Adherence is necessary to establish the effectiveness of rehabilitation exercises for those with long term conditions, but measuring it is problematic. Adherence is vital when attempting to measure and understand how an intervention works; without appropriate levels of adherence it is not possible to conduct a fair test of an intervention since failure to find benefit may be due to poor adherence rather than an ineffective intervention. There are no gold standard measures, and objective measurement devices are fallible. Without an accurate way to assess if participants are adherent to their exercise programs the efficacy of the programme cannot be assessed. Therefore the objective of this PhD was to develop a psychometrically valid and reliable self-report adherence measure to assess the likelihood of adherence for stroke survivors called the Stroke Rehabilitation Exercise Adherence Measure, (StREAM). A large systematic review was conducted to synthesise the existing evidence for self-report measures of adherence to prescribed unsupervised exercise programmes. The review found many measures but few possessed acceptable any tested psychometric properties and therefore the need for a psychometrically sound adherence measure to assess the likelihood of adherence was highlighted. To ensure StREAM had robust content validity stroke survivors (n=16), physiotherapists (n=3), and exercise professionals (n=2) were interviewed. Items were created from salient phases in the interviews which were then discussed at two focus groups of stroke survivors. This ensured items in StREAM were suitable and comprehensible to the target population.