It estimates the heterogeneity of the parameter estimates of the TSSEM objects using either the bootstrap or the delta methods.
tssemParaVar(
tssem1.obj,
tssem2.obj,
method = c("bootstrap", "delta"),
interval = 0.8,
Rep = 50,
output = c("data.frame", "matrices"),
nonPD.pop = c("replace", "nearPD", "accept")
)Either a data.frame or matrices of the output.
An object of class tssem1REM returned from
tssem1()
An object of class wls returned from
tssem2() or wls()
If it is bootstrap, random correlation matrices are
sampled from the tssem1.obj by the parametric bootstrap. If it is
delta, the delta method is used to estimate the heterogeneity of the
parameter estimates.
The desired interval, e.g., .8 or .95.
The number of parametric bootstrap. It is ignored when the method
is delta.
Either a data.frame or matrices of the output.
If it is replace, generated non-positive definite
matrices are replaced by generated new ones which are positive definite. If
it is nearPD, they are replaced by nearly positive definite matrices
by calling Matrix::nearPD(). If it is accept, they are
accepted.
Mike W.-L. Cheung <[email protected]>
The bootstrap method is based on the discussion in Cheung (2018) and Yu et al. (2016). The delta method is an alternative method to obtain the heterogeneity.
Cheung, M. W.-L. (2018). Issues in solving the problem of effect size heterogeneity in meta-analytic structural equation modeling: A commentary and simulation study on Yu, Downes, Carter, and O'Boyle (2016). Journal of Applied Psychology, 103, 787-803.
Yu, J. (Joya), Downes, P. E., Carter, K. M., & O'Boyle, E. H. (2016). The problem of effect size heterogeneity in meta-analytic structural equation modeling. Journal of Applied Psychology, 101, 1457-1473.
bootuniR1, bootuniR2,
Nohe15