probit-class: The 'probit' class and some basic methods
In bomarkussen/probit: Item response analysis with random effects.

Description Usage Arguments Note See Also

Objects of class probit as generated by probit is a list with entries as specified below.

fixed: Model for the fixed effects.
response.name: Name of response in internal representation.
weight.name: Name of weight in internal representation.
item.name: Name of item identifier in the model formula.
items.interval: Character vector with names of interval responses.
items.ordinal: Character vector with names of ordinal responses.
ordinal.levels: List of character vector with levels names of ordinal variables.
subject: Name of subject identifier.
random: List of mean models for the random effects.
dependence: Text string (marginal or joint) specifying the used correlation structure.
m.fixed: biglm-object with joint fit of fixed effects. Note that test and confidence intervals on this object does not make direct sense.
sigma2: List with estimated variances for the interval responses.
eta: List with estimated threshold parameters for the ordinal responses.
m.random: List of lm-objects for mean models on the random effects.
Gamma: Cholesky factor of estimated inverse variance of random effects.
mu: Matrix (subjects, q) of estimated approximate conditional means for the random effects, where q = number of random effects.
psi: Matrix (subjects, q*(q+1)/2) of estimated Cholesky factors of approximate inverse conditional variances for the random effects.
B: Number of replications for each subject in Monte-Carlo computation in minimization step.
BB: Number of replications for each subject in Monte-Carlo computation in maximization step.
logL: Guess at log(likelihood). Should be compared to sum(F1).
F1: Vector with subject-wise F1-values from last minimization step.
pvalue: p-value from latest T-test comparing the stochastic optimization.
iter: Number of minimization-maximization steps.
code: Iteration status: 0=iterations halted due to p-value, 1=iterations reached maximum iterations.
data: Tibble with dataset analyzed.

## S3 method for class 'probit'
print(x)

## S3 method for class 'probit'
summary(object)

## S3 method for class 'probit'
confint(object, level = 0.95)

## S3 method for class 'probit'
anova(object, BB = NULL)

## S3 method for class 'probit_confint'
print(x, digits = 4)

## S3 method for class 'probit_anova'
print(x, digits = 4)

## S3 method for class 'probit'
predict(object, re.form = TRUE)

recover_data.probit(object, ...)

emm_basis.probit(object, trms, xlev, grid, ...)

`x`	Object of class `probit`.
`object`	Object of class `probit`.
`level`	Approximative coverage for confidence interval. Defaults to `level=0.95`.
`BB`	Number of simulations per subject in `anova` or `update`. Defaults to `BB=NULL`, which corresponds to `BB` taken from the call object.
`digits`	Number of digits in print of `probit_anova` object. Defaults to `digits=4`.
`data`	Date frame with data on the wide format. Defaults to `data=NULL`, which corresponds to `data` taken from the call object.
`B`	Number of simulations in minimization step. Defaults to `B=NULL`, which corresponds to `B` taken from the call object.
`maxit`	Maximal number of minimization-maximization steps. Defaults to `maxit=20`.
`sig.level`	Significance level at which the iterative stochastic optimizations will be stopped. Defaults to `sig.level=0.60`.
`verbose`	Numeric controlling amount of convergence diagnostics. Default: `verbose=0` corresponding to no output.

anova re-simulates the underlying responses and random effects for the fixed effects model. Hence the output of the top part of anova table is stochastic.

probit

bomarkussen/probit documentation built on April 3, 2021, 7:38 p.m.