Research on children’s experiences of disability is not extensive with a number of challenges to collecting this data (Porter & Lacey 2005; Porter et al 2008). Typically researchers use qualitative and unstructured approaches to the collection of data from children and as others have noted some voices are more readily heard than others (Ruddock & Fielding 2006). Schools in the UK are required to monitor the impact of their activities on children with a disability and to actively promote equality of opportunity. The purpose of this research was to develop the tools that schools could use to enable them to collect data from children with a disability on the barriers and supports for learning. Previous research has highlighted the importance of having a variety of different strategies for collecting this data (Lewis et al 2005) noting that “one fixed approach” does not meet all needs. Additionally there is a need to leave room for schools to make choices. In the field of pupil voice multiple methods has been seen as a way of gaining a more authentic picture of what pupils really think (Flutter and Ruddock 2004). However, despite the long history of using mixed methods in applied fields (Greene 2008) there are still a number of unanswered questions, not least with respect to how and when we integrate the data, described by Teddlie & Tashakkori (2009) as the “most unchartered area”. The focus of this paper is how different methods may produce seemingly paradoxical data as different methods can result in different views being surfaced. This raises important issues concerning the analysis and representation of the data. Debates in the literature with respect to individual or social models of disability are played out in methodological decision-making, thereby raising issues of the relationship between theory/theories and the portrayal of evidence. How do we explain paradoxical data arising from the use of different methods ?

An indepth-case study of 2 schools (1 primary and 1 secondary) selecting methods from six alternatives that varied according to language content, the extent to which the child was enabled to undertake them independently; the degree of face-to-face dialogue with an adult; and whether they were individual, small or larger group activities. The choice included: “Talking Mats” an individualised picture/symbol based approach; Focus groups using nominal group technique; an interview schedule; “Point to Point” an activity based around counselling techniques using visual representation that could be used individually or in pairs or small group activity; a hard copy symbol questionnaire; and an online questionnaire. The chosen methods were used to collect data from the same group of children. The data is analysed comparing response sets, transforming one set of qualitative data into codes that are generated with reference to the qualitative and quantitative data arising from the other method.

In the secondary school, each focus group captured a consensus, through generating concerns, categorising and ranking them but did not surface individual differences. There are striking differences between the groups in relation to what issues emerged. Depending on the unit of analysis some of these differences are reflected in the questionnaire data. The questionnaire enabled individual, group and whole year data to be portrayed. Both methods appear to have prompted some self-reflection. They differ in the extent to which they ensure the child’s right to both anonymity and silence. The analysis reflects the need for contextual sensitivity in analysing the data but highlights how different theoretical stances will foreground different parts of the picture. Data from the second case study in a primary school is still to be analysed