Follow-up exploration of the injected mock catalog that most strongly (and incorrectly) excludes ζ _spike = 0 in Figure 3. Left: the joint posterior on ζ _spike and μ _eff from the selected catalog. In general, we find significant correlations between ζ _spike and the mean μ _eff of the spinning “bulk” population. Mock catalogs that most strongly disfavor ζ _spike = 0 are generally those that have the largest inferred μ _eff. Middle: the individual events comprising the selected catalog. The event with the largest measured χ _eff (and hence the event that most strongly influences μ _eff) is highlighted. Right: the solid blue curve shows the updated ζ _spike posterior after removing the highlighted high-spin event. Compared to the original ζ _spike measurement (dashed black curve), the updated posterior is now more (correctly) consistent with ζ _spike = 0.