When papaja is installed, papaja::apa_print() works on the objects
apabayes reports: a brms fit, an rstanarm fit, a lavaan or blavaan
fit, a brms hypothesis test, a LOO comparison from
loo::loo_compare(), a Bayes-factor model comparison from
bayestestR::bayesfactor_models(), inclusion Bayes factors from
bayestestR::bayesfactor_inclusion(), a table of contrasts or marginal
means from modelbased::estimate_contrasts() or
modelbased::estimate_means(), a table of Bayesian correlations from
correlation::correlation(), and a stored apabayes_tidy table.
Each
method calls apa_inline() and returns its result in the shape papaja
users address, so apa_print(fit)$full_result$wt works as it does on
an lm.
Usage
# S3 method for class 'apabayes_tidy'
apa_print(x, term = NULL, ..., in_paren = FALSE)
# S3 method for class 'brmsfit'
apa_print(x, term = NULL, ..., in_paren = FALSE)
# S3 method for class 'stanreg'
apa_print(x, term = NULL, ..., in_paren = FALSE)
# S3 method for class 'lavaan'
apa_print(x, term = NULL, ..., in_paren = FALSE)
# S3 method for class 'blavaan'
apa_print(x, term = NULL, ..., in_paren = FALSE)
# S3 method for class 'brmshypothesis'
apa_print(x, term = NULL, ..., in_paren = FALSE)
apa_print.compare.loo(x, term = NULL, ..., in_paren = FALSE)
apa_print.bayesfactor_models(x, term = NULL, ..., in_paren = FALSE)
apa_print.bayesfactor_inclusion(x, term = NULL, ..., in_paren = FALSE)
apa_print.estimate_contrasts(x, term = NULL, ..., in_paren = FALSE)
apa_print.estimate_means(x, term = NULL, ..., in_paren = FALSE)
apa_print.easycorrelation(x, term = NULL, ..., in_paren = FALSE)Arguments
- x
A
brmsfit,stanreg,lavaan,blavaan,brmshypothesis,compare.loo,bayesfactor_models,bayesfactor_inclusion,estimate_contrasts,estimate_meansoreasycorrelationobject, or an apabayes_tidy table.- term
The row to report, as in
apa_inline();NULLreports every row as named lists.- ...
Passed to
apa_inline():rhs,op,groupand the formatting options, and on a fitted model the arguments of itsapa_tidy()route.- in_paren
TRUEwrites brackets for the parentheses in the strings (χ²[24],log[*BF*~10~]), for a result quoted inside parentheses, as papaja's argument of that name does.
Value
An apa_results object.
Shape
Without term, estimate, statistic and full_result are named
lists with one string per row. The names follow papaja's rule:
parentheses, backticks and commas are removed and every other
character that is not a letter, digit or underscore becomes _, so
(Intercept) is Intercept and visual =~ x1 is visual_x1. A
parameter is named by its label, a hypothesis by its string, a fit
table and a model comparison by its model, a correlation by its pair
(mpg~~wt is mpg_wt). A name that would repeat
an earlier one gets _2, _3, …, skipping any name another row has;
a row with no usable name is row and its position.
With term, the three elements are plain strings and the result is
exactly apa_inline(x, term, ...).
The strings use apabayes' markup (see apa_inline()'s markup), not
the LaTeX math papaja writes, and print with a knit_print method, so
inline code needs no $full_result for a single result.
papaja owns apa_print() methods for emmGrid and BFBayesFactor
objects, and apabayes never replaces them; the Bayes-factor table
apabayes reads from a BFBayesFactor object is reported with
apa_print(apa_tidy(bf)). papaja's emmGrid method
reports a frequentist grid; on a grid from a Bayesian fit it prints
the estimate alone. Report such a grid through its tidy table,
apa_print(apa_tidy(grid)), which names the rows as papaja does
($full_result$cyl_f4_cyl_f6).
See also
apa_inline(), which these methods call.
Examples
model <- "visual =~ x1 + x2 + x3
textual =~ x4 + x5 + x6"
fit <- lavaan::cfa(model, lavaan::HolzingerSwineford1939)
r <- papaja::apa_print(fit, standardize = TRUE)
r$full_result$visual_x2
#> [1] ".43, 95% CI [.31, .55], *p* < .001"
papaja::apa_print(apa_tidy_sem_fit(fit), in_paren = TRUE)
#> row1: χ²[8] = 24.36, *p* = .002, CFI = .975, TLI = .953, RMSEA = .082, 90% CI [.046, .121], SRMR = .047
