This paper will describe the double-fold validation process carried out in order to obtain statistically reliable surveys to identify difficulties and best practices in catering to diversity in CLIL within the European ADiBE project. It is distinctive from other papers in that it offers the statistician’s perspective in validating instruments of this nature.
This paper describes the statistical validation process followed in the questionnaires that have been designed by the members of the European project “CLIL for All: Attention to Diversity in Bilingual Education (ADiBE)”. This validation process has several phases. In the first one, a descriptive analysis of the experts’ ratings approach was carried out. The external evaluators rated each of the items in the surveys in terms of different parameters (clarity, precision and relevance), shedding light on which must be eliminated or modified so that the instruments became clearer and more affordable for the intended respondents. Afterwards, the process of the second phase, the pilot study, is shown. It has to do with the quantitative validation of the questionnaires. The different participating cohorts (teachers, students and parents) were sampled to measure the reliability of the tools (through the internal consistency of the instruments, using Cronbach Alpha). This reliability has been calculated not only for the entire questionnaire, but also for each of the blocks into which it was divided. Moreover, the homogeneity index has been used to check whether the items really have the capacity to discriminate within each of the sections of our instruments. Finally, the results of both types of quantitative validation are crossed with the qualitative analysis of the opinions provided by experts, allowing the ADiBE project to employ reliable and valid questionnaires statistically speaking.
