<b>BANOVA</b>: An <i>R</i> Package for Hierarchical Bayesian ANOVA

Chen, Dong; Wedel, Michel

doi:10.18637/jss.v081.i09

Cited by 18 publications

(11 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The need to specify priors incorporating subjective information often hinders the recourse to Bayesian ANOVA by applied researchers (Rouder et al, 2012 ). For this reason, recently proposed software packages such as BANOVA and JASP implement default priors that can be overlooked by data analyzers that do not want to incorporate actual prior information (Dong and Wedel, 2017 ; Wagenmakers et al, 2018a ). Unfortunately, inference relying on the default priors considered by these packages (and on most of those in the literature) for the variance components can run into problems, when mixed models specified on the log of the response variable are used.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Analysis of ANOVA and Mixed Models on the Log-Transformed Response Variable

2021

View full text Add to dashboard Cite

The analysis of variance, and mixed models in general, are popular tools for analyzing experimental data in psychology. Bayesian inference for these models is gaining popularity as it allows to easily handle complex experimental designs and data dependence structures. When working on the log of the response variable, the use of standard priors for the variance parameters can create inferential problems and namely the non-existence of posterior moments of parameters and predictive distributions in the original scale of the data. The use of the generalized inverse Gaussian distributions with a careful choice of the hyper-parameters is proposed as a general purpose option for priors on variance parameters. Theoretical and simulations results motivate the proposal. A software package that implements the analysis is also discussed. As the log-transformation of the response variable is often applied when modelling response times, an empirical data analysis in this field is reported.

show abstract

Section: Introductionmentioning

confidence: 99%

Bayesian Analysis of ANOVA and Mixed Models on the Log-Transformed Response Variable

2021

View full text Add to dashboard Cite

show abstract

“…The descriptive statistic for the primary objective was presented by mean, standard deviation, and quantiles, whereas the secondary objective was presented by standard deviation, quantiles, and simple effect. Due to the limitations of standard repeated measures ANOVA for categorical variables and unbalanced data, the inductive statistics for the primary and secondary objectives as well as subgroup analysis were calculated using the Bayesian version of the repeated-measures analysis of variance (BANOVA) [127]. The p-values of the multiple comparisons were adjusted using the Bayesian model [128].…”

Section: Methodsmentioning

confidence: 99%

Immediate effect of osteopathic techniques on human resting muscle tone in healthy subjects using myotonometry: A factorial randomized trial

Bohlen¹,

Schwarze²,

Richter

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Musculoskeletal disorders (MSDs) are highly prevalent, burdensome, and putatively associated with an altered human resting muscle tone (HRMT). Osteopathic manipulative treatment (OMT) is commonly and effectively applied to treat MSDs and reputedly influences the HRMT. Arguably, OMT may modulate alterations in HRMT underlying MSDs. However, there is sparse evidence even for the effect of OMT on HRMT in healthy subjects. Methods: A 3x3 factorial randomised trial was performed to investigate the immediate-term effect of myofascial release (MRT), muscle energy (MET), and soft tissue techniques (STT) on the HRMT of the corrugator supercilii (CS), superficial masseter (SM), and upper trapezius muscles (UT) in healthy subjects in Hamburg, Germany. Participants were randomised into three groups (1:1:1 allocation ratio) receiving treatment, according to different muscle-technique pairings, over the course of three sessions with one-week washout periods. Primarily, we assessed the effect of osteopathic techniques on muscle tone (F), biomechanical (S, D), and viscoelastic properties (R, C) from baseline to follow-up (main effect) and tested if specific muscle-technique pairs modulate the effect pre- to post-intervention (interaction effect) using the MyotonPRO (at rest). Data were analysed using descriptive (mean, standard deviation, quantiles, and simple effect) and inductive statistics (Bayesian ANOVA). Results: 59 healthy participants were randomised into three groups and two subjects dropped out from one group (n=20; n=20; n=19 and n=17, respectively). The CS produced frequent measurement errors and was excluded from analysis. The main effect changed significantly for F (-0.163 [0.060]; p=0.008), S (-3.060 [1.563]; p=0.048), R (0.594 [0.141]; p<0.001), and C (0.038 [0.017]; p=0.028) but not for D (0.011 [0.017]; p=0.527). The interaction effect did not change significantly (p>0.05). No adverse events were reported. Conclusion: OMT modified the HRMT in healthy subjects which may inform future research on MSDs. In detail, MRT, MET, and STT reduced the muscle tone (F), decreased biomechanical (S not D), and increased viscoelastic properties (R and C) of the SM and UT (CS was not measurable) at immediate term. However, the effect on HRMT was not modulated by muscle-technique interaction.

show abstract

“…The posterior distribution of a function of the parameter b, say h(b), can be calculated by applying that function to each of the MCMC draws b (r) , that is, h(b (r) ), and it can be characterized via summary statistics of these draws h(b (r) ). As will be explained later, this property is very useful for computing simple effects, indirect effects in mediation models (Zhang et al, 2009), effect sizes (Dong & Wedel, 2017), and Johnson-Neyman points in floodlight analyses (Wedel & Dong, 2019).…”

Section: Posterior Distributionsmentioning

confidence: 99%

“…The BANOVA framework and its companion software package builds on prior work by Dong and Wedel (2017) and Wedel and Dong (2019). It provides a flexible approach for analyzing data collected in experimental settings in which a sample of participants is exposed to between-subjects and/or within-subject manipulations, and in which continuous covariates or mediators may have been measured as well.…”

Section: The Banova Frameworkmentioning

confidence: 99%

A tutorial on the analysis of experiments using BANOVA.

Wedel¹,

Dong²

2022

Psychological Methods

Self Cite

View full text Add to dashboard Cite

Bayesian methods are increasingly used in psychology for analyzing experimental data and for identifying mechanisms that mediate the experimental treatments. This article provides a tutorial on a Bayesian approach to the analysis of variance (BANOVA), which provides a comprehensive and coherent framework for those analyses. BANOVA encompasses the analysis of data from between, within, and mixed experimental designs with normal and non-normal dependent variables and accommodates unobserved individual differences in participants' response to the experimental manipulations. An accompanying R package allows specification of a wide range of models with a simple syntax, and can calculate planned comparisons, simple effects, floodlight ranges, indirect effects in mediation, moderated mediation, and effect sizes of direct and indirect effects. The methodology and package are illustrated with applications to three data sets from previously published studies in psychology.

show abstract

BANOVA: An R Package for Hierarchical Bayesian ANOVA

Cited by 18 publications

References 33 publications

Bayesian Analysis of ANOVA and Mixed Models on the Log-Transformed Response Variable

Bayesian Analysis of ANOVA and Mixed Models on the Log-Transformed Response Variable

Immediate effect of osteopathic techniques on human resting muscle tone in healthy subjects using myotonometry: A factorial randomized trial

A tutorial on the analysis of experiments using BANOVA.

Contact Info

Product

Resources

About