Learners as teachers? An evaluation of peer interaction and correction in a German language MOOC

The benefits of peer interaction, support, and feedback in Massive Open Online Courses (MOOCs) for Languages (LMOOCs) are well documented, but there has been little research on peer correction in MOOCs. Classroom-based research suggests that peer corrective feedback has significant potential for language development, but it also identifies a number of conditions for the feedback to be effective, notably a ‘positive classroom atmosphere’; this may be hard to achieve on a MOOC, with its diverse cohort and large number of participants. Our mixedmethod study reveals participants’ conflicting expectations of learning from their peers on the one hand and actively contributing to their peers’ learning on the other. Most participants believe they are not competent to provide helpful corrective feedback, and some think that the expectation to correct creates unwanted pressure and hinders communication. This paper encourages MOOC educators to address the challenge of creating a culture of learning through meaningful interaction whilst also finding ways of exploiting the opportunities offered by constructive peer correction.


Introduction
Many MOOCs provide open discussion forums for commenting and interacting with educators and peers. The forums in LMOOCs additionally offer participants the opportunity to implement what they have learned, that is to write in the target language. Tasks are designed to push participants to use their linguistic resources in meaningful communication, and any errors they make can open opportunities for language development -provided learners are made aware of these. This could be achieved through promoting peer corrective feedback, thereby enhancing learner engagement in courses which typically have limited educator resources.
Previous studies have established that peer interaction in discussion forums of LMOOCs has a largely positive effect (Martín-Monje, Bárcena, & Ventura, 2013;Sokolik, 2014). In this study we looked more closely at factors contributing to the effectiveness of peer interaction and, specifically, corrective feedback. Classroombased research has shown peer corrective feedback to have significant potential for language development (Philp, 2016;Sato & Ballinger, 2016;Sato & Lyster, 2012) and has identified learners' proficiency level, their social relations, and their willingness to collaborate as factors which affect the quality and quantity of feedback.

Method
The study was based on a post-beginner level MOOC for 'German at work' produced by the Open University and delivered via FutureLearn. The course included textbased and audio-visual resources, quizzes, speaking tasks, and structured writing, as well as open discussions, which are the focus of our study.
In order to explore how learners engaged with the course and how they learned from their peers, we analysed four types of data: • user analytics (4,063 learners); • learner contributions to in-course discussion forums (1,487 contributors); • pre-course survey data (1,088 respondents); and • verbal and written data from online focus groups (34 participants).
This paper focuses on learners' attitudes to 'peer corrective feedback', a term that refers to participants' replies to posts in the discussion forums, which contain any form of correction in response to an error.

Data analysis and discussion
3.1.

Peer interaction and peer correction in discussion forums
In order to establish types of peer interaction and correction in our courses, we studied learner contributions in the week with the highest interactivity (Week 1 of 'German at work Post-beginners' 1). User analytics showed that of the 4,063 active participants worldwide, 1,487 contributed to the discussion forums resulting in a total of 6,241 postings by learners. A close reading of these learner contributions to in-course discussion forums revealed that most postings did not elicit direct responses, but where learner-learner interactions did occur they included general social interactions (greetings and introductions), technical support, 'moral' peer support (e.g. encouragement to overcome challenges), and explanations of linguistic or cultural aspects. The postings also included 172 instances of peer correction, which were distributed as follows (Table 1). Peer corrective feedback included explicit correction, recasts, questioning, and translation, and was often integrated into a meaningful response, as in the following example: "Ich bin an accountant".
The corrections were provided by a small number of engaged learners (2.3% of forum contributors) including at least four first language speakers. Seventy-five percent of corrective feedback was provided entirely in the target language. Most of the corrective feedback was positively acknowledged by the recipient through thank you messages or likes, and some participants demonstrated uptake by editing the original message or in a subsequent posting. Often recipients saw the need to explain themselves ("This was my first attempt"), and there was a small number of requests for confirmation by other participants or educators. Although all corrections were potentially helpful, occasionally they introduced new errors as in the example above where the appropriate word "Steuerberater" is provided but the article "einen" is incorrect.

Attitudes to peer correction
The pre-course survey data indicated a significant discrepancy between participants' goal of learning from the expertise of others (49.5%) and their goal of sharing their own expertise (15.5%). We found a similar discrepancy when asking focus group participants about their attitudes to peer correction. Whilst many expressed at least partially positive attitudes to receiving corrective feedback by peers, attitudes to correcting others were predominantly negative. Out of the 34 focus group participants, 30 gave their opinion on being corrected by peers and 28 gave their opinion on correcting others. Their views are summarised in Figure 1.

Figure 1. Attitudes to peer correction
When asked about being corrected by peers, many participants indicated that feedback could be helpful provided it was undertaken tactfully ('positive'), but also expressed doubts about the reliability of the feedback ('mixed') and some thought that the expectation to correct peers places undue emphasis on accuracy and may even hinder communication and impede learning ('negative').
No-one was unreservedly positive about correcting others. Some participants stressed the need for feedback to be constructive and some thought it helped that they had had some training, for example, because they were themselves teachers ('mixed'). Most participants expressed reluctance or were strongly opposed to offering corrective feedback to others ('negative'). The main barriers were participants' lack of confidence in their own ability and a fear of triggering resentment, particularly given the absence of body language in the online environment and the cohort's cultural diversity. Some thought that it was 'not their place' to correct contributions by other learners who had done their best, or even that peer correction should not be touched 'with a barge-pole'. These attitudes were linked to an expectation to learn from the educators rather than from peers and a view that educators should take a more proactive role.
The discussion of peer corrective feedback led to the use of more negative emotive language ('annoying', 'conflict', 'upset', 'frustration') than any other discussion topic. For example, when talking about peer interaction more generally, most participants expressed appreciation for the opportunity to engage in a supportive way with their fellow learners from around the world. Some gave reasons for not contributing, but the kind of emotive language used in relation to peer correction was entirely absent.
On occasion, it appeared that cultural biases also led to conflicting expectations of who should correct and particularly how corrections should be undertaken. Some cultures were perceived to be more polite and sensitive, whilst others were perceived as less tactful. There is scope for further research in this area.

Conclusions
Participants reveal conflicting expectations in terms of wanting to learn from others but lacking confidence to share their own expertise. Focus group data further show discrepancies between learners' expressed preference to learn directly from the educators and the participatory nature of the course. Only a small number of LMOOC participants provide corrective feedback to their peers, and this receives mixed responses. Many respondents express strongly that they would not give corrections themselves.
As MOOC designers and educators, we need to consider how to foster the 'positive classroom atmosphere' (Philp, 2016) which is vital for successful peer learning and how to manage expectations in a large and culturally diverse cohort. We should explore ways of training our 'learners as teachers' so that the benefits of peer corrective feedback can be harnessed without causing anxiety or hindering meaningful interaction. The findings from this project can be used to develop strategies that promote constructive interaction, feedback, and error correction. These will be helpful to both participants and educators on LMOOCs.

5.
Disclaimer: Research-publishing.net does not take any responsibility for the content of the pages written by the authors of this book. The authors have recognised that the work described was not published before, or that it was not under consideration for publication elsewhere. While the information in this book is believed to be true and accurate on the date of its going to press, neither the editorial team nor the publisher can accept any legal responsibility for any errors or omissions. The publisher makes no warranty, expressed or implied, with respect to the material contained herein. While Researchpublishing.net is committed to publishing works of integrity, the words are the authors' alone.
Trademark notice: product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.
Copyrighted material: every effort has been made by the editorial team to trace copyright holders and to obtain their permission for the use of copyrighted material in this book. In the event of errors or omissions, please notify the publisher of any corrections that will need to be incorporated in future editions of this book.