Improving the Efficiency of Multimedia Learning and the Quality of Experience by Reducing Cognitive Load

Abstract

The design of multimedia teaching materials and the principles of their presentation to students strongly influence distance learning efficiency. The appropriate design of online educational multimedia content became especially important during the period of the COVID-19 crisis. This study presents a methodology for improving the efficiency of multimedia-based distance learning by reducing the cognitive load. Combining the segmentation principle with pauses and testing questions in the design of multimedia teaching materials has a positive influence on reducing the cognitive load in distance learning. Practice as an instructional element was more efficient if it was used between the multimedia educational segments than after the presentation of complete multimedia material. The obtained experimental results show an improvement in distance learning efficiency and an increase in the students’ quality of experience. The improvement of learning outcomes using the proposed methodology of multimedia segmentation combined with pauses and testing questions can be especially important for distance education in situations such as the COVID-19 pandemic.

1. Introduction

The modernization of teaching methods and the design of multimedia teaching materials represent important development trends in modern education. The common goal for both research topics is to enable students to learn faster, more efficiently, and more easily. The use of multimedia and communication technologies is fundamental in the process of the development and design of the modern learning environment [1].

Considering that, teachers who are preparing multimedia teaching materials must be well-acquainted with technical and technological potential to present relevant information more effectively to the students. There are two important aspects to this process. The first one represents the methodology of teaching material design. The second one is the principle of presenting multimedia materials to the students in the classroom or online to provide a more efficient learning process. Therefore, the framework that includes both previous aspects represents a complex research topic aiming to improve the effectiveness of methodology for the design of multimedia learning material and teaching process [2].

The effectiveness of the methodology for the design and presentation of multimedia learning material can be analyzed with two learning outcomes: the level of acquired knowledge of learning objectives and the level of students’ experience with the learning materials and teaching process. To ensure that students achieve learning objectives effectively and to measure the effectiveness of the learning experience the quality of experience (QoE) concept is mostly used. Adequate QoE metrics are often used to analyze the retention of knowledge, students’ engagement, and motivation for learning. The assessment of the students’ QoE with the presented multimedia content was usually conducted immediately after the viewing. This QoE assessment, together with the knowledge test results, is used for the assessment of the proposed learning method efficiency [3]. It is possible to use many different objective and subjective input variables for QoE assessment that illustrate the complexity of this research question.

Online teaching needs to be analyzed by relying on the appropriate usage of multimedia teaching resources, emphasizing the information processing theory. The educational video design became an important research topic that was emphasized especially during the period of the COVID-19 pandemic. The appropriate use of multimedia and different teaching methods provides an effective self-learning process, which was especially important during the period of the COVID-19 crisis [4]. Many approaches can improve the use of instructional design. Some of them are segmenting strategy, different video length usage, pairing multimedia content with user activity, using various quizzes as additional content, etc. [5].

Previous research studies showed the positive influence of the segmentation of instructional videos on the cognitive load, satisfaction, and efficiency of learning. The authors of [6] confirmed with experimental results the positive effects of segmentation as a principle of designing teaching materials. In [7], authors showed that using the videos divided into smaller units provides less cognitive load, as well as increasing user satisfaction and academic achievement. Recent research also presented that the high segmentation of multimedia education material has a significant impact on cognitive load [8].

An important research question represents the effect of combining the different principles and practices of instructional video design. Segmenting strategy can be paired with the other principles of design. The authors of [9] showed that the positive effect of segmenting strategy in instructional video design can be improved if active pauses are used between segments. Recent research investigates the possibility of pairing the segmentation strategy with the different principles of using questions, aiming to improve students’ activity, motivation, and satisfaction. If the segmentation principle is paired with using questionnaires as an additional instructional element presented to students, the methodology for its pairing should be investigated in more detail.

In line with the current research related to the design of educational multimedia materials, the purpose of this study is to investigate whether it is possible to obtain an improvement in the level of acquired students’ knowledge, the enhancement of the learning experience, and a reduction in cognitive load in multimedia learning. Therefore, this study contributes to three research questions in multimedia distance education. The first research question is related to the influence of linear multimedia segmentation on the level of acquired knowledge and students’ learning experience. The second research question aims to investigate the influence of practice, in the form of test questions within segmented multimedia content, on the level of acquired knowledge and students’ learning experience. The third research question investigates the effect of the segmentation of linear multimedia and practice on reducing cognitive load indicated through the level of acquired knowledge of previously unknown educational content.

1.1. Cognitive Load Theory and Instructional Design

Considering the wide usage of dynamic multimedia teaching materials, it is important to give special attention to the technical and cognitive aspects of that usage. Previous research shows that dynamic media enables the improvement of learning when there are limited cognitive resources and when learners’ mental representations are taken into consideration [10]. It is important to stress that the cognitive theory of multimedia learning is fundamental for the implementation of multimedia content presentation and the explanation of the significance of the modality principle in the learning process.

According to this theory, it is necessary to select relevant information and organize it into a verbal and pictorial model before their integration with prior knowledge. In this way, multimedia materials are efficiently processed in two channels, spreading through sensory memory and working memory. Sensory modality implies the receiving of auditory and visual information that further appears in the working memory as verbal and pictorial information models. As the amount of information that can be processed simultaneously is limited in the channels, it is important to use the resources rationally and in the best possible way [11,12]. There are many options to use commonly available technologies to improve the process of online learning using the principles of the cognitive theory of multimedia learning. The authors of [13] analyzed in more detail the guidance for instructional multimedia education designers. This analysis was based on the principles presented in [14].

When multimedia information is presented, it is important to process simultaneously and integrate in the working memory all types of information, so that the materials can be comprehended more effectively. In this case, the significance of working memory in the process of multimedia learning becomes apparent. One of the frequently used models of working memory is Baddeley’s model, where the central executive or the episodic buffer plays a key role in the integration of different types of information [15,16]. The current multi-component model of working memory, shown in Figure 1, represents capacities that do not change by learning. The relation of working memory components to the cognitive system capable of accumulating long-term knowledge is also illustrated [17]. As the capacity of represented buffers is limited, the process of modeling multimedia information and the manner of their presentation is especially important for efficient learning.

The central focus of the cognitive load theory is overcoming the individual limitations of working memory by using instructional manipulations. Managing the working memory load may be facilitated by changes in long-term memory. It is associated with the method of acquiring knowledge and the automation of learning. The demands of multimedia information processing may exceed the processing capacity of the cognitive system. This situation is well-known as cognitive overload. The cognitive load theory introduced three different types of cognitive load imposed on working memory. They are caused by the method of maintaining and processing information. The cognitive load imposed by the design of teaching materials may be extraneous (ineffective) and germane (effective). The third type is an intrinsic cognitive load, imposed by the number of novel elements in the teaching material that should be simultaneously processed [18].

Previous research showed the importance of the selection of the right model for multimedia content design. The relationship between different types of information in multimedia content is important, especially from the perspective of the impact of the modality effect and the redundancy effect [19]. On the other hand, the mode and dynamics of the presentation of multimedia teaching materials are very relevant in the process of education, particularly online education.

Multimedia learning aims to present a mix of static or dynamic verbal information as well as static or dynamic pictorial information. The words can be printed (on-screen text) or spoken (narration), and pictures can be static (illustration, photos, graphs) or dynamic (animation, video). The organization and usage of multimodal data represent important and complex research topics. In everyday life and especially in the multimedia learning process, the human brain processes multisensory information. The multimodal data can also be used for sentiment analysis to identify emotion-related information and, thus, improve users’ experience. Considering that multimodal emotion computing is widely used in human–computer interaction and different applications, such as e-learning, e-medicine, and e-business, the authors of [20] provided a detailed review of the multimodal emotion recognition methods. In a dynamic multimedia learning environment, the transience of information induces a high extraneous load [21]. Therefore, the cognitive load represents a central task in the design of multimedia teaching materials.

1.2. Segmentation of Multimedia Instruction

Different ways of reducing cognitive load have been proposed in previous research [22]. One of the solutions to support students in processing complex information and to reduce the high cognitive load is the segmentation of the presentation. Segmentation gives students time to efficiently process information presented within the previous segment and to be ready for the next segment.

The effect of segmentation may be analyzed regarding the influence of different factors, and one of them is the level of students’ prior knowledge. The authors of [23] have demonstrated that presenting animated teaching material in segments, rather than as continuous material, benefits novices’ learning. If students have a lower level of prior knowledge, segmented animated teaching material provides higher efficiency than the continuous one. For students with higher levels of prior knowledge, this advantage of segmented material slightly disappears. The effects of working memory capacity and the segmentation of multimedia instruction were also examined. The study presented by [24] indicates that the segmentation of multimedia instruction allows learners with lower working memory capacity to recall and apply just as those with higher working memory capacity.

The segmentation process opens new research questions related to the importance of different ways of pauses between segments and their impact on learning outcomes. Those pauses can be used to engage students to write summaries during the pauses [25], to answer different types of questions and quizzes [26], or just to have a little break between lesson segments. There are different strategies for question introduction. The authors of [27,28] analyzed the pop-up questions’ usage within educational videos. They showed that pop-up questions can increase the effectiveness of educational videos by improving student motivation and reducing cognitive load.

Another strategy assumes the use of embedded questions in educational videos. In [26], researchers experimentally analyzed the effectiveness of online video-recorded lectures presented with and without embedded questions. They showed that embedding the questions can increase the effectiveness of consumed videos. The learning mode can also influence student learning. In [29], the authors conclude that embedded questions in pre-class videos in flipped classrooms have limited effect on student learning. Interactive questions and answers (Q&A) can be an important functional resource within multimedia material in the process of developing curriculums and multimedia materials for the implementation of the microlearning approach in distance education. The authors of [30] presented that interactivity significantly influenced students’ engagement and efficiency when microlearning concepts of education are deployed.

Therefore, an important research topic can be the use of embedded questions in videos in traditional, laboratory, or online modes of learning. In this paper, we investigate the pairing of segmenting strategies with the embedding of questions in online educational videos, considering the influence of the position of questions within the educational videos.

2. Materials and Methods

The problem of improving learning efficiency can be analyzed regarding the impact of adequate teaching material modeling and the manner of their presentation to students. The multimedia teaching material modeling and its assessment, in the context of learning efficiency, has to be investigated considering the cognitive theory of multimedia learning and the influence of working memory in the perception of different types of information.

2.1. Segmentation of Multimedia Instruction Research Design

An important aspect of the methodology research design was to include and fulfill five basic steps of the instructional design process: analysis, design, development, implementation, and evaluation. First, the content of the class curriculum was analyzed to point out the main learning outcomes and to identify the influence of the student’s prior knowledge level on the learning process. In the design step, different ways of evaluating student learning efficiency were proposed considering the learning objectives. Having in mind different types of information in multimedia learning material and depending on different multimedia presentation principles, the strategy for instructional material design was developed. Based on the previous, implementation was carefully planned to motivate the students to participate in the experiment. The teaching assistants were involved during the phases of learning and testing to provide the necessary support to the research realization. The evaluation was performed by testing the level of the acquired knowledge and the student’s quality of experience.

The level of acquired knowledge and quality of learning experience depends on two important factors. The first represents the methodology of multimedia content design and the principles of content presentation, and the second represents the influence of working memory and prior knowledge. One of the main goals of this research is to investigate the paired and mutual influence of those two factors on the level of acquired knowledge and quality of experience in distance learning. Therefore, the assessment of multimedia teaching material effectiveness represents the first research topic. In this paper, a general model for the evaluation of the presented multimedia teaching material effectiveness is proposed (Figure 2).

The proposed model consists of three sections. Section 1 includes the methodology for multimedia content modeling and the principles of its presentation. Section 2 consists of the two important factors (working memory and prior knowledge influences) that influence learning efficiency. Section 3 represents the learning efficiency assessment based on measuring the level of acquired knowledge and the assessment of the quality of experience of the presented content [31,32].

As previously described, the segmentation of multimedia materials represents one of the often-used principles in the process of materials modeling. In this paper, the efficiency of the segmentation principle was investigated by considering different methods of multimedia presentation, aiming to reduce the cognitive load and improve learning efficiency and student experience. The impact of questions within the multimedia materials, as a technique to enable segmentation, was also investigated.

Considering the principles of multimedia content modeling and presentation methodology, two models were proposed for conducting the experiment and the assessment of the efficiency of learning. Figure 3 presents the efficiency assessment methodology based on segmented linear multimedia (S-LM) usage.

In this methodology, when the teaching material is segmented and presented in sequences followed by short pauses and relevant questions (Figure 3), a gradual increase in acquired knowledge is expected. This is related to an even and successive increase in overall prior knowledge. Every time a new multimedia sequence is presented to students, the new knowledge contained within the sequence is added to the student’s prior knowledge, impacting the working memory gradually. The gradual increase in the student’s knowledge acquisition can reduce cognitive overload.

The efficiency assessment methodology of non-segmented linear multimedia (NS-LM) usage is presented in Figure 4.

When the methodology within the multimedia content was presented without interruptions, pauses, and embedded questions (Figure 4), prior knowledge that affects working memory did not increase gradually. In that case, prior knowledge represents the initial knowledge that students had before they started consuming non-segmented linear multimedia. In this case, there are no pauses and questions through which working memory could be gradually filled.

This study used several research instruments to describe key research variables. The entry test questionnaire with the personal data was used to identify the level of the student’s prior knowledge and demographic data. Test questions were used to measure the level of acquired knowledge. After the knowledge level testing, the carefully defined questions and the five-level MOS (mean opinion score) scale were used to assess the student’s quality of experience. Considering the goals and methodology used in the research, six questions for assessing the QoE were selected and presented in

Table 1. The questions for assessment of QoE.

Question
What do you think about the video and audio quality?
What is the quality of the lecture presentation?
What is the quality of technical directions for the experimental session?
What is the ability to maintain the attention level during the session?
What is your confidence in correctly answering the questions?
Was it easy to answer the questions?

.

The five-grade MOS scale used for the assessment of the subjective QoE is presented in

Table 2. The five-grade MOS scale QoE assessment.

Mean Opinion Score (MOS)	Explanation
5	Excellent
4	Good
3	Fair
2	Poor
1	Bad

.

2.2. The Participants’ Characteristics and Selection Process

The research was performed in a modern, well-equipped classroom, at the university located in Bosnia and Herzegovina. The multimedia classroom is equipped for advanced multimedia presentations using multiple large-screen TV sets with an installed multi-speaker system and with high-speed internet access.

The participants involved in the experiment attended this particular lecture for the first time and had a similar level of prior knowledge about the presented content. For the experimental part of this study, 40 undergraduate students, third-year studies of computer engineering, were asked to participate. The initial group was divided into two groups of 20 students each, randomly. The students were informed that they could participate fully voluntarily with no adverse effects if they refused and no special benefits regarding course scoring if they accepted. To provide motivation, they were informed that the scores between the two groups would be compared and the group with the better score would be pronounced as winners. Another motivation was that the students expressed a very high level of interest in the topic itself. The objective of this study was to study the effectiveness of linear multimedia teaching material, so the learners’ characteristics were not analyzed as independent variables. To balance the individual differences, a random distribution of participants across sessions was used in this study. A similar approach to participant selection was used in the study [33].

All the participants had a similar level of prior knowledge about the presented content and related material, as they were all students of the same year and the same study program. However, to confirm that both groups of students had comparable prior knowledge and that there was no statistically significant difference between the two groups, participating students were asked to fill out the entry test.

Considering that this study involves student participation, approval from the University’s Ethics Committee was obtained prior to the study. Before experimental testing, the participants were asked to sign the informed consent form.

2.3. Multimedia Teaching Material Design and Presentation Methodology

New multimedia teaching content was specifically designed for this research. To achieve an experience that was as realistic as possible and to have closer contact with the instructor, previously recorded lectures and PowerPoint presentations were used to create multimedia material. The lectures used for this research were recorded using professional equipment. The content and context of the multimedia education material used for the experiment were related to the mobile computing course, because its syllabus has various objectives and different levels of complexity. Therefore, this course covers topics and objectives commonly interesting to all students who participate in testing.

The created multimedia material, by using transition effects in the form of video segments merging, enables students to receive pictorial, audio, and text information from the multimedia presentation, including the instructor’s voice within multimedia. In this manner, students have the impression that the instructor is present in the classroom, to maintain their focus and attention during the learning sessions. By combining audio and visual information in multimedia teaching materials, students can process them simultaneously and efficiently, which aligns with the cognitive theory of multimedia learning [34]. There are several ways to reduce cognitive load in multimedia learning, and one of the methods is to use the segmentation principle. This means that the use of pauses between successive segments enables the achievement of better knowledge transfer.

Two different methodologies of presentation of the teaching materials were used to analyze their effectiveness. In general, the multimedia material can consist of educational multimedia sequences, pauses, additional non-education content, and segments containing questions about the subject matter. The subject of the investigation was the influence of inserting additional content or pauses in the multimedia teaching materials on the students’ knowledge level and QoE.

For this research, two types of multimedia material were created using several educational sequences with different content, context, and complexity. Questions for the assessment of the level of acquired knowledge were also used in the design of testing multimedia material. Two methodologies for the design and presentation of the linear multimedia content are shown in Figure 5.

The first methodology was based on the principle of segmentation of multimedia content (Figure 5a). The segmented linear multimedia (S-LM) was created using educational sequences, test questions, and pauses inserted between them. In this testing methodology, after displaying each multimedia sequence, questions related to that sequence were inserted. Before and after each question, students had a 10 s break. Therefore, the educational sequences and the questions related to them are repeated evenly (Figure 5a).

In the second methodology, the non-segmented linear multimedia (NS-LM) material was presented to students (Figure 5b). In the creation of NS-LM, the smooth transition between the educational sequences was used. This methodology implies that all educational sequences are displayed one after another without pause intervals, followed by a full set of testing questions on the level of knowledge acquired by watching the sequences (Figure 5b). Before starting to answer the questions, students have a 90 s break.

Each educational multimedia sequence (educational sequences 1–5) lasts approximately 5 min. For the S-LM methodology described in Figure 5a, the time allotted for answering each group of questions is 3 min. In the case of NS-LM methodology described in Figure 5b, students have 15 min to answer all questions.

The content and context of the presented materials are part of the compulsory class subject that the students have to take during their studies. Different multimedia content from the same academic field was used to design the education sequences for the S-LM and for the NS-LM.

Multimedia teaching material was designed using educational sequences with two difficulty levels in the type of content. The first level type is the teaching material that is easier for students to comprehend and/or about which they already had some prior knowledge (used as education sequence 1, 3, 4, 5), while the other one is the content unknown to students at the time of presenting and more difficult to learn quickly (used as education sequence 2). All sections contained 3 questions, each correct answer being worth 1 point, with a range of points between 0 and 15 for the complete test.

By applying this method, it becomes possible to determine not only the level of acquired knowledge but also the influence of prior knowledge on the acquired knowledge and the student’s QoE.

2.4. The Assessment Procedures

This paper investigated the scenario in which the students are placed in an adequately equipped classroom, while the previously prepared educational multimedia material is being presented to them. Therefore, this study encompasses the situation when both models of separation (spatial and temporal) between students and instructors exist. Two teaching assistants were present during the multimedia lecture presentation and testing to provide technical support during the distance lecture and to influence students to participate in testing seriously. Before the experiment, the participants were instructed about the testing procedure, and the MOS scale for QoE testing was also described to them.

Two groups of 20 students each, previously described in the subsection related to participants’ characteristics and selection, participated in experimental testing. The first group participated in testing based on the S-LM and the second in the NS-LM methodology of design and presentation.

Before the presentation of the multimedia materials based on the described methodologies, both groups of participants answered entry test questions aiming to estimate their prior knowledge level. The entry test consisted of five questions related to the content of teaching material that would be used in further testing, one question from each of the five sections. After a 5 min recess, students from the two testing groups participated in the next phase of the experiment consisting of an online multimedia presentation and question-answering based on the S-LM and NS-LM methodology described in Figure 5. Immediately after the testing by each of the two methodologies, the participants replied to questions related to the QoE assessment.

In this study, we applied two measures of efficiency. The first one is the level of acquired knowledge represented by the number of correct answers to the test questions. The second one is the student’s QoE represented by the level of the MOS. To assess and rate the QoE, we used the absolute category rating (ACR) test method and five-level MOS scale, as standardized by ITU-T Recommendation P.910. Additionally, the three experts from the relevant field examined the questionnaire and test questions to ensure the content validity.

2.5. Data Analysis

For data analyses in this study, descriptive statistics were calculated including the mean and standard deviation. Data were also tested for normality, and t-tests were performed to establish whether there were a statistically significant differences between the groups. Data analysis was conducted using R statistical software version 4.4.2.

3. Results

After the experimental procedure, the total number of correct answers for the entry test and for both testing methodologies (S-LM and NS-LM), including the correct answers for each question and to the group of questions referring to individual video segments, were obtained and analyzed. The mean values and standard deviations of the number of correct answers were also calculated for both testing methodologies.

The entry test results showed 33% correct answers for the NS-LM group and 28% for the S-LM group. Welch’s t-test produced a p-value of 0.4737, which indicates that there was no significant statistical difference between the means of the two groups. These results validate the initial assumption that the prior knowledge levels of students from the two groups were similar enough.

To analyze the results of the complete test itself, first, it is necessary to establish whether the distributions in both groups were normal. Normal Q-Q plots for both groups are given below in Figure 6.

The results of Shapiro–Wilk tests are W = 0.9620, p-value = 0.5848 for the NS-LM group and W = 0.9575, p-value = 0.4955 for the S-LM group. Combined with Q-Q plots, the data strongly suggest that the distributions are normal.

The further analysis of the experimental results based on the collected data included the calculation of descriptive statistics. The comparative analysis of the number of correct answers obtained during the experiment, in which S-LM and NS-LM methodologies were used, is shown in

Table 3. Comparative analysis of the experiment results.

	NS-LM	S-LM
Percentage of correct answers
Correct answers in the entry test	33.00%	28.00%
Correct answers in education sequence 1 (ES1)	53.33%	75.00%
Correct answers in education sequence 2 (ES2)	40.00%	48.33%
Correct answers in education sequence 3 (ES3)	61.67%	78.33%
Correct answers in education sequence 4 (ES4)	73.33%	88.33%
Correct answers in education sequence 5 (ES5)	76.67%	80.00%
Correct answers in all education sequences (TOTAL)	61.00%	74.00%
Mean value and standard deviation
Mean value of the number of correct answers to all question	9.15	11.10
The standard deviation for the number of correct answers to all question	2.81	1.48

.

The group that participated using the NS-LM methodology had a mean value of 9.15 with a standard deviation of 2.81 for the correct answers to all questions. The group that participated using the NS-LM methodology had a mean value of 11.10 with a standard deviation of 1.48 for the correct answers to all questions. As can be seen from these values, the S-LM group not only achieved a higher score but had a smaller standard deviation as well.

The total number of correct answers and the mean value of the number of correct answers to each question indicate that the effectiveness of multimedia content with intervals and practice questions between educational sequences is much higher than presenting the entire teaching material without interruptions. The standard deviation is much smaller when questions are answered immediately after the educational sequences have been presented than when the questions are answered after the entire teaching material has been presented. This means that it was easier for all users to answer testing questions if the segmented linear multimedia (S-LM) was used, and the agreement among the users was higher.

Previous results can be also seen in the histogram plots of both groups presented in the following figure (Figure 7).

Welch’s two-sample t-test produces the following values: t = −2.7408, df = 28.7960, and p-value = 0.01042, strongly suggesting a statistically significant difference between groups.

The analysis of individual multimedia educational sequences is presented by the graphic representation of the percentage of correct answers to questions referring to each used multimedia segment (Figure 8). Figure 8 clearly shows that the S-LM group achieved better results for each of the sequences. This is illustrated by the higher percentage of correct answers to each group of questions in a case when questions follow immediately after a video sequence has been viewed.

As can be seen from Figure 9 (left), depicting the difference in the percentage of correct answers between the S-LM and NS-LM groups per sequence, the greatest difference was detected in the first sequence and the least in the fifth, with a clear trend of decreasing difference.

This is especially visible if the difference in the percentage of correct answers for each multimedia education sequence is normalized to NS-LM values, as can be seen in Figure 9 (right). For the graph on the left, the values presented are absolute differences in the percentage of correct answers with the NS-LM group as the base level, while the values presented on the right are normalized to values from the NS-LM group, and they show relative improvement in the percentage of correct answers in the S-LM group.

The largest difference in the number of correct answers, when segmented and non-segmented linear multimedia are used, is observed in the first group of questions. A lower level of correct answers to the first group of questions in NS-LM methodology is caused by the longer time between the moment of presenting the material and the moment when the questions are answered. Also, there is a tendency that the difference in the number of correct answers will drop as the time duration of the experiment increases. The previous two comments could be related to the working memory’s influence on learning efficiency.

The impact of the proposed methodologies on learning efficiency considering the complexity of the multimedia education sequence content can also be analyzed. In that sense, education sequence 2 was purposely created so that the subject matter is dense, the questions require careful examination, and the offered answers are very similar in form and meaning. This was carried out as a form of control to check how students respond to described circumstances. As such, we expected lower percentages of correct answers for this section in both test groups, which was confirmed by experimental results.

The analysis of the testing sequence with content of which it is expected that the students have lower previous knowledge (education sequence 2) shows that using the S-LM methodology provides more efficient learning. It is illustrated with more correct answers if the segmented linear multimedia methodology was used.

When it comes to the results of Welch’s two-sample t-test for each pair of sequences in S-LM and NS-LM groups, the p-value was less than 0.05 in sequences 1, 3, and 4, with values of 0.0278, 0.0414, and 0.0238, respectively. For sequences 2 and 5, the p-value was larger than 0.05 with values of 0.3955 and 0.6177, respectively. This was the expected outcome, as answering the questions from sequence 5 had the least difference between the groups, regarding the time passed before the questions needed to be answered and Section 2 being the one designated as hard, with material very unfamiliar to the students.

The efficiency of the proposed learning concepts is also illustrated in Figure 10 by presenting the percentage of participants who correctly answered the questions for both methodologies. Figure 10 shows that, in the case of the S-LM methodology, there were no respondents who scored less than 40% of correct answers, while for NS-LM, there were three respondents who scored less than 40%. The difference between the two methodologies is also illustrated by the fact that, when segmented linear multimedia was used, 18 respondents answered the questions with an accuracy higher than 60%; whereas, in the case of the non-segmented linear multimedia, the number amounted to eight respondents. It is obvious that significantly more participants had a higher percentage of correct answers when the S-LM methodology was used rather than the NS-LM methodology.

The QoE assessment was carried out by applying the previously proposed methodology, and the results obtained are given in

Table 4. The assessment of QoE of the presented multimedia material.

Question	MOSS-LM	MOSNS-LM
What do you think about the video and audio quality?	5	5
What is the quality of the lecture presentation?	5	5
What is the quality of technical directions for the experimental session?	5	5
What is the ability to maintain the attention level during the session?	4.7	4.5
What is your confidence in correctly answering the questions?	4.6	4.4
Was it easy to answer the questions?	4.4	3.7

.

The table contains MOS_S-LM and MOS_NS-LM values for the QoE assessment by applying S-LM and NS-LM methodologies, respectively. The results indicate that the students were very satisfied with the multimedia content’s technical characteristics and were fully acquainted with the testing procedures for each session. Segmented multimedia also improves students’ confidence and experience while answering questions. It can be seen that the QoE was higher in all aspects in the case of the S-LM methodology. It is important to emphasize that the students declared that it is much easier for them to answer the questions if the video segmentation methodology is used with pauses and questions between the segments.

4. Discussion

The education multimedia segmentation concept paired with the pauses and quiz questions, as additional learning content, can improve learning performance and reduce stress. Additionally, the quiz results can identify the level of acquisition of valuable information and relevant content to improve learning. This methodology can also improve different aspects of active self-learning. The efficiency of multimedia-based self-learning was especially important during the COVID-19 pandemic.

Quiz questions have a significant influence on connecting previous knowledge, working memory, and long-term memory. During the phase of question answering, processes in working memory are boosted, because the prior knowledge analysis is activated. The cognitive and mental connections between prior knowledge and working memory are established, and the processing of new information within multimedia educational content is more efficient. Thus, reducing the cognitive load can be achieved using the proposed methodology of multimedia segmentation.

Additionally, using the principle of segmentation paired with repetition in the form of quizzes improves the process of knowledge consolidation. It means that new information within educational multimedia will more efficiently transfer from working to long-term memory. It means that information repetition paired with consolidation improves multimedia learning efficiency.

The usage of pauses represents another aspect of reducing the cognitive load. Short pauses between multimedia segments and question-answering influence working memory status. The refreshment of working memory during the pauses improves the consolidation process, reduces cognitive load, and thus, improves learning efficiency.

Considering the four main components in Baddeley’s model of working memory (see Figure 1), the segmentation principle paired with the quiz and pause usage can improve several important model processes. The methodology proposed in this research can improve the processes within the central executive segment in Baddeley’s model. This segment can be more efficient in administering attention and reducing cognitive load. The process of information consolidation within long-term memory is improved due to the pause usage between multimedia segments. The pause usage also enables better information organization in episodic buffer. Within the pauses, episodic buffer can establish a mental connection between working memory and long-term memory for efficient information consolidation. Therefore, pauses enable better organization through the integration of different segments of information, as well as relaxing the users during the pauses.

Experimental results showed that the S-LM methodology is more efficient than the NS-LM methodology regarding the acquired knowledge as well as the students’ quality of experience. The best indicator of the greater efficiency of the S-LM methodology is that a considerable number of participants (18 students) gave correct answers to over 60% of questions than when NS-LM was applied. This number in the other case was more than twice as small (8 students).

The analysis of the learning process of the content for which participants have none or a low level of prior knowledge, also shows that S-LM methodology provides better results regarding the learning of completely unknown content.

The results of the quality of experience testing show that students’ satisfaction, self-confidence, and experience are higher if linear, segmented multimedia with pauses and knowledge testing questions are used.

This study has some limitations that can be eliminated in the future additional research. First, the success of the testing may depend on the content type, and some topics or courses may be better suited for the methods of multimedia teaching presented in the paper. Second, this experiment is based on a single class session. In future work, we are planning to experiment with the entire course and analyze learning efficiency for longer periods and more voluminous teaching material. Also, the efficiency of learning for different curriculums should be analyzed and compared.

5. Conclusions

This study examined the efficiency of different methodologies for multimedia teaching materials presentation in distance learning. The findings show that the appropriate selection of the methodology of teaching materials presentation results in more efficient learning.

The important aspects of knowledge acquisition in distance learning are described by the cognitive load theory. Considering that, this study emphasizes that the segmentation of multimedia materials in the learning process may improve learning by reducing cognitive load and increasing the quality of the participants’ knowledge. The findings indicate that the effectiveness of teaching materials and learning efficiency are higher if the materials are presented in multimedia segments, with shorter intervals between them and with the use of practice questions. In this manner, students’ prior knowledge is gradually expanded, while working memory is used optimally by processing materials in segments. The successive accumulation of prior knowledge is, thus, accomplished, and the learning process becomes more effective.

Based on experimental results, it can be concluded that the usage of practice questions as an instructional element increases student achievement and gives better results if it is combined with the principle of segmentation multimedia teaching material. A subjective analysis of the efficiency of learning, achieved with the application of these two methodologies of linear multimedia presentation of educational materials, was also carried out. The QoE assessment shows that students are more motivated and interested in learning if the S-LM methodology is applied. Finally, it can be concluded that the methodology of presentation in the form of multimedia sequences, with intervals or pauses between them and using practice questions, gave better results in terms of the presented content effectiveness, the efficiency of the learning process, and the students’ QoE with the learning process. The proposed methodology can be very useful in designing and presenting multimedia materials in the microlearning approach in distance education. The improvement in learning outcomes based on the proposed methodology influences distance learning efficiency, which is especially important in situations such as the COVID-19 pandemic.

Future research is planned to explore the efficiency of multimedia teaching materials presentation in segments, with intervals between them, during which advertising, educational, or entertainment materials are inserted. Demographic factors that influence the results should be investigated in detail. We will extend this research to include wider segments of the population. As continual professional development courses are a common educational possibility for students and working professionals, they may serve as a basis for further testing.