SampleSelectR is an R package developed by RTI International to support the design and implementation of common survey sampling methods. It is designed to make sample design and selection reproducible, efficient, and transparent for survey statisticians and researchers.
Currently supported sampling methods include:
Simple Random Sampling (SRS)
Systematic Sampling (sys)
Systematic Probability Proportional to Size (PPS) Sampling
Sequential PPS Sampling (i.e., Chromy’s method)
Allocation methods include:
Proportional
Power
Neyman
Optimal
Installation
You can install the development version of SampleSelectR from GitHub with either the pak or devtools package:
# install.packages("pak")
pak::pak("RTIInternational/SampleSelectR")
# install.packages("devtools")
devtools::install_github("RTIInternational/SampleSelectR")Then, you can load the package.
Included Data
SampleSelectR comes with three example datasets for illustration of realistic sample frames including:
-
ipedswhich contains a list of postsecondary educational institutions in the United States from the Integrated Postsecondary Education Data System (IPEDS) with a select set of characteristics for operating institutions in Fall 20231 -
county_2023which contains a list of the counties in the United States based on data from the 2019-2023 American Community Survey with a select set of characteristics2 -
puma_2023which contains a list of the Public Use Microdata Areas (PUMAs) in the United States based on data from the 2019-2023 American Community Survey with a select set of characteristics3
summary(ipeds)
#> UNITID INSTNM STABBR FIPS
#> Min. :100654 Length:5914 Length:5914 Min. : 1.00
#> 1st Qu.:169277 Class :character Class :character 1st Qu.:13.00
#> Median :219493 Mode :character Mode :character Median :29.00
#> Mean :284664 Mean :29.19
#> 3rd Qu.:446954 3rd Qu.:42.00
#> Max. :499723 Max. :78.00
#>
#> OBEREG
#> Southeast (AL, AR, FL, GA, KY, LA, MS, NC, SC, TN, VA, WV):1467
#> Mid East (DE, DC, MD, NJ, NY, PA) : 971
#> Far West (AK, CA, HI, NV, OR, WA) : 870
#> Great Lakes (IL, IN, MI, OH, WI) : 851
#> Southwest (AZ, NM, OK, TX) : 633
#> Plains (IA, KS, MN, MO, NE, ND, SD) : 468
#> (Other) : 654
#> ICLEVEL
#> Four or more years :2747
#> At least 2 but less than 4 years :1512
#> Less than 2 years (below associate):1655
#>
#>
#>
#>
#> SECTOR LOCALE
#> Private not-for-profit, 4-year or above:1608 Suburb: Large:1442
#> Private for-profit, less-than 2-year :1373 City: Large :1392
#> Public, 2-year : 868 City: Small : 778
#> Public, 4-year or above : 820 City: Midsize: 713
#> Private for-profit, 2-year : 517 Town: Distant: 396
#> Private for-profit, 4-year or above : 319 Rural: Fringe: 298
#> (Other) : 409 (Other) : 895
#> DEGGRANT
#> Degree-granting :3995
#> Nondegree-granting, primarily postsecondary:1919
#>
#>
#>
#>
#>
#> HLOFFER ENRTOT
#> At least 1, but less than 2 academic yrs:1514 Min. : 1
#> Doctor's degree :1232 1st Qu.: 127
#> Associate's degree :1022 Median : 588
#> Bachelor's degree : 677 Mean : 3332
#> Master's degree : 672 3rd Qu.: 2849
#> At least 2, but less than 4 academic yrs: 490 Max. :185015
#> (Other) : 307
#> EFUG EFUG1ST EFUGFT EFGRAD
#> Min. : 0.00 Min. : 0 Min. : 0 Min. : 0.0
#> 1st Qu.: 98.25 1st Qu.: 21 1st Qu.: 71 1st Qu.: 0.0
#> Median : 472.00 Median : 105 Median : 321 Median : 0.0
#> Mean : 2777.88 Mean : 508 Mean : 1706 Mean : 553.7
#> 3rd Qu.: 2455.25 3rd Qu.: 478 3rd Qu.: 1426 3rd Qu.: 120.0
#> Max. :159653.00 Max. :15343 Max. :135822 Max. :50245.0
#>
#> EFGRADFT
#> Min. : 0.00
#> 1st Qu.: 0.00
#> Median : 0.00
#> Mean : 328.37
#> 3rd Qu.: 56.75
#> Max. :49193.00
#>
summary(county_2023)
#> GEOID Name State Region
#> Length:3144 Length:3144 Length:3144 Northeast: 218
#> Class :character Class :character Class :character Midwest :1055
#> Mode :character Mode :character Mode :character South :1422
#> West : 449
#>
#>
#>
#> Division Pop_Tot Pop_Pct_White_NH Pop_Pct_Black_NH
#> West North Central:618 Min. : 43 Min. : 2.153 Min. : 0.0000
#> South Atlantic :588 1st Qu.: 10794 1st Qu.: 62.608 1st Qu.: 0.6574
#> West South Central:470 Median : 25813 Median : 81.103 Median : 2.0825
#> East North Central:437 Mean : 105721 Mean : 74.298 Mean : 8.6559
#> East South Central:364 3rd Qu.: 68379 3rd Qu.: 90.618 3rd Qu.: 9.5543
#> Mountain :281 Max. :9848406 Max. :100.000 Max. :86.0575
#> (Other) :386
#> Pop_Pct_AIAN_NH Pop_Pct_Asian_NH Pop_Pct_NHPI_NH Pop_Pct_Other_NH
#> Min. : 0.0000 Min. : 0.0000 Min. : 0.000000 Min. : 0.000
#> 1st Qu.: 0.0618 1st Qu.: 0.2780 1st Qu.: 0.000000 1st Qu.: 2.531
#> Median : 0.1546 Median : 0.6098 Median : 0.005365 Median : 3.478
#> Mean : 1.6596 Mean : 1.4257 Mean : 0.096733 Mean : 3.798
#> 3rd Qu.: 0.4637 3rd Qu.: 1.3405 3rd Qu.: 0.059577 3rd Qu.: 4.495
#> Max. :91.3520 Max. :41.9978 Max. :18.604651 Max. :25.562
#>
#> Pop_Pct_Hispanic HU_Tot HU_Pct_Occupied HU_Pct_Vacant
#> Min. : 0.000 Min. : 47 Min. :22.68 Min. : 1.707
#> 1st Qu.: 2.533 1st Qu.: 5240 1st Qu.:78.61 1st Qu.: 9.393
#> Median : 4.881 Median : 12374 Median :85.28 Median :14.721
#> Mean :10.066 Mean : 45271 Mean :83.10 Mean :16.904
#> 3rd Qu.:10.809 3rd Qu.: 31710 3rd Qu.:90.61 3rd Qu.:21.388
#> Max. :97.306 Max. :3624084 Max. :98.29 Max. :77.316
#>
#> Pop_Pct_0004 Pop_Pct_0509 Pop_Pct_1014 Pop_Pct_2544
#> Min. : 0.000 Min. : 0.000 Min. : 0.000 Min. : 0.00
#> 1st Qu.: 4.900 1st Qu.: 5.200 1st Qu.: 5.700 1st Qu.:21.90
#> Median : 5.500 Median : 6.000 Median : 6.400 Median :23.60
#> Mean : 5.529 Mean : 5.997 Mean : 6.455 Mean :23.82
#> 3rd Qu.: 6.200 3rd Qu.: 6.700 3rd Qu.: 7.200 3rd Qu.:25.60
#> Max. :20.600 Max. :21.000 Max. :13.300 Max. :58.20
#>
#> Pop_Pct_4564 Pop_Pct_6574 Pop_Pct_75plus Pop_Pct_1517
#> Min. : 8.90 Min. : 2.10 Min. : 0.700 Min. : 0.000
#> 1st Qu.:24.40 1st Qu.:10.10 1st Qu.: 6.700 1st Qu.: 3.600
#> Median :26.00 Median :11.60 Median : 8.000 Median : 4.000
#> Mean :25.75 Mean :11.87 Mean : 8.182 Mean : 3.955
#> 3rd Qu.:27.40 3rd Qu.:13.10 3rd Qu.: 9.400 3rd Qu.: 4.300
#> Max. :46.90 Max. :38.50 Max. :46.300 Max. :17.900
#>
#> Pop_Pct_1824
#> Min. : 0.000
#> 1st Qu.: 6.900
#> Median : 7.800
#> Mean : 8.453
#> 3rd Qu.: 8.900
#> Max. :49.100
#>
summary(puma_2023)
#> GEOID Name State Region
#> Length:2462 Length:2462 Length:2462 Northeast:423
#> Class :character Class :character Class :character Midwest :506
#> Mode :character Mode :character Mode :character South :952
#> West :581
#>
#>
#>
#> Division Pop_Tot Pop_Pct_White_NH Pop_Pct_Black_NH
#> South Atlantic :496 Min. : 94457 Min. : 0.6772 Min. : 0.06043
#> Pacific :392 1st Qu.:114269 1st Qu.:41.1324 1st Qu.: 2.25999
#> East North Central:340 Median :128683 Median :62.3550 Median : 5.82201
#> Middle Atlantic :313 Mean :135007 Mean :58.1157 Mean :12.06644
#> West South Central:303 3rd Qu.:151797 3rd Qu.:78.8115 3rd Qu.:15.45296
#> Mountain :189 Max. :254693 Max. :95.5992 Max. :91.54119
#> (Other) :429
#> Pop_Pct_AIAN_NH Pop_Pct_Asian_NH Pop_Pct_NHPI_NH Pop_Pct_Other_NH
#> Min. : 0.00000 Min. : 0.03401 Min. : 0.00000 Min. : 0.09704
#> 1st Qu.: 0.06733 1st Qu.: 1.11991 1st Qu.: 0.00978 1st Qu.: 3.22126
#> Median : 0.12409 Median : 2.67222 Median : 0.03879 Median : 4.07446
#> Mean : 0.54049 Mean : 5.75511 Mean : 0.16905 Mean : 4.35801
#> 3rd Qu.: 0.25500 3rd Qu.: 6.54233 3rd Qu.: 0.10762 3rd Qu.: 5.11640
#> Max. :71.05525 Max. :69.44064 Max. :23.33594 Max. :25.66375
#>
#> Pop_Pct_Hispanic HU_Tot HU_Pct_Occupied HU_Pct_Vacant
#> Min. : 0.7159 Min. : 25782 Min. :46.59 Min. : 1.123
#> 1st Qu.: 5.6869 1st Qu.: 47813 1st Qu.:87.58 1st Qu.: 4.995
#> Median :11.3890 Median : 55208 Median :92.36 Median : 7.638
#> Mean :18.9952 Mean : 57812 Mean :90.36 Mean : 9.639
#> 3rd Qu.:24.5716 3rd Qu.: 65673 3rd Qu.:95.00 3rd Qu.:12.419
#> Max. :96.0544 Max. :140746 Max. :98.88 Max. :53.406
#>
#> Pop_Pct_0004 Pop_Pct_0509 Pop_Pct_1014 Pop_Pct_2544
#> Min. : 1.80 Min. : 1.400 Min. : 1.30 Min. :11.70
#> 1st Qu.: 5.00 1st Qu.: 5.400 1st Qu.: 5.80 1st Qu.:23.90
#> Median : 5.60 Median : 6.000 Median : 6.50 Median :26.10
#> Mean : 5.69 Mean : 6.035 Mean : 6.47 Mean :26.76
#> 3rd Qu.: 6.30 3rd Qu.: 6.700 3rd Qu.: 7.20 3rd Qu.:28.70
#> Max. :16.70 Max. :14.400 Max. :12.50 Max. :55.40
#>
#> Pop_Pct_4564 Pop_Pct_6574 Pop_Pct_75plus Pop_Pct_1517
#> Min. : 9.60 Min. : 3.60 Min. : 1.700 Min. :0.60
#> 1st Qu.:23.50 1st Qu.: 8.30 1st Qu.: 5.300 1st Qu.:3.50
#> Median :25.40 Median : 9.90 Median : 6.700 Median :3.90
#> Mean :25.14 Mean :10.03 Mean : 6.837 Mean :3.92
#> 3rd Qu.:27.00 3rd Qu.:11.40 3rd Qu.: 8.100 3rd Qu.:4.30
#> Max. :34.00 Max. :30.80 Max. :27.000 Max. :6.70
#>
#> Pop_Pct_1824
#> Min. : 2.900
#> 1st Qu.: 7.400
#> Median : 8.300
#> Mean : 9.118
#> 3rd Qu.: 9.700
#> Max. :39.900
#> Example Workflow
The following are some examples of allocation and sampling on the IPEDS data. The primary functions in workflows are allocate() to allocate sample sizes though you can get a sample allocation in another manner, if you prefer and select_sample() where you specify your frame, sample sizes, sampling method, strata (if relevant), measure of size (if relevant), and sort method (if relevant).
Allocation
First, you must have a summary of your frame. In this case, we will be specifying a sample size of 500 to be proportionally allocated across region.
set.seed(8675309)
ipeds_summary <- ipeds |>
tidytable::summarize(
N = tidytable::n(),
.by = "OBEREG"
) |>
tidytable::ungroup()
ipeds_summary
#> # A tidytable: 10 × 2
#> OBEREG N
#> <fct> <int>
#> 1 U.S. Service schools 7
#> 2 New England (CT, ME, MA, NH, RI, VT) 299
#> 3 Mid East (DE, DC, MD, NJ, NY, PA) 971
#> 4 Great Lakes (IL, IN, MI, OH, WI) 851
#> 5 Plains (IA, KS, MN, MO, NE, ND, SD) 468
#> 6 Southeast (AL, AR, FL, GA, KY, LA, MS, NC, SC, TN, VA, WV) 1467
#> 7 Southwest (AZ, NM, OK, TX) 633
#> 8 Rocky Mountains (CO, ID, MT, UT, WY) 216
#> 9 Far West (AK, CA, HI, NV, OR, WA) 870
#> 10 Other U.S. jurisdictions (AS, FM, GU, MH, MP, PR, PW, VI) 132
ipeds_alloc <- ipeds_summary |>
tidytable::mutate(
sample_size = allocate("proportional", N.h = N, n.samp = 500)
)
#> Sample allocation of 500 using proportional with the relevant inputs:
#> N.h = 7, 299, 971, 851, 468, 1467, 633, 216, 870, 132
#>
#> Output:
#> 2, 25, 82, 73, 39, 124, 53, 18, 73, 11
ipeds_alloc
#> # A tidytable: 10 × 3
#> OBEREG N sample_size
#> <fct> <int> <int>
#> 1 U.S. Service schools 7 2
#> 2 New England (CT, ME, MA, NH, RI, VT) 299 25
#> 3 Mid East (DE, DC, MD, NJ, NY, PA) 971 82
#> 4 Great Lakes (IL, IN, MI, OH, WI) 851 73
#> 5 Plains (IA, KS, MN, MO, NE, ND, SD) 468 39
#> 6 Southeast (AL, AR, FL, GA, KY, LA, MS, NC, SC, TN, VA, WV) 1467 124
#> 7 Southwest (AZ, NM, OK, TX) 633 53
#> 8 Rocky Mountains (CO, ID, MT, UT, WY) 216 18
#> 9 Far West (AK, CA, HI, NV, OR, WA) 870 73
#> 10 Other U.S. jurisdictions (AS, FM, GU, MH, MP, PR, PW, VI) 132 11Sampling
Then, we are able to select a stratified sample and demonstrate this using a SRS and systematic PPS. With the systematic and sequential sampling methods, you can specify sort variables and the sort method. Often, you want to be able to re-run sampling code and get the same sample. To do this, we also set a seed.
ipeds_srs <- ipeds |>
select_sample(
method = "srs",
n = tidytable::select(ipeds_alloc, OBEREG, sample_size),
strata = "OBEREG"
)
#> Stratum: OBEREG = Southeast (AL, AR, FL, GA, KY, LA, MS, NC, SC, TN, VA, WV)
#> --Frame size: 1467
#> --Sample size: 124
#> Stratum: OBEREG = Far West (AK, CA, HI, NV, OR, WA)
#> --Frame size: 870
#> --Sample size: 73
#> Stratum: OBEREG = Southwest (AZ, NM, OK, TX)
#> --Frame size: 633
#> --Sample size: 53
#> Stratum: OBEREG = U.S. Service schools
#> --Frame size: 7
#> --Sample size: 2
#> Stratum: OBEREG = Plains (IA, KS, MN, MO, NE, ND, SD)
#> --Frame size: 468
#> --Sample size: 39
#> Stratum: OBEREG = Rocky Mountains (CO, ID, MT, UT, WY)
#> --Frame size: 216
#> --Sample size: 18
#> Stratum: OBEREG = New England (CT, ME, MA, NH, RI, VT)
#> --Frame size: 299
#> --Sample size: 25
#> Stratum: OBEREG = Mid East (DE, DC, MD, NJ, NY, PA)
#> --Frame size: 971
#> --Sample size: 82
#> Stratum: OBEREG = Great Lakes (IL, IN, MI, OH, WI)
#> --Frame size: 851
#> --Sample size: 73
#> Stratum: OBEREG = Other U.S. jurisdictions (AS, FM, GU, MH, MP, PR, PW, VI)
#> --Frame size: 132
#> --Sample size: 11
ipeds_srs
#> # A tidytable: 500 × 18
#> OBEREG UNITID INSTNM STABBR FIPS ICLEVEL SECTOR LOCALE DEGGRANT HLOFFER
#> <fct> <dbl> <chr> <chr> <dbl> <fct> <fct> <fct> <fct> <fct>
#> 1 Southeast … 101028 Chatt… AL 1 At lea… Publi… Subur… Degree-… At lea…
#> 2 Southeast … 101602 Lurle… AL 1 At lea… Publi… Rural… Degree-… Associ…
#> 3 Southeast … 101879 Unive… AL 1 Four o… Publi… City:… Degree-… Doctor…
#> 4 Southeast … 102049 Samfo… AL 1 Four o… Priva… Subur… Degree-… Doctor…
#> 5 Southeast … 102313 H Cou… AL 1 At lea… Publi… City:… Degree-… Associ…
#> 6 Southeast … 107318 Arkan… AR 5 At lea… Publi… Subur… Degree-… Associ…
#> 7 Southeast … 107460 North… AR 5 At lea… Publi… Town:… Degree-… At lea…
#> 8 Southeast … 132842 Albiz… FL 12 Four o… Priva… City:… Degree-… Doctor…
#> 9 Southeast … 133386 Dayto… FL 12 Four o… Publi… City:… Degree-… Bachel…
#> 10 Southeast … 133881 Flori… FL 12 Four o… Priva… City:… Degree-… Doctor…
#> # ℹ 490 more rows
#> # ℹ 8 more variables: ENRTOT <dbl>, EFUG <dbl>, EFUG1ST <dbl>, EFUGFT <dbl>,
#> # EFGRAD <dbl>, EFGRADFT <dbl>, SelectionProbability <dbl>,
#> # SamplingWeight <dbl>
ipeds_pps <- ipeds |>
select_sample(
method = "sys_pps",
n = tidytable::select(ipeds_alloc, OBEREG, sample_size),
strata = "OBEREG",
mos = "ENRTOT",
sort_vars = "SECTOR",
sort_method = "serpentine"
)
#> Stratum: OBEREG = Southeast (AL, AR, FL, GA, KY, LA, MS, NC, SC, TN, VA, WV)
#> --Frame size: 1467
#> --Sample size: 124
#> --Sampling interval (k): 37489.19
#> --Random start (r): 17174.08
#> Stratum: OBEREG = Far West (AK, CA, HI, NV, OR, WA)
#> --Frame size: 870
#> --Sample size: 73
#> --Sampling interval (k): 46640.48
#> --Random start (r): 23535.96
#> Stratum: OBEREG = Southwest (AZ, NM, OK, TX)
#> --Frame size: 633
#> --Sample size: 53
#> --Sampling interval (k): 49474.49
#> --Random start (r): 33072.34
#> Stratum: OBEREG = U.S. Service schools
#> --Frame size: 7
#> --Sample size: 2
#> --Sampling interval (k): 9372
#> --Random start (r): 2483.683
#> Stratum: OBEREG = Plains (IA, KS, MN, MO, NE, ND, SD)
#> --Frame size: 468
#> --Sample size: 39
#> --Sampling interval (k): 35210.1
#> --Random start (r): 301.2639
#> Stratum: OBEREG = Rocky Mountains (CO, ID, MT, UT, WY)
#> --Frame size: 216
#> --Sample size: 18
#> --Sampling interval (k): 57314.5
#> --Random start (r): 10947.04
#> Stratum: OBEREG = New England (CT, ME, MA, NH, RI, VT)
#> --Frame size: 299
#> --Sample size: 25
#> --Sampling interval (k): 44218.08
#> --Random start (r): 6258.501
#> Stratum: OBEREG = Mid East (DE, DC, MD, NJ, NY, PA)
#> --Frame size: 971
#> --Sample size: 82
#> --Sampling interval (k): 33082.89
#> --Random start (r): 11354.24
#> Stratum: OBEREG = Great Lakes (IL, IN, MI, OH, WI)
#> --Frame size: 851
#> --Sample size: 73
#> --Sampling interval (k): 35699.49
#> --Random start (r): 3529.973
#> Stratum: OBEREG = Other U.S. jurisdictions (AS, FM, GU, MH, MP, PR, PW, VI)
#> --Frame size: 132
#> --Sample size: 11
#> --Sampling interval (k): 16320.64
#> --Random start (r): 11469.85
ipeds_pps
#> # A tidytable: 474 × 19
#> OBEREG UNITID INSTNM STABBR FIPS ICLEVEL SECTOR LOCALE DEGGRANT HLOFFER
#> <fct> <dbl> <chr> <chr> <dbl> <fct> <fct> <fct> <fct> <fct>
#> 1 Southeast … 100663 Unive… AL 1 Four o… Publi… City:… Degree-… Doctor…
#> 2 Southeast … 100751 The U… AL 1 Four o… Publi… City:… Degree-… Doctor…
#> 3 Southeast … 100858 Aubur… AL 1 Four o… Publi… City:… Degree-… Doctor…
#> 4 Southeast … 101480 Jacks… AL 1 Four o… Publi… Subur… Degree-… Doctor…
#> 5 Southeast … 102368 Troy … AL 1 Four o… Publi… Town:… Degree-… Doctor…
#> 6 Southeast … 106397 Unive… AR 5 Four o… Publi… City:… Degree-… Doctor…
#> 7 Southeast … 106467 Arkan… AR 5 Four o… Publi… Town:… Degree-… Doctor…
#> 8 Southeast … 132693 Easte… FL 12 Four o… Publi… City:… Degree-… Bachel…
#> 9 Southeast … 132851 Colle… FL 12 Four o… Publi… City:… Degree-… Bachel…
#> 10 Southeast … 132903 Unive… FL 12 Four o… Publi… Subur… Degree-… Doctor…
#> # ℹ 464 more rows
#> # ℹ 9 more variables: ENRTOT <dbl>, EFUG <dbl>, EFUG1ST <dbl>, EFUGFT <dbl>,
#> # EFGRAD <dbl>, EFGRADFT <dbl>, SamplingWeight <dbl>, NumberHits <int>,
#> # ExpectedHits <dbl>
sum(ipeds_pps$NumberHits)
#> [1] 500Contributing
We’d love your help to make SampleSelectR better! Here’s the workflow we use:
Start from the dev branch
Always update your local copy first (git pull origin dev).
Create a new branch for your work
Name it clearly, e.g. fix-readme, add-tests, or feature-allocate.
Make your changes
Update code, documentation, or tests.
If adding a function, include a short example and (if possible) a unit test.
Push your branch to GitHub and open a PR into dev.