Examining the communicative efficiency of social media in the practice of digital transformation in English language education

With the popularization of social media and network technology, the teaching environment and methods are changing. How to implement efficient English education teaching in such a context has become a focus of attention for educators and researchers.

Social media is a general term for applications and technologies based on the Internet that support the participation, sharing, and creative behaviors of registered users as their main function [1]. Registered users can gather all kinds of people in a short period of time by tweeting, retweeting and commenting, thus forming an online social group organization parallel to the real space. It carries diverse forms of linguistic materials, providing students with both linguistic knowledge in textbooks and sociolinguistic resources including online language [2-4]. Social media also has the function of assisting language learning, which not only helps teachers and students to communicate with each other, but also encourages students to integrate in-class language learning with out-of-class learning [5-6]. At present, social media has become a bridge between students’ daily communication and language learning, which has strongly broadened the field of learning [7-8].

While social media plays an important role in English education, it also faces many problems such as authenticity of information, distraction, privacy security and time management [9]. It is also crucial to cultivate students’ ability to use social media correctly and reasonably. Educators should guide students to overcome the negative effects of social media, give full play to its advantages, and work to improve students’ English learning [10-12].

Li, V. counted the data on students’ use of social media for English language learning, analyzed the relationship between it and learning activities, and found that sharing and accessing course-related materials was the single most frequent activity of students on social media [13]. Siddig, B. showed that social media can provide learners with immersive environments for interacting with better language speakers, and learners can gradually improve their mastery of the language in this free communication learning process [14]. Xodabande, I. argued that social media networks help to promote learners’ learning of linguistic features, which was verified by setting up a controlled experiment on learners’ learning of English pronunciation using the social media network Telegram as an example [15]. Namaziandost, E. et al. explored the impact of social media on students’ learning of English speaking skills, where teachers enrich the teaching and learning process by integrating social media into appropriate contexts, and students are able to improve their communicative competence through social media [16]. Barrot, J. S. reviewed a large body of scientific literature on the use of social media as a language learning environment and found that language learners pay attention to well-known social media because of their flexible communication capabilities, wide geographic distribution, and large number of active users [17]. Chawinga, W. D. included social media in the higher education curriculum to study its impact on teaching and learning activities and the findings showed that the judicious use of social media will greatly enrich learner-centered teaching and learning [18].

Based on the role of social media in the digital transformation of English education, the article constructs a research model of social media knowledge exchange efficiency of English digital education based on the knowledge exchange efficiency of digital community, using data envelopment analysis method, super-efficient SBM model, non-parametric Kernel estimation method and Tobit regression model. With that, the input-output evaluation index system for the English digital community has been established. After descriptive statistical analysis of the input and output of knowledge exchange in each section of English digital community during 10 years, the SBM model is used to calculate the efficiency of knowledge exchange in each section. Tobit regression model is utilized to explore the influencing factors of knowledge exchange efficiency in the English digital community.

2

Research model on the efficiency of social media knowledge exchange in English digital education

2.1

Efficiency of knowledge exchange

Knowledge exchange is the activity of transferring knowledge through mutual communication between knowledge producers, users, and recipients. Then the definition of knowledge exchange efficiency needs to be studied from the term “efficiency”. Efficiency first appeared in physics, used to indicate the degree of energy loss in mechanical movements. Nowadays, it is often applied to the study of economics, called economic efficiency, which is used to study the maximization of output through the minimum input. Economic efficiency can usually be divided into technical efficiency, pure technical efficiency and scale efficiency, scale efficiency is equal to the ratio of technical efficiency and pure technical efficiency, of which the technical efficiency (TE) indicates that under the condition of keeping the technical level unchanged, the effective use of existing resources to control the level of inputs or the output elements unchanged, to achieve the ideal situation of the output is maximized or inputs are minimized. Pure technical efficiency (PTE) indicates the ability to minimize investment to maximize profitability under the condition of constant scale size, reflecting the degree of contribution of technical ability and management level to efficiency.Scale efficiency (SE) is a measure of the effect of scale size on efficiency, while technical capability and management level remain unchanged.It can be seen that the efficiency of knowledge exchange affects the smooth operation of knowledge exchange. Through scholars’ research on efficiency, this paper considers that knowledge exchange efficiency refers to a certain proportional relationship formed between input and output elements when carrying out knowledge exchange activities. It belongs to the category of technical efficiency, i.e., if the knowledge exchange efficiency is higher, it means that it can maximize the realization of the business objectives to be achieved under the current scale by using the existing technical capacity and management level.

2.2

Traditional DEA models

Data Envelopment Analysis (DEA) is a non-parametric method widely used in economics and operations research to calculate the production frontier, which was proposed by famous operations researchers such as A.Charnes, W.W. Cooper, and E. Rhodes in 1978 to analyze the effectiveness of decision-making units (DMUs) by calculating the input and output indexes of evaluation objectives [19]. This method is based on a mathematical planning model, which calculates whether the decision-making units in the evaluation system are at the “production frontier”, and eventually finds the “non-DEA effective” part of the decision-making units, so as to further measure the efficiency of the evaluation indexes corresponding to these decision-making units. The efficiency of the evaluation indexes corresponding to these decision units can be further evaluated. In addition, with the help of the computational model in the DEA model, the values of the relaxation variables corresponding to the inputs and outputs can be measured, and the target efficiency can be derived and the specific optimization strategy can be proposed at the same time.

2.3

SBM model

The Farrell efficiency measurement idea is the basis for traditional DEA models (BCC and CCR models), and the ability to rank is relatively weak, which may overestimate the efficiency of decision-making units. Therefore, this paper utilizes the evaluation method based on slack variables proposed by Tone, which gives full consideration to the slack of outputs and inputs, and compared with the traditional CCR model, the optimization of the SBM model is more inclined to the consideration of maximizing the efficiency of the knowledge exchange, which improves the credibility and accuracy of the efficiency evaluation [20]. Through the above summary, the SBM model is now constructed as follows:

The SBM model considers the objective function of optimizing the virtual English digital social media community as the slack variable, and the fractional planning form is shown in equation (1): (1) ${\begin{array}{l} \min δ = \frac{1 - \frac{1}{u} \sum_{u = 1}^{n} \frac{s_{u}^{-}}{x_{u 0}}}{1 + \frac{1}{v} \sum_{v = 1}^{m} \frac{s_{v}^{+}}{y_{v 0}}} \\ s . t . x_{0} = X ϕ + s^{-} \\ y_{0} = Y ϕ - s^{+} \\ ϕ \geq 0, s^{-} \geq 0, s^{+} \geq 0 \end{array}$

Where δ is the evaluation value of the efficiency of English digital social media community knowledge exchange, u is the number of input elements of English digital social media community knowledge exchange, and v is the number of output elements of English digital social media community knowledge exchange. y₀ and x₀ are the output and input vectors of the decision unit, respectively. Y and X are the output and input matrices of the decision unit, respectively. $s_{u}^{-}$ is the element of slack input s⁻, which indicates the redundancy of inputs in knowledge exchange, and $s_{v}^{+}$ is the element of slack output s⁺, which indicates the deficiency of outputs in knowledge exchange. Xϕ denotes the amount of knowledge inputs on the frontier, and Yϕ denotes the amount of knowledge outputs on the frontier. $\frac{1}{u} \sum_{u = 1}^{n} \frac{s_{u}^{-}}{x_{u 0}}$ denotes the average of the ratio of redundancy in u knowledge exchange inputs to their respective actual inputs, i.e., the average level of u inefficiency in knowledge exchange inputs, and $\frac{1}{v} \sum_{v = 1}^{m} \frac{s_{v}^{+}}{y_{v 0}}$ denotes the average of the ratio of deficiencies in v knowledge exchange outputs to their respective actual outputs, i.e., the average level of v inefficiency in knowledge outputs.

Thus the efficiency value of knowledge exchange in the English digital social media community is the product of the average efficiency level of each knowledge input and the average efficiency level of each output, and the efficiency levels of both knowledge inputs and outputs have an impact on the SBM model. Using the Charnes-Cooper transformation method, its fractional planning can be transformed into a linear planning problem, as shown in equation (2): (2) ${\begin{array}{l} \min ε = t - \frac{1}{n} \sum_{u = 1}^{n} \frac{s_{u}^{-}}{x_{u 0}} \\ s . t .1 = t + \frac{1}{m} \sum_{v = 1}^{m} \frac{s_{v}^{+}}{y_{v 0}} \\ t x_{0} = X θ + s^{-} \\ t y_{0} = Y θ + s^{+} \\ θ \geq 0, s^{-} \geq 0, s^{+} \geq 0, t > 0 \end{array}$

In equation (2), ε is the efficiency measure, s⁻=ts⁻,s⁺=ts⁺,θ=tϕ. The SBM model indicates that when the slack output and slack input are smaller, i.e., eliminating the excess of knowledge exchange inputs and the insufficiency of outputs, the higher the value of the communication efficiency of the English digital social media community. If the knowledge exchange efficiency of an English digital social media community is effective according to the evaluation criteria of the SBM model, i.e., when s⁻=s⁺=0,ε=1, the result is equivalent to the efficiency value of the traditional DEA model δ=1. In this paper, the comprehensive efficiency of knowledge exchange is decomposed into the pure technical efficiency and the scale efficiency, and the pure technical efficiency in the knowledge exchange of the English digital social media community refers to the enhancement of the knowledge exchange activities through the internal management technology of the community, and the scale efficiency means that knowledge exchange through new user registration and member posting to promote member viewing and replying.

2.4

Super-SBM model

The SBM model solves the influence generated by the slack value under the traditional DEA model, but for the case where there are multiple effective values in DMUs with relative efficiency value of 1, the effective units are not comparable and quadratic ranking, in this regard, Tone optimizes the SBM model and proposes a super-efficient SBM model with corrected slack variables, i.e., a quadratic evaluation is carried out on the basis of the original model, and the effective DMUs are remeasurement [21]. When using the super-efficient SBM model for measurement, under the premise of ensuring that the production frontier of ineffective DMUs remains unchanged, and taking the input and output indicators as constraints, the effective frontier of DMUs will move to the right, and the final efficiency value obtained will appear to be greater than or equal to 1, i.e., to realize the secondary ranking of DMUs whose efficiency values are all 1. The super-efficient SBM model’s form can be seen in equation (3): (3) ${\begin{matrix} \min ρ = k - \frac{1}{m} \sum_{i = 1}^{m} \frac{S_{i}^{-}}{x_{i 0}} \\ 1 = k + \frac{1}{S_{1} + S_{2}} (\sum_{i = 1}^{s_{1}} \frac{S_{i g}^{+}}{y_{i 0}^{g}} + \sum_{i = 1}^{s_{2}} \frac{S_{i b}^{-}}{y_{i 0}^{b}}) \\ k x_{i 0} = \sum_{j = 1}^{n}, j \neq k^{x_{i j} λ_{j} + S_{i}^{-}} \\ k y_{i 0}^{g} = \sum_{j = 1}^{n}, j \neq k^{y_{i j}^{g} λ_{j} - S_{i g}^{-}} \\ k y_{i 0}^{b} = \sum_{j = 1}^{n}, j \neq k^{v_{i j}^{b} λ_{j} + s_{i b}^{-}} \\ λ_{j} \geq 0, S_{j}^{-} \geq 0, S_{i g}^{+} \geq 0, S_{i b}^{-} \geq 0 \end{matrix}$

where ρ represents the objective function. m, S₁ and S₂ represent the specific size of the input, desired output and non-desired output elements, respectively. $x_{i j}, y_{i j}^{g}$ and $y_{i j}^{b}$ represent the i th input, g th desired output and b th non-desired output of DMU_j, respectively. λ represents the weight vector. In the specific measurement process, if the efficiency value of DMU is 1.5, this DMU indicates that the technical Q&A community remains relatively efficient even if the knowledge exchange inputs are increased by 50% in equal proportion, the DMU still remains relatively effective, i.e., the efficiency value can be greater than 1. However, if ρ is less than 1, the measured DMU is ineffective, and adjustments need to be made to the inputs and outputs.

2.5

Nonparametric Kernel Estimation

Kernel density estimation is used in probability theory to estimate an unknown density function, which belongs to one of the nonparametric test methods [22]. Assuming that the data obeys an unknown distribution, now for a randomly sampled sample set x₁,x₂,⋯x_n of this data, the probability density of the unknown distribution at data point X can be estimated as equation (4): (4) $f (x) = \frac{1}{n h} \sum_{i = 1}^{n} K (\frac{x - x_{i}}{h})$

In equation (4), K(x) is called the kernel function, commonly used kernel function has the form of Gaussian, cosine, uniform and triangular, etc., where h represents the bandwidth, the selected bandwidth is crucial to the acquisition of optimal fitting results, generally minimize the mean-square error as the principle of selection, this paper selects the automatic bandwidth, the kernel function selects the best under the meaning of mean-square error, and the loss of efficiency is also small, the Epanechnikov kernel function is very effective for the virtual academic community for dynamic knowledge exchange efficiency evolution.

2.6

Tobit model

The dependent variable of the Tobit model can only be observed partially due to some restrictions, so it is called a restricted dependent variable model or censored regression model [23]. In this paper, the value of knowledge exchange efficiency calculated by using the SBM model is between 0 and 1, which is equivalent to the value of the dependent variable is limited to the range of 0 to 1. If the least squares method is used for regression, the problem of parameter estimation bias may occur, while the Tobit model is suitable for the case of restricted dependent variable, and the maximum likelihood method is used to estimate the parameters, which can better avoid the problem of biased parameter estimation. Therefore, in this paper, Tobit regression model is chosen to calculate the environmental factors affecting the efficiency of knowledge exchange in English digital social media communities.The basic mathematical expression of Tobit model is shown in equation (5): (5) $\begin{matrix} y_{i}^{*} = β x_{i} + α + ε_{i}, ε_{i} \sim N (0, σ^{2}) \\ y_{i} = {\begin{array}{l} y_{i}^{*}, & 0 \leq y_{i}^{*} \leq 1 \\ 0, & y_{i}^{*} < 0 \\ 1, & y_{i}^{*} > 1 \end{array} \end{matrix}$

In equation (5), $y_{i}^{*}$ is the latent variable, y_i is the actual observed dependent variable, x_i is the independent variable, α is the constant term, and ε_i is the error term, independent and obeying a normal distribution: ε_i~N(0,σ²).

2.7

TF-IDF algorithm

TF-IDF is a statistical method based on word frequency proposed by Salton et al. to assess the importance of a word for a document set [24]. The higher the TF-IDF value of a word, the more important the word is to this document set, and the importance of a word increases proportionally to the number of times it appears in a document, but at the same time decreases inversely to the number of times it appears in the entire document set. The calculation method is shown in equation (6): (6) $T F I D F_{i j} = T F_{i j} * I D F_{i}$

In Eq. (6), TF is the word frequency, i.e., how often a word appears in a document. IDF, i.e., Inverse Document Frequency, measures the prevalence of the word for the whole set of documents.If a word is more common in the whole set of documents, the larger the denominator will be, and the smaller the value of IDF will be. The expressions are shown in equation (7) and equation (8): (7) $T F_{i j} = \frac{n_{i j}}{\sum_{k}^{N} n_{i k}}$ (8) $I D F_{i} = \log \frac{N}{D F_{i}}$

In Equation (7), TF_ij (word frequency) is the frequency of occurrence of the word j in the ind document, and the denominator is the normalization of TF_ij in order to avoid the excessive frequency of a word. In Equation (8), N denotes the number of all documents in the document set, and DF_i denotes the number of all documents in the document set in which the word i has appeared, which is called the characteristic document frequency. By calculating the TF-IDF values of all the words in the whole document set, the words with higher importance can be selected as the keywords of the document.

3

Analysis of factors affecting the efficiency of knowledge exchange in the English-speaking digital community

3.1

Analysis of Knowledge Exchange Efficiency in English Digital Communities

This paper combed the existing literature, used the TF-IDF algorithm to analyze the word frequency statistics of the English digital community, combined with the characteristics of the English digital community as well as the principles of constructing the indicator system, and established the input-output evaluation indicator system from the three dimensions of the breadth, depth and quality of knowledge exchange as shown in Table 1.

Table 1.

Evaluation system of knowledge exchange efficiency of English-learning community

Primary indicator	Secondary indicator	Tertiary indicator
Output indicator	Knowledge exchange span	X1 (Browsing quantity)
	Knowledge exchange depth	X2 (Response quantity)
	Knowledge exchange depth	X3 (Again-response quantity)
	Knowledge exchange quality	X4 (Thumbing-up quantity)
	Knowledge exchange quality	X5 (Collection quantity)
Input indicator	Knowledge exchange span	Y1 (User quantity)
	Knowledge exchange depth	Y2 (Post quantity)
	Knowledge exchange quality	Y3 (Knowledge viscosity)
	Knowledge exchange quality	Y4 (Knowledge density)

Based on the evaluation indexes of knowledge exchange efficiency and selected influencing factors, this section analyzes the knowledge exchange efficiency of eight selected sections of the English digital community (Community A) from 2014 to 2023.

3.1.1

Descriptive statistical analysis

In this study, after collecting the knowledge exchange related data of the selected eight sections from 2014 to 2023, the raw data were processed using Python language to get the base data, and the descriptive statistics of the input and output data of knowledge exchange of each section were carried out as shown in Table 2.

Table 2.

Descriptive statistics of 10 sections in English-learning community

Section	Statistics	Input indicator				Output indicator
Section	Statistics	Y1	Y2	Y3	Y4	X1	X2	X3	X4	X5
English linguistics	Min	206	78	23	18	120653	203	56	72	123
	Max	1104	405	76	67	681644	2848	818	719	1026
	Mean	563	283	39	33	264851	1216	376	284	486
	SD	289	88	11	15	184065	902	264	177	327
English literature	Min	1562	825	152	124	859346	1166	655	377	945
	Max	18725	4128	922	1265	3594656	27454	5615	8246	17546
	Mean	6246	2056	488	422	2264567	156428	3485	3617	4254
	SD	4295	875	238	301	894528	9034	2044	2749	4736
English translation	Min	2846	1124	249	327	152643	5016	743	923	2042
	Max	6443	2943	528	644	3014625	20468	6425	3847	8946
	Mean	3172	1743	312	506	2316582	9764	2651	2015	4759
	SD	1062	348	86	98	678316	5268	1844	907	1846
Business English	Min	1326	2046	326	726	2594524	14215	5482	2316	8756
	Max	6485	4255	452	1254	6123545	42645	22164	11065	31053
	Mean	4264	3217	368	1029	3915465	28196	8465	5269	14872
	SD	1468	516	59	219	976545	11267	7168	2913	8864
English pedagogy	Min	2524	1246	356	415	170262	5263	784	978	2133
	Max	7216	3048	684	735	3011203	23154	6924	4022	8842
	Mean	3306	1936	426	628	2265890	10246	2815	1967	4628
	SD	1244	442	112	144	681685	5746	2043	946	1922
English film & media	Min	15264	1526	425	724	4389136	22106	4263	2345	16425
	Max	27465	6652	2168	4015	17622468	88462	28467	23485	89426
	Mean	18474	4065	948	1869	10581660	56417	12416	12034	47163
	SD	7486	1588	463	972	3826491	25164	84625	7246	30247
International relations	Min	2120	1355	342	510	1426355	4012	526	342	2206
	Max	3275	2451	485	986	3254856	17856	2435	2108	14260
	Mean	3268	1845	346	745	2568955	8011	986	989	8925
	SD	859	339	72	148	621352	4352	713	789	7042
Intercultural communication	Min	2012	1106	220	426	821352	1754	402	284	1412
	Max	5126	4152	725	714	1975162	8649	968	1854	6018
	Mean	2849	2045	325	533	1220356	4252	645	721	3125
	SD	756	806	150	110	442685	2253	213	458	1546

In terms of input indicators, the section with the greatest input in terms of the number of users, the number of posts, knowledge stickiness and knowledge density is English Film and Media, and Business English and English Language and Literature are ranked at the top of the list. There are also differences in the annual input of different sections. The section with the most input in a year is English Film and Media with a total of 27,465 people, and the smallest section is English Linguistics with only 206 people; the section with the most postings in a year is English Film and Media with a total of 6,652 postings, and the least number of postings is only 78 postings, and the sections with the greatest input in knowledge stickiness and knowledge density are English Film and Media, and the smallest ones are also English Linguistics.

In terms of output indicators, the number of views reflects the breadth of knowledge dissemination, the number of replies and re-replies reflects the depth of knowledge exchange, while the number of likes and favorites reflect the degree of users’ recognition and usefulness of the post content from the quality dimension of knowledge exchange output. The top-ranked sections in knowledge exchange output are Business English, English Film and Media, and International Relations, but they differ in different aspects. In terms of the breadth of knowledge exchange, English Film and Media is the broadest, followed by Business English. From the depth of knowledge exchange, the similarity between the output rankings between the sections and the breadth of knowledge dissemination indicates that the breadth of knowledge dissemination influences the depth of knowledge exchange, and the wider the knowledge dissemination, the more likely it is to cause more users to participate in the discussion and obtain more in-depth knowledge exchange. From the dimension of the quality of knowledge exchange output, the top three sections in terms of the number of likes and favorites are English Film and Media, Business English, and English Translation, indicating that the quality of knowledge exchange output varies among sections according to their own characteristics.

3.1.2

Input elaboration indicator correlation tests

When the SBM model is used to calculate knowledge exchange efficiency, the condition of homoscedasticity needs to be satisfied between each input and output indicator. In order to test whether the selected indicators are reasonable, this paper uses SPSS software to carry out Pearson two-sided correlation test on the collected data of input and output indicators, and the variable test results are shown in Table 3. As can be seen from Table 3, the correlation coefficients between the input and output indicators selected in this paper are all positive, and they all pass the two-sided test at the 1% level, and there is a significant positive correlation between the input and output indicators, which is in line with the requirements of the SBM model.

Table 3.

The correlation test of input and output indicators

Input indicators	Output indicators
Input indicators	X1	X2	X3	X4	X5
Y1	0.865^***	0.842^***	0.563^***	0.765^***	0.682^***
Y2	0.592^***	0.552^***	0.394^***	0.378^***	0.376^***
Y3	0.696^***	0.735^***	0.594^***	0.522^***	0.482^***
Y4	0.612^***	0.610^***	0.511^***	0.371^***	0.372^***

3.1.3

Calculation of the efficiency of knowledge exchange

In this paper, we use DEA-Solver Pro5.0 software to measure the knowledge exchange efficiency of the English digital community, and we choose the input-oriented SBM model to measure the results of knowledge exchange efficiency. According to the calculation results of knowledge exchange efficiency of 8 sections in Community A from 2014 to 2023, the mean value changes of comprehensive technical efficiency, pure technical efficiency and scale efficiency of knowledge exchange of each section from 2014 to 2023 are counted as shown in Figure 1.

As can be seen in Figure 1, the mean value of the comprehensive technical efficiency of the eight sections of Community A is 0.714, with a relatively smooth change over the 10-year period, and the same direction of the comprehensive technical efficiency from 2014-2018 due to the simultaneous rise, fall and rise in pure technical efficiency and scale efficiency, and after reaching the highest efficiency value of 0.786 in 2018, the scale efficiency is still showing a smooth upward trend, while subject to the pure technical efficiency decline, the comprehensive technical efficiency continues to slowly decrease and reaches the lowest value of 0.662 in 2023.

The pure technical efficiency during the 10 years is lower than the scale efficiency, indicating that the factors restricting the improvement of the comprehensive technical efficiency of knowledge exchange in English digital communities are mainly the community system and internal management level, technology, etc. After a small increase in pure technical efficiency from 2014 to 2015, there was a decline in 2016, and in the two years after that, due to the improvement of the system and the internal management level of Community A in the process of development, the pure technical efficiency continued to rise, reaching a 10-year high of 0.831 in 2018, but pure technical efficiency continued to decline slowly after 2018, reaching a low of 0.682 in 2022, with a small increase in 2023, indicating that the internal management of Community A continued to have problems after 2018, and that the input elements of users’ knowledge exchanges could not be effectively utilized, and that the community managers had to pay attention to the technical aspects such as institutional The community managers must pay attention to the improvement of the technical level such as the construction and internal management level.

The overall level of scale efficiency in the eight sections of Community A is high, with the average value reaching 0.908 over the 10-year period, and higher than the pure technical efficiency every year, and the overall trend showing a smooth fluctuation, rising slowly in recent years and reaching a maximum value of 0.963 in 2022, which indicates that the community’s scale structure is more reasonable, and the scale efficiency has been fully utilized.

3.2

Analysis of Factors Influencing the Efficiency of Knowledge Exchange in English Digital Communities

3.2.1

Impact factor assumptions

The efficiency of knowledge exchange in English digital communities is affected by multiple external factors, such as changes in the economy, society, and policies, as well as by the users’ own attributes.In this paper, taking into account the accessibility of the data and the characteristics of the virtual digital community itself, the main internal factors affecting the efficiency of knowledge exchange in the English digital community are summarized into four factors: the size of the community, the time of the community’s establishment, the level of the community’s management, and the quality of community members. The definitions and symbols of the influencing variables are shown in Table 4.

Table 4.

The influence factors and symbols of knowledge exchange of English-learning community

Object	Interpretation variable	Variable symbol
Influence factors of English-learning community knowledge exchange efficiency	Community scale	Scale
	Community founding time	Time
	Community management participation	Manage
	Community capital investment	Capital
	Community member quality	Quality

This paper uses cross-sectional data to carry out regression analysis of influencing factors, according to the relevant data collected in the previous section of Community A from 2014-2023, the technical efficiency value of community knowledge exchange measured in the first stage is used as the independent variable, and the technical efficiency values are all truncated data greater than 0, which meets the basic requirements of the Tobit model, therefore, this paper can adopt the Tobit model for the eight sections of the English digital community of the cross-sectional data for regression analysis.

Descriptive statistics of the independent variables and dependent variables of the eight sections of the professional discipline area of Community A were conducted, and the results are shown in Table 5. The results in Table 5 show that there are 8 samples in the descriptive statistics analysis, and the mean value of the technical efficiency of knowledge exchange of the sample data is 0.714, the maximum value is 0.786, the minimum value is 0.662, and the standard deviation is 0.185, which indicates that there are some differences in the technical efficiency of knowledge exchange in the 8 sections of the professional discipline area of Community A. The efficiency is affected by several factors, resulting in different inputs and outputs of knowledge exchange activities. The efficiency is affected by several factors, resulting in different inputs and outputs of knowledge exchange activities.

Table 5.

Descriptive statistics of independent variables and dependent variables

Statistics	N	Minimum	Maximum	Mean	SD
Knowledge exchange technical efficiency	8	0.662	0.786	0.714	0.185
Community scale	8	0.452	0.926	0.665	0.142
Community founding time	8	0.108	1.012	0.754	0.338
Community management participation	8	0.051	0.242	0.098	0.074
Community capital investment	8	0.173	0.603	0.324	0.156
Community member quality	8	0.008	0.095	0.063	0.046

From the indicator of community size, the mean value of the ratio of the number of users to the number of posts of the sample data is 0.665, the maximum value is 0.926, and the minimum value is 0.452, which indicates that there is a certain gap in the strength of the investment in knowledge exchange and underinvestment in Community A from 2014 to 2023. From the perspective of the community establishment time indicator, the average value of the distance between the establishment time of each section of the sample data to 2023 is 0.754, the maximum value is 1.012, and the minimum value is 0.108, indicating that there is a large difference in the establishment time of the sections in Community A. In terms of community management level indicators, the mean value of managerial participation in each section is 0.098, the maximum value is 0.242, and the minimum value is 0.051, which indicates that overall managerial participation is relatively low, and the role of managers in community forums is not obvious. The mean value of the number of gold coins used for knowledge exchange per post in the forum is 0.324, the maximum value is 0.603, and the minimum value is 0.173. From the perspective of the indicators of community member quality, the average value of the ratio of high-quality EPI members to total users in the sample data is 0.063, the maximum value is 0.095, and the minimum value is 0.008, which can be seen that there are relatively few high-quality members in the community, and they do not form a center of strength, making it more difficult to lead the other members in effective knowledge exchange.

3.2.2

Analysis of Tobit model results

According to the relevant data collected in the previous paper, the Tobit regression analysis of the factors influencing the knowledge exchange efficiency of the English digital community was carried out using STATA 14.0 software, and the results are shown in Table 6.

Table 6.

Tobit result analysis

Variable	Coef.	Std. Err	T	P>\|t\|	[95% Conf. Interval]
Scale	0.1255642	0.1012253	1.26	0.294	-0.1348562	0.3916525
Time	-0.3816524^*	0.0925436	-4.58	0.006	-0.5815645	-0.1954258
Manage	0.2348526^*	0.0685624	4.22	0.008	0.0948252	0.3786185
Capital	0.1426528^***	0.07284	2.34	0.092	-0.0254856	0.3012526
Quality	0.162432^**	0.0624856	5.74	0.044	0.0094821	0.29482
Cons	0.2153458^**	0.0905485	2.85	0.055	0.0154827	0.425846

The results in Table 6 show that the p-value of the null hypothesis test is 0.044 < 0.05, indicating that the constructed Tobit model is valid, and the selected influencing factor variables can explain the value of knowledge exchange efficiency of the English digital community as a dependent variable better, while the p-value of the influencing factor of the size of the English digital community is 0.294, which fails to pass the test of the model’s level of significance, i.e., the knowledge exchange efficiency of English digital community exchange efficiency has no significant effect, the two influencing factors of English digital community establishment time and English digital community managers’ participation are significant at 1% confidence level, the quality of English digital community members is significant at 5% confidence level, and the financial investment in English digital community knowledge exchange management is significant at 10% confidence level.

The calculations show that, first, the size of the English digital community is positively correlated with knowledge exchange efficiency, with a correlation coefficient of 0.126, which does not pass the statistical significance test.

Second, the time factor of English digital community establishment is negatively correlated with knowledge exchange efficiency, with a correlation coefficient of -0.382. It shows that with the accumulation of time, the knowledge exchange efficiency of English digital community is decreasing, which is related to the lack of pure technical efficiency in recent years.

Thirdly, English digital community managerial involvement is positively correlated with knowledge exchange efficiency, with a correlation coefficient of 0.235. The English digital community management team plays a leading role in knowledge exchange activities, which can help to support the maintenance of high-quality knowledge exchange by users in the virtual academic community, otherwise, it will affect the activities of users’ knowledge exchange.

Fourth, English digital community knowledge exchange management capital investment is positively correlated with knowledge exchange efficiency, with a correlation coefficient of 0.143. Every unit of knowledge exchange management capital investment can be increased to improve the knowledge exchange efficiency of the virtual academic community by 0.143. Rewards and punishments should be clearly defined within the virtual academic community, with rewards as the main focus and punishments as a supplement, integrating the characteristics of different communities and maintaining the benign operation of the community itself and its own attractiveness. Recognizing and rewarding users’ knowledge exchange behaviors at both the material and spiritual levels will be conducive to the transformation of knowledge.

Fifth, the quality of English digital community members is positively correlated with knowledge exchange efficiency, with a correlation coefficient of 0.162, indicating that an increase in the proportion of high-quality members of English digital community will have a significant effect on the improvement of the knowledge exchange efficiency of English digital community, and the English digital community should make efforts to improve the professional skills and academic knowledge of its members and appropriately increase the proportion of high-quality members.

4

Conclusion

Based on the needs of digital transformation in English education, the article deeply studies the influencing factors that affect the efficiency of knowledge exchange on social media platforms. The super-efficient SBM model, nonparametric Kernel estimation and Tobit regression model are used to construct a research model of social media knowledge exchange efficiency in English digital education, and it is utilized to conduct an example study.

In the English digital community, English Film and Media has the greatest input, coming out on top in terms of the number of users, the number of posts, the knowledge stickiness and the knowledge density, and the top rankings are still the sections of Business English and English Language and Literature. English Film and Media has the highest staff input (27,465) and number of posts (6,652) among the eight sections. In terms of output, the top ones are Business English, English Film and Media, and International Relations. English Film and Media has the highest number of views, replies, re-replies, likes, and favorites.

The mean values of comprehensive technical efficiency, pure technical efficiency, and scale efficiency of the eight sections of English digital community are 0.714, 0.771, and 0.908, respectively.The correlation coefficients of community size, community establishment time, community manager participation, community knowledge exchange management capital investment, and quality of community members with knowledge exchange efficiency are 0.126, -0.382, 0.235, respectively, 0.143, and 0.162.

Język:: Angielski

Częstotliwość wydawania:: 1 razy w roku
Dziedziny czasopisma:: Nauki biologiczne, Nauki biologiczne, inne, Matematyka, Matematyka stosowana, Matematyka ogólna, Fizyka, Fizyka, inne

Kanał RSS czasopisma

Examining the communicative efficiency of social media in the practice of digital transformation in English language education

Xiao Yang

Data publikacji: 17 mar 2025

Otrzymano: 26 paź 2024

Przyjęty: 02 lut 2025

DOI: https://doi.org/10.2478/amns-2025-0154

Słowa kluczoweKnowledge exchange efficiency, SBM, nonparametric Kernel estimation method, Tobit regression modeling

© 2025 Xiao Yang, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Słowa kluczowe
Knowledge exchange efficiency, SBM, nonparametric Kernel estimation method, Tobit regression modeling