Ninoslava Savić and Zoran Radojičić
One of the published models for the calculation of digital divide index, developed in the scope of international project Empirica( Selhofer and Hüsing, 2002) starts from noticing groups within the population divided according to different criteria( for example women, older population, low educated population etc.). Taking into account the values of chosen indicators on these groups, their average value is calculated as final composite index of digital divide.
4. The aim of research
This paper studies internal digital divide in using e‐Government services within the domain of enterprises. The aim of research is definition of the model for quantitative expression of total and individual digital divide level in usage of e‐Government in economy. This model should enable monitoring the implementation of e‐ Government on enterprises, thru identification of possible polarizations caused by the different possibilities for using new technologies and services. At the end, the model will be verified on the case of Serbia, through the research conducted in the four‐year period.
5. Model for measuring digital divide in the field of e‐Government
The proposed model is based on the classification of total digital divide on the basis of assessed causes. First step in this model is a determination of possible causes of polarization in the set of observed entities( countries, enterprises, populations etc.). For each type of digital divide that is considered, the relevant set of benchmarking indicators must be chosen, and the appropriate division of observed set of objects into groups, relevant for certain type of digital divide, must be defined( Figure 3).
Figure 3: The key elements for describing certain type of digital divide
Suggested model defines the main rules for calculation of sub‐indices and the total index of digital polarization in e‐Government. The rules are based on the idea that measuring of digital polarization in the certain domain( represented by chosen set of objects) means introduction of the measure of polarization within the domain divided according to the chosen criterion into groups. In this case, as a measure for the degree of polarization we take the difference between maximum and minimum value of particular indicator in each defined group( Savić and Radojičić, 2011). In general, this difference varies between zero( when there is no polarization) and the maximum measured value of indicator( which means the maximal polarization). Dividing this difference with the maximum value of the appropriate indicator, we obtain the value in the range of [ 0, 1 ].
The final value of sub‐index is computed as an arithmetic mean of the quotients calculated for all chosen indicators. In this calculation, all indicators have the same weight, due to simplicity. It is also possible to assign different weights to sub‐indices, according to assessment of their significance for measuring total digital polarization. The total index of digital polarization is an arithmetic mean of all sub‐indices, obtained in this way. If there are no differences in their significance, sub‐indices have the same weight in computing the arithmetic mean.
The model for calculating of e‐Government polarization index( EGPI) and the corresponding sub‐indices present an opened system, flexible in changing its main components. Its main advantage is immutability of basic semantics when the set of sub‐indices, the set of indicators for its quantitative expression and / or the set of defined risk groups within a specific digital divide are broadened or changed. It is possible to introduce new noticed type of digital divide, and to calculate new appropriate sub‐indices for them. Measuring of digital polarization sub‐indices can also be performed with an expanded list of chosen indicators. Finally, it is possible to
436