lm_betaselect() and glm_betaselect()
to work when the calls used stats::lm()
or stats::glm().
(0.2.4.1)lm_betaselect() and glm_betaselect()
can accept the output of lm() or
glm() as the input. The call stored
in lm() and glm() will be retrieved
to run lm_betaselect() or glm_betaselect().
(0.2.2.2)Updated to work with tibble.
(0.2.2.1)
Fixed a harmless bug in
lm_betaselect() and
glm_betaselect(). Previously, if
.all., then all numeric variables in
the input dataset will be
standardized, including those not used
in the model. This would not affect the
results because these variables are
used anyway. Fixed and only variables
used in the model will be standardized.
(0.2.3.1)
lavaan (0.7-1).
(0.2.2)The print method of the output
of lav_betaselect() has a column
to indicate whether a parameter is
unstandardized. (0.1.3.1)
The computation of the standardized solution with bootstrapping now supports parallel processing. (0.1.3.2)
lav_betaselect() can also compute
the "standardized" intercepts, though
in the same way lavaan does, by
dividing an intercept by the
standard deviation of the outcome
variable (the y variable). This
can be enabled by setting
std_intercept to TRUE (FALSE
by default). (0.1.2.1)lav_betaselect() will no longer
check whether variables involved in
a product term have been mean-centered
by default (but can still be enabled
if necessary).
The coefficient of the so-called
"main effect" term of these variables
will now be computed correctly even
without mean-centering. (0.1.2.1, 0.1.2.2)lav_betaselect() will not compute
the coefficients of covariances and
variances that involve a product term,
if the variables involved are not
mean-centered. Without mean-centering,
it is impossible to compute them if
the joint distribution of the variables
is not multivariate normal. (0.1.2.1)lav_betaselect(): The standardized
coefficients of the component
variables of a product term are
incorrect if mean-centering is not
done. For now, lav_betaselect() will
check whether they are mena-centered.
If not, it will raise an error and
suggest users to mean-center the
involved variables first. (0.1.1)