Predicting Schoolteachers ‟ Intention and Behaviour of Promoting Cyber-Safety Awareness

School learners are attractive candidates for cyber-safety attacks. Advances in Information and Communications Technology (ICT) devices, and the surge in Internet usage, poses significant risks to users with poor cyber-safety awareness and security practices. An example of such users are school learners. Therefore, school learners should be made aware of cyber-safety or safe Internet usage, through awareness programmes as early as possible. Applying the classical Theory of Planned Behaviour in a survey of 109 schoolteachers across 18 schools in 2 districts, the results of a multiple regression analysis unveil that: 1) attitude does not predict the schoolteachers’ intention to raise cyber-safety awareness at their schools in the next 12 months, nor does it predict the schoolteachers’ behaviour of raising cyber-safety awareness in their teaching; and 2) subjective norms and perceived behavioural control predict the schoolteachers’ intention and behaviour. In light of this, the South African Department of Basic Education (DBE) is encouraged to identify influential schoolteachers (i.e. phase heads, sports and cultural coordinators, principals and their deputies, etc.), and to groom them into cyber-safety awareness advocates, for the purposes of influencing other teachers or their peers. Also, the DBE should provide teachers with the requisite resources, opportunities, and support which facilitates cyber-safety awareness in schools, to empower teachers to adopt this highly desired behaviour.


I. INTRODUCTION
The growth and range of Internet usage by children and youth is steadily rising [1]. Since mobile Internet connection is the most popular form of Internet connection in South Africa [2], with 34 million mobile Internet users out of a total of 36 million Internet users [3], a majority of South African school learners are correspondingly now able to access the Internet through their handy mobile phones [4]. With such effortless access to a global network of communication, school learners are susceptible to cyber-risks such as anxiety [5], cyber-bullying [6], human trafficking [7], identity theft [8], and phishing [9], to mention but a few. Therefore, school learners need to be made aware of cyber-safety or safe Internet usage, through awareness programmes as early as possible [10]. Such programmes should teach learners how to use Information and Communications Technology (ICT) and associated ICT devices safely and responsibly.
Cyber-safety awareness could ensure that learners have the necessary attitude towards cyber-safety, as well as the knowledge and skills to deal with real-life cyber-safety threats and risks [4]. Without cyber-safety awareness, learners will remain prone to a myriad of cyber-risks. Cyber-safety awareness and educational programmes can be seen as a defence mechanism for protecting information and other assets [11]. Given the importance of raising cyber-safety awareness, schools in general, and teachers in particular are required to participate in teaching school learners about cyber-safety [12].
Schoolteachers are cited as key role players in the cyber-safety education of learners. Their role, inter alia, is identified as the deliverers of cyber-safety awareness messages to school learners [13]. As such, this study applies the Theory of Planned Behaviour (TPB) with the objective of revealing which construct/s predict the South African schoolteachers" intention and behaviour of raising cyber-safety awareness. This insight is particularly needed and important, as it unveils the predictive power of the respective TPB constructs in our context. With this insight, school cyber-safety awareness campaigns can be enhanced as a result of improved understanding and planning.

II. THE THEORY OF PLANNED BEHAVIOUR
Although the Theory of Reasoned Action (TRA) is not the theoretical framework for the current study, it provides valuable insights into the TPB"s journey as a leading social scientific theory used to study intention and behaviour [14]. Introduced by Fishbein in 1967, the TRA explains the development of the TPB, which is the theoretical framework for this study. The TRA states that the two main determinants of intention are an individual"s attitude toward the behaviour, and the perceived pressures of subjective norms. Together, these forces determine intention [14], which subsequently determines behaviour. Generally, individuals will intend to perform a behaviour when they evaluate it positively, and when they believe important "others" think or expect them to perform the behaviour [15]. However, social scientists concede that the relative weights of attitude and subjective norms may vary, based on intention, and across individuals [15]. The TRA has been effective in a variety of settings [16]. Yet, there are concerns about its generality. For instance, the model omits explaining research findings that the best predictor of future behaviour is past behaviour [17]. Also, the model does not account for volitional control (i.e. perceived behavioural control). It is this particular deficiency which led to the extension of the TRA, called the TPB.
The TPB duly uses attitude, subjective norms, and perceived behavioural control to predict intention and behaviour relatively accurately [18]. The theory assumes that a person"s intention, plus perceived behavioural control, will help predict behaviour with greater accuracy. Both the TRA and TPB assume that behaviour is a conscious decision to act in a particular manner. Yet, there is a critical difference between these two theories. Unlike the TRA, the TPB, depicted in Fig. 1, considers volitional control as a variable. By definition, volitional control means that a person must have the requisite resources, opportunities, and support available to perform a specific behaviour. As depicted in Fig.  1, the TPB comprises three conceptually independent antecedents, namely; attitude, subjective norms, and perceived behavioural control [18].

A. Attitude
Attitude is the degree to which the performance of a particular behaviour is positively or negatively evaluated by an individual [14], [18]. This evaluation is said to encompass two main components; one that is instrumental in nature (i.e. valuable vs. worthless), and one that is related to experiential quality (i.e. pleasant vs. unpleasant) [14]. For instance, schoolteachers may commit to raising cyber-safety awareness at their schools because they might perceive it as valuable, since they are aware of its importance.

B. Subjective Norms
Peers and other key people in an individual"s life have their own perceptions of whether or not an individual should perform a particular behaviour. These perceptions are the individual"s subjective norms. Subjective norms refer to the perceived social or peer pressure to engage or not to engage in a behaviour. It is assumed that subjective norms are determined by normative beliefs concerning the expectations of important referents. These referents are certain groups of people that would approve or disapprove of the individual performing a particular behaviour [18]. For instance, schoolteachers whose parents, spouses, children, or peers approve of them raising cyber-safety awareness at their schools, will be more likely to adopt this behaviour.

C. Perceived Behavioural Control
Perceived behavioural control refers to an individual"s perceptions of whether or not they can perform a particular behaviour. This includes an assessment of the feasibility of performing a particular behaviour, considering the requisite resources, opportunities and support. The construct encompasses the perceived ease or difficulty a person associates with a specific task or behaviour [18]. For instance, schoolteachers who have received adequate cyber-safety awareness training, or those who have the corresponding instructional media or delivery methods, might perceive the task as easy, and will be more likely to raise cyber-safety awareness at their schools.

D. Intention and Behaviour
Intention is an indication of a person's readiness to perform a particular behaviour or action. Intention is considered to be the immediate antecedent of behaviour [14], [18]. This intention is based on the predictors attitude, subjective norms, and perceived behavioural control as depicted in Fig. 1. As applied in this study, intention refers to the schoolteachers" intention of raising cyber-safety awareness at their schools in the next 12 months. Behaviour is a conscious decision to act in a particular manner [14], and in this study, it refers to the schoolteachers" behaviour of raising cyber-safety awareness in their teaching. Intention predicts behaviour [18].

E. Utility of the Theory of Planned Behaviour
The TPB is a well-established and useful model, which has been applied in numerous social scientific studies. Having received substantial research support [19], the main strength of the TPB lies in its application to a variety of behaviours in different contexts, including diverse areas such as health communications, environmental concerns, risk communication, mass transit use, and technology adoption [14]. For instance, [20] used the model to examine the relationship between privacy and online purchasing. Reference [21] used the model to explore the factors which predict students" decisions to select an online course, as opposed to a face-to-face (i.e. contact) course. Reference [22] used the theory to study teachers" decisions to create and deliver lessons using computing technology. Reference [23] used the theory to examine students" intentions to use podcasts as a learning tool.
Another strength of the TPB is its accessibility. This is an important characteristic, because the simplicity of a theory is a quality associated with its strength and utility [24]. The use of the TPB has increased significantly in the last decade, with over a thousand published studies utilising this theory [14]; the current study being only one of these.

III. RESEARCH METHODOLOGY
A survey questionnaire was distributed to teachers in the Sedibeng West and East school districts of the Gauteng Province of South Africa, to explore their cyber-safety awareness intention and behaviour. The study adapted and used a questionnaire which had been piloted and used by [14] in an American doctoral study, confirming the questionnaire"s validity. Across 18 schools, 12 primary and 6 secondary schools in the 2 districts, a total of 180 questionnaires were distributed. Twenty (20) questionnaires were disseminated via email, and 160 were disseminated in hardcopies. Qualtrics was used as the data capture software. The dissemination resulted in a net of 109 completed, usable responses. This translated to a 61% positive survey response rate, with 0 responses from the email dissemination.
SPSS was used as the statistical software for the quantitative data analysis. Seventy-two percent (72%) of the responses were from public, government schools; whilst 28% were from private, independent schools. Multiple regression analysis was used to assess the predictive power of the TPB constructs (i.e. attitude, subjective norms, and perceived behavioural control), in relation to the schoolteachers" intention and behaviour of raising cyber-safety awareness. Cronbach"s coefficient α was used to measure the internal reliability of the items measuring the respective TPB constructs.
The following 7 items (α of 0.898) assessed the attitude of schoolteachers: 1) For me, raising cyber-safety awareness at my school in the next 12 months would be (circle a number for each statement): The following 5 items (α of 0.634) assessed the subjective norms of schoolteachers, using the scale below:

IV. RESULTS AND DISCUSSIONS
In the initial part of the data analysis, a standard multiple regression analysis is used to test the relationship between the TPB constructs (i.e. attitude, subjective norms, and perceived behavioural control), and the teachers" intention to raise cyber-safety awareness at their schools in the next 12 months. In essence, attitude is not revealed as a predictor of intention, whilst subjective norms and perceived behavioural control are. The results of this analysis are depicted in Table I.

A. Attitude and Intention
The multiple regression analysis between attitude and intention reveals that attitude does not predict intention [β = 0.09, n = 109, p > 0.05], nor is the relationship statistically significant [p = 0.204]. These results imply that the schoolteachers" attitude [M = 5.6820] towards raising cyber-safety awareness, is not a predictor of their intentions to raise cyber-safety awareness at their schools in the next 12 months. This means that the positive, neutral, or negative evaluation by teachers of raising cyber-safety awareness does not predict their intention of doing so in the next 12 months. The results could be attributable to the weaker correlation [r = 0.26, n = 109, p < 0.05] between attitude and intention. Contrary to this study and consistent with [25], [26] found that attitude significantly predicted the intention to accept modern agricultural biotechnology [β = 0.84, n = 220, p = 0.000].

B. Subjective Norms and Intention
The multiple regression analysis between subjective norms and intention divulges that subjective norms do predict intention [β = 0.45, n = 109, p < 0.05]. This relationship is statistically significant [p = 0.000]. These results imply that subjective norms (i.e. influences by peers and significant others) [M = 5.1982] in relation to raising cyber-safety awareness, predict the schoolteachers" intentions to raise cyber-safety awareness at their schools in the next 12 months. This means that the perceptions of the teachers" colleagues, parents, spouses, and children, etc., predict their intention of raising cyber-safety awareness in the next 12 months. These results can be attributable to the strong correlation [r = 0.66, n = 109, p < 0.05] between subjective norms and intention. Consistently, [26] found that subjective norms significantly predicted the acceptance of modern agricultural biotechnology [β = 0.40, n = 220, p = 0.000]. These findings of subjective norms being a predictor of intention are consistent with [14] and [27].

C. Perceived Behavioural Control and Intention
The multiple regression analysis between perceived behavioural control and intention shows that perceived behavioural control also predicts intention [β = 0.45, n = 109, p < 0.05] like subjective norms. This relationship is also statistically significant [p = 0.000]. These results imply that perceived behavioural control (i.e. additional resources and opportunities) [M = 5.1982] in relation to raising cyber-safety awareness, predicts the schoolteachers" intentions to raise cyber-safety awareness at their schools in the next 12 months. The results can similarly be attributable to the strong correlation [r = 0.64, n = 109, p < 0.05] between perceived behavioural control and intention. Consistently, [26] found that perceived behavioural control significantly predicted the intention to accept modern agricultural biotechnology [β = 0.53, n = 220, p = 0.000]. These findings of perceived behavioural control being a predictor of intention are consistent with [28] and [25], but contrary to [29].

D. Intention and Behaviour
Reviews of the relationship between intention and behaviour have been conducted by [15], [30], [31], and [32]. In the final part of the data analysis, a standard multiple regression analysis is similarly used to test the relationship between intention and behaviour. The results of this analysis are depicted in Table II. The relationship between the TPB constructs (i.e. attitude, subjective norms, and perceived behavioural control), and the teachers" behaviour of raising cyber-safety awareness in their teaching, is also tested here. The results of this analysis are depicted in Table III. In essence, intention is found to be a predictor of behaviour, as hypothesized by [18]. Furthermore, attitude is not found to be a predictor of behaviour, whilst subjective norms and perceived behavioural control are found to be predictors.  The R 2 value for intention and behaviour is 0.230, which means that intention accounts for 23% of the variation in the schoolteachers" behaviour of raising cyber-safety awareness in their teaching. The multiple regression analysis between intention and behaviour further reveals that intention predicts behaviour [β = 0.48, n = 109, p < 0.05]. This positive relationship is statistically significant [p = 0.000]. These results, depicted in Table II, imply that the schoolteachers" intention [α = 0.88] of raising cyber-safety awareness, is a predictor of their behaviour of raising cyber-safety awareness at their schools and in their teaching. Using multiple regression and AMOS path analysis, [26] confirmed that the intention to accept modern agricultural biotechnology significantly predicted the behaviour of accepting this technology. This finding is consistent with the current study, and supports previous studies that have found intention to be a strong predictor of behaviour [33]- [36]. The relationship between the three main constructs of the TPB (i.e. attitude, subjective norms, and perceived behavioural control) and behaviour is depicted in Table III.

E. Attitude and Behaviour
The R 2 value of 0.242 implies that attitude, subjective norms and perceived behavioural control together account for 24.2% of the variation in the schoolteachers" behaviour of raising cyber-safety awareness in their teaching. The multiple regression analysis between attitude and behaviour divulges that attitude does not predict behaviour [β = 0.14, n = 109, p > 0.05]. This positive relationship is not statistically significant [p = 0.133]. These results, depicted in Table III, imply that the schoolteachers" attitude [α = 0.90] towards raising cyber-safety awareness, is not a predictor of their behaviour of raising cyber-safety awareness at their schools and in their teaching. This finding is in the context of a weak, positive correlation between attitude and intention [r = 0.26, n = 109, p < 0.05], and that attitude does not predict intention [β = 0.09, n = 109, p > 0.05].

F. Subjective Norms and Behaviour
The multiple regression analysis between subjective norms and behaviour discloses that subjective norms predict behaviour [β = 0.25, n = 109, p < 0.05]. This positive relationship is statistically significant [p = 0.014]. These results, also depicted in Table III, suggest that the schoolteachers" subjective norms [α = 0.63] (i.e. views of significant others), are a predictor of their behaviour of raising cyber-safety awareness at their schools and in their teaching. This finding is in the context of a strong, positive correlation between subjective norms and intention [r = 0.66, n = 109, p < 0.05], and that subjective norms do predict intention [β = 0.45, n = 109, p < 0.05].

G. Perceived Behavioural Control and Behaviour
The multiple regression analysis between perceived behavioural control and behaviour unveils that perceived behavioural control similarly predicts behaviour [β = 0.28, n = 109, p < 0.05]. This positive relationship is statistically significant [p = 0.004]. These results, also depicted in Table  III, suggest that the schoolteachers" perceived behavioural control [α = 0.68] (i.e. additional resources and opportunities which facilitate cyber-safety awareness), are a predictor of the schoolteachers" behaviour of raising cyber-safety awareness at their schools and in their teaching. This finding is in the context of a strong, positive correlation between perceived behavioural control and intention [r = 0.64, n = 109, p < 0.05], and that perceived behavioural control also predicts intention [β = 0.45, n = 109, p < 0.05] along with subjective norms.

V. CONCLUSION
The results of this study confirm that social and peer pressure on the teacher by significant others and influential schoolteachers, will positively: 1) influence their intention to raise cyber-safety awareness at their schools in the next 12 months; and 2) influence their behaviour of raising cyber-safety awareness in their teaching. The results of a stepwise regression analysis depicted in Appendix A corroborate this finding, and confirm that subjective norms are the strongest predictor of intention (ΔR 2 = 0.370). In light of this, the South African Department of Basic Education (DBE) is encouraged to identify influential schoolteachers (i.e. phase heads, sports and cultural coordinators, principals and their deputies, etc.), and to groom them into cyber-safety awareness advocates, for the purposes of influencing other teachers or their peers. These advocates should be provided frequent, adequate cyber-safety awareness training, which is school grade-appropriate, for sharing with other teachers.
The results of this study further find that should the South African DBE provide teachers with requisite resources, opportunities, and support which facilitates cyber-safety awareness (i.e. media selection models, time, paraphernalia, training, incentives, recognition, etc.), this will also positively: 1) influence their intention to raise cyber-safety awareness at their schools in the next 12 months; and 2) influence their behaviour of raising cyber-safety awareness in their teaching. The results of the stepwise regression analysis corroborate this finding, and confirm perceived behavioural control as the second strongest predictor of intention (ΔR 2 = 0.168).
The evaluation (e.g., value or pleasure) of raising cyber-safety awareness by teachers was not found to predict the schoolteachers" intention to raise cyber-safety awareness at their schools in the next 12 months, nor did it predict the schoolteachers" behaviour of raising cyber-safety awareness in their teaching. The results of the stepwise regression analysis also corroborate this finding, and confirm attitude as the weakest predictor of intention (ΔR 2 = 0.067). As such, efforts should be expanded to optimise teachers" and learners" attitudes, by enhancing the instrumental evaluation (e.g., value or worth), and experiential quality (e.g., pleasure or fun) of raising cyber-safety awareness at schools. This can be achieved by highlighting the relevance and importance of cyber-safety awareness, and making it more enjoyable through the use of game-based and simulation-led delivery methods. For future research, the current study can be replicated for comparative reasons.