On Teachers’ Perceptions of Development of Expertise from the Experience of Distance Teaching and Learning During the COVID-19 Pandemic

The current study explores the factors that influence teachers’ perception of the development of their own “teacher expertise” due to the distance teaching during lockdowns because of the COVID-19 pandemic. The sample consists of 133 primary and secondary school teachers in Greece. Exploratory and confirmatory factor analyses, hierarchical regression and correlation coefficients are used for the analyses of the data. The results show that teachers’ readiness regarding the use of software and hardware is positively correlated with a. teachers’ perception of students’ learning outcomes, b. cognitive and behavioural reaction to change for primary school teachers. Cognitive reaction to change and teachers’ perception of the students’ learning outcome, in turn, influence teachers’ perception about the development of their own “teacher expertise”. The importance of readiness and the cognitive reaction to changes in the teachers’ perception of the development of their “teacher expertise” highlights the need for continuous training of teachers in computer use and technology literacy.

Introduction

Distance teaching during the COVID-19 pandemic brought about unexpected changes in the pace of life and work. According to MacDonald and Hill (2022), prompting educators around the world to transition to distant teaching methodologies. Changes were in all aspects of the teaching procedure: in teaching time, from live to asynchronous with a simultaneous reduction in teaching hours, in pedagogical techniques, in the teaching environment where verbal communication had a dominant role, whereas non-verbal was significantly limited. This unexpected challenge, while initially daunting, has proven to be a rich source of valuable experiences for teachers. It is inevitable that the post-pandemic education system considers the digital competence of teachers as an integral qualification (Hargreaves, 2021).

Literature Review

The era of the COVID-19 pandemic has unexpectedly turned education procedures upside down, but also our lives in general. For some teachers, that period provided an excellent opportunity for professional development, as they enriched their knowledge of educational technology and had the opportunity to attend several seminars from home (Hargreaves, 2021). For others, the pandemic had been a chaotic period of nonstop work as they were attempting to meet family needs while fulfilling their educational duties (MacDonald & Hill, 2022), with a major challenge being the lack of resources (Weißenfelset al., 2022).

However, distance teaching of that period contributed to the enrichment of teaching and learning methods by utilizing technology, which, until recently, was not widely used by teachers (Michaelsen, 2021). It is a fact that the technology put at the service of education at all levels worldwide has developed and evolved rapidly, annihilating distances and making the process of teaching and learning possible (Rajapriya & Kalai Arasi, 2022). Both educators and students at all levels have been called upon to respond to the new reality by embracing technology as a gift, according to Lederman (2020). A necessary condition for the effectiveness of distance teaching is the faith of teachers in the facilitating power of technology. Sticking to the traditional model of live teaching could only bring negative results to the learning process (Rajapriya & Kalai Arasi, 2022). In contrast, research by Pendergast and O’Brien (2023) has shown that, in a post-pandemic world, education stakeholders have changed what they do because they feel they are more agile, adaptable and flexible, innovations that education stakeholders would like to keep.

Three challenges that teachers were asked to face during the pandemic period and which the educational community admits that they could use, in combination, in the future in lifelong teaching are (a) teacher expertise, (b) teaching digitally, and (c) teaching naturally. Rajapriya and Kalai Arasi (2022) mention that both students and teachers perceive the usefulness of online teaching and learning and have developed positive attitudes towards it. This was mainly contributed by the multitude of online digital tools that were used to achieve the goals and objectives of the learning process. In their work, most of the teachers were trained by institutions that gained hands-on experience and look forward to efforts to make online teaching more interactive to make it fruitful and effective for future use as well.

Many researchers have worked on the psychological impact of the pandemic and remote teaching on teachers. Among others, Weißenfelset al. (2022) have found that “the burnout components depersonalization and lack of accomplishment significantly increased from the pre- to post-COVID-19 outbreak, whereas emotional exhaustion did not” (p. 1). Correlational studies have found that teacher efficacy, perceptions of administrative support and attitudes toward change were correlated with teacher resilience and burnout during the pandemic (Sokalet al., 2020).

The current study focuses on correlational analyses of responses from Greek primary and secondary school teachers about one year after the compulsory remote teaching had been completed. The purpose of this study is to explore the impact of teachers’ perception of learning outcomes for the students, resilience, readiness, attitudes toward change and demographic characteristics on the perception of the teachers about the development of their own “teacher expertise”. In addition, the relations among factors of attitudes toward change, resilience, readiness toward the use of remote teaching tools, and teachers’ perceptions of learning outcomes are explored.

Method

Measures

Teacher Attitudes toward Change scale (Kin & Kareem, 2017) was used to assess, specifically, the cognitive reaction to change (COG), the affective reaction to change (AFF) and the behavioural reaction to change (BHV). In particular, to assess COG, the items “1. I find most changes to be pleasing,” “2. Most of my co-workers benefit from change,” and “3. Change often helps me perform better” were used. AFF was assessed through the items “1. I don’t like change,” “2. Change frustrates me,” and “3. Most changes at my school are irritating”. Last, to assess BHV, the items “1. I look forward to changes at school,” “2. I often suggest new approaches to things,” and “3. Change usually benefits the school” were used. The possible responses of the items on a 7-point Likert scale, with 1 representing the response Strongly disagree and 7 meaning Strongly agree.

In addition, the resilience of the teachers was assessed through the approach of Eddyet al. (2019), which used single-item scales. More specifically, the item “How stressful is your job right now?” was used to assess stress, and the item “How effectively are you managing the stress your job is causing you right now?” was used to assess coping. Both items get possible responses on a 7-point Likert-type scale. The subtraction of the variable of stress from that of coping gives rise to the resilience of the teacher, which is used in the analyses below.

Moreover, three items were used that are related to the teachers' readiness regarding the distance learning they had to utilize: “How would you rate the distance learning you utilized in terms of (1) the hardware and software infrastructure that you had at your disposal, (2) your readiness to handle the medium you used (e.g., computer, tablet), and (3) your readiness to utilize various software?” In addition, respondents were asked to rate the distance learning they utilized in terms of the learning outcomes for the students. These items get possible responses on a 5-point scale, with 1 representing Unsuccessful and 5 being Successful.

The dependent variable used in the study is the 7-point scale (very unlikely to very likely) variable. The item “Do you think the distance teaching experience you had is a useful additional instrument you are going to integrate into teaching?” is used as an assessment of how possible utilization of experience from distance education teachers is, or in other words, the teachers’ perceptions of how useful the experience of distance teaching and learning during the COVID-19 pandemic can be and what the benefits are. The dependent variable will be referred to as perception about the development of “teacher expertise” in the following.

Data Collection and Sample

The sample consists of 133 Greek primary and secondary school teachers. The study uses a convenient snowball sampling (Goodman, 1961). The data was collected during April and May 2022 through a Google Forms questionnaire. In-service teachers of primary (students aged 6–12 years old) and secondary schools (students aged 12–18 years old) from Greece were asked to respond to the anonymous questionnaire after they consented to participate through the first question of the form. The mean age of the respondents is 42.77 years (SD = 9.84 years), ranging from 21 to 66 years. Table I shows that, of the 133 participants, the majority are female (76.70%) from urban areas (79.70%). Holders of a master’s degree correspond to 48.90% of the sample, and 64.20% are primary school teachers.

Table I. Demographic Information of the Sample

Demographic variables		Frequency	%
Gender	Female	102	76.69
	Male	31	23.31
Region of school	Urban	106	79.70
	Rural	27	20.30
Studies	Bachelor	61	45.86
	MSc	65	48.87
	PhD	7	5.26
Educational stage in-service teacher works	Primary	79	64.23
	Secondary	44	35.77

Analyses

Non-parametric Spearman’s rho is used to estimate the correlation coefficients between variables. Moreover, exploratory and confirmatory factor analyses are used on the scales of the data set.

The internal consistency of the scales is assessed through Cronbach’s alpha as well as through the omega coefficient. In recent literature, an increasing number of researchers point out that Cronbach’s alpha coefficient suffers from major problems both as a reliability index and as a measure of internal consistency (Sijtsma, 2009), suggesting the use of alternative indices that overcome alpha’s problems, like omega (ω) or greatest lower bound (Malkewitzet al., 2022; Trizano-Hermosilla & Alvarado, 2016).

A hierarchical regression model with two blocks of explanatory variables is used on the teachers’ perception of their own development of “teacher expertise” due to the distance teaching they performed. The presence of multicollinearity in the model is tested through tolerance indices.

The assumption of normality of the multiple regression is tested via the Shapiro-Wilk test and normal P-P plot; both applied to the standardized residuals of the linear model. In educational and social sciences, a common mistake of statistical analyses used, which unfortunately often appears in the literature, is that researchers test the hypothesis of normality of the dependent variable or/and the explanatory continuous covariates. In fact, general linear models’ theory has no such assumption of normality on the independent variables. Independent variables can even be dummy ones or even non-random variables. Moreover, for each set of values of all the different independent variables, the distribution of the dependent variable must be normal. That is, each value y_i of the dependent variable Y, and not Y, is supposed to follow a normal distribution with a mean value that differs, and it is a linear function of the values of all the independent variables. The only way to test the normality assumption of a regression model with continuous covariates is through the residuals of that model, which have, by default, a mean value equal to 0.

Results

Initially, an exploratory factor analysis was first used on the 9 items of the Teacher Attitudes toward Change Scale. Bartlett’s Test of Sphericity showed that X² (36) = 1174.871, p < 0.01, indicating that the correlation matrix is not an identity matrix and, thus, a latent factor model would be meaningful and can be applied to the data. Moreover, the Kaiser-Meyer-Olkin Measure of Sampling Adequacy was calculated at 0.835, close to 1, indicating the presence of strong partial correlations. The three-factor solution explains 86.02% of the total initial variance of the items. Table II shows the loading of the items on the extracted factors after a varimax rotation with Kaiser normalization. All loadings are high, indicating that all the factors are well determined. Moreover, Cronbach’s alpha and ω are higher than 0.8, indicating a strong reliability index (Cortina, 1993). Omega indices were calculated through Mplus v6.12 (Hayes & Coutts, 2020). The factors were calculated as the average of the individual scores of the factor items.

Table II. Factor Loadings of the Selected Items from Teacher Attitudes toward Change Scale and Indices

	Loadings
Item	COG	AFF	BHV
I find most changes to be pleasing.	0.868
Most of my co-workers benefit from change.	0.891
Change often helps me perform better.	0.772
I don’t like change.		0.919
Change frustrates me.		0.968
Most changes at my school are irritating.		0.962
I look forward to changes at school.			0.727
I often suggest new approaches to things.			0.906
Change usually benefits the school.			0.636
Cronbach’s alpha	0.914	0.964	0.840
ω	0.915	0.965	0.840

Furthermore, a confirmatory factor analysis was used to test the fit of the 3-factor model. The results showed a very good fit of the model (Brown & Moore, 2012); X²/df = 56.116/24 = 2.338 < 5, CFI = 0.973, TLI = 0.959, SRMR = 0.045.

Teachers’ readiness is assessed through three items: “How would you rate distance learning you utilized in terms of 1. the hardware and software infrastructure that you had at your disposal, (2) your readiness to handle the medium you used (e.g., computer, tablet) and (3) your readiness to utilize various software.” The items are positively related to each other, with Spearman’s rho correlations equal to ρ₁₂ = 0.518, ρ₁₃ = 0.497 and ρ₂₃ = 0.753 (highly statistically significant coefficients, with all p-values less than 0.01). To avoid multicollinearity problems in case these items are simultaneously used in the set on the explanatory variables in the hierarchical regression below, an exploratory factor analysis approach is used, in which a single factor is constructed and replaces the three highly correlated items. The analysis showed that 76.14% of the total variance of the items is explained by a single factor, namely Readiness, which is formed with loadings 0.809, 0.915, and 0.891, respectively. The factor to be used in the regression analysis was calculated as the average of the individual respondent scores of the three items.

A hierarchical regression model on the perception of the development of “teacher expertise” with two blocks of explanatory variables is used. The first block consists of the demographic variables age in years, gender (dummy variable with 1 denoting female and male being the baseline category) and educational stage (dummy variable with 1 denoting secondary school teacher, and primary school teacher used as a baseline category). The second block of explanatory variables consists of COG, AFF and BHV as derived from the Teacher Attitudes toward Change scale, the resilience that is assessed through the single-item scale, readiness as it has resulted from the 3 items regarding teacher’s readiness, and last, the question that is related to the “learning outcomes for the students” as a perception of the teachers. Table III presents the descriptive statistics of the variables that are involved in the hierarchical linear model. All variables’ mean values are close to the centre of the scale used respectively, except for 1. the readiness with a high mean of 3.637 and an SD of 0.983 (on a 1–5 point scale), and 2. the perception about the development of “teacher expertise”, with a mean of 4.649 and an SD of 2.011 (on a 1–7 point scale), indicating that, more than one year post to the compulsory distance teaching, teachers evaluate that experience rather positively, in terms of what they have learned out of it. Moreover, all but two coefficients of the estimated skewness and kurtosis coefficients lie outside the interval (−1,1), indicating “close to normal” shaped distributions.

Table III. Descriptive Statistics of the Dependent and Explanatory Variables Used in the Linear Model

Variable	Mean	Standard deviation	Skewness	Kurtosis
COG	3.819	1.600	0.193	−0.539
AFF	3.020	1.648	0.562	−0.509
BHV	4.221	1.431	0.013	−0.375
Resilience	0.538	2.753	0.199	−0.630
Readiness	3.637	0.983	−0.885	0.428
Perception of learning outcomes	2.970	1.066	−0.092	−1.277
Perception of the development of teacher expertise	4.649	2.011	−0.504	−1.001

In specific, on the 7-point scale, 20.6% of the respondents chose the two lowest options of the scale; 10.7% responded 1 (very unlikely there is a development of “teacher expertise”), 9.9% responded 2 (quite unlikely), whereas, on the other side of the response scale, 18.3% responded 6 (quite likely there is a development of “teacher expertise”), and 22.9% responded the maximum 7 of the scale (very likely). Table IV shows the percentages of the dependent variable of the regression model.

Table IV. Percentages of the Responses on the Dependent Variable “Do You Think the Distance Teaching Experience You Had is a Useful Additional Instrument You are Going to Integrate into Teaching?”

Levels	Frequency	%	Cumulative %
1 Very unlikely	14	10.69	10.69
2 Quite unlikely	13	9.92	20.61
3 Somewhat unlikely	11	8.40	29.01
4 Neutral	13	9.92	38.93
5 Somewhat likely	26	19.85	58.78
6 Quite likely	24	18.32	77.10
7 Very likely	30	22.90	100.00

A stepwise variable selection on each block of the hierarchical regression model is used so as to conclude a parsimonious model consisting of statistically significant predictors. Results are given in Table V.

Table V. Results of Hierarchical Regression on the Perception of the Development of “Teacher Expertise”

Explanatory variable	Beta	t	p	Tolerance
A parsimonious model of the final step
Learning outcomes for the students	0.440	5.469	<0.01	0.989
COG	0.185	2.304	0.023	0.989
Excluded variables
Age in years	−0.063	−0.787	0.433	0.989
Female (compared to male)	0.045	0.542	0.589	0.935
Secondary school teacher (compared to primary)	−0.100	−1.249	0.214	0.999
AFF	0.061	0.703	0.484	0.862
BHV	−0.012	−0.110	0.913	0.511
Resilience	0.066	0.804	0.423	0.958
Readiness	0.029	0.334	0.739	0.879

The perception of the teachers regarding the learning outcomes of distance learning on the students is a highly statistically significant predictor of the perception of their own development of “teacher expertise” (p < 0.01). Moreover, cognitive responses to change (COG) is an additional statistically significant predictor for the perception of teachers about their development of “teacher expertise” (p < 0.05). This set results in a statistically significant model: F (2,118) = 19.124, p < 0.01, with a coefficient of determination that is R² = 24.5%. No other explanatory variables, from the ones that were included in the model, are found to be significant predictors for the perceptions of the teachers.

The tolerance indices show that there are no collinearity problems among the explanatory variables of the model; the lowest estimated index is 0.511 > 0.4. Although there is no strict cut-off for tolerance that indicates a serious collinearity problem, researchers suggest 0.4 or even 0.2 as a cut-off, under which collinearity is a cause for concern (see, for example, Weisburd & Britt, 2014).

The hypothesis of the normality of the linear model is tested through the standardized residuals. Shapiro-Wilk test of normality on the standardized residuals is calculated equal to 0.984, df = 131 and p = 0.127, indicating that the hypothesis of normality of the linear regression model is not rejected. Moreover, a normal P-P plot of the standardized residuals (Fig. 1) also suggests that the normality hypothesis of the linear model is not violated; the observed cumulative probability is close to the straight line of the graph that corresponds to the expected cumulative probability of the normal distribution.

Fig. 1. Normal P-P plot of standardized residuals of regression on the perception of the development of “teacher expertise”.

Despite the high tolerance indices that indicate an absence of strong correlations among the explanatory variables, weaker but significant correlations are explored through Spearman’s rho. Each cell of Table VI shows the correlation coefficient along with p in parenthesis. The elements below the diagonal correspond to the correlations of primary school teachers’ responses, and the elements above the diagonal provide the correlations of secondary school teachers’ responses.

Table VI. Spearman’s Rho Correlations among the Explanatory Variables, Primary School Teachers below the Diagonal, Secondary School Teachers above the Diagonal

Variable	1	2	3	4	5	6
1. COG		−0.53	0.69	0.06	0.16	0.29
		(<0.01)	(<0.01)	(0.71)	(0.29)	(0.5)
2. AFF	−0.27		−0.52	0.03	0.09	−0.14
	(0.01)		(<0.01)	(0.84)	(0.57)	(0.38)
3. BHV	0.64	−0.11		0.08	0.10	0.17
	(<0.01)	(0.34)		(0.62)	(0.53)	(0.27)
4. Resilience	0.11	−0.16	−0.08		0.17	−0.14
	(0.35)	(0.15)	(0.47)		(0.28)	(0.37)
5. Readiness	0.20	−0.13	0.33	0.03		0.34
	(0.07)	(0.24)	(<0.01)	(0.82)		(0.02)
6. Perception of learning outcomes	0.01	0.12	0.25	−0.16	0.31
	(0.94)	(0.27)	(0.03)	(0.15)	(<0.01)

Based on the estimated coefficients for primary school teachers, AFF has a negative and statistically significant correlation with COG, whereas COG and BHV have a positive, highly statistically significant correlation. Readiness has a positive, weakly statistically significant correlation with COG and a positive, highly statistically significant correlation with BHV that was not observed for secondary school teachers. Last, perception of learning outcomes has a positive statistically significant correlation with BHV and a positive highly statistically significant correlation with readiness.

For secondary school teachers, AFF has a negative and highly statistically significant correlation with COG, whereas COG and BHV have a positive, highly statistically significant correlation. In addition, AFF and BHV have a negative and highly statistically significant correlation that was not observed for primary school teachers. Moreover, perception of learning outcomes has a positive statistically significant correlation with readiness.

Conclusions

The main purpose of this study was to explore the influence of teachers’ perception of learning outcomes, resilience, readiness, and attitudes toward change on the perception of the teachers about the development of their own “teacher expertise”. Readiness regarding the use of software and hardware, which was required for distance learning to take place, was found to positively correlate with teachers’ perception of students’ learning outcomes. Furthermore, readiness was found to positively correlate cognitive and behavioural reactions to change for primary school teachers. Cognitive reaction to change refers to the beliefs about the need for change, the significance of the change, the extent to which the change will be beneficial and the knowledge required to handle change (Piderit, 2000) was, in turn, found to positively influence teachers’ perception about the development of “teacher expertise” due to the remote teaching during the pandemic.

It should be noted that, as with all non-probability sampling methods, results should be interpreted with caution. Despite the fact that it is a fast and inexpensive way to collect data, respondents who choose to participate may not be representative of the reference population. However, in cases where the population has similar traits, non-probability sampling may provide useful insights.

The results show the need for continuous training of teachers on the use of computers since their readiness to use them seemed to affect their general attitudes and perceptions regarding whether their distance teaching was effective and whether they themselves achieved experience benefits. In a world with ever-increasing use of technology, training teachers in computer use and technology literacy should be a necessity, not an option.