Information and Communication Technology (ICT) has become an integral part of our economic, political, and social lives, altering the way we purchase goods, bank, communicate with one another and support our education (Ali 2011:211). Social media is one of the ICTs easily adopted and used by the younger generation for socialising. Researchers, such as Kausar and Awan (2019:120), defined social media as a means of communication, collaboration, sharing and exchanging of data through different networks. Social media has numerous benefits, such as the ability to promote and encourage higher public participation in innovations, collaboration and interaction (Haryanti & Rusfian 2018:131). Most students from the disadvantaged background are not able to benefit from the advantages offered by social media because of the digital divide (DD). Digital divide is defined as the gap between individuals, households, businesses and geographical areas at different socio-economic levels with regards to their opportunities to access ICT and to their use of the Internet for a wide variety of activities (Organization for Economic Co-operation and Development 2001).

The problem identified in this study is that despite the benefits of using social media technologies for learning, many students are still not adopting this for their learning because of various factors of the DD. The objective of this study was to investigate the impact of the DD on first-year students in using social media for learning in tertiary education institutions. The study was required to answer this research problem: ‘[w]hy are many students in higher education (HE) institutions still not adopting social media technologies for their studies, despite the benefits offered by social media?’

This study is deemed to be important because of rampant low level of academic performance in various universities of technology, despite the benefits afforded by e-learning and social media for learning. Because of this low level of academic performance, universities are laying several measures to get the learning materials to students using different digital formats, such as e-Tutors, Social media and Blackboard, and hence, there is a need to determine the factors that may prevent students from using these technologies and the impact of personal and environmental factors on students’ learning.

This article is organised as follows: first, the introduction, background, related study and literature, theoretical framework, materials and methods, ethical considerations and references. The article also covers the theory underpinned in this study, materials and methods used, and finally, highlights the findings and recommendation suggested for this study.

According to Chawinga (2017:6), the use of social media improves blended learning, which helps students to create positive contextual learning in relation to educational intentions. The application of ICT creates more efficient and effective ways for students to become proactive and competitive in the higher institution of learning (Attuquayefio 2019:117). All these opportunities are eluding the first-year students from mostly poor rural areas with no prior access to computers and ICTs. In the same view, Ilorah et al. (2017:252) explained that many rural areas are resource challenged in several ways from a lack of basic ICT infrastructure, electricity to poor network connections.

Development Support Monitor paper revealed that the affordability of ICT devices is low in all provinces of South Africa when compared with other countries. Six years later, Gillwald, Mothobi and Rademan (2018:6) stated that as much as unaffordability of devices is the primary barrier to South Africans coming online, the unaffordability of data is also one of the major factors limiting the intensity of use. Although technology is growing fast day by day and the prices of computers and other ICT devices are not as expensive as before, some families still have difficulties affording and accessing them (Smith 2015). Smith further argued that families who cannot afford Internet access are at a disadvantage compared with those who do.

According to Naidoo and Raju (2012:34), students from HE institutions in South Africa fall into two categories: some with no access or skills, whilst there are others who have had access and are skilled in using and have a vast experience with ICTs. Some students from disadvantaged backgrounds have never used the Internet.

A research study conducted by Beger and Sinha (2012) shows that there is a pronounced DD in South Africa with regards to ICTs’ ownership and access, and this is still the situation as indicated by Gillwald et al. (2018). Beger and Sinha went on to define ‘[d]igital divide’ as the gap between those who have access to ICT resources and those who do not have access. Although technology is improving, the DD still exists, thus creating a negative impact on students’ learning capabilities

Bala (2018:369) emphasised that the access and use of ICT improve teaching and learning capabilities amongst students and provides real opportunities for individualisation of instruction. However, Naidoo and Raju (2012:34) indicated that majority of the students who enroll in HE institutions only get to use computers and Internet for the first time, and hence, have difficulties even in performing the most basic computer skills.

According to Barnard and Van der Merwe (2015), much as students might have access to Internet, computers and other ICT resources whilst at school, their socio-economic backgrounds limit this access at home. In the same vein, Gudmundsdottir (2010:85) stated that although students access ICT at schools, their access to and use of the same technology at home can be adversely or positively affected by their individual circumstances. As ICT plays a greater role across our society, including public and private education, more countries are more than ever in need of high-quality, internationally comparable statistics on ICT in education (Du Toit 2015:4).

Theories and models are significant in directing the research process (Mokwena & Hlebela 2018). The study employed social cognitive theory developed by Bandura (1986) to investigate how the DD impact the adoption of social media for learning, because according to this theory, people are neither driven by inner forces nor automatically shaped and controlled by the environment (Bandura 1989).

The model is conceptualised to suit this research study. In the conceptualised model, personal and environmental factors influence each other, and they both affect the behaviour of the individual. As outlined by Bandura (1989) in social cognitive theory, personal factors include biological properties of the organism, individual’s expectation, beliefs and self-perceptions, whilst environmental factors encompass the physical and social environments. Personal and environmental factors as shown in the following model to address the issue of DD.

The constructs are explained as follows:

A quantitative study using a survey method was used to conduct the research study. Quantitative research necessitates the diminution of phenomena to arithmetical values in order to carry out statistical analysis (Gelo, Braakmann & Benetka 2008). This research study collected quantitative data using a closed-ended questionnaire because it provides the researcher with numerical data (Zohrabi 2013). The sample population for this research study included 600 first-year students from a multi-campus university in South Africa. Probability and non-probability sampling methods were used for this study. Probability random sampling was used for the selection of participants because it ensures external validity and generalisability, and reduces the possibilities of sampling biases (Fox, Hunn & Mathers 2007). Non-probability sampling was purposively used to select first-year students, and probability sampling was employed to select 300 first-year students randomly from each of the campuses. These two campuses were purposively selected because one is situated in an urban metropolitan city. Students who study there come from different backgrounds, and other campus, on the other hand, was selected because it is one of the distance campuses where majority of the students come from rural areas.

Applying the identified social cognitive theory factors and the DD elements, a questionnaire was constructed and used to collect the research data. According to Ilorah (2017), questionnaire is a data collection technique where respondents answer research questions by selecting their appropriate choices from given answers or supplying their own answers electronically or on a hard copy. They were hand delivered and later collected back. The hand-delivery method provides a high response rate, and the researcher has an opportunity to know who received the questionnaire and who did not receive them (Phellas, Bloch & Seale 2012). IBM SPSS 24.0 was used for descriptive statistics and for testing the validity and reliability of data collected. Out of the 600 questionnaires distributed, 321 were found to be useful for analysis.

In order to design the survey questionnaire, the guidelines as provided by Babbie (2005) and Kumar (2011) were followed. Each construct was represented on the survey by multiple statement items. Some of the statements and questions were adapted from the study of Dewar et al. (2012) and were formulated to suit this study.

The questionnaire was structured as a multiple-choice closed statement item where participants were asked to select their level of agreement with each statement. Before distributing the questionnaire, they were given to two PhD students, three Master students and five first-year students to pretest to check for ambiguous statements, negative, double barrelled questions, errors, instruction clarity and duplicate items. The questionnaire was then restructured using the feedback from the pretest. The instrument was designed primarily to assess the impact of DD on the adoption of social media for learning using the constructs of social cognitive theory as in the conceptual model.

The result revealed that majority of the participants (67.0%) were from the rural-based campus and others (33.0%) were from the urban-based campus. The result also showed that out of the 321 usable questionnaires, majority of the respondents were female (61.1%) and others (38.9%) were male. Most of the respondents were between the ages of 18–24 years (85.7%), followed by age group 25–34 (14.3%). The result also demonstrates that majority of the participants were from rural areas (74.1%), as compared to urban areas (25.9%). In terms of mobile ownership, the result showed that 70.1% of students own smartphones, 16.2% of them own tablets and only 13.7% own a cell phone without Internet connection. This means that 86.3% of students were able to access Internet from their phones. It also reveals that students in HE institutions are aware of social media usage, its technologies and qualities. Majority of the respondents (85.3%) showed awareness of social media tools, and 84.7% of them highlighted that they were aware that social media can be used to facilitate learning. About 75.1% of the respondents indicated that they are aware of the fact that social media can help them to interact with their fellow students.

The results of correlation analysis show a positive relationship between awareness (r = 0.253**, p < 0.01) and behavioural strategies, which means that the more the students are aware of social media usage, the more they intend to adopt it. The results of regression analysis also proved awareness (Beta = 0.213, Sig. = 0.000) as one of the main predictors of personal factors and, therefore, has an impact on adoption of social media for learning.

Most of the participants believed that if they had access to ICTs, such as computer and Internet, they would find it helpful to use social media for their studies. The analysis revealed that 75.7% of the respondents agreed that it was easy for them to learn using social media. On the other hand, 81.6% believed that they had the knowledge of using social media for their studies and 79.4% believed that they had the ability to use social media to support their learning studies. The results of correlation analysis revealed that more students believed that they had knowledge, ability and skills of using social media and intend to adopt social media for learning. The correlation showed a positive relationship between self-efficacy (r = 0.236**, p < 0.01) and behavioural strategies.

The results of regression analysis revealed self-efficacy (Beta = 0.220, Sig. = 0.000), showing that it has an impact on personal factors, which, in turn, impacts the adoption of social media for learning. The findings show that 68.0% of students responded that social media could help to improve their learning. On the other hand, 85.7% believed that it could help them interact with their classmates online, and 82.8% believed that social media could help them be updated with their course or subject. The correlation analysis results revealed that outcome expectations have no relationship (r = 0.097, p > 0.00) with behavioural strategies. Regression analysis also supported the correlation analysis results by showing that outcome expectations (Beta = -0.053, Sig. = 0.374) do not have an impact on adoption of social media for learning in HE institutions. From the analysis, although the participants agreed strongly on the benefits that social media can offer for learning, the result indicates that outcome expectation does not predict adoption.

The result also shows that out of the 46.1% of the participants who had access to computers at home, only 27.8% had computers with access to the Internet. In terms of public services, 36.5% of the students indicated that they had a public library where they could access computers connected to the Internet for free of charge.

In contrast, only 47.4% of students responded positively that they had an Internet cafe in their area where they could access pay-to-use Internet. The correlation analysis results revealed that there is a negative relationship between socio-economic factors (r = −0.045, p > 0.00) and behavioural strategies. The results of regression analysis also supported the correlation analysis result by showing that socio-economic factors (Beta = −0.120, Sig. = 0.030) are not significant predictors in this study. This could be as a result of the finding that majority of the students had smartphones or Internet-enabled phones, which suggested that they could access social media from their phones. From the observation and preliminary exploration, it shows that students sell their books and skip meals in order to buy smartphones. Beger and Sinha (2012) earlier indicated that 72% of South African youth within the age group of 15–24 years own cell phones with Internet connection.

Four hypotheses were formulated and tested using a linear regression analysis approach by reverting each of the independent variable to a dependent variable – behavioural strategies. From the four hypotheses formulated, three are predicting personal factors, which, in turn, predict behavioural strategies; that is, awareness, intended to explain that self-efficacy and outcome expectations fall under personal factors from this study’s conceptual model, and socio-economic factors are explained as environmental factors.

The findings of this research study prove that personal factors measured using awareness and self-efficacy are the main predictor of impact on adoption of social media for learning. Outcome expectations and socio-economic factors do not have a significant impact on adoption of social media.

Awareness (H1[a]) was supported. H1(a) stated that awareness has a positive impact on personal factors. Based on the result, many students responded that they are aware of social media. Correlation analysis results revealed that there is a positive relationship between awareness (r = 0.253**, p < 0.01) and behavioural strategies. Regression analysis also proved that awareness is one of the main predictors of personal factors and indirectly has an impact on adoption of social media. Therefore, awareness proved to have an impact on personal factors, and hence, an impact on adoption of social media.

Self-efficacy (H1[b]) was supported. H1(b) stated that self-efficacy has a positive impact on students’ personal factors. The outcome from the correlation analysis shows a positive relationship between self-efficacy (r = 0.236**, p < 0.01) and behavioural strategies. Regression analysis also proved self-efficacy as one of the main predicators of personal factors, and hence, has an impact on adoption of social media for learning. This means that the more the students believed that they have abilities, knowledge and skills to use social media, the more they become interested in adopting social media for their studies.

Outcome expectations (H1[c]) was not supported. H1(c) stated that outcome expectations have a positive impact on students’ personal factors. Correlation analysis results revealed a positive relationship between outcome expectations (r = 0.380**, p < 0.01) and self-efficacy and between outcome expectations (r = 0.352**, p < 0.01) and awareness. The results of regression analysis found outcome expectations (Beta = -0.053, Sig. = 0.374) to be not a significant predictor of impact on adoption of social media for learning. Therefore, outcome expectations were proven to have no impact on personal factors, and hence, no impact on the adoption of social media for learning.

Therefore, H1 was supported. H1 stated that personal factors have a positive impact on the adoption of social media to facilitate learning in HE institutions.

Socio-economic factors (H2[a]) was supported negatively. H2(a) stated that socio-economic factors have a positive impact on environmental factors. Environmental factors are described as factors that help to clarify why challenging behaviours sometimes increase or decrease following changes in where a person lives or how they spend their days. The correlation analysis results revealed a negative relationship between socio-economic factors (r = -0.045, p > 0.00) and behavioural strategy. The results of regression analysis supported the correlation analysis by proving that socio-economic factors (Beta = -0.120, Sig. = 0.030) are not significant predictors of impact on the adoption of social media for learning. Students’ physical surroundings, in addition to situations that they encounter, do not have an impact on their behaviour. Therefore, socio-economic factors do not have an impact on environmental factors, which means no impact on adoption of social media.

The result proved that H2 was not supported. H2 stated that environmental factors have a positive impact on adoption of social media to facilitate student learning in HE institutions.

This research study was required to answer the following research problem: ‘[w]hy are many students in higher education institutions still not adopting social media technologies for their studies, despite the availability of social media and its benefits?’

The general conclusion of this study, based on the findings, is as follows:

The major limitation of this study is that it was focused only on one university, which could limit the generalisation of findings, unlike if the findings compared two or more different universities. The study also focused on two campuses out of the seven campuses of Tshwane University of Technology. If more campuses were included, maybe the findings would have been different. Another limitation arising from this study was that the study focused only on three social media technologies. Maybe if more social media technologies were included, more information would have come to light.

It is recommended that future research in this field should cover at least two or more universities in a way, which will help to increase the generalisability of the findings, and studies on how ownership of computers and devices connected to the Internet, especially amongst students from disadvantaged backgrounds, can be promoted.

This research study investigated the impact of DD on adoption of social media for learning. The study reported the findings from the quantitative research study, where the data were collected from 600 first-year students of two campuses of a multi-campus university in South Africa. Only 321 of 440 questionnaires returned were found to be useful for analysis. The study proposed social cognitive theory in purpose of understanding the Socio-economic and Personal factors that have impact in the adoption of social media for learning.

From the results analysed, DD factors that were found to have an impact on adoption of social media were identified. The conclusion drawn from demographic factors showed that majority of the students own phones with Internet capability. A lack of access to computers – and devices connected to the Internet – was found to have no impact on adoption of social media. All hypotheses were tested, and three hypotheses were supported and three were not supported. Self-efficacy and awareness were shown to have a significant impact on adoption of social media, whereas outcome expectations and socio-economic factors were found to have no significant impact on adoption of social media for learning.