Results
Our aim was to explore the relationship between anxiety in learning a FL and the beliefs students have about the feedback they receive from their teachers while carrying out synchronous online speaking tasks. The differences were sought with regard to the necessity and frequency of CF; methods of CF, timing, types of errors, and choice of correctors. In the following sections, participants’ responses are reported with regard to their beliefs about CF in oral synchronous communication task.
Necessity and frequency of the feedback
With regards to the responses of the high- and low-anxiety groups to the necessity of error correction (“I want to receive corrective feedback (e.g., provide a hint for me to self-correct, tell me that I have made an error, or correct my error) when I make mistakes”), 88% of the students in the high- and 100% of the students in the low-anxiety groups responded “strongly agree” or “agree”. Furthermore, both groups agree that they take into account CF in future tasks. Regardless of their level of anxiety, both groups were in favour of receiving CF frequently. As illustrated in Table
1, there was no statistically significant difference between the two groups.
Table 1
High-/low-anxiety group responses to the Necessity and Frequency of CF
Necessity of error correction (item 1) | 4.26 | .70 | 4.40 | .13 | −1.095 | .279 | [−.55, .26] |
Take into account corrective feedback in future tasks (item 19) | 4.11 | .75 | 4.47 | .64 | .765 | .061 | [−.80, .09] |
Frequency of CB (item 2) | 3.36 | .87 | 3.16 | .82 | .827 | .132 | [−.64, .43] |
The results clearly indicate that the students in both groups, regardless of their level of anxiety in oral communication classes, were in favour of receiving CF.
Methods of feedback
As far as methods of CF are concerned, the students in both groups rated Explicit correction as the most effective method, followed by Metalinguistic feedback, Repetition by the teacher, Clarification, Elicitation and Recast. No corrective feedback was the least favoured method by the two groups.
However, as we show in Table
2,
T-test analysis showed a significant difference between the learners’ preferences only in two methods of CF: Recast and Metalinguistic feedback, which were better rated by students who reported higher levels of anxiety in oral communication classes.
Table 2
Low/high Anxiety group responses to Methods of CF
Explicit correction (item 6) | 4.12 | 1.05 | 4.27 | 1.12 | −1.249 | .154 | [−.95, 1.05] |
Metalinguistic feedback (item 9) | 3.69 | .79 | 4.07 | .961 | −1.359 | .043* | [−.83, .88] |
Repetition by the teacher (item 5) | 3.68 | 1.1 | 3.83 | 1.19 | −.375 | .258 | [−.64, .65] |
Clarification request (item 4) | 3.49 | .95 | 3.87 | 1.25 | .546 | .417 | [.45, .44] |
Elicitation (item 7) | 3.35 | .92 | 3.67 | 1.18 | −1.106 | .226 | [−.89, .92] |
Recast (item 10) | 3.20 | 1.19 | 3.37 | .74 | −.715 | .032* | [−.91, .43] |
No corrective feedback(item 8) | 1.85 | .87 | 2.03 | .66 | −.784 | .176 | [.65, .63] |
Types of errors to correct
Responses to types of errors to correct have also been analysed, showing that both High- and Low-level anxiety groups agree that Serious errors was a priority (LA: M = 3.45, SD = .61, and HA: M = 3.67, SD = .62), followed by Frequent errors (LA: M = 3.25, SD = .91, and HA: M = 3.60, SD = .63), Individual errors (LA: M = 3.37, SD = .88, and HA: M = 3.53, SD = .74), Minor errors (LA: M = 2.90, SD = .85, and HA: M = 3.00, SD = .68) and Infrequent errors (LA: M = 2.75, SD = 1.07, and HA: M = 3.00, SD = .79). However, there were not any statistically significant differences between the preferences in the two groups of participants about the types of errors to correct.
As for the timing of Feedback, both high and low anxiety groups consider that CF is most effective after finishing speaking (LA:M = 4.15, SD = .91, and HA:M = 4.07, SD = .58), CF after the activity (LA:M = 3.90, SD = 1.7, and HA: M = 3.50, SD = 1.22), then Immediate CF (LA: M = 3.40, SD = 1.04., and HA:M = 3.33, SD = 1.12), and finally CF at the conclusion of class (LA:M = 3.30, SD = 1.29., and HA:M = 2.67, SD = 1.28). In addition, the High-level anxiety group always scores higher than the Low anxiety group with only one exception, which is Immediate CF. However, there were not any statistically significant differences between the two groups.
Source of feedback
Table
3 shows the responses of both groups when it comes to the source of the correction. Regardless of their level of anxiety,
Teachers were the most valued, followed by
Myself and
Classmates, which is the least valued by the leaners.
T-test analysis showed a significant difference between the preferences related with the
Teacher as source of CF. That is, the Low-level anxiety group values Teacher feedback as more effective than the other sources of feedback when compared to the High-level anxiety group.
Table 3
Low/high Anxiety group responses to Source of CF
Teachers (item 17) | 4.70 | .48 | 4.53 | .62 | 1.01 | .048* | [−.56, .59] |
Myself (item 18) | 3.60 | 1.06 | 3.80 | 1.24 | −.612 | .837 | [−.85, .45] |
Classmates (item 16) | 3.15 | .95 | 3.13 | .973 | .058 | .715 | [−.16, .49] |
Finally, no significant differences have been observed between the sociodemographic variables and the preference for a specific type of feedback or the timing of the CF and neither for any of the types of errors nor source of feedback.
Discussion
This study aims at answering the following questions: In virtual teaching and learning environments, are students’ beliefs on their teachers’ feedback related to students’ level of anxiety when learning a FL? Also, how do students perceive different types of CF and does such perception vary as a function of their level of anxiety? In this sense, our study yielded some interesting results, which will now be discussed.
First, both High- and Low-anxiety groups strongly support provision of CF in oral tasks, which is in line with what Zhang and Rahimi (
2014) found in their study. Respondents, regardless of their anxiety level, were very interested in using CF in order to improve their future tasks, which is encouraging for teachers to see since feedback to be considered as such needs to be engaging (Carless
2007). Tseng and Tsai (
2007) concluded that reinforcing feedback that provides a positive feeling or recognition of the work and suggestive feedback provided by peers are most useful for subsequent learning. Clearly, CF is in intimate relation with a student’s emotional make-up and needs to go hand in hand for positive outcomes (Ozmen & Aydın
2015).
Interestingly, respondents in both groups seem not to like too much CF. These results are not in keeping with the ones in Zhang and Rahimi (
2014), whose respondents chose
Usually and
Immediate CF as their most preferred answers while our respondents chose
Sometimes. It is also worth mentioning that the High-level anxiety group in our study consistently scores higher than the Low-level anxiety group in Immediate CF. Such pattern may be due to the fact that Immediate CF may interrupt the natural flow of speech and that it may also undermine both the speaker’s confidence and increase anxiety, especially in the high-level anxiety group. The results agree with some of the findings in Martínez (
2013) where nearly two-thirds of his participants reported resent and worry when their teacher immediately corrected them and that almost half of the participants hated making oral mistakes, thus undermining their confidence.
As far as responses to types of errors to correct are concerned, both High- and Low-level anxiety groups agree that
Serious errors are a priority, which is also observed in Zhang and Rahimi (
2014) and Abedi et al. (
2015). It is fair to add that Zhang and Rahimi (
2014) also think that
Serious problems are a source of great anxiety, confusion and disappointment in communication, thus leading often times to communicative breakdown. We may partially agree with such explanations since our participants have registered for a communicative course in nature, which is well-grounded on constructivism (active learning in real-life situations), thus encouraging students to interact by means of tools such as Skype. Any communicative breakdown in such a context can prove problematic for the speakers.
Second, concerning appraisal of different types of CF, the students in both groups rated explicit correction as the most effective strategy, which confirms that most of the teachers’ oral CF had positive impacts on the students’ affective variables, specifically lowering their anxiety about speaking English (Lee
2016; Renko
2012; Zhang & Rahimi
2014).
More interestingly, the pattern seems to be also true across learning environments, since the participants of our study are from an online environment. In particular, our study showed that
Recast and
Metalinguistic feedback were better rated by the students who reported higher levels of anxiety in oral communication classes, which is critical information for tutors when delivering feedback. Partially contradicting results can be found in Abedi et al. (
2015), who found
elicitation,
implicit correction and
recast as the most effective methods irrespective of the anxiety level of the participants. In fact, the students’ online learning environment, in which active learning and interaction inspired by constructivism occurs, leads students to view specific CF methods as more effective such as
Metalinguistic feedback, which entails high-quality feedback and fosters uptake (Lyster & Ranta
1997).
Finally, our results also reveal that all the students prefer feedback coming from the teacher. When it comes to the source of the correction, both anxiety groups scored
Teachers the highest followed by
Myself and then
Classmates, which is the least valued. By and large, our results are in line with other researchers who also found teachers as the most preferred source for CF, considering it as more reliable (Ertmer et al.
2007; Van den Boom et al.
2007). These results may contradict Baralt and Gurzynski-Weiss (
2011) and Ahmadi and Sadeghi (
2016) where teachers may be a cause for greater anxiety when they are also the interlocutors in a task, especially due to differentials in power relationships. Therefore, teachers seem to have a critical role when delivering CF in both face-to-face and online environments since according to our study and Zhang and Rahimi (
2014), students consider teachers as their prime source of knowledge and expertise. In addition, one must consider a possible influence from the learning environment when we see that self-correction is in second position in our study while it is last in Zhang and Rahimi (
2014) since self-evaluation is encouraged in e-learning (Hopkins
2010). Furthermore, Butler and Lee (
2010) stated that teachers and students perceived the effectiveness of self-assessment differently depending on their teaching/learning contexts.
In general terms, the two research questions have been answered by providing evidence. We have evidence that confirms the fact that in virtual teaching and learning environments students’ beliefs on their teachers’ feedback is related to students’ level of anxiety when learning a FL, which is in line with similar findings in face-to-face learning environments. Furthermore, students’ perception of different types of CF does vary as a function of their level of anxiety also in line with face-to-face learning environments. However, interestingly, students in an online environment have shown slightly different preferences in types of CF, showing that the learning environment may have a significant effect on CF preference.
This result is relevant for teachers since they become critical in motivating their students in order to foster positive feelings towards learning a FL. Thus, creating an online learning environment for real life learning, social interaction and an experiential learning environment (free-anxiety environment) is very important in order to favour learning in authentic learning tasks (Herrington & Reeves
2003; Huang
2002a,
2002b). In this context, corrective feedback could be more engaging, and thus can reduce levels of anxiety in FL learning.
While the results here have validity with respect to our environment and population, they were, nonetheless, limited by a number of factors. For instance, the fact that it is about the participants’ beliefs on CF in a hypothetical situation of which students have no experience and solely rely on their face-to-face experience in speaking tasks may be one of them. Perhaps, the same study with participants who have experienced online synchronous speaking tasks with a teacher would yield different results. Secondly, the size of the population is relatively small. Also, we cannot be positive about the reasons for choosing specific answers without further looking into the matter in more depth by way of either a new survey asking participants about the reasons for choosing their answers or even an interview, which would either way allow us to collect qualitative data so as to shed some more light onto our present results. Last but not least, future research would need to include both more and face-to-face participants in the study, using the same research tools for comparison purposes. Baralt and Gurzynski-Weiss (
2011) failed to find significant differences in state anxiety between face-to-face and computer-mediated communication among students of Intermediate Spanish. However, as the same authors have suggested, there are new aspects that are in need of further study, such as the interaction between the place where the research is carried out and levels of anxiety as well as task type and levels of anxiety. Indeed, such information will be considered in our future research.