Using the DEMATEL Method to Explore Influencing Factors for Video Communication and Visual Perceptions in Social Media

Abstract

This study used expert interviews and the DEMATEL method to explore the influencing factors affecting the evaluation of the good and cognitive response of video communication and visual perception in social media and for the benefit of facilitation of the implementation in improving the overall video communication and visual quality in social media in the related fields of communication, advertising, and visual design. Correspondingly, it could provide educators and practitioners in the field with a foundation to maximize the effectiveness of allocating resources to these influencing factors. This study explored the influencing factors of video communication and visual perception in social media and evaluated the causality, relevance, and degree of materiality among different factors. The results of this study indicated that the six dimension factors, including like (X8), feeling satisfied (X9), feeling of attractiveness (X11), a good sense of shape contour (modeling) (X12), good visual characteristics (X13), and a good sense of color (X14), showed a high degree of connection (prominence) with other dimension factors. Improving the quality of these six factors could help solve the problem while promoting a good feeling about the other factors.

1. Introduction

Social media are a fast and effective platform for the public to search, share, and send instant messages in real time. The use of social media has become a powerful tool to influence the behavior of the public [1]. It can assist relevant units or organizations in providing the public with accurate information or available resources [2]. Moreover, it can provide a fast and effective communication channel at the proper time [3]. Social media has also played an intermediary role during the COVID-19 pandemic, rapidly disseminating information, acting as a public education tool, preventing misinformation, accelerating risk communication, influencing public behavior and reactions, connecting interdisciplinary teams worldwide, offering video communication solutions without physical contact [4], and many other functions.

When people communicate with video on social media, the lighting and visual perception of color, shapes, and three-dimensionality may affect the viewers’ preferences and visual perception of attraction, attention, preference, and concern. In investigating the visual perception effect of video on social media, people’s assessment of color is influenced by the light source and the visual color impressions produced by different spectral ranges. Likewise, the viewers’ visual characteristics, as well as different visual effects caused by different environmental conditions, are also crucial influencing factors [5].

Sarapin and Morris [6] were surveyed about students’ perceptions of their credibility, professionalism, and approachability in the classroom, as well as mutual connectedness with their instructors, resulting from out-of-classroom socializing with them and teacher self-disclosure on Facebook. Faculty participation in non-academic, online interaction through Facebook shows great promise for augmenting student perceptions.

Kurnia et al. [7] focused on studying visual comfort under daylight conditions in a reading hall of a building. An assessment was conducted by measuring daylight physical variables on daylighting visual comfort. Sun et al. [8] investigated the effect of the lighting environment on work performance. Their presented study suggested that it is necessary to design a personalized illumination environment for a particular workplace. Zhang et al. [9] were to better understand the impact of dynamic lighting on office occupants’ health, well-being, and experience at a living lab. Their results showed that the daytime behavioral impacts were either positive or mixed. They found additional insights into the non-visual impacts of dynamic lighting. Ali et al. [10] deal with the importance of physical environment comfort in the office workplace. They investigated the relationship between employee performance and a comfortable workplace environment. It could be concluded that an uncomfortable environment in an office workplace leads to health-related issues as well as increases the absenteeism rate.

The DEMATEL (Decision Making and Trial Evaluation Laboratory) method [11,12] is used to distinguish complex factors into cause groups and result groups and then generate a visual cause-and-effect relationship diagram. It provides researchers and executives with an easier way to find countermeasures and make decisions about complicated problems. This study adopted the DEMATEL method to achieve the subdivision of the required key factors, committing the best relevant resources to the good rating and response quality of overall video communication and visual perception in social media.

The aims of this paper were to (1) explore the causality of the assessment of video communication and visual perception factors in social media and understand the relationship between factors, (2) identify the critical influencing factors, and (3) discover the crucial influencing factors and present them as a reference for relative fields to facilitate the allocation of resources and arrange the sequencing of solving problems. This study adopted the DEMATEL method to construct a causality assessment model [13] of the video communication and visual perception factors in social media and distinguish the cause-and-effect relationships, as well as clarify the relevance and significance of the interaction from the related matrix and cause-and-effect relationship diagram.

This study investigates the influencing factors that affect the evaluation of the good and cognitive response of video visual perception in social media and discovers the cause-and-effect relationship of different dimensions and criteria between the evaluation models through the analysis too understand the video visual perception factor in social media and crucial influencing factors, as well as to explore the improvements of the quality regarding the evaluation of the response of video visual perception in social media. Therefore, the interactions and structural factors must be considered, and so are the relative importance and relevance between dimensions and factors. For decision makers, the results can be the key factors that should be noticed and identified while not only evaluating the response of video visual perception in social media but also assessing the interrelationship between the dimensions and the criteria. As a result, this study uses the DEMATEL method to identify the interactions between the different dimensions. Through the DEMATEL method, we can understand the direct and indirect relationships between the dimensions and the criteria, as well as the degree of influence and importance among dimensions.

2. Literature Review

2.1. Use of Social Media and Video Communication

The use of social media in education has many functions, including increasing participation; announcing notifications; building stronger learning communities; posting, viewing, broadcasting, and recording online lectures; promoting online discussions; creating stronger online connections; encouraging online sharing and social activities; and advertising and marketing [14].

The development of social media in recent times has resulted in new platforms for enhancing learning, facilitating interactions between learners, instructors, and peers, and participating in new distance learning environments [15,16,17]. The most-used social media platforms used for educational and academic communication are Facebook, WhatsApp, YouTube, and Wikipedia [18].

Regarding the advantages of social media used in education [19,20,21], studies have found it can provide positive and effective contributions and functions in informal academic communication, connectivity, community building, trust maintenance, satisfaction, and developing learners’ social lives. It also shows a positive impact on learners’ engagement and learning experiences [22], increasing interaction among learners, creating knowledge exchange, and providing sharing and collaboration [23,24,25]. Social media can motivate learners to promote group teamwork with peers and associates [26]. Resource sharing is an effective element of social media used for learning [27]. Sharing and collaboration are significant utility and resource tools for community interaction. Interactive communication in social media can benefit students’ learning performance [28,29,30], participation, and learning experiences [6,31].

2.2. Social Media and Visual Perception

Visual perception is the brain’s ability to receive, interpret, and act upon visual stimuli. Perception is based on the following seven elements [32]: visual discrimination, visual memory, visual-spatial relationships, visual form constancy, visual sequential memory, visual figure/ground, and visual closure. Visual perception is the ability to perceive surroundings through the light that enters our eyes. The visual perception of colors, patterns, and structures has been of particular interest in relation to graphical user interfaces because these are perceived exclusively through vision [33].

The visual effect presented by the light source casting relevant light on the user’s emotion [34,35] affects the visual perception of the feelings of comfort, naturalness, dimness, and warmth [36].

Inappropriate lighting conditions may cause fatigue and affect the feeling of comfort [37,38], while good lighting angles present better visualization for video users on social media, allowing for a better feeling of visual comfort and constancy and increasing task performance [39,40].

2.3. DEMATEL Method

The DEMATEL method was initially developed to analyze and resolve complex and interactive relationships between different problems. It allows people to improve their understanding of specific problems as well as interrelated problem factors while investigating feasible solutions and strategies [41,42]. People can study the interrelationships among factors through causality diagrams [43,44] containing horizontal and vertical axes to inspect and analyze the cause-and-effect relationships among factors and their degree of influential intensity [45].

The DEMATEL method can provide a causal model to investigate the interrelationships among factors. The interrelationships among key factors were illustrated by investigating the topic and using the cause-and-effect relationship diagram based on the DEMATEL method [46,47]. The combination of these two methods could provide insight for this study.

According to experts’ opinions [48], it can be used to explore the relevance and importance of factors and contribute to related topics (e.g., to discover potential factors and determine the effect of factors on a particular item, plans, and directions for improvement).

The DEMATEL method is superior to traditional methods commonly used to identify causal relationships and connectors among related factors. These factors are ranked in terms of their importance according to different types of relationships and are evaluated according to the degree and intensity of influence among factors [49].

The DEMATEL method consists of four operational steps [50]: (1) define the scale; (2) create a direct-relation matrix; (3) calculate the normalized direct-relation matrix; and (4) calculate the direct/indirect-relation matrix.

The cause-and-effect relationships among complex system factors are converted into comprehensible structural models [51] and mapped and evaluated using matrices and diagrams in DEMATEL. This method can acquire the causes and effects from the cause-and-effect relationship and enable the decision maker to better grasp the structural relationships [52,53] and interrelationships among system factors, allowing the decision maker to assess the association and strength of influence, degree of importance, and solving priority while discovering better solutions.

2.4. DEMATEL-ANP

ANP (Analytic network process) is a famous decision-making tool. ANP could better reflect the mutual feedback that exists between the hierarchical structure and its internal elements [54] on the basis of the Analytic Hierarchy Process [55]. The ANP can address the internal and external dependencies between alternative options and decision model elements. ANP models consist of interdependent factors, and a cluster is a collection of factors that share a set of characteristics. Each of these clusters has at least one variable associated with another cluster [56]. ANP models can provide more precise tools for dealing with interdependence [57]. The use of ANP is one of the most effective and easy decision tools to model dependence relationships and feedback between components [58]. These dependency relationships, also known as feedbacks, can be modeled using ANP methods, which realistically provide more reliable results to determine the dependency relationships that exist between all factors of the problem (i.e., alternatives, parameters, sub-criteria, and objectives) [59].

The combination of DEMATEL and ANP is often used to solve the evaluation problem of decision-making and multi-criteria objectives. This structure helps to identify the relationships among the core attributes, supporting the decision maker in formulating the plan [60].

DEMATEL-ANP can establish causality, influence the numerical relationship between indicators, and calculate the weights of each indicator. These data were collected through expert interviews and questionnaires. When internal indicators have dependency relationships, DEMATEL can be used to demonstrate the internal dependency relationships through pairwise comparisons [60]. The total influence matrix is considered the unweighted hypermatrix of ANP [61]. After normalization, the weighted hypermatrix was obtained when it is self-multiplying and has converged to become a stable hypermatrix [62,63].

3. Methodology

This study was conducted by interviewing expert scholars to find out the dimension factors from the initial questionnaire, titled “Video Communication and Visual Perception Factors in Social Media”.

The influencing factors of visual perception and response evaluation of visual effects in social media video communication were summarized and analyzed through an interview consultation panel of experts and scholars [64]. The influencing factors were compiled into 4 dimensions (48 sub-dimensions), including (1) visual perception, with 12 sub-dimensions; (2) emotional perception, with 12 sub-dimensions; (3) preference perception, with 11 sub-dimensions; and (4) shape perception, with 13 sub-dimensions. Second, 12 experts and scholars were invited to form a panel to develop the Delphi technique questionnaire. After three Delphi technique questionnaires were conducted, the mean (M), mode (Mo), and standard deviation (SD) of each response were statistically analyzed, and the one-sample Kolmogorov–Smirnov test was used to analyze the appropriateness and consistency of the Delphi technique survey results. The results indicate that 15 sub-dimensions met the criteria of appropriateness and consistency, which were used to establish 15 influencing factors for evaluating visual perception responses to social media visual communication.

After the questionnaire survey, the results of the Delphi survey were analyzed for appropriateness and consistency. The one-sample Kolmogorov–Smirnov test statistical analysis was used for the Delphi survey, and the modifications were based on the opinions of the experts and scholars.

In the statistical results of the first Delphi questionnaire, 38 sub-dimensions met these three elements: (1) mode (Mo) was above 4; (2) mean(M) ≥ 3.5 in criteria for appropriateness test; (3) standard deviation (SD) ≤ 1 in criteria of moderate consistency test and (4) Q ≤ 0.5, the group members have reached high consistency and consensus. Therefore, these sub-dimensions were kept to prepare for the second Delphi study questionnaire. In the statistical results of the second Delphi questionnaire, a total of 22 sub-dimensions met the two elements of (1) high criteria for appropriateness test with mean (M) ≥ 4; and (2) criteria for moderate consistency test with standard deviation (SD) ≤ 0.68. As a result, they were kept for the development of the third Delphi study questionnaire. In the statistical results of the third Delphi questionnaire, a total of 15 sub-dimensions met the three elements of (1) criteria for high appropriateness test with a mean(M) ≥ 4.2; (2) criteria of high consistency test with a standard deviation (SD) ≤ 0.5; (3) the one-sample Kolmogorov–Smirnov test reaching significance in the statistical analysis results; and (4) Q ≤ 0.5, the group members have reached high consistency and consensus.

Afterward, this study used the DEMATEL method to analyze the relevance and causality of developing video communication and visual perception factors in social media, find the core dimensions and directions for improvement, and identify crucial factors for decision making and problem solving.

The DEMATEL method was designed by creating a pair of comparison matrices and influence scales, which were then sent to experts and scholars. After collecting the opinions and questionnaires, the average matrix was calculated, the direct-relation matrix, the normalized direct-relation matrix, and the direct/indirect relation matrix T. The sum of each column and row of matrix T and the values of (D+R) and (D-R) were obtained. At last, the causality model was built based on the obtained results.

Research Steps

Through questionnaires from experts and scholars, the average matrix was calculated to generate the direct-relation matrix, the normalized direct-relation matrix, and the direct/indirect relation matrix T. The research steps are shown in Figure 1.

The DEMATEL Framework and execution steps [65,66] are as follows.

• Step 1: Select important defining factors and design measurement scales

Going by the opinions of experts and literature, we defined the different attributes of the possible effects with the goal of establishing the attributes and measurement scales for the degree of influence [67]. DEMATEL method, to calculate the average matrix, we asked each interviewed expert and scholar to evaluate the direct effect between any two factors using integer scores, with 0 indicating “no influence”, 1 indicating “low influence”, 2 indicating “medium influence”, 3 indicating “high influence”, and 4 indicating “extremely high influence”.

• Step 2: Establish direct-relation matrix X

This direct-relation matrix was filled out by the experts to compare and judge the degree of influence of each dimension in the questionnaire. In the next step, the values defined in Step 1 were placed in the corresponding positions to generate the direct-relation matrix, shown as follows:

$$X=\left[\begin{array}{cccc}0& x_{12}& \dots & x_{1n}\\ x_{21}& 0& \dots & x_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ x_{n1}& x_{n2}& \dots & 0\end{array}\right]$$

In the direct-relation matrix, X_ij indicated the degree of influence of attribute i on attribute j. The diagonal attribute in the direct-relation matrix X was set to 0, meaning the same influence choice of the dimension factor was reflected in the degree of influence on itself and had no influence. After the questionnaire was completed by the interviewed experts, this study obtained the degree of influence between the different factors and created a direct relationship matrix.

• Step 3: Establish normalized direct-relation matrix N

In the direct-relation matrix, this study derived the maximum value from the sum of the columns. The reciprocal of the maximum value was the λ-value, which was used as the normalization basis [68,69]. The formula for establishing the normalized direct-relation matrix [70] was as follows:

$$\lambda =\frac{1}{\begin{array}{c}Max\\ 1\le i\le n\end{array}\left({\sum }_{j=1}^n{x}_{ij}\right)}$$

By multiplying the direct-relation matrix X by the value of λ, the final normalized direct-relation matrix N was derived:

$$N=\lambda X$$

• Step 4: Establish direct/indirect relation matrix T

After obtaining the normalized relation matrix N, the direct/indirect relation matrix T and the total relation matrix were created using the unit matrix I.

In addition, matrix T provided information on how one factor could affect another factor. I was a unit matrix, i.e., a matrix in which the diagonal value was 1 and others were 0. Therefore, the direct and indirect relation matrix T was:

$$T=\underset{k\to \infty }{\lim }\left(N+N^2+\cdots +N^k\right)=N{\left(I-N\right)}^{-1}$$

• Step 5: Calculate the values of D_i and R_j, and the influencing degree of the factors from direct/indirect relation matrix T

After obtaining the direct/indirect matrix T, as it was necessary to calculate the influence of one attribute on the other attributes and the degree of influence, we defined t_ij as attribute i, j = 1, 2 ,…, n of the direct/indirect matrix T. D_i was the sum of row i, which represented the sum of the other attributes affected by attribute i. R_j was the sum of row j, which represented the sum of attribute i affected by the other attributes. D_i and R_j were obtained from direct/indirect matrix T and contained both direct and indirect effects [71,72,73].

$$D_i={\displaystyle \sum }_{j=1}^n{t}_{ij}\left(i=1,2,\cdots ,n\right)$$

$$R_j={\displaystyle \sum }_{i=1}^n{t}_{ij}\left(j=1,2,\cdots ,n\right)$$

• Step 6: Calculate the prominence(D+R) and relation(D−R)

The prominence (D+R) indicated the total degree of influence and the influence of the factor and showed the degree of the relationship between dimensions, the value of which could reveal the core degree and connection of the factor. In addition, (D−R) was the degree of cause, which showed the strength of the influence and the influence of the dimensions and represented the degree of difference between the influence and the influence of the factor [74,75,76].

• Step 7: Draw the cause-and-effect relationship diagram

The four quadrants were divided according to the center points of the X-axis and Y-axis. Cause-and-effect relationship diagrams simplify the complex relationships among factors and show the influence of each factor on the others. This study set the prominence (D+R) as the horizontal axis (X) and set the relation (D−R) as the vertical axis (Y). We obtained the average value of (D+R) and (D−R) by calculating the dimension factors affected by the “video communication and visual perception in social media”. The diagram allowed this study to interpret the degree of influence and direction from each dimension affected by the “video communication and visual perception in social media”.

The (D+R) is the degree of association, and a larger (D+R) means a greater influence on the attribute [77]. If a dimension factor is located to the right of the horizontal axis, it has a higher degree of association, higher importance, and higher priority and contains many different meanings in each quadrant.

The first quadrant is the core factor area. In this quadrant, the higher association between any dimension factor and the other dimension factors shows the intensity of influencing other dimension factors, which is greater than the strength of being influenced by the other dimension factors (high degree of relation). The factors in this region have the highest priority, as they are the key influencing factors.

The second quadrant is the driving factor area. In this quadrant, the associations among dimension factors and other dimension factors are relatively lower than those in the first quadrant area, which has a greater strength of the influence to other dimension factors than being influenced (high degree of relation). The dimension factors in this area have individual independence and can affect relatively few dimension factors compared to the first quadrant. The factors in this region can be ranked as having the second priority.

The third and the fourth quadrant (D−R) values are less than 0, so they belong to the result category, which is the affected category, that is, the result of the cause-and-effect relationship. The dimensions in these regions can be seen as the affected factors, which means the dimensions in the third and fourth quadrant can be improved continuously if they are managed well.

4. Results

The initial questionnaire was first conducted through the interview and research process from twelve experts and scholars, then executed sorting-out in statistics three times, and analyzed the key factors that influence the evaluation of the good and cognitive response of video communication and visual perception in social media according to the high suitability and consistency afterward. Eventually, the results of 15 sub-dimensions (X1–X15) related to the emotional response to video photo-chromatic visual effects in social media were obtained. These 15 sub-dimensions were as follows: feel comfortable (X1), feel relaxed (X2), sense of stability (X3), sense of brightness (X4), feel clear (X5), feel awakening (X6), feeling of a good atmosphere (X7), like (X8), feel satisfied (X9), desire for continuity (X10), feeling of attractiveness (X11), a good sense of shape contour (modeling) (X12), good visual characteristics (X13), good sense of color (X14), and a good sense of three-dimensionality (X15). The last step focused on identifying the crucial key factors as the basic data to promote the evaluation of the good and cognitive response of video communication and visual perception in social media.

4.1. Establishing Direct-Relation Matrix X

Based on the questionnaire responses and statistics from the 12 experts and scholars, we assessed the degree of influence and the cause-and-effect relationships among dimensions and compared the relationship and degree of influence of the dimensional factors, as well as the facet factors with each other and set the influence value to 0. All the experts’ and scholars’ scores were counted to two digits after the decimal point, reaching a direct-relation matrix including 15 dimensional factors of 225 grid matrices and removing 15 ineffective diagonal meshes, resulting in 210 matrices representing different degrees of influence. The results are shown in

Table 1. The initial direct-relation matrix X.

Criteria	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11	X12	X13	X14	X15
X1	0	3.00	2.17	2.08	2.42	2.08	3.00	3.33	3.00	3.00	2.75	2.00	2.00	2.00	2.00
X2	2.92	0	2.08	2.25	2.00	2.00	2.33	2.58	2.58	3.33	2.33	2.33	2.00	2.33	2.33
X3	2.92	3.00	0	2.58	2.25	2.92	2.00	2.33	2.33	2.67	2.67	2.33	2.33	2.33	2.33
X4	2.00	2.00	2.08	0	3.17	3.67	2.00	2.33	2.33	2.33	2.33	3.00	2.67	3.00	2.00
X5	2.33	1.92	1.42	2.08	0	3.00	2.00	2.00	2.00	2.67	2.00	3.00	3.00	3.00	2.00
X6	2.00	1.25	2.08	3.67	3.58	0	1.33	1.33	2.00	1.67	1.67	1.67	1.67	1.67	1.67
X7	3.17	3.00	2.17	2.75	2.67	2.17	0	3.33	3.33	3.33	3.33	2.33	2.33	2.33	2.83
X8	3.83	3.83	3.00	2.67	2.33	3.33	4.00	0	3.75	3.50	4.00	3.00	2.92	3.33	2.33
X9	3.33	3.33	2.33	2.33	1.67	3.00	3.33	3.25	0	3.75	3.00	3.08	2.42	2.75	2.92
X10	2.00	2.67	2.00	1.00	2.00	3.00	2.00	2.75	2.67	0	2.33	2.58	2.08	2.08	2.25
X11	3.50	3.50	3.00	2.33	2.33	4.00	4.00	3.92	3.33	4.00	0	4.00	2.75	3.42	3.00
X12	3.00	3.00	3.00	2.92	3.58	2.58	2.33	3.81	3.08	3.08	3.25	0	3.25	2.33	3.00
X13	3.00	2.75	3.00	2.92	3.58	2.58	2.75	3.08	2.92	3.00	2.58	3.33	0	3.25	2.33
X14	3.33	3.33	3.67	3.58	3.17	3.58	2.67	3.33	3.00	3.17	3.25	3.00	3.33	0	2.33
X15	2.75	2.75	3.50	3.58	3.58	3.17	2.33	2.58	2.42	2.33	2.33	2.67	2.83	1.75	0

.

4.2. Normalized Direct-Relation Matrix N

In this study, the maximum value of the sum of each column in the direct-relation matrix was 47.08, and the maximum value of the sum of each row was 41.83. We used the value of λ, which was the reciprocal of the maximum value of 47.08, as the normalization basis. Using the formula N = λX, the normalized direct-relation matrix N could be derived by multiplying the direct-relation matrix N by the value of λ, as shown in

Table 2. The normalized direct-relation matrix N.

Criteria	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11	X12	X13	X14	X15
X1	0	0.06	0.05	0.04	0.05	0.04	0.06	0.07	0.06	0.06	0.06	0.04	0.04	0.04	0.04
X2	0.06	0	0.04	0.05	0.04	0.04	0.05	0.05	0.05	0.07	0.05	0.05	0.04	0.05	0.05
X3	0.06	0.06	0	0.05	0.05	0.06	0.04	0.05	0.05	0.06	0.06	0.05	0.05	0.05	0.05
X4	0.04	0.04	0.04	0	0.07	0.08	0.04	0.05	0.05	0.05	0.05	0.06	0.06	0.06	0.04
X5	0.05	0.04	0.03	0.04	0	0.06	0.04	0.04	0.04	0.06	0.04	0.06	0.06	0.06	0.04
X6	0.04	0.03	0.04	0.08	0.08	0	0.03	0.03	0.04	0.04	0.04	0.04	0.04	0.04	0.04
X7	0.07	0.06	0.05	0.06	0.06	0.05	0	0.07	0.07	0.07	0.07	0.05	0.05	0.05	0.06
X8	0.08	0.08	0.06	0.06	0.05	0.07	0.08	0	0.08	0.07	0.08	0.06	0.06	0.07	0.05
X9	0.07	0.07	0.05	0.05	0.04	0.06	0.07	0.07	0	0.08	0.06	0.07	0.05	0.06	0.06
X10	0.04	0.06	0.04	0.02	0.04	0.06	0.04	0.06	0.06	0	0.05	0.05	0.04	0.04	0.05
X11	0.07	0.07	0.06	0.05	0.05	0.08	0.08	0.08	0.07	0.08	0	0.08	0.06	0.07	0.06
X12	0.06	0.06	0.06	0.06	0.08	0.05	0.05	0.08	0.07	0.07	0.07	0	0.07	0.05	0.06
X13	0.06	0.06	0.06	0.06	0.08	0.05	0.06	0.07	0.06	0.06	0.05	0.07	0	0.07	0.05
X14	0.07	0.07	0.08	0.08	0.07	0.08	0.06	0.07	0.06	0.07	0.07	0.06	0.07	0	0.05
X15	0.06	0.06	0.07	0.08	0.08	0.07	0.05	0.05	0.05	0.05	0.05	0.06	0.06	0.04	0

.

4.3. Establishing Direct/Indirect Relation Matrix T

The I was a unit matrix. The I-N matrix was created by the unit matrix I, and ${\left(I-N\right)}^{-1}$ was the inverse matrix of I-N. The direct/indirect relation matrix T was obtained from the normalized relation matrix N and ${\left(I-N\right)}^{-1}$ using the formula T = ${\left(I-N\right)}^{-1}$, as shown in

Table 3. The direct/indirect relation matrix T.

Criteria	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11	X12	X13	X14	X15
X1	0.20	0.26	0.22	0.23	0.25	0.25	0.24	0.27	0.26	0.28	0.25	0.24	0.22	0.22	0.21
X2	0.25	0.19	0.21	0.23	0.23	0.24	0.22	0.25	0.24	0.27	0.23	0.23	0.22	0.22	0.21
X3	0.26	0.26	0.18	0.24	0.25	0.26	0.22	0.25	0.24	0.27	0.25	0.24	0.23	0.23	0.22
X4	0.24	0.24	0.22	0.19	0.27	0.28	0.22	0.25	0.24	0.26	0.24	0.25	0.24	0.24	0.21
X5	0.23	0.22	0.20	0.22	0.19	0.25	0.21	0.23	0.23	0.25	0.22	0.24	0.23	0.23	0.20
X6	0.20	0.18	0.18	0.22	0.23	0.16	0.17	0.18	0.19	0.20	0.18	0.18	0.18	0.18	0.17
X7	0.29	0.28	0.24	0.26	0.28	0.27	0.21	0.29	0.29	0.31	0.28	0.26	0.25	0.25	0.25
X8	0.34	0.34	0.29	0.30	0.31	0.33	0.32	0.27	0.33	0.35	0.33	0.31	0.29	0.30	0.27
X9	0.30	0.30	0.25	0.26	0.27	0.30	0.28	0.30	0.23	0.32	0.29	0.28	0.26	0.27	0.26
X10	0.22	0.23	0.20	0.19	0.22	0.25	0.21	0.24	0.23	0.20	0.22	0.23	0.21	0.21	0.20
X11	0.34	0.34	0.30	0.30	0.32	0.35	0.32	0.35	0.33	0.37	0.26	0.34	0.30	0.31	0.29
X12	0.30	0.30	0.27	0.28	0.31	0.30	0.27	0.32	0.30	0.32	0.30	0.23	0.28	0.27	0.27
X13	0.30	0.29	0.27	0.28	0.31	0.29	0.27	0.30	0.29	0.31	0.28	0.29	0.21	0.28	0.25
X14	0.32	0.32	0.30	0.31	0.32	0.33	0.28	0.32	0.31	0.33	0.31	0.30	0.30	0.23	0.26
X15	0.27	0.27	0.26	0.28	0.29	0.29	0.24	0.27	0.26	0.28	0.26	0.26	0.25	0.23	0.19

.

4.4. Calculating the Prominence (D+R) and Relation Degree (D−R) and Drawing the Cause-and-Effect Relationship Diagram

Indirect/indirect relation matrix T, D was the sum of each column, and R was the sum of each row. D_i and R_j were inferred from the direct/indirect relation matrix T, which contained the direct-influencing and indirect-influencing matrix. Making statistics of D+R and D−R values, setting D−R as the vertical axis, D+R as the horizontal axis, which is the starting position of the vertical axis in the horizontal axis, and the average value of D+R is 7.78. The values of D+R and D−R provided the coordinate points in the quadrant, as shown in

Table 4. Prominence and relation results obtained by using the DEMATEL method.

Criteria		D	R	D+R	D−R
X1	Feel comfortable	3.61	4.06	7.67	−0.45
X2	Feel relaxed	3.46	4.01	7.47	−0.55
X3	Sense of stability	3.59	3.58	7.17	0.01
X4	Sense of brightness	3.60	3.79	7.39	−0.19
X5	Feel clear	3.37	4.05	7.42	−0.68
X6	Feel awakening	2.79	4.15	6.94	−1.36
X7	Feeling of a good atmosphere	4.02	3.69	7.71	0.33
X8	Like	4.68	4.10	8.78	0.58
X9	Feel satisfied	4.16	3.97	8.13	0.19
X10	Desire for continuity	3.27	4.33	7.60	−1.06
X11	Feeling of attractiveness	4.80	3.91	8.71	0.89
X12	A good sense of shape contour (modeling)	4.32	3.91	8.23	0.41
X13	Good visual characteristics	4.21	3.67	7.88	0.54
X14	Good sense of color	4.54	3.68	8.22	0.86
X15	Good sense of three-dimensionality	3.91	3.45	7.36	0.46
	Mean			7.78	0.00

, and finally created the cause-and-effect relationship diagram graph model, as shown in Figure 2.

4.5. The Causal Relationship between the Four Main Dimensions

By using the DEMATEL method, standardized T- matrix are used to obtain T^S-matrix by integrating them into the ANP structure analysis. Drawing the plot of the strength of the interaction between dimensions according to the T^S-matrix.

In the questionnaire, the 15 sub-dimensions are divided into four main dimensions, which are (1) visual perception, (2) emotional perception, (3) preference perception, and (4) shape perception. The sub-dimensions of Dimension 1 (Visual perception) are feel-comfortable, feel-relaxed, sense of stability, sense of brightness, feel-clear, and feel-awakening; the sub-dimensions of Dimension 2 (Emotional perception) are feeling of a good atmosphere; The sub-dimensions of Dimension 3 (Preference perception) are like feel-satisfied, desire for continuity, feeling of attractiveness; the sub-dimensions of Dimension 4 (Shape perception) are a good sense of shape contour, good visual characteristics, good sense of color, good sense of three-dimensionality.

The T^S-matrix is obtained from the standardized T-matrix of the four dimensions, which is shown in

Table 5. The total relational impact T^S-matrix for the standardization of the four main dimensions.

Main Dimension	Visual Perception	Emotional Perception	Preference Perception	Shape Perception	Total	Visual Perception	Emotional Perception	Preference Perception	Shape Perception
Visual perception	2.21	2.33	2.51	2.29	9.38	0.00	0.00	0.00	0.00
Emotional perception	2.87	2.57	3.08	2.71	11.18	2.87	0.00	3.03	2.71
Preference perception	3.00	2.96	2.87	2.85	11.68	3.00	2.96	2.87	2.85
Shape perception	2.85	2.75	2.95	2.46	11.01	2.85	2.75	2.95	0.00

. The standardized T-matrix values indicate the influencing degree of inner dependence and outer dependence among the evaluation of four dimensions. T^S-matrix can be used to draw the plot of the strength of the interaction between dimensions, shown in Figure 3.

The D-value and R-value were obtained from the T-matrices of the four main dimensions in Table 5 and calculated the degree of influencing of D+R and D−R, shown in

Table 6. The D-value and R-value of the four main dimensions.

Criteria	D	R	D+R	D−R
Visual perception	9.35	10.93	20.28	−1.59
Emotional perception	11.18	10.61	21.80	0.57
Preference perception	11.68	11.37	23.05	0.31
Shape perception	11.01	10.31	21.32	0.70
Mean			21.61	0.00

. Set (D+R) as the horizontal axis, (D−R) as the vertical axis Y, (D+R) and (D−R) as the average value of the central point of the horizontal axis X and vertical axis Y, and delineate the four quadrants of the main dimension as shown in Figure 4.

5. Discussion

The study discusses the influencing factors regarding the quality of visual perception and good and cognitive response in social media. The analysis revealed a causal relationship between the different dimensions and the criteria from the assessment model. This study offered suggestions based on the results of influencing factors regarding the quality of visual perception and good and cognitive response in social media, and it is necessary to think about how the quality of the Preference perception affects Visual perception. (X8) Like is the most relevant and important influencing factor. The results show that in order to improve the quality of visual perception and good and cognitive response in social media, (X8 )like, (X9) feel satisfied, (X11) feeling of attractiveness, (X12) a good sense of shape contour (modeling), (X13) good visual characteristics, and (X14) good sense of color are the factors that require attention.

This study has the following study limitations. First, it is possible that regional cultural constraints may influence the views of scholars and experts on this topic. For research in the future, it could extend the findings of this study to explore the evaluation views of experts from different regions on this topic. Second, this study only discusses the views on this topic based on the opinions of experts and scholars. Therefore, subsequent researchers can study the causal relationships obtained from the dimensions of crucial influencing factors constructed by scholars and experts, establish a research framework of inter-variate causality, and continue investigating users and learners in related fields for further examining whether users’ or learners’ views are consistent with those of experts and scholars.

6. Conclusions and Recommendations

6.1. Conclusions

6.1.1. Analysis of the Prominence (D+R) and Relation (D−R), as Well as the Cause-and-Effect Relationship Diagram

Below are the cause-and-effect relationships and influence degrees of the 15 factors affecting the visual perception and cognitive response in social media according to the analysis results.

• (1) High degree of prominence and a strong degree of relation:

Six factors were listed as the cause criteria, including like (X8), feeling satisfied (X9), feeling of attractiveness (X11), a good sense of shape contour (modeling) (X12), good visual characteristics (X13), and a good sense of color (X14). These factors had a high degree of connection (prominence) with the other structural factors. They were also the core dimensions affecting the other factors because the strength that affects other factors is greater than the strength that is affected by other structural factors (high degree of relation). As a result, they were the driving forces for problem solving and could be classified as having high priority.

• (2) Low degree of prominence and weak degree of relation:

Six factors fell into this category, including feeling comfortable (X1), feeling relaxed (X2), sense of brightness (X4), feeling clear (X5), feeling awakening (X6), and desire for continuity (X10). These factors were slightly influenced by other factors; therefore, they could be categorized as influenced factors, which belong to the result of the cause-effect relationship.

• (3) Low degree of prominence and a high degree of relation:

Three factors belonged to this category, including a sense of stability (X3), a feeling of a good atmosphere (X7), and a good sense of three-dimensionality (X15). These three factors were individually independent and could affect relatively few structural factors. They could be influenced by other factors. However, they could not be directly improved due to the low degree of connection with other factors.

6.1.2. Analyzing the Quadrants Using the Cause-and-Effect Relationship Diagram

The (D−R) value less than 0 is a negative value, having a tendency to the result category, feel comfortable (X1), feel relaxed (X2), sense of brightness (X4), feel clear (X5), feel awakening (X6), and desire for continuity (X10), these can be influenced by other factors, pertaining to the category of influenced factors, which belong to the result of the cause-effect relationship. They were not the main influencing factors and did not directly promote the good feeling or quality of the visual perception and cognitive response in social media.

Three factors: a sense of stability (X3), a feeling of a good atmosphere (X7), and a good sense of three-dimensionality (X15) belonged to the cause of the cause-effect relationship. Among them, the value for a sense of stability (X3) in the vertical axis (D−R) was very close to 0. Although the feeling of a good atmosphere (X7) and a good sense of three-dimensionality (X15) could also affect the quality of the other factors, when referring to resource investments, the benefits were not as critical as compared to X8, X9, X11, X12, X13, and X14.

A (D−R) value greater than 0 is a positive value, indicating the factor belongs in the cause category and has a cause-and-effect relationship. A higher (D+R) value indicates a higher degree of influence on other factors; thus, it also shows a higher degree of importance for improving or enhancing the visual quality of social media in the video. The results indicated like (X8) feel satisfied (X9), feeling of attractiveness (X11), a good sense of shape contour (modeling) (X12), good visual characteristics (X13), and a good sense of color (X14), affected relatively more aspects. The cause-and-effect diagram indicated they were the core factors affecting the other factors. In order to ameliorate the cognitive response of visual perception in social media, building the video quality that includes the following factors: like (X8), feeling satisfied (X9), feeling of attractiveness (X11), a good sense of shape contour (modeling) (X12), good visual characteristics (X13), and a good sense of color (X14), can be used as the key factors enhancing the feeling or quality of visual perception and good and cognitive response in social media. Improving the quality associated with these six factors could contribute to solving the problem and improve the good feeling about other factors.

X6 has the smallest negative value on the Y-axis and is most easily influenced by other factors; X11 has the highest value on the Y-axis and is the driving factor for realization; X8 has the highest value on the X-axis, which has the strongest influence on relevance and connectivity and can easily affect other factors; the others are X11, X12, X14, X9, and X13 sequentially, which indicates that these factors play a dominant role in influencing the quality of visual perception and good and cognitive response in social media, contributing to the greater impact on other factors.

6.1.3. Analyze the Cause-and-Effect Relationship of Mutual Influence between the Four Main Dimensions

Fifteen sub-dimensions are divided into four main dimensions. Analyze the interaction between the dimensions from the causal relationship of the four main dimensions.

Dimension 1 (Visual perception) is a sink component, and the arrows point at themselves, not at other dimensions; in other words, it simply accepts the influence from other dimensions without affecting others. Accordingly, Dimension 1 (Visual perception) is the influencing factor affected by Dimension 2 (Emotional perception), Dimension 3 (Preference perception), and Dimension 4 (Shape perception).

Dimension 2 (Emotional perception), Dimension 3 (Preference perception), and Dimension 4 (Shape perception) are provided with two characteristics of inward-facing and outward-facing arrows belonging to the intermediate component, where the structural elements influence and link with each other.

Dimension 3 (Preference perception) is inner dependence, and the arrows point at their own loop. It means there are interdependent relationship associations within the Preference perception group. The factors X8 (like), X9 (feel satisfied), X10 (desire for continuity), and X11 (feeling of attractiveness from dimensions affect each other.

The outer dependence not only equips with the characteristics of the outward-facing arrow, without the characteristics of the inward-facing arrow, but also simply makes an influence on the other dimensions. An element in a group affects an element in another group but is not affected by itself; it can be named a source component, which is no such phenomenon among these four main dimensions. Dimension 2 (Emotional perception), Dimension 3 (Preference perception), and Dimension 4 (Shape perception) influence and connect with each other.

This study can obtain the results analyzing from four main dimensions: Dimension 3 (Preference perception) is the core item with a relatively large impact on the main dimension. Enhancing the quality of Preference perception can effectively improve the good feeling of Visual perception. As far as the influencing factors regarding the quality of visual perception and good and cognitive response in social media, Dimension 3 (Preference perception) plays a dominant role and influence.

6.2. Recommendations

Based on the contribution and significance of the study, the following are the recommendations:

6.2.1. Complement the Shortcomings of Using the Delphi Method

The main purpose of this study is to explore the key influencing factor of the quality regarding the evaluation of the response of video visual perception in social media. To explore this topic, it is important to consider the interaction of dimensions and criteria and understand the relative importance and relevance of the factors.

When providing decisions to improve or enhance the quality of response evaluations, the dimensions and factors are the crucial factors that can be identified and noted, so as the assessment of the relationship between the constructs and the criteria.

Because the previous studies using Delphi to identify the interactions between different dimensions brought about the deficiency that it could only grasp the influencing factors, excluding the interactions between the dimension factors, the DEMATEL method was used to make up for defects in the causal relationships and the degree of linking influence between the dimensions.

6.2.2. Establish Evaluation Indicators in a Professional and Quantifiable Manner by Experts and Scholars

This study is an expert-guided evaluation study. The authors assemble opinions and analyses through the perspectives of experts in social media and communication, visual design, and lighting applications. By applying DEMATEL quantitative analysis tool and summarizing expert opinions, the crucial influencing factors, indicators, and models established possess academic reference and application value.

6.2.3. Provide an Application Foundation for Teaching Design Related to Social Media and Visual Communication

This study combines academic and industry perspectives to establish influencing factors for video communication and visual perceptions in social media and evaluation models, containing different dimensions and criteria, which can be used as the application basics for related teaching design, such as the reference about the development and application of visual communication and communication practical courses.