calcPtEst: Calculate point estimates and errors
In flaneuse/svywrangler: Helper Functions for Statistical Analyses of Survey Data

Description Usage Arguments Examples

Calculates point estimates and errors using the package 'survey' based on sampling weights from a survey design

1
2
3

calcPtEst(df, var, use_weights = FALSE, by_var = NA, design = NA,
  psu_var = NA, strata_var = NA, weight_var = NA, na.rm = TRUE,
  ci_factor = 2)

`df`	main data frame containing the raw data
`var`	string containing the variable name you want to average
`use_weights`	TRUE/FALSE for whether to apply a sampling frame or calculate a simple average.
`by_var`	(optional) string containing the variable name over which you want to calculate the estimates
`design`	(optional) svydesign object containing the sample frame
`psu_var`	(optional) if design isn't specified, string containing the primary sampling unit variable from the survey design (argument `id` in library('survey'))
`strata_var`	(optional) if design isn't specified, string containing the strata variable from the survey design (argument `strata` in library('survey')
`weight_var`	(optional)if design isn't specified, string containing the weights variable from the survey design (argument `weights` in library('survey')
`na.rm`	remove NAs from the mean or not
`ci_factor`	value to calculate confidence interval; in standard deviations. 1.96 standard deviations –> ~ 95 percent of the area under a normal gaussian distribution

# Generate simple random data to average
df = data.frame(region = rep(letters, 100), value = sample(100, 2600, replace = TRUE))

calcPtEst(df, 'value', by_var = 'region', use_weights = FALSE)

# generate data from a survey
data(api, package = 'survey')
dclus1 <- survey::svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)
# no weights
calcPtEst(apiclus1, var = 'api99', by_var = 'stype', use_weights = FALSE)
# specifying the sampling frame as a design object
calcPtEst(apiclus1, var = 'api99', by_var = 'stype', design = dclus1, use_weights = TRUE)