Transactional Analysis, or TA, has been used for more than four decades to enhance interpersonal relationships and promote personal growth through counselling and psychotherapy. It has been used to advantage in organisations, principally in the business world. It was thought at the outset of this study that TA could also be of benefit to schools.The aim of this study was to find out whether TA methods could be used in schools to improve student outcomes in science and mathematics, to enhance interpersonal relationships and to promote positive behaviour. The study trialled TA by using it with a group of at-risk students to discover whether its use would bring about positive change.A group of ten at-risk students became part of a mentoring programme, using TA. This programme focussed on the behaviour and academic progress of the students, and sought to empower them to make positive changes. This group referred to as the sample group, was compared with a control group. The Adult ego-state (thinking) was promoted in the students in order to shift their 'locus of control' from their Negative Adapted Child ego-state, the source of much non-productive, inappropriate and rebellious behaviour.The study upheld the reliability and validity of the questionnaires used, namely the Questionnaire on Teacher Interaction, the Coopersmith Self-Esteem Inventory (School Form) and the Mooney Problem Checklist. The study prompted the construction of an informal test, the Ego-State Questionnaire, which proved to be informative.Poor attendance and school behaviour records were good indicators of a student's 'at-risk' status. At-risk students were found to be already achieving below their potential in science and mathematics at entry to secondary school.The mentoring programme ran for six months, and at the end of this the sample group had improved behavioural records and increased self ++ / esteem. Their number of perceived problems had dropped dramatically, and their academic results were improved.Interpersonal relationships between the sample group and their science teachers were better than the interpersonal relationships with their mathematics teachers, indicating a continued difficulty with abstract ideas at the end of the programme and a need to run such programmes over a longer time span. Interpersonal relationships did improve out of school with parents and peers. Clear preferences were indicated for what students preferred in the behaviours of their ideal teacher: understanding, helping/friendly, leadership and strict behaviours.Encouragement of Adult ego-state was shown to be an appropriate and productive approach to the improvement of academic and behavioural outcomes for at-risk students in science and mathematics. The study also showed that at-risk young people had a lower than average Nurturing Parent ego-state available to them.Teachers rated their TA101 course highly, and found that it gave them a fresh perspective on classroom difficulties. Both teachers and students benefited from the use of TA in this study.