Linguistic distance effects on country TOEFL scores between 2005–2020 and the effect of morphological distance in particular

We had aggregated country TOEFL scores of at least 15 million non-immersion learners coming from 127 countries around the globe with 65 L1s covering a time span of 16 years (2005 – 2020) at our disposal. This rich database allowed us to examine not just the effects of various country level characteristics, including linguistic distances, but also their dynamic effects over this time range. We performed this analysis by applying a generalised additive mixed regression model (GAMM) to reveal intricate, non-linear interaction effects, especially those of the three linguistic distance measures we applied, and, more in particular, to map the changing role of morphological distance in relation to lexical and phonological distance. All three distances contributed in explaining L2 English country scores: the larger the distance is the lower L2 English proficiency. Morphological distance delivers its own, distinctive contribution, next to and in competition with the lexical and phonological distances. All three distances interact with year of testing. These changes seem to reflect the increasing presence and dominance of L2 English worldwide, but it shows at the same time that the impact of distance is dynamic. Morphological and phonological properties become more important the higher the L2 proficiency is, while the importance of lexical distance gradually decreases.