<p>Vierordt’s law is a well-established phenomenon in time perception research, describing systematic biases where long durations tend to be underestimated while short durations tend to be overestimated. The present study investigated whether these central tendency biases extend to temporal memory and examined how emotional context influences such estimation biases. Through three experiments employing pictorial and textual stimuli, we assessed the accuracy of temporal distance and position memories. Consistent with Vierordt’s law, we found robust regression-toward-the-mean biases across all experiments: items near the beginning of a sequence were remembered as having occurred later than they did, while items near the end were remembered as having occurred earlier; short inter-item distances were overestimated and long distances underestimated. Trial-level linear mixed-effects models using participants’ subjective valence ratings as a continuous predictor revealed that emotional valence did not reliably modulate the central tendency bias in Experiments 1a (images, blocked design) or 1b (words, blocked design). In Experiment 2, which employed a mixed-list design and collected trial-level confidence ratings, confidence robustly attenuated the central tendency bias across both temporal memory measures, whereas emotional valence did not produce a robust modulation of this bias. Overall, the central tendency bias in temporal memory proved highly replicable across stimulus types and designs, whereas emotional modulation was minimal. Our findings extend the scope of Vierordt’s law to quasi-retrospective temporal memory and highlight the role of memory confidence in shaping temporal reconstruction.</p>