Our research was initiated in an attempt to analyse the digital gap between nations. With this aim, we approached the measurement of digital disparities between countries by means of multivariate statistical methods. In particular we used factor and cluster analysis to obtain a classification of the levels of digital development. This framework was applied to the European Union (EU) before its last increase in number of nations. The analysis led to the identification of two factors and four groups of countries in the EU, showing asymmetry in the development of the information society.
At the World Summit of the Information Society in December 2003, heads of state and government from all over the world declared that the global challenge for the new millennium was to build a society where everyone could access and share information, enabling individuals and communities to achieve their full potential in promoting their development and improving their quality of life. This commitment was reaffirmed in the second phase of the summit in November 2005 [47] and [48].
To achieve this goal, however, some obstacles needed to be overcome, especially the extreme disparities of access to information technology both between and within nations. Before deciding on suitable actions that must be implemented to bridge those gaps, it was necessary to be aware of their size. Within this context, the development of accurate indicators and measures of digital disparities between groups and nations became a matter of special importance. This task, however, was not free of difficulties for many reasons, such as the absence of a clear definition of both the information society and the digital divide, the lack of a theoretical framework, and the shortage of adequate and harmonised data.
We have approached the analysis of the digital divide by means of multivariate methods. Our aim was to provide a decision support tool for policy actions to be used as a supplementary instrument to the usual approach of composite indicators. Factor analysis was found to be a useful technique for exploring the underlying dimensions of the digital divide, as well as for dealing with the complexity of this issue.
Apparently the multiple dimensions of the digital divide can be summarised in two factors: the first is related to ICT infrastructure and use and the second to costs and the availability of online public services. Moreover, factor analysis facilitates the detection of countries with similar digital profiles.
Cluster analysis offers other insights on the digital gap, classifying countries according to their levels of digital development and showing their strengths and weaknesses. It shows that France, being on top in terms of economic development, joins the less developed group together with the Southern countries. Moreover, it reveals the weakness of Belgium and Luxemburg in not providing cheap access to the Internet and online public services. Also there is a confirmation of the well-known North-South European divide. Therefore, the results of our analysis reinforce the fact that digital disparities mirror (to some important extent) social and economic imbalances across countries.
Some limitations must, however, be considered. First, our analysis refers to the digital divide at a given point of time. However, a proper appreciation of the gap requires an understanding of its evolution. Second, our empirical application consists of just 10 variables. Hence, some aspects of the information society may not be covered. Third, we perform cross-country comparisons using indicators at the national level (even though this analysis is indispensable in the definition of national strategies, it may hide internal country digital divides by income, education, activity, location). Fourth, the size of our database prevents us from defining a classification rule by means of multivariate methods such as discriminant analysis; if we were to define a classification rule in order to assign a new country to one of the clusters we should compute its scores on each factor and compare them with the average values for each cluster and then the country would be assigned to the cluster with the closest figures.
Some policy actions might be proposed from our results. In particular, the four clusters would need to consider different policies. The improvement of France and Southern countries’ position would require a multi-faceted strategy, stimulating both the demand and supply sides. Thus, ICT infrastructure (needed) would be worthless if the connection prices were still high. For Belgium and Luxemburg, a specific e-government plan could increase the provision of online public services (for instance, submitting applications, paying taxes) and lower Internet costs. For Finland, Ireland, and the United Kingdom, provision of broadband access is needed while Austria, Denmark, Germany, The Netherlands, and Sweden should put emphasis in the provision of online public services.
