Detecting Differential Transmissibilities That Affect the Size of Self-Limited Outbreaks
The layout is analogous to Figure 2 except the axes distinguish between transmission of primary and secondary cases. The inset of panel A replicates the results when and are inferred separately (our preferred model), except that the y-axis is now the ratio of to . For panels B and C, the data is shown only for cases where there was a clear record of subsequent secondary infections (as opposed to knowing that four cases lead to ten secondary cases in aggregate). The 95% confidence intervals were found by parametric bootstrap on this more limited data set.