Though emerging research supports the effectiveness of school-based coaching models utilizing motivational interviewing (MI), an examination of the specific drivers behind these effects is notably lacking in the prevention field. This study leveraged sequential analysis to examine how teachers' verbalization of change talk (i.e., language in support of change) and sustain talk (i.e., language in support of maintaining the status quo) was influenced by coaches' use of MI-consistent (i.e., collaborative language supportive of change) and MI-inconsistent (e.g., confrontational, directive) language, respectively. We also examined whether teacher and coach factors were related to coach-teacher language dynamics. Data were collected from 87 teachers in 16 elementary and middle schools randomized in a trial to the Double Check preventive intervention (see Bradshaw et al., 2018). Audio-recorded coaching feedback sessions were coded using an adapted version of the Motivational Interviewing Sequential Code for Observing Process Exchanges (MI-SCOPE). Sequential analyses indicated that MI-consistent and change talk were significantly more likely than chance to occur consecutively. Teachers' sustain talk was also more likely to occur sequentially with coach use of MI-consistent language and teacher change talk; the latter suggests teacher ambivalence. Coaches rarely used MI-inconsistent language and its occurrence was only associated with more MI-inconsistent language. Regression analyses indicated that teacher age, efficacy, burnout, classroom organization, and some design features (i.e., cohort, coach, coach-teacher racial match) were associated with different coach-teacher language dynamics. This novel school-based study illustrates how coaching MI evoked teacher change talk related to use of evidence-based programs. [This paper was published in "Prevention Science" v22 p786-798 2021.]