Teaching Digital Competence and Eco-Responsible Use of Technologies: Development and Validation of a Scale

: The environmental impact produced by digital technologies is one of the fundamental contents to be developed by teachers of the 21st century. Different investigations raise the need to redesign education towards sustainable models and promote a critical look at digital technologies. The main objective of this manuscript is to design a valid, reliable and useful scale to measure self-perceived teacher digital competence regarding the eco-responsible use of technologies. The instrument has been designed based on a detailed analysis of the main frameworks for the development of digital competence in Europe: INTEF and DigCompEdu. A content validation process has been followed through the expert judgment method. Subsequently, its reliability and validity are estimated using structural equation modeling techniques. The results obtained guarantee the reliability and validity of the model. Therefore, the need to establish environmental teacher training plans and more awareness about the eco-responsible use of technologies is established.


Introduction
Teaching digital competence is defined as the set of knowledge, skills and attitudes that are necessary for a responsible use of Digital Technologies in the educational context [1]. Therefore, it implies that the teacher masters those didactic and pedagogical principles that allow them to integrate technologies in the learning process to ensure digital competence in students [2][3][4][5]. This is evidenced by the development of different frameworks for digital competence in Europe, such as DigCompEdu [6,7]. Through these frameworks, one begins to become aware of the need for teacher training to work on different aspects of security in the virtual space.
The safe use of digital technologies and environmental problems are issues that have acquired capital importance in our society. In the educational system, this need has been translated into curricular content that is taught in Early Childhood, Primary and Secondary Education; sometimes in a transversal way, other times in complementary activities and in the same way, included in subjects such as Education for Citizenship or Education in values. These arguments raise the need to redesign education and promote a critical and responsible view of digital technologies. Furthermore, although sustainable development is a concept that according to Gómez and Díaz [8], began to be defined at the end of the eighties of the 20th century, and in which it has been worked through different national and international organizations. In the 2030 Agenda for Sustainable Development, the Sustainable Development Goals (SDGs), approved by the United Nations General Assembly in 2015 [9,10], are maintained. All this indicates that it is necessary, now more than ever, to continue researching on how to put them into teaching practice [11].
Including SDGs in the curriculum entails establishing relationships between reality and the classroom, allowing it to be observed in a critical, global and transformative way [12][13][14][15]. Currently, multiple works are being developed related to sustainability in the virtual space in the educational field [16,17]. In them, the need to train at any educational level arises. Currently, children are consumers of technological resources from a very young age. Therefore, teachers need to be trained to prepare future generations to make an eco-responsible use of technologies that guarantees sustainable development and consequently, the optimal health of the planet. This implies moving towards what has been called sustainable education [18,19].
As can be seen, there are many challenges that teachers face. Few doubt the benefits that the use of ICTs can have in any field, but are teachers prepared to assume responsibility for teaching digital security? Do they understand the environmental impact of using the Internet? Do they understand the damage that technological waste causes on our planet? These are issues which must be addressed in the development of educational programs that help raise awareness of the importance of making responsible use of technologies.
The environmental impact produced by technologies depends on the population that uses ICTs, the use made of them and the type of technology that is possessed. The manufacture of technological devices requires the extraction of raw materials and their refinement, which causes an activity with a very large environmental and landscape impact, in addition to the use and abuse of children in the extraction of materials, such as coltan, in developing countries [20]. On the other hand, the increase in data traffic in the network causes an increase in data centers. To store them, these facilities need electricity to maintain the servers that preserve the data and also require powerful cooling systems, which increases electricity consumption significantly, and contributes to the emission of greenhouse gases [21]. It is also necessary to point out the importance of the obsolescence of technological means, since technological waste is a very important polluting element. In this sense, it is necessary to indicate that several factors influence the production of technological waste. On the one hand, the consumerist zeal that our society instigates by pushing consumerism when the useful life of technology can still be extended [22]. On the other hand, dishonest practices by the industry, which leads to planned obsolescence or in other words, to program the useful life of the technology to force users to consume the new ones. This leads to the conclusion that societies have confused the relationship they should have with nature. Leon considers that he has fallen into a materialistic and consumerist pattern, in which the human being seeks to possess more, not only to satisfy basic needs but also to please luxuries, through the use of the resources offered by the Earth [23]. For Ortega, 99% of the things that are bought are thrown away in 6 months; now, is consumed twice what was consumed in the 1950s [24]. A change in awareness and in the industrial model is necessary, although, as already indicated, the industry shows little or no concern for the environment. Likewise, it is necessary to point out that currently we are working on what is called Green ICT. This concept is based on the possibilities offered by ICTs to evolve towards more efficient products and services with lower energy consumption [25].
In recent years, multiple initiatives have emerged that are interested in the research and analysis of environmental effects and technological diffusion in today's societies [20,21,26,27]. Together, the exceptional situation that COVID-19 has caused has forced digital adaptation, intensifying the use of technologies as an alternative to face-to-face activity [28,29]. In this very specific and unusual context, the lack of training in digital security and the eco-responsible use of technologies is much more pronounced. This is why, during the suspension of the in-person activity, the activity in the virtual space has increased notably. This, together with the increase in sales of technological devices, computers, mobile devices and consoles, shows the need to train for the empowerment and awareness of the importance of making a responsible use of technologies.
Many studies have emphasized the measurement of teaching digital competence [6,19,[30][31][32], but no instruments have been developed that measure such a specific dimension of it. The need to know the self-perceived level of competence that teachers have to educate in eco-responsible use has led to the development and design of a scale of measurement of this competence that serves as a basis for teachers' self-reflection. In addition, it allows assessing the strengths and weaknesses/needs in areas of digital competence and the environment. The measurement of this competence with a reliable and valid instrument also allows reflection on how to incorporate the ecological and technological dimension in the training of education professionals. Many authors defend the incorporation of environmental professional competence in the training of professionals as a need generated by the current technological and globalized society [33,34]. Therefore, it is necessary to know the previous knowledge available to the teaching staff to train them in this competence.
Based on these references, the interest of this work is linked to the design and study of the scale of measurement of the self-perceived teaching digital competence in the eco-responsible use of technology. Therefore, the main objective is to analyze the reliability and validity of the scale.

Materials and Methods
The general objective of the questionnaire is to allow a diagnosis of the "security in virtual space" competence. This encompasses environmental impact, cyberbullying, and digital addiction. Therefore, the complete instrument is made up of three scales. This article focuses on carrying out the validation of the first scale, which, as mentioned above, measures the eco-responsible use of technology.

Objectives
As previously indicated, the main purpose of this work was to design a valid, reliable and useful scale to measure self-perceived teacher digital competence regarding the eco-responsible use of technologies. To do this, the following specific objectives are proposed: (a) to guarantee that the information obtained through the analysis and the different interpretations assess the reality to be measured and (b) to analyze the reliability of the instrument. Together, we aim to (c) analyze the construct validity of the scale by analyzing the simple correlations with an exploratory factor analysis and (d) confirm the variables and the structure of the scale by showing the existing relationships between the different dimensions.

Instrument Development and Content Validation
The instrument has been designed based on a detailed analysis of the main frameworks for the development of digital competence in Europe: INTEF [7] and DigCompEdu [6]. From these frameworks, a first version of the instrument was made. The items measured the degree of agreement or disagreement with them on an 8-point Likert scale, where 1 means total disagreement and 8 means maximum agreement.
To ensure that the information obtained through the analysis and the different interpretations value the reality that is to be measured (Objective a), a first content validation is carried out through expert judgment. For this, there are five University Professors in the field of Educational Technology. They assessed the adequacy of the language, the clarity of the items and their adjustment to the proposed objectives on a scale of 1 to 8. In addition, proposals for improvement on each of them are qualitatively collected, if deemed appropriate.
Expert judgment showed that all the items used a language appropriate to their recipients, met the intended objectives and were valued with the highest score regarding their clarity-with the exception of the second, "I have heard about the environmental impact caused by the use of technologies, but could not define it clearly". It is proposed to modify the adversarial contradiction or to eliminate the second part of the sentence. On the other hand, items 3 "I have received training on the environmental impact caused by the use of technologies" and 4 "I have received information on the environmental impact caused by the use of technologies" are considered similar. Although one refers to receiving information and the other to receiving training, it was estimated, by the experts, that it is a very subtle difference to be part of two items. Regarding item 4, the proposal consists of specifying this training, for which reason the word training was changed to "courses", providing a more formal connotation to the item.
In the final instrument, item 2 was divided into two (the second part was included in an independent item (item 3), which was defined as follows, "It could clearly define what is the environmental impact caused by the use of technologies" and the original item 3 was eliminated. This covers the two proposals made without reducing the total number of questions. Then, the scale of the final questionnaire was attached together with the objectives that each item covers (Table 1). 8. I consider myself capable of identifying the technological actions that cause the greatest environmental impact. 9. I am able to establish measures that reduce the environmental impact caused by the use of technologies. 14. As a teacher, I would be able to establish the procedure to follow based on the use of technologies to prevent environmental impact.
Recognize the teaching skills to train students and families in the prevention of this impact (items 10-13) 10. I am able to help another person to manage situations in which the use of technologies is creating a great environmental impact. 11. As a teacher, I am able to inform students about the risks caused by the use of technologies on the environmental impact. 12. As a teacher, I am able to inform families about the risks caused by the use of technologies on the environmental impact. 13. As a teacher, I am able to design actions to prevent the risks caused by the use of technologies on environmental impact aimed at students and families.
After validation, the final questionnaire is adapted for online application through the Google Forms application ( Figure 1) and applied in different centers of Early Childhood, Primary and Secondary Education of the Autonomous Community of Andalusia (Spain). Previously, the management of these centers is contacted by sending them, together with the informed consent, the objectives and the link to the questionnaire for dissemination among the teaching staff during the first quarter of the current year.
Sustainability 2020, 12, x FOR PEER REVIEW 5 of 14 technologies on environmental impact aimed at students and families.
After validation, the final questionnaire is adapted for online application through the Google Forms application ( Figure 1) and applied in different centers of Early Childhood, Primary and Secondary Education of the Autonomous Community of Andalusia (Spain). Previously, the management of these centers is contacted by sending them, together with the informed consent, the objectives and the link to the questionnaire for dissemination among the teaching staff during the first quarter of the current year.

Research Subjects
The questionnaire was answered by a total of 778 Early Childhood Education (25.4%), Primary (41.7%) and Secondary (32.9) teachers from Spain, 530 women and 248 men with an average age of 37 years (Figure 2). For their selection, incidental or convenience criteria were chosen, depending on their availability to answer the questionnaire [35].

Research Subjects
The questionnaire was answered by a total of 778 Early Childhood Education (25.4%), Primary (41.7%) and Secondary (32.9) teachers from Spain, 530 women and 248 men with an average age of 37 years (Figure 2). For their selection, incidental or convenience criteria were chosen, depending on their availability to answer the questionnaire [35].
Regarding the technological profile, the vast majority claimed to have a personal computer (92.1%), a smart mobile phone (99.3%) and an internet connection (97.4%). Likewise, the most common place of internet connection is "anywhere, having mobile internet" (58.7%). At the same time, most of the participants report connecting to the Internet between 5 and 10 h (41.9%), more than 10 h a week (29.8%), between 1 and 5 h (24%) and 1 h or less (4.3%). Regarding the technological profile, the vast majority claimed to have a personal computer (92.1%), a smart mobile phone (99.3%) and an internet connection (97.4%). Likewise, the most common place of internet connection is "anywhere, having mobile internet" (58.7%). At the same time, most of the participants report connecting to the Internet between 5 and 10 hours (41.9%), more than 10 hours a week (29.8%), between 1 and 5 hours (24%) and 1 hour or less (4.3%).

Procedure and Data Analysis
To achieve the proposed objectives (b, c and d), different statistical tests were carried out as suggested by related studies [30][31][32]35]. The reliability, discriminant validity and convergent validity of the questionnaire were calculated using the coefficients: Cronbach's Alpha, McDonald's Omega, composite reliability (CR) and average variance extracted (AVE). Additionally, and to compare these results, the inferential analysis method was used to compare these results. For this, the bivariate correlational analysis technique was used by means of the Spearman correlation coefficient ρ. For its part, the construct validity of the test has been obtained through an exploratory factor analysis (EFA). The method used for the selection of the factors is the principal components method. The factors obtained are rotated orthogonally using the Varimax method with Kaiser normalization. Once the number of factors has been determined, a confirmatory factor analysis (CFA) is performed. Confirmatory factor analysis is used to check whether the theoretical measures of the model are consistent through the modeling of diagrams and the use of structural equations [36]. In other words, it is checked whether the data conform to the hypothetical measurement model produced by the exploratory factor analysis. The method used to contrast the theoretical model has been the weighted least squares (WLS), which provides consistent estimates in samples that do not conform to normality criteria [36]. For this last procedure, AMOS software has been used, capable of revealing hypothetical complex relationships between variables, using structural equation modeling (SEM). At the same time, it has been verified that the data are not normally distributed through a descriptive study in which skewness and kurtosis have been taken into account. The Kolmogorov-Smirnov goodness-offit test confirmed this finding, with a significance (p-value) equal to 0.000 for all items (non-normal distribution).

Results
The reliability of the questionnaire (objective b) was calculated using Cronbach's Alpha and McDonald's Omega coefficients globally. The results show an Alpha index of 0.979 and an Omega index of 0.981. It was established that this index is very high (>0.9), indicating a high degree of the reliability of the questionnaire [37]. Together, the coefficients of composite reliability (CR) and

Procedure and Data Analysis
To achieve the proposed objectives (b, c and d), different statistical tests were carried out as suggested by related studies [30][31][32]35]. The reliability, discriminant validity and convergent validity of the questionnaire were calculated using the coefficients: Cronbach's Alpha, McDonald's Omega, composite reliability (CR) and average variance extracted (AVE). Additionally, and to compare these results, the inferential analysis method was used to compare these results. For this, the bivariate correlational analysis technique was used by means of the Spearman correlation coefficient ρ. For its part, the construct validity of the test has been obtained through an exploratory factor analysis (EFA). The method used for the selection of the factors is the principal components method. The factors obtained are rotated orthogonally using the Varimax method with Kaiser normalization. Once the number of factors has been determined, a confirmatory factor analysis (CFA) is performed. Confirmatory factor analysis is used to check whether the theoretical measures of the model are consistent through the modeling of diagrams and the use of structural equations [36]. In other words, it is checked whether the data conform to the hypothetical measurement model produced by the exploratory factor analysis. The method used to contrast the theoretical model has been the weighted least squares (WLS), which provides consistent estimates in samples that do not conform to normality criteria [36]. For this last procedure, AMOS software has been used, capable of revealing hypothetical complex relationships between variables, using structural equation modeling (SEM). At the same time, it has been verified that the data are not normally distributed through a descriptive study in which skewness and kurtosis have been taken into account. The Kolmogorov-Smirnov goodness-of-fit test confirmed this finding, with a significance (p-value) equal to 0.000 for all items (non-normal distribution).

Results
The reliability of the questionnaire (objective b) was calculated using Cronbach's Alpha and McDonald's Omega coefficients globally. The results show an Alpha index of 0.979 and an Omega index of 0.981. It was established that this index is very high (>0.9), indicating a high degree of the reliability of the questionnaire [37]. Together, the coefficients of composite reliability (CR) and average extracted variance (AVE) are calculated. Table 2 shows the results, as well as the reference values taken for the model fit [38]. All the figures obtained adjust with the reference values. Therefore, the reliability of the model (CR) as well as its average variance extracted (AVE) were demonstrated. Then, the simple correlations of each item with the theoretical or construct dimension were analyzed. The results are shown in Table 3.
All values greater than 0.707 were considered to accept an item as a member of the dimension [39]. Therefore, all the items were integrated within the proposed model.
The construct validity of the test (objective c) was obtained through an exploratory factor analysis (Table 4). Previously, the applicability of factor analysis was confirmed through the Kaiser-Meyer-Olkin (KMO )test, with a statistically significant coefficient of 0.950, and through Bartlett's test of sphericity, with significance (p-value) equal to 0.000 (factor analysis can be applied).
The results, which explain 78.85% of the variance, determine a single theoretical factor: eco-responsible use of technologies (EUIT).
The theoretical model proposed by the exploratory factor analysis (EFA) is contrasted through a confirmatory factor analysis (CFA, objective d). Figure 3 shows the structural diagram proposed with the item-dimension and dimension-dimension correlation indices.
Sustainability 2020, 12, x FOR PEER REVIEW 7 of 14 average extracted variance (AVE) are calculated. Table 2 shows the results, as well as the reference values taken for the model fit [38]. All the figures obtained adjust with the reference values. Therefore, the reliability of the model (CR) as well as its average variance extracted (AVE) were demonstrated. Then, the simple correlations of each item with the theoretical or construct dimension were analyzed. The results are shown in Table 3.
All values greater than 0.707 were considered to accept an item as a member of the dimension [39]. Therefore, all the items were integrated within the proposed model.
The construct validity of the test (objective c) was obtained through an exploratory factor analysis (Table 4). Previously, the applicability of factor analysis was confirmed through the Kaiser-Meyer-Olkin (KMO )test, with a statistically significant coefficient of 0.950, and through Bartlett's test of sphericity, with significance (p-value) equal to 0.000 (factor analysis can be applied).
The results, which explain 78.85% of the variance, determine a single theoretical factor: ecoresponsible use of technologies (EUIT).
The theoretical model proposed by the exploratory factor analysis (EFA) is contrasted through a confirmatory factor analysis (CFA, objective d). Figure 3 shows the structural diagram proposed with the item-dimension and dimension-dimension correlation indices.    Table 3. Correlations of the items with the dimension eco-responsible use of technologies (EUIT).   I1  I2  I3  I4  I5  I6  I7  I8  I9  I10  I11  I12  I13  I14 EUIT  I1  I2  I3  I4  I5  I6  I7  I8  I9  I10  I11  I12  I13  I14 EUIT The factorial loads of the dimensions are between 0.71 and 0.94. These values denote high levels of correlation. Therefore, the results confirm the proposed theoretical model. Table 5 shows the obtained and reference values for the fit of the model [40]: chi-squared (CMIN), goodness-of-fit index (GFI), parsimony goodness-of-fit index (PGFI), normalized fit index (NFI) and the normalized parsimony fit index (PNFI). Table 5. Outcome indices of the model and reference values: chi-squared (CMIN), goodness-of-fit index (GFI), parsimony goodness-of-fit index (PGFI), normalized fit index (NFI) and normalized parsimony fit index (PNFI).

Discussion
The main objective of this manuscript was to design a valid, reliable and useful scale to measure the self-perceived teaching digital competence regarding the eco-responsible use of technologies and it has been achieved. The results obtained after the different statistical tests described above ensure the achievement of the specific objectives proposed at the beginning of this manuscript: (a) Guarantee that the information obtained through the analysis and the different interpretations value the reality that is being measured. The first of the objectives (a) was achieved following a standardized procedure for the construction of this type of scales [6,19,[30][31][32][41][42][43]; which was included from developing preliminary versions (alpha versions), to obtaining a beta version that was subjected to a content analysis by a group of experts in educational technology. After applying the Likert-type scale of eight response options, the experts' report showed that almost all the items were written in a language appropriate to their recipients, that they met the objectives set out in their design and that they were assessed with the maximum score for its clarity of presentation.
Regarding the second of its objectives (b), the reliability of the information collection instrument, the Cronbach's Alpha coefficients (0.979) and its alternative, McDonald's Omega (0.981) show very high (high) rates [37], indicating a high degree of reliability of the questionnaire that we used in our study.
To achieve the third of the proposed objectives (c), the validity of the construct of the scale designed and applied in this study, an exploratory factor analysis was carried out, the results of which explain 78.85% of the variance, determine the factor proposed theory: the eco-responsible use of technologies among the interviewed subjects.
The theoretical model proposed by the previous exploratory factor analysis was contrasted through a confirmatory factor analysis, which corresponds to the fourth objective (d). The factorial loads of the dimensions of the questionnaire were between 0.71 and 0.94, values which imply high levels of correlation. The reference values obtained for the model fit according to Lévy Mangin et al. [40]: the chi-square (CMIN), goodness-of-fit index (GFI), parsimony goodness-of-fit index (PGFI), normalized fit index (NFI) and normalized parsimony fit index (PNFI) indicate that the results confirm the theoretical model that we proposed in our study.
For future research, it is intended to expand the sample size, representing the seventeen states (autonomous communities) that make up Spain. In this way, a profile is obtained that measures and compares the digital competence of teachers at the national level with respect to the eco-responsible use of technologies in Spain. Subsequently, it is proposed to replicate the study at the level of the European Union.
From the application of the scale and after analyzing the results, the training needs of the teaching staff can be detected and reports drawn up. These can be offered to the different administrations to develop training plans, such as the Teachers' Centers in Andalusia (TC). The TCs are the institutions that organize and manage teacher training. Collaboration agreements can even be signed between the University and these organizations to carry out specific training, as regulated by specific legislation [44]. Recently, the need for teacher training in digital competence has been insisted on through the Resolution of 2 July 2020 [45]. In it, the different Autonomous Communities are urged to promote the development of these competences.
The scale of this study contributes to the detection of these needs. Specifically, with regard to Competency 4.3 (Health protection: avoiding health risks related to the use of technology in terms of threats to physical integrity and psychological well-being) and Competency 4.4 (Protection of the environment: take into account the impact of technologies on the environment).
On the other hand, the application of this scale is also viable for teachers in initial training. It can be applied during the first years of the Degrees in Early Childhood and Primary Education. In addition, it can also be implemented as an initial evaluation in the University Master's Degree in Secondary Education Training. As it is done in digital format, it is easily applicable and evaluable, allowing the establishment of specific training actions in this area based on the results obtained.
Specific training can be articulated through innovation projects in the Degrees using a transversal method. This would allow a training complement for all students. It can also be articulated with the creation of free configuration activities. In this case, it would be aimed only at some students who voluntarily want to train in these subjects.
In the case of the master in Secondary Education, it would be necessary to establish coordination mechanisms among all the teachers involved in working on this issue in a transversal way, not only in specific modules of certain specialties.

Conclusions
With this paper, it was expected to awaken the sustainable awareness of readers, as well as that of the researchers and educators of different educational levels in everything related to the eco-responsible use of technologies. This is an important aspect because the concept of sustainable education in Spain is not yet sufficiently established in the school curriculum (it is only included in a superficial way) and it is a very necessary question [18,19]. Until now, this term was seen as associated exclusively with the natural and the most basic concept of ecology.
As some authors claim, there have been several efforts to incorporate sustainable development training into university curricula [46]. There has been greater insistence from the successive guidelines and recommendations of UNESCO to contribute to its development [47][48][49]. However, it is still a long way from achieving a reorientation of the curriculum as would be necessary for the training of future teachers who contribute to the transition to sustainable societies [46]. These proposals have led to sporadic actions in different areas of knowledge. These have mainly focused on Social and Experimental Sciences, both at the national [50][51][52] and international [53,54] level. However, this transformation did not occur in a generalized way, nor as part of the teachers' curriculum. Therefore, it is expected that teacher training in this sense is not as extensive or specific as today's society demands.
The study was carried out and the results obtained provide a reliable and valid instrument that allows generating useful and applied knowledge about the training and awareness that teachers have to advance towards an ecological education. In addition, it offers a solid base that allows us to guide the training plans of teachers on this subject. The self-perception of this competence (the eco-responsible use of technologies) allows reflection on how and in what way to incorporate the ecological and technological dimension in the educational system.
Although this topic has already been studied in previous research [16,17,[41][42][43], these studies have not provided instruments to measure teaching competencies in this area. With this study, the aim was to provide a tool for the diagnosis of teachers where the competencies related to digital citizenship (at the level of responsible citizenship) and teaching competencies (ability to teach on the indicated topic) differ. This distinction is important because a first level of need could be linked to basic knowledge and skills, while for teaching it is necessary to handle more advanced skills that allow discovering the different possibilities of application in varied contexts.
The defense of the incorporation of professional environmental competence in the training of professionals is a need generated by the current technological and globalized society [33,34] and exacerbated by the recent situation that has caused COVID-19 [28,29]. Hence, this instrument represents an important advance to expand in other professional fields.