Disclaimer:

This application is currently under continuous development. While most functionality works, you may encounter unexpected behavior, please contact <email>, and we will do our best to look into it and correct it if needed. Thank you! More information will be provided shortly. Please go to Input Tab to perform the analysis.

Upload Data:

Select Variables:

Note: only unique variable selections will be allowed
WARNING: Selecting more than one independent variable will change the interpretation of the results and will assume simultaneous mediation using both variables.

Table: Raw Data

INSTRUCTIONS: Review Settings for Analysis

Selected Variables:


Selected Parameters:

MCMC:
R (# of draws)
Random seed/starting value
Keep (thinning parameter to minimize autocorrelation in draws (default is R/1000))
Burnin (number of saved MCMC draws to remove before analysis (default is 0.5*R/keep))
Number of MCMC chains for the Binary Mixture model (default is 10)
Initial MH step size for lambda (default is 0.5)


Priors: (any changes in the settings might impact impact the results of the estimation)
Aα (precision parameter for α parameters)
Aβγ (precision parameter for β and γ parameters)
Aλ (precision parameter for λ parameters)
ν (degrees of freedom for the variance parameters)
qy (scaling parameter for variance σy2, default is var(y))
qm (scaling parameter for variance σm2, default is var(m))



Results

Aggregate Model. Parameter 95% HPDIs

Aggregate Model. Proportions of Posterior Draws by Quadrant

Model fit

Aggregate Model. Plots of the Posterior Draws of Parameters

Numbers in the corners are proportions of the posterior distribution in that quadrant.

Downloads:
PDF of Plots



Results (Estimation may take 5-10 minutes)


Proportions of posterior distribution by Quadrant

Plots of the Posterior Draws of Parameters

Numbers in the corners are proportions of the posterior distribution in that quadrant
Downloads:
PDF of Plots

Parameter 95% HPDIs for the Mediating Segment (M)

Parameter 95% HPDIs for the General Segment (S)


Distribution of the individual probabilities to mediate

Histogram of the posterior distribution of parameter rho

Histogram of the posterior distributions(s) of parameter(s) Lambda

Posterior distribution of parameter RHO (95% HPDI)

Posterior distribution(s) of parameter(s) Lambda (95% HPDI)

Downloads:
PDF of Plots
Model fit and Diagnostics

Rhat for each variable (point est.)

Download

MCMC traceplots

Resources

This app is based on the working paper: Bayesian Analysis of Heterogeneous Mediation, which is currently under review process in the Journal of Consumer Research.
Authors' names are temporarily hidden to adhere to the blind review process.



References for function gelman.diag

Gelman, A and Rubin, DB (1992) Inference from iterative simulation using multiple sequences, Statistical Science, 7, 457-511.

Brooks, S P. and Gelman, A. (1998) General Methods for Monitoring Convergence of Iterative Simulations. Journal of Computational and Graphical Statistics, 7, 434-455.