github-vignettes/weight.md
In southwick-associates/sastats: Calculate Statistics
Overview
Rake weighting in R is fairly straightforward using
anesrake::anesrake(), for which sastats::rake_weight() is a thin
wrapper. The key aspect of the process is getting the two input datasets
into the right format:
-
The target population dataset needs to be a named list (one
element per demographic variable) of target distributions. The
weights::wpct() function helps with this.
-
The survey dataset needs to be a data frame with a unique ID
variable (e.g., Vrid), and one variable for each demographic
measure.
Importantly, demographic variables must be stored either as factors or
numerics and the category coding much match between the two datasets.
You can also see production examples for
B4W-19-10
and AS/HS:
rakewt-ashs.
Example Data
For demonstration, package sastats includes survey and population
datasets which include several demographic variables:
library(dplyr)
library(sastats)
data(svy, pop)
svy <- select(svy$person, -weight)
# survey to be weighted
glimpse(svy)
#> Observations: 1,252
#> Variables: 5
#> $ Vrid <chr> "98", "99", "100", "101", "102", "103", "105", "106",...
#> $ sex <fct> Female, Female, NA, Male, Female, Male, Male, Female,...
#> $ age_weight <fct> 35-54, 35-54, NA, 35-54, 35-54, 18-34, 18-34, 35-54, ...
#> $ income_weight <fct> 0-25K, 0-25K, NA, 25-35K, 50-75K, 25-35K, 50-75K, 35-...
#> $ race_weight <fct> White, Not white or Hispanic, NA, White, Not white or...
# target population (outdoor recreationists) from a genpop survey
glimpse(pop)
#> Observations: 877
#> Variables: 5
#> $ sex <fct> Female, Female, Female, Male, Female, Female, Female,...
#> $ age_weight <fct> 18-34, 18-34, 18-34, 18-34, 18-34, 35-54, 35-54, 35-5...
#> $ income_weight <fct> 75-100K, 35-50K, 0-25K, 150K+, 0-25K, 25-35K, 75-100K...
#> $ race_weight <fct> White, White, Hispanic, White, White, White, White, W...
#> $ stwt <dbl> 0.8927761, 0.8927761, 1.1913576, 1.1056643, 0.8073446...
Importantly, the demographic variables have been encoded in the same way
between the 2 datasets (as factors in this case):
wtvars <- setdiff(names(svy), "Vrid")
sapply(wtvars, function(x) all(levels(svy[[x]]) == levels(pop[[x]])))
#> sex age_weight income_weight race_weight
#> TRUE TRUE TRUE TRUE
Population Distributions
The easiest way to see the format needed for the target population is to
run weights::wpct() on a demographic variable:
weights::wpct(svy$age_weight)
#> 18-34 35-54 55+
#> 0.2766497 0.2944162 0.4289340
For our example, we are defining the target population (outdoor
recreationists) using another general population survey dataset. In this
case, we need to ensure we use the stwt variable from that survey
since it was itself weighted:
weights::wpct(pop$age_weight, pop$stwt)
#> 18-34 35-54 55+
#> 0.3514503 0.3486601 0.2998896
The weights::wpct() function returns a vector, but we need a list of
vectors (one element per demographic variables of interest). A
straightforward method involves looping over the demographic variables
with sapply():
pop_target <- sapply(wtvars, function(x) weights::wpct(pop[[x]], pop[["stwt"]]))
pop_target
#> $sex
#> Male Female
#> 0.51747 0.48253
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.3514503 0.3486601 0.2998896
#>
#> $income_weight
#> 0-25K 25-35K 35-50K 50-75K 75-100K 100-150K 150K+
#> 0.17353726 0.09405487 0.14064324 0.18922885 0.15252025 0.16103245 0.08898309
#>
#> $race_weight
#> White Hispanic Not white or Hispanic
#> 0.73559778 0.17719770 0.08720452
If we weren’t using a reference population dataset, we would need to get
the target distributions into a list in some other way. For example, by
hand:
partial_target <- list(
"sex" = c("Male" = 0.517, "Female" = 0.483),
"age_weight" = c("18-34" = 0.351, "35-54" = 0.349, "55+" = 0.300)
# etc.
)
partial_target
#> $sex
#> Male Female
#> 0.517 0.483
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.351 0.349 0.300
Using weights::wpct() also of course provides a convenient method for
comparison with the survey dataset to be weighted:
sapply(wtvars, function(x) weights::wpct(svy[[x]]))
#> $sex
#> Male Female
#> 0.4318374 0.5681626
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.2766497 0.2944162 0.4289340
#>
#> $income_weight
#> 0-25K 25-35K 35-50K 50-75K 75-100K 100-150K 150K+
#> 0.15071973 0.12785775 0.14309907 0.20491109 0.15071973 0.15410669 0.06858594
#>
#> $race_weight
#> White Hispanic Not white or Hispanic
#> 0.82387807 0.09652837 0.07959356
Rake Weighting
We can now run the rake weighting procedure. By default, rake_weight()
returns a list of 2 elements: (1) the survey dataset with the weight
variable appended, and (2) the anesrake() return object, which
includes a bunch of useful summary statistics (including “design
effect”, which may be useful for estimating confidence intervals).
svy_wts <- rake_weight(svy, pop_target, "Vrid")
#> [1] "Raking converged in 21 iterations"
svy <- svy_wts$svy
summary(svy$weight)
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 0.3568 0.7408 0.9245 1.0000 1.1662 2.6893
wt_summary <- summary(svy_wts$wts)
deff <- wt_summary$general.design.effect
deff
#> [1] 1.186242
wt_summary
#> $convergence
#> [1] "Complete convergence was achieved after 21 iterations"
#>
#> $base.weights
#> [1] "No Base Weights Were Used"
#>
#> $raking.variables
#> [1] "sex" "age_weight" "income_weight" "race_weight"
#>
#> $weight.summary
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 0.3568 0.7408 0.9245 1.0000 1.1662 2.6893
#>
#> $selection.method
#> [1] "variable selection conducted using _pctlim_ - discrepancies selected using _total_."
#>
#> $general.design.effect
#> [1] 1.186242
#>
#> $sex
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> Male 0.51747 510 0.4318374 611.2812 0.51747 0.08563258
#> Female 0.48253 671 0.5681626 570.0070 0.48253 -0.08563258
#> Total 1.00000 1181 1.0000000 1181.2882 1.00000 0.17126517
#> Resid. Disc. Orig. Disc.
#> Male 0.000000e+00 0.08563258
#> Female 5.551115e-17 -0.08563258
#> Total 5.551115e-17 0.17126517
#>
#> $age_weight
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> 18-34 0.3514503 327 0.2766497 415.4143 0.3514503 0.07480056
#> 35-54 0.3486601 348 0.2944162 412.1162 0.3486601 0.05424384
#> 55+ 0.2998896 507 0.4289340 354.4695 0.2998896 -0.12904440
#> Total 1.0000000 1182 1.0000000 1182.0000 1.0000000 0.25808881
#> Resid. Disc. Orig. Disc.
#> 18-34 0.000000e+00 0.07480056
#> 35-54 0.000000e+00 0.05424384
#> 55+ 5.551115e-17 -0.12904440
#> Total 5.551115e-17 0.25808881
#>
#> $income_weight
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> 0-25K 0.17353726 178 0.15071973 204.9975 0.17353726 0.022817533
#> 25-35K 0.09405487 151 0.12785775 111.1059 0.09405487 -0.033802878
#> 35-50K 0.14064324 169 0.14309907 166.1402 0.14064324 -0.002455831
#> 50-75K 0.18922885 242 0.20491109 223.5338 0.18922885 -0.015682244
#> 75-100K 0.15252025 178 0.15071973 180.1704 0.15252025 0.001800522
#> 100-150K 0.16103245 182 0.15410669 190.2257 0.16103245 0.006925756
#> 150K+ 0.08898309 81 0.06858594 105.1147 0.08898309 0.020397142
#> Total 1.00000000 1181 1.00000000 1181.2882 1.00000000 0.103881906
#> Resid. Disc. Orig. Disc.
#> 0-25K -5.551115e-17 0.022817533
#> 25-35K 0.000000e+00 -0.033802878
#> 35-50K 2.775558e-17 -0.002455831
#> 50-75K 0.000000e+00 -0.015682244
#> 75-100K -2.775558e-17 0.001800522
#> 100-150K 0.000000e+00 0.006925756
#> 150K+ 0.000000e+00 0.020397142
#> Total 1.110223e-16 0.103881906
#>
#> $race_weight
#> Target Unweighted N Unweighted % Wtd N Wtd %
#> White 0.73559778 973 0.82387807 868.9530 0.73559778
#> Hispanic 0.17719770 114 0.09652837 209.3216 0.17719770
#> Not white or Hispanic 0.08720452 94 0.07959356 103.0137 0.08720452
#> Total 1.00000000 1181 1.00000000 1181.2882 1.00000000
#> Change in % Resid. Disc. Orig. Disc.
#> White -0.088280289 -1.110223e-16 -0.088280289
#> Hispanic 0.080669338 2.775558e-17 0.080669338
#> Not white or Hispanic 0.007610952 0.000000e+00 0.007610952
#> Total 0.176560579 1.387779e-16 0.176560579
southwick-associates/sastats documentation built on March 27, 2020, 9:39 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Overview
Rake weighting in R is fairly straightforward using
anesrake::anesrake(), for which sastats::rake_weight() is a thin
wrapper. The key aspect of the process is getting the two input datasets
into the right format:
-
The target population dataset needs to be a named list (one element per demographic variable) of target distributions. The
weights::wpct()function helps with this. -
The survey dataset needs to be a data frame with a unique ID variable (e.g., Vrid), and one variable for each demographic measure.
Importantly, demographic variables must be stored either as factors or numerics and the category coding much match between the two datasets. You can also see production examples for B4W-19-10 and AS/HS: rakewt-ashs.
Example Data
For demonstration, package sastats includes survey and population datasets which include several demographic variables:
library(dplyr)
library(sastats)
data(svy, pop)
svy <- select(svy$person, -weight)
# survey to be weighted
glimpse(svy)
#> Observations: 1,252
#> Variables: 5
#> $ Vrid <chr> "98", "99", "100", "101", "102", "103", "105", "106",...
#> $ sex <fct> Female, Female, NA, Male, Female, Male, Male, Female,...
#> $ age_weight <fct> 35-54, 35-54, NA, 35-54, 35-54, 18-34, 18-34, 35-54, ...
#> $ income_weight <fct> 0-25K, 0-25K, NA, 25-35K, 50-75K, 25-35K, 50-75K, 35-...
#> $ race_weight <fct> White, Not white or Hispanic, NA, White, Not white or...
# target population (outdoor recreationists) from a genpop survey
glimpse(pop)
#> Observations: 877
#> Variables: 5
#> $ sex <fct> Female, Female, Female, Male, Female, Female, Female,...
#> $ age_weight <fct> 18-34, 18-34, 18-34, 18-34, 18-34, 35-54, 35-54, 35-5...
#> $ income_weight <fct> 75-100K, 35-50K, 0-25K, 150K+, 0-25K, 25-35K, 75-100K...
#> $ race_weight <fct> White, White, Hispanic, White, White, White, White, W...
#> $ stwt <dbl> 0.8927761, 0.8927761, 1.1913576, 1.1056643, 0.8073446...
Importantly, the demographic variables have been encoded in the same way between the 2 datasets (as factors in this case):
wtvars <- setdiff(names(svy), "Vrid")
sapply(wtvars, function(x) all(levels(svy[[x]]) == levels(pop[[x]])))
#> sex age_weight income_weight race_weight
#> TRUE TRUE TRUE TRUE
Population Distributions
The easiest way to see the format needed for the target population is to
run weights::wpct() on a demographic variable:
weights::wpct(svy$age_weight)
#> 18-34 35-54 55+
#> 0.2766497 0.2944162 0.4289340
For our example, we are defining the target population (outdoor
recreationists) using another general population survey dataset. In this
case, we need to ensure we use the stwt variable from that survey
since it was itself weighted:
weights::wpct(pop$age_weight, pop$stwt)
#> 18-34 35-54 55+
#> 0.3514503 0.3486601 0.2998896
The weights::wpct() function returns a vector, but we need a list of
vectors (one element per demographic variables of interest). A
straightforward method involves looping over the demographic variables
with sapply():
pop_target <- sapply(wtvars, function(x) weights::wpct(pop[[x]], pop[["stwt"]]))
pop_target
#> $sex
#> Male Female
#> 0.51747 0.48253
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.3514503 0.3486601 0.2998896
#>
#> $income_weight
#> 0-25K 25-35K 35-50K 50-75K 75-100K 100-150K 150K+
#> 0.17353726 0.09405487 0.14064324 0.18922885 0.15252025 0.16103245 0.08898309
#>
#> $race_weight
#> White Hispanic Not white or Hispanic
#> 0.73559778 0.17719770 0.08720452
If we weren’t using a reference population dataset, we would need to get the target distributions into a list in some other way. For example, by hand:
partial_target <- list(
"sex" = c("Male" = 0.517, "Female" = 0.483),
"age_weight" = c("18-34" = 0.351, "35-54" = 0.349, "55+" = 0.300)
# etc.
)
partial_target
#> $sex
#> Male Female
#> 0.517 0.483
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.351 0.349 0.300
Using weights::wpct() also of course provides a convenient method for
comparison with the survey dataset to be weighted:
sapply(wtvars, function(x) weights::wpct(svy[[x]]))
#> $sex
#> Male Female
#> 0.4318374 0.5681626
#>
#> $age_weight
#> 18-34 35-54 55+
#> 0.2766497 0.2944162 0.4289340
#>
#> $income_weight
#> 0-25K 25-35K 35-50K 50-75K 75-100K 100-150K 150K+
#> 0.15071973 0.12785775 0.14309907 0.20491109 0.15071973 0.15410669 0.06858594
#>
#> $race_weight
#> White Hispanic Not white or Hispanic
#> 0.82387807 0.09652837 0.07959356
Rake Weighting
We can now run the rake weighting procedure. By default, rake_weight()
returns a list of 2 elements: (1) the survey dataset with the weight
variable appended, and (2) the anesrake() return object, which
includes a bunch of useful summary statistics (including “design
effect”, which may be useful for estimating confidence intervals).
svy_wts <- rake_weight(svy, pop_target, "Vrid")
#> [1] "Raking converged in 21 iterations"
svy <- svy_wts$svy
summary(svy$weight)
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 0.3568 0.7408 0.9245 1.0000 1.1662 2.6893
wt_summary <- summary(svy_wts$wts)
deff <- wt_summary$general.design.effect
deff
#> [1] 1.186242
wt_summary
#> $convergence
#> [1] "Complete convergence was achieved after 21 iterations"
#>
#> $base.weights
#> [1] "No Base Weights Were Used"
#>
#> $raking.variables
#> [1] "sex" "age_weight" "income_weight" "race_weight"
#>
#> $weight.summary
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 0.3568 0.7408 0.9245 1.0000 1.1662 2.6893
#>
#> $selection.method
#> [1] "variable selection conducted using _pctlim_ - discrepancies selected using _total_."
#>
#> $general.design.effect
#> [1] 1.186242
#>
#> $sex
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> Male 0.51747 510 0.4318374 611.2812 0.51747 0.08563258
#> Female 0.48253 671 0.5681626 570.0070 0.48253 -0.08563258
#> Total 1.00000 1181 1.0000000 1181.2882 1.00000 0.17126517
#> Resid. Disc. Orig. Disc.
#> Male 0.000000e+00 0.08563258
#> Female 5.551115e-17 -0.08563258
#> Total 5.551115e-17 0.17126517
#>
#> $age_weight
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> 18-34 0.3514503 327 0.2766497 415.4143 0.3514503 0.07480056
#> 35-54 0.3486601 348 0.2944162 412.1162 0.3486601 0.05424384
#> 55+ 0.2998896 507 0.4289340 354.4695 0.2998896 -0.12904440
#> Total 1.0000000 1182 1.0000000 1182.0000 1.0000000 0.25808881
#> Resid. Disc. Orig. Disc.
#> 18-34 0.000000e+00 0.07480056
#> 35-54 0.000000e+00 0.05424384
#> 55+ 5.551115e-17 -0.12904440
#> Total 5.551115e-17 0.25808881
#>
#> $income_weight
#> Target Unweighted N Unweighted % Wtd N Wtd % Change in %
#> 0-25K 0.17353726 178 0.15071973 204.9975 0.17353726 0.022817533
#> 25-35K 0.09405487 151 0.12785775 111.1059 0.09405487 -0.033802878
#> 35-50K 0.14064324 169 0.14309907 166.1402 0.14064324 -0.002455831
#> 50-75K 0.18922885 242 0.20491109 223.5338 0.18922885 -0.015682244
#> 75-100K 0.15252025 178 0.15071973 180.1704 0.15252025 0.001800522
#> 100-150K 0.16103245 182 0.15410669 190.2257 0.16103245 0.006925756
#> 150K+ 0.08898309 81 0.06858594 105.1147 0.08898309 0.020397142
#> Total 1.00000000 1181 1.00000000 1181.2882 1.00000000 0.103881906
#> Resid. Disc. Orig. Disc.
#> 0-25K -5.551115e-17 0.022817533
#> 25-35K 0.000000e+00 -0.033802878
#> 35-50K 2.775558e-17 -0.002455831
#> 50-75K 0.000000e+00 -0.015682244
#> 75-100K -2.775558e-17 0.001800522
#> 100-150K 0.000000e+00 0.006925756
#> 150K+ 0.000000e+00 0.020397142
#> Total 1.110223e-16 0.103881906
#>
#> $race_weight
#> Target Unweighted N Unweighted % Wtd N Wtd %
#> White 0.73559778 973 0.82387807 868.9530 0.73559778
#> Hispanic 0.17719770 114 0.09652837 209.3216 0.17719770
#> Not white or Hispanic 0.08720452 94 0.07959356 103.0137 0.08720452
#> Total 1.00000000 1181 1.00000000 1181.2882 1.00000000
#> Change in % Resid. Disc. Orig. Disc.
#> White -0.088280289 -1.110223e-16 -0.088280289
#> Hispanic 0.080669338 2.775558e-17 0.080669338
#> Not white or Hispanic 0.007610952 0.000000e+00 0.007610952
#> Total 0.176560579 1.387779e-16 0.176560579
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.