anchors.data {anchors}R Documentation

Organized data from surveys with anchoring vignettes

Description

Create data object that will be used by anchors().

Usage

anchors.data(formula,
             data,
             method,
             subset,
             na.action   = na.omit,
             na.response = c(NA, 0), 
             min.response = 1,
             delete = c("minimal","maximal"), debug=0)

Arguments

formula

A list of named formulas giving a symbolic description of the model to be fit. See help(anchors) for details.

data

A data frame or matrix

method

Single string indicating method of analysis. See help(anchors) for details.

subset

Logical statement as used by function subset.

na.action

a function which indicates what should happen when the data contain 'NA's. Note also details of 'delete' in anchors.options.

na.response

vector of numeric values that should also be considered missing in self or vignette responses. Default is c(0,NA).

min.response

Default=1. This is a sanity check value. The code is designed to make sure that the user has not passed in values that are less than 1 and not included in na.response list.

delete

"minimal" deletes only cases with missing values that affect component of model

"maximal" forces listwise deletion on the basis of ALL variables in the entire formula list EVEN if vars not used by method

debug

Default: 0

Details

All of the response variables must be in the form of consequetive non-negative numeric integers, i.e., 1, 2, ... K.

** anchors currently does not support factor responses.**

The method="chopit" model does *not* require that every respondent answer all vignette or the self questions to be included in the model; only answers that are non-missing are used in the likelihood function.

The method="B", "C", and "order" requires that cases with any response that is missing be dropped.

By default, anchors only deletes those cases with missing values that affect the method of analysis requested. For example,

If a user would like to use exactly the same cases with method="chopit" as "B", "C" and "order" (i.e., drop cases with any missing responses), and vise versa (i.e., also drop any case with missing values in the tau=, tau1=, or self= formula), then the user should use the same list of formula for all methods and use the option anchors.options(delete="maximal").

Value

Return function will be of class 'anchors.data'.

anchors() and anchors.options()

Specifying no options is equivalent to

anchors(..., anchors.options(delete = "minimal"))

Alternative values include,

delete

Choose one from:

"minimal": delete only cases with missing values in components needed for current anchors analysis

"maximal": forces listwise deletion on the basis of any missing value in ALL variables in the formula list EVEN if not vars not used by method. Also deletes cases with ANY missing values in any responses (self or vignettes). This enables method="chopit" to use the same data as used by method="B" or "C", and vice versa.

Author(s)

Jonathan Wand http://wand.stanford.edu

References

Wand, Jonathan; Gary King; and Olivia Lau. (2007) “Anchors: Software for Anchoring Vignettes”. Journal of Statistical Software. Forthcoming. copy at http://wand.stanford.edu/research/anchors-jss.pdf

Wand, Jonathan and Gary King. (2007) Anchoring Vignetttes in R: A (different kind of) Vignette copy at http://wand.stanford.edu/anchors/doc/anchors.pdf

See Also

anchors

Examples

## load data and make a copy
data(mexchn)
dta <- mexchn

## insert missing values into covariates for the purpose of this demonstration
dta <- replace.value( dta, "educyrs", from=1:9  , to = NA)
dta <- replace.value( dta, "age"    , from=30:40, to = NA)

## formula that will be used throughout
fo <- list(self = xsayself ~ male + educyrs,
           vign = cbind(xsay3, xsay1) ~ 1,
           tau  = ~ educyrs)

## 'C' uses only cases with no missing responses (self, vign)
a1 <- anchors( fo, dta, method="C")
## number of cases used:
##  y0 = number of cases with self-responses
##  z0 = number of cases with all vignette responses
unlist(lapply(a1$data, NROW))[1:2]

## 'chopit' keeps case with some missing responses
##  but drops others with missing covariates
a2 <- chopit( fo, dta)
## number of cases used:
##  y0 = number of cases with self-responses
##       AND no missing covariates
##  z0 = number of cases with AT LEAST ONE vignette response
##       and no missing covariates
unlist(lapply(a2$data, NROW))[1:2]

## with ' delete = "maximal" '
## both procedures use the exact same cases 
## (IF the same formula is used for both methods)

## y0 and z0 here
a3 <- anchors( fo, dta, method="C", anchors.options(delete="maximal"))
unlist(lapply(a3$data, NROW))[1:2]

## is the same as y0 and z0 for chopit:
a4 <- chopit( fo, dta, options=anchors.options(delete="maximal"))
unlist(lapply(a4$data, NROW))[1:2]


[Package anchors version 3.0-8 Index]