Many influential educational psychologists have proposed the definition of self-regulated learning from different perspectives. Zimmerman [7] defined self-regulated learning as students’ planned regulation of their cognition, emotion, and activities to achieve their goals. Winne and Perry [10] believed that self-regulated learning was an inherent process of students’ construction and self-orientation. Continued research into self-regulated learning has also emphasized that self-regulated learning is a series of behavioral processes to achieve specific personal goals [11] , and it includes goal setting, effort process and strategy, feedback and self-evaluation. In addition, domestic scholars have expressed similar views and understandings of the concept of “self-regulated learning”. Li et al. [12] believed that self-regulated learning referred to learners’ use of cognitive, motivational and behavioral strategies to organize, guide, monitor and evaluate their own learning process. Pang [13] also demonstrated that self-regulated learning could be seen as the ability of learners to consciously set learning goals, choose learning strategies, monitor the learning process and evaluate learning results.

Self-efficacy refers to individual’s expectations of whether he or she has the ability to accomplish a certain behavior in a particular situation [14] . When referring to course learning, Soffa [15] argued that self-efficacy can be considered as students’ confidence or belief in their ability to complete a specific course task. Thus, students with high self-efficacy think that if they can learn new knowledge with serious lectures or reading, they can find ways to solve complex problems and know how to seek resource support; while students with low self-efficacy usually feel that they do not have the ability to solve problems.

Moreover, research on SRL has shown a positive correlation between self-efficacy and the use of SRL strategies. For example, Hamdan et al. carried [16] research to find out the factors and relationships that influence self-regulated learning. The results demonstrated that self-efficacy had a significant positive effect on self-regulated learning. Another study from a social cognitive perspective was conducted with a total of 184 English learners by Lee et al. [17] also indicated that there was a positive and significant relationship between self-efficacy and self-regulated learning. In addition, developing the self-regulated learning of students can help improve their self-efficacy.

Task value refers to students’ evaluation of how interesting, important, and useful a task is, and it is considered to be an important predictor of perceived effectiveness in the English learning environment [18] . When students face a new task, they will estimate the value of the task at first. Only when students determine that it is a valuable task will they take further action [19] . However, once students find that they lack of ability to complete the task, they may not take action or choose an easier goal [14] . On the contrary, if it is likely for them to succeed in something, they tend to have an optimistic and positive attitude. These are the processes of cognitive operation and affective coping, which can also be used to assess student self-efficacy.

The importance of task value in learning cannot be ignored. It is not only the source of motivation for learning but also the cornerstone for achieving learning goals [18] . Clear task value can stimulate learners’ initiative and creativity, making them more focused on learning tasks, thereby improving learning efficiency and quality. At the same time, task value can also help learners understand the practical application and significance of knowledge, and enhance the pertinence and practicality of learning [19] . Therefore, in the learning process, we should fully recognize and utilize the value of tasks, and integrate them into various stages of learning, in order to achieve better learning outcomes and broader development prospects.

According to Lee et al. [17] , task value is a positive predictor of self-regulated learning. Similarly, Joo, Lim and Kim [18] also found that there was a relationship between the task value of the course and the self-regulated learning skills employed by students. These findings provide teachers and instructional designers with new insights into how to design courses to support students’ self-regulated learning in terms of self-efficacy and task value.

The literatures suggest that task value and self-efficacy should be positively correlated. As Li and Zheng’s [20] study showed, task value may be affected by students’ self-efficacy. In other words, self-efficacy may not only directly affect students’ self-regulated learning, but also indirectly influence them via task value. For instance, Arian Far et al. [21] surveyed 320 undergraduates in 2018-19 academic year at Payame Noor University of Tehranfrom, and pointed out that task values played a mediating role between self-efficacy and self-regulated learning. In addition, Wang, Shannon and Ross [22] also identified the mediation effect of task value on the relationship between self-efficacy and self-regulated learning. Similarly, Cerezo et al. [23] demonstrated that the impact of self-efficacy on self-regulated learning may be mediated by task value. These existing studies imply that task value may function as a mediator between self-efficacy and self-regulated learning.

However, as educational researchers have pointed out, research on self-regulated learning has mostly focused on students in public universities [12] , while the self-regulated learning abilities of students in private colleges still need to be explored. In addition, most studies only focus on the relationship between self-efficacy and self-regulated learning, or between task value and self-regulated learning, and the relationship between these three needs to be further clarified among Chinese private college students [20] .

In order to fill some existing research gaps, this study will investigate the following three questions:

RQ1: Does self-efficacy have an impact on self-regulated learning among students in private colleges?

RQ2: Does task value have an impact on self-regulated learning among students in private colleges?

RQ3: Does task value mediate the self-efficacy and self-regulated learning among students in private colleges?

This study was conducted at a private college in southern China. The college has been ranked among the top three private universities and colleges for three consecutive years and has been conducting English teaching for many years, however, the self-regulated learning ability of students has yet to be improved. Therefore, it is urgent to carry out relevant research to enhance their abilities. The implementation of this study can effectively help students understand their self-efficacy and self-regulated learning abilities, and recognize the importance of task value in this process.

The college has more than 16,500 students and approximately 400 students majoring in English, Business English and Translation from the School of English Language. This group of students has just entered college from high school and needs to adjust their learning methods and strategies in time in order to better adapt to the new model of the college. A total of 343 freshmen were invited to participate in this research, whose ages ranged from 18 to 22 (M = 19.02, SD = 0.863). Due to the nature of English major students in China with more female students enrolled, the questionnaire respondents included 55 male students (16.03%) and 288 female students (83.96%).

The pilot questionnaire was administered in two parallel classes of the same grade in the selected college in China. The respondents answered a battery of questionnaires: The background information questionnaire (including gender and age, which are also the control variables of this study), the Self-Regulated Learning Questionnaire, the Self-Efficacy Scale, and the Task Value Scale. To ensure the adaptation of the measurement items in the target college, two professors from this college reviewed the validity of the questionnaires and provided feedback to refine the measurement items. Based on the good level of English of participants, all the research instruments were conducted in English.

The revised version of the Self-Regulated Learning Questionnaire developed by Barnard, Lan, To, Paton and Lai [24] was used to measure student’ level of self-regulated English learning. This 16-item SRLQ consisted of five sub-scales: 1) goal setting (GS), 2) environmental structuring (ES), 3) task strategies and time management (TSTM), 4) help-seeking (HS), and 5) self-evaluation (SEVA). The items were rated on a 7-point Likert scale ranging from “not at all true of me” (1 point) to “very true for me” (7 points). In this study, the reliability with Cronbach’s alpha value was 0.973, and Kaiser-Meyer Olkin (KMO) value was 0. 916 and significance probability was 0.000.

The Self-Efficacy Scale was adapted from the Motivated Strategies for Learning Questionnaire [25] primarily aiming to measure students’ self-efficacy levels in English learning. The items consisted of a seven-point Likert scale ranging from “not at all true of me” (1 point) to “very true of me” (7 point). The items were slightly modified to reflect the context of the English courses. For example, “I’m confident I can do an excellent job on the assignments and test in this course” was modified to “I’m confident I can do an excellent job on the assignments and test in the English courses.” The reliability with Cronbach’s alpha value was identified as 0.959, and KMO value was 0.847 and the significance probability was 0.000.

The Task Value Scale was also from the Motivated Strategies for Learning Questionnaire [25] , aiming to measure students’ views on task value in English learning. The items utilized a seven-point Likert scale ranging from “not at all true of me” (1 point) to “very true of me” (7 point). The items were slightly modified to reflect the context of the English courses. For example, “I’m very interested in the content area of this course” was modified to “I’m very interested in the content area of the English courses.” In this study, the reliability with Cronbach’s alpha value was 0.957, and KMO value was 0.867 and the significance probability was 0.000.

Exploratory factor analyses (EFA) were conducted to examine the underlying factor structure of the whole scale. The Bartlett’s spherical test provided a significant chi-square value of 673.156 (p < 0.001), and KMO value was 0.868 exceeding the minimum adequacy value of 0.50, meaning that the data were suitable for structure detection [26] . Moreover, Cronbach’s alpha value of the whole scale was 0.980, which indicated a high level of internal consistency for the whole scale.

After the design was approved by the Teaching and Research Committee of the School of English Language, the battery of questionnaires was distributed online to around 12 natural intact classes of first-year students majoring in English, which yielded 343 valid questionnaires.

Based on the purpose of the research, research questions were developed to explore the relationship between the selected variables. SPSS 26.0 was used to analyze the data collected from the questionnaire. The general data was described by using mean and standard deviation. Pearson correlation analysis and linear regression analysis were used to determine the relationship between self-efficacy, task value and self-regulated learning in terms of goal setting, environmental structuring, task strategies and time management, help-seeking and self-evaluation. Finally, the data were subjected to AMOS 26.0 using structural equation models (SEM) to investigate the structural relationships between self-efficacy, task value and self-regulated learning. The following model fit indices were used for evaluating the model fit [27] : the Chi-square statistic (χ2) and its Degrees of Freedom (df), along with the associated p-value; the Comparative Fit Index (CFI) (a value equal to or greater than 0.90 indicates acceptable model fit); Non-Normed Fit Index (NNFI) (a value equal to or greater than 0.90 indicates acceptable model fit); the Root Mean Square Error of Approximation (RMSEA) (a value between 0.05 - 0.08 indicates good fit); and the Standardized Root Mean-square Residual (SRMR) (a value less than 0.05 indicates good fit).

According to the analysis of reliability and validity on the research variables, it could be seen from Table 1 that the Cronbach’s alpha value of all variables was greater than 0.8, indicating that the questionnaire had good reliability. As illustrated in Table 1, the factor loading values of all variables were greater than 0.6, the Average Variance Extracted (AVE) values of all variables were greater than 0.5, and the Composite Reliability (CR) values of all variables were greater than

0.8, indicating that the questionnaire had a good convergent validity. In addition, discriminant validity was used to test the degree of difference between a given aspect as well as other aspects where the observed aspects could be distinguished from each other. As shown in Table 2, the AVE arithmetic square root value of the diagonal was much higher than the value of the correlation factor in the same column, indicating that each potential can distinguish the variable from others with high discriminant validity.

A Pearson’s correlation was run to determine the relationships among self-efficacy, task value, and self-regulated learning. The statistical findings shown in Table 3 supported that task value and self-efficacy had a significant and positive relationship with self-regulated learning (r = 0.861, r = 0.855, p < 0.01). In particular, the five dimensions of self-regulated learning were positively correlated with self-efficacy and task value (p < 0.01). According to Table 3, moreover, self-efficacy also had a significant and positive correlation with task value (r = 0.858, p < 0.01).

In order to eliminate the influence of other irrelevant variables, this study divided the predictors into three groups for linear regression analysis. The first group was demographic, including gender and age; the second group was self-efficacy; and the third group was task value. From the regression analysis results in Table 4, it can be seen that age and gender had no predictive effect on self-regulated learning (p > 0.05), so the effects of age and gender on SRL can be excluded. Moreover, self-efficacy and task value had significant predictive effects on self-regulated learning (p < 0.05), which could explain 79.1% of the variance of SRL. This illustrated that with the improvement of students’ self-efficacy, their self-regulated learning ability would also be enhanced, and students who were interested in the task and understood its value also had higher level of self-regulated learning.

a. Note: F1 = Goal setting; F2 = Environmental structuring; F3 = Task strategies and time management; F4 = Help-seeking; F5 = Self-evaluation; F6 = Task value; F7 = Self-efficacy.

a. Note: N = 343; * p < 0.05 ** p < 0.01.

a. Note: Dependent Variable: self-regulated learning; * p < 0.05 ** p < 0.01.

The structural equation model analysis was performed using AMOS 26.0 software to validate the proposed research questions. After analyzing the sample data using the maximum likelihood estimation method, the fit indices of the model were obtained and the standardized regression coefficients (path coefficients) and significance of each path were calculated. The results of the structural equation model SEM constructed with AMOS showed that the chi-square was 702.015 and the degrees of freedom were 244. In this study, the χ2/df value was 2.877 (χ2 = 702.015; df = 244); this meant that the test model was well-fitted (2.877 < 3.00). Model fit indices (SRMR, RMSEA, CFI, IFI, NFI, and NNFI) exceeded acceptable levels (SRMR = 0.040 < 0.05, RMSEA = 0.074 < 0.08, CFI = 0.943 > 0.90, IFI = 0.944 > 0.90, NFI = 0.916 > 0.90, NNFI = 0.936 > 0.90).

Based on the validation information, the results demonstrated that the model had a mediation effect. The details of the structural relationships in the SEM are shown in Figure 1.

The structural equation-AMOS model path analysis results demonstrated that self-efficacy and task value were both statistically significantly associated with self-regulated learning (β = 0.468, p = 0.003 < 0.01, 95% CI [0.186, 0.710]; β = 0.489, p = 0.001 < 0.01, 95% CI [0.254, 0.761]), and the path from self-efficacy to task value was also statistically significant (β = 0.921, p = 0.000 < 0.01, 95% CI [0.876, 0.956]). The mediating effect of task value was tested using bootstrapping approach. In addition, the indirect effect of self-efficacy on self-regulated learning via task value was equal to the product of the coefficients for each of the paths in the mediation chains (0.921 * 0.489 = 0.450). The 95% bias-corrected confidence interval for the mediated effect was between 0.239 and 0.715 did not include 0, with a p-value at 0.001 for the two-tailed significance test and the standard error at 0.120. These data illustrated the existence of intermediary effect. Moreover, the direct effect between self-efficacy and self-regulated learning was also significant (β = 0.468, p < 0.01). The 95% bias-corrected confidence interval for the mediated effect was between 0.186 and 0.710 did not include 0. Therefore, the effect of self-efficacy on self-regulated learning was partially mediated by task value. The total effect of self-efficacy on self-regulated learning was 0.450 + 0.468 = 0.918, and the proportion of mediating effect to total effect was 49.02%.

This study would increase understanding of the relationship among self-regulated learning, task value and self-efficacy, and also provide feasible suggestions for the development of students’ self-regulated learning in selected colleges in China. The results of this study indicated a positive correlation among self-efficacy, task value and self-regulated learning. In addition, the SEM model which explained how self-efficacy and task value affect self-regulated learning was a partial mediating model: Self-efficacy affected self-regulated learning directly as well as indirectly through the mediating effects of task value. In other words, task value played a partial mediating role between self-efficacy and self-regulated learning.

Correlation analysis found that self-efficacy had a significant positive correlation with self-regulated learning, and regression analysis also found that self-efficacy had a strong positive predictive effect on self-regulated learning. Thus, students with higher self-efficacy have stronger self-regulated ability in learning. These findings are consistent with previous studies. The stronger the sense of self-efficacy, the higher the challenge that students set for their goals, and the better they can conduct self-regulated learning, such as rationally organizing learning tasks and adjusting learning strategies. However, those with low self-efficacy will avoid making choices that exceed their own abilities, and they will adopt evasive attitudes when face with challenges. It can be seen that self-efficacy is an important factor affecting self-regulated learning. Therefore, self-efficacy can be used as a significant predictor to promote students’ self-regulated learning and finally improve their learning outcomes.

In addition, task value played a mediating role in the effect of self-efficacy on self-regulated learning. From the above research results, it can be shown that self-efficacy had a direct positive effect of 0.468 on self-regulated learning, accounting for 50.98% of the total effect, and 49.2% of the indirect effect was through task value. In order to pursue knowledge and interests, students with high self-efficacy generally have stronger intrinsic motivation and higher self-regulated learning ability. At the same time, they pay more attention to the intrinsic value of the task, which further enhances the ability of self-regulated learning.

The study offers students and practitioners in the second language acquisition and related disciplines important insights into identifying research topics, variables, and methodological approaches.

Firstly, although previous studies have examined sub-dimensions of public university students’ self-regulated learning [28] [29] , this study further illustrates that self-regulated learning is also affected by such five dimensions, including goal setting, environmental structuring, task strategies and time management, help-seeking and self-evaluation among private college students. Students with a higher level of self-regulated learning pay more attention to goal setting and will monitor, adjust, and self-evaluate their own learning process. Thus, the five sub-variables can be identified to measure self-regulated learning.

Secondly, existing literature generally studies the relationship between self-efficacy and self-regulated learning [30] [31] , and relatively few studies have explored the effect of self-efficacy on self-regulated learning via task value. On this basis, this study investigates the role of task value between self-efficacy and self-regulated learning through the construction of SEM model. The results demonstrate that self-efficacy can directly affect self-regulated learning, and can also affect self-regulated learning through task value. By improving students’ self-efficacy and helping them understand the value of tasks, students’ self-regulated learning ability can be effectively improved. Therefore, learners should actively motivate themselves and use appropriate learning strategies in the language learning process.

The results of this study provide support for how to improve self-regulated learning ability.

Firstly, students’ self-efficacy determines their choice in a given environment, that is, the possibility of choosing simple or complex tasks [32] . Students usually avoid activities and situations that are beyond their ability to cope and tend to choose what they feel they can handle [33] . Different choices lead to different abilities and experiences of students and profoundly affect the direction of personal development. Therefore, the higher a student’s self-efficacy is, the more he can challenge high goals, and thus exhibit a higher self-regulated learning ability.

Secondly, self-regulated learning is of great significance to individual development now and in the future. It is not only the need to cultivate students’ comprehensive ability but also the basis of lifelong development. In the process of learning, students should change passive acceptance into active participation, fully mobilize their enthusiasm and play a subjective role [34] .