The Role of Instagram, Facebook, and YouTube Frequency of Use in University Students’ Digital Skills Components

: The literature provides rich evidence on students’ levels of digital skills as well as on their frequency of Internet and social media use. Several studies have examined the relationships between social media use and academic achievement. However, there are not any studies investigating the relationships between social media use and digital skills of higher education students. To ﬁll this research gap, this study examines the links between the frequency of social media use and students’ digital skills. The survey was conducted with 155 university students in Greece. The instrument on six digital skills components was evaluated in terms of reliability and consistency. The research ﬁndings reveal a strong positive association between the use of YouTube and students’ digital skills on content evaluation and protection. Age and educational level differences were noted in several digital skills components, whereas age and gender were related to the frequency of Instagram use. The main conclusion is that the generic use of Facebook and Instagram does not affect students’ digital skills. However, combined literature evidence implies that when social media are used for educational purposes, they might inﬂuence students’ perceived levels of digital skills. Implications and limitations are discussed in the study.


Introduction
The emergent situation due to the COVID-19 pandemic exhibited the necessity of social media and digital skills. People in quarantine need to use digital technologies (including social media) in order to communicate, work, shop, learn, be informed, socialize, have fun, and live their lives in general. For example, in the USA, about one third of adults had an online socializing event with friends or family and one third of adults ordered food online from a local restaurant during COVID-19 [1]. Similarly, there was a 35% increase in using social media [2].
As of October 2020, there were over 4.5 billion active Internet users and over 4 billion active social media users worldwide. Specifically, there were more than 2.7 billion Facebook users, 2 billion YouTube users, 2 billion WhatsApp users, 1 billion Instagram users, 0.4 billion Pinterest users, and 0.3 billion Twitter users worldwide [30]. In Europe, over half of adults and 86% of young people 16 to 24 years old participated in social networks. According to Datareportal [31], on average Internet users spent from 1.3 h (in Japan, Germany, and France) to 4.5 h (in Brazil) on social media daily. They spent daily about 5 h on Correspondingly, when investigating the relationship of Facebook use and educational level, Van Deursen [57] found that people with low levels of education were using the Internet for more hours daily in their spare time than higher educated and employed people. Similarly, Correa [58] found that lower educated young people tended to use Facebook more frequently. She also found that digital skills were not related to the frequency of Facebook use. However, according to Pew Research Center [1], the more educated the person, the more they tend to use social media. More adults with a college degree or more (74%) use Facebook than those with a high school diploma or less (61%). Finally, Castillo de Mesa [59] investigated the moderating effect of digital skills on the relationship between closeness and resilience in a Facebook group. They found that participants' cohesion increased their resilience capacity when digital skills were high.
Motivated by the above-mentioned research gaps, this paper will investigate the following three main research questions (RQ): RQ1: Are there any significant correlations between social media use and digital skills? RQ2: Are there any significant differentiations in social media use according to students' gender, age, grade, and educational level? RQ3: Are there any significant differentiations in digital skills according to students' gender, age, grade, and educational level?
In this research, the term "digital skills" is described by six components reflecting the students' digital skills to (1) Communicate, Collaborate, Share; (2) Search, Find, Access; (3) Store, Manage, Delete; (4) Evaluate; (5) Protect; and (6) Develop, Apply, Modify, as presented in Perifanou [60,61]. The term "social media use" is defined by the frequency (hours per week) of students' use of three popular social media networks: Instagram, Facebook, and YouTube. This study aims to investigate the bivariate correlations between each network and the students' digital skills components.

Context of Study: Greece
This study was implemented in Greece. According to ELSTAT [62], Greece has a population of around 10.7 million people. In 2017-2018, there were 190,000 university students. In 2019, 75.7% of the population aged 16-74 years were accessing the Internet (i.e., 77.9% for sending/receiving emails, 66.8% for telephoning over the Internet, 74.9% for social networking, 88.3% for finding information about goods and services, 87.7% for reading/downloading newspapers and magazines, 65.9% for finding information or using health-related services, 40.3% for Internet banking, 64.6% for obtaining information from public authority websites, 39.9% for downloading official forms, and 36.5% for sending filled-in forms). In addition, almost all young Greeks 16 to 24 years old were regular Internet users.
In Greece, there were 8.3 million Internet users and 6.2 million social media users in January 2020 [63]. There were 5.2 million Facebook users, 3.3. million Instagram users, 0.7 million Twitter users, and 1.6 million LinkedIn users.
Regarding the digital skills of young Greeks 16 to 24 years old, 44.9% of them had an above basic level of digital skills, 92.3% of them had at least basic digital skills, and 35.1% of them had a low level of digital skills (missing some type of basic skills) [64].

Participants and Procedure
A questionnaire and a consent form were distributed to 300 undergraduate and postgraduate students between January and April 2020. A written form of the questionnaire was provided to students of two different undergraduate courses (e-Commerce and e-Business, Information Systems in Management) in a university in Greece, and an online version of the same questionnaire was administered in three postgraduate programs (Information Systems, e-Business and Digital Marketing, Law and Economics) in Greece.
The questionnaire items about digital skills were measured on a five-point Likert scale (1: strongly disagree to 5: strongly agree). The questionnaire also asked for information on the social and academic profile of the students (gender, age, average grade in last semester).
A total of 155 students (79 undergraduates and 76 postgraduates) voluntarily and anonymously completed the questionnaire, providing valid data (outliers were cleaned) and replying to all the items regarding digital skills components. Of those, 138 provided feedback regarding the frequency (hours per week) of Facebook use, 122 replied on Instagram, and 108 on YouTube. Table 1 presents the participants' socio-demographic profile.

Instrument Description and Evaluation
A 28-item instrument was designed based on the modern digital skills components presented in Perifanou [60,61]. The suggested digital skills extend previous models by including items related to the newly emerged technological trends like social media and mobile use. The instrument was composed of 6 components measuring the students' digital skills to (1) Communicate, Collaborate, Share; (2) Search, Find, Access; (3) Store, Manage, Delete; (4) Evaluate; (5) Protect; and (6) Develop, Apply, Modify. The structure of the instrument and the items' wording are presented in Appendix A.
The instrument was evaluated in terms of reliability, internal consistency, and convergence validity. The evaluation was applied through a partial least square (PLS) confirmatory factor analysis (CFA) approach in the SmartPLS software. As depicted in Table 2, all the values of Cronbach's alpha and composite reliability (CR) demonstrated internal consistency (>0.7). The convergence validity was evaluated through average variance extracted (AVE) that met accepted values, above the threshold of 0.7 [66] in all components. In addition, the item-total correlations were examined, and significant correlations were shown to exist between the factors (p < 0.01).

Data Analysis
The results of the normal distribution tests revealed non normality (p < 0.05) for all the measured variables across all examined groups. For this reason, we used non-parametric statistics to explore bivariate correlations and group-based differences in the measured constructs. In particular, the Spearman Rho correlation test was applied to examine the significant correlations between the frequency of social media use and digital skills. Gender and educational-level differences were examined through the Mann-Whitney Test, whereas differences in multi-level variables like age and grade were evaluated through the Kruskal-Wallis test. IBM SPSS software was used for the descriptive statistics and the statistical tests.

Descriptive Statistics Regarding Social Media Use and Digital Skills
As shown in Table 3 YouTube was the most frequently used social media network amongst Greek students. The students also spent less time (on a weekly rate) on Facebook compared to Instagram. The high values in the standard deviations reveal the students' individual differences in their expressed frequency of social media use. As shown in Table 4, students expressed high levels of digital skills in most of the examined components. However, the component of DAM received the lowest mean score (<3.0).  Table 5 presents the results of the Spearman Rho correlation test. A set of significant correlations was detected between the frequency of YouTube use and two digital skills components, EV and PR. The frequency of use of other networks (Instagram and Facebook) did not reveal any significant correlations in the examined sample. Significant group-based differences in social media use were observed for Instagram in the gender and age-related groups. Female students revealed significantly higher values of frequency use of Instagram, whereas no differences were detected in the use of Facebook or YouTube.
On the other hand, differences in the frequency of Facebook use were detected between undergraduate and postgraduate students, where postgraduate students expressed significantly higher values of frequency use.
No differences were detected in relation to the students' grades. Table 6 below shows the Mann-Whitney and Kruskal-Wallis results for the examined social media constructs across gender, age, and educational level-related groups.

Differences in Digital Skills Components
Significant group-based differences were observed in two digital skills components in relation to age, and in four components in relation to educational level. Gender and grade revealed no significant differences.
Furthermore, postgraduate students expressed significantly higher values than undergraduate students in the components of CCS, EV, PR, and DAM. Table 7 below presents the Kruskal-Wallis results for the examined digital skills components across age and educational level-related groups.

Discussion and Implications
This research study revealed several interesting findings about the role of popular social media networks (in terms of frequency use) and the students' digital skills across six digital skills components. Furthermore, some remarkable outcomes emerged in terms of group-based differences according to the students' gender, age, and educational level.
As shown in Table 4, the students expressed high levels of digital skills in most of the examined components, implying that they are quite confident to Communicate, Collaborate, Share; Search, Find, Access; Store, Manage, Delete; Evaluate; and Protect using digital technologies. These findings are in accordance with recent evidence [64] supporting that young Greeks have at least basic digital skills. However, our findings suggest that young Greeks need to be further qualified in terms of development skills (Develop, Apply, Modify). In fact, developing skills were characterized as "advanced digital skills" by Cedefop [67], and hence, Cedefop suggests designing enhanced strategies to allow citizens access to further training since "advanced digital skills, in particular programming and coding, will become a key prerequisite for entry into many jobs" (as cited in Cedefop [67].
Next, we discuss the specific findings for the stated research objectives. RQ1: Are there any significant correlations between social media use and digital skills?
Our findings revealed significant correlations between social media use and digital skills only for the YouTube network. Facebook and Instagram use did not show any relationship with the students' digital skills. This finding is in accordance with previous studies like Correa [58], who found that digital skills are not related to the frequency of Facebook use. As already stated, there are not many studies exploring this relationship since most previous studies targeted the relationship between Facebook use and students' academic achievement.
The reason that YouTube is linked (in this study) to students' digital skills might be explained by that social network's characteristics compared those of Facebook or Instagram. As explained in Bhatia [68], YouTube has emerged as an informal learning environment where users can access a huge dataset of video tutorials and how-to-related material. Researchers have pointed out that university students tend to visit YouTube regularly to learn how to solve specific problems [69]. In previous years, YouTube was recognized as a significant source of educative material for several study disciplines, especially nursing and medicine [70]. Today, researchers agree that YouTube is a rich source of online educational videos that effectively affect students' learning [71].
So, all the above imply that students tend to use YouTube to seek information (either on educational content or otherwise), which can positively affect their learning. Therefore, this seems to be the main reason why YouTube is positively associated with students' digital skills. Interestingly, in this study the frequency of YouTube use was positively associated with the digital skills of Evaluate and Protect, implying that the higher the frequency of watching YouTube videos the higher the students' skills and efficiency in identifying trusted and non-trusted web content, malicious attacks, or fake information and protecting themselves. A reasonable explanation might be Europe's recent attempts on cybersecurity awareness [72] or the recent era of the #MeToo movement and online harassment [73], which may motivate individuals to seek relevant information in the YouTube databases. RQ2: Are there any significant differentiations in social media use according to students' gender, age, grade, and educational level?
Both gender and age revealed significant differences in the frequency of Instagram use, but not of Facebook or YouTube. Regarding gender, female students expressed higher values of Instagram use compared to male students. This result aligns well with Statista's [74] recent findings that there are more female Instagram users than male ones. Previous research has also proven gender differences in online activities and social media preferences [75][76][77]; however, females' favored medium was not stated. As explained in Van Deursen [57], women use social media for communication, text messaging, and video calls more often than men. On the other hand, authors have shown that male users access social media mainly to seek information and read news, make commercial transactions, for entertainment reasons, or for personal development. All the above do not clearly explain the significant difference between Instagram and Facebook, since they share the main core characteristics compared to YouTube (or other networks, e.g., Twitter). We believe it is the network's photo-centric characteristics that might render it more favorable to the female population. However, it might be the combination of gender and age that rendered this outcome. According to Statista [78], as of January 2021 most Instagram users were young females (18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34). This is also confirmed by the age-related results of our study that demonstrated that the youngest groups expressed significantly higher values in the frequency of Instagram use. However, by examining our sample characteristics we observed that there were more young females in the groups of 18-24 and 25-35 years old than young males. Therefore, we suggest future research to re-examine the statistical significance of gender differences in the frequency of Instagram use.
Contrary to previous findings showing that academic or learning performance is related to the frequency of social media use [51][52][53][54][55], our research revealed no effect of students' average grade. Our finding can be explained by the fact that we measured the students' generic use of social media without trying to identify their learning-related activities on social media. According to the findings of Lambić [50], a correlation between students' academic performance and their frequency of Facebook use was detected only in their use for educational purposes and not in general.
The increased frequency of Facebook use that was observed in postgraduate students aligns well with the previous study by Pew Research Center [1], who found out that adults with a college degree or more use Facebook more frequently than those with a high school diploma. However, more research needs to be done since our study was focused on undergraduate and postgraduate students and not on the generic population. RQ3: Are there any significant differentiations in digital skills according to students' gender, age, grade, and educational level?
No gender-or grade-related differences were detected in the measured digital skills components. Age revealed significant group differences in two components (PR and DAM), and educational level revealed significant group differences in four (CCS, EV, PR, and DAM) out of six components.
Contrary to previous studies [79], this study did not detect any gender gap in the students' digital skills, since both female and male students expressed relatively equal mean scores even in advanced digital skills like developing (DAM). These findings are encouraging for the future of the gender gap elimination in computing; however, further research should be conducted due to the current participants' characteristics. For example, the overall descriptive statistics (Table 4) revealed a lower level of developing skills for the whole sample. This might suggest that other factors should be considered, such as, for instance, the participants' field of study of prior experience in developing tasks.
As regards age, we observed that the youngest group of 18-24 years old revealed the lowest score in protection skills (PR), whereas the group of 24-35 years old revealed the highest ones both in PR and in DAM.
The educational-level difference in four digital skills components can possibly be explained by the fact that students in postgraduate programs have greater knowledge and practical experience in technology and hence express higher values of digital skills. However, this could vary according to the study field and prior experience; hence, further research is suggested.

Research Implications and Limitations
A first implication of this work is that students and instructors would be further motivated to use YouTube for educational reasons since it is positively associated with digital skills components. As suggested in previous research [71], students and educators should carefully evaluate the quality of YouTube videos to be used as educational material and educators should carefully check for the quality of the videos before suggesting them as supplementary learning materials.
A second implication is the adoption of Facebook and Instagram in the educational context as well. Since students tend to engage strongly and spend a lot of time on those networks as well, educators could encourage their students to enhance their knowledge on the course topics through those networks. For example, educators can cultivate a connectivity approach in learning by assigning students to Facebook groups or asking them to create content and upload photo-based material to Instagram, especially for complex educational concepts that are better explained through visualized representations (e.g., in medicine courses [80]). Moreover, Instagram can be used by students and educators to share useful information and build trusted communities where affective support is also provided. Researchers [81] have already suggested that the visual-centric nature of Instagram renders it a platform that invites personal disclosure and generates feelings of connectedness. Moreover, the significant difference in gender preferences implies that gender-oriented and personalized approaches could be implemented. For instance, young female students could be encouraged by educators to participate more in Instagram-related activities, since that platform seems to engage them more than others (Facebook).
Another implication is the need for cybersecurity awareness and protection practices among young students (18-24 years old). Young students should be more informed on how to protect themselves online, for example, by frequently changing their password, keeping their transactions secure, and being well-informed about protection strategies against bulling and harassment in social media.
Last, the age-and educational level-related differences in developing digital skills should be further considered so educators and educational institutions can design policies and practices to shrink the gaps and provide those groups of students with further training and support to enhance their digital skills.
This research has some limitations. First, the survey was based on self-reported feedback and not on actual measures, so there might be critical differentiations in the frequency of social media use expressed by the students. Another limitation relates to generalizability and the student population. All the participants were attending Greek universities and did not differ significantly in terms of educational background. Finally, the sample size was quite small, and which could have led to some bias in the results.

Conclusions and Future Research
Although there is rich literature on the use of social media among young adults and on their digital skills, no research evidence exists on the relationships between those constructs. To this end, we examined the relationship between the frequency use of three popular social media networks (Instagram, Facebook, and YouTube) and students' digital skills, as well as their group-based differences according to their gender, age, grade, and educational level.
In this study we measured students' digital skills in accordance with six digital skills components as derived from recent literature. Our findings revealed a set of significant findings, especially on the role of YouTube. A strong positive correlation was found between the frequency of YouTube use and students' digital skills regarding their evaluation and protection skills. This finding raised several concerns regarding the adoption and use of YouTube as a formal educational channel in higher education.
Another interesting finding was the gender and age-related differences that were observed in the frequency of Instagram use as well as the age-related differences observed in two digital skills components. Although the students' grades did not reveal any significant outcome, the educational level (undergraduate/postgraduate) yielded several significant differentiations in the digital skills components as well as in the frequency of Facebook use.
The findings of this study have some limitations, and hence future research should be conducted. For instance, despite the generalizability limitation, these findings could contribute to specific trends in population groups, and hence more studies should be conducted to compare the results in different populations. Moreover, our findings can be extended by future research that will investigate older adults' social media use and digital skills as well as the inequalities with respect to Internet-access infrastructure and use of different social media networks (e.g., TikTok, Snapchat, LinkedIn, Twitter, Pinterest).
Overall, our findings seek to contribute to the deeper understanding of the differentiations and links between social media used and digital skills in university students. Data Availability Statement: All data that are not subjected to institutional restrictions are available within the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.