AbstractThis study reports findings from a mixed-effects analysis on item- and person-levels for five intact classes oflearners following a semester-long pragmatics instruction. Data represent multiple-choice responses on an instrument assessingpragmatic awareness from learners at three levels of formal Japanese education (lower- to post-secondary). Instructionalinterventions differed in the degree of explicitness for both pragmalinguistics and sociopragmatics. Ten speech acts representedpragmalinguistic forms (apologies, complaints, compliments, farewells, greetings, introductions, invitations, offers, requests,and suggestions), with the number of speech acts taught operationalized as part of the explicitness. Twelve relationship statuscategories represented the sociopragmatic feature (acquaintances, best friends, boss/employee, coworkers, customer/service person,family, friends, girlfriend/boyfriend, wife/husband, professional relationship, neighbors, and strangers). Item-level resultsreveal a significant effect regarding learner accuracy on items measuring sociopragmatic content relative to pragmalinguisticcontent. Person-level results reveal that increased explicitness accounted for differences between intervention groups. However,the highest-performing group did not receive the most explicit instruction. The discussion includes an argument for greater use ofvideo-based content in pragmatic instruction and assessment practices and the value of mixed-effects models when analyzinglongitudinal classroom research across multiple sites.