Descriptive statistical analysis results

Using PISA 2022 data, we approximated the average daily time students spend on the Internet across European countries. The results, depicted in Fig. 2, reveal notable variations in internet usage patterns, with Albania reporting the highest average at 3.18 h per day, closely followed by Bulgaria and Ukraine, while Switzerland records the lowest at 1.63 h. These differences may reflect disparities in digital access, educational policies, and cultural attitudes toward screen time.

Fig. 2
figure 2

Approximation of average daily time spent on the Internet, PISA 2022.

Items measuring digital skills were composed into a unified score, with thorough checks for internal consistency conducted for each country. Remarkably high Cronbach’s alpha values were observed across all countries, ranging from 0.870 for Italy to 0.969 for Albania. To ensure the robustness of the index, additional analyses were carried out, specifically an Exploratory Factor Analysis (EFA) to determine whether the items reflected a unidimensional construct or multiple dimensions of digital skills. The results supported the use of a single aggregated index, as a single factor was identified in most countries. In cases where two components were extracted, the first component was clearly dominant, underlying the digital skills index. This conclusion is viable since the first component accounted for substantially more variance than the second, capturing most of the total variance in the data. Moreover, using the eigenvalue criterion, we concluded that, in countries where a second component was detected, the first component had an eigenvalue far greater than the second one. For example, in Poland, the first component had an eigenvalue of 8.287, far exceeding the commonly used threshold of 1, while the second component had an eigenvalue of only 1.147. The significantly higher eigenvalue of the first component underscores its role as the principal factor. Additionally, the component matrix revealed that in countries where a second factor was detected, all fourteen items exhibited strong loadings (absolute values > 0.6) on the first component, demonstrating their alignment with this dominant factor. Cross-loadings on the second component were minimal, reinforcing that the items were primarily associated with the first component. This interpretation was further supported by the scree plots, which showed a sharp decline in eigenvalues after the first component, followed by a levelling-off pattern. This is a characteristic indicator of a single dominant factor, confirming the unidimensionality of the scale.

The results highlight notable variations, with Denmark achieving the highest score of 44.88, followed closely by Ireland (44.7) and Lithuania (44.62). Several other countries, including Finland, Latvia, the United Kingdom, and Croatia, also demonstrate strong digital proficiency, scoring above 43. Conversely, Albania reports the lowest score of 39.55, with Slovakia, Austria, and Sweden also falling below 42. These scores reflect students’ abilities in key digital tasks such as information searching, content creation, data management, and privacy protection, with higher scores indicating greater proficiency. Figure 3 illustrates the digital skills scores among students in the surveyed countries. Interestingly, Albania records the highest estimated daily internet usage but the lowest digital skills proficiency, suggesting that time spent online does not necessarily translate to digital competence. A similar trend is observed in Bulgaria, Slovakia, and Ukraine, underscoring the importance of not just access but also the quality and nature of digital engagement.

Fig. 3
figure 3

Digital skills scores in European adolescents, PISA 2022.

Figure 4 illustrates the varying levels of student interest in digital resources across European countries, measured on a scale from 4 to 12. Notably, Albania has the highest level of interest toward digital resources among students in the studied countries, scoring 8.88 on a scale from 4 to 12. Romania and Italy follow closely behind. Countries such as Malta, Bulgaria, and Latvia display moderate interest levels. In contrast, Iceland has the lowest interest score at 6.84 among the countries listed. This distribution highlights the diverse engagement levels with digital resources among 15-year-old students across Europe. Interestingly, countries with high digital skills scores, such as Finland, show only moderate interest in digital resources (7.39). This suggests that digital proficiency does not necessarily correlate with a high enthusiasm for digital tools, reinforcing the need to examine how technology is integrated into learning environments beyond simple access or interest levels.

Fig. 4
figure 4

Interest in digital resources, European countries, PISA 2022.

Figure 5 presents an analysis of the Distraction from digital devices scale, which ranges from 5 to 25, revealing substantial differences in the extent to which students are affected by digital distractions across the surveyed countries. The findings indicate that Malta reports the highest level of distraction among students, with a score of 19.11. Austria (16.56) and Germany (15.93) also show high levels of distraction, pointing to similar patterns where students experience frequent disruptions from digital devices. Other countries, including Slovakia (15.75), Romania (15.68), and the Czech Republic (15.58), also score above 15, indicating that digital distractions are a notable concern across various education systems. In contrast, Estonia records the lowest score (13.02), suggesting that students there are the least affected by digital distractions. Other countries with relatively low distraction levels include North Macedonia (13.15), Croatia (13.35), and Hungary (13.39), highlighting a trend where students in these regions may either have more structured digital habits, stricter regulations on device usage, or differing attitudes toward digital engagement during study hours.

Fig. 5
figure 5

Distraction from digital resources, European countries, PISA 2022.

Figure 6 presents an overview of the self-reported life satisfaction scores among 15-year-old students across the surveyed European countries, measured on a scale from 0 to 10. The results reveal notable variations in students’ overall well-being, reflecting differences in social, economic, and cultural contexts. Albania stands out with the highest mean life satisfaction score of 8.01, suggesting that students there report a relatively high level of contentment with their lives. North Macedonia (7.66) and Montenegro (7.54) follow closely behind, while Romania (7.52) and Serbia (7.46) also rank among the countries with high student life satisfaction. These findings may be influenced by strong family support networks, cultural perceptions of happiness, or lower levels of academic pressure compared to other nations. Students in Finland (7.41), Croatia (7.38), and Ukraine (7.34) report moderate levels of life satisfaction, aligning with the middle range of the surveyed countries. At the lower end of the scale, the United Kingdom reports the lowest mean life satisfaction score of 6.04, indicating a comparatively lower level of well-being among students. This finding is consistent with previous research indicating that the UK reports the lowest average levels of children’s overall life satisfaction, as measured on a scale from 0 to 10, based on data from the Children’s Worlds survey (2016–19) (Ayllón et al. 2024a, 2024b). Other countries with relatively lower scores include Malta (6.22), Poland (6.25), and Germany (6.48), suggesting that students in these regions may face higher stress levels, concerns about future prospects, or differences in how they evaluate their life satisfaction.

Fig. 6
figure 6

Overall satisfaction, European countries, PISA 2022.

Figure 7 highlights notable gender differences in internet usage across European countries, revealing variations in the time spent online by male and female students. The data suggests that in most countries, male students tend to spend more time on the internet than female students, although exceptions exist. For instance, in Albania, where the overall internet usage is among the highest, male students report an average daily internet use of 3.37 h, while female students spend slightly less at 2.93 h. A similar pattern is observed in Bulgaria, Slovakia, and Ukraine, where male students consistently report higher usage than their female peers. In contrast, Ireland shows a reverse trend, where female students spend more time online than males, suggesting potential differences in digital engagement based on gender-specific interests and online activities. Furthermore, the range of internet usage varies significantly across countries. Switzerland exhibits the lowest internet use overall, with female students spending an average of 1.5 h daily and male students 1.76 h. Other countries with low internet usage include Belgium, Germany, and Denmark, where both genders report relatively restrained online activity. On the other hand, in high-usage countries like Romania and Malta, students of both genders spend well above 2 h daily on the internet.

Fig. 7
figure 7

Time spent on the internet, Males, Females, European countries, PISA 2022.

Figure 8 illustrates significant variations in digital skill scores among male and female students across different European countries, highlighting both regional disparities and gender-based differences in digital literacy. Overall, countries such as Ireland, Latvia, and Lithuania report the highest digital skill scores for both genders, with Ireland leading at 45.26 for female students and 44.14 for male students. Similarly, Latvia and Lithuania demonstrate consistently strong self-reported digital skills, suggesting a robust digital education framework in these countries. Conversely, countries like Bulgaria and Greece exhibit comparatively lower digital skill scores, with Bulgaria showing one of the most pronounced gender gaps -female students score 42.78, whereas male students lag behind at 39.92. Gender differences in digital skills are evident across the dataset. In some countries, such as Iceland, female students outperform their male counterparts, suggesting that digital education or engagement strategies might favor female students in these regions. However, the reverse trend is observed in countries like Denmark and the UK, where male students exhibit higher digital skill scores compared to females. This variation could be influenced by differing educational opportunities, societal attitudes toward technology use, or personal interest in digital activities. While the differences are not uniform across all countries, in most countries female students tend to score slightly higher in digital skills than male students in most European countries. However, the presence of exceptions underscores the complex interplay of cultural, educational, and socio-economic factors shaping digital literacy among students.

Fig. 8
figure 8

Digital Skills, Males, Females, European countries, PISA 2022.

Inferential statistical analysis results

ANOVA results for differences between time spent on the internet and digital skills

A one-way ANOVA was performed to examine whether there was a statistically significant difference in estimated internet time and digital skill scores between female and male students.

All requirements are met except for the assumption of equal variance, so a test assuming unequal variances was employed for all countries and tests of equality of means. The analysis revealed distinct patterns, which were organized into categories based on the observed trends in gender differences in internet usage and digital skills. In all countries, except for Belgium, differences in both internet time and digital skills scores between female and male students are detected. In the first category of countries, male students spent significantly more time on the internet, while female students scored higher in digital skills. This pattern suggests that in these countries, although male students were more engaged in online activities, females exhibited stronger skills in navigating the digital landscape. The countries in this cluster include Albania, Greece, Italy, Malta, Poland, Romania, Slovakia, Slovenia, Sweden, and Ukraine. Specifically, the results of the statistical analysis for these countries are:

  • Albania: t 12.472 = −11.202, p < 0.001 for internet time and t 11.607 = 11.005, p < 0.001 for digital skill scores.

  • Greece: t 73.652 = −23.754, p < 0.001 for internet time and t 68.397 = 21.196, p < 0.001 for digital skill scores.

  • Italy: t 392.730 = −9.220, p < 0.001 for internet time and t 363.770 = 28.178, p < 0.001 for digital skill scores.

  • Malta: t 2.750 = −3.476, p < 0.001 for internet time and t 2.490 = 4.957, p < 0.001 for digital skill scores.

  • Poland: t 258.834 = −26.008, p < 0.001 for internet time and t243.759 = 16.263, p < 0.001 for digital skill scores.

  • Romania: t 117.428 = −29.131, p < 0.001 for internet time and t112.079 = 20.922, p < 0.001 for digital skill scores.

  • Slovakia: t 34.821 = −19.855, p < 0.001 for internet time and t32.935 = 10.928, p < 0.001 for digital skill scores.

  • Slovenia: t 14.853 = −15.697, p < 0.001 for internet time and t14.471 = 4.998, p < 0.001 for digital skill scores.

  • Sweden: t 79.747 = −20.260, p < 0.001 for internet time and t76.656 = 10.211, p < 0.001 for digital skill scores.

  • Ukraine: t 109.177 = −15.839, p < 0.001 for internet time and t102.279 = 23.468, p < 0.001 for digital skill scores.

In a second cluster of countries, male students not only spent significantly more time on the internet but also outperformed females in digital skills. This suggests that males in these countries exhibit both higher engagement with the internet and greater proficiency in digital environments. The countries in this cluster include the Czech Republic, Denmark, Estonia, Finland, Germany, Hungary, Iceland, Switzerland, and the United Kingdom. Specifically:

  • Czech Republic: t 76.056 = −37.037, p < 0.001 for internet time and t 70.188 = −9.967, p < 0.001 for digital skill scores.

  • Denmark: t 40.853 = −24.939, p < 0.001 for internet time and t 37.668 = −3.236, p < 0.001 for digital skill scores.

  • Estonia: t 10.996 = −7.774, p < 0.001 for internet time and t 10.747 = −0.495, p < 0.001 for digital skill scores.

  • Finland: t 43.103 = −22.875, p < 0.001 for internet time and t 41.887 = −4.623, p < 0.001 for digital skill scores.

  • Germany: t 454.887 = −72.149, p < 0.001 for internet time and t 413.709 = −12.958, p < 0.001 for digital skill scores.

  • Hungary: t 70.618 = −10.288, p < 0.001 for internet time and t 65.303 = −2.737, p < 0.001 for digital skill scores.

  • Iceland: t 2.760 = −6.732, p < 0.001 for internet time and t 2.894 = −1.809, p < 0.001 for digital skill scores.

  • Switzerland: t 56.378 = −24.373, p < 0.001 for internet time and t52.496 = −3.824, p < 0.001 for digital skill scores.

  • UK: t 56.378 = −29.845, p < 0.001 for internet time and t428.830 = −6.639, p < 0.001 for digital skill scores.

In the third category, statistically significant gender differences were observed in both internet time and digital skills, but the magnitude of these differences was smaller compared to the previous categories. This indicates that while gender disparities exist, they are less pronounced. The countries in this subcluster included Austria, Bulgaria, Croatia, Latvia, and Ireland:

  • Austria: t 57.877 = −18.633, p < 0.001 for internet time and t 55.345 = 4.394, p < 0.001 for digital skill scores.

  • Bulgaria: t 29.232 = −8.595, p < 0.001 for internet time and t 27.599 = 18.595, p < 0.001 for digital skill scores.

  • Croatia: t 27.676 = −10.768, p < 0.001 for internet time and t 26.911 = 12.430, p < 0.001 for digital skill scores.

  • Latvia: t 14.098 = −5.499, p < 0.001 for internet time and t 13.130 = 1.627, p < 0.001 for digital skill scores.

  • Ireland: t 55.089 = 3.092, p < 0.001 for internet time and t 51.519 = 13.668, p < 0.001 for digital skill scores.

Finally, in Belgium, a statistically significant difference in internet time was observed between female and male students, with male students spending more time online. However, no significant differences were found in digital skills scores between the two groups. Specifically, for estimated internet time: t 85.667 = −20.111, p < 0.001. This indicates that while males and females differ in the amount of time spent online, they possess similar digital skills. Figure 9 illustrates gender-based patterns in digital engagement and self-reported digital skills amongst students across European countries. The x-axis represents the difference in time spent on the internet between female and male students (F-M), while the y-axis denotes the difference in digital skills scores between female and male students (F-M). The top-right quadrant identifies countries where female students report higher digital skills and spend more time online than their male counterparts, with Ireland being the sole example. The bottom-right quadrant includes countries where female students exhibit higher digital skills but spend less time on the internet compared to males. Conversely, the bottom-left quadrant comprises countries where male students report higher digital skills and spend more time online. A clear trend emerges, indicating that, in nearly all countries (except Ireland), male students tend to spend more time on the internet than female students. However, notable cross-country variation exists in self-reported digital skills, with some countries exhibiting higher digital proficiency among female students and others favouring males.

Fig. 9
figure 9

Gender differences in self-reported digital skills and time spent on the internet among students across European countries, PISA, 2022.

Stepwise regression model results

We now provide the results of the stepwise regression model that was implemented for each European country. Differences are detected amongst countries and reported in this Section. Starting with six potential predictors of digital skills, i.e., interest toward ICT, distraction from digital devices, estimated time spent on digital devices, gender, parental education, and migration background that might theoretically be good predictors, a stepwise regression model was used to keep or reduce them for further exploration in further studies.

The final regression model for Albania indicates that parental education and immigrant background positively influence digital skills, while estimated time has a negative impact. The significance levels (p-values) for all predictors are below 0.05, indicating that these relationships are statistically significant.

  • Index Highest Parental Education: This predictor shows a statistically significant positive effect on digital skills. A higher level of parental education (measured in years of schooling) is associated with better digital skills in their children. The standardized coefficient (Beta = 0.100) indicates a small to moderate effect size.

  • Estimated Time: This predictor has a statistically significant negative effect on digital skills. More estimated time spent on certain activities is associated with lower digital skills. The standardized coefficient (Beta = −0.086) indicates a small negative effect size.

  • Index on Immigrant Background: This predictor shows a statistically significant positive effect on digital skills. An immigrant background (as defined by the OECD) is associated with higher digital skills. The standardized coefficient (Beta = 0.059) suggests a small positive effect size.

For Austria, all six predictors are proven statistically significant, with interest toward ICT being the most significant of all. The final regression model identifies multiple significant predictors of digital skills. Interest in digital activities is the strongest positive predictor, while estimated time and immigrant background have significant negative impacts. Parental education also positively influences digital skills, though to a lesser extent. Additionally, the gender variable highlights a small but significant gender disparity in digital skills, and distraction negatively affects digital skills. All predictors are statistically significant with p-values below 0.001, indicating robust relationships with digital skills.

  • Index Highest Parental Education: This predictor has a statistically significant positive effect on digital skills. A higher level of parental education is associated with better digital skills in their children. The standardized coefficient (Beta = 0.061) indicates a small effect size.

  • Estimated Time: This predictor shows a statistically significant negative effect on digital skills. More estimated time spent on certain activities is associated with lower digital skills. The standardized coefficient (Beta = −0.093) indicates a small negative effect size.

  • Index on Immigrant Background: This predictor has a statistically significant negative effect on digital skills. Having an immigrant background (as defined by the OECD) is associated with lower digital skills. The standardized coefficient (Beta = −0.091) suggests a small negative effect size.

  • Interest: This predictor demonstrates a stronger statistically significant positive effect on digital skills. Greater interest in digital activities is a robust predictor of better digital skills. The standardized coefficient (Beta = 0.177) indicates a moderate effect size.

  • Gender: This predictor has a statistically significant negative effect on digital skills. The gender variable indicates that there may be a gender disparity in digital skills, with one gender scoring lower on average. The standardized coefficient (Beta = −0.045) suggests a small negative effect size.

  • Distraction: This predictor has a statistically significant negative effect on digital skills. Higher levels of distraction are associated with lower digital skills. The standardized coefficient (Beta = −0.042) indicates a small negative effect size.

The final regression model for Belgium identifies several significant predictors of digital skills. Interest in digital activities is the strongest positive predictor, followed by parental education, estimated time, gender, and distraction. All predictors are statistically significant with p-values below 0.001, indicating robust relationships with digital skills. The positive influence of distraction is an interesting finding that may warrant further exploration.

  • Index Highest Parental Education: This predictor has a statistically significant positive effect on digital skills. Higher parental education (measured in years of schooling) is associated with better digital skills in their children. The standardized coefficient (Beta = 0.049) indicates a small effect size.

  • Estimated Time: This predictor shows a statistically significant positive effect on digital skills. More estimated time spent on relevant activities is associated with higher digital skills. The standardized coefficient (Beta = 0.036) suggests a small positive effect size.

  • Interest: This predictor has a stronger statistically significant positive effect on digital skills. Greater interest in digital activities strongly predicts better digital skills. The standardized coefficient (Beta = 0.137) indicates a small to moderate effect size.

  • Student (Standardized) Gender: This predictor has a statistically significant positive effect on digital skills. The gender variable suggests that there may be a gender disparity in digital skills, with one gender scoring higher on average. The standardized coefficient (Beta = 0.031) suggests a small positive effect size.

  • Distraction: This predictor also shows a statistically significant positive effect on digital skills. Despite its typically negative connotation, in this context, distraction might reflect multitasking abilities that enhance digital skills. The standardized coefficient (Beta = 0.050) indicates a small positive effect size.

In Bulgaria, predictors in the final model are estimated time spent on digital devices, interest toward ICT, gender, and parental education. The final regression model identifies several significant predictors of digital skills: higher parental education is positively associated with digital skills, time spent on digital devices negatively impacts digital skills, greater interest in digital activities strongly predicts higher digital skills, and a significant gender disparity in digital skills, with one gender scoring lower on average. All predictors are statistically significant with p-values below 0.001, indicating robust relationships with digital skills. Interest in digital activities and time spent on digital devices are particularly strong predictors, highlighting their critical roles in influencing digital skills.

In Croatia, potential predictors are interest toward ICT, parental education, gender, and estimated time spent on digital devices. The final model identifies that parental education, estimated time spent on digital devices, interest toward ICT, and distraction from digital devices are significant predictors of digital skills, with Interest being the strongest predictor. The significance values indicate that all these variables contribute meaningfully to the model.

  • Index highest parental education: This variable has a moderate, statistically significant positive effect on digital skills (B = 0.253, p = 0.036), indicating that higher parental education is associated with better digital skills in children.

  • Estimated time: This variable shows a large positive effect on digital skills (B = 0.424, p = 0.005), suggesting that more estimated time spent on digital activities is linked to higher digital skills.

  • Interest: This variable has a large and significant positive effect on digital skills (B = 0.682, p < 0.001), highlighting that greater interest in digital activities is a strong predictor of better digital skills.

  • Distraction: This variable has a small to moderate statistically significant positive effect on digital skills (B = 0.132, p = 0.020), implying that even though distractions can occur, they still correlate with improved digital skills, albeit to a lesser extent than other factors.

All predictors except migration background are proven statistically significant in the Czech Republic. The final model identifies several significant predictors of digital skills: higher parental education is positively associated with digital skills, more estimated time spent on digital activities negatively impacts digital skills, greater interest toward ICT predicts higher digital skills, higher levels of distraction are associated with lower digital skills, while there is a significant gender disparity in digital skills, with one gender scoring higher on average. All predictors are statistically significant with p-values below 0.001, indicating robust relationships with digital skills. Interest, estimated time, and gender are particularly strong predictors, highlighting their critical roles in influencing digital skills amongst students in the Czech Republic.

In Denmark, three (parental education, interest toward ICT, and gender) out of six predictors are included in the final model. The model identifies the following significant outcomes: higher parental education is positively associated with digital skills, though the effect size is relatively small, greater interest in digital activities predicts higher digital skills, with a moderate-to-large effect size, and a significant gender disparity in digital skills, with one gender scoring lower on average, though the effect size is relatively small to moderate. All predictors are statistically significant with p-values below 0.001, indicating robust relationships with digital skills. Interest emerges as the strongest predictor, highlighting its critical role in influencing digital skills.

  • Index Highest Parental Education: This predictor has a statistically significant positive effect on digital skills. Higher levels of parental education (measured in years of schooling) are associated with better digital skills in their children. The standardized coefficient (Beta = 0.045) indicates a small effect size.

  • Interest: This predictor has a statistically significant positive effect on digital skills. Greater interest in digital activities predicts better digital skills. The standardized coefficient (Beta = 0.169) indicates a small to moderate effect size.

  • Student (Standardized) Gender: This predictor has a significant negative effect on digital skills. The gender variable suggests that one gender scores lower on digital skills on average. The standardized coefficient (Beta = −0.079) indicates a small effect size.

In Estonia, the proposed model identifies several significant predictors of digital skills: parental education, time spent on digital devices, and distraction. Higher parental education is positively associated with digital skills, though the effect size is relatively small, more time spent on digital activities positively impacts digital skills, with a small effect size, greater interest in digital activities predicts higher digital skills, with a moderate effect size, and higher levels of distraction are associated with higher digital skills, with a small effect size. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. Among these predictors, interest in digital activities emerges as the strongest predictor, followed by distraction, parental education, and estimated time.

  • Index Highest Parental Education: This predictor shows a statistically significant positive effect on digital skills. A higher level of parental education (measured in years of schooling) is associated with better digital skills in their children. The standardized coefficient (Beta = 0.050) indicates a small effect size.

  • Estimated Time: This predictor also exhibits a statistically significant positive effect on digital skills. More estimated time spent on certain activities is associated with higher digital skills. The standardized coefficient (Beta = 0.068) suggests a small effect size.

  • Interest: This predictor demonstrates a statistically significant positive effect on digital skills. Greater interest in digital activities predicts better digital skills. The standardized coefficient (Beta = 0.159) indicates a small to moderate effect size.

  • Distraction: This predictor shows a statistically significant positive effect on digital skills. Higher levels of distraction are associated with higher digital skills. The standardized coefficient (Beta = 0.057) suggests a small effect size.

In Finland, the final regression model identifies several significant predictors of digital skills, namely parental education, interest toward ICT, and distraction. Higher parental education is negatively associated with digital skills, although the effect size is relatively small. Greater interest in digital activities predicts higher digital skills, with a large effect size, and higher levels of distraction are associated with lower digital skills, with a moderate effect size. All predictors are statistically significant, with p-values below 0.001, indicating robust relationships with digital skills. Among these predictors, interest in digital activities emerges as the strongest predictor, followed by distraction and parental education.

  • Index Highest Parental Education: This predictor shows a statistically significant negative effect on digital skills. A higher level of parental education (measured in years of schooling) is associated with lower digital skills in their children. The standardized coefficient (Beta = −0.043) indicates a small effect size.

  • Interest: This predictor demonstrates a statistically significant positive effect on digital skills. Greater interest in digital activities predicts better digital skills. The standardized coefficient (Beta = 0.284) indicates a moderate effect size.

  • Distraction: This predictor shows a statistically significant negative effect on digital skills. Higher levels of distraction are associated with lower digital skills. The standardized coefficient (Beta = −0.112) suggests a small to moderate effect size.

For Germany, the final regression model incorporates all predictors: higher parental education is positively associated with digital skills, more time spent on digital activities is negatively associated with digital skills, being from an immigrant background has a slightly negative association with digital skills, greater interest in digital activities predicts higher digital skills, one gender tends to score lower on digital skills on average, and higher levels of distraction are associated with higher digital skills. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, estimated time, immigrant background, interest, gender, and distraction contributes to the variance in digital skills.

  • Index Highest Parental Education: Higher parental education is positively associated with digital skills. The standardized coefficient (Beta = 0.093) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is negatively associated with digital skills. The standardized coefficient (Beta = −0.080) suggests a small effect size.

  • Index on Immigrant Background: Being from an immigrant background has a slightly negative association with digital skills. The standardized coefficient (Beta = −0.012) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.114) indicates a small to moderate effect size.

  • Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.019) suggests a small effect size.

  • Distraction: Higher levels of distraction are associated with higher digital skills. The standardized coefficient (Beta = 0.064) suggests a small effect size.

For Greece, the final regression model includes all predictors. Higher parental education is positively associated with digital skills, though the effect size is relatively small. More time spent on digital activities is negatively associated with digital skills, with a moderate effect size, being from an immigrant background has a negative association with digital skills, with a small effect size, greater interest in digital activities predicts higher digital skills, with a large effect size, one gender tends to score lower on digital skills on average, with a small effect size, and higher levels of distraction are associated with lower digital skills, with a small effect size.

All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, estimated time, immigrant background, interest, gender, and distraction contributes to the variance in digital skills.

  • Index Highest Parental Education: Higher parental education is positively associated with digital skills. The standardized coefficient (Beta = 0.045) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is negatively associated with digital skills. The standardized coefficient (Beta = −0.110) suggests a small to moderate effect size.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, has a negative association with digital skills. The standardized coefficient (Beta = −0.055) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.302) indicates a moderate-to-large effect size.

  • Student Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.075) suggests a small effect size.

  • Distraction: Higher levels of distraction are associated with lower digital skills. The standardized coefficient (Beta = −0.051) suggests a small effect size.

In Hungary, the final regression model identifies three significant predictors of digital skills: time spent on digital devices, interest, and gender. More specifically, more time spent on digital devices is associated with higher digital skills, though the effect size is relatively small. Greater interest in digital activities predicts higher digital skills, with a large effect size, and one gender tends to score lower on digital skills on average, with a small effect size. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of time spent on digital devices, interest in digital activities, and gender contributes to the variance in digital skills.

  • Time spent on digital devices: This predictor has a statistically significant positive effect on digital skills. More time spent on digital devices is associated with higher digital skills. The standardized coefficient (Beta = 0.033) suggests a small effect size.

  • Interest: This predictor has a statistically significant positive effect on digital skills. Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.213) indicates a moderate effect size.

  • Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.054) suggests a small effect size.

We have two predictors for digital skills in Iceland: parental education and interest toward ICT. Higher parental education is positively associated with digital skills, with a moderate effect size, and greater interest in digital activities predicts higher digital skills, with a moderate effect size. Both predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education and interest in digital activities contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is positively associated with digital skills. The standardized coefficient (Beta = 0.171) suggests a small to moderate effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.218) indicates a moderate effect size.

In Ireland, all six predictors are included in the final regression model: higher parental education is negatively associated with digital skills, though the effect size is relatively small, more time spent on the internet is positively associated with digital skills, but the effect size is very small, being from an immigrant background has a negative association with digital skills, with a small effect size, greater interest in digital activities predicts higher digital skills, with a large effect size, and one gender tends to score lower on digital skills on average, with a moderate effect size. Higher levels of distraction are associated with higher digital skills, with a small effect size. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, estimated time, immigrant background, interest, gender, and distraction contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is negatively associated with digital skills. The standardized coefficient (Beta = −0.034) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is positively associated with digital skills, but the effect size is small (Beta = 0.021).

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, has a negative association with digital skills. The standardized coefficient (Beta = −0.045) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.297) indicates a moderate-to-large effect size.

  • Student (Standardized) Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.128) suggests a small to moderate effect size.

  • Distraction: Higher levels of distraction are associated with higher digital skills. The standardized coefficient (Beta = 0.046) suggests a small effect size.

In Italy, the final regression model identifies six significant predictors of digital skills. Higher parental education is positively associated with digital skills, with a small to moderate effect size, more time spent on digital activities is positively associated with digital skills, being from an immigrant background has a positive association with digital skills, greater interest in digital activities predicts higher digital skills, one gender tends to score lower on digital skills on average, and higher levels of distraction are associated with higher digital skills. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, time, immigrant background, interest, gender, and distraction contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is positively associated with digital skills. The standardized coefficient (Beta = 0.079) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is positively associated with digital skills, though the effect size is relatively small (Beta = 0.036).

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, has a positive association with digital skills. The standardized coefficient (Beta = 0.061) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.337) indicates a moderate-to-large effect size.

  • Student (Standardized) Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.033) suggests a small effect size.

  • Distraction: Higher levels of distraction are associated with higher digital skills. The standardized coefficient (Beta = 0.075) suggests a small effect size.

In Latvia, the final regression model identifies four significant predictors of digital skills. Higher parental education is positively associated with digital skills, more time spent on digital activities is negatively associated with digital skills, greater interest in digital activities predicts higher digital skills, and one gender tends to score lower on digital skills on average. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, ICT time, interest, and gender contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is positively associated with digital skills. The standardized coefficient (Beta = 0.062) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is negatively associated with digital skills. The standardized coefficient (Beta = −0.048) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.200) indicates a moderate effect size.

  • Student (Standardized) Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.074) suggests a small effect size.

In Lithuania, four significant predictors of digital skills are included. Again, higher parental education is positively associated with digital skills, greater interest in digital activities predicts higher digital skills, one gender tends to score lower on digital skills on average, and higher levels of distraction are associated with lower digital skills. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, interest, gender, and distraction contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is positively associated with digital skills. The standardized coefficient (Beta = 0.072) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.259) indicates a moderate effect size.

  • Student (Standardized) Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.095) suggests a small effect size.

  • Distraction: Higher levels of distraction are associated with lower digital skills. The standardized coefficient (Beta = −0.064) suggests a small effect size.

The final regression model for Malta identifies two significant predictors of digital skills. Greater interest in digital activities predicts higher digital skills, and higher levels of distraction are associated with lower digital skills. Both predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of interest and distraction contributes to the variance in digital skills.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.243) indicates a moderate effect size.

  • Distraction: Higher levels of distraction are associated with lower digital skills. The standardized coefficient (Beta = −0.104) suggests a small to moderate effect size.

In Poland, the proposed model identifies five out of six significant predictors of digital skills. Higher parental education is positively associated with digital skills, more time spent on digital activities is negatively associated with digital skills, being from an immigrant background has a negative association with digital skills, greater interest in digital activities predicts higher digital skills, and one gender tends to score lower on digital skills on average. All predictors are statistically significant, with p-values below 0.05, indicating robust relationships with digital skills. These findings suggest that a combination of parental education, estimated time, immigrant background, interest, and gender contributes to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education, measured on the international years of schooling scale, is positively associated with digital skills. The standardized coefficient (Beta = 0.045) suggests a small effect size.

  • Estimated Time: More estimated time spent on certain activities is negatively associated with digital skills. The standardized coefficient (Beta = −0.084) suggests a small effect size.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, has a negative association with digital skills. The standardized coefficient (Beta = −0.015) suggests a small effect size.

  • Interest: Greater interest in digital activities predicts higher digital skills. The standardized coefficient (Beta = 0.282) indicates a moderate effect size.

  • Student (Standardized) Gender: Gender also plays a role, with one gender scoring lower on digital skills on average. The standardized coefficient (Beta = −0.048) suggests a small effect size.

For Romania, the regression model includes all proposed predictors of digital skills. Higher parental education predicts higher digital skills, spending more time on digital activities is associated with higher digital skills, being from an immigrant background is associated with lower digital skills, higher interest in digital activities predicts higher digital skills, gender also plays a role, with one gender scoring lower on digital skills on average, and higher levels of distraction are associated with lower digital skills. These predictors collectively contribute to the variance in digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, with a moderate effect size.

  • Estimated Time: More estimated time spent on activities is positively associated with digital skills, though the effect size is small.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, according to the OECD definition, is negatively associated with digital skills. However, the effect size is relatively small.

  • Interest: Higher interest in digital activities is strongly positively associated with digital skills, with a large effect size.

  • Student (Standardized) Gender: One gender tends to score lower on digital skills on average, with a moderate effect size.

  • Distraction: Higher levels of distraction are negatively associated with digital skills, though the effect size is relatively small.

In Slovakia, the final regression model identifies four significant predictors of digital skills. Higher parental education predicts higher digital skills. Higher interest in digital activities predicts higher digital skills, one gender scores lower on digital skills on average, and higher levels of distraction are associated with lower digital skills, though the effect size is relatively small.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, with a small effect size.

  • Interest: Higher interest in digital activities predicts higher digital skills, with a moderate effect size.

  • Student (Standardized) Gender: One gender tends to score lower on digital skills on average, with a small effect size.

  • Distraction: Higher levels of distraction are associated with lower digital skills, though the effect size is relatively small.

In Slovenia, the regression model identifies five significant predictors of digital skills. Higher parental education is negatively associated with digital skills, spending more time on digital activities is negatively associated with digital skills, being from an immigrant background is negatively associated with digital skills, higher interest in digital activities predicts higher digital skills, and gender also plays a role, with one gender scoring lower on digital skills on average.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is negatively associated with digital skills, with a small effect size.

  • Estimated Time: More estimated time spent is negatively associated with digital skills, with a small effect size.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, is negatively associated with digital skills, with a small effect size.

  • Interest: Higher interest in digital activities predicts higher digital skills, with a moderate effect size.

  • Student (Standardized) Gender: One gender tends to score lower on digital skills on average, with a small effect size.

For Sweden, the stepwise regression model identifies five significant predictors of digital skills. Higher parental education is positively associated with digital skills, being from an immigrant background is negatively associated with digital skills, higher interest in digital activities predicts higher digital skills, one gender scoring lower on digital skills on average, and higher levels of distraction are associated with higher digital skills, with a small effect size.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, with a moderate effect size.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, is negatively associated with digital skills, with a moderate effect size.

  • Interest: Higher interest in digital activities predicts higher digital skills, with a large effect size.

  • Student (Standardized) Gender: One gender tends to score lower on digital skills on average, with a moderate effect size.

  • Distraction: Higher levels of distraction are associated with higher digital skills, with a small effect size.

In Switzerland, the regression model identifies three significant predictors of digital skills. Higher parental education is positively associated with digital skills, being from an immigrant background is negatively associated with digital skills, and higher interest in digital activities predicts higher digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, with a moderate effect size.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, is negatively associated with digital skills, though the effect size is small.

  • Interest: Higher interest in digital activities predicts higher digital skills, with a large effect size.

For the UK, the proposed regression model identifies five significant predictors of digital skills. Higher parental education is positively associated with digital skills, spending more time on digital activities is positively associated with digital skills, being from an immigrant background is positively associated with digital skills, higher interest in digital activities predicts higher digital skills, and one gender scoring lower on digital skills on average, with a large effect size.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, with a large effect size.

  • Estimated Time: More estimated time spent is positively associated with digital skills, though the effect size is small.

  • Index on Immigrant Background (OECD definition): Being from an immigrant background, as per the OECD definition, is positively associated with digital skills, though the effect size is small.

  • Interest: Higher interest in digital activities predicts higher digital skills, with a very large effect size.

  • Student (Standardized) Gender: One gender tends to score lower on digital skills on average, with a large effect size.

Last, for Ukraine (18 regions out of 127), the following are made apparent from the model. Higher parental education, as measured by the International Years of Schooling scale, has a positive albeit small effect on digital skills. Spending more time on digital devices is negatively associated with digital skills. Students from an immigrant background tend to have higher digital skills, higher interest in digital activities strongly predicts higher digital skills. One gender scoring lower on digital skills on average, and higher levels of distraction are associated with higher digital skills.

  • Index Highest Parental Education (International Years of Schooling Scale): Higher parental education is positively associated with digital skills, indicating that individuals with parents who have higher levels of education tend to have better digital skills. However, the effect size is relatively small (p = 0.011).

  • Estimated Time: Spending more time on digital devices is negatively associated with digital skills. This suggests that individuals who spend more time on digital devices may have lower digital skills, possibly due to passive consumption or distraction (p < 0.001).

  • Index on Immigrant Background (OECD definition): Individuals from an immigrant background, as per the OECD definition, tend to have higher digital skills. This suggests that immigrants may possess adaptive strategies or bilingual proficiency that contribute to their digital skills (p < 0.001).

  • Interest: Higher interest in digital activities strongly predicts higher digital skills. This highlights the importance of intrinsic motivation in the development of digital skills, as individuals who are more interested in digital activities tend to have better skills (p < 0.001).

  • Student (Standardized) Gender: Gender plays a significant role in digital skills, with one gender scoring lower on average. This gender disparity suggests potential differences in access to and engagement with digital technologies (p < 0.001).

  • Distraction: Higher levels of distraction are associated with higher digital skills. This unexpected finding may suggest that individuals who are better at managing distractions are also more adept at utilizing digital technologies effectively (p < 0.001).

Source link