Post-Pandemic E-Learning Practices in Higher Education: A Cross-Sectional Survey of Students and Academic Staff Across Eight Institutions in Southeastern Poland

Abstract

The study examined e-learning practices reported by students and academic staff in the post-pandemic period and respondent characteristics associated with exploratory composite index scores. In October 2025, a cross-sectional survey included 1049 students and 189 academic staff members recruited through non-probability sampling at eight institutions in southeastern Poland. Each respondent-specific questionnaire included a ten-item block on e-learning practices, experiences, and barriers. Its first eight items formed separate exploratory composite indices for students and academic staff across the participating institutions. Analyses comprised descriptive statistics, Cronbach’s alpha, and two linear regression models with institution fixed effects and HC3 robust standard errors. Students most frequently endorsed items concerning online tests, synchronous classes, and improved time management. Academic staff most often reported online consultations and using communication platforms, but less frequently recording classes, creating LMS courses, or using specialized collaboration tools. Among students, reported awareness of training availability was positively associated with the index score (B = 0.458, p < 0.001). Among academic staff, positive associations were observed for reported training completion (B = 0.472, p < 0.001) and prior remote-teaching experience (B = 0.457, p = 0.004). These associations do not establish causality. The findings indicate a selective pattern of reported e-learning use.

1. Introduction

The COVID-19 pandemic led to the rapid transfer of a substantial proportion of higher education into digital environments. However, this shift should not automatically be equated with intentionally designed online education. Emergency remote teaching was a crisis-response measure primarily intended to ensure instructional continuity, often without sufficient time to design courses appropriately, prepare materials, or adapt assessment methods (Hodges et al., 2020; Watermeyer et al., 2021). At the same time, higher education institutions faced technical difficulties, varying levels of digital competence, challenges in maintaining student engagement, and the need to combine synchronous and asynchronous tools (Turnbull et al., 2021).

Following the most restrictive phases of the pandemic, the key question concerned not only the presence of technology but also the extent to which particular solutions were used in everyday teaching. A review of 68 studies found that models combining in-person and online instruction while maintaining organizational flexibility were considered particularly promising in the post-pandemic period (Imran et al., 2023).

Research conducted after universities returned to in-person teaching indicates that some pandemic-related changes persisted, including increased online learning activities and adaptations to assessment, although their continuation depended on institutional and individual conditions (Broadbent et al., 2023). A similar mechanism of selectively integrating technology into institutional practices was identified in research on the digital transformation of higher education institutions under crisis conditions (Vaz et al., 2026).

Blended learning is understood here as the planned combination of face-to-face teaching and online learning activities within a course or programme. The term hybrid learning is used in the same broad sense, rather than as a separately measured delivery mode (Hrastinski, 2019). In this study, e-learning refers broadly to communication platforms, learning management systems (LMSs), online tests, recordings, asynchronous materials, interactive tools, online consultations, and digital monitoring of student engagement. The questionnaires examined respondents’ reported practices and experiences; they did not classify courses or programmes as face-to-face, blended, or fully online. Consequently, reporting particular digital practices does not establish that a systematically integrated blended learning model was implemented.

An important criterion of online education quality is not merely the presence of technology but the type of interaction it enables. The classical model distinguishes student–content, student–instructor, and student–student interactions (Moore, 1989). The Community of Inquiry framework extends this perspective by incorporating cognitive, social, and teaching presence (Garrison et al., 2000). Research indicates that students particularly value regular contact with instructors, clear information, feedback, and appropriately designed collaborative tasks (Martin & Bolliger, 2018; Means & Neisler, 2021). From this perspective, a videoconference, discussion forum, or quiz may serve entirely different functions depending on how it is used.

Online education also increases the importance of self-regulation, time management, and independent organization of learning. The self-regulation of effort, learning activities, and interactions is associated with student achievement in online courses (Cho & Shen, 2013). Among academic staff, technology adoption depends on digital competence, prior experience, perceived usefulness of the tools, course characteristics, and institutional support (Damşa et al., 2021; Lee & Jung, 2021; Liu et al., 2020; Scherer et al., 2021). The DigCompEdu framework emphasizes that educators’ competence is not limited to the technical operation of software but also includes selecting resources, designing activities, conducting assessment, and supporting students (Ghomi & Redecker, 2019). Research on flexible learning design also highlights the importance of practical training combined with the collaborative development of specific teaching scenarios (Afonso et al., 2025).

In Polish higher education, academic staff have recognized both the functionality of e-learning and problems related to its usability, workload, and the support provided by their institutions (Ejdys & Szpilko, 2021). In studies involving students, flexibility and time savings were identified as advantages of remote education, whereas reduced direct contact, difficulties in conducting practical classes, and prolonged screen exposure were regarded as disadvantages (Stecuła & Wolniak, 2022). Thus, the extent to which a particular practice is used in the post-pandemic period may depend not only on its technical availability but also on the effort required to implement it, its pedagogical value, and users’ prior experiences.

Studies incorporating the perspectives of both students and academic staff indicate that these groups do not always evaluate interaction, feedback, course quality, and the role of technology in the same way (Maisano et al., 2024; Yan & Pourdavood, 2024). Academic staff may focus on organizational feasibility and control over the teaching process, whereas students more often emphasize flexibility, access to materials, and the timely provision of information. European research further indicates that the development of hybrid models requires not only appropriate infrastructure but also continuous pedagogical support and communities for sharing teaching practices (González Ruiz et al., 2025).

Few large, multi-institutional studies have described specific post-pandemic e-learning practices in Polish higher education while distinguishing the training-related characteristics reported by students and academic staff. The present study contributes a regional account of these practices and explores their associations with different respondent characteristics: students’ reported awareness of institutional training availability, academic staff members’ reported training completion, and staff members’ prior remote-teaching experience. Item-level analyses describe the reported practices, whereas the regression models examine associations with group-specific exploratory composite indices. The student analysis does not evaluate participation in training or its effectiveness.

The following research questions were formulated:

• Which post-pandemic e-learning practices are most and least frequently reported by students?
• Which post-pandemic e-learning practices are most and least frequently reported by academic staff?
• Is students’ reported awareness of training availability at their institution associated with their exploratory e-learning practice index scores?
• Is academic staff members’ reported completion of training in remote teaching associated with their exploratory e-learning practice index scores?
• Is academic staff members’ reported experience of remote teaching before the COVID-19 pandemic associated with their exploratory e-learning practice index scores?

For research questions 3–5, index scores refer to the group-specific exploratory composite indices: S-PD1–S-PD8 for students and N-PD1–N-PD8 for academic staff. These associations were examined in separate regression models for the two groups, with adjustment for the other predictors and institution fixed effects specified in Section 2.6.

2. Materials and Methods

2.1. Study Design and Setting

A cross-sectional quantitative study was conducted using a web-based survey questionnaire. Data were collected from students and academic staff between 2 and 31 October 2025. The study involved eight higher education institutions located in the Podkarpackie Voivodeship in southeastern Poland.

Both questionnaires were developed by the author for the broader research project and were not adaptations of a previously validated scale. Before the main study, the instruments did not undergo separate pilot testing or formal expert review. In the present article, the internal consistency of the exploratory composite indices constructed from selected items was assessed. These indices represent content-based summaries of reported e-learning practices and should not be treated as validated psychometric scales or as established measures of a single homogeneous construct.

A non-probability sampling strategy combining purposive and convenience sampling was used. Links to the two questionnaires, developed in Microsoft Forms, were provided to the authorities and relevant administrative units of the participating institutions, with a request that the appropriate link be distributed to students and academic staff through institutional communication channels. The author did not have direct access to the distribution lists or information on the exact number of individuals who received the invitation. Consequently, it was not possible to determine the denominator or calculate a response rate. The resulting sample should not be regarded as a random representation of the entire population of students and academic staff in Poland.

The data analyzed in this article were collected as part of a broader research project examining e-learning in higher education during the post-pandemic period, which also forms the basis of a research monograph currently in preparation. The broader project covered several distinct research areas, whereas the present study used only the complete PD1–PD10 item block concerning e-learning practices reported by students and academic staff. Separate research questions were formulated for this article, and a focused analytical strategy was adopted. This strategy included the construction of exploratory composite indices based on items PD1–PD8, assessment of their internal consistency, regression models with institution fixed effects and HC3 robust standard errors, and sensitivity analysis. The remaining areas covered by the broader research project were not included in the present article.

2.2. Data Cleaning and Participants

The original student dataset comprised 1057 responses. The completeness screen covered all 40 items in the four main questionnaire blocks, including the ten-item block analysed in this article. Eight records with at least one missing response in these blocks were excluded, leaving 1049 students. The original academic staff dataset comprised 211 responses. For this group, the completeness screen covered 50 items: ten retrospective items and 40 items in the four main questionnaire blocks. Twenty-two records with at least one missing response in these sections were excluded, leaving 189 academic staff members. These criteria were inherited from the broader project and were applied before selecting the PD1–PD10 block for the present analysis. Missing responses were not imputed.

The questionnaires did not include separate attention-check items. The main analysis also did not apply a minimum questionnaire completion-time threshold or automatically exclude records on the basis of uniform response patterns. The primary criterion for including a record in the analytical sample was the completeness of the main questionnaire blocks.

This procedure maintained the completeness criteria adopted for the broader research project. Importantly, the completeness criterion applied to the entire main questionnaire rather than solely to the item block analyzed in this article.

The student sample included 691 women (65.9%), 348 men (33.2%), and 10 respondents who selected another response option for gender (1.0%). A total of 453 students were in their first year of study, 248 in their second year, 115 in their third year, 101 in their fourth year, and 132 in their fifth year.

The student questionnaire recorded year of study using categories from 1 to 5 but did not separately record degree level or programme structure. It is therefore not possible to distinguish first-cycle, second-cycle, and integrated long-cycle programmes or to determine consistently how respondents interpreted the fourth and fifth years. Information on previous degrees, changes in institution, or breaks between study cycles was also unavailable. The reported year categories should therefore not be interpreted as a verified continuous progression from bachelor’s to master’s study.

Overall, 722 students (68.8%) answered affirmatively when asked whether their institution provided training in the use of e-learning platforms and remote-learning tools. This response was interpreted as reported awareness of institutional training provision, rather than independently verified availability. The questionnaire did not ask whether students had personally participated in or completed such training. Accordingly, this variable cannot be used as a measure of individual training participation.

The academic staff sample included 99 women (52.4%) and 90 men (47.6%). The largest age groups were 40–49 years (68 respondents) and 50–59 years (51 respondents). Completion of training in remote education was reported by 127 academic staff members (67.2%), while 59 respondents (31.2%) had taught remotely before the COVID-19 pandemic.

The exported form data contained technical identifying fields, including, for some records, respondents’ first and last names or email addresses. These fields were removed before the statistical analysis. Only research variables from which direct identifiers had been removed were analyzed.

2.3. Overview of the Analytical Procedure

The analysis followed four stages. First, the exported responses were screened using the questionnaire completeness criteria described above. Second, the ten practice-related items were examined separately through their response distributions. Third, the first eight items were averaged to construct a separate exploratory composite index for each respondent group, and internal consistency was assessed. Fourth, two separate regression models examined associations between these indices and the predictors specified in Section 2.6. Sensitivity analyses examined the treatment of “I do not know” responses and, in the student model, the coding of year of study and the inclusion of all reported gender categories. The student and academic staff samples were analysed independently.

2.4. Research Instruments and Response Coding

Both questionnaires used the following five-point response scale:

• 1—definitely not;
• 2—probably not;
• 3—I do not know;
• 4—probably yes;
• 5—definitely yes.

The response coded as 3 occupied the midpoint of the numerical coding scheme but did not represent a conventional neutral category. It indicated a lack of knowledge or an inability to provide a definite assessment of a given statement rather than an intermediate position between agreement and disagreement. In the descriptive analyses, the “I do not know” response was presented as a separate category and was not combined with either negative or positive responses. Means and standard deviations were therefore treated as supplementary information, while the interpretation of the findings was based primarily on the response distributions.

For the purposes of this article, the questionnaire items were assigned abbreviated codes:

• S—student;
• N—academic teacher (nauczyciel akademicki in Polish);
• PD—teaching practices (praktyki dydaktyczne in Polish);
• numbers 1–10—the item order in the original questionnaire block.

The questionnaires were administered in Polish. The English wording presented in Table 1 is a translation prepared for the purposes of this article.

Table 1. Questionnaire items analyzed in the present study.

Code	Student Item	Code	Academic Teacher Item
S-PD1	I most frequently participate in synchronous classes (Microsoft Teams, Zoom)	N-PD1	I regularly use Microsoft Teams or Zoom to teach classes
S-PD2	I use Moodle or similar platforms to access learning materials	N-PD2	I create courses on learning management systems (e.g., Moodle)
S-PD3	I complete online quizzes and tests provided by instructors	N-PD3	I use online quizzes and tests in my teaching
S-PD4	I receive recordings of online classes to review later	N-PD4	I record classes and make them available to students afterward
S-PD5	I participate in online discussions (forums, chats)	N-PD5	I use interactive tools (Padlet, Miro, Jamboard) during online classes
S-PD6	Online classes are interactive (questions, group work)	N-PD6	I monitor student activity and progress on e-learning platforms
S-PD7	Online group work is effective	N-PD7	I offer additional online consultations outside scheduled class time
S-PD8	Asynchronous materials are made available sufficiently promptly	N-PD8	I prepare multimedia materials specifically for online courses
S-PD9	E-learning allows me to manage my study time more effectively	N-PD9	Since the pandemic, I have reduced my delivery of asynchronous courses
S-PD10	Since the pandemic, I have used asynchronous courses less frequently	N-PD10	The main barrier to developing asynchronous courses is a lack of time, relevant competencies, or institutional support

The S-PD1–S-PD10 and N-PD1–N-PD10 codes were not included in the original questionnaires. They were introduced solely to simplify the presentation of the results. The item numbers indicate only their order in the respective questionnaire. Identical numbers in the student and academic staff versions do not imply equivalence of content or pairing of the items. Responses from the two groups were analyzed separately and were not combined into a single overall score.

2.5. Construction of Exploratory Composite Indices of E-Learning Practices

Two separate exploratory composite indices of e-learning practices were constructed for the analysis of associations. The student index was calculated as the mean response across items S-PD1–S-PD8, whereas the academic staff index was calculated as the mean response across items N-PD1–N-PD8. All items were assigned equal weight and followed the same coding direction: a higher index score indicated, on average, more positive responses across the PD1–PD8 item set. The indices were not interpreted as direct measures of the frequency or quality of e-learning practices or as measures of the intensity of a single homogeneous construct. Instead, they were treated as content-based aggregates of responses. No items were reverse-coded.

Items PD9 and PD10 were analyzed separately. They addressed perceived benefits, the reduced use or delivery of asynchronous courses, and barriers to their development and therefore differed conceptually from the items describing the occurrence of specific practices. Consequently, they were not included in the PD1–PD8 composite indices.

In the primary analysis, the original response coding from 1 to 5 was retained when calculating the exploratory composite indices, including the value of 3 assigned to the “I do not know” response. This approach was adopted to preserve the original response structure and maintain consistency with the coding used in the broader research project. However, it did not imply that the “I do not know” response was regarded as substantively neutral. The effect of this coding decision was examined in a separate sensitivity analysis in which these responses were treated as missing data. The indices for the two groups were analyzed separately and were not treated as equivalent versions of the same scale. Therefore, no direct test of the difference between the mean scores of students and academic staff was conducted.

2.6. Statistical Analysis

For each item, the mean, standard deviation, and percentages of negative responses (1–2), “I do not know” responses (3), and positive responses (4–5) were calculated. The “I do not know” response was presented as a separate category and was not included among either negative or positive responses. Given the ordinal nature of the scale and the distinct semantic meaning of category 3, means and standard deviations were treated as supplementary information, whereas the primary interpretation of the results was based on the response distributions for individual items.

Item-level response distributions formed the primary basis for interpreting the findings, while the composite indices were used as supplementary measures in the analysis of associations.

The internal consistency of the exploratory PD1–PD8 composite indices was described using Cronbach’s alpha. This coefficient was treated as a measure of response consistency rather than as evidence of the unidimensionality or psychometric validity of the indices.

Two separate linear regression models were subsequently estimated. The dependent variable was the group-specific exploratory PD1–PD8 composite index. The student model included reported awareness of the availability of training at the institution, gender, year of study, and institution fixed effects. The 10 respondents who selected another response option for gender were included in the descriptive analyses but excluded from the model, which used a binary comparison between women and men.

An additional sensitivity analysis included all 1049 students by treating the reported gender response as a categorical covariate with three categories: woman, man, and the questionnaire response labelled “Other”. Two indicator variables were included, with man as the reference category. This analysis retained the primary coding of year of study and the composite index, as well as the remaining predictors and institution fixed effects. The “Other” response was not interpreted as identifying a single homogeneous gender group.

The academic staff model included reported completion of training in remote education, remote teaching experience before the pandemic, gender, approximate age, and institution fixed effects. Age was approximated using the category midpoints of 25, 35, 45, 55, and 65 years. In each model, differences between the eight institutions were controlled for using seven indicator variables.

The original age categories were 20–29, 30–39, 40–49, 50–59, and ≥60 years.

Year of study was entered as a numeric covariate coded from 1 to 5 in the primary student model. Because degree level and educational trajectories were not recorded, the implications of this linear specification were examined in an additional sensitivity analysis. In that model, the reported years were represented by four indicator variables, with year 1 as the reference category, while the remaining predictors and the analytical sample were unchanged.

Binary predictors were coded 0/1, with “no” and “man” as the reference categories.

In the student model, the training-related variable represented respondents’ reported awareness of the availability of such training at their institution rather than verified training availability or their personal completion of training. In the academic staff model, the training predictor referred to respondents’ reported completion of training. Because institution fixed effects were included, the coefficients for these predictors represented associations between respondents within the same institutions. Institution fixed effects were included to account for average differences between the participating institutions. Their coefficients are reported as adjusted sample-level contrasts relative to the University of Rzeszów. Given the non-probability sampling and unequal institutional sample sizes, these contrasts were not interpreted as representative comparisons of institutional performance.

HC3 heteroskedasticity-consistent standard errors were used to account for potential heteroskedasticity (Long & Ervin, 2000). Unstandardized B coefficients, robust standard errors, 95% confidence intervals, p values, and coefficients of determination (R²) were reported. All tests were two-sided, and the significance level was set at α = 0.05.

Both regression models were fitted by ordinary least squares and included an intercept. Institutional affiliation was represented by seven indicator variables, with the University of Rzeszów as the reference institution. Two-sided p values and 95% confidence intervals were calculated using the t distribution with the residual degrees of freedom of the respective model.

Because the “I do not know” response differed semantically from a conventional neutral response, a sensitivity analysis was conducted. Responses coded as 3 were recoded as missing data, and index scores were calculated for respondents who had provided at least six substantive responses to the eight PD1–PD8 items. Respondents with fewer than six such responses were excluded from the corresponding model. This criterion required substantive responses for at least 75% of the items while retaining an adequate analytical sample. Both regression models were then re-estimated to determine whether the direction and statistical significance of the main associations depended on how the “I do not know” responses were coded. Missing values were not imputed.

The analyses were conducted in Python 3.12.13 using pandas 2.2.3, NumPy 2.3.5, and SciPy 1.17.0. The additional sensitivity analysis of year-of-study coding was conducted in Python 3.12.14 using the same package versions.

2.7. Ethical Considerations

Participation in the study was voluntary. Before beginning the survey, respondents were informed about the purpose of the study, the intended use of their responses, and their right to discontinue participation before submitting the form. Submission of the completed questionnaire was regarded as confirmation of informed consent to participate. Written organizational permission to distribute the questionnaires was obtained from the authorities of all participating higher education institutions.

The exported form data contained technical fields and, for some records, information that could directly identify respondents. All such fields were removed before the analysis. The analytical dataset therefore contained no direct identifiers, and the findings were reported exclusively in aggregate form. The Bioethics Committee of the University of Rzeszów issued an official statement dated 27 August 2026 (reference KB.UR.004.144.2026), confirming that the survey conducted between 2 and 31 October 2025 did not require the Committee’s opinion. According to the statement, the study did not constitute a medical experiment within the meaning of Chapter 4 of the Polish Act of 5 December 1996 on the Professions of Physician and Dentist. The statement documents the Committee’s determination regarding the requirement for review of the completed educational survey.

3. Results

3.1. Practices Reported by Students

The most frequently endorsed practice was completing online quizzes and tests, with 66.1% of students providing a positive response. A similar proportion indicated that e-learning helped them manage their study time more effectively (64.9%). Participation in synchronous classes was confirmed by 54.0% of respondents.

Students were considerably less likely to receive recordings of classes for later review, with positive responses accounting for 20.4%. The use of Moodle or a similar platform to access learning materials was reported by 28.1% of respondents. Participation in online discussions was reported by 37.8% of students.

For the item concerning the availability of asynchronous materials, half of the respondents selected the “I do not know” response. This may reflect either uncertainty about when the materials were published or limited exposure to asynchronous courses.

The distribution of the study sample by higher education institution and respondent group is presented in Table 2.

Table 2. Distribution of the study sample by higher education institution and respondent group.

Higher Education Institution	Students, n	Academic Teachers, n
University of Rzeszów	409	74
Rzeszow University of Technology	350	30
WSPiA University of Rzeszów	125	11
University of Information Technology and Management in Rzeszów	90	38
State University of Applied Sciences in Krosno	30	10
State Academy of Applied Sciences in Przemyśl	25	8
State Academy of Applied Sciences in Tarnobrzeg	10	12
Jan Grodek State University in Sanok	10	6
Total	1049	189

Detailed descriptive statistics and the distribution of student responses to individual e-learning practice items are presented in Table 3.

Table 3. Student responses to items S-PD1–S-PD10 (N = 1049).

Code	M	SD	Responses 1–2 (%)	Response 3 (%)	Responses 4–5 (%)
S-PD1	3.35	1.33	29.4	16.7	54.0
S-PD2	2.54	1.40	58.2	13.6	28.1
S-PD3	3.65	1.22	19.9	14.0	66.1
S-PD4	2.34	1.29	61.8	17.8	20.4
S-PD5	2.80	1.31	50.8	11.3	37.8
S-PD6	3.13	1.19	31.7	26.0	42.2
S-PD7	3.12	1.21	32.2	24.9	42.9
S-PD8	3.35	0.90	10.9	50.3	38.8
S-PD9	3.74	1.22	18.8	16.3	64.9
S-PD10	3.05	0.96	21.1	54.8	24.1

Note: Responses 1–2 represent “definitely not” or “probably not”, response 3 represents “I do not know”, and responses 4–5 represent “probably yes” or “definitely yes”.

3.2. Practices Reported by Academic Staff

The most frequently reported practice was offering additional online consultations outside scheduled classes. Such consultations were reported by 62.4% of academic staff members. Regular use of Microsoft Teams or Zoom was reported by 52.9% of respondents. The use of online tests was confirmed by 49.2%, while 48.7% reported preparing multimedia materials specifically for online courses.

The least frequently reported practice was the use of tools such as Padlet, Miro, or Jamboard, with 18.0% of academic staff members providing a positive response. Recording classes and making them available to students was reported by 24.3%, while 29.6% reported creating courses on LMS platforms.

A reduction in the delivery of asynchronous courses after the pandemic was reported by 45.5% of academic staff members, whereas 26.5% disagreed with this statement. A lack of time, relevant competencies, or institutional support was identified as the main barrier by 38.6% of respondents.

Detailed descriptive statistics and the distribution of academic staff responses to individual e-learning practice items are presented in Table 4.

Table 4. Academic staff responses to items N-PD1–N-PD10 (N = 189).

Code	M	SD	Responses 1–2 (%)	Response 3 (%)	Responses 4–5 (%)
N-PD1	3.13	1.55	46.6	0.5	52.9
N-PD2	2.35	1.46	69.3	1.1	29.6
N-PD3	2.95	1.45	48.1	2.6	49.2
N-PD4	2.19	1.44	72.0	3.7	24.3
N-PD5	2.01	1.27	76.7	5.3	18.0
N-PD6	2.89	1.55	49.7	3.2	47.1
N-PD7	3.42	1.58	36.5	1.1	62.4
N-PD8	2.95	1.56	48.1	3.2	48.7
N-PD9	3.29	1.25	26.5	28.0	45.5
N-PD10	3.05	1.11	28.0	33.3	38.6

Note: Responses 1–2 represent “definitely not” or “probably not”, response 3 represents “I do not know”, and responses 4–5 represent “probably yes” or “definitely yes”.

3.3. Characteristics of the Exploratory PD1–PD8 Composite Indices

The mean exploratory composite index score was 3.04 (SD = 0.76) among students and 2.74 (SD = 1.07) among academic staff. Cronbach’s alpha was α = 0.760 for the student index and α = 0.867 for the academic staff index. These values describe the internal consistency of the responses but do not provide evidence of the unidimensionality or psychometric validity of the indices.

These results should not be interpreted as a direct comparison of the extent of e-learning practices between the two groups. The student index provided an aggregate summary of responses concerning students’ experiences, whereas the academic staff index summarized the activities reported by academic staff members.

Descriptive statistics and Cronbach’s alpha coefficients for the exploratory composite indices in both respondent groups are presented in Table 5.

Table 5. Descriptive statistics and internal consistency of the exploratory PD1–PD8 composite indices.

Group	Items	N	M	SD	Cronbach’s α
Students	S-PD1–S-PD8	1049	3.04	0.76	0.760
Academic teachers	N-PD1–N-PD8	189	2.74	1.07	0.867

Note: The indices were calculated separately, and no direct test of between-group differences was conducted.

3.4. Factors Associated with Exploratory PD1–PD8 Composite Index Scores

The student model included 1039 respondents and explained 17.1% of the variance in the exploratory composite index (R² = 0.171; adjusted R² = 0.163). After accounting for gender, year of study, and institution fixed effects, students who reported awareness of the availability of training at their institution had index scores that were, on average, 0.458 points higher than those of students who did not report such awareness (B = 0.458; 95% CI [0.361, 0.554]; p < 0.001). This finding describes an association between reported awareness of training availability and the index score; it does not provide evidence that institutional training provision or participation in training affected students’ e-learning practices. Neither gender nor year of study reached the adopted level of statistical significance, with p values of approximately 0.056 in both cases.

The academic staff model included 189 respondents and explained 40.0% of the variance in the exploratory composite index (R² = 0.400; adjusted R² = 0.363). Academic staff members who reported completing training had index scores that were, on average, 0.472 points higher than those of respondents who did not report completing such training (B = 0.472; 95% CI [0.210, 0.734]; p < 0.001). Respondents with remote teaching experience before the pandemic had index scores that were, on average, 0.457 points higher (B = 0.457; 95% CI [0.149, 0.765]; p = 0.004). Both coefficients represent conditional associations and should not be interpreted as causal effects. Gender was not significantly associated with the index score. The association with age was small and negative (B = −0.014; p = 0.031), although age was only an approximation based on the midpoints of the age categories.

When all reported gender categories were included in the student model, the association between awareness of training availability and the exploratory index remained similar to the primary estimate (N = 1049; B = 0.461; HC3 SE = 0.049; 95% CI [0.365, 0.557]; p < 0.001). The year-of-study coefficient was small (B = 0.033; p = 0.041), compared with p = 0.056 in the primary model, indicating that the interpretation of this secondary association was sensitive to the model specification and analytical sample.

The results of the regression analyses for students and academic staff, including regression coefficients, HC3 robust standard errors, confidence intervals, and statistical significance levels, are presented in Table 6.

Table 6. Regression models of factors associated with the exploratory PD1–PD8 composite indices.

Group and Predictor	B	HC3 SE	95% CI	p
Students
Reported awareness of training availability at the institution	0.458	0.049	[0.361, 0.554]	<0.001
Woman (ref. man)	−0.093	0.048	[−0.188, 0.002]	0.056
Year of study	0.031	0.016	[−0.001, 0.062]	0.056
Rzeszow University of Technology	−0.103	0.056	[−0.213, 0.006]	0.065
WSPiA University of Rzeszów	0.111	0.071	[−0.028, 0.250]	0.116
University of Information Technology and Management in Rzeszów	0.709	0.083	[0.546, 0.872]	<0.001
State University of Applied Sciences in Krosno	0.040	0.148	[−0.251, 0.330]	0.789
State Academy of Applied Sciences in Przemyśl	0.190	0.124	[−0.053, 0.433]	0.125
State Academy of Applied Sciences in Tarnobrzeg	0.312	0.246	[−0.171, 0.794]	0.205
Jan Grodek State University in Sanok	0.199	0.426	[−0.638, 1.036]	0.641
Intercept	2.660	0.072	[2.518, 2.802]	<0.001
Academic Teachers
Reported completion of training	0.472	0.133	[0.210, 0.734]	<0.001
Remote teaching experience before the pandemic	0.457	0.156	[0.149, 0.765]	0.004
Woman (ref. man)	0.112	0.135	[−0.154, 0.378]	0.408
Approximate age (years)	−0.014	0.007	[−0.027, −0.001]	0.031
Rzeszow University of Technology	0.213	0.188	[−0.157, 0.583]	0.257
WSPiA University of Rzeszów	0.809	0.291	[0.234, 1.384]	0.006
University of Information Technology and Management in Rzeszów	1.284	0.190	[0.909, 1.658]	<0.001
State University of Applied Sciences in Krosno	0.252	0.380	[−0.498, 1.003]	0.507
State Academy of Applied Sciences in Przemyśl	0.609	0.382	[−0.144, 1.362]	0.112
State Academy of Applied Sciences in Tarnobrzeg	1.017	0.277	[0.470, 1.564]	<0.001
Jan Grodek State University in Sanok	1.521	0.231	[1.066, 1.976]	<0.001
Intercept	2.398	0.349	[1.709, 3.088]	<0.001

Note: Both models included seven institution indicator variables, with the University of Rzeszów as the reference institution. All institution coefficients and intercepts are reported. B denotes the unstandardized coefficient, and HC3 SE denotes the HC3 robust standard error. Institution coefficients represent adjusted contrasts between respondents in the analytical samples, rather than causal institutional effects or measures of institutional performance. Intercepts correspond to zero values of the uncentred numeric covariates and should not be interpreted as expected scores for typical participants. The student training-related variable represents reported awareness of training availability, whereas the academic staff variable represents reported training completion. The institutional contrasts are exploratory, and their p values are not adjusted for multiple comparisons.

In the sensitivity analysis, in which the “I do not know” response was treated as missing data, the direction and statistical significance of the main associations remained unchanged. Reported awareness of training availability remained positively associated with the student index (N = 779; B = 0.603; 95% CI [0.469, 0.737]; p < 0.001). Among academic staff, reported training completion (N = 185; B = 0.475; 95% CI [0.206, 0.745]; p < 0.001) and prior remote teaching experience (B = 0.471; 95% CI [0.156, 0.786]; p = 0.004) also remained significantly associated with the index. These findings indicate that the main conclusions drawn from the regression models did not depend on how the “I do not know” responses were coded.

In an additional sensitivity analysis treating reported year of study as a categorical variable, the association between awareness of training availability and the student index remained similar to that in the primary model (N = 1039; B = 0.461; HC3 SE = 0.049; 95% CI [0.364, 0.557]; p < 0.001). The gender coefficient was borderline in this specification (B = −0.096; p = 0.0498), compared with p = 0.056 in the primary model. Thus, the principal finding concerning awareness of training availability was stable, whereas the statistical significance of the secondary gender association was sensitive to the coding of year of study.

4. Discussion

4.1. Selective Profile of Post-Pandemic Digital Practices

The findings indicate clear variation in the frequency with which different forms of e-learning were reported during the post-pandemic period. The most frequently reported solutions were those that could be relatively easily integrated into conventional teaching, including videoconferencing, online tests, and online consultations. Activities requiring systematic course design, the production of recordings, or the use of additional interactive tools were reported less frequently.

The difference between the reporting of online quizzes and tests (66.1%) and the use of Moodle or similar platforms to access learning materials (28.1%) also requires clarification. These items addressed different activities: completing assessments and accessing teaching materials. The quiz item did not specify the application or access route, so the data cannot establish how the assessments were delivered or whether they were hosted within an LMS. Standalone assessment applications or links distributed through other communication channels are possible explanations, but these routes were not recorded in the questionnaire.

International evidence indicates both shared concerns and variation in post-pandemic digital practices. At an Australian university, Broadbent et al. (2023) found increased use of prerecorded videos, self-paced content, and online assessment compared with the same courses in 2019. This contrasts with the relatively limited reporting of recordings and LMS-based access to materials in the present student sample. In the United States, qualitative research by Yan and Pourdavood (2024) identified limited social interaction as a concern among students and faculty, a theme relevant to the interaction-related findings reported here. At three Spanish universities, González Ruiz et al. (2025) found that faculty valued hybrid teaching for widening access while calling for stronger dissemination of teaching resources and continued institutional support. Differences in study periods, sampling, instruments, and research designs prevent direct numerical comparisons or conclusions about national performance.

These studies nevertheless show why reported digital practices should be interpreted in their local teaching and institutional contexts. The combination of online and in-person elements can take many forms; therefore, the presence of particular tools does not itself establish a systematically designed blended learning model (Hrastinski, 2019; Müller & Mildenberger, 2021).

The institution coefficients also indicate contextual variation within the analytical samples. Relative to the University of Rzeszów, the student coefficient for the University of Information Technology and Management in Rzeszów was positive (B = 0.709; p < 0.001). In the academic staff model, positive contrasts were observed for that institution, WSPiA University of Rzeszów, the State Academy of Applied Sciences in Tarnobrzeg, and Jan Grodek State University in Sanok. These exploratory contrasts cannot identify the institutional mechanisms responsible for the differences. Attendance requirements, course delivery modes, assessment arrangements, and mandatory versus voluntary use of digital tools could be relevant, but these specific characteristics were not systematically measured. Interpretation is particularly cautious for small institutional samples, including the six academic staff respondents from Sanok.

4.2. Interaction and Contact with Instructors

Additional online consultations were the most frequently reported practice among academic staff. This may suggest that, during the study period, online communication primarily served as a supplement to in-person contact. At the same time, only 37.8% of students reported participating in online discussions, and fewer than half considered online classes to be interactive or online group work to be effective.

These findings are relevant in light of the distinction among student–instructor, student–student, and student–content interactions (Moore, 1989). The presence of an online communication channel with an instructor does not automatically foster collaboration among students. Research indicates that contact with instructors is particularly valued; however, a high-quality online environment also requires carefully designed activities that engage students and foster social presence (Garrison et al., 2000; Martin & Bolliger, 2018).

The limited use of interactive tools should not be interpreted as a complete absence of interaction. Academic staff may have used functions available directly within Microsoft Teams, Zoom, or LMS platforms. Nevertheless, the findings indicate that specialized collaboration applications had not become widely used solutions.

4.3. Flexibility and the Limited Asynchronous Component

Almost two-thirds of students endorsed the statement that e-learning helped them manage their study time more effectively (64.9%). This finding concerns a perceived time-management benefit rather than a direct assessment of flexibility in course design. The relatively limited reporting of recordings and LMS-based access to materials means that the benefit cannot be attributed to extensive asynchronous provision. Reduced travel, easier communication, or digitally completed tasks are possible explanations, but their contributions were not measured. The findings therefore support a distinction between students’ perceived organisational convenience and the availability of resources that permit learning independently of scheduled classes.

The finding concerning time management is consistent with studies in which flexibility was identified as one of the main advantages of e-learning, whereas reduced social contact and difficulties in conducting certain types of classes were regarded as its principal disadvantages (Rizun & Strzelecki, 2020; Stecuła & Wolniak, 2022). At the same time, 45.5% of academic staff reported reducing their delivery of asynchronous courses after the pandemic, while 38.6% identified a lack of time, relevant competencies, or institutional support as the main barrier.

Developing a valuable asynchronous course requires a greater initial investment of time and effort than simply conducting a class by videoconference. It requires the design of the course structure, materials, activities, communication processes, and feedback. Previous research, including a study of Polish academic staff, has similarly highlighted the importance of ease of use, functionality, and institutional conditions for implementing e-learning (Ejdys & Szpilko, 2021; Kaqinari, 2023).

4.4. Role of Training and Prior Experience

The most consistent findings from the regression models were the positive associations between the training-related variables and the exploratory composite index scores. Among students, the predictor represented reported awareness of the availability of training at their institution rather than personal participation in training. Among academic staff, the variable referred to reported completion of training. The student finding therefore concerns reported awareness of institutional provision and should not be interpreted as evidence about the association between personal training participation and e-learning practices.

Although 722 students (68.8%) reported that their institution provided training, 327 (31.2%) did not report such provision. These negative responses may reflect limited awareness, an absence of training opportunities, or both; the questionnaire cannot distinguish these explanations. The item concerned training in the operation of e-learning platforms and remote-learning tools. It did not establish whether students received preparation in how to learn online, such as planning independent study, managing time, participating in online collaboration, or using feedback. The present findings therefore leave an important question about student support unresolved. Future research should examine the content and accessibility of student training, actual participation, and whether available support addresses learning strategies as well as technical operation.

These findings do not establish that training caused higher PD1–PD8 composite index scores. Unmeasured respondent characteristics may have contributed to the observed associations: students who used e-learning more extensively may have been more aware of the available training provision, while academic staff with a stronger interest in digital pedagogy may have been both more likely to complete training and more likely to use e-learning tools. Nevertheless, these findings are consistent with the concept of educators’ digital competence, which encompasses not only technical skills but also course design and assessment (Ghomi & Redecker, 2019). Research on technology adoption indicates that the development of digital practices depends on a combination of individual competence and organizational support (Lee & Jung, 2021; Liu et al., 2020). Training based on developing a specific teaching scenario, rather than merely demonstrating software functions, may be particularly valuable (Afonso et al., 2025; Philipsen et al., 2019).

The association with pre-pandemic experience suggests that academic staff who had previously used remote teaching formats may have been better able to apply their acquired skills during the post-pandemic period. Prior experience may reduce the effort required to resume using a tool and allow its usefulness to be assessed more effectively. The association with age was small and should not lead to stereotypical assumptions that digital competence is confined to younger staff members. Opportunities to gain experience and access to appropriate support appear to be more important.

4.5. Implications for Higher Education Institutions

The findings identify issues that participating institutions may examine through local needs assessments. They do not establish the effectiveness of particular training formats, course standards, or support interventions.

For students, the association between reported awareness of training availability and the exploratory index suggests that communication about existing provision warrants further investigation alongside actual participation. Institutions could examine whether students know which training opportunities are available and whether these opportunities address their needs. The present study does not demonstrate that increasing awareness would change students’ e-learning practices.

For academic staff, the associations with training completion and prior experience provide a rationale for examining how professional development relates to specific teaching activities. However, the data do not establish which training content, duration, or delivery format would be most effective.

The relatively limited reporting of recordings, LMS course creation, and specialised collaboration tools also identifies areas for further contextual investigation. Local assessments should distinguish whether these patterns reflect course requirements, pedagogical choices, or unmet support needs. In particular, the questionnaire item combining time, competence, and institutional support does not establish which of these potential barriers predominates.

4.6. Study Limitations

The findings should be interpreted in light of several limitations. First, participants were recruited using a non-probability sampling strategy, and the study was conducted at eight higher education institutions located in a single region of Poland. The numbers of respondents from individual institutions varied considerably and were particularly small for some of the smaller institutions.

The primary student regression was restricted to respondents selecting woman or man and therefore excluded ten respondents who selected the “Other” response. Although the additional sensitivity analysis retained these respondents, their small number and the broad wording of this category limit conclusions about gender diversity. The findings should not be interpreted as representing the experiences of all gender groups.

The institution fixed effects accounted for average differences between the participating institutions. Their coefficients describe adjusted contrasts within the selected samples and should not be interpreted as representative comparisons or rankings of institutional performance. Specific contextual characteristics, including attendance requirements, course delivery modes, assessment arrangements, and mandatory versus voluntary use of digital tools, were not systematically measured. Consequently, the institutional contrasts cannot be attributed to these mechanisms.

Because the exact number of individuals who received the invitation was unknown, it was not possible to calculate a response rate or fully assess the risk of self-selection and nonresponse bias. The findings therefore cannot be regarded as representative of the entire population of students and academic staff in Poland. They nevertheless characterize the e-learning practices reported by respondents from the participating institutions in southeastern Poland. HC3 standard errors provided robustness to heteroskedasticity but did not account for possible within-institution dependence among responses. Because the study included only eight institutions, standard errors clustered at the institutional level were not used.

Second, the student and academic staff groups constituted two independent samples. Respondents were not matched according to the courses they attended or taught; therefore, student responses could not be directly linked to those of specific academic staff members. Although the items administered to the two groups addressed similar areas of e-learning practice, they were not parallel versions of the same scale. Consequently, direct tests of differences between students and academic staff were not conducted, and the findings for the two groups were interpreted in parallel.

Third, the study used author-developed questionnaires created for the broader diagnostic project. The instruments did not undergo separate pilot testing or formal expert review before the study and had not previously been standardized. The obtained Cronbach’s alpha values described the internal consistency of responses across the PD1–PD8 items but did not establish the unidimensionality or psychometric validity of the resulting indices. The indices should therefore be treated as content-based summaries of reported practices rather than as validated psychometric scales. Items PD9 and PD10 had different substantive meanings and were therefore analyzed separately rather than included in the composite indices. The academic staff item concerning barriers combined lack of time, relevant competencies, and institutional support in a single statement. Its responses therefore do not distinguish the separate contributions of these potential barriers. The construction of the instruments and the structure of the resulting indices require confirmation in future research.

Fourth, the study relied on self-reported data. Responses were not compared with LMS logs, platform activity records, or observations of actual teaching practices. The findings may therefore have been affected by recall bias, subjective interpretations of individual statements, and social desirability. The questionnaires did not directly assess the quality, comprehensiveness, or pedagogical effectiveness of digital teaching. Reported use of particular tools or activities should therefore not be interpreted as evidence of a systematically designed or effective digital learning environment. The questionnaires did not include separate attention-check items, and the primary analysis did not apply a minimum form-completion time criterion. This limits the possibility of fully assessing the quality of respondents’ engagement when completing the questionnaire.

Fifth, response category 3 represented “I do not know” rather than a conventional neutral response. High percentages of respondents selecting this category may indicate a lack of knowledge, limited experience with a particular form of e-learning, or a mismatch between a specific item and the circumstances of some participants. Means and standard deviations were therefore treated as supplementary information, while the primary interpretation was based on separately reported distributions of negative, “I do not know”, and positive responses. In the sensitivity analysis, “I do not know” responses were treated as missing data. The direction and statistical significance of the key associations remained unchanged, indicating that the main conclusions from the regression models did not depend on how this response category was coded.

Sixth, the cross-sectional design does not identify causal effects. A causal contribution of training or earlier remote-teaching experience to subsequent practices is plausible, but the observed associations alone cannot establish it. Training completion was not accompanied by information on its timing, content, or respondents’ practices before participation. Students already engaged in e-learning may have been more aware of available training, while staff interested in digital pedagogy may have been more likely both to complete training and to use digital tools. Pre-pandemic teaching experience precedes the current survey, but temporal precedence does not remove possible confounding by other respondent characteristics. Institution fixed effects and the included covariates do not address all these influences. Longitudinal measurements or evaluations of training interventions would be needed to distinguish causal contributions from selection and other explanations.

Seventh, the completeness criterion was applied to the entire main questionnaire rather than solely to the PD1–PD10 block analyzed in this article. Consequently, some excluded forms may have contained complete responses concerning e-learning practices but had missing responses in other sections of the broader study. Although retaining the criteria established during the preparation of the broader project ensured consistency within the dataset, excluding incomplete records may have affected the sample structure to some extent, particularly in the academic staff group.

Eighth, degree level, programme structure, and previous educational trajectories were not recorded. Students selecting the same year category may therefore have represented different stages of first-cycle, second-cycle, or integrated programmes. The additional analysis using categorical year indicators examined sensitivity to the linear specification but could not resolve this ambiguity or control for differences in degree level.

Ninth, the student questionnaire recorded reported awareness of institutional training provision but not personal participation or completion. The association observed in the student model therefore cannot be interpreted as evidence about training participation or effectiveness. Future studies should measure awareness, access, participation, and completion separately and collect information on the timing and content of training.

Future research should include institutions from different regions of Poland, employ random or stratified sampling methods, and subject the instruments to more comprehensive validation. Survey data could also be combined with actual indicators of LMS use, assessments of course quality, and interviews with students, academic staff, and individuals responsible for institutional e-learning support. Longitudinal studies or evaluations of specific training programs would make it possible to determine more precisely which digital solutions persist in higher education and which factors support this process.

5. Conclusions

This study contributes a multi-institutional account of post-pandemic e-learning practices reported by students and academic staff in southeastern Poland. Its principal contribution is a differentiated description of digital teaching and learning practices, showing that their reported use forms a selective pattern across activities.

The analysis also distinguishes students’ awareness of institutional training provision from academic staff members’ training completion and prior remote-teaching experience. Examining these characteristics separately provides a more precise basis for further research on institutional support and engagement with e-learning. The findings identify questions for subsequent investigation of training participation, course organisation, and the pedagogical quality of digital practices.