Towards a Composite Index for Digital Maturity: an Unsupervised Machine Learning Approach

EasyChair Preprint no. 10903

Abstract

In recent years, a considerable amount of research has explored the negative effects associated with the use of ICTs and linked it to several health issues that can pose consequences at the societal level as well as on an individual level. Despite the negative effects, the use of ICTs also provides a range of benefits and researchers are in particular interested in how we can help young people with obtaining a beneficial digital engagement with ICTs. Motivated by the advantages ICTs bring, a new multidimensional concept named digital maturity has been proposed. Digital maturity consists of three overall capacities, which are useful in a digital context. The first capacity is about making autonomous choices about using mobile devices and exercising autonomy. The second capacity involves digital literacy, individual growth, digital risk awareness, and support-seeking regarding digital problems. Finally, the third capacity consists of the regulation of negative emotions and aggressive impulses, respect towards others, and digital citizenship. To measure digital maturity based on these ten dimensions, a composite index (CI) named digital maturity inventory (DIMI) has been constructed. The DIMI can be used to gain an overview of the aggregated level of digital maturity in young people in a country or region by applying experts’ opinions on how much weight each dimension should be given. The challenge that exists with using expert’s proposed weights for the dimensions in a CI is that they not always are in line with their relative importance. In this paper, we examine the relative importance of the ten dimensions from a data-driven perspective using real-world data with an interest in optimizing the weights used to predict young peoples’ digital maturity. Our result demonstrates a misfit between experts’ opinions and the relative importance. Thus, based on our empirical evidence, we propose that an adjustment of the weights applied for the DIMI needs to take place.

Keyphrases: Children's Digital Maturity, composite index, digital maturity, ICTs, Unsupervised Machine Learning

@Booklet{EasyChair:10903,
  author = {Rikke Nyland Christensen and Aqib Siddiqui and Konstantina Valogianni and Arnd Florack and Marco Hubert},
  title = {Towards a Composite Index for Digital Maturity: an Unsupervised Machine Learning Approach},
  howpublished = {EasyChair Preprint no. 10903},

  year = {EasyChair, 2023}}