As robotics, particularly in agriculture, become more prevalent, understanding the role that different factors play on the trust levels that users have in these robots becomes crucial to facilitate their adoption and integration into the industry. In this paper we present the results of a withinsubjects study that included between-subject factors exploring how prior experience with robotics and different interaction styles with a mobile manipulator robot may affect trust levels in said robot before and after the completion of an agriculturerelated manipulation task. The results show that interacting with the robot helps improve trust levels, particularly for those without prior experience with robotics, who present a higher trust improvement score, and that an interaction style involving physical human-robot interaction (pHRI), more specifically Learning by Demonstration, was favoured versus less direct interaction styles. We found that incorporating Text-to-Speech (TTS) can be a good design choice when trying to improve trust, and that the improvement score for trust before and after interaction with the robot was significantly higher for older age groups, with these participants being more conservative with their reported trust level before the interaction. Overall, these results offer insights into different interaction styles and their effect on trust levels for an agriculture-related manipulation task, and open the door to future work exploring further interaction styles and task variations.