Vaccine information and misinformation are spread through social media in ways that may vary by platform. Understanding the role social media plays in shaping vaccine preferences is crucial for policymakers and researchers. Objective This study aims to test whether social media use is associated with changes in vaccine preferences during the COVID-19 pandemic in New Zealand, and whether trust in sources of information has a moderating role. Methods Our data consist of a balanced panel of 257 web-based respondents in New Zealand in August 2020, October-November 2020, and March-April 2021. We use a novel approach with stated choice panel data to study transitions between different vaccine preference groups. We analyze the associations between these transitions and social media use. We classify respondents as resistant (never chose a vaccine), hesitant (chose a vaccine between 1 and 5 times), and provaccine (chose a vaccine 6 out of 6 times) in each wave of data. Results We found a positive or neutral association between social media use and vaccine uptake. Facebook, Twitter (pre-2022), and TikTok users who are provaccine are less likely to become hesitant or resistant. Facebook and Instagram users who are hesitant are more likely to become pro. Some social media platforms may have a more positive association with vaccine uptake preferences for those who do not trust the government. Conclusions The paper contributes to the wider literature, which shows social media can be associated with reinforcing both pro and antivaccination sentiment, and these results depend on where individuals get their information from and their trust in such sources.