Even though many researchers have expressed the opinion that the findings of any linguistic investigation can only ever be partial and provisional (e.g. Larsen-Freeman, 2006), it is important nevertheless to attempt to make claims generalizable in age-related research, within limits, in order (1) to identify in a broad way the nature and challenges of learning foreign languages (FLs) at different ages; (2) to better understand the various factors which are typically found interacting with age effects; (3) to inform educators, teachers and policymakers about the pattern of trends of age-related learner outcomes. Such research clearly has important, nay crucial, implications for FL education, especially in relation to decision-making about early instruction in different languages at primary school, on the one hand, and, on the other, regarding later instruction in different languages at secondary school (see also Pfenninger & Singleton, 2017b). Bachman (2006) reminds us that, whatever our intentions, our audiences are not restricted to members of our own research community but also include an audience from a more public, more politically potent sphere, including people who have the power to make real decisions in the real world on the basis of our research:

Given this, one of the main goals of age researchers has to be to attach meaning to particular age-related outcomes by generalising to other individuals or groups of individuals, or other contexts. Such a goal is clearly desirable in order to be able to extrapolate beyond particular instances. It is, we would claim, perfectly possible to attain this goal without making unwarranted inferential leaps. In this respect, Duff (2010) reminds us that generalizability should not be taken for granted in quantitative research – that it must, rather, be demonstrated by the researcher.