Investigation of Chinese University Students’ Attributions of English Language Learning

Despite the importance of developing students’ learning autonomy in Chinese schools similar to Western cultured schools, many concerns are raised regarding the influence and effectiveness that learner autonomy has on students’ academic achievements. The aim of this study was to identify the attribution patterns of Chinese university students for success and failure toward students who learnt through autonomy learning (student-centered approaches) compared with students who learnt through teacher-centered approaches. Within this study, mixed research methods were adopted, and students used a reflective method to distinguish whether they were taught English through a traditional or student-centered method. The findings of the study reveal that there are no significant differences in attributional patterns between students who had learnt in high school through autonomous learning and those who learnt through teacher-centered approaches. The findings have implications for policy and practice in the Chinese Ministry of Education system and recommendations for future research.


Learner Autonomy
Developing students' learning autonomy has been accentuated by the Chinese Ministry of Education (MOE) since 2005 in the College English Curriculum. A large number of studies have focused on this in the last 20 years in China, especially in recent years; however, not many scholars have shed light on Chinese students in high schools from their perspectives in mainland China.
As requested by the MOE (2007), a major aim of English teaching in high schools was to cultivate students' learning autonomy. However, in Chinese high schools, students were reported as holding a passive attitude toward autonomous learning and having a lack of learning objectives and aims (Shao, 2012). This situation was also found in Peng et al.'s (2014) study where she claimed that in Chinese rural areas, a deficiency of teaching materials and qualified teachers, promoting students' all-round development seems rather difficult than in urban areas. Although it is not easy for all students and teachers to adapt to these requirements, a recent outline from the National Plan for Medium-and Long-Term Education Reform andDevelopment (2010-2020) reemphasizes that a better life needs to be based on "citizens' fundamental education so as to improve citizens' quality and to promote their all-round development" (MOE, 2010, p. 5). From this point of view, enhancing students' learning autonomy is not only beneficial for their quality of education but also for lifelong learning and citizens' quality in the fundamental education program.
In the last several decades, a concern with the nature and benefits of learner autonomy has been well established in the literature (Benson, 2012;Benson & Cooker, 2013;Brookes & Grundy, 1988;Dicksinson, 1992;Ellis, 2008;Harding-Esch, 1977;Holec, 1981Holec, , 1988Little, 2000;Little & Dam, 1998;Riley & Zoppis, 1985;Wenden, 1991;Willing, 1989). In previous studies, learner autonomy is also interchangeably regarded as self-directed learning and independent learning although their definitions are slightly different. Learner autonomy is generally regarded as a defining characteristic of all sustained learning that attains long-term success (Little, 1996); however, autonomy has been described and defined in a number of ways in connection with language learning. In language learning, the most often quoted definition is that of Holec (1981), who defines autonomy as "the ability to take charge of one's own learning" (p. 3). To take charge of one's own learning is to have, and to hold, the responsibility for all decisions concerning all aspects of this learning: • determining the objectives; • defining the contents and progressions; • selecting methods and techniques to be used; • monitoring the procedures of acquisition . . .; and • evaluating what has been acquired. (Holec, 1981) Learner autonomy is seen as an issue principally of students taking great control over the content and methods of learning (Holec, 1981). It grows out of the individual's acceptance of his or her own responsibility for learning. It presupposes a positive attitude to the purpose, content, and process of learning (Little, 2000). The learner is perceived as a decision-maker who has or will develop his or her capacity for selecting from available tools and resources to create what is needed for the task in hand (Dicksinson, 1992;Holec, 1985;Little, 2000). Therefore, developing positive attitudes toward learner autonomy and the necessary skills are regarded as crucial to the success of the development of learner autonomy, and is an essential goal of any course.
Compared with Western scholars' interest in learning autonomy, most papers on learning autonomy in mainland China have focused on definitions and literacy reviews, which used qualitative methodologies in the early 20th century (Gao, 2005). Liu (1991) points out that self-directed learning is lifelong, which explores four aspects, namely, the definition of selfdirected learning, the learners' identification, the learning theory, and the practice on the development of self-directed learning. This study was regarded as the first one that advocated self-access learning in mainland China (Gao, 2005). Li (1998) explains the importance of cultivating communicative competence in intercultural environments within the Chinese context. Li's (1998) discussions are based on the previous reviews (Allwright, 1988), and examined the extent of learner freedom in the Chinese context and the important factors affecting the development of learner autonomy in universities of mainland China (Walther, 2002;Wood & Smith, 2001). Li also suggests that setting up selfaccess-learning centers in universities is essential for language learning in China. Zheng (2000) claims that differences occur in cultural and social backgrounds between Western countries and China. As a consequence, Chinese educators could not totally accept the concept of "learning autonomy" without critical thinking in the Chinese educational context. Hua (2001a) elaborates the role of teachers and students in the language classroom and that the focus should be shifted from a "teacher-centered" to a "student-centered" classroom, which is beneficial for learners to develop autonomy. In the same year, Hua (2001b) introduced the rationale for self-access-learning centers and learner autonomy, and made an analysis of the major components in constructing a self-access center and related issues. Furthermore, she analyzed the practical need for setting up self-access-learning centers in Chinese universities and colleges. Other universities' experiences of utilizing self-access-learning centers in English as a Second Language (ESL) learning were also introduced in the article. Gao (2005) gives an explicit analysis on the development of learner autonomy in mainland China from 1991 to 2003. He categorized the previous reviews on learning autonomy from 1991 to 2003 in China according to the publication of journals, research participants, content of published journals, and research methodologies in papers. The findings of his study showed that (a) there was a steady increase in the number of research articles on learning autonomy, (b) special research emphasizes on non-English major students, (c) main areas of focus were introductions of definition of autonomous learning in Western countries and description of ways to develop learner autonomy in the Chinese context, and (d) the research method was non-empirical.
Overall, these researchers hold a positive view on developing learning autonomy in the Chinese context. However, as the origin of the learning autonomy is from Western countries, these papers do not provide explicit methods on how to cultivate students' learning interest by setting up the accesslearning centers and how to make it adapt to the current English curriculum designs. In addition, these Chinese researchers were still on the stage of introducing the theories on the development of learning autonomy. Few experimental studies have been undertaken to show evidence as to whether advocating Chinese students to utilize self-access-learning centers could improve their English ability or not. Hence, this empirical study aims to fill the gap in the previous research studies.

Learner Attribution
An important way to understand students' motivation about learning and achievement is through the lens of attribution theory. Attribution theorists posit that "individuals seek to understand why events have occurred" (Schuster, Forsterlung, & Weiner, 1989, p. 192). An attribution refers to "constructions imposed by perceivers to account for the relation between an action and an outcome" (Weiner, 1986, p. 22). Weiner's (1979Weiner's ( , 1986 attribution theory of motivation is one of the most cited theories for explaining an individual's affect and behavior in academic-related events (Linnenbrink & Pintrich, 2002;Salili, Chiu, & Hong, 2001;Tollefson, 2000).
In academic-related contexts, ability, effort, task difficulty, and luck are perceived as the major responsible causes for success and failure, among which ability and effort are the most dominant causes (Tollefson, 2000;Weiner, 1979;Weiner & Kukla, 1970). Weiner (1985Weiner ( , 1986 claims that to understand an individual causal belief, it is necessary to distinguish the properties of causes. He identifies three dimensions for characterizing the causes, namely, locus of causality, stability, and controllability. Locus of causality refers to the location of a cause, distinguishing whether a cause is internal or external to an individual. Stability refers to the duration of a cause, distinguishing whether a cause is constant or temporary. Controllability differentiates whether a cause is subject to an individual's volitional control. Collectively, a cause can be located in one of the eight traits, that is, two levels of locus of causality by two levels of stability by two levels of controllability (see Table 1). For instance, if a student ascribes a test success to sufficient effort, the underlying causal belief about this outcome is internal, unstable, and controllable.
The causal properties play a key role in a person's emotional consequences and expectancy that construct motivation (Weiner, 1986;Weiner, Russell, & Lerman, 1978). Locus of causality determines self-esteem and affects pride. Attribution of success to an internal cause (e.g., high ability) is more likely to increase self-esteem and the feeling of pride than is attribution of success to an external cause (e.g., easy task). Stability links to feelings of helplessness and adjusts expectancy of future outcomes. Attributing failure to a stable cause is more likely to generate the feeling of helplessness than applying to an unstable cause. Furthermore, ascribing to stable causes has much to do with increment or decrement of expectation of future success, compared with unstable causes. Finally, controllability predicts social emotions (e.g., shame, guilt, pity, and anger) that represent intra-and interpersonal judgments (Weiner, 2001). Intrapersonal judgment refers to self-directed emotions and behaviors, whereas interpersonal judgment refers to other-directed (i.e., an observer or a stakeholder) emotions and reactions to the other's performance. For example, ascribing failure to a controllable cause by an actor would likely generate guilt, whereas ascribing to an uncontrollable cause would generate feelings of shame and embarrassment. At the same time, a controllable attribution perceived by others may follow by feeling of anger and blaming behavior, whereas an uncontrollable attribution may generate sympathy and helping behavior from others.
It is suggested that students generally tend to attribute success to internal causes and failure to external (e.g., task difficulty), or internal and unstable causes (e.g., effort; Reyna, 2000;Tollefson, 2000). This is referred to as "the normal self-esteem attribution pattern" (Jacobsen, Lowery, & DuCette, 1986, p. 63). Students who attribute internal causes for academic success experience a greater level of reward and exert higher levels of goal-attaining behavior than do those who attribute to external causes. At the same time, attributing failure to lack of effort, or bad luck protects student self-esteem, and maintains the motivation and expectation for future success. In the case where low effort is ascribed, students may further elicit more effort in future tasks. Such an attributional pattern is considered positive for academic learning by the literature (e.g., Linnenbrink & Pintrich, 2002;Perry & Penner, 1990;Weiner, 1986;Weiner & Kukla, 1970).
On the contrary, abnormal attribution patterns also have been found in students with learning difficulties or disabilities (Jacobsen et al., 1986;Tabassam & Grainger, 2002;Woodcock & Vialle, 2011). These students tend to attribute success to external causes, and failure to internal and stable causes. According to Weiner's (1979) theory, external attribution for success reduces positive effects such as pleasure, satisfaction, and happiness. Similarly, internal and stable attribution for failure increases the negative affect such as being upset, displeased, and worried. Students who foster the patterns are likely to have lower self-esteem, self-image, and self-efficacy.

Attribution in Learner Autonomy
As discussed, the central tenet of autonomy is that a learner takes charge of his or her own learning (Holec, 1981). Thus, it is essential for learners to foster a belief of reasonability, meaning that a learner draws on intrinsic motivation for controlling failure and success in learning (Dickinson, 1995). Such a cognitive process can be informed by Weiner's (1979Weiner's ( , 1985 attribution theory of motivation. The cause that best represents learner reasonability is effort, whereby the locus of causality is with the learner (internal) and also under his or her control (controllable).With effort attribution, a learning success enhances motivation and stimulates more effort (unstable) for taking more challenging tasks. By analogy, a learning failure can still motivate the learner, if he or she believes that eliciting sufficient effort (unstable) would have resulted in a positive outcome.
Effort attribution is emphasized in the area of language learning autonomy (Ma & Ma, 2012;Spratt, Humphreys, & Chan, 2002). Dickinson (1995) in his review of research on autonomy in relation to motivation concludes that Task difficulty, in some situations, can also be considered as unstable. Weiner (1983Weiner ( , 1985 suggests that ease or difficulty of task is stable when the same or a similar task will be encountered in future. In other situations, such as a task is changing, the factor is unstable. . . . learning success and enhanced motivation is conditional on learners taking responsibility for their own learning, being able to control their own learning and perceiving that their learning successes or failures are to be attributed to their own efforts and strategies rather than to factors outside of their control. Each of these conditions is a characteristic of learner autonomy as it is described in applied linguistics. (p. 174) Littlewood's (1999) study that focused on East Asian leaners illustrated the "belief in effort" as follows: • innate ability does not determine how much success a person can achieve; • with effort and self-discipline, every person can achieve his or her goals; and, • failure can be retrieved by making more effort. (p. 82) The researcher proposed that effort attribution is a determinant of effective language learning in East Asian contexts. Moreover, he argued that such a belief is a key of academic success of East Asian learners who are studying in Western countries.

Attribution in English as a Foreign Language (EFL) in China
A number of studies (e.g., Mao, 2003;Wang, 2005) on high school students show that both self-reported successful and unsuccessful learners attribute internal causes (e.g., interest, effort, strategy, ability) for their current academic status. According to Weiner's (1985) attribution theory, internal attributional causes for successful and failed outcomes are associated with self-esteem. A student who associates a successful outcome to the self is likely to increase self-esteem and motivation. However, a student who attributes a failure to stable and internal causes may experience negative selfesteem and related affects. Thus, the studies indicate that the attribution pattern fostered by successful students is more positive than by unsuccessful students.
Chen's (Chen, L.-Y., 2011) investigation on learner attribution to academic success and failure further supports this conclusion. Given an academic success, the self-reported successful learners attribute more internally, whereas the unsuccessful counterparts attribute more externally. Given a failure, the former cohort mainly ascribes to internal and unstable causes (e.g., effort) and external causes (e.g., classroom environment, teacher instruction), whereas the latter cohort mainly ascribes to internal and stable causes (e.g., ability, interest) and external causes.
More relevant research has been conducted in the tertiary education. The literature (e.g., R.-Y. He & Li, 2010;Lei & Qin, 2009) that examines the general attribution of success and failure in EFL indicates that college students adapt to a mixture of internal and external attribution patterns for success and failure. Effort is ascribed as the most important cause for both outcomes. In addition, some external factors have also been reported as important causes. Causes that are related to teaching, including teacher input and the current EFL teaching model, are major external causes rated by students. In addition, task difficulty is a major external cause ascribed for success. The findings suggest that college students consider EFL learning outcomes in both internal and external manners.
The literature (e.g., R.-Y. He & Li, 2010;Hu, Shi, & Zhou, 2009;Zhang, 2011) that focuses on successful and unsuccessful learners shows distinctions between the groups. Successful learners tend to attribute successful outcomes to internal and unstable causes (e.g., effort, attention, revision) more than internal and stable (e.g., ability, interest) and external causes (e.g., teacher input, classroom environment). Their attribution for failure outcomes is effortoriented. The findings suggest that successful learners have developed positive attribution patterns. In contrast, unsuccessful learners ascribe success outcomes more externally (e.g., task difficulty, teacher input), whereas the group ascribes failed outcomes more internally, including lack of effort and stable causes (e.g., low ability, low interest). Thus, unsuccessful learners are likely to develop negative attributional patterns for both success and failure outcomes.
Taken as a whole, the research on learner attribution of EFL in China supports the pattern of effort attribution in general. Furthermore, college learners are more effort driven than their counterparts in high schools.
In addition to effort attribution, college learners tend to attribute to external causes, whereas high school learners tend to attribute to internal and stable causes. This means that high school learners are more likely to perceive themselves as being responsible for their academic outcomes. Thus, in comparison with college learners, they experience more intrinsic rewards in successful situations. However, they elicit more pressure and negative affects while they are in failing situations.
Attribution patterns adapted by college learners are sophisticated and reveal somewhat ambivalence. On one hand, they believe in the importance of effort. In particular, they believe that sufficient effort helps avoiding failure. On the other hand, they perceive that high effort is not sufficient to assure success. They are also aware of varied external factors, among which teachers and instructional factors are dominant. This reveals that college learners are more likely to take responsibility for failure than success. Such an attribution pattern may result in only "pass-oriented" learning. This means that the students are motivated to pass the minimal academic requirement, even if they have had a failure. Nonetheless, they are less motivated and self-determined to pursue more future success when external factors are not supportive.
Attributional theories suggest that a person's attributional belief builds on his or her previous experiences (Schuster et al., 1989). In this sense, previous EFL learning experiences influence college students' attribution about their current learning. However, little research has been done to investigate whether college students who have gone through autonomous learning while they were studying in high schools have fostered a better attributional belief than those who had not in EFL learning. The present study aimed to examine the attribution patterns of success and failure of each group concerned with the four main factors, namely, effort, ability, task difficulty, and luck. Specifically, it examined three hypotheses.
Hypothesis 1: The attribution patterns of college students who had or had not gone through autonomous learning in high schools are positive. Hypothesis 2: The students who had gone through autonomous learning in high schools have developed a better attributional belief than their non-autonomous learners in success. Hypothesis 3: The students who had gone through autonomous learning in high schools have developed a better attributional belief than the non-autonomous learners in failure.

Method
Mixed research methods were utilized in this project. Using mixed research methods can help bridge the schism between the qualitative and quantitative research (Johnson, Onwuegbuzie, & Turner, 2007;Onwuegbuzie, Slate, Leech, & Collins, 2009). Thus, I believe that using mixed research methods in this research could obtain more depth of information than using qualitative or quantitative methodology solely in a research study.
Also, a reflective research method was utilized in the questionnaire to gather the participants' views and perspectives. Particularly, using this method to ask the participants to choose the two groups of teaching models (studentcentered and teacher-centered) enables to enhance students' critical thinking process (Fry, Ketteridge, & Marshall, 2009)

Contextual Information
The participating university is one of the high-ranking universities in the southern part of China. The students enrolled within this university needed to finish their secondary education and achieved a high level in the College Entrance Examination. The participants involved in this study were full-time students on the main campus of the university.

Participants
A total of 100 university students took part in the pilot study, and 95 students finally handed in their questionnaires. The participants who joined in the pilot study were not included in the final study. For the final study, there were 500 university students involved in the survey and 347 students finally completed the questionnaire, which yielded a 69.4% response rate. Within these 347 participants, there were 163 male students and 184 female students who came from four main disciplines: arts and social science (N = 102), science (N = 113), engineering and technology (N = 43), and fine art and physical science (N = 67). The number of participants who experienced teacher-centered teaching model (N = 71) was three times more than that who experienced student-centered teaching model (N = 276). The ages ranged from 18 years old to 20 years old, and their average level of English in the College Entrance Examination was between 100 and 110. The details of the two groups of the participants and their majors were shown in Figures 1 and 2: In the final study, there were 10 students who volunteered to take part in the semi-structured interviews. The participants were also from the main disciplines: arts and social science (N = 3), science (N = 2), engineering and technology (N = 4), and fine art and physical science (N = 1). Half of them were female and the number of students who experienced student-centered (N = 8) were extraordinary more than those (N = 2) who experienced traditional teaching in English.
The semi-structured interview was undertaken on campus, and the total time for each student was 15 to 25 min. Within this study, all the participants took face-to-face interviews.

Instrumental Design
The research study was divided into two stages: quantitative stage and qualitative stage. At the first stage, the questionnaire was utilized to gather students' attributional responses of English learning outcomes. The questionnaire consisted of three sections, and its design was on the basis of (Weiner, Graham, & Stern, 1982) attributional theory. The first section was designed to collect participants' background information to see if any of the independent variables would affect the dependent variables in the data analysis process. The second section contained 16 scenarios, each describing a specific task and outcome. All of the tasks are common English learning tasks for college students. Below a scenario, four causes (i.e., ability, effort, task difficulty, and luck) that accounted for the outcomes were presented. The details of this section are shown in the appendix as an example of the questionnaire. Participants were asked to provide feedback on each cause on a 5-point Likert-type scale (Likert, 1932). For example, the first scenario was designed to ask students' reading ability. To understand participants' attribution in their reading ability, Question 1 was designed to ask the attribution of participants' poor reading ability, ranging from "Strongly Disagree" = 1 to "Strongly Agree" = 5.
The semi-structured interviews were utilized at the second stage to gather students' perceptions and views regarding their English learning experiences by using different teaching models. Compared with the data obtained from the questionnaires, this form of data is textural and allows researchers to have further understandings regarding students' perspectives. According to the research aim and the hypotheses, the interview questions were designed based on the participants' prior learning experience in China, and their responses were coded via NVivo 10.
Prior to the final study, the researcher invited three academic staff from the research field and three university students from various disciplines with different learning backgrounds to provide recommendations on the initial design of the interview questions in the final version. This process also ensured its reliability and validity.

Procedure
After obtaining the ethic approval, the researchers contacted the International Office of the subject university via emails. The information sheet for the university head and for the university students had been sent directly to the head of the International Office of the participating university.
Both the questionnaire items and the semi-structured interview questions were used in the pilot study to examine validity and credibility. After the pilot study, three expert academic staff from the research field and two students were invited to ask for their opinions regarding the design of the questionnaire items and the interview questions. This included the instrumental design, lexical usage, and the item content. They provided suggestive advice, so that the researchers could make slight changes in the final design of the questionnaire and the interview questions.

Results
In the pilot study, the SPSS Version 21 was adopted to ensure the validity and reliability. The reliability of the 100 question items was examined using Alpha reliability. According to Pallant (2011), internal consistency was adopted as the most frequent indicator to ensure that all the items under the same scale measure the same attributive factor. As such, Cronbach's coefficient alpha was utilized in this study.
The reliability analysis showed that Cronbach's coefficient alpha was .829. According to Nunnally (1978), if the coefficient value is above .8, it indicates that the instrument has a high inner consistency. From this point of view, the instrument adopted in this study was very reliable.
Paired samples t tests were carried out to examine learners from autonomous and teacher-centered learning backgrounds with regard to their attributional patterns toward success and failure. Furthermore, independent samples t tests were carried out to examine any differences that may occur between autonomous and teacher-centered learners. The results from the study are first presented by analyzing the attributional patterns toward success and failure for autonomous and teacher-centered learners with regard to learning English. Comparisons between the autonomous and teachercentered learners will then be shown.
The paired samples t test that compared the causes for success between learners whose prior experience was through autonomous learning and those whose prior experience was through teacher-centered learning shows that there were no significant differences between them (p < .05). Thus, the greatest causes for success were the same between both cohorts of learners.
The paired samples t test that compared the causes for failure between learners whose prior experience was through autonomous learning and those whose prior experience was through teacher-centered learning shows that there were no significant differences between them (p < .05). Thus, the greatest causes for failure were the same between both cohorts of learners. The qualitative data were analyzed using NVivo 10. The result of the textual data was very similar to the quantitative data. Most participants believed that their effort, luck, and task difficulty were three important causes of success in English learning compared with the factor of ability. One of the participants claims that he did not believe the ability as an important factor to lead to his English learning success because of the current design of examination papers and English curriculum. "Most of the question items in the English final examination paper were in the form of multiple choices. Sometimes, students who were not good at English were able to guess and get a right answer." In this case, it is very easy to understand the reason why most participants still believe luck and task difficulty are important in their success.
From the textual data, students who experienced a student-centered approach held a more positive attitude toward effort, which has a more important role in their failure in their English learning. The participants reflected that their failure experiences and all of them contributed the causes to their lack of effort rather than other three factors.

College Student Attribution in EFL
As shown in Tables 2 and 3, both cohorts developed the same attribution patterns for failure and success. Effort attribution was significantly more important than other causes in failure, indicating that college students tend to feel responsible for the outcomes. According to Weiner (1985), ascription of failure to low effort elicits feelings of guilt, but the learners may maintain the expectation for future success. It is also likely that they will put forth more effort for a better outcome.
As for success, effort, task difficulty, and luck were significantly more important than ability. Moreover, there was no significant difference among the three factors. This demonstrates that college students interpret successful outcomes in a complex manner. First, success is likely to be seen as caused by unstable factors. Task difficulty, in the present study, can also be seen as an unstable cause. It is because general EFL learning was represented by 16 different tasks, each of which is related with performance of a dominant skill (e.g., writing, speaking). Thus, the EFL learning, in general, is a diverse and unstable task. As suggested by Weiner's theory, ascribing success to unstable causes may not result in an increment and could lead to a decrement of expectancy of future success.
Second, success is likely to be attributed externally. The positive attribution or the normal self-esteem attribution pattern demonstrates that the more internal attribution the cause, the more positive the experience, and the higher self-esteem and self-image he or she will develop. Eventually, this will increase a learner's confidence in taking more challenging tasks. In the current case, the learners might have experienced less positive outcomes and were less confident in future success.
Collectively, the present study suggests that college learners have fostered a positive attribution pattern for failure and a less positive pattern for success. Effort attribution is evident for both situations, and in particular, the failed situation. A learning failure may trigger more effort expenditure in them for avoiding repeated failures. However, success might not stabilize their confidence for more success. Such a pattern is coincident with the pattern suggested by the literature that focuses on college learners (R.-Y. He & Li, 2010;Lei & Qin, 2009). This also can be supported by the result of their expectation for final-term exams (see Figure 3). Most of the learners (91.9%) did not expect to fail. However, two thirds of learners only expected the minimal pass mark.

Attribution in Learners' Experienced Autonomous Learning and Teacher-Centered Learning
The data do not support the hypotheses that college learners who had gone through autonomous EFL learning developed a better attribution pattern than their counterparts. Both cohorts did not differ in their attribution patterns of success and failure.
As suggested by the literature, learner autonomy is conditioned on learner responsibility of his or her own learning. The higher learner responsibility the learner has, the more internal attribution he or she adapts. However, in the current study, the cohort with autonomous learning experience did not see themselves solely responsible for successful outcomes because they ascribed to both internal and external causes. This might be caused by various factors, such as the educational system, English curriculum designs, cultural influence, and English assessments (Guo, 2011;Throssell & Zhao, 2011) Although the cohort tended to be more internal in the ascription of failure, it may be influenced by cultural belief in effort rather than the past learning experiences. It can be seen from the attribution pattern of the cohort with traditional learning experience, which was responsive to effort attribution as well. The literature has constantly reported that Chinese students are effort-oriented in academic-related contexts (Yan & Gaier, 1994). Effort is viewed as the most important factor of academic achievement in Chinese culture (Crittenden, 1996;Stevenson & Lee, 1996). Littlewood (1999) points out that such a belief is rooted in Confucius. Proverbs such as bènniǎoxiānfēi (a slow sparrow should make an early start) and qínnénbǔzhuó (practice makes perfect) are for encouraging people to put forth efforts, particularly for those who show less aptitude.
In conclusion, it may be reasonable to assume that although the former cohort had been guided to form learner autonomy in their secondary education, it has not fostered efficacious learner autonomy. However, although the latter cohort had been instructed under the traditional EFL teaching, it reveals signs of learner autonomy due to the cultural influence.

Recommendations and Implications
The rich data from this research study highlighted that there was a huge gap of teaching model utilized by teachers between high schools and the university. At present, most English classrooms are still "teacher-centered" in Chinese high schools, and students heavily rely on their learning process with their teachers. This might be strongly influenced by the current College Entrance Examination in the educational system.
The College Entrance Examination plays a vital role in the current educational system. That is, the higher the scores students gain, the more likely that they would be accepted to a high-ranking university in China. As a consequence, it is highly likely that students' learning is in a "test-based" process instead of "self-control," and also teachers would like to teach students based on the textbooks. In high schools, students and teachers both tried hard to realize their common goal-The College Entrance Examination. As such, it is more likely that when students failed in examinations, they would reflect and ascribe the failure to themselves instead of teachers. English tests in the College Entrance Examination are in a similar case. That is, if students could gain a higher score, they could choose their majors in a priority. Thus, English subjects have a domain position in Chinese high schools.
However, when students were enrolled in universities, they gradually found that English subjects were not as important as those in high schools because they have their own interest in their own majors, particularly for those majoring in fine art and physical science. English was a selective unit for them rather than a compulsory subject. As such, they had more time to focus on their own major study. For other students, they enjoyed being opposed to an unrestricted environment. In this open environment, students become the central part in the classrooms instead of their teachers and they could access much more resources on campus than that in high schools. In the learning process, they have to be more independent rather than relying their success on their teachers as guidance is not available for them all the time. If they have difficulties, they need to seek for help from friends at first. Under this situation, when they face success, they might believe that their luck could be more influential than their ability.
Another factor is likely to be a Confucius cultural influence on a "teacher-studentship." This is a Chinese-specific culture that emphasizes teachers' authority in a classroom (Peng et al., 2014;Throssell & Zhao, 2011). Their relationship is like Chinese "Fu" and "Zi," which means that sons should obey fathers in a traditional Chinese family. When this authority is taken to the English classrooms, students need to obey teachers so as to follow their teachers' teaching in their learning process. This climate is prominently found in Chinese high schools from this study. Evidence is also found in Peng et al.'s (2014) study, who claimed that "Confucian tradition put great emphasis on the morality of teachers and this continues to be an important aspect of teacher quality" (p. 79). From their point of view, it was easy to understand the reason high school students felt both positive in their attribution in failure and success as they have shown a higher moral authority on their teachers, and also, Note. Assessments of final English examinations used widely in Chinese universities are divided into four levels (less than 60, 61-74, 75-90, and 91-100). Students who cannot obtain 60 means that they fail in examinations. Students who can finish tasks and achieve medium level in the four English skills in examinations can be assessed 61-74, which also includes their daily performance in English classes. Students who are able to get scores between 75 and 90 means that they have a medium-advanced level in the four English skills. Students who have an advanced English competence in speaking, listening, reading, and writing can be assessed 91-100.

Part 1 and Part 2 in the Questionnaire as Examples
Part 1: Demographic Information Instruction: For each question, please select ONE answer and add a "√" or "×" in the box. they believe that their teachers have exerted effort in the whole teaching process. Consequently, if they face failure and success, they would be like to ascribe to internal attributions instead of external attributions.
Based on the previous discussions, from the researchers' perspectives, implications are strongly proposed from the following three aspects: MOE (2001) has emphasized that developing students' learning autonomy is essential in the whole learning process from an early stage to adults. However, due to the economic development in various provinces, Chinese schools would have slight differences when advocating learning autonomy according to the MOE documents. It is likely that schools located in the coastal cities could have more funds to equip teaching and learning resources, while in rural areas, students have less resources. From the data analysis of this study, there were no statistically significant differences between a "teacher-centered" model and a "student-centered" model in their attribution at university. From this point of view, it reveals that future document planners need to consider the unbalanced economic development in these areas.

Implications for Practice
As students' attribution and learning autonomy is a complex psychological behavior, it needs teachers to put more emphasis on individuals instead of groups in classrooms. The researchers strongly suggest that in the near future, the size of English classes could be controlled to under 30 students, particularly in a standard tutor classroom. This size can secure that students could have equal rights to communicate with teachers. In the meantime, teachers are more likely to be able to observe each individual student's behavior. In the long term, this classroom is beneficial for students to have positive attitudes toward learning and teaching.

Implications for Future Research
This research only focused on investigating students' attributional perspectives on EFL learning outcomes rather than both students and teachers. Also, this research did not examine differences in students coming from rural areas and urban areas. These provide an opportunity for future researchers. In a direction of future research, the researcher suggests that our focus could be shifted from students' perspectives to teachers' perspective toward their teaching and learning experiences, so as to find out whether this could be a major influence on their students' attribution.

Conclusion
The aim of this study was to identify the attribution patterns of university students comparing those who learnt through autonomy learning with those who learnt through teachercentered approaches throughout their high schools. The researcher used both quantitative and qualitative methods to conduct data collection and data analysis. The results found that no significant differences in attribution patterns were found between students who had learnt in high school through autonomous learning and those who learnt through teacher-centered approaches. Part 2: The following scenarios describe English learning in university. Each scenario has been provided with four reasons. Please select the answers that best describe your own status. 1 refers to "Strongly Disagree," 2 refers to "Disagree," 3 refers to "Not sure," 4 refers to "Agree," 5 refers to "Strongly agree." 1. Your teacher says you are doing badly in reading work. It would probably be because 2. You couldn't complete a listening task in class. It would probably be because 3. You can totally understand the reading material in the exam. It would probably be because 4. You are assigned by your teacher as the group representative in group discussion. It would probably be because 5. You didn't understand an English novel. It would probably be because: 16. You couldn't finish the duty report in class. It would probably be because

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.