A Study of Teachers’ Digital Literacy Enhancement Strategies in the Context of Information Technology in Higher Education

Teachers’ digital literacy [28] plays a crucial role in promoting the digital transformation of education and developing students’ digital literacy. In this study, the Ministry of Education issued the “Teacher Digital Literacy” industry standard based on the preparation of relevant questions, designed the questionnaire survey to grasp the first-hand information that can reflect the level of teachers’ digital literacy, and the questionnaire data were statistically analyzed in order to comprehensively understand the current situation of the level of teachers’ digital literacy. At the same time, through the summary of the relevant literature, the establishment of the influence factor model of college teachers’ digital literacy, to provide a certain theoretical basis for the subsequent study of the relevant influencing factors, and at the same time can provide a certain reference to put forward the enhancement strategy of teachers’ digital literacy.

The survey focused on selecting 352 teachers from seven well-known universities in China. In determining the survey respondents, the special situation of each university was taken into account as much as possible, and the impact of various factors on the sample of this study was fully considered to ensure that the survey data obtained in the study were typical and representative.

In the industry standard of Teachers’ Digital Literacy issued by China’s Ministry of Education, teachers’ digital literacy is categorized into five dimensions, namely digital awareness, digital technology knowledge and skills, digital application, digital social responsibility and professional development, each of which includes two to four secondary dimensions, and each of which is subdivided into multiple tertiary dimensions, which provides important theoretical support for measuring and evaluating teachers’ digital literacy.

In order to ensure the applicability of the questionnaire, a team of experts was consulted to validate it. The experts carefully and meticulously validated the questions of the questionnaire against the connotation of teachers’ digital literacy, and made comments on whether some questions matched the dimensions and whether the questions were properly formulated. After repeated revisions of the number of questions, statements and dimensional correspondences of the questionnaire, the questionnaire was finally finalized with the unanimous endorsement of the experts, which consisted of 44 questions of “Teacher Digital Literacy Survey” questionnaire. This questionnaire is divided into three parts. The first part is a survey on the basic information of individuals, including 11 questions about the gender, age, type of school district, title, professional background and other basic personal information of the respondents, which is mainly based on the range of questions available to design options for the respondents to choose from; the second part of the survey is a survey on the current status of digital literacy, which is divided into five dimensions based on the criteria: Six questions were designed in the digital awareness dimension, including questions 12 to 17; Four questions were designed in the digital technology knowledge and skills dimension, including questions 18 to 21, and 13 questions were designed in the digital applications dimension, including questions 22 to 34; Five questions were designed in the digital social responsibility dimension, including questions 35 to 39; Five questions were designed in the professional development dimension, including questions 40 to 44, and this part mainly used a Likert scale to set out response options of varying degrees for the respondents to choose from, including strongly agree, agree, unsure, disagree, and strongly disagree, and assigning a value of 5, 4, 3, 2, and 1 points to the above mentioned options, respectively. The 33 questions in the second part were numbered from Q12 to Q44.

To ensure the validity and reliability of this questionnaire, a small-scale pre-survey will be conducted by distributing the questionnaire to 100 college teachers from famous universities in China using questionnaire star. The questionnaire was distributed to 100 teachers, 100 questionnaires were recovered, and the valid number of questionnaires was 100, of which the recovery rate of the questionnaire was 100% and the validity rate of the questionnaire was 100%, which satisfied the data requirements. The questionnaires obtained from the pre-survey are now subjected to data processing to verify the validity and reliability of the questionnaires. In this study, the reliability analysis of the questionnaire recovered in the pre-survey was carried out to obtain the Cronbach’s alpha coefficient value of digital literacy overall, and the Cronbach’s alpha coefficient value of each dimension and the combined reliability CR value, and the results of the reliability analysis of the questionnaire are shown in Table 1. It can be found that the Cronbach’s alpha coefficients of digital literacy overall and at each dimension level are above 0.9, indicating that the questionnaire has very good reliability. The CR value of each dimension is also greater than 0.9, reflecting that all the questions in the questionnaire explain the dimension with very good consistency, so the questionnaire of the Survey on the Current Status of Teachers’ Digital Literacy designed in this study has a high reliability. Second, the validity analysis of the questionnaire. Validity analysis is a statistical method that evaluates the validity and accuracy of a measurement instrument, with a focus on whether it is effective in determining the concept or attribute to be measured. The extracted mean variance and factor loading values for the observed variables are compared. The standardized loading coefficients for each of the observed variables are greater than 0.55, and the AVE values of the mean variance extracted for each of the latent variables are greater than 0.6, which indicates that the questionnaire has strong explanatory power, and good internal consistency reliability and convergent validity. The questionnaire survey was scheduled from February 2024 to June 2024 Finally, 352 questionnaires were distributed and collected, 348 valid questionnaires were collected, and the validity rate of the questionnaire was 98.86%. The questionnaire’s validity rate is in compliance with the requirements, and the data from the official questionnaire can be utilized to initiate the analysis.

The basic idea of one-way analysis of variance (ANOVA) for quantitative data with a single-factor multilevel design is the decomposition of the sum of squares of the total deviations, i.e., the sum of squares of the deviations of all the data with respect to the total mean is decomposed into the sum of squares of the deviations between the groups and the sum of squares of the deviations within the groups (or errors), with a similar decomposition for the degrees of freedom. Dividing the sum of squares of the separate means of each component by their respective degrees of freedom is the variance (or mean square) of each component. Using the within-group (or error) mean square as the denominator and the between-group mean square as the numerator, a test statistic F can be constructed.

For a one-factor, multilevel design with one-dimensional quantitative information, the total sum of squared deviations from the mean SS_Total can be decomposed by the following equation: (1)SSTotal=SSIntergroup+SSError$$S{S_{{\text{Total}}}} = S{S_{{\text{Intergroup}}}} + S{S_{{\text{Error}}}}$$

The expression for the three terms of the sum of squared deviations from the mean in Eq. (1) is given below: (2)SSTotal=∑j=1k∑i=1nj(Yij¯−Y¯..)2$$S{S_{{\text{Total}}}} = \sum\limits_{j = 1}^k {\sum\limits_{i = 1}^{{n_j}} {{{\left( {{Y_{\overline {ij} }} - {{\bar Y}_{..}}} \right)}^2}} }$$ (3)SSIntergroup=∑j=1knj(Y¯j−Y¯..)2$$S{S_{{\text{Intergroup}}}} = \sum\limits_{j = 1}^k {{n_j}} {\left( {{{\bar Y}_j} - {{\bar Y}_{..}}} \right)^2}$$ (4)SSError=∑j=1k∑i=1n(Yij−Y¯..)2$$S{S_{{\text{Error}}}} = \sum\limits_{j = 1}^k {\sum\limits_{i = 1}^n {{{\left( {{Y_{ij}} - {{\bar Y}_{..}}} \right)}^2}} }$$ (5)dfIntergroup=Number of groups−1,dfError=N−1−dfIntergroup$$d{f_{{\text{Intergroup}}}} = {\text{Number of groups}} - 1,d{f_{{\text{Error}}}} = N - 1 - d{f_{{\text{Intergroup}}}}$$

Construct mean square MS based on the sum of squared off-mean deviations and degrees of freedom, see Eqs. (6) to (7): (6)MSIntergroup=SSIntergroupdfIntergroup$$M{S_{{\text{Intergroup}}}} = \frac{{S{S_{{\text{Intergroup}}}}}}{{d{f_{{\text{Intergroup}}}}}}$$ (7)MSError=SSErrordfError$$M{S_{{\text{Error}}}} = \frac{{S{S_{{\text{Error}}}}}}{{d{f_{{\text{Error}}}}}}$$

Based on the mean square construct test statistic F, see equation (8): (8)F=MSIntergroupMSError$$F = \frac{{M{S_{{\text{Intergroup}}}}}}{{M{S_{{\text{Error}}}}}}$$

In equation (8), F obeys a F distribution with df_Intergroup degrees of freedom in the numerator and df_Error degrees of freedom in the denominator.

If manual calculation is used, it is necessary to check the table of F bounded values (one-sided test) to obtain F(1−α)(dfIntergroup,dfError)$${F_{\left( {1 - \alpha } \right)\left( {d{f_{{\text{Intergroup}}}},d{f_{{\text{Error}}}}} \right)}}$$, if F≥F(1−α)(dfIntergroup,dfError)$$F \geq {F_{\left( {1 - \alpha } \right)\left( {d{f_{{\text{Intergroup}}}},d{f_{{\text{Error}}}}} \right)}}$$, then P ≤ α, and vice versa, then P > α. Finally, the P value is determined and statistical inferences are made, and then professional conclusions are given in conjunction with specialized knowledge.

Provided that the F test between treatments is significant, the least significant difference LSD_a at a significant level of α is first calculated, and then the difference (y¯i−y¯j)$$\left( {{{\bar y}_i} - {{\bar y}_j}} \right)$$ between any 2 means is calculated such that if its absolute value ≥ LSD_α is significant, the difference is significant at the α level; conversely, the difference is not significant at the α level. LSD_a The calculation formula is as follows: (9)LSDα=ta2MSen$$LSD\alpha = {t_a}\sqrt {\frac{{2M{S_e}}}{n}}$$

where MS_e is the mean square of the “within group”, “within” or “error” term in the ANOVA table; n is the number of replicates in the group; t_α is a two-tailed t value at the level of the degree of freedom α of “within group”, “within” or “error”.

The data available were analyzed by using analytical methods such as Pearson’s correlation coefficient [29-30] and mean value. Assuming X=(x1,x2,⋯,xN)$$X = \left( {{x_1},{x_2}, \cdots ,{x_N}} \right)$$, Y=(y1,y2,⋯,yN)$$Y = \left( {{y_1},{y_2}, \cdots ,{y_N}} \right)$$, the mean value is defined as: (10)x¯=∑i=1NxiN,y¯=∑i=1NyiN$$\bar x = \frac{{\sum\limits_{i = 1}^N {{x_i}} }}{N},\bar y = \frac{{\sum\limits_{i = 1}^N {{y_i}} }}{N}$$

The Pearson correlation coefficient is defined as: (11)ρXY=N∑i=1Nxiyi−∑i=1Nxi∑i=1NyiN∑i=1Nxi2−(∑i=1Nxi)2N∑i=1Nyi2−(∑i=1Nyi)2$${\rho _{XY}} = \frac{{N\sum\limits_{i = 1}^N {{x_i}} {y_i} - \sum\limits_{i = 1}^N {{x_i}} \sum\limits_{i = 1}^N {{y_i}} }}{{\sqrt {N\sum\limits_{i = 1}^N {x_i^2} - {{\left( {\sum\limits_{i = 1}^N {{x_i}} } \right)}^2}} \sqrt {N\sum\limits_{i = 1}^N {y_i^2} - {{\left( {\sum\limits_{i = 1}^N {{y_i}} } \right)}^2}} }}$$

Clearly −1 ≤ ρ_XY ≤ 1. When ρ_XY = 0, X and Y are not linearly correlated; when ρ_XY > 0, X and Y are positively correlated; when ρ_XY < 0, X and Y are negatively correlated; and when ρ_XY is closer to ±1, the correlation is higher.

In order to find an approximate mathematical expression to describe the correlation between multiple variables, the mathematical expression derived using mathematical statistics is called a regression equation model. The matrix equation of the multiple linear regression model is given below. Let random variable Y be affected by p − 1 non-random factors and random factor ε such that: (12)Y=[ Y1 Y2 ⋮ Yn],X=[ 1 X11 X12 ⋯ X1,p−1 1 X21 X22 ⋯ X2,p−1 ⋮ ⋮ ⋮ ⋮ 1 Xn,1 Xn,2 ⋯ Xn,p−1] β=[ β0 β1 ⋮ βp−1],ε=[ ε1 ε2 ⋮ εn]$$\begin{array}{*{20}{l}} {Y = \left[ {\begin{array}{*{20}{c}} {{Y_1}} \\ {{Y_2}} \\ \vdots \\ {{Y_n}} \end{array}} \right],X = \left[ {\begin{array}{*{20}{c}} 1&{{X_{11}}}&{{X_{12}}}& \cdots &{{X_{1,p - 1}}} \\ 1&{{X_{21}}}&{{X_{22}}}& \cdots &{{X_{2,p - 1}}} \\ \vdots & \vdots & \vdots &{}& \vdots \\ 1&{{X_{n,1}}}&{{X_{n,2}}}& \cdots &{{X_{n,p - 1}}} \end{array}} \right]} \\ {\beta = \left[ {\begin{array}{*{20}{c}} {{\beta _0}} \\ {{\beta _1}} \\ \vdots \\ {{\beta _{p - 1}}} \end{array}} \right],\varepsilon = \left[ {\begin{array}{*{20}{c}} {{\varepsilon _1}} \\ {{\varepsilon _2}} \\ \vdots \\ {{\varepsilon _n}} \end{array}} \right]} \end{array}$$

The matrix form of the out multiple linear regression model is as follows: (13)Y=Xβ+ε$$Y = X\beta + \varepsilon$$

In Eq. (13) Y, X are obtained from the observed data and X is assumed to be column full rank, β is the vector of unknown parameters to be estimated, and ε is the vector of unobservable random errors. 1)

Estimation of regression parameter β

The least squares estimation of regression parameter β yields the following equation as Eq: (14)XTXβ=XTY$${X^T}X\beta = {X^T}Y$$

Since X is column full rank, (XTX)−1$${\left( {{X^T}X} \right)^{ - 1}}$$ exists and solving equation (14) yields the least squares estimate of parameter β and is an unbiased estimate as equation (15): (15)β^=(XTX)−1XTY$$\hat \beta = {\left( {{X^T}X} \right)^{ - 1}}{X^T}Y$$ 2)

ANOVA table and significance test of linear regression relationship

Create an ANOVA table:

Based on the residual sum of squares SSE and regression sum of squares SSR of the data, the total deviation sum of squares of the data can be found SST: (16)SST=SSE+SSR$$SST = SSE + SSR$$

An ANOVA table was created as shown in Table 2. The J in the table represents a nnd order matrix with all elements 1, and MSE is an unbiased estimate of the error variance.

Significance of the regression equation tested by statistic F:

The test of significance of the regression equation can be done using Table 1 and the test hypothesis is given in the following equation: (17){ H0:β1=β2=⋯=βp−1=0, H1:At least some βi≠0,1≤i≤p−1$$\left\{ {\begin{array}{*{20}{l}} {{H_0}:{\beta _1} = {\beta _2} = \cdots = {\beta _{p - 1}} = 0,} \\ {{H_1}:{\text{At least some }}{\beta _i} \ne 0,1 \leq i \leq p - 1} \end{array}} \right.$$

The test statistic constructed as in Eq. (18) is: (18)F=MSRMSE$$F = \frac{{MSR}}{{MSE}}$$

When H₀ is true, F~F(p−1,n−p)$$F\sim F\left( {p - 1,n - p} \right)$$ is established; when H₀ is false, there is a tendency for the value of F to be skewed and analyzed at a set level of significance α.

Minimum, maximum, mean, and standard deviation of each dimension and overall level of teachers’ digital literacy. The results of the descriptive statistics of digital literacy are shown in Table 3. The table shows that among the five dimensions of digital literacy, the digital social responsibility dimension has the highest mean value of 4.25, which indicates that teachers are more capable of assuming digital social responsibility. The digital awareness dimension had the second highest mean value of 4.05. The professional development dimension had a mean value of 3.76, the digital technology knowledge and skills dimension had a mean value of 3.70, and the minimum value of this dimension was 1.03, indicating that some individual teachers perceived their digital technology knowledge and skills to be poor; the digital application dimension had the lowest mean value of 3.55, indicating that teachers’ self-perceived digital application competence has the most room for improvement among the five dimensions. The overall digital literacy level was 3.862, which indicates that teachers’ digital literacy level was moderately high and required further improvement.

Before performing t-test, ANOVA, Pearson correlation analysis, regression analysis, and structural equation modeling tests on the data, it is necessary to test the data for normality. This is because data conforming to a normal distribution is a prerequisite for performing the statistical analysis above. The skewness and kurtosis values of the samples were obtained through SPSS 26, and the results of the normality test of the questionnaire data are shown in Table 4. When the absolute value of skewness is less than 2 and the absolute value of kurtosis is less than 6, it can be assumed that the sample data follows a normal distribution. From the table, it can be seen that the absolute value of skewness of all question item data is less than 1, and the absolute value of kurtosis is less than 1, which is in line with the standard of normal distribution, so the sample data can be analyzed in the next test.

Teachers’ gender, which includes both male and female, is a dichotomous variable, so an independent samples t-test was used to analyze the differences in digital literacy among teachers of different genders. The results of the comparison of differences in digital literacy among teachers of different genders are shown in Table 5. It can be seen that the mean values of the digital awareness dimension of teachers of different genders are equal (M=4.063), and the mean values of the other dimensions of digital literacy and the overall level of digital literacy are slightly higher for males than for females, but whether or not the difference between males and females is significant will depend on the test of the t-statistic. The t-statistics tested for the teacher’s gender variable on the five dimensions of digital literacy and the overall level of digital literacy did not reach a significant level, and the p-values were all greater than 0.05, indicating that there is no significant difference between teachers of different genders on the five dimensions of digital literacy and the overall level of digital literacy. Overall, the digital literacy level of male teachers was 3.9974 and that of female teachers was 3.8908, which was slightly higher for males than for females, but there was no significant difference in digital literacy among teachers of different genders.

In the questionnaire of this study, age was a five-categorical variable, so one-way ANOVA was used. The results of the comparison of differences in digital literacy among teachers of different ages are shown in Table 6. It can be seen that there is no significant difference between teachers of different ages in the dimensions of digital awareness and digital technology knowledge and skills, and the F-values of the overall test are 1.873 and 2.897 (p=>0.05), respectively. There were significant differences in the three dimensions of digital application, digital social responsibility, and professional development, and the F-values of the overall test were 4.401, 4.728, and 4.330 (p < 0.05), respectively. There was a significant difference in the overall level of digital literacy among teachers of different ages, with an F-value of 3.945 for the overall test (p=0.004<0.05). Overall, teachers aged 30 and below had the highest level of digital literacy (M=4.257), teachers aged 30-40 had the second highest level of digital literacy (M=3.888), and teachers aged 40-45 and 45-50 had the lowest level of digital literacy, (M=3.605). It can be seen that there is a significant difference in the digital literacy of teachers of different ages.

In the questionnaire of this study, educational qualification was a five-categorical variable, so one-way ANOVA was used. The results of comparing the differences in digital literacy among teachers with different academic qualifications are shown in Table 7. Looking at the p-value of the F-test, p<0.05 indicates a significant difference and p>0.05 indicates no significant difference. From the table, it can be seen that the p-values of teachers with different degrees in the dimensions of digital literacy and the overall level of digital literacy are greater than 0.05, which indicates that there is no significant difference between teachers with different degrees in the five dimensions of digital literacy and the overall level of digital literacy. Overall, teachers with a doctoral degree had the highest level of digital literacy (M=4.507) and teachers with a master’s degree had the second highest level of digital literacy (M=4.417). Teachers with a high school degree or less had the lowest level of digital literacy (M=3.364). It can be seen that there is no significant difference in the digital literacy of teachers with different degrees.

Correlation analysis was used to study the correlation between digital awareness, digital technology knowledge and skills, digital application, digital social responsibility, professional development, and digital literacy, and the strength of the correlation was expressed by the Pearson correlation coefficient. The correlations between the dimensions are shown in Table 8. As can be seen from the table, there is a significant positive correlation between all dimensions (P < 0.01).

In this paper, a linear regression was conducted with digital technology knowledge and skills, digital application, digital social responsibility, professional development, and digital literacy as independent variables and digital awareness as dependent variable. The results of the linear regression analysis with digital awareness as the dependent variable are shown in Table 9. The results show that the regression coefficients of the five independent variables of “digital technology knowledge and skills, digital application, digital social responsibility, professional development and digital literacy” are 0.462, 0.173, 0.219, 0.068 and 0.002 respectively, and the t-values are 30.045, 10.077, 19.807, 4.64 and 0.078 respectively, 19.807, 4.64 and 0.078 respectively; the p-value of all dependent variables except digital literacy is 0.000. Summary: Digital technology knowledge and skills, digital application, digital social responsibility, and professional development will positively affect digital awareness, while digital literacy will not have a significant effect on digital awareness.

The results of linear regression analysis with digital technology knowledge and skills as the dependent variable are shown in Table 10. From the specific analysis, it can be seen that the regression coefficients of the five independent variables of “digital awareness, digital application, digital social responsibility, professional development and digital literacy” are 0.639, 0.012, 0.065, 0.027 and 0.278, and the t-values are 61.323, 1.171, 5.079, 3.814 and 30.604, respectively; except for digital application, the other four independent variables and digital technology knowledge and skills are linear regression, 5.079, 3.814 and 30.604; except for digital applications, the differences between the other four independent variables and digital technology knowledge and skills were significant (F=7015.538, p=0.000). It can be seen that digital awareness, digital social responsibility, professional development and digital literacy will have a significant relationship with digital technology knowledge and skills, but digital application will not have an influential relationship with digital technology knowledge and skills (P=0.135 > 0.05).

The results of the regression analysis with digital application as the dependent variable are shown in Table 11. From the specific analysis, it can be seen that the regression coefficients of the five independent variables of “digital awareness, digital knowledge and skills, digital social responsibility, professional development and digital literacy” are 0.051, 0.303, 0.085, 0.086 and 0.563, and the t-values are 5.173, 29.814, 13.827, 10.394 and 60.099, and the differences between the five independent variables and digital technology knowledge and skills are significant (P=0.000), 13.827, 10.394 and 60.099, and the difference between the five independent variables and digital technology knowledge and skills is significant (p=0.000). It can be seen that digital awareness, digital technology knowledge and skills, digital social responsibility, professional development, and digital literacy will have a significant positive impact on digital adoption.

The results of the regression analysis with digital social responsibility as the dependent variable are shown in Table 12. From the specific analysis, it can be seen that the regression coefficients of the five independent variables of “digital awareness, digital technology knowledge and skills, digital application, professional development and digital literacy” are 0.766, 0.007, 0.247, 0.037 and 0.214, and the t-values are 57.393, 0.517, 19.007, 1.528 and 5.478, and the differences between the five independent variables and the digital technology knowledge and skills are significant (P < 0.007, 1.528 and 5.478, respectively), 19.007, 1.528 and 5.478, and the differences between the five independent variables and digital technology knowledge and skills were significant (p < 0.05). Apparently, there is a highly significant positive effect between digital awareness, digital application, and digital literacy on digital social responsibility (P < 0.01), and there is also a significant effect between digital technology knowledge and skills, professional development, and digital social responsibility (P < 0.05).

The results of the regression analysis with digital social responsibility as the dependent variable are shown in Table 13. From the specific analysis, it can be seen that the regression coefficients of the five independent variables of “digital awareness, digital technology knowledge and skills, digital application, digital social responsibility and digital literacy” were 0.096, 0.045, 0.075, 0.373 and 0.447, respectively, and the t values were 14.393, 4.217, 5.007, 29.528 and 57.478, respectively, and there were significant differences between the five independent variables and digital technology knowledge and skills (P<0.01). It can be seen that digital awareness, digital technology knowledge and skills, digital applications, digital social responsibility, and familiarity with information technology hardware and software will have a significant positive resonance relationship on professional development.

The results of regression analysis with digital social responsibility as the dependent variable are shown in Table 14. From the specific analysis, it can be seen that the regression coefficients of the five independent variables of “digital awareness, digital technology knowledge and skills, digital application, professional development, and digital social responsibility” are 0.002, 0.091, 0.339, 0.011, and 0.317, and the t-values are 0.155, 3.046, 13.656, 0.772, and 14.492, respectively, 13.656, 0.772, and 14.492.The above results indicate that digital technology knowledge and skills, digital applications, and digital social responsibility will have a positive impact relationship on digital literacy (P=0.000), while digital awareness and professional development do not have a significant effect on digital literacy (P>0.05).