This article-based Master's thesis investigates how undergraduate students in Iran learn English through regular topic-based online conversations with native-speaking conversation facilitators (CFs) living in different countries. The conversations are conducted under difficult circumstances due to slow internet connections. The aim of the study was to discover what changes in the design of conversation assignments and in the training and guidance given to CFs should be implemented in order to create patterns of interaction and conditions more conducive to language learning.
The core material in the study consists of two articles using a qualitative approach. The first article is based on semi-structured interviews with eight CFs. It sought to gain insights into the ways in which the online conversations were being conducted and how oral proficiency was being developed from the point of view of the CFs. The second article was based on the results of students' responses to open questions in a broad student survey, investigating their experiences and perspectives on learning a language through the online conversations.
Although the two articles used conventional qualitative methods, the thesis as a whole was an attempt to pursue action research, which consists of cycles of: action research evaluation changes (intervention) - action etc. This meant that the researcher-practitioner has been inside the organization working as the main conversation administrator and designer while simultaneously pursuing the research. As a consequence of this position, in addition to the qualitative data used for the articles the researcher-practitioner has had access to a wide range of research material to support the analysis. The articles themselves are thus effectively ''snapshots'' of part of the research cycle.
The results of the first article showed that CFs with well-developed pedagogical and technical skills, including facility with written chat, were able to find ways to facilitate learning through their flexibility in the use of conversation assignments and variation in choice of topic. The CFs who encouraged students both to interact with one another and influence the direction of conversations were able to promote more
open dialogue, while CFs who followed conversation assignments too rigidly sometimes had difficulty in engaging students' full attention. CFs who were only able to talk with one student at a time due to technical limitations experienced particular difficulties in establishing a dialogue, especially with weaker students.
In the second article, analysis of survey responses showed that most students enjoyed and learned from the conversations, indicating that the CFs' abilities to adapt their language to make it comprehensible, and give explanations and examples where difficulties arise are essential to facilitate a smooth-running interaction. Students also stress that the CFs' appropriate use of corrective feedback, willingness and ability to vary questioning techniques and conversation topics, all help to enhance learning, as does the socio-affective support given to the students to enable them to relax in the online environment. Some students, however, saw room for improvement through an increase in the CFs' flexibility with conversation assignment questioning and through the introduction of a greater variety of topics so that conversations do not become routine and predictable. A significant number of students also thought that more explicit correction or highlighting of mistakes would be helpful. The explanations and interpretations of the different findings and the implications for changes are discussed in the two articles
The thesis itself is organized with an initial introductory chapter presenting the background for the study and an overview of relevant research followed by chapters on theory and methodology. Summaries of the articles are then presented followed by a discussion focusing on the meaning of the results of the two articles in the wider research context and the validity of the study as a whole. The conclusion outlines the implications of the study and suggests possible future directions in the field. The articles themselves are placed at the very end of the thesis.