Learning Management System-Based Evaluation to Determine Academic Efficiency Performance
Abstract
:1. Introduction
- Behavioral intention: The plan to use or not use specific technology for learning;
- Attitude towards using: The negative or positive feeling of an individual about the accomplishment;
- Perceived usefulness (PU): The degree to which people believe that the use of the system will help them to achieve progress in the performance of their work;
- Perceived ease of use (PEU): The degree of ease associated with the use of the system or technology.
- Edmodo: The use of Edmodo has allowed for evaluations of student learning through the elements that the corporate author of Edmodo facilitates on its platform, [39].
- Schoology: The objective of this free platform is to create a learning strategy for students and to motivate studying online. On this platform, teachers and students are able to create groups and courses, administer resources, set courses created as public or private, integrate resources from external platforms, and present statistics of the progress of each student [40].
- Classroom: As the author indicates, the speed of uploading files is perfect for working inside and outside the classroom. It is free for educational centers, designed for teachers and students, includes email, storage, and forums for collaboration between professors and students in the same class or different classes, and allows users to share or integrate videos, PDFs, and images. The user can also incorporate the URLs of virtual libraries or external resources. Iftakhar evaluated the use of Classroom by surveying students [41].
- Moodle: Moodle is an open-source LMS that allows users to tailor the design to each institution. In this study, the use percentage in undergraduate students was evaluated.
2. Methodology
2.1. Configuration and Architecture
- Institution and course registration: University information and six courses (Figure 1) were registered:(1) OA, (2) SE, (3) SQD, (4) ITS, (5) ITPM, and(6) IC. All were applied to the educational program of engineering in information technology.
- Academic resources: The virtual objects were created as suggested. They should be integrated into courses with LMSs; virtual objects should be created. In this study case, documents were created for each topic: (a) A PDF document with 10 sheets of content, (b) a Word document with 10 sheets of material, and (c) a 3–5 min video and a 10 slide presentation. Each LMS had a VL integrated with links to research work through a website that was shared in chat rooms and used Web 2.0 tools: Hangouts, Skype, and video chat. Each course had between one and two chat rooms. The virtual meetings were only with students for teachers to advise students.
- Virtual classes: The activities were text analysis in discussion forums on the topic studied, a real case analysis where teams of four students were assigned, one final project, and three virtual meetings. The follow-up was carried out by the teacher, who evaluated the participation of students in chat rooms with forums for assistance and homework submission, giving feedback through messages in forums and advice in the virtual chatroom.
- Hardware: The server for the Moodle LMS had a storage capacity of 480 GB and 8.0 GB RAM; in the cases of Edmodo, Google Classroom, and Schoology, the server used had 20 GB of storage. The transmission tower used for the four LMSs was provided by the region supplier. Four TP-Link Archer C9 type access points and 600 Mbps, 2.4 GHz, and 1300 Mbps, 5.0 GHz AC1900 wireless dual-band Gigabit Routers with three removable antennas were used. The workstations consisted of 75 computers with an i5 processor and with 8.0 RAM, a 500 GB hard drive, and an ethernet network card TP-link Tg 3468 at 10/100/100 Mbps and 32 bits.
- Software: For the Moodle LMS, the Linux Centos version 2.7 operating system, a Mysql database with J2EE, and Windows FTP firewall software were used. The Internet used had a download speed of 4.0 Mb/s and an upload speed of 1.0 Mb/s. The browsers used were Internet Explorer 9.0, Mozilla 3.0, and Google Chrome for Edmodo, Google Classroom, and Schoology. The online software used is found at the following URLs: https://classroom.google.com, https://new.edmodo.com, and https://www.schoology.com.
- Types of users: The user profile for the administrator is ideally a computer systems engineer with experience in the use of LMSs, a bachelor’s degree, and experience in implementing courses through an LMS, and the ideal student for this test is one with a background in Information Technology Engineering.
2.2. Statistical Metadata and Coefficients
- Metadata characteristics: The metadata for the Learning Object (LO) described by Wiley included name, author, description, initiative, and LOM, presented by date, format, and filing [45]. A metadata search was performed in each LMS. An LO search record was obtained for each course, and the categorization was done by the type of LOM.
- Statistical coefficients: For the calculation of statistical coefficients, seven questionnaires were used, as shown in Table 2, with evaluation factors and reference authors.
2.3. Factor Evaluation
2.4. Academic Efficiency
- and indicate a major impact A.
- and indicate a normal impact B.
- and indicate a minor impact C.
3. Results and Discussion
3.1. Configuration, Architecture, and LMS Users
3.2. Metadata, Statistical Coefficients, and Performance
3.3. Acceptance
3.4. Academic Efficiency
4. Conclusions
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
Abbreviations
Variable | Description | Abbreviator | Description | |
Correlation coefficient | A | Anxiety | ||
Office Automation (OA) | AI | Anxiety–Innovation | ||
Software Engineering (SE) | app | Application Program for a Particular | ||
Software Quality Development (SQD) | AS | Access Strategies | ||
Information Technology System (ITS) | office automation | |||
Information Technology Project Management (ITPM) | CLTC | Collaborative Learning Through the Computer | ||
Integrator Class (IC). | IC | Integrator Class | ||
Variation coefficient—Accuracy | ICT | Information and Communication Technologies | ||
Cronbach Reliability coefficient | ITPM | Information Technology Project Management | ||
K | Item number | ITS | Information Technology System | |
LMS Classroom | LMS | Learning Management System | ||
LMS Edmodo | LO | Learning Object | ||
LMS Moodle | LOM | Learning Objects with Metadata | ||
LMS Schoology | MU | Mobile Use | ||
Metadata Author | PEU | Perceived Ease of Use | ||
Metadata Date | PLE | Personal Learning Environments | ||
Metadata Format | PU | Perceived Usefulness | ||
Metadata Name | SE | Software Engineering | ||
n | Total number of x | SF | System Factors | |
Q1 | Anxiety and Innovation (AI) | SQD | Software Quality Development | |
Q2 | Utility and Use (UU) | TAM | Technology Acceptance Model | |
Q3 | Tools learnings (TL) | TL | Tools Learnings | |
Q4 | System Factors (SF) | TRA | Theory of Reasoned Action | |
Q5 | Access Strategies (AS) | UTAUT | Unified Theory of Acceptance and Use of Technology | |
Q6 | Virtual Library (VL) | UU | Utility and Use | |
Q7 | Mobile Use (MU) | VL | Virtual Library | |
Sum of variances of the items | VLE | Virtual Learning Environment | ||
Variance of the sum of items | ||||
x | Numerical value of data group | |||
Arithmetic mean | ||||
Reliability Alpha Cronbach | ||||
Standard deviation | ||||
Standard deviation of x | ||||
Covariance between | ||||
Standard deviation of y |
References
- Magalhães, P.; Ferreira, D.; Cunha, J.; Rosário, P. Online vs traditional homework: A systematic review on the benefits to students’ performance. Comput. Educ. 2020, 152, 103869. [Google Scholar] [CrossRef]
- Humanante, P.R.; Garcia, F.J.; Conde, M.A. Personal Learning Environments and Online Classrooms: An Experience With University Students. IEEE Rev. Iberoam. De Tecnol. Del Aprendiz. 2015, 10, 26–32. [Google Scholar]
- Masud, M. Collaborative e-learning systems using semantic data interoperability. Comput. Hum. Behav. 2016, 61, 127–135. [Google Scholar] [CrossRef]
- Denan, Z.; Munir, Z.A.; Razak, R.A.; Kamaruddin, K.; Sundram, V.P.K. Adoption of technology on E-learning effectiveness. Bull. Electr. Eng. Inform. 2020, 9, 1121–1126. [Google Scholar] [CrossRef] [Green Version]
- Polizzi, G. Digital literacy and the national curriculum for England: Learning from how the experts engage with and evaluate online content. Comput. Educ. 2020, 152, 103859. [Google Scholar] [CrossRef]
- Van-Raaij, E.M.; Schepers, J.J. The acceptance and use of a virtual learning environment in China. Comput. Educ. 2008, 50, 838–852. [Google Scholar] [CrossRef]
- Lu, X.; Liu, X.W.; Zhang, W. Diversities of learners’ interactions in different MOOC courses: How these diversities affects communication in learning. Comput. Educ. 2020, 151, 103873. [Google Scholar] [CrossRef]
- Dagger, D.; O’Connor, A.; Lawless, S.; Walsh, E.; Wade, V.P. Service-Oriented e-learning platforms: From monolithic systems to flexible services. IEEE Internet Comput. 2007, 11, 28–35. [Google Scholar] [CrossRef]
- Oprea, C.L. The Internet - A Tool for Interactive Learning. Procedia-Soc. Behav. Sci. 2014, 142, 786–792. [Google Scholar] [CrossRef] [Green Version]
- Soto, C.F.; Senra, A.I.M.; Neira, M.C.O. Ventajas del uso de las TICs en el proceso de enseñanza-aprendizaje desde la óptica de los docentes universitarios españoles. EDUTEC. Rev. Electrónica De Tecnol. Educ. 2009, 29, a119. [Google Scholar]
- Kiennert, C.; De Vos, N.; Knockaert, M.; Garcia-Alfaro, J. The influence of conception paradigms on data protection in e-learning platforms: A case study. IEEE Access 2019, 7, 64110–64119. [Google Scholar] [CrossRef]
- Cabero-Almenara, J.; Arancibia, M.; Del Prete, A. Technical and didactic knowledge of the Moodle LMS in higher education. Beyond functional use. J. New Approaches Educ. Res. (NAER J.) 2019, 8, 25–33. [Google Scholar] [CrossRef] [Green Version]
- Alghamdi, S.R.; Bayaga, A. Use and attitude towards Learning Management Systems (LMS) in Saudi Arabian universities. Eurasia J. Math. Sci. Technol. Educ. 2016, 12, 2309–2330. [Google Scholar]
- Gulzar, Z.; Leema, A.A. Proliferation of E-Learning in Indian Universities through the Analysis of Existing LMS Scenario: A Novel Approach. Indian J. Sci. Technol. 2016, 9, 21. [Google Scholar] [CrossRef]
- Ahmad, N.; Hoda, N.; Alahmari, F. Developing a Cloud-Based Mobile Learning Adoption Model to Promote Sustainable Education. Sustainability 2020, 12, 3126. [Google Scholar] [CrossRef] [Green Version]
- Radif, M.; Fan, D.; McLaughlin, D. Employment Technology Acceptance Model (TAM) To Adopt Learning Management System (LMS) In Iraqi Universities. In Proceedings of the INTED2016 Proceedings, Valencia, Spain, 7–9 March 2016; pp. 7120–7130. [Google Scholar]
- Fishbein, M.; Ajzen, I. Understanding Attitudes and Predicting Social Behavior; Prentice-Hall: Englewood Cliffs, NJ, USA, 1980. [Google Scholar]
- Ajzen, I. Attitudes, traits, and actions: Dispositional prediction of behavior in personality and social psychology. In Advances in Experimental Social Psychology; Academic Press: Cambridge, MA, USA, 1987; pp. 1–63. [Google Scholar]
- Venkatesh, V.; Morris, M.G.; Davis, G.B.; Davis, F.D. User Acceptance of Information Technology: Toward a Unified View. MIS Q. 2003, 27, 425–478. [Google Scholar] [CrossRef] [Green Version]
- Leal Filho, W.; Shiel, C.; Paço, A. Implementing and operationalising integrative approaches to sustainability in higher education: The role of project-oriented learning. J. Clean. Prod. 2016, 133, 126–135. [Google Scholar] [CrossRef]
- Staniškis, J.K.; Katiliūtė, E. Complex evaluation of sustainability in engineering education: Case & analysis. J. Clean. Prod. 2016, 120, 13–20. [Google Scholar]
- Althunibat, A. Determining the factors influencing students’ intention to use m-learning in Jordan higher education. Comput. Hum. Behav. 2015, 52, 65–71. [Google Scholar] [CrossRef]
- Šumak, B.; Heričko, M.; Pušnik, M. A meta-analysis of e-learning technology acceptance: The role of user types and e-learning technology types. Comput. Hum. Behav. 2011, 27, 2067–2077. [Google Scholar] [CrossRef]
- Liaw, S.S. Investigating students’ perceived satisfaction, behavioral intention, and effectiveness of e-learning: A case study of the Blackboard system. Comput. Educ. 2008, 51, 864–873. [Google Scholar] [CrossRef]
- Piccoli, G.; Ahmad, R.; Ives, B. Web-Based Virtual Learning Environments: A Research Framework and a Preliminary Assessment of Effectiveness in Basic IT Skills Training. MIS Q. 2001, 25, 401–426. [Google Scholar] [CrossRef] [Green Version]
- Daradoumis, T.; Puig, J.M.M.; Arguedas, M.; Liñan, L.C. Analyzing students’ perceptions to improve the design of an automated assessment tool in online distributed programming. Comput. Educ. 2019, 128, 159–170. [Google Scholar] [CrossRef]
- Ramirez-Anormaliza, R.; Sabaté, F.; Llinàs-Audet, X.; Lordan, O. Aceptación y uso de los sistemas e-learning por estudiantes de Grado de Ecuador: El caso de una universidad estatal. Intang. Cap. 2017, 13, 548–581. [Google Scholar] [CrossRef] [Green Version]
- Varela, L.A.Y.; Tova, L.A.R.; Chaparro, J. Modelo de aceptación tecnológica (TAM): Un estudio de la influencia de la cultura nacional y del perfil del usuario en el uso de las TIC. Innovar. Rev. De Cienc. Adm. Y Soc. 2010, 20, 187–203. [Google Scholar]
- Davis, F.D. Perceived usefulness, perceived ease of use, and user acceptance of information technology. Innovar. Revista de Ciencias Administrativas y Sociales 1989, 13, 319–340. [Google Scholar] [CrossRef] [Green Version]
- Mee, S. Outreach to international campuses: Removing barriers and building relationships. J. Libr. Inf. Serv. Distance Learn. 2013, 7, 1–17. [Google Scholar] [CrossRef]
- Leppisaari, I.; Lee, O. Modelling digital natives’ international collaboration: Finnish-Korean experiences of environmental education. J. Educ. Technol. Soc. 2012, 15, 244–256. [Google Scholar]
- Yang, J.; Kinshuk, Y.H.; Chen, S.J.; Huang, R. Strategies for Smooth and Effective Cross-Cultural Online Collaborative Learning. J. Educ. Technol. Soc. 2014, 17, 208–221. [Google Scholar]
- Huang, R.; Zhang, Z.; Chen, G.; Xu, C. Online learning: Does learning really happens: Comparison of Chinese and British online learning in intercultural context. Open Educ. Res. 2007, 13, 12–24. [Google Scholar]
- Taylor, S.; Todd, P. Assessing IT Usage: The Role of Prior Experience. MIS Q. 1995, 19, 561–570. [Google Scholar] [CrossRef] [Green Version]
- Navarro, A.; Cigarrán, J.; Huertas, F.; Rodríguez-Artacho, M.; Cogolludo, A. An Integration Architecture of Virtual Campuses with External e-Learning Tools. J. Educ. Technol. Soc. 2014, 17, 252–266. [Google Scholar]
- Yábar, J.M.; Hernández, J.; Roldán, P.L.; Castellà, J. The UAB virtual campus: An essential platform for a European higher education environment. J. Cases Inf. Technol. (JCIT) 2007, 9, 37–48. [Google Scholar] [CrossRef]
- Chanprasitchai, O.A.; Khlaisang, J. Inquiry-Based Learning for a Virtual Learning Community to Enhance Problem-Solving Ability of Applied Thai Traditional Medicine Students. Turk. Online J. Educ. Technol.-TOJET 2016, 15, 77–87. [Google Scholar]
- Solomou, G.; Pierrakeas, C.; Kameas, A. Characterization of educational resources in e-learning systems using an educational metadata profile. J. Educ. Technol. Soc. 2015, 18, 246. [Google Scholar]
- López, J.M.S.; Flores, M.F.; González, J.L.G. Descubriendo Edmodo: Beneficios del microblogging en educación en adultos. Campo Abierto Rev. De Educ. 2012, 31, 53–70. [Google Scholar]
- Clarenc, C.A. Análisis comparativo de LMS; Congreso Virtual Mundial de e-Learnig; Universidad Nacional del Comahue: Neuquen, Argentina, 2013. [Google Scholar]
- Iftakhar, S. Google classroom: What works and how. J. Educ. Soc. Sci. 2016, 3, 12–18. [Google Scholar]
- Cronbach, L.J. Coefficient alpha and the internal structure of tests. Psychometrika 1951, 16, 297–334. [Google Scholar] [CrossRef] [Green Version]
- Ramoni-Perazzi, J.; Orlandoni-Merli, G.; Pérez-Pulido, M.O. Evaluación de indicadores de gestión en las universidades públicas colombianas: Una aplicación de modelos de ecuaciones estructurales. Innovaciencia 2014, 2, 4–16. [Google Scholar]
- Zhang, W.; Wang, Y.; Yang, L.; Wang, C. Suspending Classes Without Stopping Learning: China’s Education Emergency Management Policy in the COVID-19 Outbreak. J. Risk Financ. Manag. 2020, 13, 55. [Google Scholar] [CrossRef] [Green Version]
- Dublin DCMI: DCMI Metadata Terms. Available online: http://www.dublincore.org/specifications/dublin-core/dcmi-terms/ (accessed on 6 June 2019).
- Moore, G.C.; Benbasat, I. Development of an Instrument to Measure the Perceptions of Adopting an Information Technology Innovation. Inf. Syst. Res. 1991, 2, 192–222. [Google Scholar] [CrossRef]
- Vargas-Arcila, A.M.; Baldassarri, S.; Arciniegas, J.L. Análisis de Esquemas de Metadatos para la Marcación de Contenidos Educativos. Form. Univ. 2016, 9, 85–96. [Google Scholar] [CrossRef] [Green Version]
- Ferran-Ferrer, N.; Casadesús, J.; Krakowska, M.; Minguillón, J. Enriching e-learning metadata through digital library usage analysis. Electron. Libr. 2007, 25. [Google Scholar] [CrossRef] [Green Version]
- Castro-GarcíA, L.; LóPez-Morteo, G. An international analysis of the extensions to the IEEE LOMv1.0 metadata standard. Comput. Stand. Interfaces 2013, 35, 567–581. [Google Scholar] [CrossRef]
- Ahn, S.H.; Baek, C.; Lee, S.; Ahn, I.S. Anisotropic tensile failure model of rapid prototyping parts - Fused Deposition Modeling (FDM). Int. J. Mod. Phys. B 2003, 17, 1510–1516. [Google Scholar] [CrossRef]
- Park, N.; Rhoads, M.; Hou, J.; Lee, K.M. Understanding the acceptance of teleconferencing systems among employees: An extension of the technology acceptance model. Comput. Hum. Behav. 2014, 39, 118–127. [Google Scholar] [CrossRef]
- Wallace, L.G.; Sheetz, S.D. The adoption of software measures: A technology acceptance model (TAM) perspective. Inf. Manag. 2014, 51, 249–259. [Google Scholar] [CrossRef]
- Cheng, S.I.; Chen, S.C.; Yen, D.C. Continuance intention of E-portfolio system: A confirmatory and multigroup invariance analysis of technology acceptance model. Comput. Stand. Interfaces 2015, 42, 17–23. [Google Scholar] [CrossRef]
- Yu, F.Y.; Pan, K.J. The Effects of Student Question-Generation with Online Prompts on Learning. J. Educ. Technol. Soc. 2014, 17, 267–279. [Google Scholar]
LMS Configuration | Architecture Resources | Users | ||||
---|---|---|---|---|---|---|
Registration of institution in LMS | Registry courses | Academic resources | Virtual classes | Hardware | Software | Types of users |
Registration of institution data: | Course data: Name, semester, and period. | Learning tools | Assignment activities | Server. | HTTPS, firewall, connection, | Content manager in LMS. |
Name, country, and sector | Style configuration | Virtual libraries | Student follow-up | Transmission tower | internet, | Teacher |
Teacher/student data record: | Interface design | Virtual rooms | Evaluations statistics | Access point | LMS platform, | User |
First name, country | Access code | Workstation | database, browser, | Students | ||
and College year | Inscription | mobile-app |
Questionnaire and Factor Evaluation | Items | Scale Answers | Coefficients Statistics |
---|---|---|---|
Q1-AI [6,19] | 10 | Likert type 1 to 5, where 1 is totally disagree, and 5 is strongly agree | Cronbach reliability. Pearson variation. Spearman correlation. |
Q2-UU [29,34] | 11 | ||
Q3-TL [2,34] | 11 | ||
Q3-SF [19,46] | 8 | ||
Q3-AS [32,47] | 23 | ||
Q3-VL [10,30] | 9 | ||
Q3-MU [22] | 4 |
LMS Metadata Search | Metadata Types | Statistical Coefficients | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
L | L | L | L | M | M | M | M | LMS | C | LMS | C | LMS | C | |
C | 105 | 104 | 105 | 51 | 134 | 82 | 66 | 83 | L | 0.73 | L | 0.69 * | L | |
C | 57 | 83 | 98 | 53 | 165 | 39 | 45 | 42 | L | 0.79 ** | L | 0.32 | L | −0.22 |
C | 45 | 62 | 60 | 36 | 124 | 26 | 5 | 48 | L | 0.74 | L | 0.29 | L | 0.12 ** |
C | 45 | 62 | 60 | 36 | 63 | 76 | 6 | 12 | L | 0.67 * | L | 0.28 ** | L | −0.22 |
C | 43 | 28 | 31 | 25 | 69 | 26 | 25 | 7 | L | 0.72 | L | 0.34 | L | −0.18 |
C | 79 | 76 | 105 | 44 | 152 | 37 | 51 | 64 | L | 0.75 | L | 0.25 *** | L | 0.17 *** |
374 * | 415 ** | 459 *** | 245 * | 707 *** | 286 ** | 198 * | 256 | |||||||
% | 25% | 28% | 31% | 16% | 49% | 20% | 14% | 18% |
Course | Scale Evaluation | |||||||
---|---|---|---|---|---|---|---|---|
0.76 | 0.55 | 0.63 | 0.89 *** | 0.45 | 0.82 ** | 0.85 * | ||
0.63 | 0.69 | 0.66 | 0.66 | 0.62 | 0.68 | 0.72 | ||
0.25 | 0.73 | 0.74 | 0.71 | 0.23 | 0.54 | 0.83 | ||
0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | ||
0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | ||
0.28 | 0.70 | 0.62 | 0.70 | 0.64 | 0.81 | 0.70 | ||
0.55 | 0.60 | 0.65 | 0.69 | 0.39 | 0.82 *** | 0.58 | ||
0.55 | 0.59 | 0.67 | 0.63 | 0.41 | 0.54 | 0.77 | ||
0.47 | 0.62 | 0.72 | 0.37 | 0.22 | 0.71 | 0.80 * | ||
0.52 | 0.58 | 0.61 | 0.60 | 0.50 | 0.84 ** | 0.73 | ||
0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | ||
0.41 | 0.69 | 0.63 | 0.75 | 0.61 | 0.64 | 0.77 | ||
0.59 | 0.64 | 0.67 | 0.63 | 0.50 | 0.76 | 0.76 | ||
0.53 | 0.59 | 0.61 | 0.61 | 0.44 | 0.51 | 0.73 | ||
0.31 | 0.68 | 0.69 | 0.58 | 0.24 | 0.60 | 0.69 | ||
0.58 | 0.57 | 0.60 | 0.60 | 0.40 | 0.67 | 0.73 | ||
0.24 | 0.78 * | 0.73 | 0.73 | 0.22 | 0.78 | 0.80 *** | ||
0.38 | 0.63 | 0.62 | 0.74 | 0.61 | 0.66 | 0.79 ** | ||
0.58 | 0.53 | 0.60 | 0.61 | 0.35 | 0.72 | 0.75 | ||
0.57 | 0.61 | 0.68 | 0.64 | 0.43 | 0.60 | 0.67 | ||
0.37 | 0.76 | 0.73 | 0.79 * | 0.32 | 0.65 | 0.80 ** | ||
0.62 | 0.58 | 0.69 | 0.70 | 0.43 | 0.77 | 0.73 | ||
0.24 | 0.81 *** | 0.71 | 0.72 | 0.30 | 0.70 | 0.77 | ||
0.46 | 0.56 | 0.65 | 0.65 | 0.60 | 0.75 | 0.80 ** | ||
0.60 | 0.53 | 0.61 | 0.62 | 0.47 | 0.74 | 0.68 | ||
0.34 | 0.28 | 0.33 | 0.28 | 0.27 | 0.27 | 0.38 | ||
0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | ||
0.57 | 0.61 | 0.66 | 0.64 | 0.53 | 0.81 ** | 0.68 | ||
0.36 | 0.80 * | 0.77 | 0.87 *** | 0.23 | 0.80 * | 0.79 | ||
0.37 | 0.62 | 0.59 | 0.64 | 0.62 | 0.68 | 0.77 |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Juárez Santiago, B.; Olivares Ramírez, J.M.; Rodríguez-Reséndiz, J.; Dector, A.; García García, R.; González-Durán, J.E.E.; Ferriol Sánchez, F. Learning Management System-Based Evaluation to Determine Academic Efficiency Performance. Sustainability 2020, 12, 4256. https://doi.org/10.3390/su12104256
Juárez Santiago B, Olivares Ramírez JM, Rodríguez-Reséndiz J, Dector A, García García R, González-Durán JEE, Ferriol Sánchez F. Learning Management System-Based Evaluation to Determine Academic Efficiency Performance. Sustainability. 2020; 12(10):4256. https://doi.org/10.3390/su12104256
Chicago/Turabian StyleJuárez Santiago, Brenda, Juan Manuel Olivares Ramírez, Juvenal Rodríguez-Reséndiz, Andrés Dector, Raúl García García, José Eli Eduardo González-Durán, and Fermín Ferriol Sánchez. 2020. "Learning Management System-Based Evaluation to Determine Academic Efficiency Performance" Sustainability 12, no. 10: 4256. https://doi.org/10.3390/su12104256
APA StyleJuárez Santiago, B., Olivares Ramírez, J. M., Rodríguez-Reséndiz, J., Dector, A., García García, R., González-Durán, J. E. E., & Ferriol Sánchez, F. (2020). Learning Management System-Based Evaluation to Determine Academic Efficiency Performance. Sustainability, 12(10), 4256. https://doi.org/10.3390/su12104256