Hi!
My data file has a column for the person-level weights (which have already been calculated for each respondent) and I was wondering how one goes about linking or "attaching" the person-level weights to the the variables of interest on Rstudio? I also have a bootstrap file, which I am assuming I would link afterwards...
Any help would be greatly appreciated as this is my first time using Rstudio and adding weights/bootstrapping.
@StatSteph your replies to some of the other posts that I have been reading have been super helpful and I was wondering if you might have some advice for me regarding my question posted?
I'm not sure what you mean by attaching. Do you mean merging? Do you have two data.frames with weights on one and responses on another? It would be very helpful to give more information and provide a reproducible example.
See FAQ: What's a reproducible example (`reprex`) and how do I create one? and FAQ: How to do a minimal reproducible example ( reprex ) for beginners
I have one file imported which contains the data. In that file, one of the columns is the person-level weights (PUMFWGHT) for each unique ID. I have tried using svredesign to merge the person-level weights to the variables of interest but have not had much luck. Not sure if this makes sense.... All my variables are categories (snapshot below)
ID PUMFWGHT Age Sex Ethic Identity
123, 25, 1, 2, 1
124, 10, 3, 1, 3
125, 14, 4 , 1, 2
126 , 23, 3 , 2 , 1
This is still not a reproducible example. You mentioned having a file with the bootstrap weights on it too. What does that look like? Please refer to the documentation on how to share your data (maybe using dput, for example) rather than how you pasted it here.
A very simple and very effective way to supply some data is to use the dput() command.
dput(mydata)
and then simply copy the output and paste it here. If you have a very large data set then a sample should be fine. To supply us with 100 rows of your data set do
dput(head(mydata , 100))
where mydata is the name of your dataframe or tibble.
Thank you both!
Is this better?
dput(APS[1:20, c(1:7)])
structure(list(PUMFID = c(30000L, 30001L, 30003L, 30005L, 30006L,
30007L, 30009L, 30011L, 30012L, 30013L, 30014L, 30015L, 30016L,
30017L, 30019L, 30020L, 30021L, 30022L, 30023L, 30026L), PUMFWGHT = c(25.3256,
10.257, 23.1295, 10.759, 42.7018, 10.9332, 25.9796, 25.5107,
43.9171, 21.1154, 60.9822, 19.3566, 48.5176, 51.538, 23.9373,
64.5496, 169.0191, 11.9627, 33.3693, 84.0422), GEO_PC = c(3L,
2L, 1L, 4L, 1L, 4L, 1L, 1L, 1L, 2L, 3L, 2L, 1L, 3L, 3L, 3L, 1L,
2L, 1L, 3L), GEO_INU = c(3L, 3L, 3L, 1L, 3L, 2L, 3L, 3L, 3L,
3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), AGE_YRSG = c(2L,
2L, 2L, 2L, 6L, 3L, 1L, 2L, 6L, 3L, 6L, 2L, 2L, 5L, 2L, 5L, 1L,
4L, 6L, 1L), SEX = c(1L, 1L, 2L, 2L, 1L, 1L, 1L, 2L, 2L, 1L,
1L, 1L, 2L, 2L, 2L, 1L, 2L, 1L, 2L, 1L), PROXY = c(2L, 2L, 2L,
1L, 2L, 2L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 2L, 2L,
1L)), row.names = c(NA, 20L), class = "data.frame")
This is the bootstrap file
dput(aps_2017_pumf_bsw_eng[1:10, c(1:5)])
structure(list(PUMFID = structure(c(30000, 30001, 30003, 30005,
30006, 30007, 30009, 30011, 30012, 30013), label = "Public Use Microdata file identification number"),
PUMFWGHT = structure(c(25.3256, 10.257, 23.1295, 10.759,
42.7018, 10.9332, 25.9796, 25.5107, 43.9171, 21.1154), label = "Survey weight of a person, for Public Use Microdata file"),
WRPP0001 = c(22.7792, 10.5725, 21.5607, 10.2069, 37.9878,
10.7163, 22.7959, 24.6361, 51.068, 24.7117), WRPP0002 = c(16.9974,
10.374, 16.6935, 9.8463, 56.3002, 13.0115, 35.7703, 37.7712,
59.1396, 13.2218), WRPP0003 = c(22.6349, 9.9881, 19.8387,
10.7217, 47.0602, 9.9361, 28.2887, 24.6291, 46.5755, 19.2514
)), row.names = c(NA, -10L), class = c("tbl_df", "tbl", "data.frame"
), label = "APS_2017_PUMF_BSW_ENG")
Given
dta <- data.frame(
PUMFID = c(
30000, 30001, 30003, 30005, 30006, 30007, 30009, 30011, 30012, 30013,
30014, 30015, 30016,
30017, 30019, 30020, 30021, 30022, 30023, 30026
),
PUMFWGHT = c(
25.3256, 10.257, 23.1295, 10.759, 42.7018, 10.9332, 25.9796,
25.5107, 43.9171, 21.1154, 60.9822, 19.3566, 48.5176, 51.538,
23.9373, 64.5496, 169.0191, 11.9627, 33.3693, 84.0422
),
GEO_PC = c(3, 2, 1, 4, 1, 4, 1, 1, 1, 2, 3, 2, 1, 3, 3, 3, 1, 2, 1, 3),
GEO_INU = c(3, 3, 3, 1, 3, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3),
AGE_YRSG = c(2, 2, 2, 2, 6, 3, 1, 2, 6, 3, 6, 2, 2, 5, 2, 5, 1, 4, 6, 1),
SEX = c(1, 1, 2, 2, 1, 1, 1, 2, 2, 1, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1),
PROXY = c(2, 2, 2, 1, 2, 2, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 1)
)
booted <- data.frame(
PUMFID = c(
30000, 30001, 30003, 30005, 30006, 30007, 30009, 30011, 30012, 30013
),
PUMFWGHT = c(
25.3256, 10.257, 23.1295, 10.759, 42.7018, 10.9332, 25.9796, 25.5107,
43.9171, 21.1154
),
WRPP0001 = c(
22.7792, 10.5725, 21.5607, 10.2069, 37.9878, 10.7163, 22.7959,
24.6361, 51.068, 24.7117
),
WRPP0002 = c(
16.9974, 10.374, 16.6935, 9.8463, 56.3002, 13.0115, 35.7703, 37.7712,
59.1396, 13.2218
),
WRPP0003 = c(
22.6349, 9.9881, 19.8387, 10.7217, 47.0602, 9.9361, 28.2887, 24.6291,
46.5755, 19.2514
)
)
dta
#> PUMFID PUMFWGHT GEO_PC GEO_INU AGE_YRSG SEX PROXY
#> 1 30000 25.3256 3 3 2 1 2
#> 2 30001 10.2570 2 3 2 1 2
#> 3 30003 23.1295 1 3 2 2 2
#> 4 30005 10.7590 4 1 2 2 1
#> 5 30006 42.7018 1 3 6 1 2
#> 6 30007 10.9332 4 2 3 1 2
#> 7 30009 25.9796 1 3 1 1 1
#> 8 30011 25.5107 1 3 2 2 2
#> 9 30012 43.9171 1 3 6 2 2
#> 10 30013 21.1154 2 3 3 1 2
#> 11 30014 60.9822 3 3 6 1 2
#> 12 30015 19.3566 2 3 2 1 2
#> 13 30016 48.5176 1 3 2 2 2
#> 14 30017 51.5380 3 3 5 2 2
#> 15 30019 23.9373 3 3 2 2 2
#> 16 30020 64.5496 3 3 5 1 2
#> 17 30021 169.0191 1 3 1 2 1
#> 18 30022 11.9627 2 3 4 1 2
#> 19 30023 33.3693 1 3 6 2 2
#> 20 30026 84.0422 3 3 1 1 1
booted
#> PUMFID PUMFWGHT WRPP0001 WRPP0002 WRPP0003
#> 1 30000 25.3256 22.7792 16.9974 22.6349
#> 2 30001 10.2570 10.5725 10.3740 9.9881
#> 3 30003 23.1295 21.5607 16.6935 19.8387
#> 4 30005 10.7590 10.2069 9.8463 10.7217
#> 5 30006 42.7018 37.9878 56.3002 47.0602
#> 6 30007 10.9332 10.7163 13.0115 9.9361
#> 7 30009 25.9796 22.7959 35.7703 28.2887
#> 8 30011 25.5107 24.6361 37.7712 24.6291
#> 9 30012 43.9171 51.0680 59.1396 46.5755
#> 10 30013 21.1154 24.7117 13.2218 19.2514
the next step is to put a little formality on the question.
Every R problem can be thought of with advantage as the interaction of three objects— an existing object, x , a desired object,y , and a function, f, that will return a value of y given x as an argument. In other words, school algebra— f(x) = y. Any of the objects can be composites.
In this case, we have two objects, dta and booted. To find, is an f
merge the person-level weights to the variables of interest
Unclear is person-level weights and variables of interest
Using the f(x) = y paradigm, what is x and y? From that f can be suggested.
Here's how I would set this up. Note you only included 10 records from the bootstrap data so I first set the data to 10 rows but you wouldn't need this with all the data.
library(tidyverse)
library(survey)
#> Loading required package: grid
#> Loading required package: Matrix
#>
#> Attaching package: 'Matrix'
#> The following objects are masked from 'package:tidyr':
#>
#> expand, pack, unpack
#> Loading required package: survival
#>
#> Attaching package: 'survey'
#> The following object is masked from 'package:graphics':
#>
#> dotchart
library(srvyr)
#>
#> Attaching package: 'srvyr'
#> The following object is masked from 'package:stats':
#>
#> filter
APS <- data.frame(
PUMFID = c(
30000, 30001, 30003, 30005, 30006, 30007, 30009, 30011, 30012, 30013,
30014, 30015, 30016,
30017, 30019, 30020, 30021, 30022, 30023, 30026
),
PUMFWGHT = c(
25.3256, 10.257, 23.1295, 10.759, 42.7018, 10.9332, 25.9796,
25.5107, 43.9171, 21.1154, 60.9822, 19.3566, 48.5176, 51.538,
23.9373, 64.5496, 169.0191, 11.9627, 33.3693, 84.0422
),
GEO_PC = c(3, 2, 1, 4, 1, 4, 1, 1, 1, 2, 3, 2, 1, 3, 3, 3, 1, 2, 1, 3),
GEO_INU = c(3, 3, 3, 1, 3, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3),
AGE_YRSG = c(2, 2, 2, 2, 6, 3, 1, 2, 6, 3, 6, 2, 2, 5, 2, 5, 1, 4, 6, 1),
SEX = c(1, 1, 2, 2, 1, 1, 1, 2, 2, 1, 1, 1, 2, 2, 2, 1, 2, 1, 2, 1),
PROXY = c(2, 2, 2, 1, 2, 2, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 2, 2, 1)
)
aps_2017_pumf_bsw_eng <- data.frame(
PUMFID = c(
30000, 30001, 30003, 30005, 30006, 30007, 30009, 30011, 30012, 30013
),
PUMFWGHT = c(
25.3256, 10.257, 23.1295, 10.759, 42.7018, 10.9332, 25.9796, 25.5107,
43.9171, 21.1154
),
WRPP0001 = c(
22.7792, 10.5725, 21.5607, 10.2069, 37.9878, 10.7163, 22.7959,
24.6361, 51.068, 24.7117
),
WRPP0002 = c(
16.9974, 10.374, 16.6935, 9.8463, 56.3002, 13.0115, 35.7703, 37.7712,
59.1396, 13.2218
),
WRPP0003 = c(
22.6349, 9.9881, 19.8387, 10.7217, 47.0602, 9.9361, 28.2887, 24.6291,
46.5755, 19.2514
)
)
# merge the data with the replicate weights
dat_wrep <- APS %>%
slice(1:10) %>% # for your example, you only included 10 rows of bootstrap weights so need this to be 1:1 merge
left_join(select(aps_2017_pumf_bsw_eng, -PUMFWGHT), by="PUMFID")
my_design <- dat_wrep %>%
as_survey_rep(weights=PUMFWGHT, type="bootstrap", repweights=starts_with("WRPP"))
# Example analysis
# See https://cran.r-project.org/web/packages/srvyr/vignettes/srvyr-vs-survey.html for more
my_design %>%
survey_count(SEX)
#> # A tibble: 2 x 3
#> SEX n n_se
#> <dbl> <dbl> <dbl>
#> 1 1 136. 8.06
#> 2 2 103. 11.2
Created on 2021-03-05 by the reprex package (v1.0.0)
Thank you! This makes sense and I tried it but received the following error (I also uploaded a picture of my environment, not sure if this helps in anyway):
booted <- aps_2017_pumf_bsw_eng
dat_wwghts <- APS %>%
slice(1:1) %>%
left_join(select(booted, -PUMFWGHT), by="PUMFID")
APSwboot <- dat_wwghts
as_survey_rep(weights=PUMFWGHT, type="bootstrap", repweights= starts_with("WRPP"))
Error in as_survey_rep(weights = PUMFWGHT, type = "bootstrap", repweights = starts_with("WRPP")) : **
** object 'PUMFWGHT' not found
This error indicates that the variable PUMFWGHT isn't on your dataset. It was there on the example so I can't really help.
Also don't use slice(1:1) as that just selects the first row of your data. Remove the line with the slice function which was just used for the example.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.
If you have a query related to it or one of the replies, start a new topic and refer back with a link.