Research on the Reform of Material Specialized Practical Training Course Based on Informatization Teaching Platform

The term “informatization” first appeared in the field of economics and industry, and was subsequently extended to the field of education. In the 1960s, the concept of “informatization” was put forward in the Japanese academic literature, and in the 1970s, Germany, the European Community and UNESCO began to realize the important role of information technology in society, and began to study and formulate relevant plans. After the 1990s, a wave of informatization construction emerged, and the United States included informatization in its national strategy and further proposed the wide application of information technology in all areas of society [1-5]. China has already proposed to promote education informatization in 2001, and the Ministry of Education has issued relevant documents, clearly stating that it should vigorously promote the extensive use of information technology in education and teaching, as well as promote the effective integration of information technology and subject curricula. In 2010, the Ministry of Education also issued the Ten-Year Development Plan for Education Informatization (2011-2020), which proposes to build a public service platform for continuing education and to improve the lifelong education system [6-8]. Subsequently, in 2015, the Ministry of Education further put forward the development goals during the "Twelfth Five-Year Plan" period, to build "three links and two platforms", including "broadband network school-to-school communication, high-quality resource class communication, online learning space for everyone", and the construction of educational resources public service platform and education management public service platform, into the rapid development of information education service platform. In 2018, the Ministry of Education formulated the Action Plan for Education Informatization 2.0, proposing to achieve the development goal of "three full, two high schools and one major" by 2022, that is, teaching applications cover all teachers, learning applications cover all school-age students, and digital campus construction covers all schools, and the level of information application and information literacy of teachers and students are generally improved [9-13]. We will build a large platform of "Internet + education", promote the transformation from special resources for education to large resources for education, from improving the ability of teachers and students to apply information technology to comprehensively improve their information literacy, and from integrated application to innovative development, and strive to build a new model of talent training, education services, and education governance under the conditions of "Internet +" [14-17]. Therefore, in recent years, with the support of national policies, China’s informatization teaching platforms have flourished. In addition to the “Smart Education Platform for Primary and Secondary Schools” created by the Ministry of Education, relevant educational enterprises and schools have also begun to actively explore new modes of teaching on informatization platforms, and have made a lot of progress in promoting the popularization of educational resources and promoting the in-depth integration of information technology and education, etc. [18-21]. Materials professional practical training is a very important practical teaching link in the teaching process of undergraduate materials majors, mainly to train and strengthen students’ professional skills, and cultivate students’ ability to use theoretical knowledge to comprehensively analyze and solve practical problems. The reform of materials professional practical training courses on the basis of information technology teaching platform can realize the accurate and effective supply of professional skilled talents, establish a two-way balance between the school and the enterprise “supply side” of the professional talent cultivation goals, which is crucial for the development of materials majors in colleges and universities [22-25].

This study sorted out the related contents of technology acceptance theory in the field of education research, proposed the main factors affecting the willingness to use live teaching, then proposed to construct a theoretical model of influencing factors on the acceptance of live teaching of college students based on the UTAUT model according to the relationship between the factors, defined the variables in the model, clarified the measurement dimensions of the study, and put forward the related research hypotheses. Subsequently, the research methods of Levene and Pearson correlation coefficients are described, and the method of PLS-SEM modeling analysis is introduced. Finally, the dimensions and question items of the questionnaire on factors influencing students’ acceptance of live teaching were designed, and the questionnaire was distributed and collected. SPSS and AMOS software were utilized to analyze questionnaire data and carry out model and hypothesis testing on the theoretical model.

2

Modeling and research hypotheses

2.1

Technology Acceptance Model for the Reform of Practical Training Courses in the Materials Program

2.1.1

Technology Acceptance and Use Integration Theory (UTAUT)

The UTAUT model is shown in Figure 1. The four core determinants of IT acceptance in the UTAUT model are: performance expectations, effort expectations, social influence, and facilitation. The “performance expectation” factor is a combination of perceived usefulness, external motivation, task matching, and relative advantage. The “effort expectations” factor is a combination of perceived ease of use, complexity and ease of use. The “Social Influence” factor is a combination of two factors: subjective norms and social factors: the “Facilitating Conditions” factor is a combination of perceived behavioral control, facilitating conditions, and compatibility, etc. The UTAUT model has been widely used to study the acceptance of specific information technologies. The UTAUT model has been widely used to study the acceptance of specific information technologies, such as students’ acceptance of MOOC platforms and their influencing factors, acceptance of mobile learning and their influencing factors, and acceptance of online learning spaces and related teaching software. This indicates that the UTAUT model is effective for the vast majority of studies, and can be flexibly applied according to the content and object of one’s own research, modifying and expanding the core variables [26-27].

2.1.2

Initial model determination presentation

1)

Summary of research hypotheses

A total of 12 research hypotheses are proposed in this study, in which “→→” represents a significant positive effect, and the initial model research hypotheses are shown in Table 1.

Table 1.

Initial model research hypothesis

Code	Research hypothesis	Hypothetical description
H1	PE→→UI	Performance expectations have a significant positive effect on the use of live teaching
H2	EE→→UI	We hope to have a significant positive effect on the use of live teaching in college students
H3	SI→→UI	The social impact has a significant positive effect on the use of live teaching in college students
H4	FC→→UI	The effect of the conditions on the use of live teaching in college students has a significant positive effect
H5	SE→→UI	The self-efficacy is a significant positive effect on the use of live teaching in college students
H6	PP→→UI	The perceived pleasure has a significant positive effect on the use of live teaching
H7	EE→→PE	We hope to have a significant positive effect on the performance expectations of the students’ live teaching in colleges and universities
H8	SI→→PE	The effect of social influence on the performance of students in college students is significantly positive
H9	FC→→EE	The efforts to promote the teaching of college students’ live teaching are expected to have a significant positive effect
H10	SE→→PP	The sense of self-efficacy has a significant positive effect on the perceived pleasure of the teaching of the students
H11	Variable hypothesis	Different gender is different in the model of the students’ live teaching acceptance model
H12		Different grades differ in the different paths of the students’ live teaching acceptance model

2)

Definition of Research Variables and Measurement Dimensions

The model of students’ acceptance of live teaching includes six independent variables: performance expectation (PE), effort expectation (EE), social influence (SI), facilitating condition (FC), self-efficacy (SE), and perceived pleasure (PP), of which EE, PE, SI, and FC come from the original UTAUT model, and SE and PP are new independent variables added after the research on the characteristics of live teaching. The following are the definitions of the six independent variables mentioned above and the preparation of the measurement question items.

(1)

Performance Expectation

Based on the UTAUT model and related acceptance studies, four questions were modified to measure the performance expectations of live teaching and learning. The design of the performance expectation measurement question items is shown in Table 2. Performance expectation (PE) refers to the extent to which users expect innovative technology to improve their learning outcomes or job performance.Students’ willingness to use live teaching will be stronger if they believe that live teaching will help them achieve their learning outcomes.

Table 2.

Performance expectations design

Independent variable	Item number	Subject content
Performance expectation (PE)	PE1	Broadcast teaching makes teaching resources better
	PE2	Learning through live teaching can take advantage of my scattered learning time
	PE3	Live teaching can greatly arouse my interest in learning
	PE4	Learning through live teaching can improve my learning efficiency

(2)

Effort Expectations

The design of the effort expectation measurement question items is shown in Table 3. Based on the combination of the UTAUT model and related acceptance studies, three question options were modified and compiled to measure the expectation of effort for live teaching and learning. Effort expectation (EE) refers to the amount of effort that college students believe they need to expend to become fluent in using live instruction. If college students believe that learning with live teaching is easy and simple, and that they can handle the difficulties that arise during the process smoothly, their willingness to use it will increase.

Table 3.

Try to expect the design of the measurement

Independent variable	Item number	Subject content
Effort expectation (EE)	EE1	It’s easy for me to learn to use live broadcasts skillfully
	EE2	Communicating with classmates or teachers in live teaching is clear and easy to understand
	EE3	I can adapt to live teaching quickly

(3)

Social Impact

The design of the social impact measurement question items is shown in Table 4. Based on the combination of the UTAUT model and related acceptance studies, three question options were modified and compiled to measure the social impact of live teaching and learning. Social influence (SI) refers to the fact that the use of live teaching by college students is influenced by the thoughts, attitudes, and behaviors of specific people or groups, thus affecting their willingness to use live teaching. Students’ willingness to use live teaching is generally enhanced if the atmosphere around them, including classmates, teachers, and school leaders, motivates them to use live teaching.

Table 4.

Social impact measurement design

Independent variable	Item number	Subject content
Social impact(SI)	SI1	If students are involved in live learning, I would like to try to use it
	SI2	People who are important to me recommend live teaching to study and I will be happy to use it
	SI3	In general, the encouragement of schools encourages me to use live learning

(4)

Facilitating conditions

The design of the enabling conditions measurement question items is shown in Table 5. Based on the combination of the UTAUT model and related acceptance studies, five question options were modified and compiled to measure the facilitating conditions for learning during live teaching.Facilitating conditions (FC) refers to the sum of all the facilities and technological conditions required for students in higher education to learn through live teaching.Students will have a stronger willingness to use live teaching if they have or obtain relevant resources, such as material security, that help them use live teaching for learning.

Table 5.

To facilitate the design of conditional measurement

Independent variable	Item number	Subject content
Facilitation condition(FC)	FC1	I have the information equipment (computer, tablet, mobile phone, etc.) that are involved in live teaching
	FC2	I have a network environment that is involved in live teaching
	FC3	The operating system of the broadcast teaching platform (registration, login, viewing) is easy and convenient
	FC4	The video image of the video that I participated in is good
	FC5	If there are software problems such as flash-back in the teaching process, I will be bothered

(5)

Self-efficacy

The self-efficacy measurement question item design is shown in Table 6. Based on the UTAUT model and related acceptance studies, three question options were modified and compiled to measure self-efficacy for learning with live teaching. Self-efficacy (SE) refers to the degree to which college students are confident that they can successfully complete learning tasks and keep up with learning progress when using live teaching for learning. If students have high self-efficacy, they are more likely to accept live teaching and are more willing to try to learn using live teaching than students with low self-efficacy.

Table 6.

Self-efficacy measurement design

Independent variable	Item number	Subject content
self-efficacy (SE)	SE1	Difficulties can be achieved in time in the course of live learning
	SE2	I am confident that I can effectively solve the difficulties encountered in the course of live learning
	SE3	I will find a solution in time

(6)

Perceived Pleasure

The perceptual pleasantness measurement question item design is shown in Table 7. Based on the combination of the UTAUT model and related acceptance studies, three question options were modified and compiled to measure the perceived pleasantness of learning with live teaching. Perceived pleasantness (PP) refers to the positive psychological state of college students towards using live teaching for learning, which is the degree to which learners feel pleasant and happy about a specific scene. If students feel that their interest is aroused in the process of learning using live teaching and that it is more interesting to use live teaching compared to traditional teaching, then students’ willingness to learn using live teaching is stronger.

Table 7.

Sensory pleasure measurement design

Independent variable	Item number	Subject content
Perceived pleasure(PP)	PP1	When I was learning live, I felt like time was going fast
	PP2	Live learning will make my learning process more interesting
	PP3	The rich information in live learning can guide me to explore new knowledge well

(7)

Intention to use

The design of the willingness to use measurement question items is shown in Table 8. Based on the combination of the UTAUT model and related acceptance studies, three question options were modified and compiled to measure the willingness to use live teaching for learning. Willingness to use (UI) refers to the likelihood that learners believe they will use live teaching for learning in the future. If a student’s willingness to use live teaching is strong, the more likely they are to use it for learning in subsequent studies.

Table 8.

Use the willingness to measure the design

Independent variable	Item number	Subject content
Use will(UI)	UI1	If the equipment resources are adequate, I would like to use live learning
	UI2	I hope to continue to use live learning in the future
	UI3	I am looking forward to the popularization of live learning in the general school

3)

Student live teaching acceptance model

According to the previous elaboration, the model of acceptance of information technology teaching platform for students in colleges and universities is shown in Figure 2. The theoretical model of this study consists of three types of variables: dependent variables, independent variables, and moderating variables. The dependent variable is willingness to use. The independent variables include: performance expectations, effort expectations, social influence, enabling conditions, and added self-efficacy and perceived pleasantness. The moderating variables were gender and grade level.

2.2

Principles and Methods

After constructing the research model, the proposed hypotheses need to be verified, and the data collected and analyzed using quantitative research methods.Firstly, the research instrument for this study needs to be designed by fully drawing on authoritative scales related to the technology acceptance model. Based on the research questions, the measurement items of the questionnaire were initially designed. Then the pre-tested survey respondents were selected and filled out by students who had conducted learning activities in the informatization teaching and training platform for materials majors.A total of 425 questionnaires were sent online, and finally, a total of 425 questionnaires were returned. Finally, SPSS software was used to organize and analyze the data results of the questionnaire to provide empirical data support for the subsequent acquisition of reliable research conclusions.

2.2.1

Levene’s test of variance alignment

Levene’s test of chi-squaredness is also known as Levene’s test. Levene’s test is mainly used to test the chi-squaredness of the variance between two or more samples.The samples are required to be randomly selected and independent of each other. The common Bartlett multi-sample variance chi-square test is mainly used for normally distributed data, for non-normally distributed data, the test effect is not ideal.Levene test can be used for normally distributed data, but also can be used for non-normally distributed data or the distribution of the data is not clear, the test effect is more ideal.

The principles and methods of the test are as follows:

1) Test hypothesis H₀:σ₁ = σ₂ =…= σ_k, that is, the variance of each treatment group is equal. H₁:σi ≠ σ_j, that the variances of the treatment groups are not all equal. Take α = 0.05 or α = 0.01.

2) Calculate the value of test statistic w (1) $W = \frac{(N - k) \sum_{i = 1}^{k} N_{i} {(Z_{i} . - Z ..)}^{2}}{(k - 1) \sum_{i = 1}^{k} \sum_{j = 1}^{N_{i}} {(Z_{i j} - Z_{i} ..)}^{2}}$ \[W=\frac{(\;N-k\;)\sum\limits_{i=1}^{k}{{{N}_{i}}}{{(\;{{Z}_{i}}.-Z..)}^{2}}}{(\;k-1\;)\sum\limits_{i=1}^{k}{\sum\limits_{j=1}^{{{N}_{i}}}{{{(\;{{Z}_{ij}}-{{Z}_{i}}..)}^{2}}}}}\]

Where: W is the Levene’s test statistic, k is the number of sample groups, N_i is the content of the ith sample, N is the sum of the contents of each sample, and Z_ij is the new value of the variable after transforming the original data by data transformation. Z_i is the mean of the ith sample, Z is the total mean of all the data.

Z_ij can be one of the following three definitions: (2) $Z_{i j} = | Y_{i j} - Y_{i} |$ \[{{Z}_{ij}}=|{{Y}_{ij}}-{{Y}_{i}}|\] Where: Y_ij is the original data and Y_i is the arithmetic mean of the ird sample in the original data. (3) $Z_{i j} = | Y_{i j} - Y_{i} |$ \[{{Z}_{ij}}=|{{Y}_{ij}}-{{Y}_{i}}|\]

Where: Y_i is the median of the ind sample in the original data. (4) $Z_{i j} = | Y_{i j} - Y_{i}^{'} |$ \[{{Z}_{ij}}=|{{Y}_{ij}}-Y_{i}^{\prime }|\]

Where: $Y_{i}^{'}$ \[Y_{i}^{\prime }\] is the 10% adjusted mean of the ind sample in the original data. 10% adjusted mean is the arithmetic mean of the data between P₅ and P₉₅ calculated by removing data less than P₅ and greater than P₉₅.

All of the above transformations of the raw data took the absolute value of the difference. These three transformations alone establish that the Levene test has a good degree of robustness and certainty. The magnitude of the calculated Levene statistic varies with the different ways of transforming the raw data. The Levene test in SAS and SPSS statistical software uses the first i.e., the data transformation of the formula [28]. The three types of data transformation can be applied to different types of data: mainly for information that is symmetrically or normally distributed, the second transformation can be used for information that is skewed, and the third transformation can be used for information that has extreme values or outliers.

3) Principle of judgment Levene test statistic W obeys a F distribution with υ₁ = k – 1,υ₂ = N – k degrees of freedom. When W ≥ F(α,k – 1,N – k), then P ≤ α, reject H₀ at the α level, accept H₁, can be considered that the variance of each sample is not all equal. When W < F(α,k – 1,N – k), then P > α, do not reject H₀, can be considered each sample variance chi-square.

2.2.2

Pearson’s correlation coefficient

In statistics, the Pearson correlation coefficient is a widely used measure of linear correlation between two variables. It assesses the strength of the relationship between two vectors based on the covariance matrix of the data. Typically, the Pearson correlation coefficient between two vectors α_i and α_j is: (5) $P (α_{i}, α_{j}) = \frac{cov(α_{i}, α_{j})}{\sqrt{var(α_{i}) \times var(α_{j})}}$ \[P({{\alpha }_{i}},{{\alpha }_{j}})=\frac{\text{cov(}{{\alpha }_{i}},{{\alpha }_{j}}\text{)}}{\sqrt{\text{var(}{{\alpha }_{i}}\text{)}\times \text{var(}{{\alpha }_{j}}\text{)}}}\] where cov(α,α_i) is the covariance, var(α_i) is the variance of vector α_i and var(α_i) is the variance of vector α_i. The Pearson’s correlation coefficient can be applied to either the sample or the aggregate. The absolute value of the Pearson’s correlation coefficient for both the sample and the aggregate is less than or equal to. In the case of sample correlation, a correlation coefficient equal to 1. Or -1 corresponds to data points that lie exactly on a line: in the case of aggregate correlation, this corresponds to a line that fully supports a bivariate distribution. The Pearson correlation coefficient is symmetric: P(α_i,α_j) = P(α_j,α_i).

2.3

Principles of PLS-SEM

2.3.1

Introduction to SEM modeling

SEM is a second-generation statistic with two major families: covariance-based structural equation modeling (CB-SEM) and variance-based structural equation modeling (SEM), also known as structural equation modeling in principal component form (SEM). Partial Least Squares (PLS) is a typical analysis method for the second type of structural equation modeling, often referred to as PLS-SEM.

Based on the above comparison, the advantages of PLS-SEM are summarized as follows

1) Small sample analysis requirements.

2) Samples need not be normally distributed.

3) Constructs can include formative indicators.

4) Complex structures can be modeled.

5) No mature theoretical structure is required, which is suitable for the exploration and development stage of the theory.

In this paper, the PLS-SEM model is needed for subsequent analysis and research.

2.3.2

PLS-SEM build types

PLS-SEM consists of a reflective construct (same as SEM) and a formative construct. Each topic of the reflective indicator is a variation of each facet, and can be exchanged and replaced, generally up to 5 to 7 topics.Formative indicators are part of the formation of each topic. They cannot be exchanged. There is no limit to the number of indicators. In principle, the more the better.

2.3.3

PLS-SEM modeling process and methodology

1)

Typical SEM model

A typical SEM model contains two parts, the measurement model, which is used to define the latent variables (i.e., CFA), and the structural model, which is used to explore the influences and interactions among the latent variables (i.e., path analysis). The structural equation modeling is shown graphically in Figure 3.

Assume that there are ξ,η two latent variables, where ξ serves as the explanatory variable and η serves as the outcome variable. These two latent variables can be defined by two sets of measurement model equations and are called endogenous latent variables since the variation in η is explained by other factors [29]. ξ as an explanatory variable that affects others is called exogenous latent variable. Latent variables are often referred to as constructs along with observed variables, which are the questionnaire questions.

(6)

x = Λ_{x} ξ + δ

\[x={{\text{ }\!\!\Lambda\!\!\text{ }}_{x}}\xi +\delta \]

(7)

y = Λ_{y} η + ε

\[y={{\text{ }\!\!\Lambda\!\!\text{ }}_{y}}\eta +\varepsilon \]

where:

x = Vector of exogenous observed variables.

y = Vector of endogenous observed variables.

Λ_x = ξ The factor loadings matrix for x.

Λ_y = η The factor loadings matrix for y.

δ,ε = vector of measurement errors.

The relationship between two latent variables can be represented by the regression equation of the formula, i.e., the structural model of SEM. The structural model consists of regression equations so that when the endogenous latent variable acts as an explanatory variable for another endogenous latent variable, a recursive causal chain is formed and path analysis is performed, and the residuals in the model are independent of each other.

(8)

η = B η + Γ ξ + ζ

\[\eta =B\eta +\text{ }\!\!\Gamma\!\!\text{ }\xi +\zeta \]

where:

ξ = Vector of exogenous latent variables.

η = Vector of endogenous latent variables.

B = Regression path coefficients for the effect between different η.

Γ = ξ Impact regression path coefficients on η.

ζ = model measurement residuals.

2)

PLS-SEM model calculation method

PLS successively performs principal component estimation and regression solution for the measurement model and structural model to compose the parameter estimation of the whole model. The relationship between observed variables and latent variables is called the measurement model in SEM and the external model in PLS, while the relationship between latent variables and potential variables is called the structural model in SEM and the internal model in PLS.

In the external model (measurement model) part, based on different assumptions about the causal relationship between the observed variables and the latent variables, PLS has two forms of setting up, the first one assumes that the variation of the observed variables is determined by the latent variables, which is called the reflective model.

The second form of external modeling assumes that the variation in the number of latent variables is determined by the observed variables and is called formative modeling: (9) $ξ = Π x + v$ \[\xi =\text{ }\!\!\Pi\!\!\text{ }x+v\] Where: ξ = Exogenous latent variables.

x = Vector of exogenous observed variables.

Π = x The weight matrix for ξ.

v = Measurement error vector.

The specific computational steps are as follows:

1) Standardize the values of all observed variables (ξ for exogenous latent variables, η for endogenous latent variables) (10) $x_{i j} = \frac{x_{i j} - \bar{x_{j}}}{\sqrt{v a r x_{j}}}$ \[{{x}_{ij}}=\frac{{{x}_{ij}}-\overline{{{x}_{j}}}}{\sqrt{var{{x}_{j}}}}\] (11) $y_{i j} = \frac{y_{i j} - \bar{y_{j}}}{\sqrt{v a r y_{j}}}$ \[{{y}_{ij}}=\frac{{{y}_{ij}}-\overline{{{y}_{j}}}}{\sqrt{var{{y}_{j}}}}\]

Among them:

i = Observation sample size.

j = Number of observed variables.

$\bar{x_{j}} = ξ$ $\overline{{{x}_{j}}}=\xi $ Potential variables j Mean of observed variables.

$\bar{y_{j}} = η$ $\overline{{{y}_{j}}}=\eta $ Potential variable j Mean of the observed variable.

varx_j = ξ Potential variable j Variance of the observed variable.

vary_i = η Potential variables j Variance of observed variables.

2) Principal component analysis using regression (12) $t_{1} = E_{0} w_{1}$ \[{{t}_{1}}={{E}_{0}}{{w}_{1}}\] (13) $u_{1} = F_{0} c_{1}$ \[{{u}_{1}}={{F}_{0}}{{c}_{1}}\]

Among them:

t₁ = ξ Extracted principal components.

u₁ = η Extracted principal components.

E₀ = Observed variables x_ij Standardized matrix.

F₀ Observed variables y_ij Normalized matrix.

w₁ Unit vector, E₀ of the first axis, ∥w₁∥ = 1.

u₁ Unit vector, F₀ of the first axis, ∥u₁∥ = 1.

The requirements for extracting the principal components are as follows:

First, they need to be maximized to include the information in their respective observed variables. (14) $v a r (t_{1}) \to m a x$ \[var({{t}_{1}})\to max\] (15) $v a r (u_{1}) \to m a x$ \[var({{u}_{1}})\to max\]

Second, the correlation between the two principal components is maximized. (16) $r (t_{1}, u_{1}) \to m a x$ \[r({{t}_{1}},{{u}_{1}})\to max\]

The above steps are repeated and the remaining information after ξ is explained by t₁ is subjected to a mrd round of component extraction until the marginal contribution of the extracted component t_m to the predictive effect of the model is not significant.

3)

Iterative analysis

Since the parameter estimation of PLS-SEM is based on the purpose of maximizing the explanation of the residuals of the endogenous latent variables, the values fitted to the observed variables need to be maximized to achieve the degree of ξ to η explanation.

External approximation: factor scores for the latent variables were calculated using regression (iterated until convergence). Assign the same initial weights to the variables, calculate new weights by treating the latent variable variance as unit variance, and then iteratively bring in the estimated weights to obtain variance explanatory power with higher explanatory power.

Internal approximation: calculating with similar latent variables as proxies is still done with regression (iterating until convergence). A new estimate of the latent variable can be obtained and then the correlation coefficient (reflective model when measuring the model) regression coefficient (formative model when measuring the model) corresponding to the new latent variable with each observed variable is calculated, this regression result is the weight of the external approximation in the next iteration, and the calculation is stopped when |wⁱ – wⁱ⁺¹|<10⁻⁵ or |(wⁱ – wⁱ⁺¹)/wⁱ|<10⁻⁵.

3

Model testing and analysis

3.1

Statistical analysis of variables related to moderator variables

3.1.1

Descriptive statistical analysis of variables

Descriptive analysis of statistics helps to integrate data and extract key information from it. In order to further understand the extent of each variable on students’ willingness to use live learning, the skewness and kurtosis values need to be checked to confirm normality. The skewness and kurtosis values of the statistics are analyzed to determine if the sample data in this questionnaire conforms to a normal distribution. According to the skewness and kurtosis assumptions, they are met if their values are between -1 and +1. The descriptive statistical analysis of each variable is shown in Table 9. The data indicate that the skewness value of each variable is between -1 and +1, and the kurtosis value is less than 1, which are within the range of the indicators of normal distribution, so the distribution of this sample data is normally distributed. Next, the mean values of each question item and the mean values of each variable were analyzed sequentially. As can be seen from the table, among all the core variables, the mean values of each variable are similar, with values ranging from 3.61 to 3.91, and a detailed analysis of each variable shows that the mean value of perceived pleasantness PP is the highest, with a value of 3.91, which indicates that the students perceive more pleasantness for the learning of the materials professional practical training course based on live teaching. The mean value of the measurement indicator FC3 (the operating system of the live teaching platform is relatively simple and convenient) is 4.09, indicating that the live teaching platform, as a new type of teaching equipment, brings college students a highly efficient learning experience. The mean value of Social Impact SI is 3.75, second only to Perceived Pleasure, indicating that students generally have better learning satisfaction in the environment of using live teaching. The third highest mean value was Performance Expectation at 3.67.

Table 9.

Descriptive statistical analysis of variables

Variable	Measuring index	Mean	Variable mean	Degree of bias		Kurtosis
Variable	Measuring index	Mean	Variable mean	Statistic	Standard error	Statistic	Standard error
PE	PE1	3.5	3.67	-0.637	0.154	-0.019	0.315
	PE2	3.81		-0.571	0.154	-0.334	0.315
	PE3	3.58		-0.67	0.154	-0.197	0.315
	PE4	3.78		-0.749	0.154	0.046	0.315
EE	EE1	3.64	3.64	-0.743	0.154	-0.259	0.315
	EE2	3.5		-0.631	0.154	0.046	0.315
	EE3	3.77		-0.656	0.154	-0.236	0.315
SI	SI1	3.68	3.75	-0.581	0.154	-0.22	0.315
	SI2	3.74		-0.637	0.154	-0.428	0.315
	SI3	3.82		-0.635	0.154	-0.064	0.315
FC	FC1	3.6	3.65	-0.54	0.154	-0.477	0.315
	FC2	3.86		-0.386	0.154	-0.468	0.315
	FC3	4.09		-0.894	0.154	0.404	0.315
	FC4	3.31		-0.529	0.154	-0.284	0.315
	FC5	3.4		-0.541	0.154	-0.285	0.315
SE	SE1	4.06	3.61	-0.622	0.154	-0.147	0.315
	SE2	3.17		-0.802	0.154	0.58	0.315
	SE3	3.6		-0.844	0.154	0.183	0.315
PP	PP1	3.75	3.91	-0.969	0.154	0.429	0.315
	PP2	4.08		-0.86	0.154	0.156	0.315
	PP3	3.89		-0.89	0.154	0.164	0.315
UI	UI1	3.72	3.62	-0.71	0.154	0.315	0.315
	UI2	3.25		-0.841	0.154	0.155	0.315
	UI3	3.44		-0.784	0.154	0.185	0.315

3.1.2

Analysis of differences in research variables

When analyzing the relationship between variables on the intention to use the information technology teaching platform, students’ gender and grade level are considered to have a moderating effect on the intention to learn behavior, so this study analyzes the moderating effect through the analysis of variance of variables. Variance study is used to compare the differences of several groups of data, the gender and grade level of this study are fixed category data, with the help of SPSS software for gender data in the t-test of independent samples t-test to analyze the data. One-way ANOVA was performed on grade category data.

1)

Gender

Gender is categorical data, and the two types of sample data, male and female, are independent of each other, which allows for comparison of whether there are significant differences in the impact factors of different gender subgroups. Setting gender as a grouping variable, an independent samples t-test was performed. Levene Levene was used to determine whether the variances of the independent samples were equal. First of all, from the results of Levene test of variance chi-square, we see that the F-value corresponding to the Sig value is less than 0.05, which rejects the hypothesis of variance chi-square, indicating that the variances are not equal in general, i.e., the variance is not chi-square. It is necessary to look at the results of the next row of the “assumption of unequal variances”, the corresponding t-value and the corresponding sig value is less than 0.05, the original hypothesis is rejected, indicating that there is a significant difference in the mean values. The results of an independent sample t-test for the gender variable are shown in Table 10.

Table 10.

Independent sample t test results of gender variables

Variable		Levene test of variance				T test of the mean equation
Variable		F	Sig.	t	df	sig.2	Mean difference	Standard error difference
PE	The variance is equal	35.462	0.000	-0.953	232	0.338	-0.11823	0.12465
PE	Assumed variance	35.462	0.000	-0.914	202.31	0.34	-0.11224	0.12008
EE	The variance is equal	29.445	0.000	-1.066	232	0.303	-0.12734	0.12526
EE	Assumed variance	29.445	0.000	-1.068	205.641	0.297	-0.12896	0.12317
SI	The variance is equal	27.166	0.000	-0.157	232	0.855	-0.01999	0.12053
SI	Assumed variance	27.166	0.000	-0.177	211.35	0.896	-0.01837	0.1197
FC	The variance is equal	30.337	0.000	-2.505	232	0.028	-0.29402	0.11951
FC	Assumed variance	30.337	0.000	-2.562	185.34	0.001	-0.29682	0.1122
SE	The variance is equal	56.103	0.000	-3.222	232	-0.014	-0.37886	0.11881
SE	Assumed variance	56.103	0.000	-3.263	206.67	0.031	-0.37806	0.11584
PP	The variance is equal	47.906	0.000	-2.592	232	0.021	-0.32811	0.12984
PP	Assumed variance	47.906	0.000	-2.665	208.34	0.006	-0.33457	0.12505
UI	The variance is equal	38.62	0.000	-2.315	232	0.025	-0.33654	0.11546
UI	Assumed variance	38.62	0.000	-2.666	205.64	0.026	-0.31521	0.13624

From the test of variance alignment, it can be seen that the sig values of the three variables of performance expectation PE, effort expectation EE and social influence SI are all greater than 0.05, and the assumption of variance alignment is valid, so the results of the analysis of variance are reliable. The Sig values corresponding to the F-values of Facilitating Conditions FC, Self-Efficacy SE and Perceived Pleasure PP are all less than 0.05, which means that except for the three variables FC, SE and PP, which are not significantly different, there are significant differences in the other three variables PE, EE and SI. Therefore, it can be concluded that there is a significant difference in the factors influencing students’ intention to learn and use the information technology teaching platform by gender. That is, the research hypothesis H11 proposed in the previous section is valid.

2)

Grade level

The moderating effect of grade level was analyzed using a one-way ANOVA in ANOVA. One-way ANOVA was used to analyze whether there was a significant difference in the means under the influence of a single control variable (grade level). The grade level one-way ANOVA and grade level ANOVA chi-square test are shown in Tables 11 and 12. From the two tables, it can be seen that based on the F-value data of the six variables and the value of significance (greater than 0.05), it means that there is no significance on these six variables. It is necessary to continue to do the analysis of Levine’s variance chi-square test, as can be seen in Table 12, the value of significance is greater than the level of significance of 0.05, which is considered to be chi-square between the sample data, which indicates that there is no significant difference in the factors influencing the students’ intention to learn and use the information technology-based teaching and learning platforms in the environment, depending on their grade level. The research hypothesis H12 that was proposed in the previous section is invalid.

Table 11.

Grade analysis

	PE	EE	SI	FC	SE	PP	UI
F	1.133	0.76	1.34	0.946	1.734	0.357
Significance	0.349	0.564	0.256	0.439	0.12	0.803

Table 12.

Grade variance test

	Levene statistic	df1	df2	Significance
PE	2.093	4	236	0.157
EE	1.141	4	236	0.324
SI	1.05	4	236	0.336
FC	1.759	4	236	0.159
SE	1.003	4	236	0.391
PP	0.914	4	236	0.438
UI	1.062	4	236	0.365

3.2

Correlation analysis

Correlation analysis explains the degree of correlation between the variables, and the degree of correlation between the research variables is analyzed by Pearson correlation coefficient, and the results of correlation analysis between the main research variables are shown in Table 13 (** indicates significant correlation at the 0.01 level (bilaterally), with a p-value <0.01). From the data in the table, it can be seen that the correlation coefficients between willingness to use and performance expectation, effort expectation, social influence, convenience, self-efficacy, and perceived pleasantness are 0.536, 0.432, 0.282, 0.478, 0.462, 0.459, respectively, and their corresponding p-values are less than 0.01, which is statistically significant, indicating that all the variables have a significant positive correlation with the live teaching platform use willingness have significant positive correlation.

Table 13.

Correlation analysis results

	PE	EE	SI	FC	SE	PP	UI
PE	1
EE	0.341**	1
SI	0.485**	0.212**	1
FC	0.187**	0.163**	0.132**	1
SE	0.326**	0.274**	0.145**	0.287**	1
PP	0.255**	0.185**	0.189**	0.364**	0.412**	1
UI	0.536**	0.432**	0.282**	0.478**	0.462**	0.459**	1

3.3

Structural equation modeling

On the basis of the UTAUT model, a research model was constructed for the materials students to receive the informatization teaching and training platform. The model was tested for fit using AMOS software. If the model meets the fit index, it is not modified and the final model is determined.If the model does not meet the fit indicators, it needs to be revised first. The final model is determined after the fit is completed and the final model is interpreted. It is important to emphasize here that since the two moderating variables, gender and grade, are categorical variables, the study cannot be well developed in the AMOS software. Therefore, the initial model constructed and revised in this section is without the two moderating variables.

3.3.1

Construction of Initial Structural Equation Modeling

In SEM, there are three types of variables: observed variables, latent variables, and residual variables. Observed variables are explicit variables that can be measured directly and are typically depicted in rectangles. Latent variables represent variables that are more abstract and difficult to measure, and are usually represented by ellipses. Residual variables are variables that are hard to measure directly, but they are still present in each latent variable and are typically depicted as circles. Causal relationships are represented by single arrows.

In conjunction with the current study, it can be seen that the observed variables in the SEM correspond to the 24 question items in the formal questionnaire of this study.The latent variables in the SEM correspond to the seven variables in this study, including six independent variables and one dependent variable that affect the physical acceptance of online learning among high school students, i.e., performance expectancy, effort expectancy, social influences, convenience conditions, self-efficacy, perceived pleasantness, and willingness to use. In addition, hypotheses H1 to H12 in this study, i.e., the positive effects of the six variables on the willingness to use the live teaching platform. These 12 hypotheses are connected with single arrows in SEM to indicate causal relationships. The residuals of each item in the initial model are standardized and defaulted to 1. The initial model SEM1 is drawn in AMOS software based on the analysis above, and the SEM1 path diagram is depicted in Figure 4.

3.3.2

Model fit test

After establishing the initial model, it needs to be tested for goodness-of-fit. Only when the indicators meet the fitting requirements can the model’s true path analysis be obtained. There are two common types of fit metrics, which are absolute fit metrics and benchmark fit metrics.

Among them, CMIN/DF (chi-square degrees of freedom ratio), RMSEA, RMR, GFI, and AGFI belong to the absolute fit metrics, and NFI, RFI, IFI, TLI, and CFI belong to the baseline fit metrics.The SEM1 fit is shown in Table 14. The data of the metrics that meet the fit criteria have been bolded in the table for ease of viewing. From the table, it can be seen that the initial model fit is not very good. Only EE currently meets the criteria, so the initial model SEM1 needs to be corrected.

Table 14.

SEM1 is a good thing

Fitting index	PE	EE	SI	FC	SE	PP
Standard	<2	<0.2	<0.06		>0.7
SEM1	7.245	0.005	0.137	0.851	0.83	0.871
Quasi fit	NO	YES	NO	NO	NO	NO

SEM4 unstandardized regression analysis is shown in Table 15. As can be seen from Table 15, the paths between the variables in SEM4 all reach the significance level.The fit of SEM4 is shown in Table 16. It can be found that all the fit indicators have met the standard, indicating that the model has been fitted at this point.In summary, in SEM4, not only are the paths between all variables significant, but also the model achieves the fit. Therefore, SEM4 is the revised and completed model.

Table 15.

SEM4 non-standardized regression analysis

	Estimate	S.E.	C.R.	P
H1	1.073	0.062	20.043	***
H2	0.387	0.037	6.18	***
H3	0.359	0.083	9.467	***
H4	0.599	0.03	12.738	***
H5	0.144	-0.004	3.641	***
H6	0.485	0.071	12.658	***
H7	0.257	0.056	7.817	***
H8	0.449	0.041	8.785	***
H9	0.317	0.089	4.232	***
H10	0.42	0.069	6.102	0.036

Table 16.

SEM4’s fitting

Fitting index	PE	EE	SI	FC	SE	PP
Standard	<4	<0.2	<0.06		>0.7
SEM3	2.965	0.045	0.045	0.966	0.934	0.945
SEM4	2.972	0.045	0.048	0.954	0.934	0.952
Standard	YES	YES	YES	YES	YES	YES

The model is constructed to study the relationship between six independent variables and one dependent variable, in order to make the final presentation of the model intuitive and concise, so the observed variables and residuals are deleted when the model is presented, and only the seven core variables are retained.The path diagram of SEM4 is shown in Fig. 5, and the numbers on the arrows indicate the path coefficients, and the larger path coefficients indicate that the degree of influence of the causal variable on the outcome variable is larger, if the path coefficient is greater than 0, it indicates that there is a positive correlation between the two, i.e., the cause variable has a positive influence on the outcome variable. From the figure, we can see the association between the seven variables: effort expectation, performance expectation, perceived pleasantness, content quality design, community influence, enabling conditions, and acceptance.

4

Conclusion

Combined with the UTAUT technology acceptance model, this study takes students of a school as the research object and explores the factors affecting the willingness of materials majors to accept the informatization teaching and training platform as well as the relationship between the factors, with a view to being able to effectively improve the willingness to use and the learning effect of students through the regulation between the potential variables. Specifically, the findings of this study are as follows:

The correlation coefficients between willingness to use and performance expectation, effort expectation, social influence, convenience, self-efficacy, and perceived pleasantness are 0.536, 0.432, 0.282, 0.478, 0.462, and 0.459, respectively, which means that the variables have a significant positive correlation with the willingness to use the informatization teaching platform. Performance expectations, effort expectations, social influence, enabling conditions, self-efficacy, and perceived pleasantness have a positive effect on the willingness to use the informationalized teaching platform. There is a significant difference in the willingness to use in different grades.

Lingua:: Inglese

Frequenza di pubblicazione:: 1 volte all'anno
Argomenti della rivista:: Scienze biologiche, Scienze della vita, altro, Matematica, Matematica applicata, Matematica generale, Fisica, Fisica, altro

Feed RSS della rivista

Research on the Reform of Material Specialized Practical Training Course Based on Informatization Teaching Platform

Qiujie Wu

Pubblicato online: 21 mar 2025

Ricevuto: 01 nov 2024

Accettato: 15 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0554

Parole chiavePearson’s correlation coefficient, Structural equation, UTAUT model, Informatized teaching platform

© 2025 Qiujie Wu, published by Sciendo

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

Parole chiave
Pearson’s correlation coefficient, Structural equation, UTAUT model, Informatized teaching platform