stature = 171.5 [8.4]cm) were observed for 2 consecutive weeks. Internal load and intensity (session rating of perceived exertion [sRPE]) were calculated using the Foster RPE for the session overall (sRPE-training load [TL]) and segmented RPE (segRPE-TL) for each training category: warm-up, striking drills, wrestling drills, Brazilian jiujitsu (BJJ) drills, striking sparring, wrestling sparring, BJJ sparring, and MMA sparring. External load and intensity were measured via Catapult OptimEye S5 for the full duration of each session using accumulated Playerload (PLdACC) and PLdACC per minute (PLdACC·min-1). Differences in loads between categories and days were assessed via Bayesian analysis of variance (BF10 ≥ 3). Predictive relationships between internal and external variables were calculated using Bayesian regression. Session overall sRPE-TL = 448.6 (191.1)arbitrary units (AU); PLdACC = 310.6 (112)AU. Category segRPE-TL range = 33.8 (22.6)AU (warm-up) to 122.8 (54.6)AU (BJJ drills). Category PLdACC range = 44 (36.3)AU (warm-up) to 125 (58.8)AU (MMA sparring). Neither sRPE-TL nor PLdACC changed between days. PLdACC was different between categories. Evidence for regressions was strong-decisive except for BJJ drills (BF10 = 7, moderate). R2 range = .50 to .77, except for warm-up (R2 = .17), BJJ drills (R2 = .27), BJJ sparring (R2 = .49), and session overall (R2 = .13). While MMA training categories may be differentiated in terms of external load, overall session external load does not change within or between weeks. Resultant regression equations may be used to appropriately plan MMA technical/tactical training loads.

Data aggregation can lead to biased inferences in Bayesian linear mixed models and Bayesian analysis of variance.

Bayesian linear mixed-effects models (LMMs) and Bayesian analysis of variance (ANOVA) are increasingly being used in the cognitive sciences to perform null hypothesis tests, where a null hypothesis that an effect is zero is compared with an alternative hypothesis that the effect exists and is different from zero. While software tools for Bayes factor null hypothesis tests are easily accessible, how to specify the data and the model correctly is often not clear. In Bayesian approaches, many authors use data aggregation at the by-subject level and estimate Bayes factors on aggregated data. Here, we use simulation-based calibration for model inference applied to several example experimental designs to demonstrate that, as with frequentist analysis, such null hypothesis tests on aggregated data can be problematic in Bayesian analysis. Specifically, when random slope variances differ (i.e., violated sphericity assumption), Bayes factors are too conservative for contrasts where the variance is small and they are too liberal for contrasts where the variance is large. Running Bayesian ANOVA on aggregated data can-if the sphericity assumption is violated-likewise lead to biased Bayes factor results. Moreover, Bayes factors for by-subject aggregated data are biased (too liberal) when random item slope variance is present but ignored in the analysis. These problems can be circumvented or reduced by running Bayesian LMMs on nonaggregated data such as on individual trials, and by explicitly modeling the full random effects structure. Reproducible code is available from https://osf.io/mjf47/. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Supplemental Material for Data Aggregation Can Lead to Biased Inferences in Bayesian Linear Mixed Models and Bayesian Analysis of Variance

Supplemental Material for Data Aggregation Can Lead to Biased Inferences in Bayesian Linear Mixed Models and Bayesian Analysis of Variance

Micromollusk families in 2 contrasting seasons in Bahia de los Angeles, Gulf of California, Mexico

Marine mollusks provide important ecosystem services. They create habitats for benthic organisms, filter water, biodeposit organic carbon in the seafloor, and serve as food sources for other organisms. Studies of mollusk diversity require time-consuming methods to process samples and identify species. Thus, it is not surprising that most studies have focused on macromollusks that can be collected and processed easily, ignoring micromollusk species. Without understanding the ecology and distributions of micromollusks properly, it is impossible to assess their populations and implement adequate conservation measures. Here we present microgastropods collected in Bahia de los Angeles at the family level for September 2013 (summer) and February 2014 (winter). During each season, we sampled 6 coastal sediment stations, and analyzed depth, salinity, temperature, granulometry, and organic matter. A total of 20,353 specimens were collected: 15,310 in summer and 5,043 in winter. Seven micromollusk families were identified: Barleeidae, Caecidae, Cerithiidae, Eulimidae, Pyramidellidae, Rissoidae, and Tornidae. Caecidae and Barleeidae exhibited the highest densities among all families in summer and winter, respectively (278,044 ind·m–2, 142,222 ind·m–2). The lowest densities for summer and winter were observed for the Tornidae family (~1,867 ind·m–2, ~1,411 ind·m–2). Barleeidae, Caecidae, Cerithiidae, Rissoidae, and Tornidae were classified as herbivorous and detritivorous; Eulimidae and Pyramidellidae, as carnivorous and ectoparasitic. Tornidae and Eulimidae showed symbiotic relationships with various invertebrate species. The Bayesian analysis of variance indicated a high probability of differences only in summer (BF > 3). When comparing the stations in both periods, all stations, except station 6, showed differences between periods (BF > 3). The canonical correlation analysis indicated some associations between family abundances and temperature. This study provides valuable information that expands the knowledge of micromollusk biodiversity in the coastal area of Bahia de los Angeles, an important site for conservation in Mexico.

Bayesian Repeated-Measures Analysis of Variance: An Updated Methodology Implemented in JASP

Analysis of variance (ANOVA) is widely used to assess the influence of one or more experimental (or quasi-experimental) manipulations on a continuous outcome. Traditionally, ANOVA is carried out in a frequentist manner using p values, but a Bayesian alternative has been proposed. Assuming that the proposed Bayesian ANOVA is closely modeled after its frequentist counterpart, one may be surprised to find that the two can yield very different conclusions when the design involves multiple repeated-measures factors. We illustrate such a discrepancy with a real data set from a two-factorial within-subject experiment. For this data set, the results of a frequentist and Bayesian ANOVA are in a disagreement about which main effect accounts for the variance in the data. The reason for this disagreement is that frequentist and the proposed Bayesian ANOVA use different model specifications. As currently implemented, the proposed Bayesian ANOVA assumes that there are no individual differences in the magnitude of effects. We suspect that this assumption is neither obvious to nor desired by most analysts because it is untenable in most applications. We argue here that the Bayesian ANOVA should be revised to allow for individual differences. As a default, we suggest the standard frequentist model specification but discuss a recently proposed alternative and provide guidance on how to choose the appropriate model specification. We end by discussing the implications of the revised model specification for previously published results of Bayesian ANOVAs.

Bayesanova: An R package for Bayesian Inference in the Analysis of Variance via Markov Chain Monte Carlo in Gaussian Mixture Models

This paper introduces the R package [bayesanova](https://CRAN.R-project.org/package=bayesanova), which performs Bayesian inference in the analysis of variance (ANOVA). Traditional ANOVA based on null hypothesis significance testing (NHST) is prone to overestimating effects and stating effects if none are present. Bayesian ANOVAs developed so far are based on Bayes factors (BF), which also enforce a hypothesis testing stance. Instead, the Bayesian ANOVA implemented in bayesanova focusses on effect size estimation and is based on a Gaussian mixture with known allocations, for which full posterior inference for the component parameters is implemented via Markov-Chain-Monte-Carlo (MCMC). Inference for the difference in means, standard deviations and effect sizes between each of the groups is obtained automatically. Estimation of the parameters instead of hypothesis testing is embraced via the region of practical equivalence (ROPE), and helper functions provide checks of the model assumptions and visualization of the results.

Vocabulary Repetition Following Multisensory Instruction Is Ineffective on L2 Sentence Comprehension: Evidence From the N400.

Putting the principles of multisensory teaching into practice, this study investigated the effect of audio-visual vocabulary repetition on L2 sentence comprehension. Forty participants were randomly assigned to experimental and control groups. A sensory-based model of instruction (i.e., emotioncy) was used to teach a list of unfamiliar vocabularies to the two groups. Following the instruction, the experimental group repeated the instructed words twice, while the control group received no vocabulary repetition. Afterward, their electrophysiological neural activities were recorded through electroencephalography while doing a sentence acceptability judgment task with 216 sentences under acceptable (correct) and unacceptable (pragmatically violated) conditions. A one-way analysis of variance (ANOVA), a multivariate analysis of variance (MANOVA), and a Bayesian repeated-measures ANOVA were used to compare the behavioral and neurocognitive responses [N400 as the main language-related event-related brain potential (ERP) effect] of the two groups. The results showed no significant N400 amplitude difference in favor of any of the groups. The findings corroborated the ineffectiveness of two repetitions preceded by multisensory instruction on L2 sentence comprehension.

A Guide to the Bayesian Analysis of Consumer Behavior Experiments With BANOVA Using Worked Examples

This article provides a step-by-step guide for the analysis of experiments using Bayesian methods, using the BANOVA R package. We provide two worked examples. First, we reanalyze data from research by Romano and Balliet, which examined reciprocity and conformity as alternative mechanisms for cooperation between partners. The study has a between-subjects design and Poisson dependent variable, and we use Bayesian floodlight analysis to explore the interaction between reciprocity/conformity and two continuous covariates. Second, we reanalyze data from a study by Perfecto, Donnelly, and Critcher, who investigated whether mental simulation could be the psychological mechanism that explains how people make volume judgments of three-dimensional objects. The study has a mixed between- and within-subjects design with a Normal dependent variable, and we use Bayesian simple effects to explore the interactions between mental simulation and the shape and orientation of cups. The applications demonstrate the versatility of BANOVA (Bayesian Analysis of Variance) in analyzing a wide range of experimental designs and reveal that the results of the Bayesian analyses differ to some degree from those of the original studies.

Variability of Bayes Factor estimates in Bayesian Analysis of Variance

Variability of Bayes Factor estimates in Bayesian Analysis of Variance

Variability of Bayes Factor estimates in Bayesian Analysis of Variance

Variability of Bayes Factor estimates in Bayesian Analysis of Variance

The co-occurrence of Attention-Deficit/Hyperactivity Disorder and mathematical difficulties: An investigation of the role of basic numerical skills

BackgroundAttention-deficit/hyperactivity disorder (ADHD) and dyscalculia, also called mathematics disorder, frequently co-occur, yet the etiology of this comorbidity is poorly understood. AimsThis study investigated whether impairments in the understanding of numbers and magnitudes (basic numerical skills) are a unique risk factor for mathematical difficulties (MD) or a shared risk factor that could help to explain the association between ADHD and MD. Methods and proceduresBasic numerical skills were assessed with eight subtests in children (age 6–10 years, N = 86) with clinically significant ADHD symptoms and/or MD and typically developing children (control group). This double dissociation design allowed to test for main and interaction effects of ADHD and MD using both classical and Bayesian analysis of variance (ANOVA). Outcomes and resultsChildren with MD were impaired in transcoding, complex number and magnitude comparison, and arithmetic fact retrieval. They were not impaired in tasks assessing core markers of numeracy, which might be explained by the sample including children with mathematical difficulties instead of a diagnosed dyscalculia. ADHD was not associated with deficits in any of the tasks. The evidence for an additive combination of cognitive profiles was weak. Conclusions and implicationsImpairments in basic numerical skills are uniquely associated with MD and do not represent a shared risk factor for ADHD symptoms and MD.

A Bayesian shared components modeling approach to develop small area indicators of social determinants of health with measures of uncertainty.

Existing Canadian social determinants of health (SDOH) indicators do not quantify uncertainty to identify priority areas. The objectives of this methodologic study were: (1) to estimate and map small area (dissemination area) shared and variable-specific SDOH indicators with measures of uncertainty using a Bayesian model that accounts for spatial dependence; (2) to quantify geographic variation in the SDOH indicators and their contribution to a shared indicator; and (3) to assess the SDOH indicators' associations with behavioural risk factors and their consistency with the Ontario Marginalization Index (ON-Marg). Lower education-, income-, unemployment-, living alone- and visible minority-related variables used in existing Canadian SDOH indices were fit as dependent variables to a Bayesian model to produce area-based SDOH indicators that were mapped with measures of uncertainty in two study areas. The fractions of spatial variation explained by the model components were computed. Bayesian analysis of variance was used to examine the SDOH indicator associations with behavioural risk factors and their consistency with ON-Marg examined using Pearson's correlation coefficient. The shared component was strongly associated with material deprivation (i.e., income) in each study area; however, variable-specific SDOH indicators were important too. The SDOH indicators were associated with behavioural risk factors for chronic disease, particularly alcohol consumption and smoking, and the shared component estimates were consistent with the ON-Marg material deprivation. The Bayesian approach to produce SDOH indicators met the three study objectives and as such provides a new approach to prioritize areas that may experience health inequalities.

A Tutorial on Conducting and Interpreting a Bayesian ANOVA in JASP

L’analyse de variance (ANOVA) est la procédure standard utilisée pour l’inférence statistique dans les plans factoriels. En règle générale, les analyses de variance sont exécutées à l’aide de statistiques fréquentistes, où les valeurs p déterminent la significativité statistique en termes de « tout ou rien ». Ces dernières années, l’approche bayésienne de la statistique inférentielle est de plus en plus considérée comme une alternative légitime à la valeur p . Toutefois, l’adoption généralisée des statistiques bayésiennes, et en particulier de l’ANOVA bayésienne, est limitée par le fait que les concepts bayésiens sont rarement enseignés dans les cours de statistiques appliquées. Par conséquent, les utilisateurs peuvent ne pas savoir comment effectuer une ANOVA bayésienne et en interpréter les résultats. Nous fournissons ici un guide pour réaliser et interpréter une ANOVA bayésienne avec JASP, un logiciel statistique open-source ayant une interface utilisateur graphique. Nous expliquons les concepts clés de l’ANOVA bayésienne à l’aide de deux exemples empiriques.

Influence of sex and resistance training status on orofacial muscle strength and morphology in healthy adults between the ages of 18 and 40: A cross-sectional study.

To compare the orofacial muscle strength and facial muscle thickness between resistance-trained and non-resistance trained men and women. Resistance-trained (25 men and 22 women) and non-resistance trained (21 men and 30 women) adults (18-40 years) had standard measurements of orofacial muscle strength (ie, anterior tongue elevation, cheek compression, and lip compression) and ultrasound measurements of facial muscle thickness. Body composition (percent fat, etc.) was estimated using ultrasound prediction Equations. A Bayesian analysis of variance (ANOVA) with between subject factors of training status (yes/no) and sex (man/woman) were used to determine differences in muscle size and strength. Body mass, height, and percent fat were similar between resistance-trained and non-resistance trained individuals, while appendicular lean mass was higher in resistance-trained. There were no differences in orofacial muscle strength between resistance-trained and non-resistance trained. However, men had greater strength in every task except for anterior tongue strength which was similar between sexes (men: 66 vs women: 64 kPa). Handgrip strength was greater in men and in those who were resistance-trained. The frontalis muscle was greater in women than in men and in those who were non-resistance-trained than those resistance-trained. None of the other muscles differed by training status, however, all were greater in men. Results from our cross-sectional study would suggest that exercise-stimulation to the facial muscles during resistance training of the limbs and trunk did not reach a level where orofacial muscle strength could be changed. Sex differences in facial muscle thickness are very unique, although the reasons are unclear.

Effect of intumescent paint coating on mechanical properties of FRP bars at elevated temperature

This paper investigates the influence of intumescent paint on the performance of FRP bars subjected to low (25–450 °C) and severe (450–800 °C) elevated temperatures. In this research, glass and carbon FRP bars with epoxy resin and coated with nitrogen-based intumescent paint were used. In addition to the temperature effects, a variety of FRP bar diameters were employed to determine its effect on the tensile behavior of FRP bars in the presence of intumescent paint. Further, Bayesian regression methods and ANOVA (ANalysis Of VAriance) were applied on the results to develop a predictive model form and quantify the contribution of the variables in the tensile performance of FRP bars, respectively. The results revealed that the intumescent paint prevented the degradation of the mechanical properties of FRP bars significantly, up to 30% of GFRP bars tensile load capacity, at the temperature range 350–600 °C.

Influence of soil properties on the distribution of Deschampsia antarctica on King George Island, Maritime Antarctica

The extremely cold and infertile Antarctic is one of the harshest terrestrial ecosystems for the growth of vegetation, except for the grass species Deschampsia antarctica. We examined the main soil variables that determine the distribution of D. antarctica in King George Island by using Bayesian analysis of variance and regression methods. This study compared the density of D. antarctica between 2 sites; the density remained relatively stable at site 1, whereas it severely decreased in site 2 over a period of 3 years. Although site 2 showed better soil conditions for the growth of D. antarctica such as organic matter content, available phosphorus, NO3-N, and extractable cations, its poor drainage and low soil pH may affected the survival of D. antarctica by altering nutrition availability and inhibiting root respiration. Poisson analysis of covariance showed that the early melting of snow was also an important factor in the distribution of D. antarctica. The results also showed that seabirds and mammals might have greatly influenced the distribution of the grass species in King George Island by transferring nutrients from the sea onto land; thus, changing the chemical characteristics of the soil.