Scribd
Upload
8 views

Transversales Script

1. The document describes strategies for stratified two-stage cluster sampling to estimate population totals. 2. It shows how to sample primary sampling units (PSUs) in the first stage and households in the second stage. Weights are calculated for each unit. 3. Estimates, standard errors, and other statistics are then calculated from the sample data using the weights. Histograms and boxplots are presented to examine the weight distributions.

Uploaded by

marco antonio diaz barragan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
8 views

Transversales Script

1. The document describes strategies for stratified two-stage cluster sampling to estimate population totals. 2. It shows how to sample primary sampling units (PSUs) in the first stage and households in the second stage. Weights are calculated for each unit. 3. Estimates, standard errors, and other statistics are then calculated from the sample data using the weights. Histograms and boxplots are presented to examine the weight distributions.

Uploaded by

marco antonio diaz barragan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Estrategias transversales

Muestreo en dos etapas estratificado

library(dplyr)
library(TeachingSampling)
data('BigCity')
FrameI <- BigCity %>% group_by(PSU) %>%
summarise(Stratum = unique(Stratum),
Persons = n(),
Income = sum(Income),
Expenditure = sum(Expenditure))
attach(FrameI)
head(FrameI)

## # A tibble: 6 x 5
## PSU Stratum Persons Income Expenditure
## <chr> <chr> <int> <dbl> <dbl>
## 1 PSU0001 idStrt001 118 70912. 44232.
## 2 PSU0002 idStrt001 136 68887. 38382.
## 3 PSU0003 idStrt001 96 37213. 19495.
## 4 PSU0004 idStrt001 88 36926. 24031.
## 5 PSU0005 idStrt001 110 57494. 31142.
## 6 PSU0006 idStrt001 116 75272. 43473.
sizes = FrameI %>% group_by(Stratum) %>%
summarise(NIh = n(),
nIh = 2,
dI = NIh/nIh)
NIh <- sizes$NIh
nIh <- sizes$nIh
head(sizes)

## # A tibble: 6 x 4
## Stratum NIh nIh dI
## <chr> <int> <dbl> <dbl>
## 1 idStrt001 9 2 4.5
## 2 idStrt002 11 2 5.5
## 3 idStrt003 7 2 3.5
## 4 idStrt004 13 2 6.5
## 5 idStrt005 11 2 5.5
## 6 idStrt006 5 2 2.5
samI <- S.STSI(Stratum, NIh, nIh)
UI <- levels(as.factor(FrameI$PSU))
sampleI <- UI[samI]
FrameII <- left_join(sizes, BigCity[which(BigCity$PSU %in% sampleI), ], by = "Stratum")
attach(FrameII)
head(FrameII)

1
## # A tibble: 6 x 15
## Stratum NIh nIh dI HHID PersonID PSU Zone Sex Age MaritalST
## <chr> <int> <dbl> <dbl> <chr> <chr> <chr> <chr> <chr> <int> <fct>
## 1 idStrt~ 9 2 4.5 idHH~ idPer01 PSU0~ Rural Male 57 Married
## 2 idStrt~ 9 2 4.5 idHH~ idPer02 PSU0~ Rural Fema~ 48 Married
## 3 idStrt~ 9 2 4.5 idHH~ idPer03 PSU0~ Rural Male 20 Single
## 4 idStrt~ 9 2 4.5 idHH~ idPer04 PSU0~ Rural Fema~ 18 Single
## 5 idStrt~ 9 2 4.5 idHH~ idPer05 PSU0~ Rural Male 16 Single
## 6 idStrt~ 9 2 4.5 idHH~ idPer06 PSU0~ Rural Fema~ 0 <NA>
## # ... with 4 more variables: Income <dbl>, Expenditure <dbl>, Employment <fct>,
## # Poverty <fct>
HHdb <- FrameII %>%
group_by(PSU) %>%
summarise(Ni = length(unique(HHID)))
Ni <- as.numeric(HHdb$Ni)
ni <- ceiling(Ni * 0.1)
sum(ni)

## [1] 702
sam = S.SI(Ni[1], ni[1])
clusterII = FrameII[which(FrameII$PSU == sampleI[1]), ]
sam.HH <- data.frame(HHID = unique(clusterII$HHID)[sam])
clusterHH <- left_join(sam.HH, clusterII, by = "HHID")
clusterHH$dki <- Ni[1]/ni[1]
clusterHH$dk <- clusterHH$dI * clusterHH$dki
data = clusterHH
for (i in 2:length(Ni)) {
sam = S.SI(Ni[i], ni[i])
clusterII = FrameII[which(FrameII$PSU == sampleI[i]), ]
sam.HH <- data.frame(HHID = unique(clusterII$HHID)[sam])
clusterHH <- left_join(sam.HH, clusterII, by = "HHID")
clusterHH$dki <- Ni[i]/ni[i]
clusterHH$dk <- clusterHH$dI * clusterHH$dki
data1 = clusterHH
data = rbind(data, data1)
}
dim(data)

## [1] 2496 17
sum(data$dk)

## [1] 146880.9
attach(data)
estima <- data.frame(Income, Expenditure)
area <- as.factor(PSU)
stratum <- as.factor(Stratum)
hist(data$dk)

2
Histogram of data$dk
700
500
Frequency

300
100
0

0 50 100 150 200 250

data$dk
boxplot(data$dk ~ data$Stratum)
250
200
150
data$dk

100
50

idStrt001 idStrt021 idStrt041 idStrt061 idStrt081 idStrt101

data$Stratum
E.UC(stratum, area, dk, estima)

## N Income Expenditure
## Estimation 1.468809e+05 8.290466e+07 5.427982e+07
## Standard Error 3.351918e+03 4.465049e+06 2.620232e+06
## CVE 2.282065e+00 5.385764e+00 4.827267e+00
## DEFF Inf 1.145389e+01 1.355435e+01

3
Muestreo autoponderado en dos etapas estratificado

data('BigCity')
FrameI <- BigCity %>% group_by(PSU) %>%
summarise(Stratum = unique(Stratum),
Households = length(unique(HHID)),
Income = sum(Income),
Expenditure = sum(Expenditure))
attach(FrameI)
head(FrameI)

## # A tibble: 6 x 5
## PSU Stratum Households Income Expenditure
## <chr> <chr> <int> <dbl> <dbl>
## 1 PSU0001 idStrt001 26 70912. 44232.
## 2 PSU0002 idStrt001 32 68887. 38382.
## 3 PSU0003 idStrt001 24 37213. 19495.
## 4 PSU0004 idStrt001 22 36926. 24031.
## 5 PSU0005 idStrt001 28 57494. 31142.
## 6 PSU0006 idStrt001 30 75272. 43473.
sizes = FrameI %>% group_by(Stratum) %>%
summarise(NIh = n(), nIh = 2)
NIh <- sizes$NIh
nIh <- sizes$nIh
head(sizes)

## # A tibble: 6 x 3
## Stratum NIh nIh
## <chr> <int> <dbl>
## 1 idStrt001 9 2
## 2 idStrt002 11 2
## 3 idStrt003 7 2
## 4 idStrt004 13 2
## 5 idStrt005 11 2
## 6 idStrt006 5 2
resI <- S.STpiPS(Stratum, Households, nIh)
head(resI)

## [,1] [,2]
## [1,] 4 0.1774194
## [2,] 6 0.2419355
## [3,] 12 0.1717791
## [4,] 20 0.1717791
## [5,] 23 0.2553191
## [6,] 26 0.2978723
samI <- resI[, 1]
piI <- resI[, 2]
UI <- levels(as.factor(FrameI$PSU))
sampleI <- data.frame(PSU = UI[samI], dI = 1/piI)
FrameII <- left_join(sampleI, BigCity[which(BigCity$PSU %in% sampleI[,1]), ])
attach(FrameII)
head(FrameII)

4
## PSU dI HHID PersonID Stratum Zone Sex Age MaritalST
## 1 PSU0004 5.636364 idHH00042 idPer01 idStrt001 Rural Male 57 Married
## 2 PSU0004 5.636364 idHH00042 idPer02 idStrt001 Rural Female 48 Married
## 3 PSU0004 5.636364 idHH00042 idPer03 idStrt001 Rural Male 20 Single
## 4 PSU0004 5.636364 idHH00042 idPer04 idStrt001 Rural Female 18 Single
## 5 PSU0004 5.636364 idHH00042 idPer05 idStrt001 Rural Male 16 Single
## 6 PSU0004 5.636364 idHH00042 idPer06 idStrt001 Rural Female 0 <NA>
## Income Expenditure Employment Poverty
## 1 527.84 309.09 Inactive NotPoor
## 2 527.84 309.09 Inactive NotPoor
## 3 527.84 309.09 Employed NotPoor
## 4 527.84 309.09 Inactive NotPoor
## 5 527.84 309.09 Inactive NotPoor
## 6 527.84 309.09 <NA> NotPoor
HHdb <- FrameII %>%
group_by(PSU) %>%
summarise(Ni = length(unique(HHID)),
ni = 5)
Ni <- as.numeric(HHdb$Ni)
ni <- 3
head(HHdb)

## # A tibble: 6 x 3
## PSU Ni ni
## <chr> <int> <dbl>
## 1 PSU0004 22 5
## 2 PSU0006 30 5
## 3 PSU0012 28 5
## 4 PSU0020 28 5
## 5 PSU0023 24 5
## 6 PSU0026 28 5
sam = S.SI(Ni[1], ni)
clusterII = FrameII[which(FrameII$PSU == sampleI$PSU[1]), ]
sam.HH <- data.frame(HHID = unique(clusterII$HHID)[sam])
clusterHH <- left_join(sam.HH, clusterII, by = "HHID")
clusterHH$dki <- Ni[1]/ni
clusterHH$dk <- clusterHH$dI * clusterHH$dki
data = clusterHH
head(data)

## HHID PSU dI PersonID Stratum Zone Sex Age MaritalST


## 1 idHH00050 PSU0004 5.636364 idPer01 idStrt001 Rural Female 41 Separated
## 2 idHH00050 PSU0004 5.636364 idPer02 idStrt001 Rural Male 19 Single
## 3 idHH00050 PSU0004 5.636364 idPer03 idStrt001 Rural Female 16 Single
## 4 idHH20687 PSU0004 5.636364 idPer01 idStrt001 Rural Male 57 Married
## 5 idHH20687 PSU0004 5.636364 idPer02 idStrt001 Rural Female 48 Married
## 6 idHH20687 PSU0004 5.636364 idPer03 idStrt001 Rural Male 20 Single
## Income Expenditure Employment Poverty dki dk
## 1 503.92 331.92 Employed NotPoor 7.333333 41.33333
## 2 503.92 331.92 Inactive NotPoor 7.333333 41.33333
## 3 503.92 331.92 Inactive NotPoor 7.333333 41.33333
## 4 527.84 309.09 Inactive NotPoor 7.333333 41.33333
## 5 527.84 309.09 Inactive NotPoor 7.333333 41.33333
## 6 527.84 309.09 Employed NotPoor 7.333333 41.33333

5
for (i in 2:length(Ni)) {
sam = S.SI(Ni[i], ni)
clusterII = FrameII[which(FrameII$PSU == sampleI$PSU[i]), ]
sam.HH <- data.frame(HHID = unique(clusterII$HHID)[sam])
clusterHH <- left_join(sam.HH, clusterII, by = "HHID")
clusterHH$dki <- Ni[i]/ni
clusterHH$dk <- clusterHH$dI * clusterHH$dki
data1 = clusterHH
data = rbind(data, data1)
}
sum(data$dk)

## [1] 148253
dim(data)

## [1] 2585 15
attach(data)
estima <- data.frame(Income, Expenditure)
area <- as.factor(PSU)
stratum <- as.factor(Stratum)
hist(data$dk)

Histogram of data$dk
800
600
Frequency

400
200
0

0 50 100 150 200 250

data$dk
boxplot(data$dk ~ data$Stratum)

6
250
200
150
data$dk

100
50

idStrt001 idStrt021 idStrt041 idStrt061 idStrt081 idStrt101

data$Stratum
E.UC(stratum, area, dk, estima)

## N Income Expenditure
## Estimation 1.482530e+05 8.382537e+07 5.363328e+07
## Standard Error 4.056097e+03 4.551031e+06 2.262998e+06
## CVE 2.735929e+00 5.429181e+00 4.219391e+00
## DEFF Inf 7.373944e+00 7.260026e+00

You might also like