Socioeconomic Status Measurement With Discrete Proxy Variables: Is Principal Component Analysis a Reliable Answer?

Abstract

The last several years have seen a growth in the number of publications in economics that use principal component analysis (PCA) in the area of welfare studies. This paper explores the ways discrete data can be incorporated into PCA. The effects of discreteness of the observed variables on the PCA are reviewed. The statistical properties of the popular Filmer and Pritchett (2001) procedure are analyzed. The concepts of polychoric and polyserial correlations are introduced with appropriate references to the existing literature demonstrating their statistical properties. A large simulation study is carried out to compare various implementations of discrete data PCA. The simulation results show that the currently used method of running PCA on a set of dummy variables as proposed by Filmer and Pritchett (2001) can be improved upon by using procedures appropriate for discrete data, such as retaining the ordinal variables without breaking them into a set of dummy variables or using polychoric correlations. An empirical example using Bangladesh 2000 Demographic and Health Survey data helps in explaining the differences between procedures.

Use, microbiological effectiveness and health impact of a household water filter intervention in rural Rwanda—A matched cohort study

“…At each visit, a household survey was administered to the primary cook of the household consisting of questions addressing household demographics and characteristics related to sanitation, hygiene, and drinking water practices. A socioeconomic status (SES) indicator was developed using polychoric principal component analysis (PCA) (Kolenikov and Angeles, 2009) based on household characteristics and durable goods ownership (household electricity, flooring material, wall material, radio, mobile phone, and bicycle ownership). The continuous PCA variable was divided into quintiles, and the variance explained by the first principal component was 0.535.…”

On Weighting the Components of the Human Development Index: A Statistical Justification

“…Kirshnakumar and Nadar (2008) provide an overview of statistical methods of deriving weights based on multivariate statistics, including PCA. On the use PCA to generate an income proxy, see also Kolenikov and Angeles (2009). 4 Brandolini (2007) states that "we should be cautious in entrusting a mathematical algorithm with a fundamentally normative task."…”

Is Timor-Leste's Growth Inclusive?

“…34 The index exclusive of housing characteristics is prepared to allow comparisons with the assets index presented in Filmer and Scott (2008). Kolenikov and Angeles (2009) found that ordinal variables perform better than the binary variables used by Filmer and Pritchett (2001) and many subsequent studies. They also found that a 'naive' ordinal coding that records a value of 1 to the lowest standard of asset with higher values given to higher standards in steps of 1, performed adequately in comparison to more computationally intensive alternatives.…”