Citation Context

...transforming the various scales of variables into one unique scale. The normalized indicators are then aggregated using specific formulas (e.g. arithmetic mean). If one indicator is more “important” than another, the former is assigned a stronger weight than the latter within the aggregation procedure. Scientifically sound methods for normalization (to make data ‘comparable’), weighting (to specify the ‘correct’ interrelationships), and aggregation (to get the ‘right’ functional relationship) are obviously pre-requisite for the construction of meaningful SD indices. However, as pointed out by Nardo et al. (2005, p. 21), already the normalization of data implies a value judgment, as different scales could not be harmonized in an meaningful manner. The same applies to weighting since this involves potentially normative ‘quotas of substitution’ (Freudenberg, 2003). For example, if the SD index weights income per capita (in 2 Following Ott (1978) we refer to the normalized variables as indicators. 4 billion US$) twice as much as emissions of carbon dioxide (in Mtons with negative sign), the SD level remains constant when a country increase per-capita-income by half a billion US$ and thereby also increas...

Citation Context

...blems and to characterize the holistic nature of the topic. The effort tossum up several indicators in a single indicator could be too risky and simplistic, sure enough useful for datasdissemination (=-=Nardo et al., 2005-=-, Jollands, 2006).sAnother issue under expert discussion is the chance of using indicators not only to have a look onssustainability performance, but also to exploit their outcomes in order to underta...

Citation Context

...tatus of the households in both small and large scale population-based surveys had been justified earlier [17,40,41]. The justification was based on the fact that it is easy to obtain information on economic proxies from households through simple questions or direct observation [40] rather than incomes which might have been falsified. Despite the popular use of assets, there are concerns about various issues such as the choice of indicators, the weight of individual indicators, the treatment of missing values and the choice of aggregation of indicators in the construction of composite indices [42,43]. We used a total of 32 assets in this analysis to arrive at the WQs. This number is quite higher than the eight assets used in a Mauritania study [15]. This study also showed that WI was sensitive to the types of each of the assets involved in the analysis. For example, assets such as motorcycle and radio which are owned by most households irrespective of living standards had nearly same proportions in the five quintiles. This distribution is typical of Nigerian socio-economic life. While nearly every household in the northern states of Nigeria possess the radio, the motorcycle is indeed the ...

Citation Context

...ariation in the data. Using the varimax rotated component loading matrix the factor 28sloading of each indicator is squared and then divided by the explained variance (see Johnson and Wichern (2002), =-=Nardo et al. 2005-=-) and then re-scaled so that the final weights add up to one. If an indicator has comparative strong capacity to explain the variation in the data, it would be expected to receive a relatively high we...