Speech accommodation is influenced by various factors, such as interlocutor traits, social identity, context, and opinion [Pardo, 2022 JPhon95]. Although the effects of many of these factors have been extensively researched, the relationship between opinion convergence or divergence and speaker traits is a relatively new area of interest [Ma et al., 2023, HISPCSL]. Another important factor affecting speech is the timepoint within a conversation, as individuals may modify their speech patterns as the conversation progresses. This study aimed to investigate whether the stage of a conversation (initial vs. later) influences the way individuals express opinion convergence or divergence in relation to their interlocutors. Using Praat, DARLA, and statistical methods in R software, we analyzed speech data from an Ellen Fisher Podcast YouTube debate on "Plant vs. Animal Regenerative Farming." Audio clips from the first and last 10 minutes of the debate were transcribed manually into words, automatically converted into phonemes using DARLA, and then manually verified. Each word was further coded for the speaker, interlocutor, whether they expressed convergence or divergence, and the strength of their convergence or divergence on a numerical scale. Preliminary findings revealed that speakers tended to raise their fundamental frequency (F0) when expressing divergence in opinion during deeper stages of the conversation, suggesting a relationship between speech patterns and opinion divergence in different conversation stages. However, we did not observe any significant effect of conversation stage on opinion convergence. Furthermore, no differences were found in terms of vowel quality, as measured by F1 and F2. These preliminary results shed light on the complex interplay between conversation stage and the expression of opinion divergence in speech. Further analysis and exploration of these findings will provide a deeper understanding of the dynamics of speech accommodation in relation to opinion expression within conversations.