Motivational Interviewing (MI) is an evidence-based, directive, client-centered therapy designed to develop discrepancy and resolve ambivalence by eliciting and reinforcing client Change Talk. However, the exact link between the process engaged in during MI and outcome is only starting to be uncovered. The present thesis has replicated and expanded on the current knowledge of the relationship between Therapist and Client Behaviours during a MI-based intervention (Motivational Enhancement Therapy; MET) and outcome, and has provided support for the emergent theory of the inner workings of MI. This was achieved by coding 106 audiotaped MET sessions primarily by the methods outlined in the Motivational Interviewing Skill Code Version 2.0. Data was drawn from 28 participants who received 3-4 sessions of MET within the context of a randomised controlled trial for mild-moderate alcohol dependence at the Community Alcohol and Drug Service of Christchurch. Therapist and Client Behaviours were analysed within sessions (categorised into Early, Mid, or End Intervals) and across sessions, and compared with whether the client had drank within national drinking guidelines during the 6-months after MET (Controlled Drinkers). In terms of Client Behaviours during MET it was found that Uncontrolled Drinkers (compared with Controlled Drinkers) uttered a significantly higher frequency of Sustain Talk, lower Ability Language strength (over all MET and during End Intervals), and lower Commitment Language strength (during Session 2 and 4, and change over MET). Giving Information was the only Therapist Behaviour where significant differences were observed over all MET, with a higher frequency given to the Uncontrolled Drinkers. However, during End Intervals within MET Sessions, Controlled Drinkers received a significantly higher frequency of Advise without Permission and a lower frequency of Emphasise Control statements. In most instances MI-Consistent Therapist Behaviours were associated with higher strength of Ability and Commitment Language, and a lower frequency of Sustain Talk. MI-Inconsistent Therapist Behaviour, Direct, was associated with lower Client Language strength. Limitations to these results include small sample, limited ability to make inferences about causality, coder biases, and uneven reliability. However, this exploratory study was unique in investigating the relationship between Therapist Behaviours and the strength of Client Language, and in examining these factors within and across multiple sessions, and has produced a number of potentially valuable findings that warrant further investigation.