Assessment of Regional Digital Divide in Türkiye

The world is changing rapidly, especially in the last decade. This change has accelerated with the help of new technological developments. Graham (2002) emphasised the two leading forces shaping the current era: urbanisation and digital information and communication technologies (ICTs). As of January 2021, 59.5% of the global population (4.66 billion active Internet users) are online (Statista 2022). Likewise, the urban population reached 55% of the world’s population (4.2 billion inhabitants) in 2020. By 2050, it is estimated that approximately seven out of 10 individuals will live in cities (The World Bank 2020). Given the information, these two topics have become crucial for sustainable development policies. Not surprisingly, the United Nations (2015) pointed out their significance in the Sustainable Development Goals via many targets.

Despite these rapid developments, inequalities still exist. In 2019, a global breakpoint showed how vulnerable our world and cities are. ICTs have become even more critical of the influence of the lockdown period due to the unexpected COVID-19 pandemic. Improved digital infrastructure and society with digital abilities increased the potential to cope with several emergencies from COVID-19. Quite the reverse, the challenges worsened in countries with fewer digital opportunities. According to the OECD report (2021), the percentage of teleworking doubled from 16% in 2019 to 37% of employees in April 2020. However, Türkiye only had 3% of employees working at home during lockdowns in 2020. Another critical impact was observed in education. The State of the Global Education Crisis: A Path to Recovery Report stated that full and partial school closures lasted about 224 days worldwide (The World Bank, UNESCO, UNICEF 2021). According to Ministry of Education statistics¹, more than seven million students actively used online education platforms during the lockdown period in Türkiye. According to the 2018 Programme for International Student Assessment (PISA) data, in Türkiye, 67% of students reported having a computer they could use for school work, which is lower than the OECD average (89%). Urbanisation can also be a determinant. The OECD Regions and Cities at a Glance Report (2020) states that “in 26 OECD countries, 1 in 3 households in rural areas do not have access to high-speed broadband on average” (OECD 2020a,b: 52). The COVID-19 pandemic proves that the digital divide still exists in different scopes of everyday life.

In the light of all information mentioned above, it is evident that digital inequalities can affect our lives; consequently, it becomes a vital theme for countries to monitor digital progress to have inclusive policies. Thus, this research will focus mainly on the literature on digital divide. Various studies are being conducted to discover the reasons behind digital inequalities. Primarily, researchers concentrated on having an Internet connection or not (the first-level digital divide); after that, the research focus shifted to ICT skills and usage (the second-level digital divide); lately, the research focus has turned to the outcomes of ICT use or tangible benefits (the third-level digital divide). Because of the complex nature of the digital divide, indexes deal with various aspects. The most recent indexes are the ICT Development Index (IDI) (ITU 2017), Digital Economy and Society Index (DESI) (EU 2018), and Going Digital Toolkit (OECD 2021a,b,c). Despite the fact that all the existing monitoring tools are examining digital inequalities on an international level, digital divide could differ between regions within the same country (Vicente, Lopez 2011). Besides, new research indicates that regional instruments support national measures to mitigate the digital divide (Szeles 2018). There are a few studies to reveal regional disparities in Türkiye in terms of the digital divide (Guz 2019; Koramaz et al. 2019; Ozcan Alp, Baycan 2024). Yet, no tool has been developed to monitor digital disparities between Türkiye’s regions. As an emerging economy, Türkiye needs to overcome the digital divide to maximise tangible benefits and boost innovation. This study aims to formulate a new index for monitoring regional digital divide in Türkiye. To do so, research starts with an inevitable question: What are the digital divide indicators in the case of Türkiye? Then goes further to reveal the digital divide between regions in Türkiye by using these indicators. Thus, the research involves three steps: finding digital divide indicators, formulating an index, and applying it for Türkiye.

The research is structured in five sections. The following section focuses on the literature on digital divide and the evolution of the term since the beginning. The second section also includes investigating recent indexes to understand digital divide indicators for all levels. The third section explores the indicators related to digital disparities in Türkiye and develops a new regional monitoring index. This allows us in the fourth section to use the new tool for revealing regional disparities in Türkiye. The final part provides brief conclusions and future research topics.

The digital divide has emerged as an umbrella term for various aspects of digital disparities between individuals. The term entered common usage by the beginning of the 20th century and has been studied by several scholars. The digital divide was initially used in an official report Falling Through the Net by the US Department of Commerce’s National Telecommunications and Information Administration (NTIA 1995), in which the term is not used as ‘digital divide’ but instead referred to those who have access to the Internet and who have not (the first-level digital divide). Conversely, the 1999 report Falling Through the Net did refer to the digital divide by counting all new technologies (NTIA 1999). After much research on material access, the term adjusted and comprehended additional aspects. According to Scopus, there were over 1,000 publications on this phenomenon since 2005 (Scopus 2022). In Table 1, three different phases of the digital divide are shown.

Initially, the term referred to material access (having Internet or not) (NTIA 1999), transformed into a more social issue, and included not only material access but also skills to use the Internet or computer (DiMaggio, Hargittai 2001). As a final step, the scope is broadened by outcomes of ICT use or tangible benefits (Warschauer 2011; van Deursen, Helsper 2015).

In the 2000s, some scholars claimed that the digital divide was shrinking because of the rapid penetration of the mobile Internet (Stump et al. 2008). However, in 2019, a global pandemic has shown that the digital divide still exists; as van Dijk explained, “The digital divide cannot be closed completely. When the whole world population would reach access to the digital media such as the Internet, inequalities of digital skills, usage and outcomes or benefits remain and even tend to grow” (van Dijk 2020: 1). Thus, the digital divide is needed to be considered for comprehensive development policies.

These theoretical digital divide considerations have affected the indexes that track countries’ digital progress. Because of its complex nature, the indexes have many indicators related to all three levels. Table 2 lists the most comprehensive and recent ones, which include different aspects of the digital divide.

The International Telecommunication Union (ITU) formulated the IDI to meet ITU Member States’ demand to establish an overall ICT index in 2008. The IDI aggregates several indicators into a single number to capture the complexity of the digital divide. It comprises 12 indicators within three sub-indexes corresponding to technical infrastructure, usage, and skills (ITU 2017). DESI has been an annual tool for monitoring European Union (EU) member states’ digital progress since 2014 (EU 2021). Its purpose is to support the member states in identifying areas for priority action each year. Table 2 shows that DESI has four main themes: technical infrastructure, ICT usage among individuals and enterprises, e-government, and e-commerce. The scope of DESI is more comprehensive than IDI and more complex as well. Not only material access but also usage, skills, and outcomes are targeted in DESI. However, it boosts the number of indicators (there are 33 indicators).

As the last index, OECD identifies seven policy dimensions in the Going Digital Toolkit (OECD 2021a,b,c) to shape digital transformation. The Going Digital Toolkit tackles various areas, such as embracing education, innovation, trade, and socio-economic outcomes. The index aims to assist governments with a complete governmental approach to the digital economy strategy. Unlike the previous two indexes, the Going Digital Toolkit recognises trust issues in digital technology. It uses some indexes as an indicator, such as OECD Digital Services Trade Restrictiveness Index and OECD Foreign Direct Investment Regulatory Restrictiveness Index, which are calculated on a country scale. In Appendix A, all indicators are listed in detail for each index. In the next section, indicators related to three levels of the digital divide will be discovered with the help of the indexes aforementioned and several scholars.

In this study, the digital divide is considered in four phases. Even if it seems similar to the existing literature, there are some unique considerations. The first step represents material access and skills together because conscious usage of them is possible with proper strategic, informational, and instrumental digital skills (van Dijk 2005). The second step represents the usage in terms of variety and regularity (DiMaggio, Hargittai 2001; Warschauer 2003; van Deursen, van Dijk 2018). The third phase represents the outcome of Internet usage (van Deursen, Helsper 2015); the consequences can create benefits for individual users or an innovation driver for communities or countries. In Türkiye’s case, using technology as a booster of innovation is crucial to overcome the digital divide in terms of Singer’s (1970) theory of technological dualism. Thus, here in this study, innovation is considered the primary effect.

As a final comment, to have digital inclusions, all three phases should be equal for all individuals in various socio-demographic and socio-economic conditions (ITU 2017; EU 2018; OECD 2021a,b,c). Thus, the fourth phase represents socio-demographic and socio-economic divides.

In the previous section, three recent and international indexes, IDI (ITU 2017), DESI (EU 2021), and Going Digital Toolkit (OECD 2021a,b,c), were analysed. According to the proposed conceptual framework, the second step is identifying digital inequality indicators that can be used in Türkiye’s case. After identifying appropriate indicators, the data-gathering process starts via various data sources. While TurkStat is the leading data provider, various other public and private authorities, such as the Information Technologies and Communications Authority, Ministry of Industry and Technology, and TurkPatent, were used as statistical data resources for the study. All indicators with data sources and units are listed in Appendix C. Thanks to principal component analysis (PCA), a new index for Türkiye will be constructed. As a final step, QGIS does the spatial analysis of digital inequalities at the regional scale (NUTS1 – 12 sub-regions) and provincial scale (NUTS3 – 81 provinces).

All indicators are divided into four categories explained in the conceptual framework. There are 19 indicators associated with the first-level digital divide, 30 indicators are connected to the second-level digital divide, and 17 hands are mainly about the third-level digital divide. Finally, 21 of them are primarily corresponding to other divides. All indicators are listed in Appendix B. After the categorisation, the next step is to evaluate them in the case of Türkiye. To do so, three questions help with the identification. They are as follows:

Does the indicator differ within the regions in Türkiye?

First, indicators that are no different within the country are eliminated, such as the broadband price index, OECD digital government index, health data sharing intensity, OECD digital services trade restrictiveness index, and OECD foreign direct investment regulatory restrictiveness index. Then, indicators that are irrelevant to digital inequalities in Türkiye are removed. For example, 5G mobile technology is not available in Türkiye, yet instead, the availability of 3G/4G mobile technology indicates digital inequality. Another decent example is a fixed telephone subscription, which is decreasing with mobile phone expansion, thus unrelating to digital inequality.

Some indicators are substituted comparable statistical data, such as techno parks, R&D, and design centres are represented as digital-intensive sectors. Since 2001, Türkiye has supported them with specific laws: Law No. 5746 on Technology Development Zones (TDZ) and Law No. 5746 on Supporting Research, Development, and Design Activities. Both laws aim to support and encourage the production of technological knowledge to make the country’s economy internationally competitive (Technology Development Zones 2001; Supporting Research, Development, and Design Activities 2008). Thus, this study accepts techno parks, R&D, and design centres as digital-intensive sectors. The number of businesses making e-commerce, ICT goods, and services as a share of international trade is also used as third-level digital divide indicators. Additionally, demographic and socio-economic factors play a vital role in Türkiye because low purchasing power creates an economic barrier to digital technologies (Ozcan Alp, Baycan 2024). ITU (2017) states that computer ownership in Türkiye is 20% lower than the European average. Additionally, Türkiye’s human capital score is nearly half of the average score of the EU member states (EU 2021). Another interesting aspect is gender inequality; the gap in Internet use between men and women in the western part is smaller compared to the eastern part of Türkiye being 6% and 22%, respectively (TurkStat 2021). Demographic and socio-economic factors include disparity in Internet use between men and women, age dependency, mean year of schooling, secondary gross enrolment ratio, percentage of tertiary graduates, poverty ratio, urban and rural households, business R&D expenditure, and GDP.

As a final step, repeated indicators are gathered, such as the percentage of households with Internet access, the share of households with fixed broadband access, and the rate of households with mobile broadband access. Similarly, indicators related to Internet speed are simplified at the length of fibre optic cable since there is no other suitable statistic for Internet speed in Türkiye.

Unluckily, many indicators could not be used because of a lack of precise or equivalent statistical data. No region- or city-scale data are available since many are measured countrywide, such as individuals’ basic digital and software skills, enterprises providing ICT training, and electronic information sharing. Furthermore, the study has limited the need for regional statistics such as e-government and e-health usage. Even if the ratio of e-government users is increasing in Türkiye (58.9%), it is still not equal among all regions (TurkStat 2021). Another significant limitation of the study was that there were no data on digital literacy and skills at a regional scale, which is the most problematic issue in Türkiye. In particular, digital skills are missing in Level 1. According to DESI (2020), Türkiye’s human capital score (23.0) was almost half the EU member states’ average (41.8). The final list of indicators can be seen in Table 3. All indicators with data sources and units are listed in Appendix C (Table 3).

Principal component analysis is used to determine the final indicators. PCA helps to have a relatively simple index by dropping the number of variables. It has been performed for all levels using the Statistical Package for the Social Sciences (SPSS). Before PCA, Pearson correlation is applied to all data, and indicators with a high correlation are eliminated by checking the correlation matrix score (r > 0.80). Six indicators have been removed as the high correlation score are the percentage of households with fixed broadband access, the share of the population covered by at least a 4G mobile network, mean year of schooling, percentage of tertiary graduates, and GDP. Then, PCA is applied to the remaining indicators. Three main steps are as follows:

The significance of KMO and Bartlett’s test was checked (KMO > 0.6 and Bartlett sig < 0.05).

ICT goods and services as a share of international trade and poverty ratio are eliminated because of PCA. After that, for each level, PCA was applied with the final indicators for each level. The results derived from the PCA are shown in Table 4.

Thanks to PCA, final indicators are determined. Then, to have a precise index, the weights are computed with the help of the IDI methodology. Finally, the following three steps are performed:

The component loadings were squared and divided by the share of variance explained by the component.

The contribution of all four levels to the new index has been determined to be equal. As a final step, the ideal value is calculated by adding two standard deviations to the mean value of the indicator (ITU 2009). After various stages, the last index can be seen in Table 5.

The new index is applied at regional (NUTS1 – 12 sub-regions) and city scales (NUTS3 – 81 provinces). The results indicate a digital divide among Türkiye regions. While regions generally converge in terms of access and use of technology (Levels 1 and 2), it is seen that the tangible benefits obtained from technology (Level 3) differ from each other in Türkiye. Levels 1 and 2 score between 0.25-0.18 and 0.49-0.28, respectively. However, the score range varies for Level 3, between 1.98 and 0.01. At the same time, it is observed that the socio-economic and demographic factors (Level 4) that are effectual in outcomes differ between regions. The score range varies for Level 4, between 0.15 and 0.07. Since Level 3 represents the outcome of Internet usage (van Deursen, Helsper 2015), to be specific, economic outcomes are considered the primary outcome in this study; the disparities are conspicuous. It can be related to socio-economic and demographic factors and unequal ability to use ICT (Warschauer 2011) to have offline outcomes (van Deursen, Helsper 2015). Supportively, Türkiye’s human capital score is lower than most EU members (EU 2021). Another critical factor is R&D expenditure as a percentage of GDP and the disparity in Internet use between men and women. That indicators show relatively unequal distribution between the regions.

Furthermore, two distinctive indicators in Level 3 are ICT companies in the business population and techno parks, R&D, and design centres for the regions in Türkiye. Therefore, these two indicators need to be targeted to boost the outcomes of ICTs and overcome technological dualism.

The results indicate that fostering the ability to use technology, supporting gender equality, encouraging R&D expenditures, and supporting initiatives, especially ICT initiatives, will assist in reducing digital inequalities in Türkiye. Tables 6 and 7 represent the new index scores on regional and city scales.

Regional scale (NUTS1 – 12 sub-regions)

Digital development differences between regions within Türkiye are seen in Figure 2. TR1 Istanbul has the highest digitisation value (2.86), followed by the TR5 Western Anatolia region, which has almost half the highest value (1.55). The score ranges from 1.55 to 0.53 within 12 sub-regions, excluding TR1 Istanbul. Digital progress is more significant in the western part of Türkiye than in the eastern region. However, TR2 Western Marmara has lower values, similar to the eastern part of Türkiye. Each region has its strengths and weaknesses related to the overall score. For example, TR5 West Anatolia region stands out with its excessive R&D expenditures, total number of techno parks, R&D, and design centres and high e-commerce usage; TR4 East Marmara region has the second highest R&D expenditures. There are some unpredicted results, as well. Even if TR5 West Anatolia and TR4 East Marmara have higher R&D expenditures, TR3 Aegean region has left them behind with the abundance of ICT initiatives and high e-commerce usage.

Fig. 1.

Fig. 2.

Fig. 3.

The most critical problems in the eastern regions (TRA Northeast Anatolia, TRB Middle East Anatolia, and TRC Southeast Anatolia) that have the lowest values are the difference in Internet use between men and women, the low rate of computer use, and the low R&D expenditures.

This article proposes a new digital development monitoring index to understand and observe the digital development differences in Türkiye. With the help of the current literature, a comprehensive monitoring index is presented. The monitoring tool consists of three different levels of digital divide: Level 1 represents material access and skills (van Dijk 2005), Level 2 represents usage in terms of variety and regularity (DiMaggio, Hargittai 2001; Warschauer 2003; van Deursen, van Dijk 2018), and Level 3 represents the outcome of Internet usage (van Deursen, Helsper 2015) and economic outcomes; specifically, innovation is considered the primary outcome (Shakina et al. 2021). As a further divide, all individuals in various socio-demographic and socio-economic conditions (ITU 2017; EU 2018; OECD 2021a,b,c) should be considered equal for digital inclusion.

The indicators encompassing all aspects of the digital divide are analysed for Türkiye. The maps indicate the digital divide at regional level (NUTS1 – 12 Sub-Regions) and city level (NUTS3 – 81 provinces); however, disparities are more significant at the city scale. While regions in Türkiye generally converge in terms of access and use of technology (the first and second-level digital divide), the tangible benefits of technology (the third-level digital divide) are divergent. The primary findings of the study indicate the reasons, which are uneven digital skills, socio-economic and demographic divides, and R&D expenditure as a percentage of GDP. All of them gain imbalanced scores between the regions. To enhance digital inclusion, Türkiye needs to focus on fostering the ability to use technology, supporting gender equality, encouraging R&D expenditures, and supporting initiatives, especially ICT initiatives.

This study has developed a comprehensive index containing all digital divide aspects. In addition, the results demonstrate how digital inequalities exist in regions within the same country. Similar to the previous studies, the findings show that there are many determinants of outcomes of technology usage. This study’s determinants are uneven digital skills, socio-economic and demographic factors, and R&D expenditure. However, the study has some limitations owing to a lack of precise or equivalent statistical data at a regional scale, such as individuals’ basic digital and software skills, ICT usage in enterprises, e-government usage, and digital literacy. Despite these study limitations, digital development differences at the regional scale have been revealed. More than that, the regional strengths and weaknesses have been exposed. Future studies can be focused on the physical effects of digitalisation in cities.