imputeSE: Impute missing standard errors

In MJAlexander/distortr: Temporal smoothing methods for demographic time series

Description

Impute missing standard errors by country based on other standard errors in that country. Imputed value is assumed to be the maximum standard error observed in that country multiplied by a specificed factor. If all standard errors within a country are missing, a pre-specified value can be imputed.

Usage

imputeSE(d, iso.column, data.column, se.column, transform.se.column = NULL,
  mult.factor = 1.5, impute.value = 0.025, transform.impute.value = 0.1)

Arguments

d	a dataframe that contains country codes/names, data column and standard error column
iso.column	string specifying name of column which contains country codes/names
data.column	string specifying name of column which contains observations
se.column	string specifying name of column which contains standard errors
transform.se.column	string specifying name of column which contains standard errors of transformed variable, if using
mult.factor	multiplication factor to multiply the maximum observed standard error
impute.value	value of standard error to impute if no information available in country.
transform.impute.value	value of transformed standard error to impute if no information available in country.

Value

A data frame of final standard error values and a column indicating whether the standard error value was imputed.

MJAlexander/distortr documentation built on July 17, 2020, 4:06 p.m.