1. Introduction
Today, the Internet and mobile phone influence various aspects of our lives. With only a few clicks, one can find the information required and can disseminate information with an unspecified number of the general public. Additionally, individuals build networks with others through SNS (social network service) and by communicating and exchanging ideas, sometimes even creating a new identity of their own. The development of the Information and Communication Technology (ICT) most importantly accomplishes a change of communication form to enable communication that transcends time and space.
The forthcoming new digital technologies and communication platforms of the ‘Fourth Industrial Revolution’ (or called Digital Transformation) era (the Fourth Industrial Revolution can be defined as an era of technology convergence in which the boundary between physical space, digital space and biological space is diluted based on the digital revolution (Third Industrial Revolution), such as IT technology), will not deliver much more drastic changes than hitherto accomplished changes to our living and consciousness, economy and culture, but also to the national policy decision-making.
In the transitional trend, this study highlights Digital Social Innovation (DSI) that is defined as a form of social and cooperative innovation where innovators, activists, and organizations utilize digital technology in order to produce ideas and deliver solutions for demanded social needs at a rate that is thought to be unimaginable before the Internet Renaissance [
1,
2]. In particular, the DSI can be regarded as the collective results of digital governance that refers to the new mechanism that administrates market and society that are based on convergence of digital technology, in which people, the state, and firms form new relations by utilizing the ICT, and decide and manage the future of the communities [
3,
4,
5].
The DSI is an attempt of guaranteeing citizen participation in policy-making and its execution process by utilizing digital technology, such as crowd-sourcing (engaging consumers in production and service processes and share profits with the participants) and online petitions. A series of process of the DSI, which enhances democracy of the policy per se as well as policy formulation process through active engagement of the citizens, has effects on rendering ‘throughput legitimacy’ (including the condition and efficacy of governance, as well as the process and method of participation) to be strengthened [
6,
7]. Moreover, it is seen that the DSI encompasses the following six sectors (categories): ‘Collaborative Economy’, ‘New Ways of Making’, ‘Open Democracy’, ‘Open Access’, ‘Awareness Network’, and ‘Funding, Acceleration and Incubation’ [
1,
8].
Meanwhile, as digital social innovation (DSI) is of a social innovation, it requires taking an approach from the aspect of the socio-technical system transition [
9]. As the previous discussions of digital social innovation have focused on utilizing digital technologies that have been used in industry to solve social challenges, they have only been expanded in short term and local views [
4,
10]. However, this study underlines that the introduction of the perspective of the socio-technical system transition is required in order to fundamentally solve social problems by making use of digital technology. This is not brought to sufficient attention from previous studies. Unlike researches in some related academic fields, this study tries to reveal the fact that the DSI can be more properly constituted by the multi-lateral and holistic approach of the socio-technical system transition that goes beyond the top-down and fragmentary approach of existing innovation researches. This study aims to determine how certain factors that comprise the results of DSI (digital social innovation) are composed in a combined way while considering the contexts of the socio-technical system transition and the digital governance. Particularly, it sets up the following research questions:
Which of the causal variables—democracy and e-participation (in politics); GDP and business environment (in economy); social capital and social expenditure (in society); and, ICT development and patent applications (in technology)—that have been set for the outcome condition of the DESI (Digital Economic and Social Index), are significant through the necessary condition verification?
Which of the causal sets/arrangements for the outcome condition of the DESI explain the high combined causal relationship with digital social innovation?
Subsequently, it delivers the findings of the causal sets of digital social innovation, while using the fuzzy-set multiple conjunctural analysis (STATA (Software for Statistics and Data Science) version 12.0 used) that can analyze and explain the collective contexts (arrangements) for this research, rather than the results between individual variables and of linear relationship that the existing qualitative methods highlight. For this study, thirty-four OECD countries (valid data out of all thirty-six OECD member countries, excluding Netherlands and Lithuania, where the data are not available) were included as the cases by extracting the relevant and recent data (eight variables) on the six categories of DSI, as described above.
This paper consists of four sections: in
Section 2, it discusses the theoretical background of socio-technical system transition, digital governance, and digital social innovation. In
Section 3, it clarifies the rationale for using the method of the fuzzy-set analysis and organizes the measurement frameworks, including the variable composition of outcome and causal sets. In
Section 4, it provides the findings of fussy-set multiple conjunctural analysis, including the necessary and sufficient conditions, and also the implications and conclusions.
3. Materials and Methods
This research examines thirty-four OECD countries through comparative analysis by utilizing fuzzy-set multi-conjunctural analysis. The fussy-set analysis method consists of the fussy-set multiple conjunctural analysis and the fussy-set ideal type analysis [
34,
35] (the fuzzy-set analysis is a methodology developed from qualitative comparative analysis). Without setting limits within the existing traditional two membership scores, 1 or 0, fuzzy set analysis, which permits various fuzzy membership scores ranging between 0 and 1, not only represents the partial membership but also the differences in degree [
34,
36].
Moreover, in the causal relationship of causes and consequences (effects) performed in the conventional quantitative analysis, a certain outcome (consequences) seems to be not clearly identified with a certain cause. In other words, it was difficult to distinguish between a correlation and a causal relationship [
34,
37]. However, it is an important advantage to be able to overcome such problems through the fussy-set multiple conjunctural analysis that defines the relationship between cause and outcome as a necessary condition or/and a sufficient condition. In addition, it can be a powerful tool in analyzing causal complexity in medium case studies [
37].
In fact, this study underlines the fact that we can clearly analyze the combinations of causal variables (conditions) for small to medium-sized (usually 15 to 50) cases through the fussy-set multiple conjunctural analysis. The fussy-set multiple conjunctural analysis is able to analyze the combination of causal conditions when compared to the existing qualitative methods, without an assumption of the independence between variables and the linear relationship. As a result, through the fussy-set multiple conjunctural analysis, this study has the strength of analyzing causal relationship by integrating not only the unilineal effects by certain variables, but also multiple causal conditions [
38,
39,
40].
As shown in
Table 1, below, this study with fussy-set multiple analysis presumes the Digital Economy and Social Index (DESI, published by the European Commission) as the outcome set. The DESI (and Intl.-DESI) is a composite index, exploring how the digital economy (economic activity conducted through digital technologies and innovations) can benefit all of society [
41,
42]. It summarizes the relevant indicators on the digital performance and digital competitiveness of Europe and major countries around the world [
41,
43]. The DESI (and Intl.-DESI) is composed of five principal policy areas with overall thirty-four indicators, as follows: ‘connectivity’ (including fixed broadband and mobile broadband), ‘human capital’ (including basic skills and internet use), ‘use of internet service’ (including citizens’ use of content), ‘integration of digital technology’ (including business digitization and e-commerce), and ‘digital public services’ (including e-Government and e-Health) [
44].
For the causal sets, the eight variables were selected, respectively, based on the four areas (‘Politics’, ‘Economy’, ‘Society’, and ‘Technology’). As
Table 1 shows, this study tried to set up the four-stranded areas of the system of socio-technology transition with each of the eight main categories of DSI (digital social innovation) being rooted and activated. Particularly, the area of technology is characterized by a cross-cutting one, in which all of the six categories (variable (1)~(6)) of DSI can be rooted.
The first P (politics) area is characterized by both ‘democracy index’ and ‘e-participation index’. The democracy index was produced by the Economist, having over 70 years’ experience, which consists of the five categories: ’electoral process and pluralism’, ‘civil liberties’, ‘the functioning of government’, ‘political participation’, and ‘political culture’ [
45]. The e-participation index was developed from United Nations as additional index to the UN E-Government survey with ‘e-information’ with public information and access to it, ‘e-consultation’ to engage people in public policies and services, ‘e-decision making’ for policy options, and co-production through co-design with the citizens [
46].
The second E (economy) area is composed of ‘GDP’ and ‘business environment’. The variable of GDP underlines the size of economy (USD, constant prices, from OECD statistics), and the business environment represents the quality of economy. In particular, the latter variable as one of ‘Legatum Prosperity Index’, as developed from the Legatum Institute, measures a country’s entrepreneurial environment, its business infrastructure, barriers to innovation, and labor market flexibility [
47]. Those variables that are aligned in the DSI categories are ‘Collaborative Economy’, ‘New Ways of Making’, and ‘Funding/Acceleration/Incubation’, reviewed in
Section 2.2.
The third S (society) area consists of the both variables of ‘social capital’ and ‘social expenditure’. As the social capital represents social networks between people and different groups, and also the level of trust, it can be a substantial catalyst in social innovation that utilizes the power of the citizen, who is a party to social problems [
3,
50]. The variable of social capital as one of ‘Legatum Prosperity Index’ measures the strength of personal trust (relationships), social network support, and social norms [
47]. The variable of social expenditure stands for net social expenditure, including public social welfare expenditure, mandatory private social expenditure, voluntary private social expenditure, and tax benefits; excluding tax burden (from OECD statistics).
The final T (technology) area includes both variables of ‘ICT Development Index’ (IDI) and ‘patent applications’. IDI from the International Telecommunication Union (ITU) has ‘ICT readiness’, which is accessed by the quality, level, and access to ICTs and network infrastructure, ‘ICT use’ assessed by the intensity level of ICTs in a given society, and ‘ICT impact’ implied in the results of ICT use [
48]. The ‘patent application’ variable represents the context that patents are an integral part of global technology innovation and an important intangible asset that accounts for most of the enterprise value, and also that the intellectual property that encompasses patents plays a role of a pioneer of breakthrough of technology innovation and of breakwater of technology protection [
49].
This study focuses on the cases of thirty-four OECD countries out of all thirty-six member countries, in which the above eight variables (the most recent) data in the four areas (described in
Table 1 above) are available. It attempted to collect the most recent data that may best show the characteristics of the eight variables.
4. Findings and Results
This study develops the analysis model that consists of the outcome set of the DESI (Digital Economy and Society Index) of thirty-four OECD countries, and of the causal set of the eight variables in the four categories. All of the variables are marked in uppercase, as shown below. Lower case indicates the complement of the set (d, p, g, b, s, e, i, r).
Model: DESI = D (Democracy Index) + P (E-participation Index) + G (GDP) + B (Business Environment) + S (Social Capital) + E (Social Expenditure) + I (ICT Development Index) + R (Intellectual Property).
In this study, the Y-consistency and N-consistency test (“settest (yvv yvn”) was conducted to look at the necessary conditions (if the outcome set is a subset of the causal set, the fuzzy membership score of the causal set is higher than the fuzzy membership score of outcome set). The concept of consistency indicates to what extent the relationship between causal conditions (sets) and outcome conditions (sets) is significant and influential. It shows whether empirical connections, such as ‘significance’, may or may not be worthy of the researcher’s attention, and, in particular, is a measure of how well the researcher’s theoretical arguments can be supported. In the case of the necessary condition and the sufficient condition, the value criteria, 0.65~0.8 is generally used.
As shown in
Table 2, the seven variables (D(0.909), P(0.826), G(0.967), B(0.841), E(0.812.), I(0.927), and R(0.968) were satisfied with the principles of the necessary conditions (‘Y-cons > 0.8’ (benchmark proportion, ‘almost always’) and ‘Y-cons > N-cons’), and the one variable (S(0.746)) was found to agree with ‘Y-cons > 0.65’ (benchmark, ‘usually’).
As shown in
Table 3 (the truth table of the causal set for the outcome set), below, there are seven sets (of the total 256 sets) that satisfy the condition of the existence of cases and the significance level,
p < 0.05. They are also above Y-consistency 0.80 (benchmark proportion, ‘almost always’). In other words, according to this criterion, a total of seven arrangements (sets) indicating a high outcome upper case set are identified (‘Y’ = DESI (Digital Economy and Society Index). Among the seven sets (arrangements), there are three sets with more than two case countries (best fit), respectively.
For example, the first arrangement is ‘DPGBSEIR’ (high democracy (D), high e-participation (P), high GDP (G), high business environment (B), high social capital (S), high social expenditure (E), high ICT development (S), and high patent applications (R)), the best fits of which (cases) are AUS, DEN, GER, NOR, and SWE (Y-cons. 0.963).
The second is ‘DPGBsEIR’ (high democracy (D), high e-participation (P), high GDP (G), high business environment (B), low social capital (s), high social expenditure (E), high ICT development (S), and high patent applications (R)), the best fits of which (cases) are NZL, ISR (Y-cons. 0.938). Additionally, the arrangement of ‘DpGBSEIR’ (high democracy (D), low e-participation (p), high GDP (G), high business environment (B), high social capital (S), high social expenditure (E), high ICT development (S), and high patent applications (R)) has the two case countries: AUT, SUI (Y-cons. 0.930).
This study carried out the verification of both Y-consistency and N-consistency in order to examine the sufficient condition (the fuzzy (membership) score of the outcome set > the fuzzy score of the causal set). This paper follows the usual criteria: the benchmark as 0.80 and the significance level of 0.01 (
p < 0.01). As described in
Table 4 below, in short, both sets were identified, satisfying the Y-consistency and N-consistency validation of the ‘sufficient condition’ for the outcome set (DESI), and also include the best fits (a total of eight country cases).
As a result, the five causal sets, DPGBSEIR, and DPGBsEIR were found as the significant arrangements. Through the ‘reduction process’ (minimum configuration reduction sets that pass both Y and N consistency tests), the arrangements are reduced as the one causal set: D × P × G × B × E × I × R. The solution consistency of the causal set of the sufficient condition is 0.947 and the total coverage (explanatory range, strength) is 0.659 (see
Table 5). The result of solution consistency, 0.947, means that the empirical connections (‘significance’) are very worthy (in comparison with 0.65~0.8 generally used). Additionally, the total coverage that explains the importance of empirical relevance and verifies how the cases that were included in the study can be explained by the theories (assumptions) of this study, are significantly high (explanatory range 0.659).
DESI = D × P × G × B × E × I × R = (Democracy Index × E-participation Index × GDP × Business Environment × Social Expenditure × ICT Dev. Index × Patent Applications).
To summarize, the causal set of the sufficient condition explains, as follows. Among the OECD member countries, if ones have high level of democracy and e-participation, high GDP and business environment, high social expenditure, and high level of ICT development and patent applications, they are highly likely to achieve a sufficient level of digital social innovation.
As shown in
Figure 3, below, the analysis of each case can be more concretely illustrated through the scatter plot that places the fuzzy score of the outcome set, DESI (Digital Economy and Society Index), on the X axis, and the fuzzy score of the causal set of the sufficient condition (‘DPGBEIR’) on the Y axis. As a result, except for a small number of cases, most of the cases are located at the upper left corner of the diagonal line, showing high consistency. It shows that the typical scatter plot of sufficient condition of the fuzzy-set. The countries that best describe the causal set of the sufficient conditions are the UK, Sweden, and Australia, which are located in the right edge of the diagonal.
5. Discussion and Conclusions
This study reviewed the influencing factors of multidisciplinary digital social innovation (DSI) beyond the existing top-down and static approaches of the innovation system theory through the socio-technical system transition theory. Essentially, it concludes that the seven factors in the four areas (politics, economy, society, and technology) of OECD countries—the political maturity, the active e-participation, the large scale of economy (market), the high quality of business environment, the large size of social expenditure, the high development level of ICT, and the multi-faced technology and intellectual property innovation—have a very significant impact on the DSI in a combined way. The OECD countries with high levels of the seven factors in the four areas, reflecting on the contexts of the socio-technical system transition (STST) and the digital governance, can be seen to be more effectively completing the DSI. In particular, the result of combined arrangement explains that the DSI can be more properly characterized by the multi-level (particularly meso- and micro-levels of STST) and structured approach of STST that goes beyond the fragmentary approach of existing innovation theory.
Meanwhile, there comes a suggestion from a slightly different perspective that in the combination sets satisfying the sufficient condition, the seven variables of all eight ones show the high bonding conditions, which indicates a significance of the socio-technological transition, while the S (social capital) variable is excluded from the combination for the sufficient condition. In fact, as described in
Table 2 previously, this paper notes that the consistency value of the social capital (S) variable (0.746) is lower than that of the other seven variables. Additionally, the social expenditure (E) variable, which is another variable of the social domain, the result of the consistency value (0.812) is the second lowest one, followed by the social capital variable. Moreover, the statistical significance level of the both variables (S, E) in the social domain are relatively lower than the other six variables in the other three domains (politics, economy, and technology); the
p-values of S and E are 0.075 and 0.013, respectively, on the other hand, the rest six variables satisfy the criteria,
p-value < 0.01.
This study argues that these results suggest that the existing social factors are relatively ineffective in comparison with other domains’ ones (politics, economy, and technology) in the DSI that promotes social problem solving through technological innovation. In particular, it can be seen that social capital—referring to the ability of individuals to mobilize the resources of others through social relations—as a result of accumulation in the existing systems may have little effect on the DSI. On the other hand, as the DSI evolves, it shows the possibility that social capital might be able to be constituted in other dimensions by becoming a new aggregation in the ‘fourth industrial revolution’ (digital transformation) era [
26]. As a result, this study assumes that these results describe an emerging aspect of the mutual structure of social change through the innovative development of new technologies. For example, the emergence and popularization of new technologies change the shape and form of people’s lives, which in turn changes people’s thinking, transforms into a new community of people with the corresponding way of thinking, and subsequently again requires a new technology. It may make the overall structure a full circulation.
In this context, as compared to the citizens and citizenship in the existing societies, the discussions about ‘digital citizen’ and ‘digital citizenship’ in the new era of the fourth industrial revolution will need to be considered more multi-dimensionally. In the digital (information) society, where the massive amount of information that corresponds to various topics and interests can be circulated quickly and at low cost to almost everyone, the mindset and behavior of citizens will be greatly transformed. This will ultimately lead to the search for new possibilities for the DSI through digital citizens who are actively engaged in social interaction in online/offline by using various new technologies in the digital world with interconnected rights, responsibilities, and opportunities for life. In line with it, Dufva and Dufva [
27] argue the conception of ‘digi-grasping’ that can empower people to outline the relationship with the digital world, and also understand the choices and motivations behind current digital structures and create new structures.
In the meantime, this study has limitations in that it could not reveal (further) implications of the result cases (of OECD countries), including the best fit countries of the arrangement of causal set; UK, Sweden, and Australia, which are located on the right edge of the diagonal of the scatter plot, as illustrated in
Figure 3. In addition, while the results of this paper, based on the relevant key data (eight variables) from 34 of 36 OECD countries, are sufficient to be generalized, among non-OECD countries, there is also the possibility of different results from this study. Therefore, the limitations of the study can be supplemented by subsequent studies to provide more concrete implications. It is also required for further studies to apply the analytical framework of this study to non-OECD countries.