rakeadj {rake}R Documentation

Adjust the marginal total weights of a "rake" class object

Description

The marginal total weights of the rake object are adjusted to match the population marginal total weights specified by marg. As a result, the total weights of each class of rake will be adjusted, and these adjusted weights can be used for estimation.

This is the second function in the rake, rakeadj, predict.rake Rake analysis series.

Usage

rakeadj(rake, marg, verbose)

Arguments

rake Object of class "rake" created with rake.
marg The population marginal total weights. This may be specified in any of the following ways:
numeric: vector of marginal total weights, matched to the values in rake by name.
data.frame: column "weight" must contain the weights, matched by names specified in column "name".
character: filename of an appropriate table that will be read as a data.frame.
verbose logical: Should the number of iterations for the rake adjustment to converge be printed?

Details

Rake adjustment via rakeadj is accomplished by, for each class, dividing by the marginal total weights of the class, then multiplying by the marginal total weight specified by marg.

After performing the rake adjustment above on each class in the rows of rake, the marginal totals of the rows of rake match the target marginal totals specified by marg, but the marginal totals of the columns of rake do not match the target marginal totals. So, the rake adjustment is then performed on each class of the columns of rake, which then syncs the column marginal totals but unsyncs the row marginal totals.

By repeatedly performing the rake adjustment, the marginal totals of the columns and rows of rake eventually converge to the marginal totals specified by marg.

Value

Object of class "rake", with adjusted weights.

Author(s)

Toby Dylan Hocking <tdhock@ocf.berkeley.edu>

References

Sharon L. Lohr. Sampling: Design and Analysis, pp. 269-271. Brooks/Cole, Pacific Grove, CA, 1999.

See Also

rake for creating the initial "rake" class object, predict.rake for making estimates with the adjusted weight values, simpleRake for performing the entire Raking process.

Examples

example(rake)
r <- rakeadj( r, statpoptotal, TRUE )
print(r)

[Package rake version 1.0 Index]