DIGITAL INEQUALITY

DIGITAL INEQUALITY: AN ANALYSIS OF THE CONNECTION BETWEEN TECHNOLOGICAL AND EDUCATIONAL INEQUALITIES
KAMILA KOLPASHNIKOVA
Graduate School of Interdisciplinary Information Studies, the University of Tokyo, 7-3-1 Hongo
Bunkyo-ku, Tokyo, 113-0033, Japan
E-mail: ippanemail@gmail.com
This paper represents a correlation analysis between technological infrastructure and human capital, taken in aggregated form, with the help of a digital divide index. The analysis was made with a view to see into the character of the relationship between technological and educational inequalities. The research employs ordinary least squares method, using GNU software – Gretl. It was found that the correlation between ICT infrastructure and human capital is strong, despite the fact that their distribution patterns are not that similar. This suggests the idea that for the development of human capital of a country, it is necessary to have an extensive ICT infrastructure.
1. Introduction
“The future has already arrived. It's just not evenly distributed yet.” William Gibson, American-Canadian writer.
As Gibson had noted, we live in the age of technology. A lot of opportunities are now available off- as well as on-line, of course, the latter has in general more to offer. Eminent universities, such as Massachusetts Institute of Technology, provide their courses through the web (MIT Open Course Ware), and that makes people who has access to the Internet in a way educationally more fortunate than the rest of the world.
Not all the people of the globe have the same opportunities of access to the Internet, or other information and communication technologies (ICTs), this difference in opportunities is called digital divide (or digital inequality). And there exists a close relationship between education and technology. The former is made easier and has more opportunities if the latter is provided. This means that the knowledge may become more easily accessible with the development of ICT infrastructure as it increases the information exchange.
Such infrastructure is necessary for the economic development of any country, which may largely depend on the said technology, but also, of course, on the human capital (Becker, 1967; Schultz, 1972), that is on education and training.
Human capital plays an important role in a country's development. However, before we manage human capital (and hence, knowledge) and decide how to improve it, we have to decide what knowledge is to be developed in a state. In the modern age, it is mostly the technology related skills that need to be increased, rather than any other skills, as technology becomes more and more intertwined with our everyday lives.
These skills are helping people to handle technology and software packages. They are very important for the productivity of those people, as the more they know about technologies, the faster they operate them.
As there is a connection between ICTs and productivity (Jorgenson, 1999; Brynjolfsson and Hitt, 2003; Shafer and Byrd, 2000), managing knowledge (human capital) connected with ICTs becomes important not only on the company's level, but on the state's level as well, as productivity of a nation in whole depends on the productivity of all individuals combined.
This research concentrates on the state level, trying to estimate the relationship between technological infrastructure and human capital.
2. Research Background
There are two main directions which are of interest to the present research. The notion of digital inequality must be made clear and the empirical ways to measure it should be found from the research done in the area before. First, we will start with the definition of digital inequality, then, the indices on digital divide will be brought to show what indicators are used in the measurement of the digital inequality. The terms “digital inequality” and “digital divide” are used here interchangeably.
2.1. The Definition of Digital Inequality
There are numerous different research directions in the connection with the concept of inequality. To narrow down our quest, we have to answer the question, which was pointed out by Sen and Foster (1997): “equality of what?” or in this case the inequality of what. The two notions, equality and inequality, are to be treated as the opposites.
When we talk about the digital inequality, we mean the inequality or equality of “the access to computers and the Internet” Van Dijk (2006, p. 178). Some authors use it in narrower meaning as the access only to the Internet (Castells, 2002, p. 248), while there are others who use it as the access to all types of ICTs in general.
Van Dijk in the work “Digital divide research, achievements and shortcomings” (2006) comes up with 5 different types of inequality:
 Technological – “technological opportunities.”
 Immaterial - “life chances, freedom.”
 Material - “capital (economic, social, cultural), resources.”
 Social - “positions, power, participation.”
 Educational - “capabilities, skills.”
The inequalities, that are directly connected with two matters under discussion, ICT and education, are represented in three of the types of this classification: namely, in technological, material, and educational.
However, the material (tangible) side of inequality, in case of the digital divide, may be represented either as money (investment either in technology or in education and training) or products (here, only technologies). Hence, it is already covered by the notions of investment into factors of production and technological inequality.
That leaves only two sides of inequality that is significant for the present research: technological and educational inequality.
The technological inequality, defined as the technological opportunities, is represented by info-tech-infrastructure discrepancies. Those discrepancies are defined by the availability of the infrastructure to the population of a given country. The infrastructure that is relevant in this research is the infrastructure, represented by computers and the Internet, according to the definition of digital inequality. In other words, this type of inequality is defined not only by, as Kling (1998) has put it, in the availability of “computers of adequate speed and equipped with appropriate software for a given activity,” but also by the availability of the access to the Internet, which corresponds with the definition by van Dijk (2005).
The educational inequality may be represented in the two aspects: formal education itself and skills that can be classified according to the types ICTs in question: computers and the Internet. Those skills will include: software literacy, hardware literacy (in whole called IT literacy - “the ability to use IT for a range of purposes” (Servon, 2002, p.7)), and language (Brooks et al., 2005), as a necessary skill for maximizing the utility of using the Internet.
Hardware literacy is the skills that are necessary to manipulate hardware. In analogy, the software literacy is the skills to handle with the PC applications and programs. They are defined by the access to hardware and software.
Formal education, on the other hand, is requiring a little more consideration. In order to comprehend the notion of education, in the context of the present framework it is defined by the the access to it. As the primary and secondary education is more or less equally distributed around most of the world (UNDP, 2007), the education which matters in economic way is generally the tertiary education. Many researchers connect education with the indicators of welfare. DiMaggio and Hargittai (2001) “hypothesize that, in the long run, education will be a strong predictor of the use of the Internet for the enhancement of human capital, the development of social capital, and political participation.” Or others claim that “the residual benefits from having a better educated populace can often include the development of a more highly skilled workforce and an improved economy overall” (Brooks et al., 2005). Or also, as Duncan Campbell (2001) has stated, “no developing country has made substantial progress in the information economy or achieved entry into global value chains in information and knowledge-based services without an educated, skilled workforce.”
Additionally, researchers like van Dijk and de Haan (2000; in van Dijk, 2006, p. 229) have concluded that “a striking result is that those having a high level of traditional literacy also possess a high level of digital information skills.” That means that most of the differentiation in digital divide will be within the workforce that has tertiary education (the highest level of education, among the three traditionally defined), as they are the ones that handle the ICTs the best (Tien and Fu, 2008). Not only highly-educated people deal with ICTs the best way, but they also “use the Internet more often for information, while less educated people use the Internet more frequently for entertainment (Bonfadelli, 2002; van Dijk, 1999)” (in Peter and Valkenburg, 2006).
But not only the levels of education define the computer skills, but also the computer skills and availability of ICTs define in their turn the overall information possession. “one of the most unfortunate by-products of the digital divide is its negative impact on educational efforts throughout the developing world. Digital technologies provide exciting new opportunities for students in the industrialized world to obtain large amounts of current information on almost any topic, to communicate their thoughts in dynamic new ways, and to work more efficiently than ever before possible. Without access to the benefits of ICT..., students in less developed countries may fall even further behind their peers in other nations” (Tiene, 2002).
How those two inequalities are connected with the digital divide? As the digital divide is not only access to tangible technologies, but also reflects the immanent characteristics of the person who uses them, it is here assumed that digital divide includes the both notions of technological and educational inequalities.
To summarize, the digital inequality is defined by access discrepancies to computers, the Internet, education and training.
2.2. Digital Divide Indicesa
There are ways to calculate the approximate level of technological infrastructure and education levels. For this purpose aggregate indices are used. They represent a framework for analyzing the digital inequality.
There are three indices that are worth to be noted. They are Digital Opportunity Index (DOI), ICT Opportunity Index (ICT-OI) and Digital Divide Index (DIDIX).
2.2.1. Digital Opportunity Index (DOI)
The Digital Opportunity Index was an outcome of the World Summit on the Information Society (WSIS) and it was meant to see how things have progressed in overcoming the digital divide (ITU and UNCTAD, 2007, p. 35). It consists of three groups that cumulatively include 11 indicators, DOI is presented in the following Figure 1.


Source: ITU and UNCTAD. World Information Society Report 2007: Beyond WSIS.
Figure 1. The Digital Opportunity Index

The indicators of DOI index are focused on solely technological inequality: mobiles, Internet, computers and fixed lines, with an emphasis on different types of Internet access. As in the present research only the technological infrastructure that is connected with computers and the Internet is analyzed, the ICTs that are computers or can provide access to the Internet are regarded. In the case of DOI, those are Internet access, broadband subscribers, fixed lines (as one can connect to the Internet through fixed lines) and mobile phones (as they also can be used as the tools for accessing the world wide web).
2.2.2. ICT – Opportunity Index (ICT-OI)
The ICT Opportunity Index (ICT-OI) was made based on two indices Digital Access Index (DAI) and Digital Divide Index (DDI), it was influenced by the collaboration with Canadian company Orbicom (Ibid., p. 119). DDI project was launched by Orbicom and the Canadian International Development Agency (CIDA) in 2002-2003. ICT-OI is very alike the DDI, rather than DAI. ICT-OI consist of 2 groups of indicators: Info-Density and Info-Use. While Info-Density reflects productive capacity, Info-Use reflects consumption. These two groups are divided into 4 sub-groups: Networks, Skills, and Uptake, Intensity, correspondingly.

Source: ITU and UNCTAD. World Information Society Report 2007: Beyond WSIS.
Figure 2. The ICT Opportunity Index

The indicators, additionally to the technological infrastructure (main telephone lines, mobiles, Internet, computers and TVs [however, TVs cannot be used as the access tool to the Internet]), draw attention to the education inequality component of digital divide and introduce literacy and enrollment rates.
2.2.3. DIDIX
The Digital Divide Index (DIDIX) is an index introduced by German consulting firm Empirica (Empirica, 2007). The European Commission and the Eurostat are among the clients that use Empirica's consulting services. The index is built on survey results and uses grouping by gender, age, income and education. The DIDIX has divided the results into four “risk groups” accordingly.
In the Figure 3, DIDIX is shown. In the lest column, the groups of the social strata under research are brought. On the right side, the three dimensions (three indicators) of digital divide are shown with their respective weights.

Source: Empirica. Benchmarking in a Policy Perspective. Dec., 2006.
Figure 3. The Digital Divide Index (DIDIX)

The indicators of the DIDIX concentrate on two components of technological infrastructure: computers (namely, hardware) and the Internet use. In the risk group, the education inequality, emphasizing the formal education (here secondary).
Overall, these three indices are developed to estimate the technological and educational inequalities, using a number of variables.
3. Research Methodology
For the analysis data from an aggregate index was employed. Several variables were aggregated to be finalized in two, according to two main inequality types: technological and educational. The index is called global digital divide index, it was constructed by Kolpashnikova K (in 2009).
Global Digital Divide Index (GDDI) consists of data of 63 countries for which the data was obtainable.
In parallel to its definition, the generated digital divide index has two sub-indices: Technological Infrastructure (TI) and Human Capital (HC). They represent the levels of technological and educational inequalities accordingly.

Figure 4. Global Digital Divide Index.

In the left side of the Figure 4, there are the indicators of inequality which were obtained through the correlation analysis with GDP (only those variables were regarded as inequality indicators that had correlation coefficient with GDP per capita higher than 0.7). The arrows indicate sub-indices, the right side gives the whole global digital divide index. Each sub-index was calculated using geometric average. The GDDI itself represents geometric mean of its two sub-groups.
In the present research, the sub-indices of GDDI are of more importance than the index itself, because they deal precisely with the phenomena which this research tries to investigate: technological and educational inequality. The research tries to find out the relationship between the sub-indices in order to estimate the correlation between technological and educational inequalities. This will help to understand what kind of countries tend to have better educational and technological opportunities. The focus is on the countries with the same infrastructural or educational levels.
For the further analysis (OLS method), statistical program Gretl 1.8.1. (GNU Regression, Econometric and Time-series Library) was used.
With the view explained above, namely, to see into the connection between the Technological Infrastructure and Human Capital, firstly, the distributions of the sub-indices were analyzed, taking the values of their means, medians, maximum and minimum values as the main attention spots.
Secondly, the correlation between the two sub-indices, that is, Technological Infrastructure and Human Capital, was made.
Their correlation, which will be discussed in Results and Discussions section, if represented in graphical way, will look the following way (see Figure 5):

Figure 5. The Correlation of TI with HC

As we can see from the Figure 5, Technological Infrastructure (horizontal) and Human Capital (vertical) are highly correlated (R2 = 0.81).
4. Limitations
As we can see from the Figure 5, the externalities of the usage of geometric mean with the dimension index appeared in the graph. Croatia (studendized residual = -4.397) and Greece (studendized residual = 2.954) seem to have fallen out of the frame. The distortion seems to be originated within the Human Capital sub-index.
The reason for such a low number of a Croatian Human Capital lies in the fact that the ICT expenditure variable was equal 0, which dragged down the whole sub-index. The author believes that there was a mistake in the data on Croatia, because it is hard to believe that the ICT expenditure is this low in a contemporary European country. Hence, the data provided by the World Bank might have contained a mistake. In fact the same data for Croatia was 364 in the year of 2002, which may reflect a better indicator. That is why the data on Croatia should be neglected.
The Greece's success in Human Capital indicator, on the other hand, seems to be explained by high levels of tertiary enrollment together with comparatively high levels of piracy.
Another limitation of the present index is that it is based on the indicators most of which are brought to a proportion with the population of a country, hence, they are heavily favor the small countries, and sometimes disregarding the network effects that are enjoyed by the bigger countries, that is perhaps, one of the explanations of why the countries like China and India are so much far from the leaders.
And even United States may be considered as a big country which is discriminated against by the favoring of small European states in the GDDI.
As it concerns the research on the whole, we also have to take into consideration that the indicators in the present research do not consider the concepts of human capital and access as a definition of the phenomena but as names for sub-indices, therefore, the conclusions on the sub-indices might not correspond accurately with the concepts themselves.
Moreover, the present research is based on the assumption that ICTs are one of the main factors contributing to the economic development of countries, disregarding, perhaps, the trade-off decisions faced by policymakers.
5. Results and Discussions
Now let us see into the analysis of sub-indices.
5.1. The First Sub-Index (Technological Infrastructure)
On the Table 1, the distribution analysis of technological infrastructure is summarized.







Table 1. The Distribution Analysis of Technological Infrastructure (TI)
Statistics for the variable TI
Mean 0.33518
Median 0.26316
Minimum 0.00042
Maximum 0.81578
Standard Deviation 0.25520
C.V. 0.76137
Skewness 0.32169
Ex. kurtosis -1.29770

The mean of the Technological Infrastructure distribution is 0.33518, which means that Technological Infrastructure sub-index's average is 0.33518 around the world. The standard deviation is 0.25520.
The median is 0.26316, which means that more than half of the countries have less than average 0.33518 level.
The minimum value is 0.00042 (Bangladesh), and maximum is 0.81578 (Sweden). The range is 0.81537.
Divided into 9 bins the distribution will look in the way, like it is represented in the graph in the Figure 6:
Figure 6. The Distribution of Technological Infrastructure

Table 2. Distribution Data for TI
interval midpt frequency rel. cum.
< 0.10192 0.050960 14 22.22% 22.22% *******
0.10192 - 0.20384 0.15288 12 19.05% 41.27% ******
0.20384 - 0.30576 0.25480 8 12.70% 53.97% ****
0.30576 - 0.40768 0.35672 6 9.52% 63.49% ***
0.40768 - 0.50960 0.45864 2 3.17% 66.67% *
0.50960 - 0.61152 0.56056 7 11.11% 77.78% ***
0.61152 - 0.71344 0.66248 9 14.29% 92.06% *****
0.71344 - 0.81536 0.76440 4 6.35% 98.41% **
>= 0.81536 0.86632 1 1.59% 100.00%

As we can see from the Figure 6, the distribution is positively skewed, while the whole distribution in some ways repeats the distribution of the GDDI itself. The modal group is in the first bin which contains 14 countries. There is one clear leader – Sweden, which occupies the last 9th bin. The fifth bin (the one in the middle) seem to represent where the divide between technology affluent and technology poor countries lie.
5.2. The Second Sub-Index (Human Capital)
On the Table 3, the distribution analysis for Human Capital sub-index is presented.

Table 3. The Distribution Analysis of Human Capital (HC)
Statistics for the variable HC
Mean 0.31539
Median 0.31047
Minimum 0.03954
Maximum 0.58888
Standard Deviation 0.15407
C.V. 0.48852
Skewness -0.07772
Ex. kurtosis -1.17500

The mean of the Human Capital distribution is 0.31539, that means that Human Capital subindex's average is 0.31539 around the world. The standard deviation is 0.15407.
The median is 0.31047, which is very close to the mean and signifies more or less harmonized distribution.
The minimum value is 0.03954 (Croatia), and maximum is 0.58888 (Norway). The range is 0.54934.
In the Figure 7, the graph of its distribution is shown, divided into 9 bins:
Figure 7. The Distribution of Human Capital (HC)






Table 4. Distribution Data for HC
interval midpt frequency rel. cum.
< 0.073878 0.039544 3 4.76% 4.76% *
0.073878 - 0.14254 0.10821 8 12.70% 17.46% ****
0.14254 - 0.21121 0.17688 5 7.94% 25.40% **
0.21121 - 0.27988 0.24554 11 17.46% 42.86% ******
0.27988 - 0.34854 0.31421 9 14.29% 57.14% *****
0.34854 - 0.41721 0.38288 6 9.52% 66.67% ***
0.41721 - 0.48588 0.45154 10 15.87% 82.54% *****
0.48588 - 0.55454 0.52021 10 15.87% 98.41% *****
>= 0.55454 0.58888 1 1.59% 100.00%

As we can see from Figure 7, the distribution is almost not skewed. The modal group is in the fourth bin which contains 11 countries, but the bins #7 and 8 also have a big number of countries – 10 each. There is one clear leader – Norway, which occupies the last 9th bin.
5.3. Correlation Analysis
The correlation between these two sub-indices is quite strong (R2 is equal 0.81, which means that 81% of human capital sub-index may be explained by technological infrastructure sub-index).
Although the indices do not exactly describe the phenomena of technological and educational inequality, they however are their quantitative approximation. Overall, it can be stated out of the results that the countries which has better technological infrastructure tend to have better educational levels. However, some countries such as Argentina, although has close educational levels to Slovenia, is comparatively far from it in the Technological Infrastructure sub-index. Or Greece and Romania, having similar technological levels, are very far from each other in human capital sub-index levels.
We could also draw a line perpendicular to regression line between Czech Republic and Spain, which could graphically show the divide between educated and technologized and less-educated and less technologically blessed.
As we can see from the results, there is a distinct relationship between ICTs and human capital (technological and educational inequalities), hence, for the development of human capital of a country, it might be essential to provide ICTs to its population.
6. Conclusions
As technological inequality is highly correlated with human capital, in order for the human capital to increase, technological inequality must be overcome. (However, we have to remember that in the present research “human capital” and “technological inequality” are considered not as the respective concepts themselves, but as the sub-indices, which consist of quantitative indicators of those concepts, based on the data available.) This tight connection of these two inequalities may be better understood in the following way: ICTs provide enormous amounts of information, and information not only the component of economic success and dominance, but also it provides resources for the development and sharing of knowledge between people. Therefore, for the successful knowledge management on a state level, countries should consider the development of technological infrastructure as well. However, there are country-specific differences and the further analysis of all the countries and regions is advisable. For example, we can see from the Figure 5, there are differences between the current development of the countries of the same region (like Eastern Europe and Latin America), or between developed countries as well (we can see the higher educational levels in the Scandinavian countries, compared to, for instance, technologically similar Israel or Hong Kong). Nonetheless, overall for the world in general, population of technologically developed countries has also more educational opportunities.
References
Becker, G. S. (1967) Human Capital and Personal Distribution of Income: An analytical approach, Institute of Public Administration.
Bonfadelli, H. (2002) “The Internet and knowledge gaps: A theoretical and empirical investigation,” European Journal of Communication, 17: 65-84.
Brooks, S, Donovan, P., and Rumble, C. (2005, December) “Developing nations, the digital divide and research databases,” Serials Review, 31(4): 270-278.
Brynjolfsson, E., and Hitt, L. M. (2003) “Computing productivity: Firm-level evidence,” Review of Economics and Statistics, 85(4): 793-808.
Castells, M. (1996) The Rise of the Network Society, Vol. 1 of The Information Age: Economy, Society and Culture. Boston: Blackwell.
Empirica. (2007, September) “Benchmarking in a policy perspective – e-inclusion,” Retrieved January 3, 2009, from http://www.empirica.biz/publikationen/documents/No06 2007_BenchPol_eInclusion.pdf International Telecommunication Union and UNCTAD (2007, June). World Information Society Report 2007: Beyond WSIS. Geneva: International Telecommunication Union.
Jorgenson, D. (1999) “Information technology and growth,” The American Economic Review, 89(2): 109-115.
Kling, R. (1998) “Technological and social access on computing, information and communication technologies,” A White Paper for the Presidential Advisory Committee on High-Performance Computing and Communications, Information Technology, and the Next Generation Internet.
Kolpashnikova K. (2008, September 12-14) “On digital divide indices,” Proceedings of the Conference of the Japan Association for Social Informatics (JASI&JSIS) 2008: 88-93.
Peter, J., and Valkenburg, P.M. (2006) “Adolescents’ Internet use: Testing the ‘‘ disappearing digital divide’’ versus the ‘‘emerging digital differentiation’’ approach,” Poetics, 34: 293-305.
Sen A., & Foster, J. (1997) On Economic Inequality After a Quarter Century, Oxford: Clarendon Press.
Servon, L. (2002) Bridging the Digital Divide: Technology, Community and Public Policy, Oxford: Blackwell Publishing.
Shafer, S.M., and Byrd, T.A. (2000) “A framework for measuring the efficiency of organizational investments in information technology using data envelopment analysis,” Omega, 28: 125-41.
Schultz, T.W. (1972) Human Resources (Human Capital: Policy Issues and Research Opportunities), New York: National Bureau of Economic Research.
Tien, F. F., and Fu, T.-T. (2008, January) “The correlates of the digital divide and their impact on college student learning,” Computers & Education, 50(1): 421-436.
Tiene D. (2002, September) “Addressing the global digital divide and its impact on educational opportunity,” Educational Media International, 39(3/4): 211-222.
Van Dijk, J. (1999) The Network Society, London: Sage.
Van Dijk, J. (2002) “A framework for digital divide research,” Electronic Journal of Communication, 12(1,2).
Van Dijk, J., and Hacker K. (2003) “The digital divide as a complex and dynamic phenomenon,” The Information Society, 19: 315-326.
Van Dijk, J. (2005) The Deepening Divide: Inequality in the Information Society, London: Sage.
Van Dijk, J. (2006) “Digital divide research, achievements and shortcomings,” Poetics, 34: 221-235.