Figure 2: Scatterplot of all judged activities from seven subjects who made inconsistent judgements of percentage of harmless time, i.e., their fitting parameters L50 and/or S were non‑significant