Some researchers have focused on identifying factors that affect dropout behavior, or factors that affect students’ learning progress. Woodley（Woodley，2004）summed up 46 influencing factors drawn from learners' personal characteristics and characteristics of the educational institution. In Woodley’s view, the study of dropout behavior involves the study of the students’ academic progress itself and the study of educational output produced as a function of educational investment. Yorke and Davies（Yorke，1999；Davies，2003）drew similar conclusions about the reasons for student dropouts from their independent case studies, which include: poor choice of fields or courses, academic difficulty, financial difficulty, unfavorable background, dissatisfaction with the service of the school, or poor choice of educational institution.

The study by UK scholar Ormond Simpson（Simpson，2007）on the practices of UK’s Open University is a notable recent achievement in the research of school dropouts. Simpson paid attention to the timing of the dropouts and suggested that the dropout rate consists of two parts: dropouts preventable by the educational institution, and dropouts that are not preventable by the educational institution. The non-preventable part is also called the minimum dropout rate. Even students with a high education level and strong learning motivation will give up schooling because of illness, family- and job-related factors. The preventable dropouts should be the focus of educational institutions in lowering the dropout rate. He also calculated the financial benefits to the university, students and the financial supporters of the university, namely, the UK government, if UK's Open University were to apply proactive and interventional student support services.

Despite a great many overseas studies on distance education dropouts, related studies in mainland China are insufficient. Influential domestic case studies include those by CAO Wen (2005), Zhang Miaohua (2006) and Li Ying (2009, 2010). These studies were based on data from various education management systems, using educational statistics, econometric models, and investigations via interviews. They revealed general rules for distance learner dropouts, such as the peak times of dropouts, the difference in dropout rate between associate degree programmes and bachelor's degree programmes, the study performance of student dropouts and the causes of dropout. However, none of the studies, either using statistical data or using interviews, disclosed the inevitable link between the school dropouts and their programmes. In open education, what are the characteristics of the dropouts in a particular programme? Which of these characteristics are common or similar to dropouts of other programmes and which of them are closely related to the course features of the programme itself? When the correlation is established, distance education institutions can adopt more tailored, effective measures to lower the dropout rates for specific programmes.

The radio and TV universities in China accommodate about two-thirds of the students enrolled in ODL nationally. This study applies both quantitative and qualitative methods in researching the high dropout rate for English majors at radio and TV universities, and brings to light the inevitable link between dropouts and specific programmes.

Research Method

Research design
As stated above, English is one of the programmes with the highest dropout rate in radio and TV universities in China. To investigate the connection between the dropouts and the programmes, the English programme was chosen for research. 14 provincial radio and TV universities (representing the eastern, central and western regions) with relatively high enrollment in the English programme were selected as sample observation sites. These 14 observation sites include 8 English bachelor's degree programmes and 6 associate degree programmes.

The biggest difficulty for researchers is to find or contact school dropouts scattered throughout society. That is also the reason why domestic studies tend to be based on static data from educational system databases rather than on personal surveys. In order to collect fresh information reflecting the real-life views of student dropouts and not solely rely on static institutional data, the research team chose the students admitted for the spring semester of 2010 from the 14 universities as subjects for the study. The research plan is to track the students for three semesters², recording and analyzing the dropout phenomenon for each semester. Investigative data are collected each semester from three sources: questionnaires filled out by the dropouts, academic transcripts from the education management system databases and interviews with the dropouts, their class supervisors and English tutors. The multiple data sources and interviews allow for an accurate, comprehensive and in-depth examination of the English programme dropout phenomenon.

Study implementation
With the start of student enrollment in the spring semester of 2010, the research team officially began their work with the help of research partners at the 14 observation sites. During the spring semester, research partners at each site increased their communication with the class supervisors of the freshmen, obtained contact information for the students through phone calls, QQ and MSN and created files for the students. A total of 2,237 students were included for observation.

In the autumn semester of 2010, partners from the 14 sites used video conferencing and face-to-face meetings to unify research content and implementation procedures. They reached consensus on the topic of “What is dropout?”, or the work definition of dropout, and set out to investigate dropouts during the autumn semester. The investigation lasted 4 months from September to December. A uniform investigation package was used, including questionnaires, transcripts of the dropouts, and outlines and plans for student and tutors interviews (including English tutors and class supervisors). Interviews were recorded and converted into text later.

Data collection
This study combined quantitative and qualitative research methods. In the quantitative aspect, the study used online questionnaires to survey the dropouts' general profiles, academic progress and causes for quitting school; academic transcripts for the dropouts were also obtained from the education management system databases. In the qualitative aspect, qualitative interview methods were used to conduct face-to-face individual and group interviews, telephone interviews and online interviews with dropouts who could be contacted. Interview contents covered five aspects, including the students' individual background, academic progress, interaction between teachers and students, causes for quitting school and suggestions for improving teaching. As dropping out constitutes a failure or unhappy experience, some dropouts refused to accept questionnaires or interviews. Among the 204 dropouts in the first semester, transcripts were obtained for 195 of them, of which 118 filled in online questionnaires, and 98 of those who filled in questionnaires took part in the interviews. To explore different angles on why the students dropped out, face-to-face interviews with English tutors and class supervisors (also called counselors, but hereinafter called class supervisors) were also included. A total of 40 English tutors and class supervisors for the first semester dropouts were interviewed from the 14 provincial radio and TV universities.

² Earlier research shows that the first 3 semesters out of the 8 years of study are usually the peak period for dropouts. See Reference [4].