Pesquisa de Orcamentos Familiares (POF)
Contributed by Dr. Djalma Pessoa <pessoad@gmail.com>
The Pesquisa de Orcamentos Familiares is Brazil’s national survey of household budgets.
One table of survey responses per sampled household. Additional tables, many containing one record per expenditure.
A complex sample survey designed to generalize to the civilian population of Brazil.
Released at irregular intervals, with only 2002-2003 and 2008-2009 microdata available.
Administered by the Instituto Brasileiro de Geografia e Estatistica.
Simplified Download and Importation
The R lodown package easily downloads and imports all available POF microdata by simply specifying "pof" with an output_dir = parameter in the lodown() function. Depending on your internet connection and computer processing speed, you might prefer to run this step overnight.
library(lodown)
lodown( "pof" , output_dir = file.path( path.expand( "~" ) , "POF" ) )lodown also provides a catalog of available microdata extracts with the get_catalog() function. After requesting the POF catalog, you could pass a subsetted catalog through the lodown() function in order to download and import specific extracts (rather than all available extracts).
library(lodown)
# examine all available POF microdata files
pof_cat <-
get_catalog( "pof" ,
output_dir = file.path( path.expand( "~" ) , "POF" ) )
# 2008-2009 only
pof_cat <- subset( pof_cat , period == "2008_2009" )
# download the microdata to your local computer
pof_cat <- lodown( "pof" , pof_cat )Analysis Examples with the survey library
Construct a complex sample survey design:
options( survey.lonely.psu = "adjust" )
library(survey)
poststr <-
readRDS(
file.path( path.expand( "~" ) , "POF" ,
"2008_2009/poststr.rds" )
)
t_morador_s <-
readRDS(
file.path( path.expand( "~" ) , "POF" ,
"2008_2009/t_morador_s.rds" )
)
t_morador_s <-
transform(
t_morador_s ,
control = paste0( cod_uf , num_seq , num_dv )
)
pof_df <- merge( t_morador_s , poststr )
stopifnot( nrow( pof_df ) == nrow( t_morador_s ) )
pre_stratified_design <-
svydesign(
id = ~control ,
strata = ~estrato_unico ,
weights = ~fator_expansao1 ,
data = pof_df ,
nest = TRUE
)
population_totals <-
data.frame(
pos_estrato = unique( pof_df$pos_estrato ) ,
Freq = unique( pof_df$tot_pop )
)
pof_design <-
postStratify(
pre_stratified_design ,
~ pos_estrato ,
population_totals
)Variable Recoding
Add new columns to the data set:
pof_design <-
update(
pof_design ,
one = 1 ,
# centimeters instead of meters
altura_imputado = altura_imputado / 100 ,
age_categories =
factor(
1 + findInterval( idade_anos ,
c( 20 , 25 , 30 , 35 , 45 , 55 , 65 , 75 ) ) ,
levels = 1:9 , labels = c( "under 20" , "20-24" , "25-29" ,
"30-34" , "35-44" , "45-54" , "55-64" , "65-74" , "75+" )
) ,
# create a body mass index (bmi) variable, excluding babies (who have altura_imputado==0)
body_mass_index = ifelse( altura_imputado == 0 , 0 , peso_imputado / ( altura_imputado ^ 2 ) ) ,
sexo = ifelse( cod_sexo == '01' , "masculino" , ifelse( cod_sexo == '02' , "feminino" , NA ) )
)
pof_design <-
transform(
pof_design ,
# individuals with a low bmi - underweight
underweight = ifelse( body_mass_index < 18.5 , 1 , 0 ) ,
# individuals with a high bmi - overweight
overweight = ifelse( body_mass_index >= 25 , 1 , 0 ) ,
# individuals with a very high bmi - obese
obese = ifelse( body_mass_index >= 30 , 1 , 0 )
)Unweighted Counts
Count the unweighted number of records in the survey sample, overall and by groups:
sum( weights( pof_design , "sampling" ) != 0 )
svyby( ~ one , ~ sexo , pof_design , unwtd.count )Weighted Counts
Count the weighted size of the generalizable population, overall and by groups:
svytotal( ~ one , pof_design )
svyby( ~ one , ~ sexo , pof_design , svytotal )Descriptive Statistics
Calculate the mean (average) of a linear variable, overall and by groups:
svymean( ~ body_mass_index , pof_design , na.rm = TRUE )
svyby( ~ body_mass_index , ~ sexo , pof_design , svymean , na.rm = TRUE )Calculate the distribution of a categorical variable, overall and by groups:
svymean( ~ age_categories , pof_design )
svyby( ~ age_categories , ~ sexo , pof_design , svymean )Calculate the sum of a linear variable, overall and by groups:
svytotal( ~ body_mass_index , pof_design , na.rm = TRUE )
svyby( ~ body_mass_index , ~ sexo , pof_design , svytotal , na.rm = TRUE )Calculate the weighted sum of a categorical variable, overall and by groups:
svytotal( ~ age_categories , pof_design )
svyby( ~ age_categories , ~ sexo , pof_design , svytotal )Calculate the median (50th percentile) of a linear variable, overall and by groups:
svyquantile( ~ body_mass_index , pof_design , 0.5 , na.rm = TRUE )
svyby(
~ body_mass_index ,
~ sexo ,
pof_design ,
svyquantile ,
0.5 ,
ci = TRUE ,
keep.var = TRUE ,
na.rm = TRUE
)Estimate a ratio:
svyratio(
numerator = ~ peso_imputado ,
denominator = ~ altura_imputado ,
pof_design ,
na.rm = TRUE
)Subsetting
Restrict the survey design to :
sub_pof_design <- subset( pof_design , underweight == 1 )Calculate the mean (average) of this subset:
svymean( ~ body_mass_index , sub_pof_design , na.rm = TRUE )Measures of Uncertainty
Extract the coefficient, standard error, confidence interval, and coefficient of variation from any descriptive statistics function result, overall and by groups:
this_result <- svymean( ~ body_mass_index , pof_design , na.rm = TRUE )
coef( this_result )
SE( this_result )
confint( this_result )
cv( this_result )
grouped_result <-
svyby(
~ body_mass_index ,
~ sexo ,
pof_design ,
svymean ,
na.rm = TRUE
)
coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )Calculate the degrees of freedom of any survey design object:
degf( pof_design )Calculate the complex sample survey-adjusted variance of any statistic:
svyvar( ~ body_mass_index , pof_design , na.rm = TRUE )Include the complex sample design effect in the result for a specific statistic:
# SRS without replacement
svymean( ~ body_mass_index , pof_design , na.rm = TRUE , deff = TRUE )
# SRS with replacement
svymean( ~ body_mass_index , pof_design , na.rm = TRUE , deff = "replace" )Compute confidence intervals for proportions using methods that may be more accurate near 0 and 1. See ?svyciprop for alternatives:
svyciprop( ~ obese , pof_design ,
method = "likelihood" , na.rm = TRUE )Regression Models and Tests of Association
Perform a design-based t-test:
svyttest( body_mass_index ~ obese , pof_design )Perform a chi-squared test of association for survey data:
svychisq(
~ obese + age_categories ,
pof_design
)Perform a survey-weighted generalized linear model:
glm_result <-
svyglm(
body_mass_index ~ obese + age_categories ,
pof_design
)
summary( glm_result )Analysis Examples with srvyr
The R srvyr library calculates summary statistics from survey data, such as the mean, total or quantile using dplyr-like syntax. srvyr allows for the use of many verbs, such as summarize, group_by, and mutate, the convenience of pipe-able functions, the tidyverse style of non-standard evaluation and more consistent return types than the survey package. This vignette details the available features. As a starting point for POF users, this code replicates previously-presented examples:
library(srvyr)
pof_srvyr_design <- as_survey( pof_design )Calculate the mean (average) of a linear variable, overall and by groups:
pof_srvyr_design %>%
summarize( mean = survey_mean( body_mass_index , na.rm = TRUE ) )
pof_srvyr_design %>%
group_by( sexo ) %>%
summarize( mean = survey_mean( body_mass_index , na.rm = TRUE ) )