lival: Living Standards Measurement Survey Albania 2012

In modi: Multivariate Outlier Detection and Imputation for Incomplete Survey Data

Living Standards Measurement Survey Albania 2012

Description

The dataset is an extended version of the public micro data file of the LSMS 2012 of Albania available at (https://www.instat.gov.al/en/figures/micro-data/, accessed 13 February 2023). Documentation of the LSMS 2012 of Albania is from the World Bank (https://microdata.worldbank.org/index.php/catalog/1970, accessed 5 November 2020). The data set is ported to R and updated with approximate survey design information derived from the data itself. The units are households and the variables are expenditures on main categories, poverty measures and structural information including weights and sample design.

Usage

lival

Format

A data frame with 6671 rows and 26 variables

psu

primary sampling unit (psu)

hhid

unique household identifier (100*psu+hh)

hh

household number per psu

prefectu

prefecture

urban

urbanicity (Urban=1, Rural=2)

strat

stratum

region

region

totcons

total consumption of hh

rcons

real mean per capita consumption

rfood

real food consumption per capita

rtotnfoo

real non food consumption per capita

reduexpp

real education consumption per capita

rdurcons

real durable consumption per capita

rtotutil

real utilities consumption per capita

egap0

extreme headcount poverty

egap1

extreme poverty gap

egap2

extreme poverty depth

agap0

absolute headcount poverty

agap1

absolute poverty gap

agap2

absolute poverty depth

weight

final cross-sectional weight

nph

number of psu in stratum population

mph

number of households in stratum population

mphi

number of households in sampled psu

pi1

psu inclusion probability

pi2

household inclusion probability

Details

Absolute poverty measures use a poverty line of Lek 4891 (2002 prices). Extreme poverty measures use a poverty line where the basic nutritional needs are difficult to meet. The headcount poverty variable is an indicator for the income of the household y_i being below the (absolute or extreme) poverty line z. The poverty gap variable measures the relative distance to the poverty line: (z-y_i)/z. The poverty depth variable is the square of the poverty gap variable, i.e. [(z-y_i)/z]^2, giving more weight to the poorer among the poor and thus describing the inequality among the poor.

The survey design is a stratified clustered two stage design. The primary sampling units are enumeration zones. The strata are the crossing of prefecture and urbanicity and the allocation of the psu sample to the strata is proportional to the number of households. Within strata the psu are sampled with probability proportional to number of households. Within psu a simple random sample of 8 households was selected. The weights are calibrated to population margins. All survey design informations except the strata and the weights are approximated through the weights using assumptions on the design. Since the data set has undergone data protection measures and the survey design is approximate only, inference to the population does not yield exact results. However, the complexity of the data and of the survey design are realistic.

The size of the household is not on the original data set. However, the transformation capita <- round(0.07527689 * totcons/rcons, 0) yields the number of persons in the household.

Note

With R package survey a survey design object can be built with, e.g., svydesign(~psu + hhid , strata= ~strat, fpc= ~pi1 +pi2, weight= ~weight, data=lival, pps="brewer").

References

https://www.instat.gov.al/en/figures/micro-data/

Examples

data(lival)
lival$capita <- with(lival, round(0.07527689 * totcons / rcons, 0))
## Not run: 
library(survey)
lival.des <- svydesign(~psu + hhid , strata= ~strat, fpc= ~pi1 +pi2,
                      weight= ~weight, data=lival, pps="brewer")
svymean(~totcons, lival.des, deff=TRUE)

## End(Not run)

modi documentation built on March 31, 2023, 8:35 p.m.