Microblogging sites, like Twitter, have become a common means to share information, participate in topical debate, and express opinions about events or entities. However, the way people search for information in microblogs, and interact with the bearers of this information, is still under-explored, especially when it comes to looking for opinions for a decision-making process. We propose a task-based user study to investigate the search behaviour of users when looking for opinions. We capture users’ clicks and ratings and create a query log dataset. The analysis of this data can provide insights about what elements of a tweet are deemed relevant while searching for opinionated information.