1 Introduction

The 2022 World MOOC and Online Education Conference was held online from December 8th to 9th under the theme of "Digital Education Leads the Future." The Chinese Ministry of Education reviewed the achievements of China's MOOC and online education development and proposed initiatives in four areas: accelerating resource opening and sharing, deepening technological applications, improving standards and specifications, and promoting innovative and healthy development of online education while building a global education community. Over the past two decades, online education has offered advantages such as breaking through time and space limitations, flexible learning modes, supporting synchronous or asynchronous teaching, reviewing learning content, and tracing the learning process. High-quality resources like MOOCs and SPOCs provide new opportunities to promote student learning and will gradually become a beneficial supplement to classroom teaching in universities. The vigorous development of new online teaching models will continue to promote the acceleration of online teaching development in universities worldwide. According to the fourth edition of "The Changing Landscape of Online Education" (CHLOE) report jointly released by the online learning quality assurance organization Quality Matters (QM) and Edu Ventures Research, since 2019, fully online learning and blended learning have become the mainstream form of online courses in American universities. In early 2020, due to the impact of the COVID-19 pandemic, Chinese universities fully utilized various high-quality online resources and online learning platforms to actively carry out online teaching activities, in accordance with the requirements of "no suspension of teaching and learning." Students conducted large-scale, long-term, and restrictive online learning under the organization of their schools, which gave them a preliminary understanding and cognition of online learning, and accumulated certain online learning experience. This creates opportunities for the reform of classroom teaching models in Chinese universities during the digital transformation period, providing a foundation and conditions for the implementation of online teaching and online-offline hybrid teaching models.

At the technological level, online education provides some extent promoting educational equity and narrowing the education gap at the technological level. However, from the perspective of the learning process, online education places learners in a virtual environment, creating a ubiquitous learning field that transcends time and space. This can easily cause students to lose focus and have poor learning adaptation. Despite the potential benefits, there are also continuous doubts about online education, including low social recognition and poor teaching satisfaction. While many people in China view large-scale online teaching as a "temporary mode" under the background of sudden epidemics, online education has accelerated the transformation of global higher education. Factors in the online learning process, including the external environment, teaching quality, and learner performance, have become widely discussed topics in society. It must be acknowledged that the "space separation" of teaching and learning in online education has weakened the objective requirements of teachers' teaching to some extent and has placed higher demands on students' learning adaptation. Similarly, many countries, such as the United States and Australia, face various problems in online education, such as the digital divide, poor student autonomy, insufficient teacher information literacy, low teaching efficiency, and poor student adaptability.

Compared to traditional learning environments, online settings demand a higher degree of student autonomy and a more pronounced deployment of self-regulated learning strategies (Broadbent, 2017; Kizilcec et al., 2017). Moreover, the high-pressure and uncertain circumstances during the COVID-19 pandemic necessitate students to mobilize heightened personal resources to effectively adapt to the new learning conditions. Students must strive to sustain their motivation and engagement while identifying methods to facilitate their adjustment to the novel learning environment (Besser et al., 2020). For instance, Besser et al. (2020) observed that during the COVID-19 online learning period, students reported elevated levels of negative emotions, stress, and feelings of isolation, alongside reduced attention, motivation, and comprehension of the materials, compared to their prior face-to-face learning experiences.Adaptability represents a self-regulatory construct encompassing adjustments in cognition (e.g., devising new solutions), behavior (e.g., approaching tasks differently), and emotion (e.g., effectively managing negative emotions) in new, uncertain, and ever-changing situations (Martin et al., 2012, 2013). Unlike resilience, coping, or perseverance, adaptability goes beyond handling setbacks and adversity, focusing more on overcoming novelty and uncertainty in the environment (Martin et al., 2012, 2013). Specifically, amid the pandemic, students must adapt to numerous novelties and changes, such as altered daily routines (e.g., staying at home due to lockdowns), modified learning modes (e.g., hybrid or online learning), social distancing norms, and uncertainties regarding school reopenings or parental employment status. On an individual level, continually devising new approaches to address these changes, regulating negative emotions, and adjusting learning strategies are pivotal for successful learning (Waters et al., 2021). In essence, adaptability can predict students' academic engagement and achievement during this period (Martin et al., 2012, 2013). Similarly, adaptability (as a personal resource) exhibits a significant correlation with higher levels of online self-efficacy, facilitating mastery and efficacy experiences in online learning (Collie & Martin, 2017). Thus, adaptability stands as a critical factor influencing how students navigate online learning during times characterized by substantial novelty, variability, and uncertainty (Martin et al., 2012, 2013). The adjustments required of students within these uncertain environments are realized through the malleable psychological attributes of adaptability (Martin et al., 2013), representing a viable direction for enhancing students' online learning experiences. Whether in online or offline learning contexts, adaptability exerts a robust positive influence on students' developmental progress.

Existing research has identified individual characteristics, family background, and school education features as pivotal factors influencing university students' learning adaptability. Firstly, in terms of individual characteristics, factors such as gender, household registration, self-awareness, and values impact university students' adaptability to learning. Secondly, school-related factors, including learning resources (such as majors, courses, learning activities), teaching methods, instructional approaches by teachers, learning environment (Feng et al., 2006), institutional management, and campus culture (Liu, 2015), influence students' learning adaptability. For instance, Zhang et al. (2017a, b) pointed out that learning platforms and curriculum design significantly positively affect adaptability. Thirdly, interpersonal interactions and dynamics, such as teacher-student roles and school social relationships, affect university students' adaptability to learning. In summary, learning adaptability is influenced by the confluence of various factors. Current research often delves into the influences on traditional learning adaptability from perspectives like the learner, teacher, and environment. These studies explore the internal and external factors affecting adaptability using multi-angled or localized approaches. In contrast to traditional offline classroom learning, the integration of digital technology makes factors such as users' adoption of information technology significant in influencing the learning process. Faced with new learning modalities and heightened requirements, there is a lack of research exploring the combined influences of new and existing factors on individuals' adaptability to online learning. Furthermore, little is known about how to enhance individual adaptability to online learning in this new paradigm. This study's focal point centers on these aspects.

Simultaneously, driven by the robust engine of higher education's digital transformation, the "emergency remote teaching" mode during the pandemic is gradually transitioning into a normalized "blended online teaching" mode. In contrast to the wealth of learning outcome data, the potential of process data that can be extracted from online learning experiences remains to be further explored. Against this backdrop, this study integrates Task-Technology Fit Theory with the Technology Acceptance Model (TAM) (Davis, 1989) to explain individuals' technological cognition, technology adoption behavior, and technological adoption performance. Building upon previous findings and incorporating relevant variable measurements, adaptability is positioned as a crucial outcome variable and quality criterion. Consequently, a comprehensive model of key factors influencing university students' online learning process is constructed, considering environmental support, teaching satisfaction, and perceived ease of use as pivotal foundational elements, with online learning task-technology fit being a critical juncture or process element. The synergy between process and foundational elements contributes to learning outcomes. Learners must adapt their learning experience approaches based on changes in the environment and process, given that these external factors influence university students' learning process. The study explores the impact of three key foundational variables on task-technology fit, thereby further investigating the influence of task-technology fit on learning adaptability. The ultimate aim is to holistically present the entire process and outcomes of online learning, elucidating its causal pathways and mechanisms. The study's ultimate objective is to inform policy making pertaining to online learning, providing empirical evidence for enhancing teacher instruction, fostering online technological capabilities, boosting students' adaptability, and improving academic performance. This endeavor ultimately benefits students.

2 Theoretical basis and research hypotheses

2.1 Theoretical basis

The sustained usage intention of college students towards online learning essentially belongs to the research scope of user adoption of information technology. Davis's Technology Acceptance Model (TAM) has long been considered the theoretical paradigm for studying user adoption of information technology. In 1985, Davis (1985) first proposed the concept model of TAM. He believed that the use of a system is a response that can be explained or predicted by user motivation, which is directly influenced by the characteristics of the system and individual capabilities. However, scholars have also pointed out limitations in using the unadjusted TAM to explain user choices in complex environments. For instance, Wen et al. (2004) argued that the explanatory power of the TAM diminishes significantly when users experience resistance towards a product, necessitating continuous adjustments based on the research subjects. Additionally, Farahat (2012) found that besides perceived usefulness and perceived ease of use, students' attitudes towards online learning and social influence also play crucial roles in determining their willingness to engage in online learning. Moreover, research by Bazelais et al. (2018) emphasized the role of external influencing variables. Based on the concept model, Davis (1989) further proposed that user motivation beliefs in technology acceptance can be explained by four factors: perceived ease of use, perceived usefulness, attitudes towards use, and usage intention. The user's usage intention of the system is the main factor that determines whether the user actually uses or rejects the system, and it is also influenced by perceived usefulness and attitudes towards use, which are in turn influenced by perceived ease of use.

However, the Technology Acceptance Model (TAM) is difficult to answer three questions: first, users' application of a certain technology system is actually in a specific task context, but information technology cannot meet all of users' task needs; second, users may perceive a technology as useful or easy to use but may not necessarily use it; third, the generation of user usage behavior is not always voluntary. To better explain these issues, Goodhue and Thompson (1995) proposed the Task-Technology Fit (TTF) theory. They believe that users may perceive a technology as useful and easy to use, but if they think the technology is not suitable for their tasks and cannot improve their performance, they will not adopt the technology. In other words, users' perception of technology usage will change with the task. When the use of the system can improve work performance, some potential users will use the information system, but the behavior is not necessarily voluntary. Staples and Seddon (2004) focused on the forced use of library central catalog systems by librarians and the voluntary use of word processing and spreadsheet software for coursework and personal tasks by students. They discovered that in both contexts, task-technology fit significantly influences usage performance and attitude beliefs. However, in the context of forced use, attitude beliefs fail to effectively predict usage behavior. From the perspective of variable relationships, the TTF theory believes that user behavior is influenced by usage performance, and usage performance is determined by task-technology fit, while the degree of fit is jointly determined by the characteristics of information technology and task requirements. In addition, personal characteristics such as training experience, computer use experience, and motivation also play a significant moderating role (Goodhue, 1995). However, the final empirical results show that the direct impact of task-technology fit on usage behavior is weak.

Dishaw and Strong (1999) pointed out that one important reason for this is that TTF ignores the impact of fit on users' psychology, and thus fails to fully construct the internal influence mechanism of task-technology fit on individual behavior. TAM (Technology Acceptance Model) primarily focuses on the formation of attitude beliefs based on individual perceptions of technological attributes, while Task-Technology Fit Theory emphasizes the alignment between task requirements and technological capabilities. By integrating these two theories, a stronger explanatory power could potentially be achieved for usage behavior. To verify this hypothesis, Dishaw and Strong (1998) were pioneers in attempting such integration within a work environment. They constructed an integrated model by combining viewpoints from Task-Technology Fit Theory and TAM. After surveying implementers of 60 software maintenance projects, they found that the explanatory power of the integrated model, Task-Technology Fit Theory, and TAM for usage behavior were 51%, 41%, and 36%, respectively. This finding demonstrated that the integrated model's explanatory power surpassed that of any individual model. They believe that TAM and TTF respectively explain the influencing factors and mechanisms of user adoption of information technology and the generation of usage performance from different perspectives, and both models have good explanatory power for user usage behavior. If the two models are combined organically, their shortcomings can be compensated. Similarly, in a workplace context, Pagani (2006) constructed a model for predicting enterprise employees' adoption of high-speed wireless data services by integrating Task-Technology Fit Theory and TAM. The research findings also supported the notion that the integrated model offered better predictions for user behavior. Therefore, it is necessary to incorporate TTF as an important external factor based on TAM to systematically analyze its impact on perceived usefulness, perceived ease of use, attitudes towards use, and usage intention, and then elucidate its effect on actual usage behavior.

2.2 Research model and hypotheses

2.2.1 Introduction

As the main learning mode for college students during the COVID-19 pandemic, online learning is influenced by numerous factors. This study attempts to summarize the key elements of college students' online learning process to establish a solid theoretical foundation. Based on investigations and analysis of online education in various universities, this study defines "online learning" as college students' online classroom learning and learning engagement based on various online platforms, smart terminals, and digital resources. Previous literature searches found that research on online learning processes is mainly focused on environmental support, teaching interaction, learning engagement, and learning outcomes. Scholars have discussed college students' learning processes from different perspectives. Shi (2015) found that a considerable portion of college students spent three years from their freshman to junior year adapting from high school to university. Zhao et al. (2019) studied college students using online learning spaces at a university and found that learners' information literacy, teacher support, and learning environment are key factors affecting their adaptation to online learning. Learning adaptation is the foundation and prerequisite for achieving better learning outcomes. It is the process by which learners rely on the current learning environment to complete relevant learning tasks based on their own learning needs and continuously adjust their mind and body to seek balance with the learning environment (Zhao & Xu, 2015). Currently, research on learning adaptation has made significant progress both domestically and internationally. Foreign researchers mainly focus on exploring single or certain influencing factors of learning adaptation. A study investigates the learner and instructional factors that influence learning outcomes in a blended learning environment. Drawing on past research, the authors propose that learner factors include learning style, motivation, attitude, and strategy, while instructional factors include instructional design, strategy, and technology. The article also explores how these factors interact and how instructional practices can be optimized in a blended learning environment to enhance learning outcomes (Lim & Morris, 2009). Larose and Roy (1995) designed the College Adaptation Test to measure students' personal learning tendencies and demonstrated the factors influencing learning adaptation from three aspects: learning beliefs, learning behavior, and learning emotions. Simon et al. (2007) explored the influence of student characteristics (such as motivation, self-efficacy, and communication skills) on learning adaptation. Ruan and Deng (2014) discussed the problems, influencing factors, and significance of learning adaptation in new graduate students' online learning and proposed improvement strategies.

In contrast, more scholars in China focus on the study of multiple influential factors, regarding student characteristics, teachers, information technology, and school support as key factors affecting adaptive learning in online education. Jiang (2005) constructed a "three-stage, four-module" relationship model based on foreign dropout theories, after analyzing the influential factors of adaptive learning in depth. Some scholars propose that media richness, which refers to the amount of information and feedback available through a communication channel, and flow, which is the state of deep engagement and absorption in an activity, are important factors that influence individuals' acceptance of e-learning technology (Liu et al., 2009). Deng et al. (2012) survey research shows that learners, learning resources, learning environments, and teachers all have an impact on the self-directed learning adaptability of college English learners in online environments. Zhang et al. (2017) integrated eight influential factors including self-directed learning, learning support, learning environment, and course design based on MOOCs. Qin et al. (2018) analyzed the influential factors of learning adaptability from three aspects: learning initiative, information literacy, and cognitive prerequisites, with flipped classroom as the starting point. Zhao et al. (2019)categorized the influential factors of learning adaptability in online environments into four aspects: learners, teachers, courses, and environments, and believed that the key to learners' adaptive learning in online learning spaces lies in the learners themselves.

In summary, scholars have conducted research on online teaching during the pandemic from different perspectives, focusing on online learning quality, satisfaction, platform usage, and challenges faced by teachers and students. However, there is still a lack of in-depth research on the online learning process of Chinese university students, including learning engagement, teaching feedback behavior, environment support, and learning adaptability. Nonetheless, researchers are increasingly focusing on the cultivation process and learning adaptability, which constitute key observation dimensions of the online learning process for college students. However, there is no overall consensus on the key elements of the online learning process. Therefore, this study takes a results-oriented approach, using learning adaptability as a variable to measure the results of learning in the online space. By combining classroom and external environments, the study explores the factors that influence the online learning process from multiple perspectives. Focusing on the key elements of the online learning process for college students, the study constructs a key factor impact model for the learning process, using representative students from domestic universities as research subjects to verify the real online learning process during the pandemic through a questionnaire survey. The study also explores the impact pathways and mechanisms between various elements of the learning process, aiming to provide insights for more effective online or blended teaching reforms in domestic universities.

2.2.2 Research hypotheses

  1. 1)

    Environment support

The influencing factors of learners' online learning process mainly include online environment factors, learner factors and teacher teaching. Environmental factors refer to the general environmental perception of online learning, which refers to students' feelings about the overall policy atmosphere of online teaching, especially whether students feel the support, tolerance, and resource guarantee of the surrounding environment including the school, family and related hardware and software environment for online teaching. Therefore, environment support refers to the attitudes of individuals who are considered important by college students (such as parents, teachers, classmates, etc.) towards their online learning, as well as the degree of convenience provided by the relevant technology and equipment that college students feel during their online learning process. Based on the above analysis, this study proposes hypotheses H1 and H2:

  • H1: There is a significant positive correlation between environment support and task-technology fit in online learning;

  • H2: There is a significant positive correlation between environment support and learning adaptability in online learning.

  1. 2)

    Teaching satisfaction

Online teaching satisfaction, as a specific environmental perception, is an important factor that affects college students' online learning experience. It refers to individuals' perception of specific online teaching tools and activities. Some scholars have found that in online education, teachers generally play the roles of educators, managers, social facilitators, and technicians. Teachers' guidance and feedback to students' learning, as well as the roles they play in the process, will affect students' learning engagement and overall learning experience. Many teachers still use traditional lecture-style teaching methods. Li & Zhong (2020) conducted interviews with 32 teachers and students in distance higher education online courses to explore the impact of online teachers' teaching engagement (knowledge explanation, teaching design, teacher-student interaction, teacher-student relationship) on students' learning performance (grades, satisfaction, student engagement), and found that the interactive support role played by online teachers has a positive impact on students' behavior and emotional participation through its impact on students' perception of course goals. Li et al. (2020) constructed a model of satisfaction with online learning for college students based on customer satisfaction theory, learning condition theory, and teaching system element theory. The results showed that factors such as online teaching quality, students' perception of task value, online self-efficacy, online usage ability, learning motivation, online interaction, and perceived social support have a significant predictive effect on satisfaction. Based on the above analysis, this study proposes hypotheses H3 and H4:

  • H3: There is a significant positive correlation between teaching satisfaction and task-technology fit in online learning;

  • H4: There is a significant positive correlation between teaching satisfaction and learning adaptability in online learning.

  1. 3)

    Perceived ease of use

Perceived usefulness in the online learning environment refers to individuals' subjective perception of the effectiveness of adopting the information platform in improving or enhancing learning outcomes. Perceived ease of use refers to individuals' perception of the ease of using the online learning platform for online learning. According to the technology acceptance model, if learners perceive the online learning platform as useful and easy to use, they will have a more positive attitude towards online learning, which will in turn enhance their online learning experience and satisfaction, and increase the likelihood of continued participation in online learning in the future (Arbaugh & Duray, 2002; Wang, 2007). The technology acceptance model mainly uses the three variables of perceived ease of use, perceived usefulness, and attitude to predict users' intention to adopt information systems (Davis, 1989). In fact, many studies have shown that perceived usefulness has a strong impact on intention to continue using, with an impact coefficient typically stable at 0.6 (Davis, 1989). Since perceived ease of use and the technical quality of the system exist in a cross-cutting relationship in the technical dimension, this study only analyzes the impact of perceived usefulness. Based on the above analysis, this study proposes hypotheses H5 and H6:

  • H5: There is a significant positive correlation between perceived ease of use and task-technology fit in online learning;

  • H6: There is a significant positive correlation between perceived ease of use and learning adaptability in online learning.

  1. 4)

    Task-technology fit

Task-technology fit is an important factor that affects the recognition of online learning by remote learners. It is the basic guarantee for the normal development of college students' online learning, and it is also a key variable that affects college students' willingness to continue using online learning. Qin et al. (2020) found that task-technology fit has the greatest impact on college students' willingness to continue using online learning among many influencing factors, highlighting the important value of task-technology fit in improving college students' online learning. Due to the impact of the COVID-19 pandemic, in response to the policy requirements of "suspension of classes without suspension of teaching and suspension of classes without suspension of learning", it has been found that organizational support, technical training, and policy support are indispensable conditions for promoting college students' participation in online learning (Liu et al., 2021). However, more and more results show that with the increasing maturity of online learning technology itself, the basic contradiction that affects the willingness of online learning to continue using has gradually shifted from the technical aspect to the dialectical unity of technology function supply and users' learning needs. Based on the above analysis, this study proposes hypothesis H7:

  • H7: There is a significant positive correlation between task-technology fit and learning adaptability in online learning.

3 Research design

3.1 Research design and participants

The data for this study were collected from the "National Survey of Undergraduate Education Teaching Quality" conducted by Xiamen University in 2021. This survey comprehensively investigated the learning situations of Chinese university students under the backdrop of the pandemic. From April to June 2021, electronic survey questionnaires were distributed to undergraduate institutions nationwide. Students completed the questionnaires by logging into an online survey platform. Only after answering all questions were they allowed to submit the questionnaire. A total of 133,069 electronic questionnaires were collected. After excluding incomplete and systematically answered samples, a total of 123,894 questionnaires from 272 universities were retained as the sample for analysis in this study, resulting in an effective response rate of 93.10%.

The response rate and effective rate of this study were both ideal, meeting the basic requirements of statistical analysis in social sciences. The sample covered different genders, grades, school types, and disciplines. The total sample size of this study was 123,894, with 38.60% male and 61.40% female. In addition, freshmen accounted for 39.70%, sophomores 28.40%, juniors 23.40%, and seniors and above 8.50%. Among the surveyed student population, 49.20% were in the field of humanities and social sciences, and 50.8% were in the fields of science, engineering, medicine, and agriculture. The average age of the participants in this study was 20.67 years old.

3.2 Research instruments

This study employed an online survey assessment tool and developed a structured questionnaire comprising several sections. The first section of the questionnaire collected demographic data, such as age, gender, and academic level, from the participants. Additionally, five latent variables were designed, encompassing teaching satisfaction, environmental support, perceived ease of use, task-technology fit, and learning adaptability. The items included in the questionnaire are outlined as follows:

The Teaching Satisfaction Scale, which surveys student satisfaction with various aspects of online learning, consists of 3 questions scored on a 5-point Likert scale (1 = very dissatisfied, 2 = dissatisfied, 3 = mostly satisfied, 4 = satisfied, 5 = very satisfied). The scale consists of a one-factor scale that investigates student satisfaction with the instructor's teaching mastery, classroom student–teacher interaction, and overall online teaching effectiveness.

The Environmental Support Scale was used to investigate students' perceived environmental support for online learning styles, including the attitudes of people around them toward online learning styles, online teaching service guarantee including support in terms of e-book resources, and support in terms of school policies. A 5-point Likert scale was used, with scores ranging from 1 to 5 indicating "not at all" to "completely", with higher scores indicating higher perceived environmental support.

The Perceived Ease of Use Scale is an adaptation of the Perceived Ease of Use subscale of the Technology Acceptance Scale proposed by Davis et al. (1989), which measures the degree to which college students believe that online learning is easy to implement. The scores are based on a 5-point Likert scale ranging from 1 to 5, from "not at all" to "completely", with higher scores indicating higher perceived ease of use.

The Task-technology Fit refers to the extent to which the learning needs of college students can be met with the help of online learning technology, and this study focuses on the single core dimension of matching to measure this scale. Through the previous literature review, the task needs of college students for online learning in the context of the epidemic are focused on quality resource sharing, personalized learning, and learning data analysis, so this study adapts the task technology matching scale proposed by Goodhue and Thompson (1995) to the contextualization of the above three dimensions. The online learning Task-technology Fit consists of three questions, all of which are scored on a 5-point Likert scale, with scores ranging from 1 to 5 indicating "not at all" to "completely", and the higher the score, the higher the degree of matching between online learning's functionality provision and college students' task needs. The higher the score, the higher the degree of matching between e-learning function provision and college students' task needs.

Finally, previous studies have mainly constructed indicators of learning adaptability from two dimensions: individual factors (e.g., learning attitude, learning ability, learning method, etc.) and environmental factors (e.g., curriculum, teaching method, learning environment, etc.). Similarly, the measurement of learning adaptability in this study should focus on the students' subjective initiative and the adaptation of the way in the process of online learning, which consists of three questions, all of which are scored using the Likert 5-point scale, with the scores ranging from 1 to 5 indicating "not at all in line with" to "completely in line with", and the higher the score, the higher the score means that the students' attitude is not in line with their learning ability. "The higher the score, the better the students' adaptation.

This study focuses on the measurement of college students' adaptability to online learning and its influencing factors, so the variables in the questionnaire are designed to help the authors adequately address the research questions.

The reliability and validity were tested in this study, and the rationality and effectiveness of the latent variables and questionnaire items were verified. As shown in Table 1, the Cronbach's alpha coefficients of each latent variable were all above 0.90, indicating that the scale had good reliability. Secondly, the structural validity of the questionnaire was tested using factor analysis, and the KMO value was above 0.70 (p < 0.001). Therefore, it can be concluded that the reliability and validity of the sample were good.

Table 1 Measurement of key variables
Full size table

3.3 Correlation analysis

This study used Pearson correlation coefficients to calculate the correlation between latent variables. Table 2 shows that there is a significant moderate correlation between the latent variables, indicating that it is suitable to use structural equation modeling to test the influence paths and path coefficients between the latent variables.

Table 2 Correlation matrix, reliability and validity, and descriptive statistics of latent variables (N = 123,894)
Full size table

3.4 Path model and analysis

Based on the theoretical framework, using AMOS 25 software, this study constructed a structural equation model of the key elements of college students' online learning process. Due to the large sample size of the study, other fit indices were used to judge the model fit. The fit results in Table 3 show that the structural equation model has good fit with the sample data, and the theoretical assumptions of the model are reasonable (Fig. 1).

Table 3 Fit indices of the structural equation model
Full size table
Fig. 1
figure 1

A theoretical model of the factors influencing college students’ online learning adaptability

Full size image

Through Fig. 2 can be seen that the path coefficients of each latent variable indicate significant effects, validating the interaction between the latent variables. Overall, the hypotheses of direct effects of the latent variables are supported. Except for the significant weak negative correlation between environmental support and task-technology fit, with an effect size of 0.07, which can be considered negligible. Different degrees of significant positive correlations exist between other latent variables. Perceived ease of use positively predicts online task-technology fit, and teaching satisfaction also positively predicts online learning task-technology fit. Task-technology fit in online learning also positively predicts learning adaptability. Perceived ease of use and teaching satisfaction both indirectly positively predict learning adaptability through identification (Table 4).

Fig. 2
figure 2

The path of the role of key elements of college students’ online learning adaptability. Note: ES Environmental Support; PEU Perceived ease of use; TS Teaching satisfaction; TTF Task-technology fit; LA Learning adaptability

Full size image
Table 4 Path coefficients of the structural equation model
Full size table

In addition to the direct predictive effects in the model, the results of the mediation analysis showed that: first, online task-technology fit has a significant partial mediating effect (β = 0.549, 95% CI: [0.556–0.602], p < 0.01) between perceived ease of use and learning adaptability; second, online task-technology fit has a significant partial mediating effect (β = 0.388, 95% CI: [0.394–0.413], p < 0.01) between teaching satisfaction and learning adaptability. Meanwhile, because the direct effect of the environmental support variable in the article is relatively small, we did not discuss it temporarily.

To test the applicability of the model, this study performed a multi-group structural equation modeling analysis by adding variables such as gender, grade, subject, and school type to restrict the path coefficients of the model. The results showed that there was no significant difference in fit indices between the restricted model and the non-restricted model (ΔCFI < 0.001, ΔIFI < 0.001, ΔNFI < 0.001), indicating that the model in this study has universality among different groups.

4 Discussion

The results indicate that there are significant correlations among variables such as environmental support, teaching satisfaction, and perceived ease of use. The statistical analysis reveals that environmental support, teaching satisfaction, perceived ease of use, and task-technology fit generally fall within the upper-middle range. Specifically, the aspects of the online learning environment, learner attributes, teaching processes, and task-technology fit all score above the theoretical median, suggesting that the current online learning environment effectively caters to the task demands of university students. This to some extent validates the efficacy of the continuous efforts made by the government, relevant industry sectors, and universities in enhancing educational technology infrastructure development and strategic adjustments during the pandemic period. However, it's important to note that several process elements of online learning have not reached the level of "very satisfied" for students. This implies that there is still room for upgrading and iteration in terms of technology, systems, and teaching. Beyond the pandemic, it will be crucial to enhance the comprehensibility of learning resources, foster an open sharing approach, cater to personalized learning needs, provide robust technological support, optimize teaching processes, innovate instructional models, and better meet the learning requirements of "digital natives." This will undoubtedly steer the development of online education in a direction that aligns with the needs of the digitally savvy learners (Fang, 2021).

Despite relatively positive findings reported in the questionnaire, structural equation testing revealed an unexpected result: environmental support even negatively predicted subsequent paths. To better comprehend this phenomenon, further research is warranted. This observation aligns with the recent developments in JD-R theory and research, which pinpoint individual differences in how needs and resources are perceived (Bakker & Demerouti, 2017; Han et al., 2020; Martin et al., 2016). Some individuals perceive resources as barriers rather than aids. In our study, it's conceivable that perceived control from school/family assistance might be viewed as an impediment, indicating adverse impacts on achievement. Similar counterintuitive effects of parental involvement and attitudes on student academic performance have been noted in other studies as well. Murayama et al. (2016) posit that overly positive parental evaluations might be disadvantageous due to their association with excessive involvement, controlling behavior, and heightened stress. Other investigations delve into parents' "intrusive support." For instance, Gunderson et al. (2012) explain how parental expectations grounded in their own anxiety and stereotypes can lead to lowered performance through intrusive support during homework. While environmental support in our study extends beyond parents, the logic remains consistent: excessive arrangements and undue attention from schools and peers might disrupt university students' online learning experiences.

The most significant finding is that the structural equation also revealed task-technology fit as a crucial mediating variable, where perceived ease of use exerts a positive impact on university students' online learning adaptability through task-technology fit. These research findings largely align with existing conclusions from studies on integrated models of TAM and TTF. For instance, Wu and Chen (2016) indicated that the more MOOC learning meets students' learning needs and the stronger the perceived practicality and convenience of MOOC, the more enduring their willingness to participate in learning. It's important to note that previous research on integrated models has already confirmed the positive facilitating role of perceived ease of use on adaptability (Davis, 1989). Many students exert effort to sustain their motivation and participation while finding ways to adapt to new learning environments (Besser et al., 2020; Perets et al., 2020). However, fewer studies have focused on the relationship between adaptability and task-technology fit. The results of this research uncover the noteworthy influence of task-technology fit on enhancing university students' online learning adaptability. As online learning technologies mature, the fundamental contradiction influencing online learning adaptability is gradually shifting from the technological aspect to the dialectical unity of technological capability provision and users' learning needs.

Finally, the research findings also highlight that teaching satisfaction serves as a fundamental element influencing the online learning process. Even in today's digitalized education landscape, students maintain higher expectations and a path dependency on teaching. It is generally believed that online teaching can better facilitate students' self-directed learning and self-management (Wu, 2020a, b), as the characteristics of online education seemingly enable more personalized learning. However, the results of this study reflect that students still possess a strong path dependency on teacher-led instruction, and they have elevated expectations for teaching organization, attitudes, and design. Consequently, using instructional interaction as the practical pathway to enhance the effectiveness of online teaching becomes imperative: the practice logic of instructional interaction and technological alignment. Technology has transformed the ways in which people interact with each other. Online teaching has become a medium for knowledge dissemination, emotional interaction, cultural transmission, and spiritual resonance, expanding the meaning and value inherent in traditional education. If online teaching only focuses on the unidirectional delivery of content while neglecting the role of emotional communication and sense of belonging, it can easily lead to limited teacher-student communication and weak interaction. This in turn hampers the improvement of instructional interaction effectiveness and impacts the quality of online teaching.

5 Conclusion

Based on a review of existing literature, this study constructed a model that focuses on the key elements of college students' online learning process, based on the effectiveness orientation, and verified the real online learning process of representative domestic university students during the epidemic period through questionnaire surveys. The study explored the impact pathways and mechanisms of various elements of the learning process and comprehensively discussed the relationships between environmental support, teaching satisfaction, perceived ease of use, task-technology fit, and learning adaptability. The structural equation model supported the hypothesis of this study, which is that college students' learning adaptability is affected by teaching satisfaction, perceived ease of use, and task-technology fit. At the same time, teaching satisfaction and perceived ease of use during college students' online learning process play a role in learning adaptability with the mediating variable of students' online learning technology task fit. The greater the perceived ease of use, the more helpful it is to improve college students' online task-technology fit, and thus adapt to online learning more easily, resulting in better learning outcomes. Similarly, the higher the teaching satisfaction, the more helpful it is to improve college students' task-technology fit in online learning, and thus adapt to online learning more easily, resulting in better learning outcomes. The results of multi-group analysis further support the relationship between the key elements of the online learning process.

6 Recommendations

Firstly, it is necessary to improve the technical support for online learning, grasp the task demands of college students, and improve the degree of task-technology fit. On the one hand, the development department of online learning platforms and the network technology department of universities need to continuously enhance the technical support for online learning. Relevant industry departments can meet the diverse and high-level learning needs of students by upgrading machine learning and big data analysis technologies, establishing teaching platforms that integrate resource distribution, resource management, resource evaluation, and knowledge management, etc. Universities can also form online learning technology service teams to provide "one-to-one" targeted support, answer various technical questions, and provide remote assistance to students. On the other hand, university managers and teachers should accurately grasp the intrinsic task demands of college students for online learning. In the mixed teaching environment of the digital age, meeting the diverse needs of students is still the focus and difficulty of improving teaching quality. This requires universities to comprehensively grasp the online learning demands of college students during the epidemic period through empirical research, fully tap and analyze learners' previous learning-related data, depict learner profiles, enhance the degree of task-technology fit, and thus better promote students' sustained participation in online teaching.

Furthermore, educational administrators need to facilitate a deep understanding of the essence of online teaching among teachers and students, fostering new concepts of instructional interaction. Teachers should be empowered to take a leading role, enhance the construction of teaching communities, encourage students' initiative and enthusiasm, mobilize their active participation, and increase interaction between teachers and students. "Online teaching will reform the accustomed passive learning state of students, accelerating a series of transformations from learning knowledge to cultivating abilities and exploring values" (Li & Xiong, 2020). However, looking at teaching satisfaction in the context of online learning, there are still shortcomings. Teaching satisfaction is closely related to the organizational competence of teachers in online teaching. Technology itself cannot replace teachers in providing students with positive emotional experiences. Teachers remain a crucial factor in education, and their interaction with students influences emotional experiences and learning outcomes. Yet, in online teaching, most teachers lack relevant experience and are unfamiliar with the patterns and organization of online instruction, resulting in insufficient interaction with students. Hence, higher education institutions need to strengthen teacher training, guide teachers in effectively utilizing digital technology, encourage the evolution of educational perspectives, and encourage the exploration of innovative teaching methods such as dialogic, seminar-style, and interactive teaching methods to enhance online teaching capabilities. Simultaneously, teachers should pay attention to students' experiences and perceptions of online learning, genuinely involving them in the process of educational and technological reform. As different disciplines often have varying technological requirements for online learning experiences (e.g., experimental operations for science subjects and contextual experiences for humanities), empowering students as active learners enables them to adapt to the external environment while continually constructing practical experiences, thus fostering their own development.

Lastly, it is crucial to strengthen technical training, enhance the quality of online teaching, and cultivate students' awareness of "usability." Designers of online learning technologies need to lower the complexity of manipulation and functionality of online learning technologies (including auxiliary technologies, intelligent learning systems, collaborative learning technologies, and big data technologies) and platforms (adaptation environment, text design, menus, navigation, etc.). This approach assists students in achieving learning objectives with minimal effort. Concurrently, higher education institutions should continuously reinforce the cultivation of students' information literacy. Through various forms of technical training, students' confidence and proficiency in using information technology can be elevated. These efforts will play a pivotal role in transforming our country's higher education institutions and society as a whole into the "online era" of educational productivity. This transformation will provide sustained momentum for national economic development and technological innovation. The commitment to advancement in the next phase of teaching practice is essential to achieving these goals.

7 Limitations

Some limitations need to be considered in our study. Firstly, our sample is drawn from the region of China, primarily focusing on the adaptability of Chinese university students to online learning. The normalization of online learning is a global trend, and this research should be replicated among more diverse student populations to enhance the generalizability of our findings. Secondly, despite our study being cross-sectional and utilizing a single research method, we cannot establish explicit causal conclusions. Future endeavors should involve rigorous experimental designs or mixed-method research approaches to further substantiate our claims. Furthermore, we lack detailed data on student performance in the classroom, and studying the adaptability of online learning from both the perspectives of teachers and students is imperative. Therefore, the next step of research will center on a comparative analysis of teachers and students in real classroom settings. For instance, evaluating teaching satisfaction only from the students' viewpoint regarding the online instructional process, without adequately addressing teachers' experiences, neglects an essential aspect. Teachers, as a pivotal part of the teaching process, deserve focused attention. Lastly, the prevailing majority of studies continue to treat adaptability as an individual resource to explain its impact on learning outcomes, while investigations that consider adaptability as an outcome variable to explore its mechanisms are scarce. Subsequent research should strengthen examinations of the mechanisms influencing adaptive learning and enhance studies on the impacts of various independent variables in online learning, thereby enhancing the explanatory power of integrated theories concerning real-world issues in China.