The Bayesian ANCOVA allows the user to analyze the difference between multiple group means, while taking into account the effect of variables that have an influence on the dependent variable but are not part of the experimental manipulation (i.e., covariates).
Effects: By selecting this option, the inclusion probability of each component (i.e., model term) in the models will be calculated.
Effects
is selected, this method is used by default.Estimates: By selecting this option, a table with the model averaged posterior summary will be displayed. This table includes information about the model averaged posterior mean, the standard deviation, and the credible interval for each level of the fixed factors, their interactions, and the covariates.
Model
will be compared to the model containing the grand mean and the random factors, called the null model. This option is selected by default.Model
will be compared to the best model included in the analysis.
GIF demonstration: Select order
estimate
option in Output, since the estimates
option provides the posterior summary averaged over all the models included in the analysis, while this option gives the posterior summary for the single specified model only. Then it is possible to select: - Correction - Null control: When selecting this option, the prior odds will be corrected for multiple testing. This option is selected by default. At the moment, no output will be generated for the post hoc test when this option is not selected.
Model Comparison - Dependent Variable:
- Models: The first column contains all the models included in the analysis.
- Null model: This model contains the grand mean and the random factors.
- Independent Variable model: This model adds the effect of the independent variable.
- P(M): This column contains the prior model probability.
- P(M|data): This column contains the updated probability of the model given the data. This is called the posterior model probability.
- BFM : This column contains the posterior model odds. This is the change from the prior odds to the posterior odds for the model.
- BF10 : This column contains the Bayes factor that quantifies evidence for the alternative hypothesis relative to the null hypothesis/null model. However, when the option Compare to best model
is selected, the column will contain the Bayes factor that quantifies evidence for this model relative to the best model.
- BF01 : This column contains the Bayes factor that quantifies evidence for the null hypothesis/null model relative to the alternative hypothesis. However, when the option Compare to best model
is selected, the column will contain the Bayes factor that quantifies evidence for the best model relative to this model.
- error % : The error of the Gaussian quadrature integration routine used by the BayesFactor
package for the computation of the Bayes factor.
Analysis of Effects - Dependent Variable: - Effects: This column contains the components included in the models, such as independent variables and their interactions. - P(incl): This column contains the prior inclusion probability. This is the prior probability summed across all models that include the component. - P(incl|data): This column contains the posterior inclusion probability. This is the summed posterior probability over all models that include the component. - BFinclusion : This column contains the change from prior inclusion odds to posterior inclusion odds for each component averaged by all the models that includes the component.
Model Averaged Posterior Summary: - Variable: This column contains all the fixed factors, their interactions, and covariates included in the models. The first row contains information about the intercept. - Level: Each level of the factor and combination of levels of the interactions that are included in the model. - Mean: The model averaged mean. For the factors, this is the deviation from the intercept for each level of the factor. The level means for a factor sum to zero. - SD: The standard deviation of the model averaged mean. - % Credible interval: The credible interval of the mean. By default, this is set to 95%. - Lower: The lower bound of the credible interval of the mean. - Upper: The upper bound of the credible interval of the mean.
For each factor, interaction, and covariate, the model averaged posterior distributions per level are displayed, with on the x-axis the factor and on the y-axis the density. The posterior distribution for each level can either be displayed in the same plot, or in different plots for each level.
With the Q-Q plot, the normality of the residuals can be inspected visually. The theoretical quantiles are presented on the x-axis and standardized residuals on y-axis. The closer the dots are to the diagonal, the more evidence that the residuals are normally distributed.
The model averaged density of the R2 (i.e., explained variance).
Post Hoc Comparisons - Independent Variable: - The first columns contain the levels of the independent variable that are compared to each other. - Prior Odds: This column contains the prior odds. The prior odds are corrected for multiple testing (Westfall, Johnson, & Utts, 1997). - Posterior Odds: This column contains the posterior odds. The posterior odds are the prior odds multiplied by the Bayes factor. - BF10, U : This column contains the Bayes factor quantifying evidence for the alternative hypothesis relative to the null hypothesis/null model. The Bayes factor is uncorrected for multiple testing. - BF01, U : This column contains the Bayes factor quantifying evidence for the null hypothesis/null model relative to the alternative hypothesis. The Bayes factor is uncorrected for multiple testing. - error % : The error of the Gaussian quadrature integration routine used for the computation of the Bayes factor.
Single Model Posterior Summary: - Variable: This column contains all the factors, interactions, and covariates included in the models. The first row contains information about the intercept. - Level: Each level of the factors and combination of levels of the interactions that are included in the single model. - Mean: The single model mean. For the factors, this is the deviation from the intercept for each level of the factor. The level means for a factor sum to zero. - SD: The standard deviation of the single model mean. - % Credible Interval: The credible interval of the mean. By default, this is set to 95%. - Lower: The lower bound of the credible interval of the mean. - Upper: The upper bound of the credible interval of the mean.
Posterior Distributions: - For each factor, interaction, and covariate, the single model posterior distributions per level are displayed. The posterior distribution for each level can either be displayed in the same plot, or by different plots for each level.
Q-Q plot: - With the Q-Q plot, the normality of the residuals can be inspected visually. The theoretical quantiles are presented on the x-axis and standardized residuals on y-axis. The closer the dots are to the diagonal, the more evidence that the residuals are normally distributed.
Posterior R2: - The single model density of the R2 (i.e., explained variance), with the R2 on the x-axis and the density on the y-axis.
Descriptives - dependent variable: - Independent variables: The levels of the independent variable(s) included in the analysis. If more than 1, the descriptives will be displayed for each combination of levels of the independent variables. - Mean: The mean per level or, if more than 1 independent variable, the mean per combination of levels. - SD: The standard deviation. - N: The sample size. - % Credible interval: The credible interval of the mean. By default, this is set to 95%. - Lower: The lower bound of the credible interval of the mean. - Upper: The upper bound of the credible interval of the mean.
Descriptives Plot: - Independent variable on the x-axis and dependent variable on the y-axis. If other independent variables are included, either different lines representing different values of the other independent variable are displayed in the same plot, or different plots representing different values of the other independent variable are displayed.
Bar Plot: - Independent variable on the x-axis and dependent variable on the y-axis. If other independent variables are included, different plots representing different values of the other independent variable are displayed.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.