How might you describe the “population” that de Winter et al.’s study was reporting on? Do you consider that de Winter et al.’s approach to sampling to be biased, and—if you do – might that undermine the validity of their conclusions?
We can only judge the applicability of a sample in terms of the purpose of the particular study. So often in ERP1 research, the aim is often to find out what is usual, normal or typical in a population (e.g., how much the average student might learn in given circumstances), and a biased sample will bias the results. In general, when a study claims to look at what is typical in a population, then the sample needs to be representative of the population the study claims to be about.
However, de Winter, Winterbottom, and Wilson were looking to produce guidance to support teachers in adopting new technologies in their teaching, and so deliberately selected their participants in order to work with those in good positions to help them identify effective approaches. In such a study, it is quite appropriate to purposefully select participants who are not typical of a wider population. We could argue here that the population actually being sampled was not teachers in general, but only those who had interest and competence in using new classroom technologies. Alternatively, we could suggest that in this type of study, the very notion of sampling a population is simply not relevant. The important point is that the selection of participants in the study was sensible, and so could be justified, in terms of the purposes of that study.
In reading about the examples of research discussed in this book, the reader should be alert to the size, mode of selection, and composition of samples of informants providing the data upon which findings are based (perhaps keeping a note when reading studies, using a form such as that in Figure 1.1). Where studies attempt to offer general findings, the reader should ask themselves whether the sample could be