1. Introduction
Traditional learning has been the staple for hundreds of years, from the very founding of higher education institutions (Dewey, 2009). Learners meet together in real time and in a specified location with the educator present. This educator centered approach is the predominant modality of instruction in higher education. Information and Communication Technologies (ICT) can be used to enhance the educational experience with the use of technology delivering the instruction and maintaining learner records.
Distance Learning is the evolution of correspondence courses since learners and educators are separated by time, location, or both in this model of instruction. The materials may be delivered to remote locations via print or ICT, but this form of instruction does not preclude the use of remote classrooms. This form of instruction may be done synchronously or asynchronously. The emphasis of this mode of instruction is the separation of the learner from the educator and not the delivery mechanism [8].
Elearning had its beginnings in late 1994. Elearning describes the ability to electronically transfer, manage, support, and supervise learning and learning materials (Normark and Cetindamar, 2005). Elearning has been viewed as synonymous with webbased learning, Internetbased training, advanced distributed learning, webbased instruction, online learning and open/flexible learning (Khan, 2001). Elearning is the effective learning process created by combining digitally delivered content with learning support services [12]. Many authors have discussed the way in which elearning can be used for the delivery of training, assessment, and support [4].
Learning Management Systems (LMSs) are software tools designed to manage user learning interventions and offer an extensive range of complementary functionality [6]. They are extensively used nowadays and they provide a variety of information and communication channels for the users [30]. Among the features they provide are the development, management, distribution, diffusion and presentation of the educational material as well as tools for the management of users and courses (Folden, 2012).
In the blended learning mode, parts of the instruction are delivered in a traditional format while other parts of it are delivered using ICT. It is this blending of the delivery modality that attempts to use the strengths of both formats to enhance the educational experience. The most used Blended Learning definition is the combination of online and face to face instruction [23,26,27,34,36]. Other studies [7], (Nor Azian 2015) also mention that blended learning is a popular learning approach in higher institutions and more and more universities are adopting blended learning (BL). Blended learning uses the tools of the provincial learning management system (LMS) to teach and support learning in a facetoface class. Therefore, it combines traditional learning, distance learning and elearning in such way, in order to produce high quality courses (fig. 1).
Statistics is a mathematical science concerning the collection, analysis, interpretation or explanation, and presentation of data (Freedman et al., 2007). It is relatively easy to get basic descriptive statistics from statistical software, such as SPSS. It is also used with educational data; this descriptive analysis can provide such global data characteristics as summaries and reports about learner's behavior (Wu and Leung, 2002).
It is not surprising that teachers prefer pedagogically oriented statistics (overall success rate, mastery levels, typical misconceptions, percentage of exercises tackled, and material read) that are easy to interpret. Applying descriptive statistics to educational data is the easiest way that educators prefer [37]. Sometimes, teachers find the finegrained statistics in log data too cumbersome to inspect or too timeconsuming to interpret. Statistical analysis of educational data (logs files/databases) can tell us things such as: where students enter and exit, the most popular pages, the browsers students tend to use, and patterns of use over time (Ingram, 1999); the number of visits, origin of visitors, number of hits, and patterns of use throughout various time periods (Gibbs and Rice, 2003); number of visits and duration per quarter, top search terms, and number of downloads of elearning resources (Grob et al., 2004); number of different pages browsed and total time for browsing different pages [13]; usage summaries and reports on weekly and monthly user trends and activities [18]; session statistics and session patterns [21]; the time a student dedicates to the course or a particular part of it [21]; the learners' behavior and time distribution and the distribution of network traffic over time [38];
Statistical analysis is also very useful to obtain reports assessing how many minutes the student has worked, how many minutes he has worked in a particular day, how many problems he has resolved, and his correct percentage, our prediction of his score, and his performance level [2]. Information visualization uses graphic techniques to help people to understand and analyze data (Mazza, 2006); mean values of attributes analyzed in data to measure mathematical skills [39]; and higher education studentevaluation data [11].
Providing feedback for supporting instructors is a very crucial feature for any learning system. The objective sare to provide feedback to support course educators/teachers/administrators in decision making (about how to improve students' learning, organize instructional resources more efficiently, etc.) and enable them to take appropriate proactive and/or remedial action [16]. It is important to point out that this task is different than data analyzing and visualizing tasks, which only provide basic information directly from data (reports, statistics, etc.). Moreover, providing feedback divulges completely new, hidden, and interesting information found in data [32].
There are studies regarding these features such as: to identify interesting and unexpected learning patterns, which in turn may provide decision lines, enabling teachers to more efficiently organize their teaching structure [33]; to provide feedback to the course instructor about how to improve courseware [25]; to analyze the user's access log in improve eelearning and to support the analysis of trends (Baruque, 2007); to help the teacher to discover beneficial or detrimental relationships between the use of webbased educational resources and student's learning [6]; to reveal information about university students' enrolment(Sanjeev and Zytkow, 1995); learning decomposition and logistic regression to compare the impact of different educational interventions on learning [3]; and usage data analysis to improve the effectiveness of the learning process in elearning systems [19].
The role of educational content in learning is crucial. The key factor to success is high quality educational content, appropriate for learning and able to fulfil course educational aims and objectives [16]. Most of the times it is not an easy task to give feedback to instructors about the online educational content [32]. However, some course characteristics and students' actions may reflect the quality and quantity of the educational content [29].
This study analyzes the effects of blended learning in an institute and its students. The approach goes backward to examine whether this courses usage by the learners is strongly affected by the educational content exposed by the instructors. It proposes the variables Richness and Usefulness, taking into account statistics concerning the courses. These variables aim to help course instructors and/or platform administrators review course usage and find course weaknesses.
2. Research Method
In order to ascertain the views of students about blended learning in TEI of WM, a self administrative questionnaire was chosen to collect the data. Students study the courses for 12 weeks. At the end of the 12th week they were asked to evaluate the three courses they had studied by filling in, an online questionnaire for each course they had studied. The experimental measurements took place at the first semester of 2016 and the qualitative data were gathered.
In the questionnaire we chose closed questions for quick completing and data processing. Moreover, the questions were multiple choice, where it is possible to choose among several predefined answers and most of them were questions of scale or preference, where the degree of preference of the respondent stated. They are the most important questions in a questionnaire, because they allow classification of the views or attitudes of respondents.
The chosen scales in this research are the Likert and the Gutman. In Likert scale, the attitudes have five response categories "Totally agree" = 5, "Agree" = 4, " Neither agree nor disagree" = 3, "Disagree" = 2, "Totally disagree" = 1. In the Gutman scale the attitudes have two response categories "Yes" and "No".
Furthermore we attended the questionnaire marked by clearness and clarity. The questions were short and clear. The negative questions were avoided because they are often misunderstood, since the negative keyword is ignored and the respondent gives an answer that is contrary to his real opinion. Also we did not include questions with double meaning, because they require the respondent to answer two separate ideas with a single answer.
The questionnaire consists of 22 questions and is divided in four parts. The first part consists of 5 questions and examines the students' personal profile. Knowledge of the students profile is important, in order to understand the needs and the personal characteristics of the learners. The second part has 1 question concerning the blended learning. The third part consists of 7 questions and deals with the attitudes of respondents in relation to traditional learning. The fourth part consists of 9 questions concerning the attitudes of respondents in relation to elearning. Eight questions are about richness of the course and nine questions are about usefulness of the course in both education types (traditional and elearning).
The questionnaire was completed by sixty nine (69) learners. All statistical analysis was conducted with the application of the SPSS version 19 software package. The deterministic variables are the questions of our questionnaire. After a thorough analysis, the questionnaire responses show us if the deterministic variables are suitable to measure the hidden variables (richness and usefulness) and how they affect the formation of the students' intention in the adoption and use of blended learning. The analysis is focused on the two hidden variables, each determined by the deterministic variables.
We will investigate possible correlations between specific course properties. Specifically:
1. We compare the educational content in both education types traditional and electronic learning.
2. We correlate gender and with richness and usefulness.
3. We compare courses in terms of richness.
4. We compare courses in terms of usefulness.
5. We discover thoroughly the differences among the courses.
3. Results
The variables rich Content and Useful Content are the hidden variables of the questionnaire. Their values are derived from the deterministic variables which are the questions of our questionnaire. The richness is translated into rich Content and usefulness is translated into Useful Content according to the meaning of the questions.
3.1. Comparison of the Educational Contents between Traditional and Electronic Learning for the 3 Courses
Table 1


Edu Type  N  Mean  Std. Deviation  Std. Error Mean  
Rich Content  1.00  69  3.6087  .91100  .10967 
2.00  69  3.4348  .86566  .10421  
Useful Content  1.00  69  3.6522  .88826  .10693 
2.00  69  3.3333  .90207  .10860 
The content in traditional education (edutype=1) is more useful with a significant statistical difference. Also, they felt slightly richer (no statistically significant difference). This means that they prefer to reuse content from traditional education for their course. The instructors should improve quality the supplied content via Internet.
Table 2
3.2. Correlation between Gender with Richness and Usefulness
Table 3


Gender  N  Mean  Std. Deviation  Std. Error Mean  
Rich Content  1.00  84  3.6548  .85720  .09353 
2.00  54  3.3148  .90750  .12349  
Useful Content  1.00  84  3.6310  .95413  .10410 
2.00  54  3.2778  .78708  .10711 
There is a difference between women and men. The men (gender = 1) find the content more useful (mean 3.63) than women (mean 3.2778) and there is a statistically significant difference. Despite the fact that men perceive it slightly richer, there is a statistically significant difference. This leads to the conclusion that men can exploit this content and learn better from it. However, women have difficulties in understanding; this justifies they do not rate it, as useful as men. However, it is quite a good content since the rating is above 3. Therefore, the instructors should make it more understandable to women, perhaps with more and better examples.
3.3. Comparison of Courses in Terms of Richness
We have the overall course evaluation in both traditional and electronic and the comparison among the courses about the richness of the educational content. In table 4, we have the descriptive statistics for the results
Table 4
For a valid measurement, the criterion Levene Statistic has to be not significant, namely to have equal variances.
From the above table 5, the value is 0.456, so our measure is valid. We can see in the following results, if there is a statistically significant difference among the courses. Through the next table 6, we see that there is a statistically significant difference with p=0.002<0.05 among some courses (even do not know yet which courses).
Table 6


Sum of Squares  df  Mean Square  F  Sig.  
Between Groups  9.599  2  4.799  6.556  .002 
Within Groups  98.836  135  .732  
Total  108.435  137 
Moving towards post hoctests, we see through the table 7 below in which courses there is significant statistical difference about the richness of their content
Table 7
From the above table 7 we see that the first course is better than the second course (difference score by 0.615 in the answers) with statistically significant difference (p=0.009). The second course is also worse than the third course (worse score at 0.84545 units) with a statistically significant difference of 0.002. The courses 1 and 3 have no significant statistical difference between them. This means that in second course, it is necessary to add new content.
3.4. Comparison of Courses in Terms of Usefulness
In the same test for the usefulness of the content, we do not observe significant statistical changes. In the table 8, we have the mean values.
Table 8
For a valid measurement, the criterion Levene Statistic has to be not significant, namely to have equal variances. From the table 9 below, the value is 0.143, so our measure is valid. We can see in the following results, if there is a statistically significant difference among the courses. Through the next table, we see that there is not a statistically significant difference with p = 0.332 > 0.05 among some courses. So, it is not necessary to analyze thoroughly the particular courses.
Table 10


Sum of Squares  df  Mean Square  F  Sig.  
Between Groups  1.823  2  .912  1.112  .332 
Within Groups  110.670  135  .820  
Total  112.493  137 
Table 11
3.5. Thorough Analysis for the Differences among the Courses
We apply ANOVA for all the variables and we have the following descriptive statistics.
Table 12
Table 13


Levene Statistic  df1  df2  Sig.  
t Rich Content  4.014  2  66  .023 
t Useful Content  .363  2  66  .697 
e Rich Content  .971  2  66  .384 
e Useful Content  .866  2  66  .425 
The Levene Statistic (table 13) indicates a problem for traditional Rich Content, since it is statistically significantly differentiated the variation of the sample (with p=0.023<0.05); in this case we should use, instead of ANOVA, the KruskalWallis H test for the analysis of traditional Rich Content.
Table 14
In the table 14, we see the significant statistical difference only for variable e Rich Content (with p=0.003<0.005) between courses.
Table 15
Table 15 shows that the courses 1 and 3 are significantly better than the course 2. Especially, for the first course we have p=0.023; it has better rating at 0.73784 units than second course and third course with p=0.002 is better at 1.13939 units than second course. So, instructors need to add more content in the second course. We also test the traditional Rich content among the courses with KruskalWallisTest since Levene test was significant differentiated at the ANOVA test. The statistics per course are shown at Table 15.
The table 16 shows that there is no significant statistical difference between courses since p=0.447>0.05. So, it is not necessary to add more content.
4. Discussion and Conclusions
This study evaluates the use of blended learning in TEI of West Macedonia with the use of structured questionnaires exposed to the learners. The learners express their attitude about how useful the blended learning is and how this blended means facilitates their studies. It proposes two variables Richness and Usefulness, taking into account statistics concerning the courses. These variables aim to help course instructors and administrators review course usage and find course weaknesses.
There are assessed both traditional and elearning. There were assessed taking into account the gender of the respondents for the three courses.
From the overall comparison of the educational contents between traditional and electronic learning for the 3 courses, it is discovered that the content in traditional learning (edutype=1) is more useful with a significant statistical difference. They think that it is slightly richer than elearning. This means that they prefer to reuse content from traditional learning for their course. The instructors should improve quality the supplied content in elearning.
Significant differences were found between the attitudes of women and men. The men find the content more useful than women with a significant statistically difference. This leads to the conclusion that men can exploit this content and learn better from it. Maybe, women have difficulties in understanding; this justifies they do not rate it, as useful as men. However, it is quite a good content since the rating is above 3. Therefore, the instructors should make it more understandable to women, perhaps with more and better examples.
From the results regarding richness, we conclude that the first course is better than the second course. The third course is also better than the second course with a statistically significant difference. The courses 1 and 3 have no significant statistical difference between them. This means that in second course, it is necessary to add new content.
From the results regarding richness, we conclude that the first course is better than the second course. The third course is also better than the second course with a statistically significant difference. The courses 1 and 3 have no significant statistical difference between them. This means that in second course, it is necessary to add new content. In contrast, from the results regarding usefulness, there were not discovered statistically significant differences among the courses.
From the thorough analysis of the courses, we conclude that the first and third courses are significantly better than the second course. So, the instructors need to add more content in the second course.
The research of this study in no way could be considered full, since the sample of respondents is small, in one only institute and for three only courses. Enlargement and reopening of the investigation in future will pay more accurately the views of learners in order to provide safer conclusions. The sample of respondents will consist of many different institutes. There will be a second questionnaire for educators/facilitators/administrators for a twofold evaluation. A causal model will be adopted which will explain the factors which affect the usage of blended learning. A confirmatory factor analysis will be performed in the model.