ON THE MEASUREMENT OF ECONOMIC DEVELOPMENT USING SCALOGRAM ANALYSIS

Abstract

MOST D A T A wi th which economists deal are measurable in te rms of universa l ly accepted s tandards. For this reason, economists tend to employ techniques of pa ramet r i c s ta t i s t ics . Despi te the measurab i l i ty of data, somet imes i t is be t t e r to resor t to a non-parametr ic s ta t i s t ica l method because of the na ture of the data. Usually, non-paramet r ic methods of s ta t i s t i cs are adopted when the da ta do not sa t i s fy the under ly ing assumptions of pa ramet r i c s ta t i s t ics such as normal i ty and homogenei ty of variance. There are other reasons for using non-parametr ic methods. In many instances, where approximat ions and est imat ions are involved in the process of col lect ing the data, these da ta become inaccura te and the marg in of e r ror is widened. Here magni tudes lose the i r significance and become more qua l i ta t ive than quant i ta t ive in nature. In fact , where such wide marg ins of e r ror are involved, we can descr ibe the da ta as semi-qual i ta t ive (or semi-quant i ta t ive) . In these cases a non-parametr ic s ta t i s t ica l method is more appropr ia te . Moreover, these methods are more justif iable when we know tha t the er rors are not randomly d is t r ibuted . * Scal ing techniques are a class of non-parametr ic methods which have been widely used in the fields of sociology and social psychology. In this paper, we propose to apply one of these scal ing techniques, namely scalogram analysis , to the classiLfication of economic data. The p r imary focus of the s tudy is to use scalogram analysis to develop an index consis t ing of more than one variable, for measur ing the degree of economic development, or r a t h e r to use i t to c lassify countr ies wi th respect to their s tages of development.