Most studies investigating the role of age in L2 acquisition compare L2 children and adults in terms of ultimate attainment (e.g. Johnson & Newport 1989). This paper addresses some of the methodological and conceptual issues which arise when, following Schwartz (1992), such a comparison is made in terms of development. The paper considers three variables carefully controlled for in any such comparison: L1 transfer, age at time of testing and L2 proficiency. A new proficiency measure, the Age-Sensitive Composite Proficiency Score (ASCOPS), is presented. Unlike previous measures, ASCOPS can be used with both L2 children and adults, and it takes into account the intrinsic differences between these two groups relating to their age at time of testing.