drawPVs {Dire}

Draw plausible values (PVs) from an mml fit

Description

Draw plausible values (PVs) from an mml fit

Usage

drawPVs(x, npv, pvVariableNameSuffix = "_dire", ...)

## S3 method for class 'summary.mmlMeans'
drawPVs(x, npv = 5L, pvVariableNameSuffix = "_dire", ...)

## S3 method for class 'mmlMeans'
drawPVs(
  x,
  npv = 5L,
  pvVariableNameSuffix = "_dire",
  stochasticBeta = FALSE,
  normalApprox = TRUE,
  newStuDat = NULL,
  newStuItems = NULL,
  returnPosterior = FALSE,
  construct = NULL,
  ...
)

## S3 method for class 'mmlCompositeMeans'
drawPVs(
  x,
  npv = 5L,
  pvVariableNameSuffix = "_dire",
  stochasticBeta = FALSE,
  normalApprox = TRUE,
  newStuDat = NULL,
  newStuItems = NULL,
  verbose = TRUE,
  ...
)

Arguments

x

a fit from a call to mml

npv

integer indicating the number of plausible values to draw

pvVariableNameSuffix

suffix added to new PV variables after construct name and before the plausible value ID. For example, if there is a construct math and the suffix is the default _dire, then the fourth plausible value would have a column name, math_dire4.

...

additional parameters

stochasticBeta

logical when TRUE the regression coefficients will be drawn from their posterior distribution. Can also be a data frame of values (see Details).

normalApprox

logical must be TRUE to use the normal approximation to the posterior distribution rather than drawing from the actual posterior distribution.

newStuDat

new stuDat object, (see mml) for which plausible values will be drawn

newStuItems

new stuItems object, (see mml); unlike in mml students with no items can be passed to this function

returnPosterior

logical set to TRUE to change output to include two additional data frames (see Value).

construct

character, changes the name of the columns in the final data frame

verbose

logical set to TRUE to see the status of the processing

Details

When the argument passed to stocasticBeta is a data frame then each column is an element that will be used as a regression coefficient for that index of the coefficients vector. The row index used for the nth PV will be the nth row.

Value

when returnPosterior is FALSE returns a object of class DirePV which is a list of two elements. first, a data frame with a row for every row of newStuDat (or the original stuDat object)

• id the value of idVar in the model run

• [construct][pvVariableNameSuffix][L] every other column is a plausible value of this format. The [construct] is the name of the construct, the [pvVariableNameSuffix] is the value of the pvVariableNameSuffix argument, and the [L] part is the plausible value index, from 1 to npv.

The second argument is named newpvvars and is a list with an element for each set of construct that lists all of the variables in that construct.

When returnPosterior is TRUE returns list with three elements. One is named posterior and has one row per idvar level in the newStuDat argument and three columns:

• id the value of idVar in the model run

• mu the posterior mean

• sd the posterior standard deviation

the second list element is named X that is the design matrix for newStuDat (see Value for mml). The third list element is the rr1 element returned from mml with one column for each individual in newStuDat (see Value in mml).

Author(s)

Paul Bailey, Sun-joo Lee, and Eric Buehler

Examples

# See Examples in mml