Ms. Gilthorpe, THE IMPORTANCE OF NORMALIZATION IN THE CONSTRUCTION OF DEPRIVATION INDEXES, Journal of epidemiology and community health, 49, 1995, pp. 45-50
Study objectives - Measuring socioeconomic deprivation is a major chal
lenge usually addressed through the use of composite indices. This pap
er aims to clarify the technical details regarding composite index con
struction. The distribution of some variables, for example unemploymen
t, varies over time, and these variations must be considered when comp
osite indices are periodically re-evaluated. The process of normalisat
ion is examined in detail and particular attention is paid to the impo
rtance of symmetry and skewness of the composite variable distribution
s. Design - Four different solutions of the Townsend index of socioeco
nomic deprivation are compared to reveal the effects that differing tr
ansformation processes have on the meaning or interpretation of the fi
nal index values. Differences in the rank order and the relative separ
ation between values are investigated. Main results - Constituent vari
ables which have been transformed to yield a more symmetric distributi
on provide indices that behave similarly, irrespective of the actual t
ransformation methods adopted. Normalisation is seen to be of less imp
ortance than the removal of variable skewness. Furthermore, the degree
of success of the transformation in removing skewness has a major eff
ect in determining the variation between the individual electoral ward
scores. Constituent variables undergoing no transformation produce an
index that is distorted by the inherent variable skewness, and this i
ndex is not consistent between re-evaluations, either temporally or sp
atially. Conclusions - Effective transformation of constituent variabl
es should always be undertaken when generating a composite index. The
most important aspect is the removal of variable skewness. There is no
need for the transformed variables to be normally distributed, only s
ymmetrically distributed, before standardisation. Even where additiona
l parameter weights are to be applied, which significantly alter the f
inal index, appropriate transformation procedures should be adopted fo
r the purpose of consistency over time and between different geographi
cal areas.