Introduction

Informal online learning is defined as “unstructured, ubiquitous and spontaneous learning that occurs in daily life when learners are accessing the internet” with mobile devices, e.g., smartphones, tablets and other portal devices (Holland, 2019, p.215). In fact, informal online learning supports learners’ collaborative learning and knowledge sharing when they are staying in a shared mobile learning community (Cook & Smith, 2004; Mills et al., 2014), supporting their synchronous or asynchronous academic activities anytime, anywhere with the mobile devices (Li, 2023a). More importantly, learners can instill social network applications (Apps) directly, e.g., WeChat, Tencent QQ and WhatsApp, etc., to engage in informal online learning activities (Li, 2022a, b). As a result, informal online learning has to date attracted considerable attention in language learning (e.g., Balouchi & Samad, 2021; Jurkovič, 2019), healthcare (Treasure-Jones et al., 2019), and teacher education (Grosemans et al., 2015), among others.

In China, the rapid advances of social network Apps have provided in-service teachers the ubiquitous, multimodal and engaging informal learning environments, and Chinese teachers tend to positively adopt informal online learning (Huang et al., 2020; Yu et al., 2021). For instance, Huang et al. (2020) explored how contextual (school resources, instructional creativity, and colleague collaboration) and psychological factor (psychological capital) influence Chinese in-service teachers’ informal online learning. Results of their study indicated that both contextual and psychological factors are crucial for teachers’ informal learning activities. In another recent study, Yu et al. (2021) surveyed the effect of Chinese teachers’ informal learning on innovative teaching during the COVID-19 pandemic and found that motivational factors, viz. personal teaching efficacy, general teaching efficacy and social network efficacy, are the major moderators for the use of informal online learning.

Notwithstanding the aforementioned research that may suggest Chinese teachers’ active reaction to informal online learning, their perceptions of informal online learning in the shared mobile learning community have been seldom addressed, let alone factors lying behind their continuance intention (Li, 2021). As posited by Liu et al. (2010), learners might fail to continue the use of educational technology if they did not perceive any long-term benefits. Huang (2016) also argued that Chinese learners’ continuance intention has been gradually taken as an important variable, because they may discontinue using the technology even if they have initially accepted it. Given the research gap, this paper aims to identify factors that influence Chinese in-service teachers’ informal online learning continuance in a mobile learning community, based on the expectation confirmation model (ECM) (Bhattacherjee, 2001). From Deci and Ryan’s (1985) self-determination perspective, Lin (2007) asserted that Chinese teachers’ extrinsic and intrinsic motivation are the main incentives of using informal online learning. In self-determination theory, motivation would be basically divided into extrinsic and intrinsic types (Deci & Ryan, 2000). The former refers to doing something to earn external rewards or avoid punishments, while the latter implies doing something driven by internal rewards (e.g., happiness or enjoyment) (Deci & Ryan, 2000; Feng et al., 2016). While Chinese in-service teachers might accept informal online learning due to the possibility of achieving utility and enjoyment from it, it remains largely unclear how and in what ways intrinsic and extrinsic motivation affect their informal online learning continuance (Feng et al., 2016; Lee et al., 2019).

Therefore, it is of vital importance to justify the motivational factors that influence Chinese teachers’ informal online learning continuance in a mobile learning community through an integrated model consisting of intrinsic–extrinsic motivation. To this end, the present study developed a structural model by incorporating Chinese teachers’ intrinsic–extrinsic motivation as antecedents of the ECM. In addition, the structural relationship between antecedents and the original ECM constructs was determined by our empirical investigations. Questionnaire surveys based on the proposed model were designed to collect data on Chinese teachers’ perceptions about informal online learning continuance, from which factors behind their continuance intention could be identified.

The remainder of the paper is structured as follows: First, to gain a better understanding of the theoretical underpinnings, literature review on the related studies of informal online learning, the ECM and intrinsic–extrinsic motivation was presented, followed by the proposed research model and hypotheses. Then, to test the proposed model, research design regarding confirmatory factor analysis was carried out. Finally, empirical findings and implications were discussed.

Literature review

Informal online learning

Unlike formal learning in schools or universities, informal online learning, free from situational or temporal barriers, is about open-ended, flexible and self-directed learning activities that consider both the contextual affordances and motivational factors most relevant to learners (Beach, 2018; Gomes Junior, 2020). In this study, informal online learning refers to those informal learning activities—teachers’ autonomous sharing of academic knowledge or resources, inquiry of academic questions and provision of academic feedbacks—that occurred in a mobile learning community entitled “Chinese Education Youth Scholar Group”. To date, drawing on the contextual affordances and motivational factors of informal learning, a growing body of empirical studies into informal online learning underlines the importance of understanding the education potentials of online learning beyond classroom (Li, 2022c; Lecat et al., 2020).

Arising from an ecological view, one strand of studies examines the contextual affordances of informal learning. Informal learning environments enable to provide learners with meaningful opportunities, as long as they interact with and within the environments that meanings become available, and affordances are perceived (Li et al., 2021; Gomes Junior, 2020). As such, researchers have explored teachers’ perceptions regarding the affordances of informal online learning environments, e.g., perceived usefulness (Dorner & Kumar, 2016; Lai, 2015; Wang et al., 2013), satisfaction (Limongelli et al., 2011; Tadesse et al., 2020), and expected cognitions and emotions (Hoekstra & Korthagen, 2011; Huang et al., 2020). For instance, to help teachers integrate educational technology in their teaching practice, Dorner and Kumar (2016) developed an informal online learning model that combines online modular of pedagogical training with informal online learning community to examine the sharing experience of learning resources during the coursework. Results based on the survey of 116 pre-service teachers supported the usefulness of the informal online learning, as indexed by the improved technology self-efficacy and satisfaction. Likewise, Tadesse et al. (2020) adopted a quasi-experimental design by comparing the results between the treatment group that used the informal online learning approach and the control group that used the traditional lecture instruction in terms of teaching effectiveness, task orientation, and learning satisfaction. Results of their study indicated that those in the informal learning group tend to have higher perceptions of teaching effectiveness, more task orientation and great satisfaction than those in the control group. To gain a deeper understanding of the associations between five different informal learning activities (e.g., learning through media, colleague interaction, stakeholder interaction, student interaction and individual reflection) and three expected motivations (e.g., enjoyment, anxiety and anger) among 2880 primary teachers. Huang et al. (2020) argued that teacher enjoyment was positively associated with all the five informal learning activities, whereas anxiety and anger were negatively related to some of the informal learning activities. Specifically, anxiety was negatively associated with learning colleague interaction and individual reflection, and anger was negatively related to learning through student interaction.

Apart from the focus of the contextual affordances, the other strand of studies investigates users’ perceptions of informal learning, with a particular eye on their intrinsic–extrinsic motivation, which can help explain users’ incentives of participating in the informal online learning (Dorner & Kumar, 2016; Huang et al., 2020). Driven by the extrinsic motivation, teachers are likely to focus on the satisfaction of their personal gains or rewards earned, while those with the intrinsic motivation seem more willing to share knowledge or information with other group members in the informal online learning environments (Lee et al., 2019; Lin, 2007; Shin, 2010, 2018). For instance, Lin (2007) found that the extrinsic motivation (e.g., expected organizational rewards and reciprocal benefits) could only partially predict knowledge sharing attitude and intention of informal learning, whereas intrinsic motivation (e.g., knowledge self-efficacy and enjoyment in helping others) could positively predict both variables. In a qualitative research, Hur et al. (2012) analyzed teachers’ knowledge sharing in a self-generated informal online learning community and suggested that teachers’ knowledge sharing should be supported to promote their engagement in the online community.

Although the aforementioned studies have been helpful in affording insights into informal online learning, few studies published to date have directly addressed factors that influence Chinese teachers’ informal online learning continuance, which suggests an urgent need to revisit the effects of those perceptions and examine whether intrinsic and extrinsic motivation altogether influence their informal online learning continuance. The present study, therefore, was promoted by a need to obtain a perspective of the ECM that may influence Chinese teachers’ informal online learning continuance in a mobile learning community.

Expectation confirmation model

Over the years, the ECM has been considered as a robust and well-established model to predict learners’ continuance intention of using a new educational technology, e.g., mobile social learning platforms (Ooi et al., 2018), MOOCs (Alraimi et al., 2015; Dai et al., 2020; Joo et al., 2018), and cloud services (Huang, 2016), etc. Drawing on this line of inquiry, the ECM is also taken as the theoretical framework to examine how factors influencing teachers’ informal online learning continuance in a mobile learning community.

The theoretical underpinnings of the ECM originate from Oliver’s (1980) expectation confirmation theory to examine the effects of consumers’ satisfaction on their purchase intention. Later, motivated by technology acceptance model (Davis, 1989), innovation diffusion theory (Rogers, 1995) and theory of planned behavior (Ajzen, 1991), Bhattacherjee (2001) extended Oliver’s (1980) expectation confirmation theory and proposed the ECM in the information system (IS) context. The ECM includes four constructs, i.e., perceived usefulness, confirmation, satisfaction and IS continuance intention. Perceived usefulness refers to the usefulness of a particular IS system to improve users’ performance (Li, 2021; Li et al., 2019), which refers to the educational affordances (e.g., collaborative learning and knowledge sharing) provided by mobile learning community in this study. Confirmation is understood as the gap between the actual performances and users’ expectations (Bhattacherjee, 2001), which refers to the comparison between teachers’ actual performances (e.g., sense of achievement and belongness) and their expectations. Similarly, satisfaction is defined as users’ satisfied feelings towards the IS system (Bhattacherjee, 2001), which refers to teachers’ satisfied feelings towards the educational affordances of mobile learning community in this study. Continuance intention refers to users’ intention to use the IS system in the future (Bhattacherjee, 2001), which is also taken as an index of teachers’ informal online learning continuance.

With regard to the causal relations of four constructs in the original ECM, some existing studies hold that perceived usefulness has a positive effect on continuance intention (Chang & Zhu, 2012; Chen et al., 2015; Huang, 2016) and satisfaction (Chang & Zhu, 2012; Chen et al., 2015; Joo et al., 2018). Some others insist that confirmation is the major determinant that influences users’ perceived usefulness and satisfaction (Bhattacherjee, 2001; Chang & Zhu, 2012; Chen et al., 2015; Dai et al., 2020). Likewise, others posit that users’ continuance intention is positively dependent on their satisfaction (Chen et al., 2015; Joo et al., 2018). Based on the literature reviewed above and research model shown in Fig. 1, the study proposes the following research hypotheses to examine the causal relations between perceived usefulness, confirmation, satisfaction and continuance intention when informal online learning is conducted:

Fig. 1
figure 1

The proposed research model

Full size image

H1a

Perceived usefulness has a positive impact on teachers’ continuance intention.

H1b

Perceived usefulness has a positive impact on teachers’ satisfaction.

H2a

Confirmation has a positive impact on teachers’ perceived usefulness.

H2b

Confirmation has a positive impact on teachers’ satisfaction.

H3

Satisfaction has a positive impact on teachers’ continuance intention.

A critical review of the ECM has revealed that there is an urgent need to add some external factors into the original model to improve its explanatory power. Researchers (e.g., Chang & Zhu, 2012; Chen et al., 2015; Dai et al., 2020; Huang, 2016; Joo et al., 2018) seek for solutions to extend the original ECM with some external factors by incorporating the antecedents into the key ECM constructs, i.e., perceived usefulness, confirmation, satisfaction and continuance intention. For instance, Cho and Lee (2020) criticized the simplicity of ECM in explaining users’ continuance intention of smart device use and proposed an expanded version composed of five major components of confirmation, perceived usefulness, perceived ease of use, satisfaction, and continuance use intention. After critically reviewing some key empirical studies regarding the IS continuance literature published in leading journals, Osatuyi et al. (2020) argued that the original ECM may be restrictive in fully capturing the complex dynamics of IS continuance and extended the ECM by adding some motivational antecedents, i.e., social benefits, hedonic benefits and utility benefits, to understand the complexities of social commerce continuance. Similarly, from self-determination perspective, Dai et al. (2020) posited that the original ECM is insufficient to explain future use of educational technology and highlighted that both extrinsic motivation (viz. perceived usefulness) and intrinsic motivation (viz. curiosity) are important reasons to initiate and continue technology use behavior. While Dai et al. (2020) may provide valuable insights into the incorporation of intrinsic–extrinsic motivation antecedents to the ECM, there is still a need to distinguish the subtle difference between perceived usefulness and extrinsic motivation, as the former refers to the perceived affordances provided by an IS system (Li et al., 2019; Davis, 1989), and the latter the motivate of doing something “as a means to achieve self-benefit” (Lee et al., 2019, p.182).

Intrinsic–extrinsic motivation

Drawing on self-determination theory perspective, the current study attempts to examine factors affecting Chinese teachers’ informal online learning continuance in a mobile learning community by adopting an extended ECM combined with intrinsic and extrinsic motivation for two considerations. On one hand, as motivations, which are the impetus or physiological drive of learning behaviors (Deci & Ryan, 2000), are essential to informal learning as well. Thus, the inclusion of intrinsic–extrinsic motivation may shed some light on the psychological mechanisms of teachers’ informal learning behaviors. On the other hand, previous studies on mobile social learning platforms continuance (Ooi et al., 2018), MOOCs continuance (Dai et al., 2020), and social commerce continuance (Osatuyi et al., 2020) have claimed that an extended ECM might outperform the original ECM in relation to the explanatory power. What remains unclear is whether the extended ECM by incorporating intrinsic–extrinsic motivation into the ECM can be applied to predict teachers’ informal online learning continuance. In this sense, results of this study might also contribute to validating the theoretical underpinnings of the extended ECM in turn.

For the two antecedents, extrinsic motivation refers to “individual behavior is driven by its perceived values and the benefits of the action” (Lin, 2007, p.139), and intrinsic motivation is defined as “motivation to do something due to intrinsic and inherent satisfaction or enjoyment” (Yoo et al., 2012, p.944). More specifically, the fundamental informal learning goals of extrinsically motivated teachers are to understand or solve academic questions from other group members and inquire for valuable academic resources they need. Driven by intrinsic motivation, teachers are likely to feel enjoyment and pleasure when sharing their knowledge, ideas or academic resources to other group members in the mobile learning community for the purpose of informal learning. In the literature, studies on the importance of both extrinsic and intrinsic motivation are well acknowledged (Deci & Ryan, 2000; Shibchurn & Yan, 2015; Yoo et al., 2012). Prior study (e.g., Deci & Ryan, 2000) suggested that extrinsic motivation reflects learners’ external control or true self-regulation on their learning activities, and intrinsic motivation reflects the natural human propensity to learn and assimilate. As such, researchers tend to take intrinsic and extrinsic motivation as the crucial antecedents that influence educational technology use behaviors, e.g., social networking sites (Shibchurn & Yan, 2015), web-based educational environment (Sánchez-Franco et al., 2014), e-learning (Roca & Gagné, 2008) and knowledge sharing (Lin, 2007). For instance, Roca and Gagné (2008) examined the factors of e-learning continuance intention in the workplace and argued that both intrinsic and extrinsic motivation play some positive and direct roles in the continuance intention. Similarly, Chang et al. (2013) investigated the influence of learners’ intrinsic and extrinsic motivation on English mobile learning systems continuance. Results of their study showed that both intrinsic and extrinsic motivation positively predict perceived usefulness and continuance intention. Sánchez-Franco et al. (2014) explored user’ intrinsic and extrinsic motivation on a web-based educational environment use, and found that both intrinsic and extrinsic motivation positively predict satisfaction. As Deci and Ryan (2000, p.57) put it, all behaviors are motivated by physiological drives, we could assume that both intrinsic and extrinsic motivation positively predict confirmation. Accordingly, as displayed in Fig. 1, the following hypotheses regarding two motivational antecedents were proposed:

H4a

Extrinsic motivation has a positive impact on perceived usefulness.

H4b

Extrinsic motivation has a positive impact on teachers’ continuance intention.

H4c

Extrinsic motivation has a positive impact on teachers’ satisfaction.

H4d

Extrinsic motivation has a positive impact on confirmation.

H5a

Intrinsic motivation has a positive impact on perceived usefulness.

H5b

Intrinsic motivation has a positive impact on teachers’ continuance intention.

H5c

Intrinsic motivation has a positive impact on teachers’ satisfaction.

H5d

Intrinsic motivation has a positive impact on confirmation.

Research design

Participants

A total of 303 Chinese in-service teachers recruited from the mobile learning community entitled “Chinese Education Youth Scholar Group” (see Appendix 1) volunteered and consented to participate in the study. There was neither incentive for completing the survey, nor was there any penalty for not completing the questionnaire. As a result, 72 invalid questionnaires due to failure of answering the trap questions that were randomly distributed with the questionnaire items were eliminated, resulting in 231 valid questionnaires. The illustration of demographic information is shown in Table 1.

Table 1 Demographic data
Full size table

The results infer that 158 (68.4%) are males and 73 (31.6%) are females. Over half of the responses (50.20%) are from the teachers of age ranging from 26 to 35 years old, followed by 36–45 years (35.10%), as the majority of the teachers are youth scholars. With regard to their educational background, 53.2% of the youth scholars hold doctorate degrees, followed by master degrees (36.80%) and bachelor degrees (9.50%), respectively, suggesting that those teachers have higher educational backgrounds. Most of the responses (79.70%) are from university teachers, followed by vocational teachers (8.20%) and primary or secondary teachers (6.50%). As far as professional title is concerned, most of the responses are from lecturers (40.30%), followed by associate professors or professors (29.90%), others (21.20%) and teaching assistants (8.70%), respectively. Moreover, most of the participants (83.50%) support the use of real names rather than fake names, indicating that they intend to be responsible for their informal learning behaviors and activities with their real names.

Mobile learning community

The mobile learning community entitled “Chinese Education Youth Scholar Group” refers to one of the popular social network platforms, i.e., Tencent QQ group, which provides teachers with a personalized, collaborative and ubiquitous learning environment (see Appendix 1). After a careful selection of educational technologies available in this study, Tencent QQ group was finally applied as the mobile learning community for two reasons. On one hand, Tencent QQ was one of the popular instant communication Apps that was familiar among Chinese users, as Benson (2019) recommended that adopting user familiar Apps was beneficial to learners. And the collaborative features of Tencent QQ were also confirmed in Zeng (2017), who tried to examine its educational affordances. On the other, Tencent QQ can be installed on mobile devices with either Apple iOS or Android operating systems (Li et al., 2019). Mobile devices can be used to promote ubiquitous learning anywhere and anytime, to foster social learning and knowledge sharing for the informal learning purposes.

Instruments

Two questionnaires in the instruments were designed to measure the constructs of ECM, intrinsic and extrinsic motivation, respectively. A summary of each construct is shown in Appendix 2.

Items regarding four constructs of ECM were adapted based on the questionnaire developed by Alraimi et al., (2015) and Bhattacherjee (2001), including perceived usefulness (e.g., ‘I can learn a lot from other group members.’), confirmation (e.g., ‘My sense of achievement is better than I expected.’), satisfaction (e.g., ‘I feel satisfied with other members’ feedback to my academic questions.’) and continuance intention (e.g., ‘I intend to inquire academic problems in the group in the future.’).

Regarding two antecedents of ECM, six items of intrinsic motivation and extrinsic motivation were adapted from the questionnaire items developed by Lee et al. (2019) and Lin (2007), including three items of extrinsic motivation (e.g., ‘I join this group to understand an academic problem.’) and another three items of intrinsic motivation (e.g., ‘I am confident in my academic ability to share my knowledge with the group members.’).

The procedures of questionnaire development were as follows. First, questionnaires adapted from the literature reviewed above were translated into simplified Chinese. Second, the Chinese version was back translated to English by an English teacher majoring in translation to ensure its translation quality. The high similarity between two versions confirmed its accuracy. Third, five Chinese postgraduate students pre-tested and commented on the survey to evaluate the face validity. Minor adjustments to wording and formatting were made accordingly. The final Chinese version of questionnaires (see column of ‘Chinese translations’ in Appendix 2) was used in data collection.

Data collection procedure

Before collecting data, participants were first asked to fill in the web-based consent forms. Then, they were given web-based questionnaires with two sub-sections (section 1: demographic information; and section 2: 7-point Likert scale surveys for ECM, intrinsic and extrinsic motivation anchored on ‘1 = strongly disagree’ and ‘7 = strongly agree’) to fill, without any previous notice.

Data analysis

Data of the questionnaire had been analyzed to validate the proposed model (see Fig. 1). To follow Anderson and Gerbing’s (1988) instruction, a two-step structural equation modeling (SEM) approach was adopted: First, the confirmatory factor analysis (CFA) was carried out to assess reliability (e.g., Cronbach’s α and composite reliability) and convergent validity of the scale. Second, the SEM was employed to empirically examine path coefficients of the proposed model. Since the sample size of 231 is over the suggested minimum sample size of 150 with seven constructs or less (Hair et al., 2006), the proposed theoretical model could then be tested. By using the maximum likelihood, the overall goodness-of-fit, the relative strengths of each cause-effect path as well as the explanatory power of the proposed model were analyzed. The descriptive statistics and Cronbach’s α coefficients that could reveal an overview of the descriptive statistic results were obtained through SPSS 24.0, and the structural model that could suggest the causal relationships of the proposed model was evaluated through AMOS 24.0.

Results

Descriptive statistics

Table 2 shows the mean (M), standard deviation (SD), skewness and kurtosis of perceived usefulness, confirmation, satisfaction, continuance intention, intrinsic and extrinsic motivation. The values of each construct range from 1 to 7 with 4 as the mid-point. It can be seen in Table 2 that the mean values of the constructs range from 4.42 to 5.64, with intrinsic motivation being the lowest and continuance intention being the highest, indicating teachers’ overall positive responses to each of these variables, continuance intention in particular. The normality of the data is checked by observing the maximum likelihood estimation of Skewness (within ± 3) and Kurtosis (within ± 10) requirements (Kline, 2005; Wong, 2016). It is obvious to observe in Table 2 that the data are normally distributed, as Skewness and Kurtosis values of all items fall within the desirable range.

Table 2 Descriptive statistics of each construct
Full size table

Reliability and validity analysis

Reliability analysis consists of two reliability index values: Cronbach’s α (> 0.70) and composite reliability (CR > 0.6) (Hair et al., 2006). Likewise, validity analysis involves two validity index values: convergent validity (the values of factor loadings > 0.5, and average variance extracted > 0.5) and discriminant validity of each construct (Fornell & Larcker, 1981). Results of reliability and validity analysis are summarized in Tables 3 and 4.

Table 3 Reliability and convergent validity of the constructs
Full size table
Table 4 Correlation matrix and discriminant validity
Full size table

In Table 3, all values of Cronbach’s α are more than 0.70 and the values of CR are over the minimum value of 0.60, suggesting the satisfactory reliability of the measurement model (Hair et al., 2006). And the convergent validity is ensured in that loadings of each item are beyond 0.5 and AVEs of each construct exceed the threshold value of 0.50 (Fornell & Larcker, 1981).

In Table 4, the discriminant validity was further tested by comparing the square roots of AVE values with the correlations between the constructs (Fornell & Larcker, 1981). Correlations between constructs lower than the squared root of AVEs shown in bold type could be verified for discriminant validity. All criterions of the measurement model, in this sense, are met in all cases.

Structural model

After confirming the reliability and validity of the measurement model, structural model was further performed to evaluate the structural relations among the proposed constructs. As reported in Table 5, the structural model has a good model fit (Hair et al., 2006), suggesting that a further exploration of the structural relations among the proposed constructs can be conducted. The estimates and significance of each hypothesis in the SEM was summarized in Table 6. The structural model with path coefficients of each construct was presented in Fig. 2.

Table 5 CFA model fit indices
Full size table
Table 6 Estimates and significance of each hypothesis in the SEM
Full size table
Fig. 2
figure 2

The structural model

Full size image

Discussion

Since the field is in urgent need of an empirical research on in-service teachers’ continuance intention of informal online learning, this study would constitute a timely initiative that fills a significant gap. This study not only contributes to promoting the theoretical model of ECM in general, but also deepens the understanding of possible factors that influence Chinese teachers’ informal learning continuance in particular. More specifically, the main purpose of the paper is to examine the motivational factors that influence teachers’ informal online learning continuance in a mobile learning community. To this end, intrinsic and extrinsic motivation, as the antecedent variables, were incorporated into the extended ECM. The theoretical model was developed and the research hypotheses were also proposed. To further testify the hypotheses, 231 valid data were collected from Chinese in-service teachers. The results validated the proposed model and confirmed all the hypotheses within the ECM constructs, except for H4b and H4c that assume extrinsic motivation positively affects continuance intention and satisfaction, along with H5a, H5b and H5d that assume intrinsic motivation positively influences confirmation, perceived usefulness and satisfaction, respectively. The findings are discussed in the rest of this section.

For the causal relationship of the original ECM constructs, the hypotheses of four constructs, i.e., perceived usefulness, confirmation, satisfaction and continuance intention, are confirmed. First, both satisfaction and perceived usefulness are positive predictors of teachers’ continuance intention, with satisfaction being stronger than that of perceived usefulness (β SAT→CI vs. β PU→CI = 0.44 vs. 0.43), indicating that teachers’ major concern might lie in whether the educational affordances of mobile learning community could satisfy their personalized needs (Feng et al., 2016). This result was consistent with those reported in some earlier studies that investigated MOOCs (Alraimi et al., 2015), social networking sites (Chang & Zhu, 2012), and blogs (Chen et al., 2015) continuance. This finding, however, contradicted findings of Cho and Lee (2020) who regarded perceived usefulness as a determinant of users’ smart device continuance (β SAT→CI vs. β PU→CI = 0.33 vs. 0.48). A plausible explanation for the discrepancy might be the educational affordances of the technologies involved. Since there is no such thing as the physical existence of MOOCs, social networking sites and blog, etc., users might paid more attention to the satisfied services afforded by these technologies, compared with smart devices that have physical existence. As such, mobile learning community that has no physical existence further corroborates the argument drawn above. Second, as expected, both confirmation and perceived usefulness positively predicts satisfaction, which means the expectation gap between the desired outcomes and the actual use experiences might directly influence language teachers’ satisfied affects or feelings of use experiences, supported by results reported in the existing studies (Alraimi et al., 2015; Bhattacherjee & Lin, 2015; Huang, 2016), indicating that the higher an individual’s confirmed expectation, the higher an individual’s satisfied feelings and perceived usefulness of the informal learning. Third, confirmation plays a positive role in perceived usefulness, resonating a number of studies (Bhattacherjee, 2001; Chen et al., 2015; Dai et al., 2020). This result suggests that once teachers’ expectation for the informal online learning was satisfied and confirmed, they were prone to feel it useful.

Aside from the causal relationship among the original ECM constructs, two motivational antecedents are considered for analysis as well. For the first antecedent, it was found that, while extrinsic motivation positively predicts perceived usefulness (β = 0.27, p < 0.001) and confirmation (β = 0.67, p < 0.001), the nonsignificant associations between extrinsic motivation and continuance intention and satisfaction are not obtained. Increasing teachers’ extrinsic motivation could lead to higher usefulness perceptions and confirmation, consistent with those findings in existing research (Chang et al., 2013). That is, those teachers who join the informal online learning for their personal academic gains rather than obtaining enjoyment in helping others are more likely to highlight the expected gains (i.e., confirmation and perceived usefulness) they could achieve in the informal online learning activities. Thus, it could be reasonable to speculate that, autonomous and unstructured informal learning activities occurred in this study could not be possible, if teachers have little willingness to engage, share knowledge and exchange ideas collaboratively (Hur et al., 2012; Lin, 2007). When it comes to the effects of the second antecedent as hypothesized, counterintuitively, the results showed that, intrinsic motivation does not positively affects confirmation, perceived usefulness and satisfaction; rather, it directly and positively influences continuance intention (β = − 0.20, p < 0.001), a finding that warrants detailed attention. It could be interpreted that, those teachers who are intrinsically driven to engage in the activity are more likely to continue the informal online learning, because they are willing to help other in knowledge sharing (Lin, 2007), and enjoy constantly inherent satisfaction or enjoyment in the informal learning process (Yoo et al., 2012).

Implications

Theoretically speaking, this study has proposed the extended ECM and examined the factors that influence teachers’ informal online learning continuance by the inclusion of intrinsic-extrinsic motivation together with the original ECM. The causal relations among the four ECM constructs, the roles of intrinsic and extrinsic motivation are discussed. First, satisfaction is the major determinant of teachers’ continuance intention, followed by perceived usefulness and intrinsic motivation. In addition, their extrinsic motivation positively predicts perceived usefulness and confirmation, shedding some light on the causal relations among intrinsic-extrinsic motivation and the original ECM. On the other, the theoretical soundness of the extended ECM proposed in this study has been confirmed using the quantitative SEM method. Drawing on intrinsic-extrinsic motivation perspective, this study integrates the prior studies based on the ECM to develop the extended ECM, which will provide some theoretical framework for the subsequent research.

Some practical implications for programmers and developers, educational managers and group members are inferred from the major findings as follows. First, as satisfaction is the major determinant of teachers’ continuance intention, programmers and developers should develop friendly mobile learning community to ensure learners’ satisfied feelings towards it. Educational managers should also try to improve group members’ trust by encouraging them to use real names, since real name would enjoy a higher level of online interpersonal trust than fake name (Zürn & Topolinski, 2017). The higher the sense of trust, the high the satisfaction they would feel. Second, as perceived usefulness is also an important determinant, programmers and developers should enhance the functionalities of the mobile learning community to improve users’ continuance intention (Roca & Gagné, 2008). Third, as intrinsic motivation positively predicts continuance intention, and extrinsic motivation positively predicts perceived usefulness and confirmation, group members should try to develop their informal online learning autonomy and foster the transformation from extrinsic to intrinsic motivation. It is argued that the improved internalization “comes greater persistence, more positive self-perceptions, and better quality of engagement” in informal online learning activities (Ryan & Deci, 2000, p.60–61).

Conclusion

To the best of our knowledge, existing studies fail to examine the underlying factors of informal online learning based on ECM. To fill the gap, this study added motivational antecedents (i.e., intrinsic and extrinsic motivation) with four ECM constructs. Results indicated that satisfaction is the major determinant of continuance intention, followed by perceived usefulness and intrinsic motivation. In addition, extrinsic motivation positively predicts perceived usefulness and confirmation.

Apart from the meaningful findings, limitations of the study remain to be addressed in the future. First, as this is a cross-sectional investigation, the dynamic development of factors from the longitudinal and comparative approach remains open for debate. Future research could try to compare the causal relationships between the pre-use and post-use behaviors (Terzis et al., 2013). Second, as a preliminary study of informal online learning continuance focusing on intrinsic-extrinsic motivation perspective, further attempt needs to be done from other insightful perspectives to gain a more comprehensive understanding. Third, this study fails to take the modulating effects of some moderators, e.g., gender, age, and educational background, etc. into account (Li, 2023b). Future research could seek for multi-group analysis technique to examine the modulating effects the aforementioned moderators. Lastly, while the quantitative SEM method adopted in the study could reveal the causal relationships of the informal online learning continuance variables in the proposed model, some mixed-methods attempts that could combine both the qualitative and methods still need to be done to triangulate the quantitative results in the future.