Retrospective symptom reports serve as an important source of information for clinical diagnosis, treatment choice and assessment of treatment effects. Previous research has shown that symptom reports, both concurrent and retrospective, are subject to various emotional and cognitive biases, leading to inaccurate (mostly exaggerated) symptom recall. The purpose of this study was to investigate the effect of attentional direction and focus on biased symptom reporting. Healthy participants (N=47) participated in one laboratory and three follow-up sessions. During the laboratory session, two cold pain induction trials were administered via a thermode. In one trial, participants focused attention internally (IF) on either sensory or affective components of the experience; in the other trial they were distracted from pain by a neutral video clip. Pain (pain on average, pain at the peak, and pain at the end of the trial) and affect ratings were collected after each trial, as well as during the follow-up sessions at 2, 4, and 8 weeks after the experiment. Distraction resulted in lower state NA and unpleasantness than IF during the pain episode. Average and peak pain was overall lower in distraction than in IF, but differences changed over time. While average pain remained at the same level over time, peak pain in the affective IF condition decreased in the course of 2 months. Pain at the end did not differ between conditions, but increased over time. Finally, fear of pain interacted with some of the abovementioned effects. The results are discussed in light of previous findings regarding the influence of attentional and memory processes on symptom reports.