The study is an evaluation of the ‘Science Teacher Development Project’, which was undertaken in a rural area of the Mpumalanga Province of South Africa. The intervention of almost four years was an outreach programme located at eMpuluzi, a distance of 375 km from the University of Pretoria. The project was funded by a grant of the Belgian Council of Flemish Universities. Funding was administered by the South African Science Education Network (SASEN) which was located at the University of the Western Cape. The Centre for Science Education of the University of Pretoria designed, managed and implemented the intervention. A ‘Further Diploma in Education’ (FDE) from the University of Pretoria was the intended outcome for individual participating teachers and 26 of 28 teachers succeeded and obtained the diploma. The project was particularly unique through being funded and implemented by a partnership of universities, rather than through state initiatives or aid, and as such must be seen as a small project with high impact within its limited scope. This thesis presents an evaluation of the project in terms of its impacts, achievements and failures, using qualitative paradigms. The evaluation process by the participant observer was an ongoing programme and included a post-completion summative evaluation for which the last interview was conducted two years after completion of the project. The evaluation proceeded in two phases. The first was a traditional analysis based on proxies derived from the research question, namely self reported change in educator attitude and change in school and classroom practices of educators. The second investigation which was begun when a successful conclusion to the intervention was already evident established the degree to which expectations of stakeholders were met, and factors that contributed to the success. For this a novel adaptation of Checkland’s Soft Systems Analysis with six stages was used. This ensured that the nature of the project was modelled and assessed holistically. The model led to the identification of several perspectives for analysis. These included the competence criteria of the Department of Education known collectively as ‘the seven roles of educators’, and whether the project was a ‘learning organisation’. A major factor which contributed to the successes of the project was the sense of ownership that developed amongst the participants. This work presents a new approach to the in-depth evaluation of single interventions, particularly where these are presented over a long period and the roles of participants at all levels may change or develop over time. The Soft Systems Analysis approach has the distinct advantage of not imposing an assumption of structure onto the analysis of the project. The potentially generalisable success factors which were identified (most importantly the sense of ownership developed by the participating teachers which exceeded the concept of partnership envisaged in the Jomtien (1990) declaration), the novel adaptation of Soft Systems Methodology, and the value of strongly academically focussed INSET are significant contributions to the body of knowledge of those planning, implementing and evaluating teacher development projects amid considerable complexity. These all lead to further research.