ABSTRACT
Most theoretical and empirical explanations of the generation of digital divides have been integrated into the resources and appropriation theory, which proposes a sequential model reflecting a socially unequally distributed digital divide. The unequal social distribution is reflected in internet use that is sequentially influenced by motivations/attitudes, physical access, and digital skills. We extend the sequential model by exploring the complementary role of information security concerns in producing the digital divide. Using a predictive approach, we tested a comprehensive partial least squares-structural equation model with data from a European Union survey, finding that information security concern is another significant determiner of the digital divide. Heterogeneity in social internet appropriation can be summarized in social mechanisms explained by education and age among well-educated Europeans, and by country digital development among less well-educated Europeans. We conclude with a discussion of theoretical and policy implications of our findings.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Data collection details are available at: https://circabc.europa.eu/ui/group/4f80b004-7f0a-4e5a-ba91-a7bb40cc0304/library/8bc71641-bd53-4039-b9f0-71d87822749d/details.
2 When variables are categorical, the identification of factors describing the interdependence between indicators cannot be assessed by applying principal component analysis (PCA) or factor analysis (FA). A better choice is MCA, a multivariate method of analysis used to describe, explore, summarize, and visualize the interdependence among a set of indicators contained within a data table of n individuals described by q categorical variables. It can be seen as an analogue of PCA for categorical variables (rather than quantitative variables). MCA reduces the dimensionality of a table and the new dimensions can be understood as ‘latent’ characteristics. The coordinates (scores) are linear combinations of the categorical indicators. The dimensions are defined to maximize the variability of the original indicators. Therefore, with few dimensions it is possible to retain the original variation, with the principal benefit of reducing dimensionality. From a practical point of view, MCA dimension coordinates can be interpreted as an optimal numeric scale where each coordinate represents individual scores. For MCA, Greenacre (Citation1993) has shown that the scores of individuals form an optimal scale when those scores are far apart, thereby maximizing differences between individuals.
Additional information
Funding
Notes on contributors
Giuseppe Lamberti
Giuseppe Lamberti is a postdoctoral researcher at the Autonomous University of Barcelona (UAB), Spain. He presented his PhD thesis in 2015 at the Technical University of Catalonia (UPC), Spain. His research interests cover methodological research related to traditional multivariate analyzes, structural equation modeling (SEM), and partial least squares (PLS) path modeling algorithms, with a particular interest in the issue of heterogeneity in SEM.
Jordi Lopez-Sintas
Jordi Lopez-Sintas is a full professor in the Department of Business of the Autonomous University of Barcelona (UAB), Spain. His research focuses on consumer research, the sociology of consumption, culture consumption, media studies, qualitative research, and information, communication, and society. He is involved in several projects that concern culture, leisure, and the digital technologies, and has published numerous articles on various facets of culture, leisure, and the digital technologies.
Jakkapong Sukphan
Jakkapong Sukphan holds a PhD from Autonomous University of Barcelona (UAB), Spain. Currently he holds a position as lecturer in marketing in Maejo University, Chiang Mai, Thailand.