BF.default | R Documentation |
The BF
function can be used for hypothesis testing and
model selection using the Bayes factor. By default exploratory hypothesis tests are
performed of whether each model parameter equals zero, is negative, or is
positive. Confirmatory hypothesis tests can be executed by specifying hypotheses with
equality and/or order constraints on the parameters of interest. Depending on the
class of the fitted model different Bayes factors are used as described in the output.
## Default S3 method:
BF(
x,
hypothesis = NULL,
prior.hyp.explo = NULL,
prior.hyp.conf = NULL,
prior.hyp = NULL,
complement = TRUE,
log = FALSE,
cov.prob = 0.95,
Sigma,
n,
...
)
## S3 method for class 'lm'
BF(
x,
hypothesis = NULL,
prior.hyp.explo = NULL,
prior.hyp.conf = NULL,
prior.hyp = NULL,
complement = TRUE,
log = FALSE,
cov.prob = 0.95,
BF.type = NULL,
iter = 10000,
...
)
## S3 method for class 'rma.uni'
BF(
x,
hypothesis = NULL,
prior.hyp.explo = NULL,
prior.hyp.conf = NULL,
prior.hyp = NULL,
complement = TRUE,
log = FALSE,
cov.prob = 0.95,
BF.type,
iter = 20000,
...
)
## S3 method for class 't_test'
BF(
x,
hypothesis = NULL,
prior.hyp.explo = NULL,
prior.hyp.conf = NULL,
prior.hyp = NULL,
complement = TRUE,
log = FALSE,
cov.prob = 0.95,
BF.type = NULL,
iter = 1e+06,
...
)
x |
An R object containing the outcome of a statistical analysis.
An R object containing the outcome of a statistical analysis. Currently, the
following objects can be processed: t_test(), bartlett_test(), lm(), aov(),
manova(), cor_test(), lmer() (only for testing random intercep variances),
glm(), coxph(), survreg(), polr(), zeroinfl(), rma(), ergm(), bergm(), or named
vector objects. In the case |
hypothesis |
A character string containing the constrained (informative) hypotheses to
evaluate in a confirmatory test. The default is NULL, which will result in standard exploratory testing
under the model |
prior.hyp.explo |
The prior probabilities of the hypotheses in the exploratory tests. Except for
objects of class |
prior.hyp.conf |
The prior probabilities of the constrained hypotheses in the confirmatory test. |
prior.hyp |
Deprecated. Please use the argument |
complement |
a logical specifying whether the complement should be added
to the tested hypothesis under |
log |
a logical specifying whether the Bayes factors should be computed on a log scale.
Default is |
cov.prob |
coverage probability of the Bayesian credibility interval in the |
Sigma |
An approximate posterior covariance matrix (e.g,. error covariance
matrix) of the parameters of interest. This argument is only required when |
n |
The (effective) sample size that was used to acquire the estimates in the named vector
|
... |
Parameters passed to and from other functions. |
BF.type |
For certain object classes of |
iter |
Number of iterations that are used to compute the Monte Carlo estimates.
Only used for certain hypothesis tests of class |
The function requires a fitted modeling object. Current analyses
that are supported: t_test
,
bartlett_test
,
aov
, manova
,
lm
, mlm
,
glm
, hetcor
,
lmer
, coxph
,
survreg
, ergm
,
bergm
,
zeroinfl
, rma
and polr
.
For testing parameters from the results of t_test(), lm(), aov(),
manova(), and bartlett_test(), hypothesis testing is done using
adjusted fractional Bayes factors are computed (using minimal fractions).
For testing measures of association (e.g., correlations) via cor_test()
,
Bayes factors are computed using joint uniform priors under the correlation
matrices. For testing intraclass correlations (random intercept variances) via
lmer()
, Bayes factors are computed using uniform priors for the intraclass
correlations. For all other tests, approximate adjusted fractional Bayes factors
(with minimal fractions) are computed using Gaussian approximations, similar as
a classical Wald test.
The output is an object of class BF
. The object has elements:
BFtu_exploratory
: The Bayes factors of the constrained hypotheses against
the unconstrained hypothesis in the exploratory test.
BFtu_main
(only for aov
objects with predictors of class factor
):
The Bayes factors of a constrained model where all levels of a factor
are assumed
to have the same effect on the outcome variable versus an unconstrained (full) model with
no constraints.
BFtu_interaction
(only for aov
objects with interaction effects with
predictors of class factor
): The Bayes factors of a constrained model where the effect
of the dummy variables corresponding to an interaction effects are assumed to be zero versus
an unconstrained (full) model with no constraints.
PHP_exploratory:
The posterior probabilities of the constrained hypotheses
in the exploratory test.
PHP_main
(only for aov
objects with predictors of class factor
):
The posterior probabilities a constrained model where all levels of a factor
are assumed
to have the same effect on the outcome variable versus an unconstrained (full) model with
no constraints.
PHP_interaction
(only for aov
objects with interaction effects with
predictors of class factor
): The posterior probabilities of a constrained model where the
effect of the dummy variables corresponding to an interaction effects are assumed to be zero versus
an unconstrained (full) model with no constraints.
BFtu_confirmatory
: The Bayes factors of the constrained hypotheses against
the unconstrained hypothesis in the confirmatory test using the hypothesis
argument.
PHP_confirmatory
: The posterior probabilities of the constrained hypotheses
in the confirmatory test using the hypothesis
argument.
BFmatrix_confirmatory
: The evidence matrix which contains the Bayes factors
between all possible pairs of hypotheses in the confirmatory test.
BFtable_confirmatory
: The Specification table
(output when printing the
summary
of a BF
for a confirmatory test) which contains the different
elements of the extended Savage Dickey density ratio where
The first column 'complex=
' quantifies the relative complexity of the
equality constraints of a hypothesis (the prior density at the equality constraints in the
extended Savage Dickey density ratio).
The second column 'complex>
' quantifies the relative complexity of the
order constraints of a hypothesis (the prior probability of the order constraints in the extended
Savage Dickey density ratio).
The third column 'fit=
' quantifies the relative fit of the equality
constraints of a hypothesis (the posterior density at the equality constraints in the extended
Savage Dickey density ratio).
The fourth column 'fit>
' quantifies the relative fit of the order
constraints of a hypothesis (the posterior probability of the order constraints in the extended
Savage Dickey density ratio)
The fifth column 'BF=
' contains the Bayes factor of the equality constraints
against the unconstrained hypothesis.
The sixth column 'BF>
' contains the Bayes factor of the order constraints
against the unconstrained hypothesis.
The seventh column 'BF
' contains the Bayes factor of the constrained hypothesis
against the unconstrained hypothesis.
The eighth column 'PHP
' contains the posterior probabilities of the hypotheses.
prior.hyp.explo
: The prior probabilities of the constrained hypotheses in the exploratory tests.
prior.hyp.conf
: The prior probabilities of the constrained hypotheses in the confirmatory test.
hypotheses
: The tested constrained hypotheses in a confirmatory test.
estimates
: Descriptives of unconstrained estimates based on flat priors (also for rma.uni
objects for
Bayesian meta-analyses).
model
: The input model x
.
bayesfactor
: The type of Bayes factor that is used for this model.
parameter
: The type of parameter that is tested.
log
: logical
whether the Bayes factors were reported on a log scale.
fraction_number_groupIDs
(only for objects of class lm
): The number of
'group identifiers' that were identified based on the number of unique combinations of level
s
of predictor variables of class factor
in the data. These group identifiers are used to automatically
specify the minimal fractions that are used to compute (adjusted) fractional Bayes factors.
fraction_groupID_observations
(only for objects of class lm
): A vector that
specifies to which 'group identifier' an observation belongs. The group identifiers are constructed
based on the unique combination of the levels
based on the predictor variables of class factor
of the observations.
call
: The call of the BF
function.
BF(default)
: S3 method for a named vector 'x'
BF(lm)
: S3 method for an object of class 'lm'
BF(rma.uni)
: BF S3 method for an object of class 'rma.uni'
BF(t_test)
: BF S3 method for an object of class 't_test'
Mulder, Williams, Gu, Tomarken, Böing-Messing, Olsson-Collentine, Meyerink, Menke, Fox, Rosseel, Wagenmakers, Hoijtink, and van Lissa (2021). BFpack: Flexible Bayes Factor Testing of Scientific Theories in R. Journal of Statistical Software, 100. <https://doi.org/10.18637/jss.v100.i18>
Mulder and Xin (2022). Bayesian Testing of Scientific Expectations under Multivariate Normal Linear Models. <https://doi.org/10.1080/00273171.2021.1904809>
Mulder and Gelissen (2021). Bayes factor testing of equality and order constraints on measures of association in social research. Journal of Applied Statistics, 50. <https://doi.org/10.1080/02664763.2021.1992360>
Mulder and Fox (2019). Bayes Factor Testing of Multiple Intraclass Correlations. Bayesian Analysis, 14. <http://doi.org/10.1214/18-BA1115>
Hoijtink, Mulder, van Lissa, and Gu (2018). A tutorial on testing hypotheses using the Bayes factor. Psychological Methods, 24(5), 539–556. <http://doi.org/10.1037/met0000201>
Boeing-Messing, van Assen, Hofman, Hoijtink, and Mulder (2017). Bayesian evaluation of constrained hypotheses on variances of multiple independent groups. Psycholological Methods, 22(2), 262-287. <https://doi.org/10.1037/met0000116>
van Aert and Mulder (2021). Bayesian hypothesis testing and estimation under the marginalized random-effects meta-analysis model. Psychonomic Bulletin and Review, 29, 55–69. <https://doi.org/10.3758/s13423-021-01918-9>
# EXAMPLE 1. One-sample t test
ttest1 <- t_test(therapeutic, mu = 5)
print(ttest1)
# confirmatory Bayesian one sample t test
BF1 <- BF(ttest1, hypothesis = "mu = 5")
summary(BF1)
# exploratory Bayesian one sample t test
BF(ttest1)
# EXAMPLE 2. ANOVA
aov1 <- aov(price ~ anchor * motivation,data = tvprices)
BF1 <- BF(aov1, hypothesis = "anchorrounded = motivationlow;
anchorrounded < motivationlow")
summary(BF1)
# EXAMPLE 3. linear regression
lm1 <- lm(mpg ~ cyl + hp + wt, data = mtcars)
BF(lm1, hypothesis = "wt < cyl < hp = 0")
# EXAMPLE 4. Logistic regression
fit <- glm(sent ~ ztrust + zfWHR + zAfro + glasses + attract + maturity +
tattoos, family = binomial(), data = wilson)
BF1 <- BF(fit, hypothesis = "ztrust > zfWHR > 0;
ztrust > 0 & zfWHR = 0")
summary(BF1)
# EXAMPLE 5. Correlation analysis
set.seed(123)
cor1 <- cor_test(memory[1:20,c(1,2,6)])
BF1 <- BF(cor1)
summary(BF1)
BF2 <- BF(cor1, hypothesis = "Rat_with_Im > Rat_with_Del > 0;
Rat_with_Im = Rat_with_Del = 0")
summary(BF2)
# correlations can also be computed between continuous/ordinal variables
memory_test <- memory[1:20,c(1,2,6)]
memory_test[,3] <- as.ordered(memory_test[,3])
cor2 <- cor_test(memory_test)
BF(cor2)
# EXAMPLE 6. Bayes factor testing on a named vector
# A Poisson regression model is used to illustrate the computation
# of Bayes factors with a named vector as input
poisson1 <- glm(formula = breaks ~ wool + tension,
data = datasets::warpbreaks, family = poisson)
# extract estimates, error covariance matrix, and sample size:
estimates <- poisson1$coefficients
covmatrix <- vcov(poisson1)
samplesize <- nobs(poisson1)
# compute Bayes factors on equal/order constrained hypotheses on coefficients
BF1 <- BF(estimates, Sigma = covmatrix, n = samplesize, hypothesis =
"woolB > tensionM > tensionH; woolB = tensionM = tensionH")
summary(BF1)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.