Mobile-assisted prosody training for limited english proficiency: Learner background and speech learning pattern

The use of Mobile-Assisted Pronunciation Training (MAPT) has been increasing drastically due to the personal and interactive nature of mobile devices. However, MAPT applications lack support from empirical evidence as research on MAPT-based acquisition, particularly related to prosody, has been rare. The present study employs a MAPT application with lessons on lexical stress and prominence with Limited English Proficiency (LEP) users (n = 31) of mixed ages and first languages. Then, 16 experienced raters conducted discourse-based prosodic analysis on unconstrained speech collected at the beginning and the end of the intervention. A series of mixed-effect model analyses were conducted on learner effort, improvement and learner background to investigate their relationship with accentedness and comprehensibility. The results indicated that present MAPT prosody interventions were effective for comprehensibility but not accentedness, however, learner effort on lexical stress and prominence exhibit differing patterns. Similar to previous findings, learner age impacts production more than the length of residency or history of language study. Implications include a prosody-based MAPT application; support for the treatment of accentedness and comprehensibility as separate, but related constructs; and a further understanding of the role of learner-related factors in prosody intervention.