Paranormal beliefs (PBs) are common in adults. There are numerous psychological correlates of PBs and associated theories, yet, we do not know whether such correlates reinforce or result from PBs. To understand causality, we developed an experimental design in which participants experience supposedly paranormal events. Thus, we can test an event's impact on PBs and PB-associated correlates. Here, 419 naïve students saw a performer making contact with a confederate's deceased kin. We tested participants' opinions and feelings about this performance, and whether these predicted how participants explain the performance. We assessed participants' PBs and repetition avoidance (PB related cognitive correlate) before and after the performance. Afterwards, participants rated explanations of the event and described their opinions and feelings (open-ended question). Overall, 65% of participants reported having witnessed a genuine paranormal event. The open-ended question revealed distinct opinion and affect groups, with reactions commonly characterized by doubt and mixed feelings. Importantly, paranormal explanations were more likely when participants reported their feelings than when not reported. Beyond these results, we replicated that 1) higher pre-existing PBs were associated with more psychic explanations (confirmation bias), and 2) PBs and repetition avoidance did not change from before to after the performance. Yet, PBs reminiscent of the actual performance (spiritualism) increased. Results showed that young adults easily endorse PBs and paranormal explanations for events, and that their affective reactions matter. Future studies should use participants' subjective experiences to target PBs in causal designs (e.g., adding control conditions).