Functions that Compute Error Levels Given Audit Data

Description

Calculate the error amounts for all precincts in Z that were audited from the audit data, given as overstatement errors for all candidates.

Usage

1
2
3
4
5
compute.audit.errors(Z, audit = NULL, calc.e_p = calc.pairwise.e_p, w_p =
                 weight.function("no.weight"), bound.col = "tot.votes",
                 err.override = NULL)
calc.pairwise.e_p(Z, audit=NULL, err.override = NULL)
calc.overstatement.e_p(Z)

Arguments

Z

elec.data object

err.override

Assume a baserate of this amount of error everywhere, ignoring audit data. If non-null, use this as the found error in votes rather than the actual errors found in the audit.

bound.col

This is the vector (in audit) containing the maximum number of votes possible in the various precincts.

calc.e_p

Calculate e\_p or take as given.

w_p

The weight function to use to reweight the errors of precincts.

audit

The audit object, if it is not in the Z object, or if some other object other than the one in the Z object is desired to be considered as the audit object. Used by the simulation functions to generate errors for some fixed amount of error in conjunction with the err.override.

Details

compute.audit.errors uses the calc functions and the weight functions in a 1-2 combination.

calc.pairwise.e\_p() is often used with an err.override for simulation studies and whatnot to see what a fixed vote impact would have on taints for trinomial.

Value

compute.audit.errors returns a new audit table from Z with two new columns, err and err.weighted, corresponding to the errors found in each audited precinct before and after the weight function has been applied to them.

Note

Z must have an audit component, or one must be passed, for this function to make sense! Remember that audit objects have overstatements, NOT total votes for candidates. With err.override being set this is less relevant as the actual votes are usually ignored.

Author(s)

Luke W. Miratrix

See Also

See audit.totals.to.OS for a utility function that handles processing of audit data.

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.