Page 169 - Contributed Paper Session (CPS) - Volume 6

CPS1866 Milica Maricic et al.
where k is the number of indicators in the composite indicators. After each
exclusion, the rank of entities R k h is obtained. In the third step the Spearman

correlation coefficients between the R and all R k h are obtained. If the
k
Spearman correlation coefficient is close to 1 that indicates that an indicator
can be removed. As a stopping rule Marozzi (2009) suggests when the value
of Spearman correlation coefficient drops below 0.9 and 0.8.
For example, Markovic et al. (2016) proposed a post hoc I-distance
approach to reduce the number of indicators which make the OECD Better life
index. Their approach is based on the I-distance method, a multivariate
statistical analysis which is able to synthetize indicators without assigning
weights (Ivanovic, 1977). The method stands out as attempts to minimize the
duplicity of information (Jeremic et al.). The post hoc I-distance undermines
the application of I-distance and removal of the least important variable for
the ranking process after each iteration. The importance of the variable for the
ranking process was measured through the coefficient of determination.
This literature review should indicate that the dimension reduction and
framework reduction is a topic of high interest in the field of composite
indicators and that statistical methodologies have so far been employed with
a lot of success to solve the issue.

3. Methodology
3.1 Sustainable Society Index (SSI)
Sustainable Society Index (SSI) is a multi-layered composite indicator
consisted of 21 indicators divided into seven categories which make three
dimensions: Human wellbeing, Environmental wellbeing, and Economic
wellbeing. In our research, we focused solely on indicator data. The list of
indicators which are used to compute the SSI are listed in Table 1. For more
information on the description of the indicators please consult the official
framework description (SSI, 2018).
The structure of the SSI is based on three pillars of sustainability as
suggested in the Brundtland report (1987): Human, Economic, and
Environmental pillar. Kaivo-oja et al. (2014) observe that the structure of the
SSI is therefore quite conventional. The Joint Research Centre (JRC) conducted
an audit of the SSI in 2012 (Saisana & Philippas, 2012) and stated that there is
conceptual coherence of the structure of the SSI, that there are few imbalances
within categories, that the marginal weights do not differ too much and that
the ranking is robust. On the other hand, there are studies which showed that
the SSI could be modified. Maricic et al. (2014) used the I-distance method
and provided an alternative ranking of the countries based on the SSI
indicators which is free from weights. Savic et al. (2016) attempted to revise
the number of indicator within the SSI using the post hoc I-distance. They

158 | I S I W S C 2 0 1 9