Emerging social media platforms such as Twitter and its Chinese equivalent Weibo have become important in information-sharing and communication. They are also gradually becoming stronger in guiding public opinion. When compared with traditional media, these platforms have salient characteristics, such as highly efficient dissemination of information and interactive commentary, which can contribute to information overload. In earlier research, only the effect of social media on attitude change has been studied, but the specific mechanism of this effect in the context of information overload has not been found. To answer this question, we measured the attitude change of participants after they read Weibo posts about street vendors. A 2 (post-attitude: positive posts vs. negative posts) x 4 (reading time: 35 vs. 25 vs. 15 vs. 5 min) experiment was set up, and the Single Category Implicit Attitude Test was used to measure the implicit attitudes. The interaction effect revealed that in both positive and negative posts, less reading time (i.e., information overload) had a stronger influence. Users were more easily persuaded by posts under high overload. Furthermore, the changes in the attitudes of users were not simply stronger with more information. We found three stages, namely, obedience, resistance, and acceptance, with different mechanisms. Therefore, in the positive information overload condition, the attitudes of individuals eventually change in a positive way. In the negative information overload condition, individuals tend to be biased against the group being reported.